1 @c Copyright (C) 1988,1989,1992,1993,1994,1995,1996,1997,1998,1999,2000,2001,
2 @c 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012
3 @c Free Software Foundation, Inc.
4 @c This is part of the GCC manual.
5 @c For copying conditions, see the file gcc.texi.
8 @chapter Target Description Macros and Functions
9 @cindex machine description macros
10 @cindex target description macros
11 @cindex macros, target description
12 @cindex @file{tm.h} macros
14 In addition to the file @file{@var{machine}.md}, a machine description
15 includes a C header file conventionally given the name
16 @file{@var{machine}.h} and a C source file named @file{@var{machine}.c}.
17 The header file defines numerous macros that convey the information
18 about the target machine that does not fit into the scheme of the
19 @file{.md} file. The file @file{tm.h} should be a link to
20 @file{@var{machine}.h}. The header file @file{config.h} includes
21 @file{tm.h} and most compiler source files include @file{config.h}. The
22 source file defines a variable @code{targetm}, which is a structure
23 containing pointers to functions and data relating to the target
24 machine. @file{@var{machine}.c} should also contain their definitions,
25 if they are not defined elsewhere in GCC, and other functions called
26 through the macros defined in the @file{.h} file.
29 * Target Structure:: The @code{targetm} variable.
30 * Driver:: Controlling how the driver runs the compilation passes.
31 * Run-time Target:: Defining @samp{-m} options like @option{-m68000} and @option{-m68020}.
32 * Per-Function Data:: Defining data structures for per-function information.
33 * Storage Layout:: Defining sizes and alignments of data.
34 * Type Layout:: Defining sizes and properties of basic user data types.
35 * Registers:: Naming and describing the hardware registers.
36 * Register Classes:: Defining the classes of hardware registers.
37 * Old Constraints:: The old way to define machine-specific constraints.
38 * Stack and Calling:: Defining which way the stack grows and by how much.
39 * Varargs:: Defining the varargs macros.
40 * Trampolines:: Code set up at run time to enter a nested function.
41 * Library Calls:: Controlling how library routines are implicitly called.
42 * Addressing Modes:: Defining addressing modes valid for memory operands.
43 * Anchored Addresses:: Defining how @option{-fsection-anchors} should work.
44 * Condition Code:: Defining how insns update the condition code.
45 * Costs:: Defining relative costs of different operations.
46 * Scheduling:: Adjusting the behavior of the instruction scheduler.
47 * Sections:: Dividing storage into text, data, and other sections.
48 * PIC:: Macros for position independent code.
49 * Assembler Format:: Defining how to write insns and pseudo-ops to output.
50 * Debugging Info:: Defining the format of debugging output.
51 * Floating Point:: Handling floating point for cross-compilers.
52 * Mode Switching:: Insertion of mode-switching instructions.
53 * Target Attributes:: Defining target-specific uses of @code{__attribute__}.
54 * Emulated TLS:: Emulated TLS support.
55 * MIPS Coprocessors:: MIPS coprocessor support and how to customize it.
56 * PCH Target:: Validity checking for precompiled headers.
57 * C++ ABI:: Controlling C++ ABI changes.
58 * Named Address Spaces:: Adding support for named address spaces
59 * Misc:: Everything else.
62 @node Target Structure
63 @section The Global @code{targetm} Variable
65 @cindex target functions
67 @deftypevar {struct gcc_target} targetm
68 The target @file{.c} file must define the global @code{targetm} variable
69 which contains pointers to functions and data relating to the target
70 machine. The variable is declared in @file{target.h};
71 @file{target-def.h} defines the macro @code{TARGET_INITIALIZER} which is
72 used to initialize the variable, and macros for the default initializers
73 for elements of the structure. The @file{.c} file should override those
74 macros for which the default definition is inappropriate. For example:
77 #include "target-def.h"
79 /* @r{Initialize the GCC target structure.} */
81 #undef TARGET_COMP_TYPE_ATTRIBUTES
82 #define TARGET_COMP_TYPE_ATTRIBUTES @var{machine}_comp_type_attributes
84 struct gcc_target targetm = TARGET_INITIALIZER;
88 Where a macro should be defined in the @file{.c} file in this manner to
89 form part of the @code{targetm} structure, it is documented below as a
90 ``Target Hook'' with a prototype. Many macros will change in future
91 from being defined in the @file{.h} file to being part of the
92 @code{targetm} structure.
94 Similarly, there is a @code{targetcm} variable for hooks that are
95 specific to front ends for C-family languages, documented as ``C
96 Target Hook''. This is declared in @file{c-family/c-target.h}, the
97 initializer @code{TARGETCM_INITIALIZER} in
98 @file{c-family/c-target-def.h}. If targets initialize @code{targetcm}
99 themselves, they should set @code{target_has_targetcm=yes} in
100 @file{config.gcc}; otherwise a default definition is used.
102 Similarly, there is a @code{targetm_common} variable for hooks that
103 are shared between the compiler driver and the compilers proper,
104 documented as ``Common Target Hook''. This is declared in
105 @file{common/common-target.h}, the initializer
106 @code{TARGETM_COMMON_INITIALIZER} in
107 @file{common/common-target-def.h}. If targets initialize
108 @code{targetm_common} themselves, they should set
109 @code{target_has_targetm_common=yes} in @file{config.gcc}; otherwise a
110 default definition is used.
113 @section Controlling the Compilation Driver, @file{gcc}
115 @cindex controlling the compilation driver
117 @c prevent bad page break with this line
118 You can control the compilation driver.
120 @defmac DRIVER_SELF_SPECS
121 A list of specs for the driver itself. It should be a suitable
122 initializer for an array of strings, with no surrounding braces.
124 The driver applies these specs to its own command line between loading
125 default @file{specs} files (but not command-line specified ones) and
126 choosing the multilib directory or running any subcommands. It
127 applies them in the order given, so each spec can depend on the
128 options added by earlier ones. It is also possible to remove options
129 using @samp{%<@var{option}} in the usual way.
131 This macro can be useful when a port has several interdependent target
132 options. It provides a way of standardizing the command line so
133 that the other specs are easier to write.
135 Do not define this macro if it does not need to do anything.
138 @defmac OPTION_DEFAULT_SPECS
139 A list of specs used to support configure-time default options (i.e.@:
140 @option{--with} options) in the driver. It should be a suitable initializer
141 for an array of structures, each containing two strings, without the
142 outermost pair of surrounding braces.
144 The first item in the pair is the name of the default. This must match
145 the code in @file{config.gcc} for the target. The second item is a spec
146 to apply if a default with this name was specified. The string
147 @samp{%(VALUE)} in the spec will be replaced by the value of the default
148 everywhere it occurs.
150 The driver will apply these specs to its own command line between loading
151 default @file{specs} files and processing @code{DRIVER_SELF_SPECS}, using
152 the same mechanism as @code{DRIVER_SELF_SPECS}.
154 Do not define this macro if it does not need to do anything.
158 A C string constant that tells the GCC driver program options to
159 pass to CPP@. It can also specify how to translate options you
160 give to GCC into options for GCC to pass to the CPP@.
162 Do not define this macro if it does not need to do anything.
165 @defmac CPLUSPLUS_CPP_SPEC
166 This macro is just like @code{CPP_SPEC}, but is used for C++, rather
167 than C@. If you do not define this macro, then the value of
168 @code{CPP_SPEC} (if any) will be used instead.
172 A C string constant that tells the GCC driver program options to
173 pass to @code{cc1}, @code{cc1plus}, @code{f771}, and the other language
175 It can also specify how to translate options you give to GCC into options
176 for GCC to pass to front ends.
178 Do not define this macro if it does not need to do anything.
182 A C string constant that tells the GCC driver program options to
183 pass to @code{cc1plus}. It can also specify how to translate options you
184 give to GCC into options for GCC to pass to the @code{cc1plus}.
186 Do not define this macro if it does not need to do anything.
187 Note that everything defined in CC1_SPEC is already passed to
188 @code{cc1plus} so there is no need to duplicate the contents of
189 CC1_SPEC in CC1PLUS_SPEC@.
193 A C string constant that tells the GCC driver program options to
194 pass to the assembler. It can also specify how to translate options
195 you give to GCC into options for GCC to pass to the assembler.
196 See the file @file{sun3.h} for an example of this.
198 Do not define this macro if it does not need to do anything.
201 @defmac ASM_FINAL_SPEC
202 A C string constant that tells the GCC driver program how to
203 run any programs which cleanup after the normal assembler.
204 Normally, this is not needed. See the file @file{mips.h} for
207 Do not define this macro if it does not need to do anything.
210 @defmac AS_NEEDS_DASH_FOR_PIPED_INPUT
211 Define this macro, with no value, if the driver should give the assembler
212 an argument consisting of a single dash, @option{-}, to instruct it to
213 read from its standard input (which will be a pipe connected to the
214 output of the compiler proper). This argument is given after any
215 @option{-o} option specifying the name of the output file.
217 If you do not define this macro, the assembler is assumed to read its
218 standard input if given no non-option arguments. If your assembler
219 cannot read standard input at all, use a @samp{%@{pipe:%e@}} construct;
220 see @file{mips.h} for instance.
224 A C string constant that tells the GCC driver program options to
225 pass to the linker. It can also specify how to translate options you
226 give to GCC into options for GCC to pass to the linker.
228 Do not define this macro if it does not need to do anything.
232 Another C string constant used much like @code{LINK_SPEC}. The difference
233 between the two is that @code{LIB_SPEC} is used at the end of the
234 command given to the linker.
236 If this macro is not defined, a default is provided that
237 loads the standard C library from the usual place. See @file{gcc.c}.
241 Another C string constant that tells the GCC driver program
242 how and when to place a reference to @file{libgcc.a} into the
243 linker command line. This constant is placed both before and after
244 the value of @code{LIB_SPEC}.
246 If this macro is not defined, the GCC driver provides a default that
247 passes the string @option{-lgcc} to the linker.
250 @defmac REAL_LIBGCC_SPEC
251 By default, if @code{ENABLE_SHARED_LIBGCC} is defined, the
252 @code{LIBGCC_SPEC} is not directly used by the driver program but is
253 instead modified to refer to different versions of @file{libgcc.a}
254 depending on the values of the command line flags @option{-static},
255 @option{-shared}, @option{-static-libgcc}, and @option{-shared-libgcc}. On
256 targets where these modifications are inappropriate, define
257 @code{REAL_LIBGCC_SPEC} instead. @code{REAL_LIBGCC_SPEC} tells the
258 driver how to place a reference to @file{libgcc} on the link command
259 line, but, unlike @code{LIBGCC_SPEC}, it is used unmodified.
262 @defmac USE_LD_AS_NEEDED
263 A macro that controls the modifications to @code{LIBGCC_SPEC}
264 mentioned in @code{REAL_LIBGCC_SPEC}. If nonzero, a spec will be
265 generated that uses --as-needed and the shared libgcc in place of the
266 static exception handler library, when linking without any of
267 @code{-static}, @code{-static-libgcc}, or @code{-shared-libgcc}.
271 If defined, this C string constant is added to @code{LINK_SPEC}.
272 When @code{USE_LD_AS_NEEDED} is zero or undefined, it also affects
273 the modifications to @code{LIBGCC_SPEC} mentioned in
274 @code{REAL_LIBGCC_SPEC}.
277 @defmac STARTFILE_SPEC
278 Another C string constant used much like @code{LINK_SPEC}. The
279 difference between the two is that @code{STARTFILE_SPEC} is used at
280 the very beginning of the command given to the linker.
282 If this macro is not defined, a default is provided that loads the
283 standard C startup file from the usual place. See @file{gcc.c}.
287 Another C string constant used much like @code{LINK_SPEC}. The
288 difference between the two is that @code{ENDFILE_SPEC} is used at
289 the very end of the command given to the linker.
291 Do not define this macro if it does not need to do anything.
294 @defmac THREAD_MODEL_SPEC
295 GCC @code{-v} will print the thread model GCC was configured to use.
296 However, this doesn't work on platforms that are multilibbed on thread
297 models, such as AIX 4.3. On such platforms, define
298 @code{THREAD_MODEL_SPEC} such that it evaluates to a string without
299 blanks that names one of the recognized thread models. @code{%*}, the
300 default value of this macro, will expand to the value of
301 @code{thread_file} set in @file{config.gcc}.
304 @defmac SYSROOT_SUFFIX_SPEC
305 Define this macro to add a suffix to the target sysroot when GCC is
306 configured with a sysroot. This will cause GCC to search for usr/lib,
307 et al, within sysroot+suffix.
310 @defmac SYSROOT_HEADERS_SUFFIX_SPEC
311 Define this macro to add a headers_suffix to the target sysroot when
312 GCC is configured with a sysroot. This will cause GCC to pass the
313 updated sysroot+headers_suffix to CPP, causing it to search for
314 usr/include, et al, within sysroot+headers_suffix.
318 Define this macro to provide additional specifications to put in the
319 @file{specs} file that can be used in various specifications like
322 The definition should be an initializer for an array of structures,
323 containing a string constant, that defines the specification name, and a
324 string constant that provides the specification.
326 Do not define this macro if it does not need to do anything.
328 @code{EXTRA_SPECS} is useful when an architecture contains several
329 related targets, which have various @code{@dots{}_SPECS} which are similar
330 to each other, and the maintainer would like one central place to keep
333 For example, the PowerPC System V.4 targets use @code{EXTRA_SPECS} to
334 define either @code{_CALL_SYSV} when the System V calling sequence is
335 used or @code{_CALL_AIX} when the older AIX-based calling sequence is
338 The @file{config/rs6000/rs6000.h} target file defines:
341 #define EXTRA_SPECS \
342 @{ "cpp_sysv_default", CPP_SYSV_DEFAULT @},
344 #define CPP_SYS_DEFAULT ""
347 The @file{config/rs6000/sysv.h} target file defines:
351 "%@{posix: -D_POSIX_SOURCE @} \
352 %@{mcall-sysv: -D_CALL_SYSV @} \
353 %@{!mcall-sysv: %(cpp_sysv_default) @} \
354 %@{msoft-float: -D_SOFT_FLOAT@} %@{mcpu=403: -D_SOFT_FLOAT@}"
356 #undef CPP_SYSV_DEFAULT
357 #define CPP_SYSV_DEFAULT "-D_CALL_SYSV"
360 while the @file{config/rs6000/eabiaix.h} target file defines
361 @code{CPP_SYSV_DEFAULT} as:
364 #undef CPP_SYSV_DEFAULT
365 #define CPP_SYSV_DEFAULT "-D_CALL_AIX"
369 @defmac LINK_LIBGCC_SPECIAL_1
370 Define this macro if the driver program should find the library
371 @file{libgcc.a}. If you do not define this macro, the driver program will pass
372 the argument @option{-lgcc} to tell the linker to do the search.
375 @defmac LINK_GCC_C_SEQUENCE_SPEC
376 The sequence in which libgcc and libc are specified to the linker.
377 By default this is @code{%G %L %G}.
380 @defmac LINK_COMMAND_SPEC
381 A C string constant giving the complete command line need to execute the
382 linker. When you do this, you will need to update your port each time a
383 change is made to the link command line within @file{gcc.c}. Therefore,
384 define this macro only if you need to completely redefine the command
385 line for invoking the linker and there is no other way to accomplish
386 the effect you need. Overriding this macro may be avoidable by overriding
387 @code{LINK_GCC_C_SEQUENCE_SPEC} instead.
390 @defmac LINK_ELIMINATE_DUPLICATE_LDIRECTORIES
391 A nonzero value causes @command{collect2} to remove duplicate @option{-L@var{directory}} search
392 directories from linking commands. Do not give it a nonzero value if
393 removing duplicate search directories changes the linker's semantics.
396 @hook TARGET_ALWAYS_STRIP_DOTDOT
398 @defmac MULTILIB_DEFAULTS
399 Define this macro as a C expression for the initializer of an array of
400 string to tell the driver program which options are defaults for this
401 target and thus do not need to be handled specially when using
402 @code{MULTILIB_OPTIONS}.
404 Do not define this macro if @code{MULTILIB_OPTIONS} is not defined in
405 the target makefile fragment or if none of the options listed in
406 @code{MULTILIB_OPTIONS} are set by default.
407 @xref{Target Fragment}.
410 @defmac RELATIVE_PREFIX_NOT_LINKDIR
411 Define this macro to tell @command{gcc} that it should only translate
412 a @option{-B} prefix into a @option{-L} linker option if the prefix
413 indicates an absolute file name.
416 @defmac MD_EXEC_PREFIX
417 If defined, this macro is an additional prefix to try after
418 @code{STANDARD_EXEC_PREFIX}. @code{MD_EXEC_PREFIX} is not searched
419 when the compiler is built as a cross
420 compiler. If you define @code{MD_EXEC_PREFIX}, then be sure to add it
421 to the list of directories used to find the assembler in @file{configure.in}.
424 @defmac STANDARD_STARTFILE_PREFIX
425 Define this macro as a C string constant if you wish to override the
426 standard choice of @code{libdir} as the default prefix to
427 try when searching for startup files such as @file{crt0.o}.
428 @code{STANDARD_STARTFILE_PREFIX} is not searched when the compiler
429 is built as a cross compiler.
432 @defmac STANDARD_STARTFILE_PREFIX_1
433 Define this macro as a C string constant if you wish to override the
434 standard choice of @code{/lib} as a prefix to try after the default prefix
435 when searching for startup files such as @file{crt0.o}.
436 @code{STANDARD_STARTFILE_PREFIX_1} is not searched when the compiler
437 is built as a cross compiler.
440 @defmac STANDARD_STARTFILE_PREFIX_2
441 Define this macro as a C string constant if you wish to override the
442 standard choice of @code{/lib} as yet another prefix to try after the
443 default prefix when searching for startup files such as @file{crt0.o}.
444 @code{STANDARD_STARTFILE_PREFIX_2} is not searched when the compiler
445 is built as a cross compiler.
448 @defmac MD_STARTFILE_PREFIX
449 If defined, this macro supplies an additional prefix to try after the
450 standard prefixes. @code{MD_EXEC_PREFIX} is not searched when the
451 compiler is built as a cross compiler.
454 @defmac MD_STARTFILE_PREFIX_1
455 If defined, this macro supplies yet another prefix to try after the
456 standard prefixes. It is not searched when the compiler is built as a
460 @defmac INIT_ENVIRONMENT
461 Define this macro as a C string constant if you wish to set environment
462 variables for programs called by the driver, such as the assembler and
463 loader. The driver passes the value of this macro to @code{putenv} to
464 initialize the necessary environment variables.
467 @defmac LOCAL_INCLUDE_DIR
468 Define this macro as a C string constant if you wish to override the
469 standard choice of @file{/usr/local/include} as the default prefix to
470 try when searching for local header files. @code{LOCAL_INCLUDE_DIR}
471 comes before @code{NATIVE_SYSTEM_HEADER_DIR} (set in
472 @file{config.gcc}, normally @file{/usr/include}) in the search order.
474 Cross compilers do not search either @file{/usr/local/include} or its
478 @defmac NATIVE_SYSTEM_HEADER_COMPONENT
479 The ``component'' corresponding to @code{NATIVE_SYSTEM_HEADER_DIR}.
480 See @code{INCLUDE_DEFAULTS}, below, for the description of components.
481 If you do not define this macro, no component is used.
484 @defmac INCLUDE_DEFAULTS
485 Define this macro if you wish to override the entire default search path
486 for include files. For a native compiler, the default search path
487 usually consists of @code{GCC_INCLUDE_DIR}, @code{LOCAL_INCLUDE_DIR},
488 @code{GPLUSPLUS_INCLUDE_DIR}, and
489 @code{NATIVE_SYSTEM_HEADER_DIR}. In addition, @code{GPLUSPLUS_INCLUDE_DIR}
490 and @code{GCC_INCLUDE_DIR} are defined automatically by @file{Makefile},
491 and specify private search areas for GCC@. The directory
492 @code{GPLUSPLUS_INCLUDE_DIR} is used only for C++ programs.
494 The definition should be an initializer for an array of structures.
495 Each array element should have four elements: the directory name (a
496 string constant), the component name (also a string constant), a flag
497 for C++-only directories,
498 and a flag showing that the includes in the directory don't need to be
499 wrapped in @code{extern @samp{C}} when compiling C++. Mark the end of
500 the array with a null element.
502 The component name denotes what GNU package the include file is part of,
503 if any, in all uppercase letters. For example, it might be @samp{GCC}
504 or @samp{BINUTILS}. If the package is part of a vendor-supplied
505 operating system, code the component name as @samp{0}.
507 For example, here is the definition used for VAX/VMS:
510 #define INCLUDE_DEFAULTS \
512 @{ "GNU_GXX_INCLUDE:", "G++", 1, 1@}, \
513 @{ "GNU_CC_INCLUDE:", "GCC", 0, 0@}, \
514 @{ "SYS$SYSROOT:[SYSLIB.]", 0, 0, 0@}, \
521 Here is the order of prefixes tried for exec files:
525 Any prefixes specified by the user with @option{-B}.
528 The environment variable @code{GCC_EXEC_PREFIX} or, if @code{GCC_EXEC_PREFIX}
529 is not set and the compiler has not been installed in the configure-time
530 @var{prefix}, the location in which the compiler has actually been installed.
533 The directories specified by the environment variable @code{COMPILER_PATH}.
536 The macro @code{STANDARD_EXEC_PREFIX}, if the compiler has been installed
537 in the configured-time @var{prefix}.
540 The location @file{/usr/libexec/gcc/}, but only if this is a native compiler.
543 The location @file{/usr/lib/gcc/}, but only if this is a native compiler.
546 The macro @code{MD_EXEC_PREFIX}, if defined, but only if this is a native
550 Here is the order of prefixes tried for startfiles:
554 Any prefixes specified by the user with @option{-B}.
557 The environment variable @code{GCC_EXEC_PREFIX} or its automatically determined
558 value based on the installed toolchain location.
561 The directories specified by the environment variable @code{LIBRARY_PATH}
562 (or port-specific name; native only, cross compilers do not use this).
565 The macro @code{STANDARD_EXEC_PREFIX}, but only if the toolchain is installed
566 in the configured @var{prefix} or this is a native compiler.
569 The location @file{/usr/lib/gcc/}, but only if this is a native compiler.
572 The macro @code{MD_EXEC_PREFIX}, if defined, but only if this is a native
576 The macro @code{MD_STARTFILE_PREFIX}, if defined, but only if this is a
577 native compiler, or we have a target system root.
580 The macro @code{MD_STARTFILE_PREFIX_1}, if defined, but only if this is a
581 native compiler, or we have a target system root.
584 The macro @code{STANDARD_STARTFILE_PREFIX}, with any sysroot modifications.
585 If this path is relative it will be prefixed by @code{GCC_EXEC_PREFIX} and
586 the machine suffix or @code{STANDARD_EXEC_PREFIX} and the machine suffix.
589 The macro @code{STANDARD_STARTFILE_PREFIX_1}, but only if this is a native
590 compiler, or we have a target system root. The default for this macro is
594 The macro @code{STANDARD_STARTFILE_PREFIX_2}, but only if this is a native
595 compiler, or we have a target system root. The default for this macro is
599 @node Run-time Target
600 @section Run-time Target Specification
601 @cindex run-time target specification
602 @cindex predefined macros
603 @cindex target specifications
605 @c prevent bad page break with this line
606 Here are run-time target specifications.
608 @defmac TARGET_CPU_CPP_BUILTINS ()
609 This function-like macro expands to a block of code that defines
610 built-in preprocessor macros and assertions for the target CPU, using
611 the functions @code{builtin_define}, @code{builtin_define_std} and
612 @code{builtin_assert}. When the front end
613 calls this macro it provides a trailing semicolon, and since it has
614 finished command line option processing your code can use those
617 @code{builtin_assert} takes a string in the form you pass to the
618 command-line option @option{-A}, such as @code{cpu=mips}, and creates
619 the assertion. @code{builtin_define} takes a string in the form
620 accepted by option @option{-D} and unconditionally defines the macro.
622 @code{builtin_define_std} takes a string representing the name of an
623 object-like macro. If it doesn't lie in the user's namespace,
624 @code{builtin_define_std} defines it unconditionally. Otherwise, it
625 defines a version with two leading underscores, and another version
626 with two leading and trailing underscores, and defines the original
627 only if an ISO standard was not requested on the command line. For
628 example, passing @code{unix} defines @code{__unix}, @code{__unix__}
629 and possibly @code{unix}; passing @code{_mips} defines @code{__mips},
630 @code{__mips__} and possibly @code{_mips}, and passing @code{_ABI64}
631 defines only @code{_ABI64}.
633 You can also test for the C dialect being compiled. The variable
634 @code{c_language} is set to one of @code{clk_c}, @code{clk_cplusplus}
635 or @code{clk_objective_c}. Note that if we are preprocessing
636 assembler, this variable will be @code{clk_c} but the function-like
637 macro @code{preprocessing_asm_p()} will return true, so you might want
638 to check for that first. If you need to check for strict ANSI, the
639 variable @code{flag_iso} can be used. The function-like macro
640 @code{preprocessing_trad_p()} can be used to check for traditional
644 @defmac TARGET_OS_CPP_BUILTINS ()
645 Similarly to @code{TARGET_CPU_CPP_BUILTINS} but this macro is optional
646 and is used for the target operating system instead.
649 @defmac TARGET_OBJFMT_CPP_BUILTINS ()
650 Similarly to @code{TARGET_CPU_CPP_BUILTINS} but this macro is optional
651 and is used for the target object format. @file{elfos.h} uses this
652 macro to define @code{__ELF__}, so you probably do not need to define
656 @deftypevar {extern int} target_flags
657 This variable is declared in @file{options.h}, which is included before
658 any target-specific headers.
661 @hook TARGET_DEFAULT_TARGET_FLAGS
662 This variable specifies the initial value of @code{target_flags}.
663 Its default setting is 0.
666 @cindex optional hardware or system features
667 @cindex features, optional, in system conventions
669 @hook TARGET_HANDLE_OPTION
670 This hook is called whenever the user specifies one of the
671 target-specific options described by the @file{.opt} definition files
672 (@pxref{Options}). It has the opportunity to do some option-specific
673 processing and should return true if the option is valid. The default
674 definition does nothing but return true.
676 @var{decoded} specifies the option and its arguments. @var{opts} and
677 @var{opts_set} are the @code{gcc_options} structures to be used for
678 storing option state, and @var{loc} is the location at which the
679 option was passed (@code{UNKNOWN_LOCATION} except for options passed
683 @hook TARGET_HANDLE_C_OPTION
684 This target hook is called whenever the user specifies one of the
685 target-specific C language family options described by the @file{.opt}
686 definition files(@pxref{Options}). It has the opportunity to do some
687 option-specific processing and should return true if the option is
688 valid. The arguments are like for @code{TARGET_HANDLE_OPTION}. The
689 default definition does nothing but return false.
691 In general, you should use @code{TARGET_HANDLE_OPTION} to handle
692 options. However, if processing an option requires routines that are
693 only available in the C (and related language) front ends, then you
694 should use @code{TARGET_HANDLE_C_OPTION} instead.
697 @hook TARGET_OBJC_CONSTRUCT_STRING_OBJECT
699 @hook TARGET_STRING_OBJECT_REF_TYPE_P
701 @hook TARGET_CHECK_STRING_OBJECT_FORMAT_ARG
703 @hook TARGET_OVERRIDE_OPTIONS_AFTER_CHANGE
704 This target function is similar to the hook @code{TARGET_OPTION_OVERRIDE}
705 but is called when the optimize level is changed via an attribute or
706 pragma or when it is reset at the end of the code affected by the
707 attribute or pragma. It is not called at the beginning of compilation
708 when @code{TARGET_OPTION_OVERRIDE} is called so if you want to perform these
709 actions then, you should have @code{TARGET_OPTION_OVERRIDE} call
710 @code{TARGET_OVERRIDE_OPTIONS_AFTER_CHANGE}.
713 @defmac C_COMMON_OVERRIDE_OPTIONS
714 This is similar to the @code{TARGET_OPTION_OVERRIDE} hook
715 but is only used in the C
716 language frontends (C, Objective-C, C++, Objective-C++) and so can be
717 used to alter option flag variables which only exist in those
721 @hook TARGET_OPTION_OPTIMIZATION_TABLE
722 Some machines may desire to change what optimizations are performed for
723 various optimization levels. This variable, if defined, describes
724 options to enable at particular sets of optimization levels. These
725 options are processed once
726 just after the optimization level is determined and before the remainder
727 of the command options have been parsed, so may be overridden by other
728 options passed explicitly.
730 This processing is run once at program startup and when the optimization
731 options are changed via @code{#pragma GCC optimize} or by using the
732 @code{optimize} attribute.
735 @hook TARGET_OPTION_INIT_STRUCT
737 @hook TARGET_OPTION_DEFAULT_PARAMS
739 @defmac SWITCHABLE_TARGET
740 Some targets need to switch between substantially different subtargets
741 during compilation. For example, the MIPS target has one subtarget for
742 the traditional MIPS architecture and another for MIPS16. Source code
743 can switch between these two subarchitectures using the @code{mips16}
744 and @code{nomips16} attributes.
746 Such subtargets can differ in things like the set of available
747 registers, the set of available instructions, the costs of various
748 operations, and so on. GCC caches a lot of this type of information
749 in global variables, and recomputing them for each subtarget takes a
750 significant amount of time. The compiler therefore provides a facility
751 for maintaining several versions of the global variables and quickly
752 switching between them; see @file{target-globals.h} for details.
754 Define this macro to 1 if your target needs this facility. The default
758 @node Per-Function Data
759 @section Defining data structures for per-function information.
760 @cindex per-function data
761 @cindex data structures
763 If the target needs to store information on a per-function basis, GCC
764 provides a macro and a couple of variables to allow this. Note, just
765 using statics to store the information is a bad idea, since GCC supports
766 nested functions, so you can be halfway through encoding one function
767 when another one comes along.
769 GCC defines a data structure called @code{struct function} which
770 contains all of the data specific to an individual function. This
771 structure contains a field called @code{machine} whose type is
772 @code{struct machine_function *}, which can be used by targets to point
773 to their own specific data.
775 If a target needs per-function specific data it should define the type
776 @code{struct machine_function} and also the macro @code{INIT_EXPANDERS}.
777 This macro should be used to initialize the function pointer
778 @code{init_machine_status}. This pointer is explained below.
780 One typical use of per-function, target specific data is to create an
781 RTX to hold the register containing the function's return address. This
782 RTX can then be used to implement the @code{__builtin_return_address}
783 function, for level 0.
785 Note---earlier implementations of GCC used a single data area to hold
786 all of the per-function information. Thus when processing of a nested
787 function began the old per-function data had to be pushed onto a
788 stack, and when the processing was finished, it had to be popped off the
789 stack. GCC used to provide function pointers called
790 @code{save_machine_status} and @code{restore_machine_status} to handle
791 the saving and restoring of the target specific information. Since the
792 single data area approach is no longer used, these pointers are no
795 @defmac INIT_EXPANDERS
796 Macro called to initialize any target specific information. This macro
797 is called once per function, before generation of any RTL has begun.
798 The intention of this macro is to allow the initialization of the
799 function pointer @code{init_machine_status}.
802 @deftypevar {void (*)(struct function *)} init_machine_status
803 If this function pointer is non-@code{NULL} it will be called once per
804 function, before function compilation starts, in order to allow the
805 target to perform any target specific initialization of the
806 @code{struct function} structure. It is intended that this would be
807 used to initialize the @code{machine} of that structure.
809 @code{struct machine_function} structures are expected to be freed by GC@.
810 Generally, any memory that they reference must be allocated by using
811 GC allocation, including the structure itself.
815 @section Storage Layout
816 @cindex storage layout
818 Note that the definitions of the macros in this table which are sizes or
819 alignments measured in bits do not need to be constant. They can be C
820 expressions that refer to static variables, such as the @code{target_flags}.
821 @xref{Run-time Target}.
823 @defmac BITS_BIG_ENDIAN
824 Define this macro to have the value 1 if the most significant bit in a
825 byte has the lowest number; otherwise define it to have the value zero.
826 This means that bit-field instructions count from the most significant
827 bit. If the machine has no bit-field instructions, then this must still
828 be defined, but it doesn't matter which value it is defined to. This
829 macro need not be a constant.
831 This macro does not affect the way structure fields are packed into
832 bytes or words; that is controlled by @code{BYTES_BIG_ENDIAN}.
835 @defmac BYTES_BIG_ENDIAN
836 Define this macro to have the value 1 if the most significant byte in a
837 word has the lowest number. This macro need not be a constant.
840 @defmac WORDS_BIG_ENDIAN
841 Define this macro to have the value 1 if, in a multiword object, the
842 most significant word has the lowest number. This applies to both
843 memory locations and registers; see @code{REG_WORDS_BIG_ENDIAN} if the
844 order of words in memory is not the same as the order in registers. This
845 macro need not be a constant.
848 @defmac REG_WORDS_BIG_ENDIAN
849 On some machines, the order of words in a multiword object differs between
850 registers in memory. In such a situation, define this macro to describe
851 the order of words in a register. The macro @code{WORDS_BIG_ENDIAN} controls
852 the order of words in memory.
855 @defmac FLOAT_WORDS_BIG_ENDIAN
856 Define this macro to have the value 1 if @code{DFmode}, @code{XFmode} or
857 @code{TFmode} floating point numbers are stored in memory with the word
858 containing the sign bit at the lowest address; otherwise define it to
859 have the value 0. This macro need not be a constant.
861 You need not define this macro if the ordering is the same as for
865 @defmac BITS_PER_UNIT
866 Define this macro to be the number of bits in an addressable storage
867 unit (byte). If you do not define this macro the default is 8.
870 @defmac BITS_PER_WORD
871 Number of bits in a word. If you do not define this macro, the default
872 is @code{BITS_PER_UNIT * UNITS_PER_WORD}.
875 @defmac MAX_BITS_PER_WORD
876 Maximum number of bits in a word. If this is undefined, the default is
877 @code{BITS_PER_WORD}. Otherwise, it is the constant value that is the
878 largest value that @code{BITS_PER_WORD} can have at run-time.
881 @defmac UNITS_PER_WORD
882 Number of storage units in a word; normally the size of a general-purpose
883 register, a power of two from 1 or 8.
886 @defmac MIN_UNITS_PER_WORD
887 Minimum number of units in a word. If this is undefined, the default is
888 @code{UNITS_PER_WORD}. Otherwise, it is the constant value that is the
889 smallest value that @code{UNITS_PER_WORD} can have at run-time.
893 Width of a pointer, in bits. You must specify a value no wider than the
894 width of @code{Pmode}. If it is not equal to the width of @code{Pmode},
895 you must define @code{POINTERS_EXTEND_UNSIGNED}. If you do not specify
896 a value the default is @code{BITS_PER_WORD}.
899 @defmac POINTERS_EXTEND_UNSIGNED
900 A C expression that determines how pointers should be extended from
901 @code{ptr_mode} to either @code{Pmode} or @code{word_mode}. It is
902 greater than zero if pointers should be zero-extended, zero if they
903 should be sign-extended, and negative if some other sort of conversion
904 is needed. In the last case, the extension is done by the target's
905 @code{ptr_extend} instruction.
907 You need not define this macro if the @code{ptr_mode}, @code{Pmode}
908 and @code{word_mode} are all the same width.
911 @defmac PROMOTE_MODE (@var{m}, @var{unsignedp}, @var{type})
912 A macro to update @var{m} and @var{unsignedp} when an object whose type
913 is @var{type} and which has the specified mode and signedness is to be
914 stored in a register. This macro is only called when @var{type} is a
917 On most RISC machines, which only have operations that operate on a full
918 register, define this macro to set @var{m} to @code{word_mode} if
919 @var{m} is an integer mode narrower than @code{BITS_PER_WORD}. In most
920 cases, only integer modes should be widened because wider-precision
921 floating-point operations are usually more expensive than their narrower
924 For most machines, the macro definition does not change @var{unsignedp}.
925 However, some machines, have instructions that preferentially handle
926 either signed or unsigned quantities of certain modes. For example, on
927 the DEC Alpha, 32-bit loads from memory and 32-bit add instructions
928 sign-extend the result to 64 bits. On such machines, set
929 @var{unsignedp} according to which kind of extension is more efficient.
931 Do not define this macro if it would never modify @var{m}.
934 @hook TARGET_PROMOTE_FUNCTION_MODE
935 Like @code{PROMOTE_MODE}, but it is applied to outgoing function arguments or
936 function return values. The target hook should return the new mode
937 and possibly change @code{*@var{punsignedp}} if the promotion should
938 change signedness. This function is called only for scalar @emph{or
941 @var{for_return} allows to distinguish the promotion of arguments and
942 return values. If it is @code{1}, a return value is being promoted and
943 @code{TARGET_FUNCTION_VALUE} must perform the same promotions done here.
944 If it is @code{2}, the returned mode should be that of the register in
945 which an incoming parameter is copied, or the outgoing result is computed;
946 then the hook should return the same mode as @code{promote_mode}, though
947 the signedness may be different.
949 @var{type} can be NULL when promoting function arguments of libcalls.
951 The default is to not promote arguments and return values. You can
952 also define the hook to @code{default_promote_function_mode_always_promote}
953 if you would like to apply the same rules given by @code{PROMOTE_MODE}.
956 @defmac PARM_BOUNDARY
957 Normal alignment required for function parameters on the stack, in
958 bits. All stack parameters receive at least this much alignment
959 regardless of data type. On most machines, this is the same as the
963 @defmac STACK_BOUNDARY
964 Define this macro to the minimum alignment enforced by hardware for the
965 stack pointer on this machine. The definition is a C expression for the
966 desired alignment (measured in bits). This value is used as a default
967 if @code{PREFERRED_STACK_BOUNDARY} is not defined. On most machines,
968 this should be the same as @code{PARM_BOUNDARY}.
971 @defmac PREFERRED_STACK_BOUNDARY
972 Define this macro if you wish to preserve a certain alignment for the
973 stack pointer, greater than what the hardware enforces. The definition
974 is a C expression for the desired alignment (measured in bits). This
975 macro must evaluate to a value equal to or larger than
976 @code{STACK_BOUNDARY}.
979 @defmac INCOMING_STACK_BOUNDARY
980 Define this macro if the incoming stack boundary may be different
981 from @code{PREFERRED_STACK_BOUNDARY}. This macro must evaluate
982 to a value equal to or larger than @code{STACK_BOUNDARY}.
985 @defmac FUNCTION_BOUNDARY
986 Alignment required for a function entry point, in bits.
989 @defmac BIGGEST_ALIGNMENT
990 Biggest alignment that any data type can require on this machine, in
991 bits. Note that this is not the biggest alignment that is supported,
992 just the biggest alignment that, when violated, may cause a fault.
995 @defmac MALLOC_ABI_ALIGNMENT
996 Alignment, in bits, a C conformant malloc implementation has to
997 provide. If not defined, the default value is @code{BITS_PER_WORD}.
1000 @defmac ATTRIBUTE_ALIGNED_VALUE
1001 Alignment used by the @code{__attribute__ ((aligned))} construct. If
1002 not defined, the default value is @code{BIGGEST_ALIGNMENT}.
1005 @defmac MINIMUM_ATOMIC_ALIGNMENT
1006 If defined, the smallest alignment, in bits, that can be given to an
1007 object that can be referenced in one operation, without disturbing any
1008 nearby object. Normally, this is @code{BITS_PER_UNIT}, but may be larger
1009 on machines that don't have byte or half-word store operations.
1012 @defmac BIGGEST_FIELD_ALIGNMENT
1013 Biggest alignment that any structure or union field can require on this
1014 machine, in bits. If defined, this overrides @code{BIGGEST_ALIGNMENT} for
1015 structure and union fields only, unless the field alignment has been set
1016 by the @code{__attribute__ ((aligned (@var{n})))} construct.
1019 @defmac ADJUST_FIELD_ALIGN (@var{field}, @var{computed})
1020 An expression for the alignment of a structure field @var{field} if the
1021 alignment computed in the usual way (including applying of
1022 @code{BIGGEST_ALIGNMENT} and @code{BIGGEST_FIELD_ALIGNMENT} to the
1023 alignment) is @var{computed}. It overrides alignment only if the
1024 field alignment has not been set by the
1025 @code{__attribute__ ((aligned (@var{n})))} construct.
1028 @defmac MAX_STACK_ALIGNMENT
1029 Biggest stack alignment guaranteed by the backend. Use this macro
1030 to specify the maximum alignment of a variable on stack.
1032 If not defined, the default value is @code{STACK_BOUNDARY}.
1034 @c FIXME: The default should be @code{PREFERRED_STACK_BOUNDARY}.
1035 @c But the fix for PR 32893 indicates that we can only guarantee
1036 @c maximum stack alignment on stack up to @code{STACK_BOUNDARY}, not
1037 @c @code{PREFERRED_STACK_BOUNDARY}, if stack alignment isn't supported.
1040 @defmac MAX_OFILE_ALIGNMENT
1041 Biggest alignment supported by the object file format of this machine.
1042 Use this macro to limit the alignment which can be specified using the
1043 @code{__attribute__ ((aligned (@var{n})))} construct. If not defined,
1044 the default value is @code{BIGGEST_ALIGNMENT}.
1046 On systems that use ELF, the default (in @file{config/elfos.h}) is
1047 the largest supported 32-bit ELF section alignment representable on
1048 a 32-bit host e.g. @samp{(((unsigned HOST_WIDEST_INT) 1 << 28) * 8)}.
1049 On 32-bit ELF the largest supported section alignment in bits is
1050 @samp{(0x80000000 * 8)}, but this is not representable on 32-bit hosts.
1053 @defmac DATA_ALIGNMENT (@var{type}, @var{basic-align})
1054 If defined, a C expression to compute the alignment for a variable in
1055 the static store. @var{type} is the data type, and @var{basic-align} is
1056 the alignment that the object would ordinarily have. The value of this
1057 macro is used instead of that alignment to align the object.
1059 If this macro is not defined, then @var{basic-align} is used.
1062 One use of this macro is to increase alignment of medium-size data to
1063 make it all fit in fewer cache lines. Another is to cause character
1064 arrays to be word-aligned so that @code{strcpy} calls that copy
1065 constants to character arrays can be done inline.
1068 @defmac CONSTANT_ALIGNMENT (@var{constant}, @var{basic-align})
1069 If defined, a C expression to compute the alignment given to a constant
1070 that is being placed in memory. @var{constant} is the constant and
1071 @var{basic-align} is the alignment that the object would ordinarily
1072 have. The value of this macro is used instead of that alignment to
1075 If this macro is not defined, then @var{basic-align} is used.
1077 The typical use of this macro is to increase alignment for string
1078 constants to be word aligned so that @code{strcpy} calls that copy
1079 constants can be done inline.
1082 @defmac LOCAL_ALIGNMENT (@var{type}, @var{basic-align})
1083 If defined, a C expression to compute the alignment for a variable in
1084 the local store. @var{type} is the data type, and @var{basic-align} is
1085 the alignment that the object would ordinarily have. The value of this
1086 macro is used instead of that alignment to align the object.
1088 If this macro is not defined, then @var{basic-align} is used.
1090 One use of this macro is to increase alignment of medium-size data to
1091 make it all fit in fewer cache lines.
1093 If the value of this macro has a type, it should be an unsigned type.
1096 @hook TARGET_VECTOR_ALIGNMENT
1098 @defmac STACK_SLOT_ALIGNMENT (@var{type}, @var{mode}, @var{basic-align})
1099 If defined, a C expression to compute the alignment for stack slot.
1100 @var{type} is the data type, @var{mode} is the widest mode available,
1101 and @var{basic-align} is the alignment that the slot would ordinarily
1102 have. The value of this macro is used instead of that alignment to
1105 If this macro is not defined, then @var{basic-align} is used when
1106 @var{type} is @code{NULL}. Otherwise, @code{LOCAL_ALIGNMENT} will
1109 This macro is to set alignment of stack slot to the maximum alignment
1110 of all possible modes which the slot may have.
1112 If the value of this macro has a type, it should be an unsigned type.
1115 @defmac LOCAL_DECL_ALIGNMENT (@var{decl})
1116 If defined, a C expression to compute the alignment for a local
1117 variable @var{decl}.
1119 If this macro is not defined, then
1120 @code{LOCAL_ALIGNMENT (TREE_TYPE (@var{decl}), DECL_ALIGN (@var{decl}))}
1123 One use of this macro is to increase alignment of medium-size data to
1124 make it all fit in fewer cache lines.
1126 If the value of this macro has a type, it should be an unsigned type.
1129 @defmac MINIMUM_ALIGNMENT (@var{exp}, @var{mode}, @var{align})
1130 If defined, a C expression to compute the minimum required alignment
1131 for dynamic stack realignment purposes for @var{exp} (a type or decl),
1132 @var{mode}, assuming normal alignment @var{align}.
1134 If this macro is not defined, then @var{align} will be used.
1137 @defmac EMPTY_FIELD_BOUNDARY
1138 Alignment in bits to be given to a structure bit-field that follows an
1139 empty field such as @code{int : 0;}.
1141 If @code{PCC_BITFIELD_TYPE_MATTERS} is true, it overrides this macro.
1144 @defmac STRUCTURE_SIZE_BOUNDARY
1145 Number of bits which any structure or union's size must be a multiple of.
1146 Each structure or union's size is rounded up to a multiple of this.
1148 If you do not define this macro, the default is the same as
1149 @code{BITS_PER_UNIT}.
1152 @defmac STRICT_ALIGNMENT
1153 Define this macro to be the value 1 if instructions will fail to work
1154 if given data not on the nominal alignment. If instructions will merely
1155 go slower in that case, define this macro as 0.
1158 @defmac PCC_BITFIELD_TYPE_MATTERS
1159 Define this if you wish to imitate the way many other C compilers handle
1160 alignment of bit-fields and the structures that contain them.
1162 The behavior is that the type written for a named bit-field (@code{int},
1163 @code{short}, or other integer type) imposes an alignment for the entire
1164 structure, as if the structure really did contain an ordinary field of
1165 that type. In addition, the bit-field is placed within the structure so
1166 that it would fit within such a field, not crossing a boundary for it.
1168 Thus, on most machines, a named bit-field whose type is written as
1169 @code{int} would not cross a four-byte boundary, and would force
1170 four-byte alignment for the whole structure. (The alignment used may
1171 not be four bytes; it is controlled by the other alignment parameters.)
1173 An unnamed bit-field will not affect the alignment of the containing
1176 If the macro is defined, its definition should be a C expression;
1177 a nonzero value for the expression enables this behavior.
1179 Note that if this macro is not defined, or its value is zero, some
1180 bit-fields may cross more than one alignment boundary. The compiler can
1181 support such references if there are @samp{insv}, @samp{extv}, and
1182 @samp{extzv} insns that can directly reference memory.
1184 The other known way of making bit-fields work is to define
1185 @code{STRUCTURE_SIZE_BOUNDARY} as large as @code{BIGGEST_ALIGNMENT}.
1186 Then every structure can be accessed with fullwords.
1188 Unless the machine has bit-field instructions or you define
1189 @code{STRUCTURE_SIZE_BOUNDARY} that way, you must define
1190 @code{PCC_BITFIELD_TYPE_MATTERS} to have a nonzero value.
1192 If your aim is to make GCC use the same conventions for laying out
1193 bit-fields as are used by another compiler, here is how to investigate
1194 what the other compiler does. Compile and run this program:
1213 printf ("Size of foo1 is %d\n",
1214 sizeof (struct foo1));
1215 printf ("Size of foo2 is %d\n",
1216 sizeof (struct foo2));
1221 If this prints 2 and 5, then the compiler's behavior is what you would
1222 get from @code{PCC_BITFIELD_TYPE_MATTERS}.
1225 @defmac BITFIELD_NBYTES_LIMITED
1226 Like @code{PCC_BITFIELD_TYPE_MATTERS} except that its effect is limited
1227 to aligning a bit-field within the structure.
1230 @hook TARGET_ALIGN_ANON_BITFIELD
1231 When @code{PCC_BITFIELD_TYPE_MATTERS} is true this hook will determine
1232 whether unnamed bitfields affect the alignment of the containing
1233 structure. The hook should return true if the structure should inherit
1234 the alignment requirements of an unnamed bitfield's type.
1237 @hook TARGET_NARROW_VOLATILE_BITFIELD
1238 This target hook should return @code{true} if accesses to volatile bitfields
1239 should use the narrowest mode possible. It should return @code{false} if
1240 these accesses should use the bitfield container type.
1242 The default is @code{!TARGET_STRICT_ALIGN}.
1245 @defmac MEMBER_TYPE_FORCES_BLK (@var{field}, @var{mode})
1246 Return 1 if a structure or array containing @var{field} should be accessed using
1249 If @var{field} is the only field in the structure, @var{mode} is its
1250 mode, otherwise @var{mode} is VOIDmode. @var{mode} is provided in the
1251 case where structures of one field would require the structure's mode to
1252 retain the field's mode.
1254 Normally, this is not needed.
1257 @defmac ROUND_TYPE_ALIGN (@var{type}, @var{computed}, @var{specified})
1258 Define this macro as an expression for the alignment of a type (given
1259 by @var{type} as a tree node) if the alignment computed in the usual
1260 way is @var{computed} and the alignment explicitly specified was
1263 The default is to use @var{specified} if it is larger; otherwise, use
1264 the smaller of @var{computed} and @code{BIGGEST_ALIGNMENT}
1267 @defmac MAX_FIXED_MODE_SIZE
1268 An integer expression for the size in bits of the largest integer
1269 machine mode that should actually be used. All integer machine modes of
1270 this size or smaller can be used for structures and unions with the
1271 appropriate sizes. If this macro is undefined, @code{GET_MODE_BITSIZE
1272 (DImode)} is assumed.
1275 @defmac STACK_SAVEAREA_MODE (@var{save_level})
1276 If defined, an expression of type @code{enum machine_mode} that
1277 specifies the mode of the save area operand of a
1278 @code{save_stack_@var{level}} named pattern (@pxref{Standard Names}).
1279 @var{save_level} is one of @code{SAVE_BLOCK}, @code{SAVE_FUNCTION}, or
1280 @code{SAVE_NONLOCAL} and selects which of the three named patterns is
1281 having its mode specified.
1283 You need not define this macro if it always returns @code{Pmode}. You
1284 would most commonly define this macro if the
1285 @code{save_stack_@var{level}} patterns need to support both a 32- and a
1289 @defmac STACK_SIZE_MODE
1290 If defined, an expression of type @code{enum machine_mode} that
1291 specifies the mode of the size increment operand of an
1292 @code{allocate_stack} named pattern (@pxref{Standard Names}).
1294 You need not define this macro if it always returns @code{word_mode}.
1295 You would most commonly define this macro if the @code{allocate_stack}
1296 pattern needs to support both a 32- and a 64-bit mode.
1299 @hook TARGET_LIBGCC_CMP_RETURN_MODE
1300 This target hook should return the mode to be used for the return value
1301 of compare instructions expanded to libgcc calls. If not defined
1302 @code{word_mode} is returned which is the right choice for a majority of
1306 @hook TARGET_LIBGCC_SHIFT_COUNT_MODE
1307 This target hook should return the mode to be used for the shift count operand
1308 of shift instructions expanded to libgcc calls. If not defined
1309 @code{word_mode} is returned which is the right choice for a majority of
1313 @hook TARGET_UNWIND_WORD_MODE
1314 Return machine mode to be used for @code{_Unwind_Word} type.
1315 The default is to use @code{word_mode}.
1318 @defmac ROUND_TOWARDS_ZERO
1319 If defined, this macro should be true if the prevailing rounding
1320 mode is towards zero.
1322 Defining this macro only affects the way @file{libgcc.a} emulates
1323 floating-point arithmetic.
1325 Not defining this macro is equivalent to returning zero.
1328 @defmac LARGEST_EXPONENT_IS_NORMAL (@var{size})
1329 This macro should return true if floats with @var{size}
1330 bits do not have a NaN or infinity representation, but use the largest
1331 exponent for normal numbers instead.
1333 Defining this macro only affects the way @file{libgcc.a} emulates
1334 floating-point arithmetic.
1336 The default definition of this macro returns false for all sizes.
1339 @hook TARGET_MS_BITFIELD_LAYOUT_P
1340 This target hook returns @code{true} if bit-fields in the given
1341 @var{record_type} are to be laid out following the rules of Microsoft
1342 Visual C/C++, namely: (i) a bit-field won't share the same storage
1343 unit with the previous bit-field if their underlying types have
1344 different sizes, and the bit-field will be aligned to the highest
1345 alignment of the underlying types of itself and of the previous
1346 bit-field; (ii) a zero-sized bit-field will affect the alignment of
1347 the whole enclosing structure, even if it is unnamed; except that
1348 (iii) a zero-sized bit-field will be disregarded unless it follows
1349 another bit-field of nonzero size. If this hook returns @code{true},
1350 other macros that control bit-field layout are ignored.
1352 When a bit-field is inserted into a packed record, the whole size
1353 of the underlying type is used by one or more same-size adjacent
1354 bit-fields (that is, if its long:3, 32 bits is used in the record,
1355 and any additional adjacent long bit-fields are packed into the same
1356 chunk of 32 bits. However, if the size changes, a new field of that
1357 size is allocated). In an unpacked record, this is the same as using
1358 alignment, but not equivalent when packing.
1360 If both MS bit-fields and @samp{__attribute__((packed))} are used,
1361 the latter will take precedence. If @samp{__attribute__((packed))} is
1362 used on a single field when MS bit-fields are in use, it will take
1363 precedence for that field, but the alignment of the rest of the structure
1364 may affect its placement.
1367 @hook TARGET_DECIMAL_FLOAT_SUPPORTED_P
1368 Returns true if the target supports decimal floating point.
1371 @hook TARGET_FIXED_POINT_SUPPORTED_P
1372 Returns true if the target supports fixed-point arithmetic.
1375 @hook TARGET_EXPAND_TO_RTL_HOOK
1376 This hook is called just before expansion into rtl, allowing the target
1377 to perform additional initializations or analysis before the expansion.
1378 For example, the rs6000 port uses it to allocate a scratch stack slot
1379 for use in copying SDmode values between memory and floating point
1380 registers whenever the function being expanded has any SDmode
1384 @hook TARGET_INSTANTIATE_DECLS
1385 This hook allows the backend to perform additional instantiations on rtl
1386 that are not actually in any insns yet, but will be later.
1389 @hook TARGET_MANGLE_TYPE
1390 If your target defines any fundamental types, or any types your target
1391 uses should be mangled differently from the default, define this hook
1392 to return the appropriate encoding for these types as part of a C++
1393 mangled name. The @var{type} argument is the tree structure representing
1394 the type to be mangled. The hook may be applied to trees which are
1395 not target-specific fundamental types; it should return @code{NULL}
1396 for all such types, as well as arguments it does not recognize. If the
1397 return value is not @code{NULL}, it must point to a statically-allocated
1400 Target-specific fundamental types might be new fundamental types or
1401 qualified versions of ordinary fundamental types. Encode new
1402 fundamental types as @samp{@w{u @var{n} @var{name}}}, where @var{name}
1403 is the name used for the type in source code, and @var{n} is the
1404 length of @var{name} in decimal. Encode qualified versions of
1405 ordinary types as @samp{@w{U @var{n} @var{name} @var{code}}}, where
1406 @var{name} is the name used for the type qualifier in source code,
1407 @var{n} is the length of @var{name} as above, and @var{code} is the
1408 code used to represent the unqualified version of this type. (See
1409 @code{write_builtin_type} in @file{cp/mangle.c} for the list of
1410 codes.) In both cases the spaces are for clarity; do not include any
1411 spaces in your string.
1413 This hook is applied to types prior to typedef resolution. If the mangled
1414 name for a particular type depends only on that type's main variant, you
1415 can perform typedef resolution yourself using @code{TYPE_MAIN_VARIANT}
1418 The default version of this hook always returns @code{NULL}, which is
1419 appropriate for a target that does not define any new fundamental
1424 @section Layout of Source Language Data Types
1426 These macros define the sizes and other characteristics of the standard
1427 basic data types used in programs being compiled. Unlike the macros in
1428 the previous section, these apply to specific features of C and related
1429 languages, rather than to fundamental aspects of storage layout.
1431 @defmac INT_TYPE_SIZE
1432 A C expression for the size in bits of the type @code{int} on the
1433 target machine. If you don't define this, the default is one word.
1436 @defmac SHORT_TYPE_SIZE
1437 A C expression for the size in bits of the type @code{short} on the
1438 target machine. If you don't define this, the default is half a word.
1439 (If this would be less than one storage unit, it is rounded up to one
1443 @defmac LONG_TYPE_SIZE
1444 A C expression for the size in bits of the type @code{long} on the
1445 target machine. If you don't define this, the default is one word.
1448 @defmac ADA_LONG_TYPE_SIZE
1449 On some machines, the size used for the Ada equivalent of the type
1450 @code{long} by a native Ada compiler differs from that used by C@. In
1451 that situation, define this macro to be a C expression to be used for
1452 the size of that type. If you don't define this, the default is the
1453 value of @code{LONG_TYPE_SIZE}.
1456 @defmac LONG_LONG_TYPE_SIZE
1457 A C expression for the size in bits of the type @code{long long} on the
1458 target machine. If you don't define this, the default is two
1459 words. If you want to support GNU Ada on your machine, the value of this
1460 macro must be at least 64.
1463 @defmac CHAR_TYPE_SIZE
1464 A C expression for the size in bits of the type @code{char} on the
1465 target machine. If you don't define this, the default is
1466 @code{BITS_PER_UNIT}.
1469 @defmac BOOL_TYPE_SIZE
1470 A C expression for the size in bits of the C++ type @code{bool} and
1471 C99 type @code{_Bool} on the target machine. If you don't define
1472 this, and you probably shouldn't, the default is @code{CHAR_TYPE_SIZE}.
1475 @defmac FLOAT_TYPE_SIZE
1476 A C expression for the size in bits of the type @code{float} on the
1477 target machine. If you don't define this, the default is one word.
1480 @defmac DOUBLE_TYPE_SIZE
1481 A C expression for the size in bits of the type @code{double} on the
1482 target machine. If you don't define this, the default is two
1486 @defmac LONG_DOUBLE_TYPE_SIZE
1487 A C expression for the size in bits of the type @code{long double} on
1488 the target machine. If you don't define this, the default is two
1492 @defmac SHORT_FRACT_TYPE_SIZE
1493 A C expression for the size in bits of the type @code{short _Fract} on
1494 the target machine. If you don't define this, the default is
1495 @code{BITS_PER_UNIT}.
1498 @defmac FRACT_TYPE_SIZE
1499 A C expression for the size in bits of the type @code{_Fract} on
1500 the target machine. If you don't define this, the default is
1501 @code{BITS_PER_UNIT * 2}.
1504 @defmac LONG_FRACT_TYPE_SIZE
1505 A C expression for the size in bits of the type @code{long _Fract} on
1506 the target machine. If you don't define this, the default is
1507 @code{BITS_PER_UNIT * 4}.
1510 @defmac LONG_LONG_FRACT_TYPE_SIZE
1511 A C expression for the size in bits of the type @code{long long _Fract} on
1512 the target machine. If you don't define this, the default is
1513 @code{BITS_PER_UNIT * 8}.
1516 @defmac SHORT_ACCUM_TYPE_SIZE
1517 A C expression for the size in bits of the type @code{short _Accum} on
1518 the target machine. If you don't define this, the default is
1519 @code{BITS_PER_UNIT * 2}.
1522 @defmac ACCUM_TYPE_SIZE
1523 A C expression for the size in bits of the type @code{_Accum} on
1524 the target machine. If you don't define this, the default is
1525 @code{BITS_PER_UNIT * 4}.
1528 @defmac LONG_ACCUM_TYPE_SIZE
1529 A C expression for the size in bits of the type @code{long _Accum} on
1530 the target machine. If you don't define this, the default is
1531 @code{BITS_PER_UNIT * 8}.
1534 @defmac LONG_LONG_ACCUM_TYPE_SIZE
1535 A C expression for the size in bits of the type @code{long long _Accum} on
1536 the target machine. If you don't define this, the default is
1537 @code{BITS_PER_UNIT * 16}.
1540 @defmac LIBGCC2_LONG_DOUBLE_TYPE_SIZE
1541 Define this macro if @code{LONG_DOUBLE_TYPE_SIZE} is not constant or
1542 if you want routines in @file{libgcc2.a} for a size other than
1543 @code{LONG_DOUBLE_TYPE_SIZE}. If you don't define this, the
1544 default is @code{LONG_DOUBLE_TYPE_SIZE}.
1547 @defmac LIBGCC2_HAS_DF_MODE
1548 Define this macro if neither @code{DOUBLE_TYPE_SIZE} nor
1549 @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is
1550 @code{DFmode} but you want @code{DFmode} routines in @file{libgcc2.a}
1551 anyway. If you don't define this and either @code{DOUBLE_TYPE_SIZE}
1552 or @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is 64 then the default is 1,
1556 @defmac LIBGCC2_HAS_XF_MODE
1557 Define this macro if @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is not
1558 @code{XFmode} but you want @code{XFmode} routines in @file{libgcc2.a}
1559 anyway. If you don't define this and @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE}
1560 is 80 then the default is 1, otherwise it is 0.
1563 @defmac LIBGCC2_HAS_TF_MODE
1564 Define this macro if @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is not
1565 @code{TFmode} but you want @code{TFmode} routines in @file{libgcc2.a}
1566 anyway. If you don't define this and @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE}
1567 is 128 then the default is 1, otherwise it is 0.
1570 @defmac LIBGCC2_GNU_PREFIX
1571 This macro corresponds to the @code{TARGET_LIBFUNC_GNU_PREFIX} target
1572 hook and should be defined if that hook is overriden to be true. It
1573 causes function names in libgcc to be changed to use a @code{__gnu_}
1574 prefix for their name rather than the default @code{__}. A port which
1575 uses this macro should also arrange to use @file{t-gnu-prefix} in
1576 the libgcc @file{config.host}.
1583 Define these macros to be the size in bits of the mantissa of
1584 @code{SFmode}, @code{DFmode}, @code{XFmode} and @code{TFmode} values,
1585 if the defaults in @file{libgcc2.h} are inappropriate. By default,
1586 @code{FLT_MANT_DIG} is used for @code{SF_SIZE}, @code{LDBL_MANT_DIG}
1587 for @code{XF_SIZE} and @code{TF_SIZE}, and @code{DBL_MANT_DIG} or
1588 @code{LDBL_MANT_DIG} for @code{DF_SIZE} according to whether
1589 @code{DOUBLE_TYPE_SIZE} or
1590 @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is 64.
1593 @defmac TARGET_FLT_EVAL_METHOD
1594 A C expression for the value for @code{FLT_EVAL_METHOD} in @file{float.h},
1595 assuming, if applicable, that the floating-point control word is in its
1596 default state. If you do not define this macro the value of
1597 @code{FLT_EVAL_METHOD} will be zero.
1600 @defmac WIDEST_HARDWARE_FP_SIZE
1601 A C expression for the size in bits of the widest floating-point format
1602 supported by the hardware. If you define this macro, you must specify a
1603 value less than or equal to the value of @code{LONG_DOUBLE_TYPE_SIZE}.
1604 If you do not define this macro, the value of @code{LONG_DOUBLE_TYPE_SIZE}
1608 @defmac DEFAULT_SIGNED_CHAR
1609 An expression whose value is 1 or 0, according to whether the type
1610 @code{char} should be signed or unsigned by default. The user can
1611 always override this default with the options @option{-fsigned-char}
1612 and @option{-funsigned-char}.
1615 @hook TARGET_DEFAULT_SHORT_ENUMS
1616 This target hook should return true if the compiler should give an
1617 @code{enum} type only as many bytes as it takes to represent the range
1618 of possible values of that type. It should return false if all
1619 @code{enum} types should be allocated like @code{int}.
1621 The default is to return false.
1625 A C expression for a string describing the name of the data type to use
1626 for size values. The typedef name @code{size_t} is defined using the
1627 contents of the string.
1629 The string can contain more than one keyword. If so, separate them with
1630 spaces, and write first any length keyword, then @code{unsigned} if
1631 appropriate, and finally @code{int}. The string must exactly match one
1632 of the data type names defined in the function
1633 @code{init_decl_processing} in the file @file{c-decl.c}. You may not
1634 omit @code{int} or change the order---that would cause the compiler to
1637 If you don't define this macro, the default is @code{"long unsigned
1641 @defmac PTRDIFF_TYPE
1642 A C expression for a string describing the name of the data type to use
1643 for the result of subtracting two pointers. The typedef name
1644 @code{ptrdiff_t} is defined using the contents of the string. See
1645 @code{SIZE_TYPE} above for more information.
1647 If you don't define this macro, the default is @code{"long int"}.
1651 A C expression for a string describing the name of the data type to use
1652 for wide characters. The typedef name @code{wchar_t} is defined using
1653 the contents of the string. See @code{SIZE_TYPE} above for more
1656 If you don't define this macro, the default is @code{"int"}.
1659 @defmac WCHAR_TYPE_SIZE
1660 A C expression for the size in bits of the data type for wide
1661 characters. This is used in @code{cpp}, which cannot make use of
1666 A C expression for a string describing the name of the data type to
1667 use for wide characters passed to @code{printf} and returned from
1668 @code{getwc}. The typedef name @code{wint_t} is defined using the
1669 contents of the string. See @code{SIZE_TYPE} above for more
1672 If you don't define this macro, the default is @code{"unsigned int"}.
1676 A C expression for a string describing the name of the data type that
1677 can represent any value of any standard or extended signed integer type.
1678 The typedef name @code{intmax_t} is defined using the contents of the
1679 string. See @code{SIZE_TYPE} above for more information.
1681 If you don't define this macro, the default is the first of
1682 @code{"int"}, @code{"long int"}, or @code{"long long int"} that has as
1683 much precision as @code{long long int}.
1686 @defmac UINTMAX_TYPE
1687 A C expression for a string describing the name of the data type that
1688 can represent any value of any standard or extended unsigned integer
1689 type. The typedef name @code{uintmax_t} is defined using the contents
1690 of the string. See @code{SIZE_TYPE} above for more information.
1692 If you don't define this macro, the default is the first of
1693 @code{"unsigned int"}, @code{"long unsigned int"}, or @code{"long long
1694 unsigned int"} that has as much precision as @code{long long unsigned
1698 @defmac SIG_ATOMIC_TYPE
1704 @defmacx UINT16_TYPE
1705 @defmacx UINT32_TYPE
1706 @defmacx UINT64_TYPE
1707 @defmacx INT_LEAST8_TYPE
1708 @defmacx INT_LEAST16_TYPE
1709 @defmacx INT_LEAST32_TYPE
1710 @defmacx INT_LEAST64_TYPE
1711 @defmacx UINT_LEAST8_TYPE
1712 @defmacx UINT_LEAST16_TYPE
1713 @defmacx UINT_LEAST32_TYPE
1714 @defmacx UINT_LEAST64_TYPE
1715 @defmacx INT_FAST8_TYPE
1716 @defmacx INT_FAST16_TYPE
1717 @defmacx INT_FAST32_TYPE
1718 @defmacx INT_FAST64_TYPE
1719 @defmacx UINT_FAST8_TYPE
1720 @defmacx UINT_FAST16_TYPE
1721 @defmacx UINT_FAST32_TYPE
1722 @defmacx UINT_FAST64_TYPE
1723 @defmacx INTPTR_TYPE
1724 @defmacx UINTPTR_TYPE
1725 C expressions for the standard types @code{sig_atomic_t},
1726 @code{int8_t}, @code{int16_t}, @code{int32_t}, @code{int64_t},
1727 @code{uint8_t}, @code{uint16_t}, @code{uint32_t}, @code{uint64_t},
1728 @code{int_least8_t}, @code{int_least16_t}, @code{int_least32_t},
1729 @code{int_least64_t}, @code{uint_least8_t}, @code{uint_least16_t},
1730 @code{uint_least32_t}, @code{uint_least64_t}, @code{int_fast8_t},
1731 @code{int_fast16_t}, @code{int_fast32_t}, @code{int_fast64_t},
1732 @code{uint_fast8_t}, @code{uint_fast16_t}, @code{uint_fast32_t},
1733 @code{uint_fast64_t}, @code{intptr_t}, and @code{uintptr_t}. See
1734 @code{SIZE_TYPE} above for more information.
1736 If any of these macros evaluates to a null pointer, the corresponding
1737 type is not supported; if GCC is configured to provide
1738 @code{<stdint.h>} in such a case, the header provided may not conform
1739 to C99, depending on the type in question. The defaults for all of
1740 these macros are null pointers.
1743 @defmac TARGET_PTRMEMFUNC_VBIT_LOCATION
1744 The C++ compiler represents a pointer-to-member-function with a struct
1751 ptrdiff_t vtable_index;
1758 The C++ compiler must use one bit to indicate whether the function that
1759 will be called through a pointer-to-member-function is virtual.
1760 Normally, we assume that the low-order bit of a function pointer must
1761 always be zero. Then, by ensuring that the vtable_index is odd, we can
1762 distinguish which variant of the union is in use. But, on some
1763 platforms function pointers can be odd, and so this doesn't work. In
1764 that case, we use the low-order bit of the @code{delta} field, and shift
1765 the remainder of the @code{delta} field to the left.
1767 GCC will automatically make the right selection about where to store
1768 this bit using the @code{FUNCTION_BOUNDARY} setting for your platform.
1769 However, some platforms such as ARM/Thumb have @code{FUNCTION_BOUNDARY}
1770 set such that functions always start at even addresses, but the lowest
1771 bit of pointers to functions indicate whether the function at that
1772 address is in ARM or Thumb mode. If this is the case of your
1773 architecture, you should define this macro to
1774 @code{ptrmemfunc_vbit_in_delta}.
1776 In general, you should not have to define this macro. On architectures
1777 in which function addresses are always even, according to
1778 @code{FUNCTION_BOUNDARY}, GCC will automatically define this macro to
1779 @code{ptrmemfunc_vbit_in_pfn}.
1782 @defmac TARGET_VTABLE_USES_DESCRIPTORS
1783 Normally, the C++ compiler uses function pointers in vtables. This
1784 macro allows the target to change to use ``function descriptors''
1785 instead. Function descriptors are found on targets for whom a
1786 function pointer is actually a small data structure. Normally the
1787 data structure consists of the actual code address plus a data
1788 pointer to which the function's data is relative.
1790 If vtables are used, the value of this macro should be the number
1791 of words that the function descriptor occupies.
1794 @defmac TARGET_VTABLE_ENTRY_ALIGN
1795 By default, the vtable entries are void pointers, the so the alignment
1796 is the same as pointer alignment. The value of this macro specifies
1797 the alignment of the vtable entry in bits. It should be defined only
1798 when special alignment is necessary. */
1801 @defmac TARGET_VTABLE_DATA_ENTRY_DISTANCE
1802 There are a few non-descriptor entries in the vtable at offsets below
1803 zero. If these entries must be padded (say, to preserve the alignment
1804 specified by @code{TARGET_VTABLE_ENTRY_ALIGN}), set this to the number
1805 of words in each data entry.
1809 @section Register Usage
1810 @cindex register usage
1812 This section explains how to describe what registers the target machine
1813 has, and how (in general) they can be used.
1815 The description of which registers a specific instruction can use is
1816 done with register classes; see @ref{Register Classes}. For information
1817 on using registers to access a stack frame, see @ref{Frame Registers}.
1818 For passing values in registers, see @ref{Register Arguments}.
1819 For returning values in registers, see @ref{Scalar Return}.
1822 * Register Basics:: Number and kinds of registers.
1823 * Allocation Order:: Order in which registers are allocated.
1824 * Values in Registers:: What kinds of values each reg can hold.
1825 * Leaf Functions:: Renumbering registers for leaf functions.
1826 * Stack Registers:: Handling a register stack such as 80387.
1829 @node Register Basics
1830 @subsection Basic Characteristics of Registers
1832 @c prevent bad page break with this line
1833 Registers have various characteristics.
1835 @defmac FIRST_PSEUDO_REGISTER
1836 Number of hardware registers known to the compiler. They receive
1837 numbers 0 through @code{FIRST_PSEUDO_REGISTER-1}; thus, the first
1838 pseudo register's number really is assigned the number
1839 @code{FIRST_PSEUDO_REGISTER}.
1842 @defmac FIXED_REGISTERS
1843 @cindex fixed register
1844 An initializer that says which registers are used for fixed purposes
1845 all throughout the compiled code and are therefore not available for
1846 general allocation. These would include the stack pointer, the frame
1847 pointer (except on machines where that can be used as a general
1848 register when no frame pointer is needed), the program counter on
1849 machines where that is considered one of the addressable registers,
1850 and any other numbered register with a standard use.
1852 This information is expressed as a sequence of numbers, separated by
1853 commas and surrounded by braces. The @var{n}th number is 1 if
1854 register @var{n} is fixed, 0 otherwise.
1856 The table initialized from this macro, and the table initialized by
1857 the following one, may be overridden at run time either automatically,
1858 by the actions of the macro @code{CONDITIONAL_REGISTER_USAGE}, or by
1859 the user with the command options @option{-ffixed-@var{reg}},
1860 @option{-fcall-used-@var{reg}} and @option{-fcall-saved-@var{reg}}.
1863 @defmac CALL_USED_REGISTERS
1864 @cindex call-used register
1865 @cindex call-clobbered register
1866 @cindex call-saved register
1867 Like @code{FIXED_REGISTERS} but has 1 for each register that is
1868 clobbered (in general) by function calls as well as for fixed
1869 registers. This macro therefore identifies the registers that are not
1870 available for general allocation of values that must live across
1873 If a register has 0 in @code{CALL_USED_REGISTERS}, the compiler
1874 automatically saves it on function entry and restores it on function
1875 exit, if the register is used within the function.
1878 @defmac CALL_REALLY_USED_REGISTERS
1879 @cindex call-used register
1880 @cindex call-clobbered register
1881 @cindex call-saved register
1882 Like @code{CALL_USED_REGISTERS} except this macro doesn't require
1883 that the entire set of @code{FIXED_REGISTERS} be included.
1884 (@code{CALL_USED_REGISTERS} must be a superset of @code{FIXED_REGISTERS}).
1885 This macro is optional. If not specified, it defaults to the value
1886 of @code{CALL_USED_REGISTERS}.
1889 @defmac HARD_REGNO_CALL_PART_CLOBBERED (@var{regno}, @var{mode})
1890 @cindex call-used register
1891 @cindex call-clobbered register
1892 @cindex call-saved register
1893 A C expression that is nonzero if it is not permissible to store a
1894 value of mode @var{mode} in hard register number @var{regno} across a
1895 call without some part of it being clobbered. For most machines this
1896 macro need not be defined. It is only required for machines that do not
1897 preserve the entire contents of a register across a call.
1901 @findex call_used_regs
1904 @findex reg_class_contents
1905 @hook TARGET_CONDITIONAL_REGISTER_USAGE
1906 This hook may conditionally modify five variables
1907 @code{fixed_regs}, @code{call_used_regs}, @code{global_regs},
1908 @code{reg_names}, and @code{reg_class_contents}, to take into account
1909 any dependence of these register sets on target flags. The first three
1910 of these are of type @code{char []} (interpreted as Boolean vectors).
1911 @code{global_regs} is a @code{const char *[]}, and
1912 @code{reg_class_contents} is a @code{HARD_REG_SET}. Before the macro is
1913 called, @code{fixed_regs}, @code{call_used_regs},
1914 @code{reg_class_contents}, and @code{reg_names} have been initialized
1915 from @code{FIXED_REGISTERS}, @code{CALL_USED_REGISTERS},
1916 @code{REG_CLASS_CONTENTS}, and @code{REGISTER_NAMES}, respectively.
1917 @code{global_regs} has been cleared, and any @option{-ffixed-@var{reg}},
1918 @option{-fcall-used-@var{reg}} and @option{-fcall-saved-@var{reg}}
1919 command options have been applied.
1921 @cindex disabling certain registers
1922 @cindex controlling register usage
1923 If the usage of an entire class of registers depends on the target
1924 flags, you may indicate this to GCC by using this macro to modify
1925 @code{fixed_regs} and @code{call_used_regs} to 1 for each of the
1926 registers in the classes which should not be used by GCC@. Also define
1927 the macro @code{REG_CLASS_FROM_LETTER} / @code{REG_CLASS_FROM_CONSTRAINT}
1928 to return @code{NO_REGS} if it
1929 is called with a letter for a class that shouldn't be used.
1931 (However, if this class is not included in @code{GENERAL_REGS} and all
1932 of the insn patterns whose constraints permit this class are
1933 controlled by target switches, then GCC will automatically avoid using
1934 these registers when the target switches are opposed to them.)
1937 @defmac INCOMING_REGNO (@var{out})
1938 Define this macro if the target machine has register windows. This C
1939 expression returns the register number as seen by the called function
1940 corresponding to the register number @var{out} as seen by the calling
1941 function. Return @var{out} if register number @var{out} is not an
1945 @defmac OUTGOING_REGNO (@var{in})
1946 Define this macro if the target machine has register windows. This C
1947 expression returns the register number as seen by the calling function
1948 corresponding to the register number @var{in} as seen by the called
1949 function. Return @var{in} if register number @var{in} is not an inbound
1953 @defmac LOCAL_REGNO (@var{regno})
1954 Define this macro if the target machine has register windows. This C
1955 expression returns true if the register is call-saved but is in the
1956 register window. Unlike most call-saved registers, such registers
1957 need not be explicitly restored on function exit or during non-local
1962 If the program counter has a register number, define this as that
1963 register number. Otherwise, do not define it.
1966 @node Allocation Order
1967 @subsection Order of Allocation of Registers
1968 @cindex order of register allocation
1969 @cindex register allocation order
1971 @c prevent bad page break with this line
1972 Registers are allocated in order.
1974 @defmac REG_ALLOC_ORDER
1975 If defined, an initializer for a vector of integers, containing the
1976 numbers of hard registers in the order in which GCC should prefer
1977 to use them (from most preferred to least).
1979 If this macro is not defined, registers are used lowest numbered first
1980 (all else being equal).
1982 One use of this macro is on machines where the highest numbered
1983 registers must always be saved and the save-multiple-registers
1984 instruction supports only sequences of consecutive registers. On such
1985 machines, define @code{REG_ALLOC_ORDER} to be an initializer that lists
1986 the highest numbered allocable register first.
1989 @defmac ADJUST_REG_ALLOC_ORDER
1990 A C statement (sans semicolon) to choose the order in which to allocate
1991 hard registers for pseudo-registers local to a basic block.
1993 Store the desired register order in the array @code{reg_alloc_order}.
1994 Element 0 should be the register to allocate first; element 1, the next
1995 register; and so on.
1997 The macro body should not assume anything about the contents of
1998 @code{reg_alloc_order} before execution of the macro.
2000 On most machines, it is not necessary to define this macro.
2003 @defmac HONOR_REG_ALLOC_ORDER
2004 Normally, IRA tries to estimate the costs for saving a register in the
2005 prologue and restoring it in the epilogue. This discourages it from
2006 using call-saved registers. If a machine wants to ensure that IRA
2007 allocates registers in the order given by REG_ALLOC_ORDER even if some
2008 call-saved registers appear earlier than call-used ones, this macro
2012 @defmac IRA_HARD_REGNO_ADD_COST_MULTIPLIER (@var{regno})
2013 In some case register allocation order is not enough for the
2014 Integrated Register Allocator (@acronym{IRA}) to generate a good code.
2015 If this macro is defined, it should return a floating point value
2016 based on @var{regno}. The cost of using @var{regno} for a pseudo will
2017 be increased by approximately the pseudo's usage frequency times the
2018 value returned by this macro. Not defining this macro is equivalent
2019 to having it always return @code{0.0}.
2021 On most machines, it is not necessary to define this macro.
2024 @node Values in Registers
2025 @subsection How Values Fit in Registers
2027 This section discusses the macros that describe which kinds of values
2028 (specifically, which machine modes) each register can hold, and how many
2029 consecutive registers are needed for a given mode.
2031 @defmac HARD_REGNO_NREGS (@var{regno}, @var{mode})
2032 A C expression for the number of consecutive hard registers, starting
2033 at register number @var{regno}, required to hold a value of mode
2034 @var{mode}. This macro must never return zero, even if a register
2035 cannot hold the requested mode - indicate that with HARD_REGNO_MODE_OK
2036 and/or CANNOT_CHANGE_MODE_CLASS instead.
2038 On a machine where all registers are exactly one word, a suitable
2039 definition of this macro is
2042 #define HARD_REGNO_NREGS(REGNO, MODE) \
2043 ((GET_MODE_SIZE (MODE) + UNITS_PER_WORD - 1) \
2048 @defmac HARD_REGNO_NREGS_HAS_PADDING (@var{regno}, @var{mode})
2049 A C expression that is nonzero if a value of mode @var{mode}, stored
2050 in memory, ends with padding that causes it to take up more space than
2051 in registers starting at register number @var{regno} (as determined by
2052 multiplying GCC's notion of the size of the register when containing
2053 this mode by the number of registers returned by
2054 @code{HARD_REGNO_NREGS}). By default this is zero.
2056 For example, if a floating-point value is stored in three 32-bit
2057 registers but takes up 128 bits in memory, then this would be
2060 This macros only needs to be defined if there are cases where
2061 @code{subreg_get_info}
2062 would otherwise wrongly determine that a @code{subreg} can be
2063 represented by an offset to the register number, when in fact such a
2064 @code{subreg} would contain some of the padding not stored in
2065 registers and so not be representable.
2068 @defmac HARD_REGNO_NREGS_WITH_PADDING (@var{regno}, @var{mode})
2069 For values of @var{regno} and @var{mode} for which
2070 @code{HARD_REGNO_NREGS_HAS_PADDING} returns nonzero, a C expression
2071 returning the greater number of registers required to hold the value
2072 including any padding. In the example above, the value would be four.
2075 @defmac REGMODE_NATURAL_SIZE (@var{mode})
2076 Define this macro if the natural size of registers that hold values
2077 of mode @var{mode} is not the word size. It is a C expression that
2078 should give the natural size in bytes for the specified mode. It is
2079 used by the register allocator to try to optimize its results. This
2080 happens for example on SPARC 64-bit where the natural size of
2081 floating-point registers is still 32-bit.
2084 @defmac HARD_REGNO_MODE_OK (@var{regno}, @var{mode})
2085 A C expression that is nonzero if it is permissible to store a value
2086 of mode @var{mode} in hard register number @var{regno} (or in several
2087 registers starting with that one). For a machine where all registers
2088 are equivalent, a suitable definition is
2091 #define HARD_REGNO_MODE_OK(REGNO, MODE) 1
2094 You need not include code to check for the numbers of fixed registers,
2095 because the allocation mechanism considers them to be always occupied.
2097 @cindex register pairs
2098 On some machines, double-precision values must be kept in even/odd
2099 register pairs. You can implement that by defining this macro to reject
2100 odd register numbers for such modes.
2102 The minimum requirement for a mode to be OK in a register is that the
2103 @samp{mov@var{mode}} instruction pattern support moves between the
2104 register and other hard register in the same class and that moving a
2105 value into the register and back out not alter it.
2107 Since the same instruction used to move @code{word_mode} will work for
2108 all narrower integer modes, it is not necessary on any machine for
2109 @code{HARD_REGNO_MODE_OK} to distinguish between these modes, provided
2110 you define patterns @samp{movhi}, etc., to take advantage of this. This
2111 is useful because of the interaction between @code{HARD_REGNO_MODE_OK}
2112 and @code{MODES_TIEABLE_P}; it is very desirable for all integer modes
2115 Many machines have special registers for floating point arithmetic.
2116 Often people assume that floating point machine modes are allowed only
2117 in floating point registers. This is not true. Any registers that
2118 can hold integers can safely @emph{hold} a floating point machine
2119 mode, whether or not floating arithmetic can be done on it in those
2120 registers. Integer move instructions can be used to move the values.
2122 On some machines, though, the converse is true: fixed-point machine
2123 modes may not go in floating registers. This is true if the floating
2124 registers normalize any value stored in them, because storing a
2125 non-floating value there would garble it. In this case,
2126 @code{HARD_REGNO_MODE_OK} should reject fixed-point machine modes in
2127 floating registers. But if the floating registers do not automatically
2128 normalize, if you can store any bit pattern in one and retrieve it
2129 unchanged without a trap, then any machine mode may go in a floating
2130 register, so you can define this macro to say so.
2132 The primary significance of special floating registers is rather that
2133 they are the registers acceptable in floating point arithmetic
2134 instructions. However, this is of no concern to
2135 @code{HARD_REGNO_MODE_OK}. You handle it by writing the proper
2136 constraints for those instructions.
2138 On some machines, the floating registers are especially slow to access,
2139 so that it is better to store a value in a stack frame than in such a
2140 register if floating point arithmetic is not being done. As long as the
2141 floating registers are not in class @code{GENERAL_REGS}, they will not
2142 be used unless some pattern's constraint asks for one.
2145 @defmac HARD_REGNO_RENAME_OK (@var{from}, @var{to})
2146 A C expression that is nonzero if it is OK to rename a hard register
2147 @var{from} to another hard register @var{to}.
2149 One common use of this macro is to prevent renaming of a register to
2150 another register that is not saved by a prologue in an interrupt
2153 The default is always nonzero.
2156 @defmac MODES_TIEABLE_P (@var{mode1}, @var{mode2})
2157 A C expression that is nonzero if a value of mode
2158 @var{mode1} is accessible in mode @var{mode2} without copying.
2160 If @code{HARD_REGNO_MODE_OK (@var{r}, @var{mode1})} and
2161 @code{HARD_REGNO_MODE_OK (@var{r}, @var{mode2})} are always the same for
2162 any @var{r}, then @code{MODES_TIEABLE_P (@var{mode1}, @var{mode2})}
2163 should be nonzero. If they differ for any @var{r}, you should define
2164 this macro to return zero unless some other mechanism ensures the
2165 accessibility of the value in a narrower mode.
2167 You should define this macro to return nonzero in as many cases as
2168 possible since doing so will allow GCC to perform better register
2172 @hook TARGET_HARD_REGNO_SCRATCH_OK
2173 This target hook should return @code{true} if it is OK to use a hard register
2174 @var{regno} as scratch reg in peephole2.
2176 One common use of this macro is to prevent using of a register that
2177 is not saved by a prologue in an interrupt handler.
2179 The default version of this hook always returns @code{true}.
2182 @defmac AVOID_CCMODE_COPIES
2183 Define this macro if the compiler should avoid copies to/from @code{CCmode}
2184 registers. You should only define this macro if support for copying to/from
2185 @code{CCmode} is incomplete.
2188 @node Leaf Functions
2189 @subsection Handling Leaf Functions
2191 @cindex leaf functions
2192 @cindex functions, leaf
2193 On some machines, a leaf function (i.e., one which makes no calls) can run
2194 more efficiently if it does not make its own register window. Often this
2195 means it is required to receive its arguments in the registers where they
2196 are passed by the caller, instead of the registers where they would
2199 The special treatment for leaf functions generally applies only when
2200 other conditions are met; for example, often they may use only those
2201 registers for its own variables and temporaries. We use the term ``leaf
2202 function'' to mean a function that is suitable for this special
2203 handling, so that functions with no calls are not necessarily ``leaf
2206 GCC assigns register numbers before it knows whether the function is
2207 suitable for leaf function treatment. So it needs to renumber the
2208 registers in order to output a leaf function. The following macros
2211 @defmac LEAF_REGISTERS
2212 Name of a char vector, indexed by hard register number, which
2213 contains 1 for a register that is allowable in a candidate for leaf
2216 If leaf function treatment involves renumbering the registers, then the
2217 registers marked here should be the ones before renumbering---those that
2218 GCC would ordinarily allocate. The registers which will actually be
2219 used in the assembler code, after renumbering, should not be marked with 1
2222 Define this macro only if the target machine offers a way to optimize
2223 the treatment of leaf functions.
2226 @defmac LEAF_REG_REMAP (@var{regno})
2227 A C expression whose value is the register number to which @var{regno}
2228 should be renumbered, when a function is treated as a leaf function.
2230 If @var{regno} is a register number which should not appear in a leaf
2231 function before renumbering, then the expression should yield @minus{}1, which
2232 will cause the compiler to abort.
2234 Define this macro only if the target machine offers a way to optimize the
2235 treatment of leaf functions, and registers need to be renumbered to do
2239 @findex current_function_is_leaf
2240 @findex current_function_uses_only_leaf_regs
2241 @code{TARGET_ASM_FUNCTION_PROLOGUE} and
2242 @code{TARGET_ASM_FUNCTION_EPILOGUE} must usually treat leaf functions
2243 specially. They can test the C variable @code{current_function_is_leaf}
2244 which is nonzero for leaf functions. @code{current_function_is_leaf} is
2245 set prior to local register allocation and is valid for the remaining
2246 compiler passes. They can also test the C variable
2247 @code{current_function_uses_only_leaf_regs} which is nonzero for leaf
2248 functions which only use leaf registers.
2249 @code{current_function_uses_only_leaf_regs} is valid after all passes
2250 that modify the instructions have been run and is only useful if
2251 @code{LEAF_REGISTERS} is defined.
2252 @c changed this to fix overfull. ALSO: why the "it" at the beginning
2253 @c of the next paragraph?! --mew 2feb93
2255 @node Stack Registers
2256 @subsection Registers That Form a Stack
2258 There are special features to handle computers where some of the
2259 ``registers'' form a stack. Stack registers are normally written by
2260 pushing onto the stack, and are numbered relative to the top of the
2263 Currently, GCC can only handle one group of stack-like registers, and
2264 they must be consecutively numbered. Furthermore, the existing
2265 support for stack-like registers is specific to the 80387 floating
2266 point coprocessor. If you have a new architecture that uses
2267 stack-like registers, you will need to do substantial work on
2268 @file{reg-stack.c} and write your machine description to cooperate
2269 with it, as well as defining these macros.
2272 Define this if the machine has any stack-like registers.
2275 @defmac STACK_REG_COVER_CLASS
2276 This is a cover class containing the stack registers. Define this if
2277 the machine has any stack-like registers.
2280 @defmac FIRST_STACK_REG
2281 The number of the first stack-like register. This one is the top
2285 @defmac LAST_STACK_REG
2286 The number of the last stack-like register. This one is the bottom of
2290 @node Register Classes
2291 @section Register Classes
2292 @cindex register class definitions
2293 @cindex class definitions, register
2295 On many machines, the numbered registers are not all equivalent.
2296 For example, certain registers may not be allowed for indexed addressing;
2297 certain registers may not be allowed in some instructions. These machine
2298 restrictions are described to the compiler using @dfn{register classes}.
2300 You define a number of register classes, giving each one a name and saying
2301 which of the registers belong to it. Then you can specify register classes
2302 that are allowed as operands to particular instruction patterns.
2306 In general, each register will belong to several classes. In fact, one
2307 class must be named @code{ALL_REGS} and contain all the registers. Another
2308 class must be named @code{NO_REGS} and contain no registers. Often the
2309 union of two classes will be another class; however, this is not required.
2311 @findex GENERAL_REGS
2312 One of the classes must be named @code{GENERAL_REGS}. There is nothing
2313 terribly special about the name, but the operand constraint letters
2314 @samp{r} and @samp{g} specify this class. If @code{GENERAL_REGS} is
2315 the same as @code{ALL_REGS}, just define it as a macro which expands
2318 Order the classes so that if class @var{x} is contained in class @var{y}
2319 then @var{x} has a lower class number than @var{y}.
2321 The way classes other than @code{GENERAL_REGS} are specified in operand
2322 constraints is through machine-dependent operand constraint letters.
2323 You can define such letters to correspond to various classes, then use
2324 them in operand constraints.
2326 You must define the narrowest register classes for allocatable
2327 registers, so that each class either has no subclasses, or that for
2328 some mode, the move cost between registers within the class is
2329 cheaper than moving a register in the class to or from memory
2332 You should define a class for the union of two classes whenever some
2333 instruction allows both classes. For example, if an instruction allows
2334 either a floating point (coprocessor) register or a general register for a
2335 certain operand, you should define a class @code{FLOAT_OR_GENERAL_REGS}
2336 which includes both of them. Otherwise you will get suboptimal code,
2337 or even internal compiler errors when reload cannot find a register in the
2338 class computed via @code{reg_class_subunion}.
2340 You must also specify certain redundant information about the register
2341 classes: for each class, which classes contain it and which ones are
2342 contained in it; for each pair of classes, the largest class contained
2345 When a value occupying several consecutive registers is expected in a
2346 certain class, all the registers used must belong to that class.
2347 Therefore, register classes cannot be used to enforce a requirement for
2348 a register pair to start with an even-numbered register. The way to
2349 specify this requirement is with @code{HARD_REGNO_MODE_OK}.
2351 Register classes used for input-operands of bitwise-and or shift
2352 instructions have a special requirement: each such class must have, for
2353 each fixed-point machine mode, a subclass whose registers can transfer that
2354 mode to or from memory. For example, on some machines, the operations for
2355 single-byte values (@code{QImode}) are limited to certain registers. When
2356 this is so, each register class that is used in a bitwise-and or shift
2357 instruction must have a subclass consisting of registers from which
2358 single-byte values can be loaded or stored. This is so that
2359 @code{PREFERRED_RELOAD_CLASS} can always have a possible value to return.
2361 @deftp {Data type} {enum reg_class}
2362 An enumerated type that must be defined with all the register class names
2363 as enumerated values. @code{NO_REGS} must be first. @code{ALL_REGS}
2364 must be the last register class, followed by one more enumerated value,
2365 @code{LIM_REG_CLASSES}, which is not a register class but rather
2366 tells how many classes there are.
2368 Each register class has a number, which is the value of casting
2369 the class name to type @code{int}. The number serves as an index
2370 in many of the tables described below.
2373 @defmac N_REG_CLASSES
2374 The number of distinct register classes, defined as follows:
2377 #define N_REG_CLASSES (int) LIM_REG_CLASSES
2381 @defmac REG_CLASS_NAMES
2382 An initializer containing the names of the register classes as C string
2383 constants. These names are used in writing some of the debugging dumps.
2386 @defmac REG_CLASS_CONTENTS
2387 An initializer containing the contents of the register classes, as integers
2388 which are bit masks. The @var{n}th integer specifies the contents of class
2389 @var{n}. The way the integer @var{mask} is interpreted is that
2390 register @var{r} is in the class if @code{@var{mask} & (1 << @var{r})} is 1.
2392 When the machine has more than 32 registers, an integer does not suffice.
2393 Then the integers are replaced by sub-initializers, braced groupings containing
2394 several integers. Each sub-initializer must be suitable as an initializer
2395 for the type @code{HARD_REG_SET} which is defined in @file{hard-reg-set.h}.
2396 In this situation, the first integer in each sub-initializer corresponds to
2397 registers 0 through 31, the second integer to registers 32 through 63, and
2401 @defmac REGNO_REG_CLASS (@var{regno})
2402 A C expression whose value is a register class containing hard register
2403 @var{regno}. In general there is more than one such class; choose a class
2404 which is @dfn{minimal}, meaning that no smaller class also contains the
2408 @defmac BASE_REG_CLASS
2409 A macro whose definition is the name of the class to which a valid
2410 base register must belong. A base register is one used in an address
2411 which is the register value plus a displacement.
2414 @defmac MODE_BASE_REG_CLASS (@var{mode})
2415 This is a variation of the @code{BASE_REG_CLASS} macro which allows
2416 the selection of a base register in a mode dependent manner. If
2417 @var{mode} is VOIDmode then it should return the same value as
2418 @code{BASE_REG_CLASS}.
2421 @defmac MODE_BASE_REG_REG_CLASS (@var{mode})
2422 A C expression whose value is the register class to which a valid
2423 base register must belong in order to be used in a base plus index
2424 register address. You should define this macro if base plus index
2425 addresses have different requirements than other base register uses.
2428 @defmac MODE_CODE_BASE_REG_CLASS (@var{mode}, @var{address_space}, @var{outer_code}, @var{index_code})
2429 A C expression whose value is the register class to which a valid
2430 base register for a memory reference in mode @var{mode} to address
2431 space @var{address_space} must belong. @var{outer_code} and @var{index_code}
2432 define the context in which the base register occurs. @var{outer_code} is
2433 the code of the immediately enclosing expression (@code{MEM} for the top level
2434 of an address, @code{ADDRESS} for something that occurs in an
2435 @code{address_operand}). @var{index_code} is the code of the corresponding
2436 index expression if @var{outer_code} is @code{PLUS}; @code{SCRATCH} otherwise.
2439 @defmac INDEX_REG_CLASS
2440 A macro whose definition is the name of the class to which a valid
2441 index register must belong. An index register is one used in an
2442 address where its value is either multiplied by a scale factor or
2443 added to another register (as well as added to a displacement).
2446 @defmac REGNO_OK_FOR_BASE_P (@var{num})
2447 A C expression which is nonzero if register number @var{num} is
2448 suitable for use as a base register in operand addresses.
2451 @defmac REGNO_MODE_OK_FOR_BASE_P (@var{num}, @var{mode})
2452 A C expression that is just like @code{REGNO_OK_FOR_BASE_P}, except that
2453 that expression may examine the mode of the memory reference in
2454 @var{mode}. You should define this macro if the mode of the memory
2455 reference affects whether a register may be used as a base register. If
2456 you define this macro, the compiler will use it instead of
2457 @code{REGNO_OK_FOR_BASE_P}. The mode may be @code{VOIDmode} for
2458 addresses that appear outside a @code{MEM}, i.e., as an
2459 @code{address_operand}.
2462 @defmac REGNO_MODE_OK_FOR_REG_BASE_P (@var{num}, @var{mode})
2463 A C expression which is nonzero if register number @var{num} is suitable for
2464 use as a base register in base plus index operand addresses, accessing
2465 memory in mode @var{mode}. It may be either a suitable hard register or a
2466 pseudo register that has been allocated such a hard register. You should
2467 define this macro if base plus index addresses have different requirements
2468 than other base register uses.
2470 Use of this macro is deprecated; please use the more general
2471 @code{REGNO_MODE_CODE_OK_FOR_BASE_P}.
2474 @defmac REGNO_MODE_CODE_OK_FOR_BASE_P (@var{num}, @var{mode}, @var{address_space}, @var{outer_code}, @var{index_code})
2475 A C expression which is nonzero if register number @var{num} is
2476 suitable for use as a base register in operand addresses, accessing
2477 memory in mode @var{mode} in address space @var{address_space}.
2478 This is similar to @code{REGNO_MODE_OK_FOR_BASE_P}, except
2479 that that expression may examine the context in which the register
2480 appears in the memory reference. @var{outer_code} is the code of the
2481 immediately enclosing expression (@code{MEM} if at the top level of the
2482 address, @code{ADDRESS} for something that occurs in an
2483 @code{address_operand}). @var{index_code} is the code of the
2484 corresponding index expression if @var{outer_code} is @code{PLUS};
2485 @code{SCRATCH} otherwise. The mode may be @code{VOIDmode} for addresses
2486 that appear outside a @code{MEM}, i.e., as an @code{address_operand}.
2489 @defmac REGNO_OK_FOR_INDEX_P (@var{num})
2490 A C expression which is nonzero if register number @var{num} is
2491 suitable for use as an index register in operand addresses. It may be
2492 either a suitable hard register or a pseudo register that has been
2493 allocated such a hard register.
2495 The difference between an index register and a base register is that
2496 the index register may be scaled. If an address involves the sum of
2497 two registers, neither one of them scaled, then either one may be
2498 labeled the ``base'' and the other the ``index''; but whichever
2499 labeling is used must fit the machine's constraints of which registers
2500 may serve in each capacity. The compiler will try both labelings,
2501 looking for one that is valid, and will reload one or both registers
2502 only if neither labeling works.
2505 @hook TARGET_PREFERRED_RENAME_CLASS
2507 @hook TARGET_PREFERRED_RELOAD_CLASS
2508 A target hook that places additional restrictions on the register class
2509 to use when it is necessary to copy value @var{x} into a register in class
2510 @var{rclass}. The value is a register class; perhaps @var{rclass}, or perhaps
2511 another, smaller class.
2513 The default version of this hook always returns value of @code{rclass} argument.
2515 Sometimes returning a more restrictive class makes better code. For
2516 example, on the 68000, when @var{x} is an integer constant that is in range
2517 for a @samp{moveq} instruction, the value of this macro is always
2518 @code{DATA_REGS} as long as @var{rclass} includes the data registers.
2519 Requiring a data register guarantees that a @samp{moveq} will be used.
2521 One case where @code{TARGET_PREFERRED_RELOAD_CLASS} must not return
2522 @var{rclass} is if @var{x} is a legitimate constant which cannot be
2523 loaded into some register class. By returning @code{NO_REGS} you can
2524 force @var{x} into a memory location. For example, rs6000 can load
2525 immediate values into general-purpose registers, but does not have an
2526 instruction for loading an immediate value into a floating-point
2527 register, so @code{TARGET_PREFERRED_RELOAD_CLASS} returns @code{NO_REGS} when
2528 @var{x} is a floating-point constant. If the constant can't be loaded
2529 into any kind of register, code generation will be better if
2530 @code{TARGET_LEGITIMATE_CONSTANT_P} makes the constant illegitimate instead
2531 of using @code{TARGET_PREFERRED_RELOAD_CLASS}.
2533 If an insn has pseudos in it after register allocation, reload will go
2534 through the alternatives and call repeatedly @code{TARGET_PREFERRED_RELOAD_CLASS}
2535 to find the best one. Returning @code{NO_REGS}, in this case, makes
2536 reload add a @code{!} in front of the constraint: the x86 back-end uses
2537 this feature to discourage usage of 387 registers when math is done in
2538 the SSE registers (and vice versa).
2541 @defmac PREFERRED_RELOAD_CLASS (@var{x}, @var{class})
2542 A C expression that places additional restrictions on the register class
2543 to use when it is necessary to copy value @var{x} into a register in class
2544 @var{class}. The value is a register class; perhaps @var{class}, or perhaps
2545 another, smaller class. On many machines, the following definition is
2549 #define PREFERRED_RELOAD_CLASS(X,CLASS) CLASS
2552 Sometimes returning a more restrictive class makes better code. For
2553 example, on the 68000, when @var{x} is an integer constant that is in range
2554 for a @samp{moveq} instruction, the value of this macro is always
2555 @code{DATA_REGS} as long as @var{class} includes the data registers.
2556 Requiring a data register guarantees that a @samp{moveq} will be used.
2558 One case where @code{PREFERRED_RELOAD_CLASS} must not return
2559 @var{class} is if @var{x} is a legitimate constant which cannot be
2560 loaded into some register class. By returning @code{NO_REGS} you can
2561 force @var{x} into a memory location. For example, rs6000 can load
2562 immediate values into general-purpose registers, but does not have an
2563 instruction for loading an immediate value into a floating-point
2564 register, so @code{PREFERRED_RELOAD_CLASS} returns @code{NO_REGS} when
2565 @var{x} is a floating-point constant. If the constant can't be loaded
2566 into any kind of register, code generation will be better if
2567 @code{TARGET_LEGITIMATE_CONSTANT_P} makes the constant illegitimate instead
2568 of using @code{TARGET_PREFERRED_RELOAD_CLASS}.
2570 If an insn has pseudos in it after register allocation, reload will go
2571 through the alternatives and call repeatedly @code{PREFERRED_RELOAD_CLASS}
2572 to find the best one. Returning @code{NO_REGS}, in this case, makes
2573 reload add a @code{!} in front of the constraint: the x86 back-end uses
2574 this feature to discourage usage of 387 registers when math is done in
2575 the SSE registers (and vice versa).
2578 @hook TARGET_PREFERRED_OUTPUT_RELOAD_CLASS
2579 Like @code{TARGET_PREFERRED_RELOAD_CLASS}, but for output reloads instead of
2582 The default version of this hook always returns value of @code{rclass}
2585 You can also use @code{TARGET_PREFERRED_OUTPUT_RELOAD_CLASS} to discourage
2586 reload from using some alternatives, like @code{TARGET_PREFERRED_RELOAD_CLASS}.
2589 @defmac LIMIT_RELOAD_CLASS (@var{mode}, @var{class})
2590 A C expression that places additional restrictions on the register class
2591 to use when it is necessary to be able to hold a value of mode
2592 @var{mode} in a reload register for which class @var{class} would
2595 Unlike @code{PREFERRED_RELOAD_CLASS}, this macro should be used when
2596 there are certain modes that simply can't go in certain reload classes.
2598 The value is a register class; perhaps @var{class}, or perhaps another,
2601 Don't define this macro unless the target machine has limitations which
2602 require the macro to do something nontrivial.
2605 @hook TARGET_SECONDARY_RELOAD
2606 Many machines have some registers that cannot be copied directly to or
2607 from memory or even from other types of registers. An example is the
2608 @samp{MQ} register, which on most machines, can only be copied to or
2609 from general registers, but not memory. Below, we shall be using the
2610 term 'intermediate register' when a move operation cannot be performed
2611 directly, but has to be done by copying the source into the intermediate
2612 register first, and then copying the intermediate register to the
2613 destination. An intermediate register always has the same mode as
2614 source and destination. Since it holds the actual value being copied,
2615 reload might apply optimizations to re-use an intermediate register
2616 and eliding the copy from the source when it can determine that the
2617 intermediate register still holds the required value.
2619 Another kind of secondary reload is required on some machines which
2620 allow copying all registers to and from memory, but require a scratch
2621 register for stores to some memory locations (e.g., those with symbolic
2622 address on the RT, and those with certain symbolic address on the SPARC
2623 when compiling PIC)@. Scratch registers need not have the same mode
2624 as the value being copied, and usually hold a different value than
2625 that being copied. Special patterns in the md file are needed to
2626 describe how the copy is performed with the help of the scratch register;
2627 these patterns also describe the number, register class(es) and mode(s)
2628 of the scratch register(s).
2630 In some cases, both an intermediate and a scratch register are required.
2632 For input reloads, this target hook is called with nonzero @var{in_p},
2633 and @var{x} is an rtx that needs to be copied to a register of class
2634 @var{reload_class} in @var{reload_mode}. For output reloads, this target
2635 hook is called with zero @var{in_p}, and a register of class @var{reload_class}
2636 needs to be copied to rtx @var{x} in @var{reload_mode}.
2638 If copying a register of @var{reload_class} from/to @var{x} requires
2639 an intermediate register, the hook @code{secondary_reload} should
2640 return the register class required for this intermediate register.
2641 If no intermediate register is required, it should return NO_REGS.
2642 If more than one intermediate register is required, describe the one
2643 that is closest in the copy chain to the reload register.
2645 If scratch registers are needed, you also have to describe how to
2646 perform the copy from/to the reload register to/from this
2647 closest intermediate register. Or if no intermediate register is
2648 required, but still a scratch register is needed, describe the
2649 copy from/to the reload register to/from the reload operand @var{x}.
2651 You do this by setting @code{sri->icode} to the instruction code of a pattern
2652 in the md file which performs the move. Operands 0 and 1 are the output
2653 and input of this copy, respectively. Operands from operand 2 onward are
2654 for scratch operands. These scratch operands must have a mode, and a
2655 single-register-class
2656 @c [later: or memory]
2659 When an intermediate register is used, the @code{secondary_reload}
2660 hook will be called again to determine how to copy the intermediate
2661 register to/from the reload operand @var{x}, so your hook must also
2662 have code to handle the register class of the intermediate operand.
2664 @c [For later: maybe we'll allow multi-alternative reload patterns -
2665 @c the port maintainer could name a mov<mode> pattern that has clobbers -
2666 @c and match the constraints of input and output to determine the required
2667 @c alternative. A restriction would be that constraints used to match
2668 @c against reloads registers would have to be written as register class
2669 @c constraints, or we need a new target macro / hook that tells us if an
2670 @c arbitrary constraint can match an unknown register of a given class.
2671 @c Such a macro / hook would also be useful in other places.]
2674 @var{x} might be a pseudo-register or a @code{subreg} of a
2675 pseudo-register, which could either be in a hard register or in memory.
2676 Use @code{true_regnum} to find out; it will return @minus{}1 if the pseudo is
2677 in memory and the hard register number if it is in a register.
2679 Scratch operands in memory (constraint @code{"=m"} / @code{"=&m"}) are
2680 currently not supported. For the time being, you will have to continue
2681 to use @code{SECONDARY_MEMORY_NEEDED} for that purpose.
2683 @code{copy_cost} also uses this target hook to find out how values are
2684 copied. If you want it to include some extra cost for the need to allocate
2685 (a) scratch register(s), set @code{sri->extra_cost} to the additional cost.
2686 Or if two dependent moves are supposed to have a lower cost than the sum
2687 of the individual moves due to expected fortuitous scheduling and/or special
2688 forwarding logic, you can set @code{sri->extra_cost} to a negative amount.
2691 @defmac SECONDARY_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2692 @defmacx SECONDARY_INPUT_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2693 @defmacx SECONDARY_OUTPUT_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2694 These macros are obsolete, new ports should use the target hook
2695 @code{TARGET_SECONDARY_RELOAD} instead.
2697 These are obsolete macros, replaced by the @code{TARGET_SECONDARY_RELOAD}
2698 target hook. Older ports still define these macros to indicate to the
2699 reload phase that it may
2700 need to allocate at least one register for a reload in addition to the
2701 register to contain the data. Specifically, if copying @var{x} to a
2702 register @var{class} in @var{mode} requires an intermediate register,
2703 you were supposed to define @code{SECONDARY_INPUT_RELOAD_CLASS} to return the
2704 largest register class all of whose registers can be used as
2705 intermediate registers or scratch registers.
2707 If copying a register @var{class} in @var{mode} to @var{x} requires an
2708 intermediate or scratch register, @code{SECONDARY_OUTPUT_RELOAD_CLASS}
2709 was supposed to be defined be defined to return the largest register
2710 class required. If the
2711 requirements for input and output reloads were the same, the macro
2712 @code{SECONDARY_RELOAD_CLASS} should have been used instead of defining both
2715 The values returned by these macros are often @code{GENERAL_REGS}.
2716 Return @code{NO_REGS} if no spare register is needed; i.e., if @var{x}
2717 can be directly copied to or from a register of @var{class} in
2718 @var{mode} without requiring a scratch register. Do not define this
2719 macro if it would always return @code{NO_REGS}.
2721 If a scratch register is required (either with or without an
2722 intermediate register), you were supposed to define patterns for
2723 @samp{reload_in@var{m}} or @samp{reload_out@var{m}}, as required
2724 (@pxref{Standard Names}. These patterns, which were normally
2725 implemented with a @code{define_expand}, should be similar to the
2726 @samp{mov@var{m}} patterns, except that operand 2 is the scratch
2729 These patterns need constraints for the reload register and scratch
2731 contain a single register class. If the original reload register (whose
2732 class is @var{class}) can meet the constraint given in the pattern, the
2733 value returned by these macros is used for the class of the scratch
2734 register. Otherwise, two additional reload registers are required.
2735 Their classes are obtained from the constraints in the insn pattern.
2737 @var{x} might be a pseudo-register or a @code{subreg} of a
2738 pseudo-register, which could either be in a hard register or in memory.
2739 Use @code{true_regnum} to find out; it will return @minus{}1 if the pseudo is
2740 in memory and the hard register number if it is in a register.
2742 These macros should not be used in the case where a particular class of
2743 registers can only be copied to memory and not to another class of
2744 registers. In that case, secondary reload registers are not needed and
2745 would not be helpful. Instead, a stack location must be used to perform
2746 the copy and the @code{mov@var{m}} pattern should use memory as an
2747 intermediate storage. This case often occurs between floating-point and
2751 @defmac SECONDARY_MEMORY_NEEDED (@var{class1}, @var{class2}, @var{m})
2752 Certain machines have the property that some registers cannot be copied
2753 to some other registers without using memory. Define this macro on
2754 those machines to be a C expression that is nonzero if objects of mode
2755 @var{m} in registers of @var{class1} can only be copied to registers of
2756 class @var{class2} by storing a register of @var{class1} into memory
2757 and loading that memory location into a register of @var{class2}.
2759 Do not define this macro if its value would always be zero.
2762 @defmac SECONDARY_MEMORY_NEEDED_RTX (@var{mode})
2763 Normally when @code{SECONDARY_MEMORY_NEEDED} is defined, the compiler
2764 allocates a stack slot for a memory location needed for register copies.
2765 If this macro is defined, the compiler instead uses the memory location
2766 defined by this macro.
2768 Do not define this macro if you do not define
2769 @code{SECONDARY_MEMORY_NEEDED}.
2772 @defmac SECONDARY_MEMORY_NEEDED_MODE (@var{mode})
2773 When the compiler needs a secondary memory location to copy between two
2774 registers of mode @var{mode}, it normally allocates sufficient memory to
2775 hold a quantity of @code{BITS_PER_WORD} bits and performs the store and
2776 load operations in a mode that many bits wide and whose class is the
2777 same as that of @var{mode}.
2779 This is right thing to do on most machines because it ensures that all
2780 bits of the register are copied and prevents accesses to the registers
2781 in a narrower mode, which some machines prohibit for floating-point
2784 However, this default behavior is not correct on some machines, such as
2785 the DEC Alpha, that store short integers in floating-point registers
2786 differently than in integer registers. On those machines, the default
2787 widening will not work correctly and you must define this macro to
2788 suppress that widening in some cases. See the file @file{alpha.h} for
2791 Do not define this macro if you do not define
2792 @code{SECONDARY_MEMORY_NEEDED} or if widening @var{mode} to a mode that
2793 is @code{BITS_PER_WORD} bits wide is correct for your machine.
2796 @hook TARGET_CLASS_LIKELY_SPILLED_P
2797 A target hook which returns @code{true} if pseudos that have been assigned
2798 to registers of class @var{rclass} would likely be spilled because
2799 registers of @var{rclass} are needed for spill registers.
2801 The default version of this target hook returns @code{true} if @var{rclass}
2802 has exactly one register and @code{false} otherwise. On most machines, this
2803 default should be used. Only use this target hook to some other expression
2804 if pseudos allocated by @file{local-alloc.c} end up in memory because their
2805 hard registers were needed for spill registers. If this target hook returns
2806 @code{false} for those classes, those pseudos will only be allocated by
2807 @file{global.c}, which knows how to reallocate the pseudo to another
2808 register. If there would not be another register available for reallocation,
2809 you should not change the implementation of this target hook since
2810 the only effect of such implementation would be to slow down register
2814 @hook TARGET_CLASS_MAX_NREGS
2815 A target hook returns the maximum number of consecutive registers
2816 of class @var{rclass} needed to hold a value of mode @var{mode}.
2818 This is closely related to the macro @code{HARD_REGNO_NREGS}. In fact,
2819 the value returned by @code{TARGET_CLASS_MAX_NREGS (@var{rclass},
2820 @var{mode})} target hook should be the maximum value of
2821 @code{HARD_REGNO_NREGS (@var{regno}, @var{mode})} for all @var{regno}
2822 values in the class @var{rclass}.
2824 This target hook helps control the handling of multiple-word values
2827 The default version of this target hook returns the size of @var{mode}
2831 @defmac CLASS_MAX_NREGS (@var{class}, @var{mode})
2832 A C expression for the maximum number of consecutive registers
2833 of class @var{class} needed to hold a value of mode @var{mode}.
2835 This is closely related to the macro @code{HARD_REGNO_NREGS}. In fact,
2836 the value of the macro @code{CLASS_MAX_NREGS (@var{class}, @var{mode})}
2837 should be the maximum value of @code{HARD_REGNO_NREGS (@var{regno},
2838 @var{mode})} for all @var{regno} values in the class @var{class}.
2840 This macro helps control the handling of multiple-word values
2844 @defmac CANNOT_CHANGE_MODE_CLASS (@var{from}, @var{to}, @var{class})
2845 If defined, a C expression that returns nonzero for a @var{class} for which
2846 a change from mode @var{from} to mode @var{to} is invalid.
2848 For the example, loading 32-bit integer or floating-point objects into
2849 floating-point registers on the Alpha extends them to 64 bits.
2850 Therefore loading a 64-bit object and then storing it as a 32-bit object
2851 does not store the low-order 32 bits, as would be the case for a normal
2852 register. Therefore, @file{alpha.h} defines @code{CANNOT_CHANGE_MODE_CLASS}
2856 #define CANNOT_CHANGE_MODE_CLASS(FROM, TO, CLASS) \
2857 (GET_MODE_SIZE (FROM) != GET_MODE_SIZE (TO) \
2858 ? reg_classes_intersect_p (FLOAT_REGS, (CLASS)) : 0)
2862 @node Old Constraints
2863 @section Obsolete Macros for Defining Constraints
2864 @cindex defining constraints, obsolete method
2865 @cindex constraints, defining, obsolete method
2867 Machine-specific constraints can be defined with these macros instead
2868 of the machine description constructs described in @ref{Define
2869 Constraints}. This mechanism is obsolete. New ports should not use
2870 it; old ports should convert to the new mechanism.
2872 @defmac CONSTRAINT_LEN (@var{char}, @var{str})
2873 For the constraint at the start of @var{str}, which starts with the letter
2874 @var{c}, return the length. This allows you to have register class /
2875 constant / extra constraints that are longer than a single letter;
2876 you don't need to define this macro if you can do with single-letter
2877 constraints only. The definition of this macro should use
2878 DEFAULT_CONSTRAINT_LEN for all the characters that you don't want
2879 to handle specially.
2880 There are some sanity checks in genoutput.c that check the constraint lengths
2881 for the md file, so you can also use this macro to help you while you are
2882 transitioning from a byzantine single-letter-constraint scheme: when you
2883 return a negative length for a constraint you want to re-use, genoutput
2884 will complain about every instance where it is used in the md file.
2887 @defmac REG_CLASS_FROM_LETTER (@var{char})
2888 A C expression which defines the machine-dependent operand constraint
2889 letters for register classes. If @var{char} is such a letter, the
2890 value should be the register class corresponding to it. Otherwise,
2891 the value should be @code{NO_REGS}. The register letter @samp{r},
2892 corresponding to class @code{GENERAL_REGS}, will not be passed
2893 to this macro; you do not need to handle it.
2896 @defmac REG_CLASS_FROM_CONSTRAINT (@var{char}, @var{str})
2897 Like @code{REG_CLASS_FROM_LETTER}, but you also get the constraint string
2898 passed in @var{str}, so that you can use suffixes to distinguish between
2902 @defmac CONST_OK_FOR_LETTER_P (@var{value}, @var{c})
2903 A C expression that defines the machine-dependent operand constraint
2904 letters (@samp{I}, @samp{J}, @samp{K}, @dots{} @samp{P}) that specify
2905 particular ranges of integer values. If @var{c} is one of those
2906 letters, the expression should check that @var{value}, an integer, is in
2907 the appropriate range and return 1 if so, 0 otherwise. If @var{c} is
2908 not one of those letters, the value should be 0 regardless of
2912 @defmac CONST_OK_FOR_CONSTRAINT_P (@var{value}, @var{c}, @var{str})
2913 Like @code{CONST_OK_FOR_LETTER_P}, but you also get the constraint
2914 string passed in @var{str}, so that you can use suffixes to distinguish
2915 between different variants.
2918 @defmac CONST_DOUBLE_OK_FOR_LETTER_P (@var{value}, @var{c})
2919 A C expression that defines the machine-dependent operand constraint
2920 letters that specify particular ranges of @code{const_double} values
2921 (@samp{G} or @samp{H}).
2923 If @var{c} is one of those letters, the expression should check that
2924 @var{value}, an RTX of code @code{const_double}, is in the appropriate
2925 range and return 1 if so, 0 otherwise. If @var{c} is not one of those
2926 letters, the value should be 0 regardless of @var{value}.
2928 @code{const_double} is used for all floating-point constants and for
2929 @code{DImode} fixed-point constants. A given letter can accept either
2930 or both kinds of values. It can use @code{GET_MODE} to distinguish
2931 between these kinds.
2934 @defmac CONST_DOUBLE_OK_FOR_CONSTRAINT_P (@var{value}, @var{c}, @var{str})
2935 Like @code{CONST_DOUBLE_OK_FOR_LETTER_P}, but you also get the constraint
2936 string passed in @var{str}, so that you can use suffixes to distinguish
2937 between different variants.
2940 @defmac EXTRA_CONSTRAINT (@var{value}, @var{c})
2941 A C expression that defines the optional machine-dependent constraint
2942 letters that can be used to segregate specific types of operands, usually
2943 memory references, for the target machine. Any letter that is not
2944 elsewhere defined and not matched by @code{REG_CLASS_FROM_LETTER} /
2945 @code{REG_CLASS_FROM_CONSTRAINT}
2946 may be used. Normally this macro will not be defined.
2948 If it is required for a particular target machine, it should return 1
2949 if @var{value} corresponds to the operand type represented by the
2950 constraint letter @var{c}. If @var{c} is not defined as an extra
2951 constraint, the value returned should be 0 regardless of @var{value}.
2953 For example, on the ROMP, load instructions cannot have their output
2954 in r0 if the memory reference contains a symbolic address. Constraint
2955 letter @samp{Q} is defined as representing a memory address that does
2956 @emph{not} contain a symbolic address. An alternative is specified with
2957 a @samp{Q} constraint on the input and @samp{r} on the output. The next
2958 alternative specifies @samp{m} on the input and a register class that
2959 does not include r0 on the output.
2962 @defmac EXTRA_CONSTRAINT_STR (@var{value}, @var{c}, @var{str})
2963 Like @code{EXTRA_CONSTRAINT}, but you also get the constraint string passed
2964 in @var{str}, so that you can use suffixes to distinguish between different
2968 @defmac EXTRA_MEMORY_CONSTRAINT (@var{c}, @var{str})
2969 A C expression that defines the optional machine-dependent constraint
2970 letters, amongst those accepted by @code{EXTRA_CONSTRAINT}, that should
2971 be treated like memory constraints by the reload pass.
2973 It should return 1 if the operand type represented by the constraint
2974 at the start of @var{str}, the first letter of which is the letter @var{c},
2975 comprises a subset of all memory references including
2976 all those whose address is simply a base register. This allows the reload
2977 pass to reload an operand, if it does not directly correspond to the operand
2978 type of @var{c}, by copying its address into a base register.
2980 For example, on the S/390, some instructions do not accept arbitrary
2981 memory references, but only those that do not make use of an index
2982 register. The constraint letter @samp{Q} is defined via
2983 @code{EXTRA_CONSTRAINT} as representing a memory address of this type.
2984 If the letter @samp{Q} is marked as @code{EXTRA_MEMORY_CONSTRAINT},
2985 a @samp{Q} constraint can handle any memory operand, because the
2986 reload pass knows it can be reloaded by copying the memory address
2987 into a base register if required. This is analogous to the way
2988 an @samp{o} constraint can handle any memory operand.
2991 @defmac EXTRA_ADDRESS_CONSTRAINT (@var{c}, @var{str})
2992 A C expression that defines the optional machine-dependent constraint
2993 letters, amongst those accepted by @code{EXTRA_CONSTRAINT} /
2994 @code{EXTRA_CONSTRAINT_STR}, that should
2995 be treated like address constraints by the reload pass.
2997 It should return 1 if the operand type represented by the constraint
2998 at the start of @var{str}, which starts with the letter @var{c}, comprises
2999 a subset of all memory addresses including
3000 all those that consist of just a base register. This allows the reload
3001 pass to reload an operand, if it does not directly correspond to the operand
3002 type of @var{str}, by copying it into a base register.
3004 Any constraint marked as @code{EXTRA_ADDRESS_CONSTRAINT} can only
3005 be used with the @code{address_operand} predicate. It is treated
3006 analogously to the @samp{p} constraint.
3009 @node Stack and Calling
3010 @section Stack Layout and Calling Conventions
3011 @cindex calling conventions
3013 @c prevent bad page break with this line
3014 This describes the stack layout and calling conventions.
3018 * Exception Handling::
3023 * Register Arguments::
3025 * Aggregate Return::
3030 * Stack Smashing Protection::
3034 @subsection Basic Stack Layout
3035 @cindex stack frame layout
3036 @cindex frame layout
3038 @c prevent bad page break with this line
3039 Here is the basic stack layout.
3041 @defmac STACK_GROWS_DOWNWARD
3042 Define this macro if pushing a word onto the stack moves the stack
3043 pointer to a smaller address.
3045 When we say, ``define this macro if @dots{}'', it means that the
3046 compiler checks this macro only with @code{#ifdef} so the precise
3047 definition used does not matter.
3050 @defmac STACK_PUSH_CODE
3051 This macro defines the operation used when something is pushed
3052 on the stack. In RTL, a push operation will be
3053 @code{(set (mem (STACK_PUSH_CODE (reg sp))) @dots{})}
3055 The choices are @code{PRE_DEC}, @code{POST_DEC}, @code{PRE_INC},
3056 and @code{POST_INC}. Which of these is correct depends on
3057 the stack direction and on whether the stack pointer points
3058 to the last item on the stack or whether it points to the
3059 space for the next item on the stack.
3061 The default is @code{PRE_DEC} when @code{STACK_GROWS_DOWNWARD} is
3062 defined, which is almost always right, and @code{PRE_INC} otherwise,
3063 which is often wrong.
3066 @defmac FRAME_GROWS_DOWNWARD
3067 Define this macro to nonzero value if the addresses of local variable slots
3068 are at negative offsets from the frame pointer.
3071 @defmac ARGS_GROW_DOWNWARD
3072 Define this macro if successive arguments to a function occupy decreasing
3073 addresses on the stack.
3076 @defmac STARTING_FRAME_OFFSET
3077 Offset from the frame pointer to the first local variable slot to be allocated.
3079 If @code{FRAME_GROWS_DOWNWARD}, find the next slot's offset by
3080 subtracting the first slot's length from @code{STARTING_FRAME_OFFSET}.
3081 Otherwise, it is found by adding the length of the first slot to the
3082 value @code{STARTING_FRAME_OFFSET}.
3083 @c i'm not sure if the above is still correct.. had to change it to get
3084 @c rid of an overfull. --mew 2feb93
3087 @defmac STACK_ALIGNMENT_NEEDED
3088 Define to zero to disable final alignment of the stack during reload.
3089 The nonzero default for this macro is suitable for most ports.
3091 On ports where @code{STARTING_FRAME_OFFSET} is nonzero or where there
3092 is a register save block following the local block that doesn't require
3093 alignment to @code{STACK_BOUNDARY}, it may be beneficial to disable
3094 stack alignment and do it in the backend.
3097 @defmac STACK_POINTER_OFFSET
3098 Offset from the stack pointer register to the first location at which
3099 outgoing arguments are placed. If not specified, the default value of
3100 zero is used. This is the proper value for most machines.
3102 If @code{ARGS_GROW_DOWNWARD}, this is the offset to the location above
3103 the first location at which outgoing arguments are placed.
3106 @defmac FIRST_PARM_OFFSET (@var{fundecl})
3107 Offset from the argument pointer register to the first argument's
3108 address. On some machines it may depend on the data type of the
3111 If @code{ARGS_GROW_DOWNWARD}, this is the offset to the location above
3112 the first argument's address.
3115 @defmac STACK_DYNAMIC_OFFSET (@var{fundecl})
3116 Offset from the stack pointer register to an item dynamically allocated
3117 on the stack, e.g., by @code{alloca}.
3119 The default value for this macro is @code{STACK_POINTER_OFFSET} plus the
3120 length of the outgoing arguments. The default is correct for most
3121 machines. See @file{function.c} for details.
3124 @defmac INITIAL_FRAME_ADDRESS_RTX
3125 A C expression whose value is RTL representing the address of the initial
3126 stack frame. This address is passed to @code{RETURN_ADDR_RTX} and
3127 @code{DYNAMIC_CHAIN_ADDRESS}. If you don't define this macro, a reasonable
3128 default value will be used. Define this macro in order to make frame pointer
3129 elimination work in the presence of @code{__builtin_frame_address (count)} and
3130 @code{__builtin_return_address (count)} for @code{count} not equal to zero.
3133 @defmac DYNAMIC_CHAIN_ADDRESS (@var{frameaddr})
3134 A C expression whose value is RTL representing the address in a stack
3135 frame where the pointer to the caller's frame is stored. Assume that
3136 @var{frameaddr} is an RTL expression for the address of the stack frame
3139 If you don't define this macro, the default is to return the value
3140 of @var{frameaddr}---that is, the stack frame address is also the
3141 address of the stack word that points to the previous frame.
3144 @defmac SETUP_FRAME_ADDRESSES
3145 If defined, a C expression that produces the machine-specific code to
3146 setup the stack so that arbitrary frames can be accessed. For example,
3147 on the SPARC, we must flush all of the register windows to the stack
3148 before we can access arbitrary stack frames. You will seldom need to
3152 @hook TARGET_BUILTIN_SETJMP_FRAME_VALUE
3153 This target hook should return an rtx that is used to store
3154 the address of the current frame into the built in @code{setjmp} buffer.
3155 The default value, @code{virtual_stack_vars_rtx}, is correct for most
3156 machines. One reason you may need to define this target hook is if
3157 @code{hard_frame_pointer_rtx} is the appropriate value on your machine.
3160 @defmac FRAME_ADDR_RTX (@var{frameaddr})
3161 A C expression whose value is RTL representing the value of the frame
3162 address for the current frame. @var{frameaddr} is the frame pointer
3163 of the current frame. This is used for __builtin_frame_address.
3164 You need only define this macro if the frame address is not the same
3165 as the frame pointer. Most machines do not need to define it.
3168 @defmac RETURN_ADDR_RTX (@var{count}, @var{frameaddr})
3169 A C expression whose value is RTL representing the value of the return
3170 address for the frame @var{count} steps up from the current frame, after
3171 the prologue. @var{frameaddr} is the frame pointer of the @var{count}
3172 frame, or the frame pointer of the @var{count} @minus{} 1 frame if
3173 @code{RETURN_ADDR_IN_PREVIOUS_FRAME} is defined.
3175 The value of the expression must always be the correct address when
3176 @var{count} is zero, but may be @code{NULL_RTX} if there is no way to
3177 determine the return address of other frames.
3180 @defmac RETURN_ADDR_IN_PREVIOUS_FRAME
3181 Define this if the return address of a particular stack frame is accessed
3182 from the frame pointer of the previous stack frame.
3185 @defmac INCOMING_RETURN_ADDR_RTX
3186 A C expression whose value is RTL representing the location of the
3187 incoming return address at the beginning of any function, before the
3188 prologue. This RTL is either a @code{REG}, indicating that the return
3189 value is saved in @samp{REG}, or a @code{MEM} representing a location in
3192 You only need to define this macro if you want to support call frame
3193 debugging information like that provided by DWARF 2.
3195 If this RTL is a @code{REG}, you should also define
3196 @code{DWARF_FRAME_RETURN_COLUMN} to @code{DWARF_FRAME_REGNUM (REGNO)}.
3199 @defmac DWARF_ALT_FRAME_RETURN_COLUMN
3200 A C expression whose value is an integer giving a DWARF 2 column
3201 number that may be used as an alternative return column. The column
3202 must not correspond to any gcc hard register (that is, it must not
3203 be in the range of @code{DWARF_FRAME_REGNUM}).
3205 This macro can be useful if @code{DWARF_FRAME_RETURN_COLUMN} is set to a
3206 general register, but an alternative column needs to be used for signal
3207 frames. Some targets have also used different frame return columns
3211 @defmac DWARF_ZERO_REG
3212 A C expression whose value is an integer giving a DWARF 2 register
3213 number that is considered to always have the value zero. This should
3214 only be defined if the target has an architected zero register, and
3215 someone decided it was a good idea to use that register number to
3216 terminate the stack backtrace. New ports should avoid this.
3219 @hook TARGET_DWARF_HANDLE_FRAME_UNSPEC
3220 This target hook allows the backend to emit frame-related insns that
3221 contain UNSPECs or UNSPEC_VOLATILEs. The DWARF 2 call frame debugging
3222 info engine will invoke it on insns of the form
3224 (set (reg) (unspec [@dots{}] UNSPEC_INDEX))
3228 (set (reg) (unspec_volatile [@dots{}] UNSPECV_INDEX)).
3230 to let the backend emit the call frame instructions. @var{label} is
3231 the CFI label attached to the insn, @var{pattern} is the pattern of
3232 the insn and @var{index} is @code{UNSPEC_INDEX} or @code{UNSPECV_INDEX}.
3235 @defmac INCOMING_FRAME_SP_OFFSET
3236 A C expression whose value is an integer giving the offset, in bytes,
3237 from the value of the stack pointer register to the top of the stack
3238 frame at the beginning of any function, before the prologue. The top of
3239 the frame is defined to be the value of the stack pointer in the
3240 previous frame, just before the call instruction.
3242 You only need to define this macro if you want to support call frame
3243 debugging information like that provided by DWARF 2.
3246 @defmac ARG_POINTER_CFA_OFFSET (@var{fundecl})
3247 A C expression whose value is an integer giving the offset, in bytes,
3248 from the argument pointer to the canonical frame address (cfa). The
3249 final value should coincide with that calculated by
3250 @code{INCOMING_FRAME_SP_OFFSET}. Which is unfortunately not usable
3251 during virtual register instantiation.
3253 The default value for this macro is
3254 @code{FIRST_PARM_OFFSET (fundecl) + crtl->args.pretend_args_size},
3255 which is correct for most machines; in general, the arguments are found
3256 immediately before the stack frame. Note that this is not the case on
3257 some targets that save registers into the caller's frame, such as SPARC
3258 and rs6000, and so such targets need to define this macro.
3260 You only need to define this macro if the default is incorrect, and you
3261 want to support call frame debugging information like that provided by
3265 @defmac FRAME_POINTER_CFA_OFFSET (@var{fundecl})
3266 If defined, a C expression whose value is an integer giving the offset
3267 in bytes from the frame pointer to the canonical frame address (cfa).
3268 The final value should coincide with that calculated by
3269 @code{INCOMING_FRAME_SP_OFFSET}.
3271 Normally the CFA is calculated as an offset from the argument pointer,
3272 via @code{ARG_POINTER_CFA_OFFSET}, but if the argument pointer is
3273 variable due to the ABI, this may not be possible. If this macro is
3274 defined, it implies that the virtual register instantiation should be
3275 based on the frame pointer instead of the argument pointer. Only one
3276 of @code{FRAME_POINTER_CFA_OFFSET} and @code{ARG_POINTER_CFA_OFFSET}
3280 @defmac CFA_FRAME_BASE_OFFSET (@var{fundecl})
3281 If defined, a C expression whose value is an integer giving the offset
3282 in bytes from the canonical frame address (cfa) to the frame base used
3283 in DWARF 2 debug information. The default is zero. A different value
3284 may reduce the size of debug information on some ports.
3287 @node Exception Handling
3288 @subsection Exception Handling Support
3289 @cindex exception handling
3291 @defmac EH_RETURN_DATA_REGNO (@var{N})
3292 A C expression whose value is the @var{N}th register number used for
3293 data by exception handlers, or @code{INVALID_REGNUM} if fewer than
3294 @var{N} registers are usable.
3296 The exception handling library routines communicate with the exception
3297 handlers via a set of agreed upon registers. Ideally these registers
3298 should be call-clobbered; it is possible to use call-saved registers,
3299 but may negatively impact code size. The target must support at least
3300 2 data registers, but should define 4 if there are enough free registers.
3302 You must define this macro if you want to support call frame exception
3303 handling like that provided by DWARF 2.
3306 @defmac EH_RETURN_STACKADJ_RTX
3307 A C expression whose value is RTL representing a location in which
3308 to store a stack adjustment to be applied before function return.
3309 This is used to unwind the stack to an exception handler's call frame.
3310 It will be assigned zero on code paths that return normally.
3312 Typically this is a call-clobbered hard register that is otherwise
3313 untouched by the epilogue, but could also be a stack slot.
3315 Do not define this macro if the stack pointer is saved and restored
3316 by the regular prolog and epilog code in the call frame itself; in
3317 this case, the exception handling library routines will update the
3318 stack location to be restored in place. Otherwise, you must define
3319 this macro if you want to support call frame exception handling like
3320 that provided by DWARF 2.
3323 @defmac EH_RETURN_HANDLER_RTX
3324 A C expression whose value is RTL representing a location in which
3325 to store the address of an exception handler to which we should
3326 return. It will not be assigned on code paths that return normally.
3328 Typically this is the location in the call frame at which the normal
3329 return address is stored. For targets that return by popping an
3330 address off the stack, this might be a memory address just below
3331 the @emph{target} call frame rather than inside the current call
3332 frame. If defined, @code{EH_RETURN_STACKADJ_RTX} will have already
3333 been assigned, so it may be used to calculate the location of the
3336 Some targets have more complex requirements than storing to an
3337 address calculable during initial code generation. In that case
3338 the @code{eh_return} instruction pattern should be used instead.
3340 If you want to support call frame exception handling, you must
3341 define either this macro or the @code{eh_return} instruction pattern.
3344 @defmac RETURN_ADDR_OFFSET
3345 If defined, an integer-valued C expression for which rtl will be generated
3346 to add it to the exception handler address before it is searched in the
3347 exception handling tables, and to subtract it again from the address before
3348 using it to return to the exception handler.
3351 @defmac ASM_PREFERRED_EH_DATA_FORMAT (@var{code}, @var{global})
3352 This macro chooses the encoding of pointers embedded in the exception
3353 handling sections. If at all possible, this should be defined such
3354 that the exception handling section will not require dynamic relocations,
3355 and so may be read-only.
3357 @var{code} is 0 for data, 1 for code labels, 2 for function pointers.
3358 @var{global} is true if the symbol may be affected by dynamic relocations.
3359 The macro should return a combination of the @code{DW_EH_PE_*} defines
3360 as found in @file{dwarf2.h}.
3362 If this macro is not defined, pointers will not be encoded but
3363 represented directly.
3366 @defmac ASM_MAYBE_OUTPUT_ENCODED_ADDR_RTX (@var{file}, @var{encoding}, @var{size}, @var{addr}, @var{done})
3367 This macro allows the target to emit whatever special magic is required
3368 to represent the encoding chosen by @code{ASM_PREFERRED_EH_DATA_FORMAT}.
3369 Generic code takes care of pc-relative and indirect encodings; this must
3370 be defined if the target uses text-relative or data-relative encodings.
3372 This is a C statement that branches to @var{done} if the format was
3373 handled. @var{encoding} is the format chosen, @var{size} is the number
3374 of bytes that the format occupies, @var{addr} is the @code{SYMBOL_REF}
3378 @defmac MD_FALLBACK_FRAME_STATE_FOR (@var{context}, @var{fs})
3379 This macro allows the target to add CPU and operating system specific
3380 code to the call-frame unwinder for use when there is no unwind data
3381 available. The most common reason to implement this macro is to unwind
3382 through signal frames.
3384 This macro is called from @code{uw_frame_state_for} in
3385 @file{unwind-dw2.c}, @file{unwind-dw2-xtensa.c} and
3386 @file{unwind-ia64.c}. @var{context} is an @code{_Unwind_Context};
3387 @var{fs} is an @code{_Unwind_FrameState}. Examine @code{context->ra}
3388 for the address of the code being executed and @code{context->cfa} for
3389 the stack pointer value. If the frame can be decoded, the register
3390 save addresses should be updated in @var{fs} and the macro should
3391 evaluate to @code{_URC_NO_REASON}. If the frame cannot be decoded,
3392 the macro should evaluate to @code{_URC_END_OF_STACK}.
3394 For proper signal handling in Java this macro is accompanied by
3395 @code{MAKE_THROW_FRAME}, defined in @file{libjava/include/*-signal.h} headers.
3398 @defmac MD_HANDLE_UNWABI (@var{context}, @var{fs})
3399 This macro allows the target to add operating system specific code to the
3400 call-frame unwinder to handle the IA-64 @code{.unwabi} unwinding directive,
3401 usually used for signal or interrupt frames.
3403 This macro is called from @code{uw_update_context} in @file{unwind-ia64.c}.
3404 @var{context} is an @code{_Unwind_Context};
3405 @var{fs} is an @code{_Unwind_FrameState}. Examine @code{fs->unwabi}
3406 for the abi and context in the @code{.unwabi} directive. If the
3407 @code{.unwabi} directive can be handled, the register save addresses should
3408 be updated in @var{fs}.
3411 @defmac TARGET_USES_WEAK_UNWIND_INFO
3412 A C expression that evaluates to true if the target requires unwind
3413 info to be given comdat linkage. Define it to be @code{1} if comdat
3414 linkage is necessary. The default is @code{0}.
3417 @node Stack Checking
3418 @subsection Specifying How Stack Checking is Done
3420 GCC will check that stack references are within the boundaries of the
3421 stack, if the option @option{-fstack-check} is specified, in one of
3426 If the value of the @code{STACK_CHECK_BUILTIN} macro is nonzero, GCC
3427 will assume that you have arranged for full stack checking to be done
3428 at appropriate places in the configuration files. GCC will not do
3429 other special processing.
3432 If @code{STACK_CHECK_BUILTIN} is zero and the value of the
3433 @code{STACK_CHECK_STATIC_BUILTIN} macro is nonzero, GCC will assume
3434 that you have arranged for static stack checking (checking of the
3435 static stack frame of functions) to be done at appropriate places
3436 in the configuration files. GCC will only emit code to do dynamic
3437 stack checking (checking on dynamic stack allocations) using the third
3441 If neither of the above are true, GCC will generate code to periodically
3442 ``probe'' the stack pointer using the values of the macros defined below.
3445 If neither STACK_CHECK_BUILTIN nor STACK_CHECK_STATIC_BUILTIN is defined,
3446 GCC will change its allocation strategy for large objects if the option
3447 @option{-fstack-check} is specified: they will always be allocated
3448 dynamically if their size exceeds @code{STACK_CHECK_MAX_VAR_SIZE} bytes.
3450 @defmac STACK_CHECK_BUILTIN
3451 A nonzero value if stack checking is done by the configuration files in a
3452 machine-dependent manner. You should define this macro if stack checking
3453 is required by the ABI of your machine or if you would like to do stack
3454 checking in some more efficient way than the generic approach. The default
3455 value of this macro is zero.
3458 @defmac STACK_CHECK_STATIC_BUILTIN
3459 A nonzero value if static stack checking is done by the configuration files
3460 in a machine-dependent manner. You should define this macro if you would
3461 like to do static stack checking in some more efficient way than the generic
3462 approach. The default value of this macro is zero.
3465 @defmac STACK_CHECK_PROBE_INTERVAL_EXP
3466 An integer specifying the interval at which GCC must generate stack probe
3467 instructions, defined as 2 raised to this integer. You will normally
3468 define this macro so that the interval be no larger than the size of
3469 the ``guard pages'' at the end of a stack area. The default value
3470 of 12 (4096-byte interval) is suitable for most systems.
3473 @defmac STACK_CHECK_MOVING_SP
3474 An integer which is nonzero if GCC should move the stack pointer page by page
3475 when doing probes. This can be necessary on systems where the stack pointer
3476 contains the bottom address of the memory area accessible to the executing
3477 thread at any point in time. In this situation an alternate signal stack
3478 is required in order to be able to recover from a stack overflow. The
3479 default value of this macro is zero.
3482 @defmac STACK_CHECK_PROTECT
3483 The number of bytes of stack needed to recover from a stack overflow, for
3484 languages where such a recovery is supported. The default value of 75 words
3485 with the @code{setjmp}/@code{longjmp}-based exception handling mechanism and
3486 8192 bytes with other exception handling mechanisms should be adequate for
3490 The following macros are relevant only if neither STACK_CHECK_BUILTIN
3491 nor STACK_CHECK_STATIC_BUILTIN is defined; you can omit them altogether
3492 in the opposite case.
3494 @defmac STACK_CHECK_MAX_FRAME_SIZE
3495 The maximum size of a stack frame, in bytes. GCC will generate probe
3496 instructions in non-leaf functions to ensure at least this many bytes of
3497 stack are available. If a stack frame is larger than this size, stack
3498 checking will not be reliable and GCC will issue a warning. The
3499 default is chosen so that GCC only generates one instruction on most
3500 systems. You should normally not change the default value of this macro.
3503 @defmac STACK_CHECK_FIXED_FRAME_SIZE
3504 GCC uses this value to generate the above warning message. It
3505 represents the amount of fixed frame used by a function, not including
3506 space for any callee-saved registers, temporaries and user variables.
3507 You need only specify an upper bound for this amount and will normally
3508 use the default of four words.
3511 @defmac STACK_CHECK_MAX_VAR_SIZE
3512 The maximum size, in bytes, of an object that GCC will place in the
3513 fixed area of the stack frame when the user specifies
3514 @option{-fstack-check}.
3515 GCC computed the default from the values of the above macros and you will
3516 normally not need to override that default.
3520 @node Frame Registers
3521 @subsection Registers That Address the Stack Frame
3523 @c prevent bad page break with this line
3524 This discusses registers that address the stack frame.
3526 @defmac STACK_POINTER_REGNUM
3527 The register number of the stack pointer register, which must also be a
3528 fixed register according to @code{FIXED_REGISTERS}. On most machines,
3529 the hardware determines which register this is.
3532 @defmac FRAME_POINTER_REGNUM
3533 The register number of the frame pointer register, which is used to
3534 access automatic variables in the stack frame. On some machines, the
3535 hardware determines which register this is. On other machines, you can
3536 choose any register you wish for this purpose.
3539 @defmac HARD_FRAME_POINTER_REGNUM
3540 On some machines the offset between the frame pointer and starting
3541 offset of the automatic variables is not known until after register
3542 allocation has been done (for example, because the saved registers are
3543 between these two locations). On those machines, define
3544 @code{FRAME_POINTER_REGNUM} the number of a special, fixed register to
3545 be used internally until the offset is known, and define
3546 @code{HARD_FRAME_POINTER_REGNUM} to be the actual hard register number
3547 used for the frame pointer.
3549 You should define this macro only in the very rare circumstances when it
3550 is not possible to calculate the offset between the frame pointer and
3551 the automatic variables until after register allocation has been
3552 completed. When this macro is defined, you must also indicate in your
3553 definition of @code{ELIMINABLE_REGS} how to eliminate
3554 @code{FRAME_POINTER_REGNUM} into either @code{HARD_FRAME_POINTER_REGNUM}
3555 or @code{STACK_POINTER_REGNUM}.
3557 Do not define this macro if it would be the same as
3558 @code{FRAME_POINTER_REGNUM}.
3561 @defmac ARG_POINTER_REGNUM
3562 The register number of the arg pointer register, which is used to access
3563 the function's argument list. On some machines, this is the same as the
3564 frame pointer register. On some machines, the hardware determines which
3565 register this is. On other machines, you can choose any register you
3566 wish for this purpose. If this is not the same register as the frame
3567 pointer register, then you must mark it as a fixed register according to
3568 @code{FIXED_REGISTERS}, or arrange to be able to eliminate it
3569 (@pxref{Elimination}).
3572 @defmac HARD_FRAME_POINTER_IS_FRAME_POINTER
3573 Define this to a preprocessor constant that is nonzero if
3574 @code{hard_frame_pointer_rtx} and @code{frame_pointer_rtx} should be
3575 the same. The default definition is @samp{(HARD_FRAME_POINTER_REGNUM
3576 == FRAME_POINTER_REGNUM)}; you only need to define this macro if that
3577 definition is not suitable for use in preprocessor conditionals.
3580 @defmac HARD_FRAME_POINTER_IS_ARG_POINTER
3581 Define this to a preprocessor constant that is nonzero if
3582 @code{hard_frame_pointer_rtx} and @code{arg_pointer_rtx} should be the
3583 same. The default definition is @samp{(HARD_FRAME_POINTER_REGNUM ==
3584 ARG_POINTER_REGNUM)}; you only need to define this macro if that
3585 definition is not suitable for use in preprocessor conditionals.
3588 @defmac RETURN_ADDRESS_POINTER_REGNUM
3589 The register number of the return address pointer register, which is used to
3590 access the current function's return address from the stack. On some
3591 machines, the return address is not at a fixed offset from the frame
3592 pointer or stack pointer or argument pointer. This register can be defined
3593 to point to the return address on the stack, and then be converted by
3594 @code{ELIMINABLE_REGS} into either the frame pointer or stack pointer.
3596 Do not define this macro unless there is no other way to get the return
3597 address from the stack.
3600 @defmac STATIC_CHAIN_REGNUM
3601 @defmacx STATIC_CHAIN_INCOMING_REGNUM
3602 Register numbers used for passing a function's static chain pointer. If
3603 register windows are used, the register number as seen by the called
3604 function is @code{STATIC_CHAIN_INCOMING_REGNUM}, while the register
3605 number as seen by the calling function is @code{STATIC_CHAIN_REGNUM}. If
3606 these registers are the same, @code{STATIC_CHAIN_INCOMING_REGNUM} need
3609 The static chain register need not be a fixed register.
3611 If the static chain is passed in memory, these macros should not be
3612 defined; instead, the @code{TARGET_STATIC_CHAIN} hook should be used.
3615 @hook TARGET_STATIC_CHAIN
3616 This hook replaces the use of @code{STATIC_CHAIN_REGNUM} et al for
3617 targets that may use different static chain locations for different
3618 nested functions. This may be required if the target has function
3619 attributes that affect the calling conventions of the function and
3620 those calling conventions use different static chain locations.
3622 The default version of this hook uses @code{STATIC_CHAIN_REGNUM} et al.
3624 If the static chain is passed in memory, this hook should be used to
3625 provide rtx giving @code{mem} expressions that denote where they are stored.
3626 Often the @code{mem} expression as seen by the caller will be at an offset
3627 from the stack pointer and the @code{mem} expression as seen by the callee
3628 will be at an offset from the frame pointer.
3629 @findex stack_pointer_rtx
3630 @findex frame_pointer_rtx
3631 @findex arg_pointer_rtx
3632 The variables @code{stack_pointer_rtx}, @code{frame_pointer_rtx}, and
3633 @code{arg_pointer_rtx} will have been initialized and should be used
3634 to refer to those items.
3637 @defmac DWARF_FRAME_REGISTERS
3638 This macro specifies the maximum number of hard registers that can be
3639 saved in a call frame. This is used to size data structures used in
3640 DWARF2 exception handling.
3642 Prior to GCC 3.0, this macro was needed in order to establish a stable
3643 exception handling ABI in the face of adding new hard registers for ISA
3644 extensions. In GCC 3.0 and later, the EH ABI is insulated from changes
3645 in the number of hard registers. Nevertheless, this macro can still be
3646 used to reduce the runtime memory requirements of the exception handling
3647 routines, which can be substantial if the ISA contains a lot of
3648 registers that are not call-saved.
3650 If this macro is not defined, it defaults to
3651 @code{FIRST_PSEUDO_REGISTER}.
3654 @defmac PRE_GCC3_DWARF_FRAME_REGISTERS
3656 This macro is similar to @code{DWARF_FRAME_REGISTERS}, but is provided
3657 for backward compatibility in pre GCC 3.0 compiled code.
3659 If this macro is not defined, it defaults to
3660 @code{DWARF_FRAME_REGISTERS}.
3663 @defmac DWARF_REG_TO_UNWIND_COLUMN (@var{regno})
3665 Define this macro if the target's representation for dwarf registers
3666 is different than the internal representation for unwind column.
3667 Given a dwarf register, this macro should return the internal unwind
3668 column number to use instead.
3670 See the PowerPC's SPE target for an example.
3673 @defmac DWARF_FRAME_REGNUM (@var{regno})
3675 Define this macro if the target's representation for dwarf registers
3676 used in .eh_frame or .debug_frame is different from that used in other
3677 debug info sections. Given a GCC hard register number, this macro
3678 should return the .eh_frame register number. The default is
3679 @code{DBX_REGISTER_NUMBER (@var{regno})}.
3683 @defmac DWARF2_FRAME_REG_OUT (@var{regno}, @var{for_eh})
3685 Define this macro to map register numbers held in the call frame info
3686 that GCC has collected using @code{DWARF_FRAME_REGNUM} to those that
3687 should be output in .debug_frame (@code{@var{for_eh}} is zero) and
3688 .eh_frame (@code{@var{for_eh}} is nonzero). The default is to
3689 return @code{@var{regno}}.
3693 @defmac REG_VALUE_IN_UNWIND_CONTEXT
3695 Define this macro if the target stores register values as
3696 @code{_Unwind_Word} type in unwind context. It should be defined if
3697 target register size is larger than the size of @code{void *}. The
3698 default is to store register values as @code{void *} type.
3702 @defmac ASSUME_EXTENDED_UNWIND_CONTEXT
3704 Define this macro to be 1 if the target always uses extended unwind
3705 context with version, args_size and by_value fields. If it is undefined,
3706 it will be defined to 1 when @code{REG_VALUE_IN_UNWIND_CONTEXT} is
3707 defined and 0 otherwise.
3712 @subsection Eliminating Frame Pointer and Arg Pointer
3714 @c prevent bad page break with this line
3715 This is about eliminating the frame pointer and arg pointer.
3717 @hook TARGET_FRAME_POINTER_REQUIRED
3718 This target hook should return @code{true} if a function must have and use
3719 a frame pointer. This target hook is called in the reload pass. If its return
3720 value is @code{true} the function will have a frame pointer.
3722 This target hook can in principle examine the current function and decide
3723 according to the facts, but on most machines the constant @code{false} or the
3724 constant @code{true} suffices. Use @code{false} when the machine allows code
3725 to be generated with no frame pointer, and doing so saves some time or space.
3726 Use @code{true} when there is no possible advantage to avoiding a frame
3729 In certain cases, the compiler does not know how to produce valid code
3730 without a frame pointer. The compiler recognizes those cases and
3731 automatically gives the function a frame pointer regardless of what
3732 @code{TARGET_FRAME_POINTER_REQUIRED} returns. You don't need to worry about
3735 In a function that does not require a frame pointer, the frame pointer
3736 register can be allocated for ordinary usage, unless you mark it as a
3737 fixed register. See @code{FIXED_REGISTERS} for more information.
3739 Default return value is @code{false}.
3742 @findex get_frame_size
3743 @defmac INITIAL_FRAME_POINTER_OFFSET (@var{depth-var})
3744 A C statement to store in the variable @var{depth-var} the difference
3745 between the frame pointer and the stack pointer values immediately after
3746 the function prologue. The value would be computed from information
3747 such as the result of @code{get_frame_size ()} and the tables of
3748 registers @code{regs_ever_live} and @code{call_used_regs}.
3750 If @code{ELIMINABLE_REGS} is defined, this macro will be not be used and
3751 need not be defined. Otherwise, it must be defined even if
3752 @code{TARGET_FRAME_POINTER_REQUIRED} always returns true; in that
3753 case, you may set @var{depth-var} to anything.
3756 @defmac ELIMINABLE_REGS
3757 If defined, this macro specifies a table of register pairs used to
3758 eliminate unneeded registers that point into the stack frame. If it is not
3759 defined, the only elimination attempted by the compiler is to replace
3760 references to the frame pointer with references to the stack pointer.
3762 The definition of this macro is a list of structure initializations, each
3763 of which specifies an original and replacement register.
3765 On some machines, the position of the argument pointer is not known until
3766 the compilation is completed. In such a case, a separate hard register
3767 must be used for the argument pointer. This register can be eliminated by
3768 replacing it with either the frame pointer or the argument pointer,
3769 depending on whether or not the frame pointer has been eliminated.
3771 In this case, you might specify:
3773 #define ELIMINABLE_REGS \
3774 @{@{ARG_POINTER_REGNUM, STACK_POINTER_REGNUM@}, \
3775 @{ARG_POINTER_REGNUM, FRAME_POINTER_REGNUM@}, \
3776 @{FRAME_POINTER_REGNUM, STACK_POINTER_REGNUM@}@}
3779 Note that the elimination of the argument pointer with the stack pointer is
3780 specified first since that is the preferred elimination.
3783 @hook TARGET_CAN_ELIMINATE
3784 This target hook should returns @code{true} if the compiler is allowed to
3785 try to replace register number @var{from_reg} with register number
3786 @var{to_reg}. This target hook need only be defined if @code{ELIMINABLE_REGS}
3787 is defined, and will usually be @code{true}, since most of the cases
3788 preventing register elimination are things that the compiler already
3791 Default return value is @code{true}.
3794 @defmac INITIAL_ELIMINATION_OFFSET (@var{from-reg}, @var{to-reg}, @var{offset-var})
3795 This macro is similar to @code{INITIAL_FRAME_POINTER_OFFSET}. It
3796 specifies the initial difference between the specified pair of
3797 registers. This macro must be defined if @code{ELIMINABLE_REGS} is
3801 @node Stack Arguments
3802 @subsection Passing Function Arguments on the Stack
3803 @cindex arguments on stack
3804 @cindex stack arguments
3806 The macros in this section control how arguments are passed
3807 on the stack. See the following section for other macros that
3808 control passing certain arguments in registers.
3810 @hook TARGET_PROMOTE_PROTOTYPES
3811 This target hook returns @code{true} if an argument declared in a
3812 prototype as an integral type smaller than @code{int} should actually be
3813 passed as an @code{int}. In addition to avoiding errors in certain
3814 cases of mismatch, it also makes for better code on certain machines.
3815 The default is to not promote prototypes.
3819 A C expression. If nonzero, push insns will be used to pass
3821 If the target machine does not have a push instruction, set it to zero.
3822 That directs GCC to use an alternate strategy: to
3823 allocate the entire argument block and then store the arguments into
3824 it. When @code{PUSH_ARGS} is nonzero, @code{PUSH_ROUNDING} must be defined too.
3827 @defmac PUSH_ARGS_REVERSED
3828 A C expression. If nonzero, function arguments will be evaluated from
3829 last to first, rather than from first to last. If this macro is not
3830 defined, it defaults to @code{PUSH_ARGS} on targets where the stack
3831 and args grow in opposite directions, and 0 otherwise.
3834 @defmac PUSH_ROUNDING (@var{npushed})
3835 A C expression that is the number of bytes actually pushed onto the
3836 stack when an instruction attempts to push @var{npushed} bytes.
3838 On some machines, the definition
3841 #define PUSH_ROUNDING(BYTES) (BYTES)
3845 will suffice. But on other machines, instructions that appear
3846 to push one byte actually push two bytes in an attempt to maintain
3847 alignment. Then the definition should be
3850 #define PUSH_ROUNDING(BYTES) (((BYTES) + 1) & ~1)
3853 If the value of this macro has a type, it should be an unsigned type.
3856 @findex current_function_outgoing_args_size
3857 @defmac ACCUMULATE_OUTGOING_ARGS
3858 A C expression. If nonzero, the maximum amount of space required for outgoing arguments
3859 will be computed and placed into the variable
3860 @code{current_function_outgoing_args_size}. No space will be pushed
3861 onto the stack for each call; instead, the function prologue should
3862 increase the stack frame size by this amount.
3864 Setting both @code{PUSH_ARGS} and @code{ACCUMULATE_OUTGOING_ARGS}
3868 @defmac REG_PARM_STACK_SPACE (@var{fndecl})
3869 Define this macro if functions should assume that stack space has been
3870 allocated for arguments even when their values are passed in
3873 The value of this macro is the size, in bytes, of the area reserved for
3874 arguments passed in registers for the function represented by @var{fndecl},
3875 which can be zero if GCC is calling a library function.
3876 The argument @var{fndecl} can be the FUNCTION_DECL, or the type itself
3879 This space can be allocated by the caller, or be a part of the
3880 machine-dependent stack frame: @code{OUTGOING_REG_PARM_STACK_SPACE} says
3883 @c above is overfull. not sure what to do. --mew 5feb93 did
3884 @c something, not sure if it looks good. --mew 10feb93
3886 @defmac OUTGOING_REG_PARM_STACK_SPACE (@var{fntype})
3887 Define this to a nonzero value if it is the responsibility of the
3888 caller to allocate the area reserved for arguments passed in registers
3889 when calling a function of @var{fntype}. @var{fntype} may be NULL
3890 if the function called is a library function.
3892 If @code{ACCUMULATE_OUTGOING_ARGS} is defined, this macro controls
3893 whether the space for these arguments counts in the value of
3894 @code{current_function_outgoing_args_size}.
3897 @defmac STACK_PARMS_IN_REG_PARM_AREA
3898 Define this macro if @code{REG_PARM_STACK_SPACE} is defined, but the
3899 stack parameters don't skip the area specified by it.
3900 @c i changed this, makes more sens and it should have taken care of the
3901 @c overfull.. not as specific, tho. --mew 5feb93
3903 Normally, when a parameter is not passed in registers, it is placed on the
3904 stack beyond the @code{REG_PARM_STACK_SPACE} area. Defining this macro
3905 suppresses this behavior and causes the parameter to be passed on the
3906 stack in its natural location.
3909 @hook TARGET_RETURN_POPS_ARGS
3910 This target hook returns the number of bytes of its own arguments that
3911 a function pops on returning, or 0 if the function pops no arguments
3912 and the caller must therefore pop them all after the function returns.
3914 @var{fundecl} is a C variable whose value is a tree node that describes
3915 the function in question. Normally it is a node of type
3916 @code{FUNCTION_DECL} that describes the declaration of the function.
3917 From this you can obtain the @code{DECL_ATTRIBUTES} of the function.
3919 @var{funtype} is a C variable whose value is a tree node that
3920 describes the function in question. Normally it is a node of type
3921 @code{FUNCTION_TYPE} that describes the data type of the function.
3922 From this it is possible to obtain the data types of the value and
3923 arguments (if known).
3925 When a call to a library function is being considered, @var{fundecl}
3926 will contain an identifier node for the library function. Thus, if
3927 you need to distinguish among various library functions, you can do so
3928 by their names. Note that ``library function'' in this context means
3929 a function used to perform arithmetic, whose name is known specially
3930 in the compiler and was not mentioned in the C code being compiled.
3932 @var{size} is the number of bytes of arguments passed on the
3933 stack. If a variable number of bytes is passed, it is zero, and
3934 argument popping will always be the responsibility of the calling function.
3936 On the VAX, all functions always pop their arguments, so the definition
3937 of this macro is @var{size}. On the 68000, using the standard
3938 calling convention, no functions pop their arguments, so the value of
3939 the macro is always 0 in this case. But an alternative calling
3940 convention is available in which functions that take a fixed number of
3941 arguments pop them but other functions (such as @code{printf}) pop
3942 nothing (the caller pops all). When this convention is in use,
3943 @var{funtype} is examined to determine whether a function takes a fixed
3944 number of arguments.
3947 @defmac CALL_POPS_ARGS (@var{cum})
3948 A C expression that should indicate the number of bytes a call sequence
3949 pops off the stack. It is added to the value of @code{RETURN_POPS_ARGS}
3950 when compiling a function call.
3952 @var{cum} is the variable in which all arguments to the called function
3953 have been accumulated.
3955 On certain architectures, such as the SH5, a call trampoline is used
3956 that pops certain registers off the stack, depending on the arguments
3957 that have been passed to the function. Since this is a property of the
3958 call site, not of the called function, @code{RETURN_POPS_ARGS} is not
3962 @node Register Arguments
3963 @subsection Passing Arguments in Registers
3964 @cindex arguments in registers
3965 @cindex registers arguments
3967 This section describes the macros which let you control how various
3968 types of arguments are passed in registers or how they are arranged in
3971 @hook TARGET_FUNCTION_ARG
3972 Return an RTX indicating whether a function argument is passed in a
3973 register and if so, which register.
3975 The arguments are @var{ca}, which summarizes all the previous
3976 arguments; @var{mode}, the machine mode of the argument; @var{type},
3977 the data type of the argument as a tree node or 0 if that is not known
3978 (which happens for C support library functions); and @var{named},
3979 which is @code{true} for an ordinary argument and @code{false} for
3980 nameless arguments that correspond to @samp{@dots{}} in the called
3981 function's prototype. @var{type} can be an incomplete type if a
3982 syntax error has previously occurred.
3984 The return value is usually either a @code{reg} RTX for the hard
3985 register in which to pass the argument, or zero to pass the argument
3988 The value of the expression can also be a @code{parallel} RTX@. This is
3989 used when an argument is passed in multiple locations. The mode of the
3990 @code{parallel} should be the mode of the entire argument. The
3991 @code{parallel} holds any number of @code{expr_list} pairs; each one
3992 describes where part of the argument is passed. In each
3993 @code{expr_list} the first operand must be a @code{reg} RTX for the hard
3994 register in which to pass this part of the argument, and the mode of the
3995 register RTX indicates how large this part of the argument is. The
3996 second operand of the @code{expr_list} is a @code{const_int} which gives
3997 the offset in bytes into the entire argument of where this part starts.
3998 As a special exception the first @code{expr_list} in the @code{parallel}
3999 RTX may have a first operand of zero. This indicates that the entire
4000 argument is also stored on the stack.
4002 The last time this hook is called, it is called with @code{MODE ==
4003 VOIDmode}, and its result is passed to the @code{call} or @code{call_value}
4004 pattern as operands 2 and 3 respectively.
4006 @cindex @file{stdarg.h} and register arguments
4007 The usual way to make the ISO library @file{stdarg.h} work on a
4008 machine where some arguments are usually passed in registers, is to
4009 cause nameless arguments to be passed on the stack instead. This is
4010 done by making @code{TARGET_FUNCTION_ARG} return 0 whenever
4011 @var{named} is @code{false}.
4013 @cindex @code{TARGET_MUST_PASS_IN_STACK}, and @code{TARGET_FUNCTION_ARG}
4014 @cindex @code{REG_PARM_STACK_SPACE}, and @code{TARGET_FUNCTION_ARG}
4015 You may use the hook @code{targetm.calls.must_pass_in_stack}
4016 in the definition of this macro to determine if this argument is of a
4017 type that must be passed in the stack. If @code{REG_PARM_STACK_SPACE}
4018 is not defined and @code{TARGET_FUNCTION_ARG} returns nonzero for such an
4019 argument, the compiler will abort. If @code{REG_PARM_STACK_SPACE} is
4020 defined, the argument will be computed in the stack and then loaded into
4024 @hook TARGET_MUST_PASS_IN_STACK
4025 This target hook should return @code{true} if we should not pass @var{type}
4026 solely in registers. The file @file{expr.h} defines a
4027 definition that is usually appropriate, refer to @file{expr.h} for additional
4031 @hook TARGET_FUNCTION_INCOMING_ARG
4032 Define this hook if the target machine has ``register windows'', so
4033 that the register in which a function sees an arguments is not
4034 necessarily the same as the one in which the caller passed the
4037 For such machines, @code{TARGET_FUNCTION_ARG} computes the register in
4038 which the caller passes the value, and
4039 @code{TARGET_FUNCTION_INCOMING_ARG} should be defined in a similar
4040 fashion to tell the function being called where the arguments will
4043 If @code{TARGET_FUNCTION_INCOMING_ARG} is not defined,
4044 @code{TARGET_FUNCTION_ARG} serves both purposes.
4047 @hook TARGET_ARG_PARTIAL_BYTES
4048 This target hook returns the number of bytes at the beginning of an
4049 argument that must be put in registers. The value must be zero for
4050 arguments that are passed entirely in registers or that are entirely
4051 pushed on the stack.
4053 On some machines, certain arguments must be passed partially in
4054 registers and partially in memory. On these machines, typically the
4055 first few words of arguments are passed in registers, and the rest
4056 on the stack. If a multi-word argument (a @code{double} or a
4057 structure) crosses that boundary, its first few words must be passed
4058 in registers and the rest must be pushed. This macro tells the
4059 compiler when this occurs, and how many bytes should go in registers.
4061 @code{TARGET_FUNCTION_ARG} for these arguments should return the first
4062 register to be used by the caller for this argument; likewise
4063 @code{TARGET_FUNCTION_INCOMING_ARG}, for the called function.
4066 @hook TARGET_PASS_BY_REFERENCE
4067 This target hook should return @code{true} if an argument at the
4068 position indicated by @var{cum} should be passed by reference. This
4069 predicate is queried after target independent reasons for being
4070 passed by reference, such as @code{TREE_ADDRESSABLE (type)}.
4072 If the hook returns true, a copy of that argument is made in memory and a
4073 pointer to the argument is passed instead of the argument itself.
4074 The pointer is passed in whatever way is appropriate for passing a pointer
4078 @hook TARGET_CALLEE_COPIES
4079 The function argument described by the parameters to this hook is
4080 known to be passed by reference. The hook should return true if the
4081 function argument should be copied by the callee instead of copied
4084 For any argument for which the hook returns true, if it can be
4085 determined that the argument is not modified, then a copy need
4088 The default version of this hook always returns false.
4091 @defmac CUMULATIVE_ARGS
4092 A C type for declaring a variable that is used as the first argument
4093 of @code{TARGET_FUNCTION_ARG} and other related values. For some
4094 target machines, the type @code{int} suffices and can hold the number
4095 of bytes of argument so far.
4097 There is no need to record in @code{CUMULATIVE_ARGS} anything about the
4098 arguments that have been passed on the stack. The compiler has other
4099 variables to keep track of that. For target machines on which all
4100 arguments are passed on the stack, there is no need to store anything in
4101 @code{CUMULATIVE_ARGS}; however, the data structure must exist and
4102 should not be empty, so use @code{int}.
4105 @defmac OVERRIDE_ABI_FORMAT (@var{fndecl})
4106 If defined, this macro is called before generating any code for a
4107 function, but after the @var{cfun} descriptor for the function has been
4108 created. The back end may use this macro to update @var{cfun} to
4109 reflect an ABI other than that which would normally be used by default.
4110 If the compiler is generating code for a compiler-generated function,
4111 @var{fndecl} may be @code{NULL}.
4114 @defmac INIT_CUMULATIVE_ARGS (@var{cum}, @var{fntype}, @var{libname}, @var{fndecl}, @var{n_named_args})
4115 A C statement (sans semicolon) for initializing the variable
4116 @var{cum} for the state at the beginning of the argument list. The
4117 variable has type @code{CUMULATIVE_ARGS}. The value of @var{fntype}
4118 is the tree node for the data type of the function which will receive
4119 the args, or 0 if the args are to a compiler support library function.
4120 For direct calls that are not libcalls, @var{fndecl} contain the
4121 declaration node of the function. @var{fndecl} is also set when
4122 @code{INIT_CUMULATIVE_ARGS} is used to find arguments for the function
4123 being compiled. @var{n_named_args} is set to the number of named
4124 arguments, including a structure return address if it is passed as a
4125 parameter, when making a call. When processing incoming arguments,
4126 @var{n_named_args} is set to @minus{}1.
4128 When processing a call to a compiler support library function,
4129 @var{libname} identifies which one. It is a @code{symbol_ref} rtx which
4130 contains the name of the function, as a string. @var{libname} is 0 when
4131 an ordinary C function call is being processed. Thus, each time this
4132 macro is called, either @var{libname} or @var{fntype} is nonzero, but
4133 never both of them at once.
4136 @defmac INIT_CUMULATIVE_LIBCALL_ARGS (@var{cum}, @var{mode}, @var{libname})
4137 Like @code{INIT_CUMULATIVE_ARGS} but only used for outgoing libcalls,
4138 it gets a @code{MODE} argument instead of @var{fntype}, that would be
4139 @code{NULL}. @var{indirect} would always be zero, too. If this macro
4140 is not defined, @code{INIT_CUMULATIVE_ARGS (cum, NULL_RTX, libname,
4141 0)} is used instead.
4144 @defmac INIT_CUMULATIVE_INCOMING_ARGS (@var{cum}, @var{fntype}, @var{libname})
4145 Like @code{INIT_CUMULATIVE_ARGS} but overrides it for the purposes of
4146 finding the arguments for the function being compiled. If this macro is
4147 undefined, @code{INIT_CUMULATIVE_ARGS} is used instead.
4149 The value passed for @var{libname} is always 0, since library routines
4150 with special calling conventions are never compiled with GCC@. The
4151 argument @var{libname} exists for symmetry with
4152 @code{INIT_CUMULATIVE_ARGS}.
4153 @c could use "this macro" in place of @code{INIT_CUMULATIVE_ARGS}, maybe.
4154 @c --mew 5feb93 i switched the order of the sentences. --mew 10feb93
4157 @hook TARGET_FUNCTION_ARG_ADVANCE
4158 This hook updates the summarizer variable pointed to by @var{ca} to
4159 advance past an argument in the argument list. The values @var{mode},
4160 @var{type} and @var{named} describe that argument. Once this is done,
4161 the variable @var{cum} is suitable for analyzing the @emph{following}
4162 argument with @code{TARGET_FUNCTION_ARG}, etc.
4164 This hook need not do anything if the argument in question was passed
4165 on the stack. The compiler knows how to track the amount of stack space
4166 used for arguments without any special help.
4169 @defmac FUNCTION_ARG_OFFSET (@var{mode}, @var{type})
4170 If defined, a C expression that is the number of bytes to add to the
4171 offset of the argument passed in memory. This is needed for the SPU,
4172 which passes @code{char} and @code{short} arguments in the preferred
4173 slot that is in the middle of the quad word instead of starting at the
4177 @defmac FUNCTION_ARG_PADDING (@var{mode}, @var{type})
4178 If defined, a C expression which determines whether, and in which direction,
4179 to pad out an argument with extra space. The value should be of type
4180 @code{enum direction}: either @code{upward} to pad above the argument,
4181 @code{downward} to pad below, or @code{none} to inhibit padding.
4183 The @emph{amount} of padding is not controlled by this macro, but by the
4184 target hook @code{TARGET_FUNCTION_ARG_ROUND_BOUNDARY}. It is
4185 always just enough to reach the next multiple of that boundary.
4187 This macro has a default definition which is right for most systems.
4188 For little-endian machines, the default is to pad upward. For
4189 big-endian machines, the default is to pad downward for an argument of
4190 constant size shorter than an @code{int}, and upward otherwise.
4193 @defmac PAD_VARARGS_DOWN
4194 If defined, a C expression which determines whether the default
4195 implementation of va_arg will attempt to pad down before reading the
4196 next argument, if that argument is smaller than its aligned space as
4197 controlled by @code{PARM_BOUNDARY}. If this macro is not defined, all such
4198 arguments are padded down if @code{BYTES_BIG_ENDIAN} is true.
4201 @defmac BLOCK_REG_PADDING (@var{mode}, @var{type}, @var{first})
4202 Specify padding for the last element of a block move between registers and
4203 memory. @var{first} is nonzero if this is the only element. Defining this
4204 macro allows better control of register function parameters on big-endian
4205 machines, without using @code{PARALLEL} rtl. In particular,
4206 @code{MUST_PASS_IN_STACK} need not test padding and mode of types in
4207 registers, as there is no longer a "wrong" part of a register; For example,
4208 a three byte aggregate may be passed in the high part of a register if so
4212 @hook TARGET_FUNCTION_ARG_BOUNDARY
4213 This hook returns the alignment boundary, in bits, of an argument
4214 with the specified mode and type. The default hook returns
4215 @code{PARM_BOUNDARY} for all arguments.
4218 @hook TARGET_FUNCTION_ARG_ROUND_BOUNDARY
4220 @defmac FUNCTION_ARG_REGNO_P (@var{regno})
4221 A C expression that is nonzero if @var{regno} is the number of a hard
4222 register in which function arguments are sometimes passed. This does
4223 @emph{not} include implicit arguments such as the static chain and
4224 the structure-value address. On many machines, no registers can be
4225 used for this purpose since all function arguments are pushed on the
4229 @hook TARGET_SPLIT_COMPLEX_ARG
4230 This hook should return true if parameter of type @var{type} are passed
4231 as two scalar parameters. By default, GCC will attempt to pack complex
4232 arguments into the target's word size. Some ABIs require complex arguments
4233 to be split and treated as their individual components. For example, on
4234 AIX64, complex floats should be passed in a pair of floating point
4235 registers, even though a complex float would fit in one 64-bit floating
4238 The default value of this hook is @code{NULL}, which is treated as always
4242 @hook TARGET_BUILD_BUILTIN_VA_LIST
4243 This hook returns a type node for @code{va_list} for the target.
4244 The default version of the hook returns @code{void*}.
4247 @hook TARGET_ENUM_VA_LIST_P
4248 This target hook is used in function @code{c_common_nodes_and_builtins}
4249 to iterate through the target specific builtin types for va_list. The
4250 variable @var{idx} is used as iterator. @var{pname} has to be a pointer
4251 to a @code{const char *} and @var{ptree} a pointer to a @code{tree} typed
4253 The arguments @var{pname} and @var{ptree} are used to store the result of
4254 this macro and are set to the name of the va_list builtin type and its
4256 If the return value of this macro is zero, then there is no more element.
4257 Otherwise the @var{IDX} should be increased for the next call of this
4258 macro to iterate through all types.
4261 @hook TARGET_FN_ABI_VA_LIST
4262 This hook returns the va_list type of the calling convention specified by
4264 The default version of this hook returns @code{va_list_type_node}.
4267 @hook TARGET_CANONICAL_VA_LIST_TYPE
4268 This hook returns the va_list type of the calling convention specified by the
4269 type of @var{type}. If @var{type} is not a valid va_list type, it returns
4273 @hook TARGET_GIMPLIFY_VA_ARG_EXPR
4274 This hook performs target-specific gimplification of
4275 @code{VA_ARG_EXPR}. The first two parameters correspond to the
4276 arguments to @code{va_arg}; the latter two are as in
4277 @code{gimplify.c:gimplify_expr}.
4280 @hook TARGET_VALID_POINTER_MODE
4281 Define this to return nonzero if the port can handle pointers
4282 with machine mode @var{mode}. The default version of this
4283 hook returns true for both @code{ptr_mode} and @code{Pmode}.
4286 @hook TARGET_REF_MAY_ALIAS_ERRNO
4288 @hook TARGET_SCALAR_MODE_SUPPORTED_P
4289 Define this to return nonzero if the port is prepared to handle
4290 insns involving scalar mode @var{mode}. For a scalar mode to be
4291 considered supported, all the basic arithmetic and comparisons
4294 The default version of this hook returns true for any mode
4295 required to handle the basic C types (as defined by the port).
4296 Included here are the double-word arithmetic supported by the
4297 code in @file{optabs.c}.
4300 @hook TARGET_VECTOR_MODE_SUPPORTED_P
4301 Define this to return nonzero if the port is prepared to handle
4302 insns involving vector mode @var{mode}. At the very least, it
4303 must have move patterns for this mode.
4306 @hook TARGET_ARRAY_MODE_SUPPORTED_P
4308 @hook TARGET_SMALL_REGISTER_CLASSES_FOR_MODE_P
4309 Define this to return nonzero for machine modes for which the port has
4310 small register classes. If this target hook returns nonzero for a given
4311 @var{mode}, the compiler will try to minimize the lifetime of registers
4312 in @var{mode}. The hook may be called with @code{VOIDmode} as argument.
4313 In this case, the hook is expected to return nonzero if it returns nonzero
4316 On some machines, it is risky to let hard registers live across arbitrary
4317 insns. Typically, these machines have instructions that require values
4318 to be in specific registers (like an accumulator), and reload will fail
4319 if the required hard register is used for another purpose across such an
4322 Passes before reload do not know which hard registers will be used
4323 in an instruction, but the machine modes of the registers set or used in
4324 the instruction are already known. And for some machines, register
4325 classes are small for, say, integer registers but not for floating point
4326 registers. For example, the AMD x86-64 architecture requires specific
4327 registers for the legacy x86 integer instructions, but there are many
4328 SSE registers for floating point operations. On such targets, a good
4329 strategy may be to return nonzero from this hook for @code{INTEGRAL_MODE_P}
4330 machine modes but zero for the SSE register classes.
4332 The default version of this hook returns false for any mode. It is always
4333 safe to redefine this hook to return with a nonzero value. But if you
4334 unnecessarily define it, you will reduce the amount of optimizations
4335 that can be performed in some cases. If you do not define this hook
4336 to return a nonzero value when it is required, the compiler will run out
4337 of spill registers and print a fatal error message.
4340 @hook TARGET_FLAGS_REGNUM
4343 @subsection How Scalar Function Values Are Returned
4344 @cindex return values in registers
4345 @cindex values, returned by functions
4346 @cindex scalars, returned as values
4348 This section discusses the macros that control returning scalars as
4349 values---values that can fit in registers.
4351 @hook TARGET_FUNCTION_VALUE
4353 Define this to return an RTX representing the place where a function
4354 returns or receives a value of data type @var{ret_type}, a tree node
4355 representing a data type. @var{fn_decl_or_type} is a tree node
4356 representing @code{FUNCTION_DECL} or @code{FUNCTION_TYPE} of a
4357 function being called. If @var{outgoing} is false, the hook should
4358 compute the register in which the caller will see the return value.
4359 Otherwise, the hook should return an RTX representing the place where
4360 a function returns a value.
4362 On many machines, only @code{TYPE_MODE (@var{ret_type})} is relevant.
4363 (Actually, on most machines, scalar values are returned in the same
4364 place regardless of mode.) The value of the expression is usually a
4365 @code{reg} RTX for the hard register where the return value is stored.
4366 The value can also be a @code{parallel} RTX, if the return value is in
4367 multiple places. See @code{TARGET_FUNCTION_ARG} for an explanation of the
4368 @code{parallel} form. Note that the callee will populate every
4369 location specified in the @code{parallel}, but if the first element of
4370 the @code{parallel} contains the whole return value, callers will use
4371 that element as the canonical location and ignore the others. The m68k
4372 port uses this type of @code{parallel} to return pointers in both
4373 @samp{%a0} (the canonical location) and @samp{%d0}.
4375 If @code{TARGET_PROMOTE_FUNCTION_RETURN} returns true, you must apply
4376 the same promotion rules specified in @code{PROMOTE_MODE} if
4377 @var{valtype} is a scalar type.
4379 If the precise function being called is known, @var{func} is a tree
4380 node (@code{FUNCTION_DECL}) for it; otherwise, @var{func} is a null
4381 pointer. This makes it possible to use a different value-returning
4382 convention for specific functions when all their calls are
4385 Some target machines have ``register windows'' so that the register in
4386 which a function returns its value is not the same as the one in which
4387 the caller sees the value. For such machines, you should return
4388 different RTX depending on @var{outgoing}.
4390 @code{TARGET_FUNCTION_VALUE} is not used for return values with
4391 aggregate data types, because these are returned in another way. See
4392 @code{TARGET_STRUCT_VALUE_RTX} and related macros, below.
4395 @defmac FUNCTION_VALUE (@var{valtype}, @var{func})
4396 This macro has been deprecated. Use @code{TARGET_FUNCTION_VALUE} for
4397 a new target instead.
4400 @defmac LIBCALL_VALUE (@var{mode})
4401 A C expression to create an RTX representing the place where a library
4402 function returns a value of mode @var{mode}.
4404 Note that ``library function'' in this context means a compiler
4405 support routine, used to perform arithmetic, whose name is known
4406 specially by the compiler and was not mentioned in the C code being
4410 @hook TARGET_LIBCALL_VALUE
4411 Define this hook if the back-end needs to know the name of the libcall
4412 function in order to determine where the result should be returned.
4414 The mode of the result is given by @var{mode} and the name of the called
4415 library function is given by @var{fun}. The hook should return an RTX
4416 representing the place where the library function result will be returned.
4418 If this hook is not defined, then LIBCALL_VALUE will be used.
4421 @defmac FUNCTION_VALUE_REGNO_P (@var{regno})
4422 A C expression that is nonzero if @var{regno} is the number of a hard
4423 register in which the values of called function may come back.
4425 A register whose use for returning values is limited to serving as the
4426 second of a pair (for a value of type @code{double}, say) need not be
4427 recognized by this macro. So for most machines, this definition
4431 #define FUNCTION_VALUE_REGNO_P(N) ((N) == 0)
4434 If the machine has register windows, so that the caller and the called
4435 function use different registers for the return value, this macro
4436 should recognize only the caller's register numbers.
4438 This macro has been deprecated. Use @code{TARGET_FUNCTION_VALUE_REGNO_P}
4439 for a new target instead.
4442 @hook TARGET_FUNCTION_VALUE_REGNO_P
4443 A target hook that return @code{true} if @var{regno} is the number of a hard
4444 register in which the values of called function may come back.
4446 A register whose use for returning values is limited to serving as the
4447 second of a pair (for a value of type @code{double}, say) need not be
4448 recognized by this target hook.
4450 If the machine has register windows, so that the caller and the called
4451 function use different registers for the return value, this target hook
4452 should recognize only the caller's register numbers.
4454 If this hook is not defined, then FUNCTION_VALUE_REGNO_P will be used.
4457 @defmac APPLY_RESULT_SIZE
4458 Define this macro if @samp{untyped_call} and @samp{untyped_return}
4459 need more space than is implied by @code{FUNCTION_VALUE_REGNO_P} for
4460 saving and restoring an arbitrary return value.
4463 @hook TARGET_RETURN_IN_MSB
4464 This hook should return true if values of type @var{type} are returned
4465 at the most significant end of a register (in other words, if they are
4466 padded at the least significant end). You can assume that @var{type}
4467 is returned in a register; the caller is required to check this.
4469 Note that the register provided by @code{TARGET_FUNCTION_VALUE} must
4470 be able to hold the complete return value. For example, if a 1-, 2-
4471 or 3-byte structure is returned at the most significant end of a
4472 4-byte register, @code{TARGET_FUNCTION_VALUE} should provide an
4476 @node Aggregate Return
4477 @subsection How Large Values Are Returned
4478 @cindex aggregates as return values
4479 @cindex large return values
4480 @cindex returning aggregate values
4481 @cindex structure value address
4483 When a function value's mode is @code{BLKmode} (and in some other
4484 cases), the value is not returned according to
4485 @code{TARGET_FUNCTION_VALUE} (@pxref{Scalar Return}). Instead, the
4486 caller passes the address of a block of memory in which the value
4487 should be stored. This address is called the @dfn{structure value
4490 This section describes how to control returning structure values in
4493 @hook TARGET_RETURN_IN_MEMORY
4494 This target hook should return a nonzero value to say to return the
4495 function value in memory, just as large structures are always returned.
4496 Here @var{type} will be the data type of the value, and @var{fntype}
4497 will be the type of the function doing the returning, or @code{NULL} for
4500 Note that values of mode @code{BLKmode} must be explicitly handled
4501 by this function. Also, the option @option{-fpcc-struct-return}
4502 takes effect regardless of this macro. On most systems, it is
4503 possible to leave the hook undefined; this causes a default
4504 definition to be used, whose value is the constant 1 for @code{BLKmode}
4505 values, and 0 otherwise.
4507 Do not use this hook to indicate that structures and unions should always
4508 be returned in memory. You should instead use @code{DEFAULT_PCC_STRUCT_RETURN}
4512 @defmac DEFAULT_PCC_STRUCT_RETURN
4513 Define this macro to be 1 if all structure and union return values must be
4514 in memory. Since this results in slower code, this should be defined
4515 only if needed for compatibility with other compilers or with an ABI@.
4516 If you define this macro to be 0, then the conventions used for structure
4517 and union return values are decided by the @code{TARGET_RETURN_IN_MEMORY}
4520 If not defined, this defaults to the value 1.
4523 @hook TARGET_STRUCT_VALUE_RTX
4524 This target hook should return the location of the structure value
4525 address (normally a @code{mem} or @code{reg}), or 0 if the address is
4526 passed as an ``invisible'' first argument. Note that @var{fndecl} may
4527 be @code{NULL}, for libcalls. You do not need to define this target
4528 hook if the address is always passed as an ``invisible'' first
4531 On some architectures the place where the structure value address
4532 is found by the called function is not the same place that the
4533 caller put it. This can be due to register windows, or it could
4534 be because the function prologue moves it to a different place.
4535 @var{incoming} is @code{1} or @code{2} when the location is needed in
4536 the context of the called function, and @code{0} in the context of
4539 If @var{incoming} is nonzero and the address is to be found on the
4540 stack, return a @code{mem} which refers to the frame pointer. If
4541 @var{incoming} is @code{2}, the result is being used to fetch the
4542 structure value address at the beginning of a function. If you need
4543 to emit adjusting code, you should do it at this point.
4546 @defmac PCC_STATIC_STRUCT_RETURN
4547 Define this macro if the usual system convention on the target machine
4548 for returning structures and unions is for the called function to return
4549 the address of a static variable containing the value.
4551 Do not define this if the usual system convention is for the caller to
4552 pass an address to the subroutine.
4554 This macro has effect in @option{-fpcc-struct-return} mode, but it does
4555 nothing when you use @option{-freg-struct-return} mode.
4558 @hook TARGET_GET_RAW_RESULT_MODE
4560 @hook TARGET_GET_RAW_ARG_MODE
4563 @subsection Caller-Saves Register Allocation
4565 If you enable it, GCC can save registers around function calls. This
4566 makes it possible to use call-clobbered registers to hold variables that
4567 must live across calls.
4569 @defmac CALLER_SAVE_PROFITABLE (@var{refs}, @var{calls})
4570 A C expression to determine whether it is worthwhile to consider placing
4571 a pseudo-register in a call-clobbered hard register and saving and
4572 restoring it around each function call. The expression should be 1 when
4573 this is worth doing, and 0 otherwise.
4575 If you don't define this macro, a default is used which is good on most
4576 machines: @code{4 * @var{calls} < @var{refs}}.
4579 @defmac HARD_REGNO_CALLER_SAVE_MODE (@var{regno}, @var{nregs})
4580 A C expression specifying which mode is required for saving @var{nregs}
4581 of a pseudo-register in call-clobbered hard register @var{regno}. If
4582 @var{regno} is unsuitable for caller save, @code{VOIDmode} should be
4583 returned. For most machines this macro need not be defined since GCC
4584 will select the smallest suitable mode.
4587 @node Function Entry
4588 @subsection Function Entry and Exit
4589 @cindex function entry and exit
4593 This section describes the macros that output function entry
4594 (@dfn{prologue}) and exit (@dfn{epilogue}) code.
4596 @hook TARGET_ASM_FUNCTION_PROLOGUE
4597 If defined, a function that outputs the assembler code for entry to a
4598 function. The prologue is responsible for setting up the stack frame,
4599 initializing the frame pointer register, saving registers that must be
4600 saved, and allocating @var{size} additional bytes of storage for the
4601 local variables. @var{size} is an integer. @var{file} is a stdio
4602 stream to which the assembler code should be output.
4604 The label for the beginning of the function need not be output by this
4605 macro. That has already been done when the macro is run.
4607 @findex regs_ever_live
4608 To determine which registers to save, the macro can refer to the array
4609 @code{regs_ever_live}: element @var{r} is nonzero if hard register
4610 @var{r} is used anywhere within the function. This implies the function
4611 prologue should save register @var{r}, provided it is not one of the
4612 call-used registers. (@code{TARGET_ASM_FUNCTION_EPILOGUE} must likewise use
4613 @code{regs_ever_live}.)
4615 On machines that have ``register windows'', the function entry code does
4616 not save on the stack the registers that are in the windows, even if
4617 they are supposed to be preserved by function calls; instead it takes
4618 appropriate steps to ``push'' the register stack, if any non-call-used
4619 registers are used in the function.
4621 @findex frame_pointer_needed
4622 On machines where functions may or may not have frame-pointers, the
4623 function entry code must vary accordingly; it must set up the frame
4624 pointer if one is wanted, and not otherwise. To determine whether a
4625 frame pointer is in wanted, the macro can refer to the variable
4626 @code{frame_pointer_needed}. The variable's value will be 1 at run
4627 time in a function that needs a frame pointer. @xref{Elimination}.
4629 The function entry code is responsible for allocating any stack space
4630 required for the function. This stack space consists of the regions
4631 listed below. In most cases, these regions are allocated in the
4632 order listed, with the last listed region closest to the top of the
4633 stack (the lowest address if @code{STACK_GROWS_DOWNWARD} is defined, and
4634 the highest address if it is not defined). You can use a different order
4635 for a machine if doing so is more convenient or required for
4636 compatibility reasons. Except in cases where required by standard
4637 or by a debugger, there is no reason why the stack layout used by GCC
4638 need agree with that used by other compilers for a machine.
4641 @hook TARGET_ASM_FUNCTION_END_PROLOGUE
4642 If defined, a function that outputs assembler code at the end of a
4643 prologue. This should be used when the function prologue is being
4644 emitted as RTL, and you have some extra assembler that needs to be
4645 emitted. @xref{prologue instruction pattern}.
4648 @hook TARGET_ASM_FUNCTION_BEGIN_EPILOGUE
4649 If defined, a function that outputs assembler code at the start of an
4650 epilogue. This should be used when the function epilogue is being
4651 emitted as RTL, and you have some extra assembler that needs to be
4652 emitted. @xref{epilogue instruction pattern}.
4655 @hook TARGET_ASM_FUNCTION_EPILOGUE
4656 If defined, a function that outputs the assembler code for exit from a
4657 function. The epilogue is responsible for restoring the saved
4658 registers and stack pointer to their values when the function was
4659 called, and returning control to the caller. This macro takes the
4660 same arguments as the macro @code{TARGET_ASM_FUNCTION_PROLOGUE}, and the
4661 registers to restore are determined from @code{regs_ever_live} and
4662 @code{CALL_USED_REGISTERS} in the same way.
4664 On some machines, there is a single instruction that does all the work
4665 of returning from the function. On these machines, give that
4666 instruction the name @samp{return} and do not define the macro
4667 @code{TARGET_ASM_FUNCTION_EPILOGUE} at all.
4669 Do not define a pattern named @samp{return} if you want the
4670 @code{TARGET_ASM_FUNCTION_EPILOGUE} to be used. If you want the target
4671 switches to control whether return instructions or epilogues are used,
4672 define a @samp{return} pattern with a validity condition that tests the
4673 target switches appropriately. If the @samp{return} pattern's validity
4674 condition is false, epilogues will be used.
4676 On machines where functions may or may not have frame-pointers, the
4677 function exit code must vary accordingly. Sometimes the code for these
4678 two cases is completely different. To determine whether a frame pointer
4679 is wanted, the macro can refer to the variable
4680 @code{frame_pointer_needed}. The variable's value will be 1 when compiling
4681 a function that needs a frame pointer.
4683 Normally, @code{TARGET_ASM_FUNCTION_PROLOGUE} and
4684 @code{TARGET_ASM_FUNCTION_EPILOGUE} must treat leaf functions specially.
4685 The C variable @code{current_function_is_leaf} is nonzero for such a
4686 function. @xref{Leaf Functions}.
4688 On some machines, some functions pop their arguments on exit while
4689 others leave that for the caller to do. For example, the 68020 when
4690 given @option{-mrtd} pops arguments in functions that take a fixed
4691 number of arguments.
4693 @findex current_function_pops_args
4694 Your definition of the macro @code{RETURN_POPS_ARGS} decides which
4695 functions pop their own arguments. @code{TARGET_ASM_FUNCTION_EPILOGUE}
4696 needs to know what was decided. The number of bytes of the current
4697 function's arguments that this function should pop is available in
4698 @code{crtl->args.pops_args}. @xref{Scalar Return}.
4703 @findex current_function_pretend_args_size
4704 A region of @code{current_function_pretend_args_size} bytes of
4705 uninitialized space just underneath the first argument arriving on the
4706 stack. (This may not be at the very start of the allocated stack region
4707 if the calling sequence has pushed anything else since pushing the stack
4708 arguments. But usually, on such machines, nothing else has been pushed
4709 yet, because the function prologue itself does all the pushing.) This
4710 region is used on machines where an argument may be passed partly in
4711 registers and partly in memory, and, in some cases to support the
4712 features in @code{<stdarg.h>}.
4715 An area of memory used to save certain registers used by the function.
4716 The size of this area, which may also include space for such things as
4717 the return address and pointers to previous stack frames, is
4718 machine-specific and usually depends on which registers have been used
4719 in the function. Machines with register windows often do not require
4723 A region of at least @var{size} bytes, possibly rounded up to an allocation
4724 boundary, to contain the local variables of the function. On some machines,
4725 this region and the save area may occur in the opposite order, with the
4726 save area closer to the top of the stack.
4729 @cindex @code{ACCUMULATE_OUTGOING_ARGS} and stack frames
4730 Optionally, when @code{ACCUMULATE_OUTGOING_ARGS} is defined, a region of
4731 @code{current_function_outgoing_args_size} bytes to be used for outgoing
4732 argument lists of the function. @xref{Stack Arguments}.
4735 @defmac EXIT_IGNORE_STACK
4736 Define this macro as a C expression that is nonzero if the return
4737 instruction or the function epilogue ignores the value of the stack
4738 pointer; in other words, if it is safe to delete an instruction to
4739 adjust the stack pointer before a return from the function. The
4742 Note that this macro's value is relevant only for functions for which
4743 frame pointers are maintained. It is never safe to delete a final
4744 stack adjustment in a function that has no frame pointer, and the
4745 compiler knows this regardless of @code{EXIT_IGNORE_STACK}.
4748 @defmac EPILOGUE_USES (@var{regno})
4749 Define this macro as a C expression that is nonzero for registers that are
4750 used by the epilogue or the @samp{return} pattern. The stack and frame
4751 pointer registers are already assumed to be used as needed.
4754 @defmac EH_USES (@var{regno})
4755 Define this macro as a C expression that is nonzero for registers that are
4756 used by the exception handling mechanism, and so should be considered live
4757 on entry to an exception edge.
4760 @defmac DELAY_SLOTS_FOR_EPILOGUE
4761 Define this macro if the function epilogue contains delay slots to which
4762 instructions from the rest of the function can be ``moved''. The
4763 definition should be a C expression whose value is an integer
4764 representing the number of delay slots there.
4767 @defmac ELIGIBLE_FOR_EPILOGUE_DELAY (@var{insn}, @var{n})
4768 A C expression that returns 1 if @var{insn} can be placed in delay
4769 slot number @var{n} of the epilogue.
4771 The argument @var{n} is an integer which identifies the delay slot now
4772 being considered (since different slots may have different rules of
4773 eligibility). It is never negative and is always less than the number
4774 of epilogue delay slots (what @code{DELAY_SLOTS_FOR_EPILOGUE} returns).
4775 If you reject a particular insn for a given delay slot, in principle, it
4776 may be reconsidered for a subsequent delay slot. Also, other insns may
4777 (at least in principle) be considered for the so far unfilled delay
4780 @findex current_function_epilogue_delay_list
4781 @findex final_scan_insn
4782 The insns accepted to fill the epilogue delay slots are put in an RTL
4783 list made with @code{insn_list} objects, stored in the variable
4784 @code{current_function_epilogue_delay_list}. The insn for the first
4785 delay slot comes first in the list. Your definition of the macro
4786 @code{TARGET_ASM_FUNCTION_EPILOGUE} should fill the delay slots by
4787 outputting the insns in this list, usually by calling
4788 @code{final_scan_insn}.
4790 You need not define this macro if you did not define
4791 @code{DELAY_SLOTS_FOR_EPILOGUE}.
4794 @hook TARGET_ASM_OUTPUT_MI_THUNK
4795 A function that outputs the assembler code for a thunk
4796 function, used to implement C++ virtual function calls with multiple
4797 inheritance. The thunk acts as a wrapper around a virtual function,
4798 adjusting the implicit object parameter before handing control off to
4801 First, emit code to add the integer @var{delta} to the location that
4802 contains the incoming first argument. Assume that this argument
4803 contains a pointer, and is the one used to pass the @code{this} pointer
4804 in C++. This is the incoming argument @emph{before} the function prologue,
4805 e.g.@: @samp{%o0} on a sparc. The addition must preserve the values of
4806 all other incoming arguments.
4808 Then, if @var{vcall_offset} is nonzero, an additional adjustment should be
4809 made after adding @code{delta}. In particular, if @var{p} is the
4810 adjusted pointer, the following adjustment should be made:
4813 p += (*((ptrdiff_t **)p))[vcall_offset/sizeof(ptrdiff_t)]
4816 After the additions, emit code to jump to @var{function}, which is a
4817 @code{FUNCTION_DECL}. This is a direct pure jump, not a call, and does
4818 not touch the return address. Hence returning from @var{FUNCTION} will
4819 return to whoever called the current @samp{thunk}.
4821 The effect must be as if @var{function} had been called directly with
4822 the adjusted first argument. This macro is responsible for emitting all
4823 of the code for a thunk function; @code{TARGET_ASM_FUNCTION_PROLOGUE}
4824 and @code{TARGET_ASM_FUNCTION_EPILOGUE} are not invoked.
4826 The @var{thunk_fndecl} is redundant. (@var{delta} and @var{function}
4827 have already been extracted from it.) It might possibly be useful on
4828 some targets, but probably not.
4830 If you do not define this macro, the target-independent code in the C++
4831 front end will generate a less efficient heavyweight thunk that calls
4832 @var{function} instead of jumping to it. The generic approach does
4833 not support varargs.
4836 @hook TARGET_ASM_CAN_OUTPUT_MI_THUNK
4837 A function that returns true if TARGET_ASM_OUTPUT_MI_THUNK would be able
4838 to output the assembler code for the thunk function specified by the
4839 arguments it is passed, and false otherwise. In the latter case, the
4840 generic approach will be used by the C++ front end, with the limitations
4845 @subsection Generating Code for Profiling
4846 @cindex profiling, code generation
4848 These macros will help you generate code for profiling.
4850 @defmac FUNCTION_PROFILER (@var{file}, @var{labelno})
4851 A C statement or compound statement to output to @var{file} some
4852 assembler code to call the profiling subroutine @code{mcount}.
4855 The details of how @code{mcount} expects to be called are determined by
4856 your operating system environment, not by GCC@. To figure them out,
4857 compile a small program for profiling using the system's installed C
4858 compiler and look at the assembler code that results.
4860 Older implementations of @code{mcount} expect the address of a counter
4861 variable to be loaded into some register. The name of this variable is
4862 @samp{LP} followed by the number @var{labelno}, so you would generate
4863 the name using @samp{LP%d} in a @code{fprintf}.
4866 @defmac PROFILE_HOOK
4867 A C statement or compound statement to output to @var{file} some assembly
4868 code to call the profiling subroutine @code{mcount} even the target does
4869 not support profiling.
4872 @defmac NO_PROFILE_COUNTERS
4873 Define this macro to be an expression with a nonzero value if the
4874 @code{mcount} subroutine on your system does not need a counter variable
4875 allocated for each function. This is true for almost all modern
4876 implementations. If you define this macro, you must not use the
4877 @var{labelno} argument to @code{FUNCTION_PROFILER}.
4880 @defmac PROFILE_BEFORE_PROLOGUE
4881 Define this macro if the code for function profiling should come before
4882 the function prologue. Normally, the profiling code comes after.
4886 @subsection Permitting tail calls
4889 @hook TARGET_FUNCTION_OK_FOR_SIBCALL
4890 True if it is ok to do sibling call optimization for the specified
4891 call expression @var{exp}. @var{decl} will be the called function,
4892 or @code{NULL} if this is an indirect call.
4894 It is not uncommon for limitations of calling conventions to prevent
4895 tail calls to functions outside the current unit of translation, or
4896 during PIC compilation. The hook is used to enforce these restrictions,
4897 as the @code{sibcall} md pattern can not fail, or fall over to a
4898 ``normal'' call. The criteria for successful sibling call optimization
4899 may vary greatly between different architectures.
4902 @hook TARGET_EXTRA_LIVE_ON_ENTRY
4903 Add any hard registers to @var{regs} that are live on entry to the
4904 function. This hook only needs to be defined to provide registers that
4905 cannot be found by examination of FUNCTION_ARG_REGNO_P, the callee saved
4906 registers, STATIC_CHAIN_INCOMING_REGNUM, STATIC_CHAIN_REGNUM,
4907 TARGET_STRUCT_VALUE_RTX, FRAME_POINTER_REGNUM, EH_USES,
4908 FRAME_POINTER_REGNUM, ARG_POINTER_REGNUM, and the PIC_OFFSET_TABLE_REGNUM.
4911 @hook TARGET_SET_UP_BY_PROLOGUE
4913 @node Stack Smashing Protection
4914 @subsection Stack smashing protection
4915 @cindex stack smashing protection
4917 @hook TARGET_STACK_PROTECT_GUARD
4918 This hook returns a @code{DECL} node for the external variable to use
4919 for the stack protection guard. This variable is initialized by the
4920 runtime to some random value and is used to initialize the guard value
4921 that is placed at the top of the local stack frame. The type of this
4922 variable must be @code{ptr_type_node}.
4924 The default version of this hook creates a variable called
4925 @samp{__stack_chk_guard}, which is normally defined in @file{libgcc2.c}.
4928 @hook TARGET_STACK_PROTECT_FAIL
4929 This hook returns a tree expression that alerts the runtime that the
4930 stack protect guard variable has been modified. This expression should
4931 involve a call to a @code{noreturn} function.
4933 The default version of this hook invokes a function called
4934 @samp{__stack_chk_fail}, taking no arguments. This function is
4935 normally defined in @file{libgcc2.c}.
4938 @hook TARGET_SUPPORTS_SPLIT_STACK
4941 @section Implementing the Varargs Macros
4942 @cindex varargs implementation
4944 GCC comes with an implementation of @code{<varargs.h>} and
4945 @code{<stdarg.h>} that work without change on machines that pass arguments
4946 on the stack. Other machines require their own implementations of
4947 varargs, and the two machine independent header files must have
4948 conditionals to include it.
4950 ISO @code{<stdarg.h>} differs from traditional @code{<varargs.h>} mainly in
4951 the calling convention for @code{va_start}. The traditional
4952 implementation takes just one argument, which is the variable in which
4953 to store the argument pointer. The ISO implementation of
4954 @code{va_start} takes an additional second argument. The user is
4955 supposed to write the last named argument of the function here.
4957 However, @code{va_start} should not use this argument. The way to find
4958 the end of the named arguments is with the built-in functions described
4961 @defmac __builtin_saveregs ()
4962 Use this built-in function to save the argument registers in memory so
4963 that the varargs mechanism can access them. Both ISO and traditional
4964 versions of @code{va_start} must use @code{__builtin_saveregs}, unless
4965 you use @code{TARGET_SETUP_INCOMING_VARARGS} (see below) instead.
4967 On some machines, @code{__builtin_saveregs} is open-coded under the
4968 control of the target hook @code{TARGET_EXPAND_BUILTIN_SAVEREGS}. On
4969 other machines, it calls a routine written in assembler language,
4970 found in @file{libgcc2.c}.
4972 Code generated for the call to @code{__builtin_saveregs} appears at the
4973 beginning of the function, as opposed to where the call to
4974 @code{__builtin_saveregs} is written, regardless of what the code is.
4975 This is because the registers must be saved before the function starts
4976 to use them for its own purposes.
4977 @c i rewrote the first sentence above to fix an overfull hbox. --mew
4981 @defmac __builtin_next_arg (@var{lastarg})
4982 This builtin returns the address of the first anonymous stack
4983 argument, as type @code{void *}. If @code{ARGS_GROW_DOWNWARD}, it
4984 returns the address of the location above the first anonymous stack
4985 argument. Use it in @code{va_start} to initialize the pointer for
4986 fetching arguments from the stack. Also use it in @code{va_start} to
4987 verify that the second parameter @var{lastarg} is the last named argument
4988 of the current function.
4991 @defmac __builtin_classify_type (@var{object})
4992 Since each machine has its own conventions for which data types are
4993 passed in which kind of register, your implementation of @code{va_arg}
4994 has to embody these conventions. The easiest way to categorize the
4995 specified data type is to use @code{__builtin_classify_type} together
4996 with @code{sizeof} and @code{__alignof__}.
4998 @code{__builtin_classify_type} ignores the value of @var{object},
4999 considering only its data type. It returns an integer describing what
5000 kind of type that is---integer, floating, pointer, structure, and so on.
5002 The file @file{typeclass.h} defines an enumeration that you can use to
5003 interpret the values of @code{__builtin_classify_type}.
5006 These machine description macros help implement varargs:
5008 @hook TARGET_EXPAND_BUILTIN_SAVEREGS
5009 If defined, this hook produces the machine-specific code for a call to
5010 @code{__builtin_saveregs}. This code will be moved to the very
5011 beginning of the function, before any parameter access are made. The
5012 return value of this function should be an RTX that contains the value
5013 to use as the return of @code{__builtin_saveregs}.
5016 @hook TARGET_SETUP_INCOMING_VARARGS
5017 This target hook offers an alternative to using
5018 @code{__builtin_saveregs} and defining the hook
5019 @code{TARGET_EXPAND_BUILTIN_SAVEREGS}. Use it to store the anonymous
5020 register arguments into the stack so that all the arguments appear to
5021 have been passed consecutively on the stack. Once this is done, you can
5022 use the standard implementation of varargs that works for machines that
5023 pass all their arguments on the stack.
5025 The argument @var{args_so_far} points to the @code{CUMULATIVE_ARGS} data
5026 structure, containing the values that are obtained after processing the
5027 named arguments. The arguments @var{mode} and @var{type} describe the
5028 last named argument---its machine mode and its data type as a tree node.
5030 The target hook should do two things: first, push onto the stack all the
5031 argument registers @emph{not} used for the named arguments, and second,
5032 store the size of the data thus pushed into the @code{int}-valued
5033 variable pointed to by @var{pretend_args_size}. The value that you
5034 store here will serve as additional offset for setting up the stack
5037 Because you must generate code to push the anonymous arguments at
5038 compile time without knowing their data types,
5039 @code{TARGET_SETUP_INCOMING_VARARGS} is only useful on machines that
5040 have just a single category of argument register and use it uniformly
5043 If the argument @var{second_time} is nonzero, it means that the
5044 arguments of the function are being analyzed for the second time. This
5045 happens for an inline function, which is not actually compiled until the
5046 end of the source file. The hook @code{TARGET_SETUP_INCOMING_VARARGS} should
5047 not generate any instructions in this case.
5050 @hook TARGET_STRICT_ARGUMENT_NAMING
5051 Define this hook to return @code{true} if the location where a function
5052 argument is passed depends on whether or not it is a named argument.
5054 This hook controls how the @var{named} argument to @code{TARGET_FUNCTION_ARG}
5055 is set for varargs and stdarg functions. If this hook returns
5056 @code{true}, the @var{named} argument is always true for named
5057 arguments, and false for unnamed arguments. If it returns @code{false},
5058 but @code{TARGET_PRETEND_OUTGOING_VARARGS_NAMED} returns @code{true},
5059 then all arguments are treated as named. Otherwise, all named arguments
5060 except the last are treated as named.
5062 You need not define this hook if it always returns @code{false}.
5065 @hook TARGET_PRETEND_OUTGOING_VARARGS_NAMED
5066 If you need to conditionally change ABIs so that one works with
5067 @code{TARGET_SETUP_INCOMING_VARARGS}, but the other works like neither
5068 @code{TARGET_SETUP_INCOMING_VARARGS} nor @code{TARGET_STRICT_ARGUMENT_NAMING} was
5069 defined, then define this hook to return @code{true} if
5070 @code{TARGET_SETUP_INCOMING_VARARGS} is used, @code{false} otherwise.
5071 Otherwise, you should not define this hook.
5075 @section Trampolines for Nested Functions
5076 @cindex trampolines for nested functions
5077 @cindex nested functions, trampolines for
5079 A @dfn{trampoline} is a small piece of code that is created at run time
5080 when the address of a nested function is taken. It normally resides on
5081 the stack, in the stack frame of the containing function. These macros
5082 tell GCC how to generate code to allocate and initialize a
5085 The instructions in the trampoline must do two things: load a constant
5086 address into the static chain register, and jump to the real address of
5087 the nested function. On CISC machines such as the m68k, this requires
5088 two instructions, a move immediate and a jump. Then the two addresses
5089 exist in the trampoline as word-long immediate operands. On RISC
5090 machines, it is often necessary to load each address into a register in
5091 two parts. Then pieces of each address form separate immediate
5094 The code generated to initialize the trampoline must store the variable
5095 parts---the static chain value and the function address---into the
5096 immediate operands of the instructions. On a CISC machine, this is
5097 simply a matter of copying each address to a memory reference at the
5098 proper offset from the start of the trampoline. On a RISC machine, it
5099 may be necessary to take out pieces of the address and store them
5102 @hook TARGET_ASM_TRAMPOLINE_TEMPLATE
5103 This hook is called by @code{assemble_trampoline_template} to output,
5104 on the stream @var{f}, assembler code for a block of data that contains
5105 the constant parts of a trampoline. This code should not include a
5106 label---the label is taken care of automatically.
5108 If you do not define this hook, it means no template is needed
5109 for the target. Do not define this hook on systems where the block move
5110 code to copy the trampoline into place would be larger than the code
5111 to generate it on the spot.
5114 @defmac TRAMPOLINE_SECTION
5115 Return the section into which the trampoline template is to be placed
5116 (@pxref{Sections}). The default value is @code{readonly_data_section}.
5119 @defmac TRAMPOLINE_SIZE
5120 A C expression for the size in bytes of the trampoline, as an integer.
5123 @defmac TRAMPOLINE_ALIGNMENT
5124 Alignment required for trampolines, in bits.
5126 If you don't define this macro, the value of @code{FUNCTION_ALIGNMENT}
5127 is used for aligning trampolines.
5130 @hook TARGET_TRAMPOLINE_INIT
5131 This hook is called to initialize a trampoline.
5132 @var{m_tramp} is an RTX for the memory block for the trampoline; @var{fndecl}
5133 is the @code{FUNCTION_DECL} for the nested function; @var{static_chain} is an
5134 RTX for the static chain value that should be passed to the function
5137 If the target defines @code{TARGET_ASM_TRAMPOLINE_TEMPLATE}, then the
5138 first thing this hook should do is emit a block move into @var{m_tramp}
5139 from the memory block returned by @code{assemble_trampoline_template}.
5140 Note that the block move need only cover the constant parts of the
5141 trampoline. If the target isolates the variable parts of the trampoline
5142 to the end, not all @code{TRAMPOLINE_SIZE} bytes need be copied.
5144 If the target requires any other actions, such as flushing caches or
5145 enabling stack execution, these actions should be performed after
5146 initializing the trampoline proper.
5149 @hook TARGET_TRAMPOLINE_ADJUST_ADDRESS
5150 This hook should perform any machine-specific adjustment in
5151 the address of the trampoline. Its argument contains the address of the
5152 memory block that was passed to @code{TARGET_TRAMPOLINE_INIT}. In case
5153 the address to be used for a function call should be different from the
5154 address at which the template was stored, the different address should
5155 be returned; otherwise @var{addr} should be returned unchanged.
5156 If this hook is not defined, @var{addr} will be used for function calls.
5159 Implementing trampolines is difficult on many machines because they have
5160 separate instruction and data caches. Writing into a stack location
5161 fails to clear the memory in the instruction cache, so when the program
5162 jumps to that location, it executes the old contents.
5164 Here are two possible solutions. One is to clear the relevant parts of
5165 the instruction cache whenever a trampoline is set up. The other is to
5166 make all trampolines identical, by having them jump to a standard
5167 subroutine. The former technique makes trampoline execution faster; the
5168 latter makes initialization faster.
5170 To clear the instruction cache when a trampoline is initialized, define
5171 the following macro.
5173 @defmac CLEAR_INSN_CACHE (@var{beg}, @var{end})
5174 If defined, expands to a C expression clearing the @emph{instruction
5175 cache} in the specified interval. The definition of this macro would
5176 typically be a series of @code{asm} statements. Both @var{beg} and
5177 @var{end} are both pointer expressions.
5180 To use a standard subroutine, define the following macro. In addition,
5181 you must make sure that the instructions in a trampoline fill an entire
5182 cache line with identical instructions, or else ensure that the
5183 beginning of the trampoline code is always aligned at the same point in
5184 its cache line. Look in @file{m68k.h} as a guide.
5186 @defmac TRANSFER_FROM_TRAMPOLINE
5187 Define this macro if trampolines need a special subroutine to do their
5188 work. The macro should expand to a series of @code{asm} statements
5189 which will be compiled with GCC@. They go in a library function named
5190 @code{__transfer_from_trampoline}.
5192 If you need to avoid executing the ordinary prologue code of a compiled
5193 C function when you jump to the subroutine, you can do so by placing a
5194 special label of your own in the assembler code. Use one @code{asm}
5195 statement to generate an assembler label, and another to make the label
5196 global. Then trampolines can use that label to jump directly to your
5197 special assembler code.
5201 @section Implicit Calls to Library Routines
5202 @cindex library subroutine names
5203 @cindex @file{libgcc.a}
5205 @c prevent bad page break with this line
5206 Here is an explanation of implicit calls to library routines.
5208 @defmac DECLARE_LIBRARY_RENAMES
5209 This macro, if defined, should expand to a piece of C code that will get
5210 expanded when compiling functions for libgcc.a. It can be used to
5211 provide alternate names for GCC's internal library functions if there
5212 are ABI-mandated names that the compiler should provide.
5215 @findex set_optab_libfunc
5216 @findex init_one_libfunc
5217 @hook TARGET_INIT_LIBFUNCS
5218 This hook should declare additional library routines or rename
5219 existing ones, using the functions @code{set_optab_libfunc} and
5220 @code{init_one_libfunc} defined in @file{optabs.c}.
5221 @code{init_optabs} calls this macro after initializing all the normal
5224 The default is to do nothing. Most ports don't need to define this hook.
5227 @hook TARGET_LIBFUNC_GNU_PREFIX
5229 @defmac FLOAT_LIB_COMPARE_RETURNS_BOOL (@var{mode}, @var{comparison})
5230 This macro should return @code{true} if the library routine that
5231 implements the floating point comparison operator @var{comparison} in
5232 mode @var{mode} will return a boolean, and @var{false} if it will
5235 GCC's own floating point libraries return tristates from the
5236 comparison operators, so the default returns false always. Most ports
5237 don't need to define this macro.
5240 @defmac TARGET_LIB_INT_CMP_BIASED
5241 This macro should evaluate to @code{true} if the integer comparison
5242 functions (like @code{__cmpdi2}) return 0 to indicate that the first
5243 operand is smaller than the second, 1 to indicate that they are equal,
5244 and 2 to indicate that the first operand is greater than the second.
5245 If this macro evaluates to @code{false} the comparison functions return
5246 @minus{}1, 0, and 1 instead of 0, 1, and 2. If the target uses the routines
5247 in @file{libgcc.a}, you do not need to define this macro.
5250 @cindex @code{EDOM}, implicit usage
5253 The value of @code{EDOM} on the target machine, as a C integer constant
5254 expression. If you don't define this macro, GCC does not attempt to
5255 deposit the value of @code{EDOM} into @code{errno} directly. Look in
5256 @file{/usr/include/errno.h} to find the value of @code{EDOM} on your
5259 If you do not define @code{TARGET_EDOM}, then compiled code reports
5260 domain errors by calling the library function and letting it report the
5261 error. If mathematical functions on your system use @code{matherr} when
5262 there is an error, then you should leave @code{TARGET_EDOM} undefined so
5263 that @code{matherr} is used normally.
5266 @cindex @code{errno}, implicit usage
5267 @defmac GEN_ERRNO_RTX
5268 Define this macro as a C expression to create an rtl expression that
5269 refers to the global ``variable'' @code{errno}. (On certain systems,
5270 @code{errno} may not actually be a variable.) If you don't define this
5271 macro, a reasonable default is used.
5274 @cindex C99 math functions, implicit usage
5275 @defmac TARGET_C99_FUNCTIONS
5276 When this macro is nonzero, GCC will implicitly optimize @code{sin} calls into
5277 @code{sinf} and similarly for other functions defined by C99 standard. The
5278 default is zero because a number of existing systems lack support for these
5279 functions in their runtime so this macro needs to be redefined to one on
5280 systems that do support the C99 runtime.
5283 @cindex sincos math function, implicit usage
5284 @defmac TARGET_HAS_SINCOS
5285 When this macro is nonzero, GCC will implicitly optimize calls to @code{sin}
5286 and @code{cos} with the same argument to a call to @code{sincos}. The
5287 default is zero. The target has to provide the following functions:
5289 void sincos(double x, double *sin, double *cos);
5290 void sincosf(float x, float *sin, float *cos);
5291 void sincosl(long double x, long double *sin, long double *cos);
5295 @defmac NEXT_OBJC_RUNTIME
5296 Set this macro to 1 to use the "NeXT" Objective-C message sending conventions
5297 by default. This calling convention involves passing the object, the selector
5298 and the method arguments all at once to the method-lookup library function.
5299 This is the usual setting when targeting Darwin/Mac OS X systems, which have
5300 the NeXT runtime installed.
5302 If the macro is set to 0, the "GNU" Objective-C message sending convention
5303 will be used by default. This convention passes just the object and the
5304 selector to the method-lookup function, which returns a pointer to the method.
5306 In either case, it remains possible to select code-generation for the alternate
5307 scheme, by means of compiler command line switches.
5310 @node Addressing Modes
5311 @section Addressing Modes
5312 @cindex addressing modes
5314 @c prevent bad page break with this line
5315 This is about addressing modes.
5317 @defmac HAVE_PRE_INCREMENT
5318 @defmacx HAVE_PRE_DECREMENT
5319 @defmacx HAVE_POST_INCREMENT
5320 @defmacx HAVE_POST_DECREMENT
5321 A C expression that is nonzero if the machine supports pre-increment,
5322 pre-decrement, post-increment, or post-decrement addressing respectively.
5325 @defmac HAVE_PRE_MODIFY_DISP
5326 @defmacx HAVE_POST_MODIFY_DISP
5327 A C expression that is nonzero if the machine supports pre- or
5328 post-address side-effect generation involving constants other than
5329 the size of the memory operand.
5332 @defmac HAVE_PRE_MODIFY_REG
5333 @defmacx HAVE_POST_MODIFY_REG
5334 A C expression that is nonzero if the machine supports pre- or
5335 post-address side-effect generation involving a register displacement.
5338 @defmac CONSTANT_ADDRESS_P (@var{x})
5339 A C expression that is 1 if the RTX @var{x} is a constant which
5340 is a valid address. On most machines the default definition of
5341 @code{(CONSTANT_P (@var{x}) && GET_CODE (@var{x}) != CONST_DOUBLE)}
5342 is acceptable, but a few machines are more restrictive as to which
5343 constant addresses are supported.
5346 @defmac CONSTANT_P (@var{x})
5347 @code{CONSTANT_P}, which is defined by target-independent code,
5348 accepts integer-values expressions whose values are not explicitly
5349 known, such as @code{symbol_ref}, @code{label_ref}, and @code{high}
5350 expressions and @code{const} arithmetic expressions, in addition to
5351 @code{const_int} and @code{const_double} expressions.
5354 @defmac MAX_REGS_PER_ADDRESS
5355 A number, the maximum number of registers that can appear in a valid
5356 memory address. Note that it is up to you to specify a value equal to
5357 the maximum number that @code{TARGET_LEGITIMATE_ADDRESS_P} would ever
5361 @hook TARGET_LEGITIMATE_ADDRESS_P
5362 A function that returns whether @var{x} (an RTX) is a legitimate memory
5363 address on the target machine for a memory operand of mode @var{mode}.
5365 Legitimate addresses are defined in two variants: a strict variant and a
5366 non-strict one. The @var{strict} parameter chooses which variant is
5367 desired by the caller.
5369 The strict variant is used in the reload pass. It must be defined so
5370 that any pseudo-register that has not been allocated a hard register is
5371 considered a memory reference. This is because in contexts where some
5372 kind of register is required, a pseudo-register with no hard register
5373 must be rejected. For non-hard registers, the strict variant should look
5374 up the @code{reg_renumber} array; it should then proceed using the hard
5375 register number in the array, or treat the pseudo as a memory reference
5376 if the array holds @code{-1}.
5378 The non-strict variant is used in other passes. It must be defined to
5379 accept all pseudo-registers in every context where some kind of
5380 register is required.
5382 Normally, constant addresses which are the sum of a @code{symbol_ref}
5383 and an integer are stored inside a @code{const} RTX to mark them as
5384 constant. Therefore, there is no need to recognize such sums
5385 specifically as legitimate addresses. Normally you would simply
5386 recognize any @code{const} as legitimate.
5388 Usually @code{PRINT_OPERAND_ADDRESS} is not prepared to handle constant
5389 sums that are not marked with @code{const}. It assumes that a naked
5390 @code{plus} indicates indexing. If so, then you @emph{must} reject such
5391 naked constant sums as illegitimate addresses, so that none of them will
5392 be given to @code{PRINT_OPERAND_ADDRESS}.
5394 @cindex @code{TARGET_ENCODE_SECTION_INFO} and address validation
5395 On some machines, whether a symbolic address is legitimate depends on
5396 the section that the address refers to. On these machines, define the
5397 target hook @code{TARGET_ENCODE_SECTION_INFO} to store the information
5398 into the @code{symbol_ref}, and then check for it here. When you see a
5399 @code{const}, you will have to look inside it to find the
5400 @code{symbol_ref} in order to determine the section. @xref{Assembler
5403 @cindex @code{GO_IF_LEGITIMATE_ADDRESS}
5404 Some ports are still using a deprecated legacy substitute for
5405 this hook, the @code{GO_IF_LEGITIMATE_ADDRESS} macro. This macro
5409 #define GO_IF_LEGITIMATE_ADDRESS (@var{mode}, @var{x}, @var{label})
5413 and should @code{goto @var{label}} if the address @var{x} is a valid
5414 address on the target machine for a memory operand of mode @var{mode}.
5416 @findex REG_OK_STRICT
5417 Compiler source files that want to use the strict variant of this
5418 macro define the macro @code{REG_OK_STRICT}. You should use an
5419 @code{#ifdef REG_OK_STRICT} conditional to define the strict variant in
5420 that case and the non-strict variant otherwise.
5422 Using the hook is usually simpler because it limits the number of
5423 files that are recompiled when changes are made.
5426 @defmac TARGET_MEM_CONSTRAINT
5427 A single character to be used instead of the default @code{'m'}
5428 character for general memory addresses. This defines the constraint
5429 letter which matches the memory addresses accepted by
5430 @code{TARGET_LEGITIMATE_ADDRESS_P}. Define this macro if you want to
5431 support new address formats in your back end without changing the
5432 semantics of the @code{'m'} constraint. This is necessary in order to
5433 preserve functionality of inline assembly constructs using the
5434 @code{'m'} constraint.
5437 @defmac FIND_BASE_TERM (@var{x})
5438 A C expression to determine the base term of address @var{x},
5439 or to provide a simplified version of @var{x} from which @file{alias.c}
5440 can easily find the base term. This macro is used in only two places:
5441 @code{find_base_value} and @code{find_base_term} in @file{alias.c}.
5443 It is always safe for this macro to not be defined. It exists so
5444 that alias analysis can understand machine-dependent addresses.
5446 The typical use of this macro is to handle addresses containing
5447 a label_ref or symbol_ref within an UNSPEC@.
5450 @hook TARGET_LEGITIMIZE_ADDRESS
5451 This hook is given an invalid memory address @var{x} for an
5452 operand of mode @var{mode} and should try to return a valid memory
5455 @findex break_out_memory_refs
5456 @var{x} will always be the result of a call to @code{break_out_memory_refs},
5457 and @var{oldx} will be the operand that was given to that function to produce
5460 The code of the hook should not alter the substructure of
5461 @var{x}. If it transforms @var{x} into a more legitimate form, it
5462 should return the new @var{x}.
5464 It is not necessary for this hook to come up with a legitimate address.
5465 The compiler has standard ways of doing so in all cases. In fact, it
5466 is safe to omit this hook or make it return @var{x} if it cannot find
5467 a valid way to legitimize the address. But often a machine-dependent
5468 strategy can generate better code.
5471 @defmac LEGITIMIZE_RELOAD_ADDRESS (@var{x}, @var{mode}, @var{opnum}, @var{type}, @var{ind_levels}, @var{win})
5472 A C compound statement that attempts to replace @var{x}, which is an address
5473 that needs reloading, with a valid memory address for an operand of mode
5474 @var{mode}. @var{win} will be a C statement label elsewhere in the code.
5475 It is not necessary to define this macro, but it might be useful for
5476 performance reasons.
5478 For example, on the i386, it is sometimes possible to use a single
5479 reload register instead of two by reloading a sum of two pseudo
5480 registers into a register. On the other hand, for number of RISC
5481 processors offsets are limited so that often an intermediate address
5482 needs to be generated in order to address a stack slot. By defining
5483 @code{LEGITIMIZE_RELOAD_ADDRESS} appropriately, the intermediate addresses
5484 generated for adjacent some stack slots can be made identical, and thus
5487 @emph{Note}: This macro should be used with caution. It is necessary
5488 to know something of how reload works in order to effectively use this,
5489 and it is quite easy to produce macros that build in too much knowledge
5490 of reload internals.
5492 @emph{Note}: This macro must be able to reload an address created by a
5493 previous invocation of this macro. If it fails to handle such addresses
5494 then the compiler may generate incorrect code or abort.
5497 The macro definition should use @code{push_reload} to indicate parts that
5498 need reloading; @var{opnum}, @var{type} and @var{ind_levels} are usually
5499 suitable to be passed unaltered to @code{push_reload}.
5501 The code generated by this macro must not alter the substructure of
5502 @var{x}. If it transforms @var{x} into a more legitimate form, it
5503 should assign @var{x} (which will always be a C variable) a new value.
5504 This also applies to parts that you change indirectly by calling
5507 @findex strict_memory_address_p
5508 The macro definition may use @code{strict_memory_address_p} to test if
5509 the address has become legitimate.
5512 If you want to change only a part of @var{x}, one standard way of doing
5513 this is to use @code{copy_rtx}. Note, however, that it unshares only a
5514 single level of rtl. Thus, if the part to be changed is not at the
5515 top level, you'll need to replace first the top level.
5516 It is not necessary for this macro to come up with a legitimate
5517 address; but often a machine-dependent strategy can generate better code.
5520 @hook TARGET_MODE_DEPENDENT_ADDRESS_P
5521 This hook returns @code{true} if memory address @var{addr} can have
5522 different meanings depending on the machine mode of the memory
5523 reference it is used for or if the address is valid for some modes
5526 Autoincrement and autodecrement addresses typically have mode-dependent
5527 effects because the amount of the increment or decrement is the size
5528 of the operand being addressed. Some machines have other mode-dependent
5529 addresses. Many RISC machines have no mode-dependent addresses.
5531 You may assume that @var{addr} is a valid address for the machine.
5533 The default version of this hook returns @code{false}.
5536 @defmac GO_IF_MODE_DEPENDENT_ADDRESS (@var{addr}, @var{label})
5537 A C statement or compound statement with a conditional @code{goto
5538 @var{label};} executed if memory address @var{x} (an RTX) can have
5539 different meanings depending on the machine mode of the memory
5540 reference it is used for or if the address is valid for some modes
5543 Autoincrement and autodecrement addresses typically have mode-dependent
5544 effects because the amount of the increment or decrement is the size
5545 of the operand being addressed. Some machines have other mode-dependent
5546 addresses. Many RISC machines have no mode-dependent addresses.
5548 You may assume that @var{addr} is a valid address for the machine.
5550 These are obsolete macros, replaced by the
5551 @code{TARGET_MODE_DEPENDENT_ADDRESS_P} target hook.
5554 @hook TARGET_LEGITIMATE_CONSTANT_P
5555 This hook returns true if @var{x} is a legitimate constant for a
5556 @var{mode}-mode immediate operand on the target machine. You can assume that
5557 @var{x} satisfies @code{CONSTANT_P}, so you need not check this.
5559 The default definition returns true.
5562 @hook TARGET_DELEGITIMIZE_ADDRESS
5563 This hook is used to undo the possibly obfuscating effects of the
5564 @code{LEGITIMIZE_ADDRESS} and @code{LEGITIMIZE_RELOAD_ADDRESS} target
5565 macros. Some backend implementations of these macros wrap symbol
5566 references inside an @code{UNSPEC} rtx to represent PIC or similar
5567 addressing modes. This target hook allows GCC's optimizers to understand
5568 the semantics of these opaque @code{UNSPEC}s by converting them back
5569 into their original form.
5572 @hook TARGET_CONST_NOT_OK_FOR_DEBUG_P
5573 This hook should return true if @var{x} should not be emitted into
5577 @hook TARGET_CANNOT_FORCE_CONST_MEM
5578 This hook should return true if @var{x} is of a form that cannot (or
5579 should not) be spilled to the constant pool. @var{mode} is the mode
5582 The default version of this hook returns false.
5584 The primary reason to define this hook is to prevent reload from
5585 deciding that a non-legitimate constant would be better reloaded
5586 from the constant pool instead of spilling and reloading a register
5587 holding the constant. This restriction is often true of addresses
5588 of TLS symbols for various targets.
5591 @hook TARGET_USE_BLOCKS_FOR_CONSTANT_P
5592 This hook should return true if pool entries for constant @var{x} can
5593 be placed in an @code{object_block} structure. @var{mode} is the mode
5596 The default version returns false for all constants.
5599 @hook TARGET_BUILTIN_RECIPROCAL
5600 This hook should return the DECL of a function that implements reciprocal of
5601 the builtin function with builtin function code @var{fn}, or
5602 @code{NULL_TREE} if such a function is not available. @var{md_fn} is true
5603 when @var{fn} is a code of a machine-dependent builtin function. When
5604 @var{sqrt} is true, additional optimizations that apply only to the reciprocal
5605 of a square root function are performed, and only reciprocals of @code{sqrt}
5609 @hook TARGET_VECTORIZE_BUILTIN_MASK_FOR_LOAD
5610 This hook should return the DECL of a function @var{f} that given an
5611 address @var{addr} as an argument returns a mask @var{m} that can be
5612 used to extract from two vectors the relevant data that resides in
5613 @var{addr} in case @var{addr} is not properly aligned.
5615 The autovectorizer, when vectorizing a load operation from an address
5616 @var{addr} that may be unaligned, will generate two vector loads from
5617 the two aligned addresses around @var{addr}. It then generates a
5618 @code{REALIGN_LOAD} operation to extract the relevant data from the
5619 two loaded vectors. The first two arguments to @code{REALIGN_LOAD},
5620 @var{v1} and @var{v2}, are the two vectors, each of size @var{VS}, and
5621 the third argument, @var{OFF}, defines how the data will be extracted
5622 from these two vectors: if @var{OFF} is 0, then the returned vector is
5623 @var{v2}; otherwise, the returned vector is composed from the last
5624 @var{VS}-@var{OFF} elements of @var{v1} concatenated to the first
5625 @var{OFF} elements of @var{v2}.
5627 If this hook is defined, the autovectorizer will generate a call
5628 to @var{f} (using the DECL tree that this hook returns) and will
5629 use the return value of @var{f} as the argument @var{OFF} to
5630 @code{REALIGN_LOAD}. Therefore, the mask @var{m} returned by @var{f}
5631 should comply with the semantics expected by @code{REALIGN_LOAD}
5633 If this hook is not defined, then @var{addr} will be used as
5634 the argument @var{OFF} to @code{REALIGN_LOAD}, in which case the low
5635 log2(@var{VS}) @minus{} 1 bits of @var{addr} will be considered.
5638 @hook TARGET_VECTORIZE_BUILTIN_MUL_WIDEN_EVEN
5639 This hook should return the DECL of a function @var{f} that implements
5640 widening multiplication of the even elements of two input vectors of type @var{x}.
5642 If this hook is defined, the autovectorizer will use it along with the
5643 @code{TARGET_VECTORIZE_BUILTIN_MUL_WIDEN_ODD} target hook when vectorizing
5644 widening multiplication in cases that the order of the results does not have to be
5645 preserved (e.g.@: used only by a reduction computation). Otherwise, the
5646 @code{widen_mult_hi/lo} idioms will be used.
5649 @hook TARGET_VECTORIZE_BUILTIN_MUL_WIDEN_ODD
5650 This hook should return the DECL of a function @var{f} that implements
5651 widening multiplication of the odd elements of two input vectors of type @var{x}.
5653 If this hook is defined, the autovectorizer will use it along with the
5654 @code{TARGET_VECTORIZE_BUILTIN_MUL_WIDEN_EVEN} target hook when vectorizing
5655 widening multiplication in cases that the order of the results does not have to be
5656 preserved (e.g.@: used only by a reduction computation). Otherwise, the
5657 @code{widen_mult_hi/lo} idioms will be used.
5660 @hook TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST
5661 Returns cost of different scalar or vector statements for vectorization cost model.
5662 For vector memory operations the cost may depend on type (@var{vectype}) and
5663 misalignment value (@var{misalign}).
5666 @hook TARGET_VECTORIZE_VECTOR_ALIGNMENT_REACHABLE
5667 Return true if vector alignment is reachable (by peeling N iterations) for the given type.
5670 @hook TARGET_VECTORIZE_VEC_PERM_CONST_OK
5671 Return true if a vector created for @code{vec_perm_const} is valid.
5674 @hook TARGET_VECTORIZE_BUILTIN_CONVERSION
5675 This hook should return the DECL of a function that implements conversion of the
5676 input vector of type @var{src_type} to type @var{dest_type}.
5677 The value of @var{code} is one of the enumerators in @code{enum tree_code} and
5678 specifies how the conversion is to be applied
5679 (truncation, rounding, etc.).
5681 If this hook is defined, the autovectorizer will use the
5682 @code{TARGET_VECTORIZE_BUILTIN_CONVERSION} target hook when vectorizing
5683 conversion. Otherwise, it will return @code{NULL_TREE}.
5686 @hook TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION
5687 This hook should return the decl of a function that implements the
5688 vectorized variant of the builtin function with builtin function code
5689 @var{code} or @code{NULL_TREE} if such a function is not available.
5690 The value of @var{fndecl} is the builtin function declaration. The
5691 return type of the vectorized function shall be of vector type
5692 @var{vec_type_out} and the argument types should be @var{vec_type_in}.
5695 @hook TARGET_VECTORIZE_SUPPORT_VECTOR_MISALIGNMENT
5696 This hook should return true if the target supports misaligned vector
5697 store/load of a specific factor denoted in the @var{misalignment}
5698 parameter. The vector store/load should be of machine mode @var{mode} and
5699 the elements in the vectors should be of type @var{type}. @var{is_packed}
5700 parameter is true if the memory access is defined in a packed struct.
5703 @hook TARGET_VECTORIZE_PREFERRED_SIMD_MODE
5704 This hook should return the preferred mode for vectorizing scalar
5705 mode @var{mode}. The default is
5706 equal to @code{word_mode}, because the vectorizer can do some
5707 transformations even in absence of specialized @acronym{SIMD} hardware.
5710 @hook TARGET_VECTORIZE_AUTOVECTORIZE_VECTOR_SIZES
5711 This hook should return a mask of sizes that should be iterated over
5712 after trying to autovectorize using the vector size derived from the
5713 mode returned by @code{TARGET_VECTORIZE_PREFERRED_SIMD_MODE}.
5714 The default is zero which means to not iterate over other vector sizes.
5717 @hook TARGET_VECTORIZE_BUILTIN_TM_LOAD
5719 @hook TARGET_VECTORIZE_BUILTIN_TM_STORE
5721 @hook TARGET_VECTORIZE_BUILTIN_GATHER
5722 Target builtin that implements vector gather operation. @var{mem_vectype}
5723 is the vector type of the load and @var{index_type} is scalar type of
5724 the index, scaled by @var{scale}.
5725 The default is @code{NULL_TREE} which means to not vectorize gather
5729 @node Anchored Addresses
5730 @section Anchored Addresses
5731 @cindex anchored addresses
5732 @cindex @option{-fsection-anchors}
5734 GCC usually addresses every static object as a separate entity.
5735 For example, if we have:
5739 int foo (void) @{ return a + b + c; @}
5742 the code for @code{foo} will usually calculate three separate symbolic
5743 addresses: those of @code{a}, @code{b} and @code{c}. On some targets,
5744 it would be better to calculate just one symbolic address and access
5745 the three variables relative to it. The equivalent pseudocode would
5751 register int *xr = &x;
5752 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
5756 (which isn't valid C). We refer to shared addresses like @code{x} as
5757 ``section anchors''. Their use is controlled by @option{-fsection-anchors}.
5759 The hooks below describe the target properties that GCC needs to know
5760 in order to make effective use of section anchors. It won't use
5761 section anchors at all unless either @code{TARGET_MIN_ANCHOR_OFFSET}
5762 or @code{TARGET_MAX_ANCHOR_OFFSET} is set to a nonzero value.
5764 @hook TARGET_MIN_ANCHOR_OFFSET
5765 The minimum offset that should be applied to a section anchor.
5766 On most targets, it should be the smallest offset that can be
5767 applied to a base register while still giving a legitimate address
5768 for every mode. The default value is 0.
5771 @hook TARGET_MAX_ANCHOR_OFFSET
5772 Like @code{TARGET_MIN_ANCHOR_OFFSET}, but the maximum (inclusive)
5773 offset that should be applied to section anchors. The default
5777 @hook TARGET_ASM_OUTPUT_ANCHOR
5778 Write the assembly code to define section anchor @var{x}, which is a
5779 @code{SYMBOL_REF} for which @samp{SYMBOL_REF_ANCHOR_P (@var{x})} is true.
5780 The hook is called with the assembly output position set to the beginning
5781 of @code{SYMBOL_REF_BLOCK (@var{x})}.
5783 If @code{ASM_OUTPUT_DEF} is available, the hook's default definition uses
5784 it to define the symbol as @samp{. + SYMBOL_REF_BLOCK_OFFSET (@var{x})}.
5785 If @code{ASM_OUTPUT_DEF} is not available, the hook's default definition
5786 is @code{NULL}, which disables the use of section anchors altogether.
5789 @hook TARGET_USE_ANCHORS_FOR_SYMBOL_P
5790 Return true if GCC should attempt to use anchors to access @code{SYMBOL_REF}
5791 @var{x}. You can assume @samp{SYMBOL_REF_HAS_BLOCK_INFO_P (@var{x})} and
5792 @samp{!SYMBOL_REF_ANCHOR_P (@var{x})}.
5794 The default version is correct for most targets, but you might need to
5795 intercept this hook to handle things like target-specific attributes
5796 or target-specific sections.
5799 @node Condition Code
5800 @section Condition Code Status
5801 @cindex condition code status
5803 The macros in this section can be split in two families, according to the
5804 two ways of representing condition codes in GCC.
5806 The first representation is the so called @code{(cc0)} representation
5807 (@pxref{Jump Patterns}), where all instructions can have an implicit
5808 clobber of the condition codes. The second is the condition code
5809 register representation, which provides better schedulability for
5810 architectures that do have a condition code register, but on which
5811 most instructions do not affect it. The latter category includes
5814 The implicit clobbering poses a strong restriction on the placement of
5815 the definition and use of the condition code, which need to be in adjacent
5816 insns for machines using @code{(cc0)}. This can prevent important
5817 optimizations on some machines. For example, on the IBM RS/6000, there
5818 is a delay for taken branches unless the condition code register is set
5819 three instructions earlier than the conditional branch. The instruction
5820 scheduler cannot perform this optimization if it is not permitted to
5821 separate the definition and use of the condition code register.
5823 For this reason, it is possible and suggested to use a register to
5824 represent the condition code for new ports. If there is a specific
5825 condition code register in the machine, use a hard register. If the
5826 condition code or comparison result can be placed in any general register,
5827 or if there are multiple condition registers, use a pseudo register.
5828 Registers used to store the condition code value will usually have a mode
5829 that is in class @code{MODE_CC}.
5831 Alternatively, you can use @code{BImode} if the comparison operator is
5832 specified already in the compare instruction. In this case, you are not
5833 interested in most macros in this section.
5836 * CC0 Condition Codes:: Old style representation of condition codes.
5837 * MODE_CC Condition Codes:: Modern representation of condition codes.
5838 * Cond Exec Macros:: Macros to control conditional execution.
5841 @node CC0 Condition Codes
5842 @subsection Representation of condition codes using @code{(cc0)}
5846 The file @file{conditions.h} defines a variable @code{cc_status} to
5847 describe how the condition code was computed (in case the interpretation of
5848 the condition code depends on the instruction that it was set by). This
5849 variable contains the RTL expressions on which the condition code is
5850 currently based, and several standard flags.
5852 Sometimes additional machine-specific flags must be defined in the machine
5853 description header file. It can also add additional machine-specific
5854 information by defining @code{CC_STATUS_MDEP}.
5856 @defmac CC_STATUS_MDEP
5857 C code for a data type which is used for declaring the @code{mdep}
5858 component of @code{cc_status}. It defaults to @code{int}.
5860 This macro is not used on machines that do not use @code{cc0}.
5863 @defmac CC_STATUS_MDEP_INIT
5864 A C expression to initialize the @code{mdep} field to ``empty''.
5865 The default definition does nothing, since most machines don't use
5866 the field anyway. If you want to use the field, you should probably
5867 define this macro to initialize it.
5869 This macro is not used on machines that do not use @code{cc0}.
5872 @defmac NOTICE_UPDATE_CC (@var{exp}, @var{insn})
5873 A C compound statement to set the components of @code{cc_status}
5874 appropriately for an insn @var{insn} whose body is @var{exp}. It is
5875 this macro's responsibility to recognize insns that set the condition
5876 code as a byproduct of other activity as well as those that explicitly
5879 This macro is not used on machines that do not use @code{cc0}.
5881 If there are insns that do not set the condition code but do alter
5882 other machine registers, this macro must check to see whether they
5883 invalidate the expressions that the condition code is recorded as
5884 reflecting. For example, on the 68000, insns that store in address
5885 registers do not set the condition code, which means that usually
5886 @code{NOTICE_UPDATE_CC} can leave @code{cc_status} unaltered for such
5887 insns. But suppose that the previous insn set the condition code
5888 based on location @samp{a4@@(102)} and the current insn stores a new
5889 value in @samp{a4}. Although the condition code is not changed by
5890 this, it will no longer be true that it reflects the contents of
5891 @samp{a4@@(102)}. Therefore, @code{NOTICE_UPDATE_CC} must alter
5892 @code{cc_status} in this case to say that nothing is known about the
5893 condition code value.
5895 The definition of @code{NOTICE_UPDATE_CC} must be prepared to deal
5896 with the results of peephole optimization: insns whose patterns are
5897 @code{parallel} RTXs containing various @code{reg}, @code{mem} or
5898 constants which are just the operands. The RTL structure of these
5899 insns is not sufficient to indicate what the insns actually do. What
5900 @code{NOTICE_UPDATE_CC} should do when it sees one is just to run
5901 @code{CC_STATUS_INIT}.
5903 A possible definition of @code{NOTICE_UPDATE_CC} is to call a function
5904 that looks at an attribute (@pxref{Insn Attributes}) named, for example,
5905 @samp{cc}. This avoids having detailed information about patterns in
5906 two places, the @file{md} file and in @code{NOTICE_UPDATE_CC}.
5909 @node MODE_CC Condition Codes
5910 @subsection Representation of condition codes using registers
5914 @defmac SELECT_CC_MODE (@var{op}, @var{x}, @var{y})
5915 On many machines, the condition code may be produced by other instructions
5916 than compares, for example the branch can use directly the condition
5917 code set by a subtract instruction. However, on some machines
5918 when the condition code is set this way some bits (such as the overflow
5919 bit) are not set in the same way as a test instruction, so that a different
5920 branch instruction must be used for some conditional branches. When
5921 this happens, use the machine mode of the condition code register to
5922 record different formats of the condition code register. Modes can
5923 also be used to record which compare instruction (e.g. a signed or an
5924 unsigned comparison) produced the condition codes.
5926 If other modes than @code{CCmode} are required, add them to
5927 @file{@var{machine}-modes.def} and define @code{SELECT_CC_MODE} to choose
5928 a mode given an operand of a compare. This is needed because the modes
5929 have to be chosen not only during RTL generation but also, for example,
5930 by instruction combination. The result of @code{SELECT_CC_MODE} should
5931 be consistent with the mode used in the patterns; for example to support
5932 the case of the add on the SPARC discussed above, we have the pattern
5936 [(set (reg:CC_NOOV 0)
5938 (plus:SI (match_operand:SI 0 "register_operand" "%r")
5939 (match_operand:SI 1 "arith_operand" "rI"))
5946 together with a @code{SELECT_CC_MODE} that returns @code{CC_NOOVmode}
5947 for comparisons whose argument is a @code{plus}:
5950 #define SELECT_CC_MODE(OP,X,Y) \
5951 (GET_MODE_CLASS (GET_MODE (X)) == MODE_FLOAT \
5952 ? ((OP == EQ || OP == NE) ? CCFPmode : CCFPEmode) \
5953 : ((GET_CODE (X) == PLUS || GET_CODE (X) == MINUS \
5954 || GET_CODE (X) == NEG) \
5955 ? CC_NOOVmode : CCmode))
5958 Another reason to use modes is to retain information on which operands
5959 were used by the comparison; see @code{REVERSIBLE_CC_MODE} later in
5962 You should define this macro if and only if you define extra CC modes
5963 in @file{@var{machine}-modes.def}.
5966 @defmac CANONICALIZE_COMPARISON (@var{code}, @var{op0}, @var{op1})
5967 On some machines not all possible comparisons are defined, but you can
5968 convert an invalid comparison into a valid one. For example, the Alpha
5969 does not have a @code{GT} comparison, but you can use an @code{LT}
5970 comparison instead and swap the order of the operands.
5972 On such machines, define this macro to be a C statement to do any
5973 required conversions. @var{code} is the initial comparison code
5974 and @var{op0} and @var{op1} are the left and right operands of the
5975 comparison, respectively. You should modify @var{code}, @var{op0}, and
5976 @var{op1} as required.
5978 GCC will not assume that the comparison resulting from this macro is
5979 valid but will see if the resulting insn matches a pattern in the
5982 You need not define this macro if it would never change the comparison
5986 @defmac REVERSIBLE_CC_MODE (@var{mode})
5987 A C expression whose value is one if it is always safe to reverse a
5988 comparison whose mode is @var{mode}. If @code{SELECT_CC_MODE}
5989 can ever return @var{mode} for a floating-point inequality comparison,
5990 then @code{REVERSIBLE_CC_MODE (@var{mode})} must be zero.
5992 You need not define this macro if it would always returns zero or if the
5993 floating-point format is anything other than @code{IEEE_FLOAT_FORMAT}.
5994 For example, here is the definition used on the SPARC, where floating-point
5995 inequality comparisons are always given @code{CCFPEmode}:
5998 #define REVERSIBLE_CC_MODE(MODE) ((MODE) != CCFPEmode)
6002 @defmac REVERSE_CONDITION (@var{code}, @var{mode})
6003 A C expression whose value is reversed condition code of the @var{code} for
6004 comparison done in CC_MODE @var{mode}. The macro is used only in case
6005 @code{REVERSIBLE_CC_MODE (@var{mode})} is nonzero. Define this macro in case
6006 machine has some non-standard way how to reverse certain conditionals. For
6007 instance in case all floating point conditions are non-trapping, compiler may
6008 freely convert unordered compares to ordered one. Then definition may look
6012 #define REVERSE_CONDITION(CODE, MODE) \
6013 ((MODE) != CCFPmode ? reverse_condition (CODE) \
6014 : reverse_condition_maybe_unordered (CODE))
6018 @hook TARGET_FIXED_CONDITION_CODE_REGS
6019 On targets which do not use @code{(cc0)}, and which use a hard
6020 register rather than a pseudo-register to hold condition codes, the
6021 regular CSE passes are often not able to identify cases in which the
6022 hard register is set to a common value. Use this hook to enable a
6023 small pass which optimizes such cases. This hook should return true
6024 to enable this pass, and it should set the integers to which its
6025 arguments point to the hard register numbers used for condition codes.
6026 When there is only one such register, as is true on most systems, the
6027 integer pointed to by @var{p2} should be set to
6028 @code{INVALID_REGNUM}.
6030 The default version of this hook returns false.
6033 @hook TARGET_CC_MODES_COMPATIBLE
6034 On targets which use multiple condition code modes in class
6035 @code{MODE_CC}, it is sometimes the case that a comparison can be
6036 validly done in more than one mode. On such a system, define this
6037 target hook to take two mode arguments and to return a mode in which
6038 both comparisons may be validly done. If there is no such mode,
6039 return @code{VOIDmode}.
6041 The default version of this hook checks whether the modes are the
6042 same. If they are, it returns that mode. If they are different, it
6043 returns @code{VOIDmode}.
6046 @node Cond Exec Macros
6047 @subsection Macros to control conditional execution
6048 @findex conditional execution
6051 There is one macro that may need to be defined for targets
6052 supporting conditional execution, independent of how they
6053 represent conditional branches.
6055 @defmac REVERSE_CONDEXEC_PREDICATES_P (@var{op1}, @var{op2})
6056 A C expression that returns true if the conditional execution predicate
6057 @var{op1}, a comparison operation, is the inverse of @var{op2} and vice
6058 versa. Define this to return 0 if the target has conditional execution
6059 predicates that cannot be reversed safely. There is no need to validate
6060 that the arguments of op1 and op2 are the same, this is done separately.
6061 If no expansion is specified, this macro is defined as follows:
6064 #define REVERSE_CONDEXEC_PREDICATES_P (x, y) \
6065 (GET_CODE ((x)) == reversed_comparison_code ((y), NULL))
6070 @section Describing Relative Costs of Operations
6071 @cindex costs of instructions
6072 @cindex relative costs
6073 @cindex speed of instructions
6075 These macros let you describe the relative speed of various operations
6076 on the target machine.
6078 @defmac REGISTER_MOVE_COST (@var{mode}, @var{from}, @var{to})
6079 A C expression for the cost of moving data of mode @var{mode} from a
6080 register in class @var{from} to one in class @var{to}. The classes are
6081 expressed using the enumeration values such as @code{GENERAL_REGS}. A
6082 value of 2 is the default; other values are interpreted relative to
6085 It is not required that the cost always equal 2 when @var{from} is the
6086 same as @var{to}; on some machines it is expensive to move between
6087 registers if they are not general registers.
6089 If reload sees an insn consisting of a single @code{set} between two
6090 hard registers, and if @code{REGISTER_MOVE_COST} applied to their
6091 classes returns a value of 2, reload does not check to ensure that the
6092 constraints of the insn are met. Setting a cost of other than 2 will
6093 allow reload to verify that the constraints are met. You should do this
6094 if the @samp{mov@var{m}} pattern's constraints do not allow such copying.
6096 These macros are obsolete, new ports should use the target hook
6097 @code{TARGET_REGISTER_MOVE_COST} instead.
6100 @hook TARGET_REGISTER_MOVE_COST
6101 This target hook should return the cost of moving data of mode @var{mode}
6102 from a register in class @var{from} to one in class @var{to}. The classes
6103 are expressed using the enumeration values such as @code{GENERAL_REGS}.
6104 A value of 2 is the default; other values are interpreted relative to
6107 It is not required that the cost always equal 2 when @var{from} is the
6108 same as @var{to}; on some machines it is expensive to move between
6109 registers if they are not general registers.
6111 If reload sees an insn consisting of a single @code{set} between two
6112 hard registers, and if @code{TARGET_REGISTER_MOVE_COST} applied to their
6113 classes returns a value of 2, reload does not check to ensure that the
6114 constraints of the insn are met. Setting a cost of other than 2 will
6115 allow reload to verify that the constraints are met. You should do this
6116 if the @samp{mov@var{m}} pattern's constraints do not allow such copying.
6118 The default version of this function returns 2.
6121 @defmac MEMORY_MOVE_COST (@var{mode}, @var{class}, @var{in})
6122 A C expression for the cost of moving data of mode @var{mode} between a
6123 register of class @var{class} and memory; @var{in} is zero if the value
6124 is to be written to memory, nonzero if it is to be read in. This cost
6125 is relative to those in @code{REGISTER_MOVE_COST}. If moving between
6126 registers and memory is more expensive than between two registers, you
6127 should define this macro to express the relative cost.
6129 If you do not define this macro, GCC uses a default cost of 4 plus
6130 the cost of copying via a secondary reload register, if one is
6131 needed. If your machine requires a secondary reload register to copy
6132 between memory and a register of @var{class} but the reload mechanism is
6133 more complex than copying via an intermediate, define this macro to
6134 reflect the actual cost of the move.
6136 GCC defines the function @code{memory_move_secondary_cost} if
6137 secondary reloads are needed. It computes the costs due to copying via
6138 a secondary register. If your machine copies from memory using a
6139 secondary register in the conventional way but the default base value of
6140 4 is not correct for your machine, define this macro to add some other
6141 value to the result of that function. The arguments to that function
6142 are the same as to this macro.
6144 These macros are obsolete, new ports should use the target hook
6145 @code{TARGET_MEMORY_MOVE_COST} instead.
6148 @hook TARGET_MEMORY_MOVE_COST
6149 This target hook should return the cost of moving data of mode @var{mode}
6150 between a register of class @var{rclass} and memory; @var{in} is @code{false}
6151 if the value is to be written to memory, @code{true} if it is to be read in.
6152 This cost is relative to those in @code{TARGET_REGISTER_MOVE_COST}.
6153 If moving between registers and memory is more expensive than between two
6154 registers, you should add this target hook to express the relative cost.
6156 If you do not add this target hook, GCC uses a default cost of 4 plus
6157 the cost of copying via a secondary reload register, if one is
6158 needed. If your machine requires a secondary reload register to copy
6159 between memory and a register of @var{rclass} but the reload mechanism is
6160 more complex than copying via an intermediate, use this target hook to
6161 reflect the actual cost of the move.
6163 GCC defines the function @code{memory_move_secondary_cost} if
6164 secondary reloads are needed. It computes the costs due to copying via
6165 a secondary register. If your machine copies from memory using a
6166 secondary register in the conventional way but the default base value of
6167 4 is not correct for your machine, use this target hook to add some other
6168 value to the result of that function. The arguments to that function
6169 are the same as to this target hook.
6172 @defmac BRANCH_COST (@var{speed_p}, @var{predictable_p})
6173 A C expression for the cost of a branch instruction. A value of 1 is
6174 the default; other values are interpreted relative to that. Parameter
6175 @var{speed_p} is true when the branch in question should be optimized
6176 for speed. When it is false, @code{BRANCH_COST} should return a value
6177 optimal for code size rather than performance. @var{predictable_p} is
6178 true for well-predicted branches. On many architectures the
6179 @code{BRANCH_COST} can be reduced then.
6182 Here are additional macros which do not specify precise relative costs,
6183 but only that certain actions are more expensive than GCC would
6186 @defmac SLOW_BYTE_ACCESS
6187 Define this macro as a C expression which is nonzero if accessing less
6188 than a word of memory (i.e.@: a @code{char} or a @code{short}) is no
6189 faster than accessing a word of memory, i.e., if such access
6190 require more than one instruction or if there is no difference in cost
6191 between byte and (aligned) word loads.
6193 When this macro is not defined, the compiler will access a field by
6194 finding the smallest containing object; when it is defined, a fullword
6195 load will be used if alignment permits. Unless bytes accesses are
6196 faster than word accesses, using word accesses is preferable since it
6197 may eliminate subsequent memory access if subsequent accesses occur to
6198 other fields in the same word of the structure, but to different bytes.
6201 @defmac SLOW_UNALIGNED_ACCESS (@var{mode}, @var{alignment})
6202 Define this macro to be the value 1 if memory accesses described by the
6203 @var{mode} and @var{alignment} parameters have a cost many times greater
6204 than aligned accesses, for example if they are emulated in a trap
6207 When this macro is nonzero, the compiler will act as if
6208 @code{STRICT_ALIGNMENT} were nonzero when generating code for block
6209 moves. This can cause significantly more instructions to be produced.
6210 Therefore, do not set this macro nonzero if unaligned accesses only add a
6211 cycle or two to the time for a memory access.
6213 If the value of this macro is always zero, it need not be defined. If
6214 this macro is defined, it should produce a nonzero value when
6215 @code{STRICT_ALIGNMENT} is nonzero.
6218 @defmac MOVE_RATIO (@var{speed})
6219 The threshold of number of scalar memory-to-memory move insns, @emph{below}
6220 which a sequence of insns should be generated instead of a
6221 string move insn or a library call. Increasing the value will always
6222 make code faster, but eventually incurs high cost in increased code size.
6224 Note that on machines where the corresponding move insn is a
6225 @code{define_expand} that emits a sequence of insns, this macro counts
6226 the number of such sequences.
6228 The parameter @var{speed} is true if the code is currently being
6229 optimized for speed rather than size.
6231 If you don't define this, a reasonable default is used.
6234 @defmac MOVE_BY_PIECES_P (@var{size}, @var{alignment})
6235 A C expression used to determine whether @code{move_by_pieces} will be used to
6236 copy a chunk of memory, or whether some other block move mechanism
6237 will be used. Defaults to 1 if @code{move_by_pieces_ninsns} returns less
6238 than @code{MOVE_RATIO}.
6241 @defmac MOVE_MAX_PIECES
6242 A C expression used by @code{move_by_pieces} to determine the largest unit
6243 a load or store used to copy memory is. Defaults to @code{MOVE_MAX}.
6246 @defmac CLEAR_RATIO (@var{speed})
6247 The threshold of number of scalar move insns, @emph{below} which a sequence
6248 of insns should be generated to clear memory instead of a string clear insn
6249 or a library call. Increasing the value will always make code faster, but
6250 eventually incurs high cost in increased code size.
6252 The parameter @var{speed} is true if the code is currently being
6253 optimized for speed rather than size.
6255 If you don't define this, a reasonable default is used.
6258 @defmac CLEAR_BY_PIECES_P (@var{size}, @var{alignment})
6259 A C expression used to determine whether @code{clear_by_pieces} will be used
6260 to clear a chunk of memory, or whether some other block clear mechanism
6261 will be used. Defaults to 1 if @code{move_by_pieces_ninsns} returns less
6262 than @code{CLEAR_RATIO}.
6265 @defmac SET_RATIO (@var{speed})
6266 The threshold of number of scalar move insns, @emph{below} which a sequence
6267 of insns should be generated to set memory to a constant value, instead of
6268 a block set insn or a library call.
6269 Increasing the value will always make code faster, but
6270 eventually incurs high cost in increased code size.
6272 The parameter @var{speed} is true if the code is currently being
6273 optimized for speed rather than size.
6275 If you don't define this, it defaults to the value of @code{MOVE_RATIO}.
6278 @defmac SET_BY_PIECES_P (@var{size}, @var{alignment})
6279 A C expression used to determine whether @code{store_by_pieces} will be
6280 used to set a chunk of memory to a constant value, or whether some
6281 other mechanism will be used. Used by @code{__builtin_memset} when
6282 storing values other than constant zero.
6283 Defaults to 1 if @code{move_by_pieces_ninsns} returns less
6284 than @code{SET_RATIO}.
6287 @defmac STORE_BY_PIECES_P (@var{size}, @var{alignment})
6288 A C expression used to determine whether @code{store_by_pieces} will be
6289 used to set a chunk of memory to a constant string value, or whether some
6290 other mechanism will be used. Used by @code{__builtin_strcpy} when
6291 called with a constant source string.
6292 Defaults to 1 if @code{move_by_pieces_ninsns} returns less
6293 than @code{MOVE_RATIO}.
6296 @defmac USE_LOAD_POST_INCREMENT (@var{mode})
6297 A C expression used to determine whether a load postincrement is a good
6298 thing to use for a given mode. Defaults to the value of
6299 @code{HAVE_POST_INCREMENT}.
6302 @defmac USE_LOAD_POST_DECREMENT (@var{mode})
6303 A C expression used to determine whether a load postdecrement is a good
6304 thing to use for a given mode. Defaults to the value of
6305 @code{HAVE_POST_DECREMENT}.
6308 @defmac USE_LOAD_PRE_INCREMENT (@var{mode})
6309 A C expression used to determine whether a load preincrement is a good
6310 thing to use for a given mode. Defaults to the value of
6311 @code{HAVE_PRE_INCREMENT}.
6314 @defmac USE_LOAD_PRE_DECREMENT (@var{mode})
6315 A C expression used to determine whether a load predecrement is a good
6316 thing to use for a given mode. Defaults to the value of
6317 @code{HAVE_PRE_DECREMENT}.
6320 @defmac USE_STORE_POST_INCREMENT (@var{mode})
6321 A C expression used to determine whether a store postincrement is a good
6322 thing to use for a given mode. Defaults to the value of
6323 @code{HAVE_POST_INCREMENT}.
6326 @defmac USE_STORE_POST_DECREMENT (@var{mode})
6327 A C expression used to determine whether a store postdecrement is a good
6328 thing to use for a given mode. Defaults to the value of
6329 @code{HAVE_POST_DECREMENT}.
6332 @defmac USE_STORE_PRE_INCREMENT (@var{mode})
6333 This macro is used to determine whether a store preincrement is a good
6334 thing to use for a given mode. Defaults to the value of
6335 @code{HAVE_PRE_INCREMENT}.
6338 @defmac USE_STORE_PRE_DECREMENT (@var{mode})
6339 This macro is used to determine whether a store predecrement is a good
6340 thing to use for a given mode. Defaults to the value of
6341 @code{HAVE_PRE_DECREMENT}.
6344 @defmac NO_FUNCTION_CSE
6345 Define this macro if it is as good or better to call a constant
6346 function address than to call an address kept in a register.
6349 @defmac RANGE_TEST_NON_SHORT_CIRCUIT
6350 Define this macro if a non-short-circuit operation produced by
6351 @samp{fold_range_test ()} is optimal. This macro defaults to true if
6352 @code{BRANCH_COST} is greater than or equal to the value 2.
6355 @hook TARGET_RTX_COSTS
6356 This target hook describes the relative costs of RTL expressions.
6358 The cost may depend on the precise form of the expression, which is
6359 available for examination in @var{x}, and the fact that @var{x} appears
6360 as operand @var{opno} of an expression with rtx code @var{outer_code}.
6361 That is, the hook can assume that there is some rtx @var{y} such
6362 that @samp{GET_CODE (@var{y}) == @var{outer_code}} and such that
6363 either (a) @samp{XEXP (@var{y}, @var{opno}) == @var{x}} or
6364 (b) @samp{XVEC (@var{y}, @var{opno})} contains @var{x}.
6366 @var{code} is @var{x}'s expression code---redundant, since it can be
6367 obtained with @code{GET_CODE (@var{x})}.
6369 In implementing this hook, you can use the construct
6370 @code{COSTS_N_INSNS (@var{n})} to specify a cost equal to @var{n} fast
6373 On entry to the hook, @code{*@var{total}} contains a default estimate
6374 for the cost of the expression. The hook should modify this value as
6375 necessary. Traditionally, the default costs are @code{COSTS_N_INSNS (5)}
6376 for multiplications, @code{COSTS_N_INSNS (7)} for division and modulus
6377 operations, and @code{COSTS_N_INSNS (1)} for all other operations.
6379 When optimizing for code size, i.e.@: when @code{speed} is
6380 false, this target hook should be used to estimate the relative
6381 size cost of an expression, again relative to @code{COSTS_N_INSNS}.
6383 The hook returns true when all subexpressions of @var{x} have been
6384 processed, and false when @code{rtx_cost} should recurse.
6387 @hook TARGET_ADDRESS_COST
6388 This hook computes the cost of an addressing mode that contains
6389 @var{address}. If not defined, the cost is computed from
6390 the @var{address} expression and the @code{TARGET_RTX_COST} hook.
6392 For most CISC machines, the default cost is a good approximation of the
6393 true cost of the addressing mode. However, on RISC machines, all
6394 instructions normally have the same length and execution time. Hence
6395 all addresses will have equal costs.
6397 In cases where more than one form of an address is known, the form with
6398 the lowest cost will be used. If multiple forms have the same, lowest,
6399 cost, the one that is the most complex will be used.
6401 For example, suppose an address that is equal to the sum of a register
6402 and a constant is used twice in the same basic block. When this macro
6403 is not defined, the address will be computed in a register and memory
6404 references will be indirect through that register. On machines where
6405 the cost of the addressing mode containing the sum is no higher than
6406 that of a simple indirect reference, this will produce an additional
6407 instruction and possibly require an additional register. Proper
6408 specification of this macro eliminates this overhead for such machines.
6410 This hook is never called with an invalid address.
6412 On machines where an address involving more than one register is as
6413 cheap as an address computation involving only one register, defining
6414 @code{TARGET_ADDRESS_COST} to reflect this can cause two registers to
6415 be live over a region of code where only one would have been if
6416 @code{TARGET_ADDRESS_COST} were not defined in that manner. This effect
6417 should be considered in the definition of this macro. Equivalent costs
6418 should probably only be given to addresses with different numbers of
6419 registers on machines with lots of registers.
6423 @section Adjusting the Instruction Scheduler
6425 The instruction scheduler may need a fair amount of machine-specific
6426 adjustment in order to produce good code. GCC provides several target
6427 hooks for this purpose. It is usually enough to define just a few of
6428 them: try the first ones in this list first.
6430 @hook TARGET_SCHED_ISSUE_RATE
6431 This hook returns the maximum number of instructions that can ever
6432 issue at the same time on the target machine. The default is one.
6433 Although the insn scheduler can define itself the possibility of issue
6434 an insn on the same cycle, the value can serve as an additional
6435 constraint to issue insns on the same simulated processor cycle (see
6436 hooks @samp{TARGET_SCHED_REORDER} and @samp{TARGET_SCHED_REORDER2}).
6437 This value must be constant over the entire compilation. If you need
6438 it to vary depending on what the instructions are, you must use
6439 @samp{TARGET_SCHED_VARIABLE_ISSUE}.
6442 @hook TARGET_SCHED_VARIABLE_ISSUE
6443 This hook is executed by the scheduler after it has scheduled an insn
6444 from the ready list. It should return the number of insns which can
6445 still be issued in the current cycle. The default is
6446 @samp{@w{@var{more} - 1}} for insns other than @code{CLOBBER} and
6447 @code{USE}, which normally are not counted against the issue rate.
6448 You should define this hook if some insns take more machine resources
6449 than others, so that fewer insns can follow them in the same cycle.
6450 @var{file} is either a null pointer, or a stdio stream to write any
6451 debug output to. @var{verbose} is the verbose level provided by
6452 @option{-fsched-verbose-@var{n}}. @var{insn} is the instruction that
6456 @hook TARGET_SCHED_ADJUST_COST
6457 This function corrects the value of @var{cost} based on the
6458 relationship between @var{insn} and @var{dep_insn} through the
6459 dependence @var{link}. It should return the new value. The default
6460 is to make no adjustment to @var{cost}. This can be used for example
6461 to specify to the scheduler using the traditional pipeline description
6462 that an output- or anti-dependence does not incur the same cost as a
6463 data-dependence. If the scheduler using the automaton based pipeline
6464 description, the cost of anti-dependence is zero and the cost of
6465 output-dependence is maximum of one and the difference of latency
6466 times of the first and the second insns. If these values are not
6467 acceptable, you could use the hook to modify them too. See also
6468 @pxref{Processor pipeline description}.
6471 @hook TARGET_SCHED_ADJUST_PRIORITY
6472 This hook adjusts the integer scheduling priority @var{priority} of
6473 @var{insn}. It should return the new priority. Increase the priority to
6474 execute @var{insn} earlier, reduce the priority to execute @var{insn}
6475 later. Do not define this hook if you do not need to adjust the
6476 scheduling priorities of insns.
6479 @hook TARGET_SCHED_REORDER
6480 This hook is executed by the scheduler after it has scheduled the ready
6481 list, to allow the machine description to reorder it (for example to
6482 combine two small instructions together on @samp{VLIW} machines).
6483 @var{file} is either a null pointer, or a stdio stream to write any
6484 debug output to. @var{verbose} is the verbose level provided by
6485 @option{-fsched-verbose-@var{n}}. @var{ready} is a pointer to the ready
6486 list of instructions that are ready to be scheduled. @var{n_readyp} is
6487 a pointer to the number of elements in the ready list. The scheduler
6488 reads the ready list in reverse order, starting with
6489 @var{ready}[@var{*n_readyp} @minus{} 1] and going to @var{ready}[0]. @var{clock}
6490 is the timer tick of the scheduler. You may modify the ready list and
6491 the number of ready insns. The return value is the number of insns that
6492 can issue this cycle; normally this is just @code{issue_rate}. See also
6493 @samp{TARGET_SCHED_REORDER2}.
6496 @hook TARGET_SCHED_REORDER2
6497 Like @samp{TARGET_SCHED_REORDER}, but called at a different time. That
6498 function is called whenever the scheduler starts a new cycle. This one
6499 is called once per iteration over a cycle, immediately after
6500 @samp{TARGET_SCHED_VARIABLE_ISSUE}; it can reorder the ready list and
6501 return the number of insns to be scheduled in the same cycle. Defining
6502 this hook can be useful if there are frequent situations where
6503 scheduling one insn causes other insns to become ready in the same
6504 cycle. These other insns can then be taken into account properly.
6507 @hook TARGET_SCHED_DEPENDENCIES_EVALUATION_HOOK
6508 This hook is called after evaluation forward dependencies of insns in
6509 chain given by two parameter values (@var{head} and @var{tail}
6510 correspondingly) but before insns scheduling of the insn chain. For
6511 example, it can be used for better insn classification if it requires
6512 analysis of dependencies. This hook can use backward and forward
6513 dependencies of the insn scheduler because they are already
6517 @hook TARGET_SCHED_INIT
6518 This hook is executed by the scheduler at the beginning of each block of
6519 instructions that are to be scheduled. @var{file} is either a null
6520 pointer, or a stdio stream to write any debug output to. @var{verbose}
6521 is the verbose level provided by @option{-fsched-verbose-@var{n}}.
6522 @var{max_ready} is the maximum number of insns in the current scheduling
6523 region that can be live at the same time. This can be used to allocate
6524 scratch space if it is needed, e.g.@: by @samp{TARGET_SCHED_REORDER}.
6527 @hook TARGET_SCHED_FINISH
6528 This hook is executed by the scheduler at the end of each block of
6529 instructions that are to be scheduled. It can be used to perform
6530 cleanup of any actions done by the other scheduling hooks. @var{file}
6531 is either a null pointer, or a stdio stream to write any debug output
6532 to. @var{verbose} is the verbose level provided by
6533 @option{-fsched-verbose-@var{n}}.
6536 @hook TARGET_SCHED_INIT_GLOBAL
6537 This hook is executed by the scheduler after function level initializations.
6538 @var{file} is either a null pointer, or a stdio stream to write any debug output to.
6539 @var{verbose} is the verbose level provided by @option{-fsched-verbose-@var{n}}.
6540 @var{old_max_uid} is the maximum insn uid when scheduling begins.
6543 @hook TARGET_SCHED_FINISH_GLOBAL
6544 This is the cleanup hook corresponding to @code{TARGET_SCHED_INIT_GLOBAL}.
6545 @var{file} is either a null pointer, or a stdio stream to write any debug output to.
6546 @var{verbose} is the verbose level provided by @option{-fsched-verbose-@var{n}}.
6549 @hook TARGET_SCHED_DFA_PRE_CYCLE_INSN
6550 The hook returns an RTL insn. The automaton state used in the
6551 pipeline hazard recognizer is changed as if the insn were scheduled
6552 when the new simulated processor cycle starts. Usage of the hook may
6553 simplify the automaton pipeline description for some @acronym{VLIW}
6554 processors. If the hook is defined, it is used only for the automaton
6555 based pipeline description. The default is not to change the state
6556 when the new simulated processor cycle starts.
6559 @hook TARGET_SCHED_INIT_DFA_PRE_CYCLE_INSN
6560 The hook can be used to initialize data used by the previous hook.
6563 @hook TARGET_SCHED_DFA_POST_CYCLE_INSN
6564 The hook is analogous to @samp{TARGET_SCHED_DFA_PRE_CYCLE_INSN} but used
6565 to changed the state as if the insn were scheduled when the new
6566 simulated processor cycle finishes.
6569 @hook TARGET_SCHED_INIT_DFA_POST_CYCLE_INSN
6570 The hook is analogous to @samp{TARGET_SCHED_INIT_DFA_PRE_CYCLE_INSN} but
6571 used to initialize data used by the previous hook.
6574 @hook TARGET_SCHED_DFA_PRE_ADVANCE_CYCLE
6575 The hook to notify target that the current simulated cycle is about to finish.
6576 The hook is analogous to @samp{TARGET_SCHED_DFA_PRE_CYCLE_INSN} but used
6577 to change the state in more complicated situations - e.g., when advancing
6578 state on a single insn is not enough.
6581 @hook TARGET_SCHED_DFA_POST_ADVANCE_CYCLE
6582 The hook to notify target that new simulated cycle has just started.
6583 The hook is analogous to @samp{TARGET_SCHED_DFA_POST_CYCLE_INSN} but used
6584 to change the state in more complicated situations - e.g., when advancing
6585 state on a single insn is not enough.
6588 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD
6589 This hook controls better choosing an insn from the ready insn queue
6590 for the @acronym{DFA}-based insn scheduler. Usually the scheduler
6591 chooses the first insn from the queue. If the hook returns a positive
6592 value, an additional scheduler code tries all permutations of
6593 @samp{TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD ()}
6594 subsequent ready insns to choose an insn whose issue will result in
6595 maximal number of issued insns on the same cycle. For the
6596 @acronym{VLIW} processor, the code could actually solve the problem of
6597 packing simple insns into the @acronym{VLIW} insn. Of course, if the
6598 rules of @acronym{VLIW} packing are described in the automaton.
6600 This code also could be used for superscalar @acronym{RISC}
6601 processors. Let us consider a superscalar @acronym{RISC} processor
6602 with 3 pipelines. Some insns can be executed in pipelines @var{A} or
6603 @var{B}, some insns can be executed only in pipelines @var{B} or
6604 @var{C}, and one insn can be executed in pipeline @var{B}. The
6605 processor may issue the 1st insn into @var{A} and the 2nd one into
6606 @var{B}. In this case, the 3rd insn will wait for freeing @var{B}
6607 until the next cycle. If the scheduler issues the 3rd insn the first,
6608 the processor could issue all 3 insns per cycle.
6610 Actually this code demonstrates advantages of the automaton based
6611 pipeline hazard recognizer. We try quickly and easy many insn
6612 schedules to choose the best one.
6614 The default is no multipass scheduling.
6617 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD
6619 This hook controls what insns from the ready insn queue will be
6620 considered for the multipass insn scheduling. If the hook returns
6621 zero for @var{insn}, the insn will be not chosen to
6624 The default is that any ready insns can be chosen to be issued.
6627 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_BEGIN
6628 This hook prepares the target backend for a new round of multipass
6632 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_ISSUE
6633 This hook is called when multipass scheduling evaluates instruction INSN.
6636 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_BACKTRACK
6637 This is called when multipass scheduling backtracks from evaluation of
6641 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_END
6642 This hook notifies the target about the result of the concluded current
6643 round of multipass scheduling.
6646 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_INIT
6647 This hook initializes target-specific data used in multipass scheduling.
6650 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_FINI
6651 This hook finalizes target-specific data used in multipass scheduling.
6654 @hook TARGET_SCHED_DFA_NEW_CYCLE
6655 This hook is called by the insn scheduler before issuing @var{insn}
6656 on cycle @var{clock}. If the hook returns nonzero,
6657 @var{insn} is not issued on this processor cycle. Instead,
6658 the processor cycle is advanced. If *@var{sort_p}
6659 is zero, the insn ready queue is not sorted on the new cycle
6660 start as usually. @var{dump} and @var{verbose} specify the file and
6661 verbosity level to use for debugging output.
6662 @var{last_clock} and @var{clock} are, respectively, the
6663 processor cycle on which the previous insn has been issued,
6664 and the current processor cycle.
6667 @hook TARGET_SCHED_IS_COSTLY_DEPENDENCE
6668 This hook is used to define which dependences are considered costly by
6669 the target, so costly that it is not advisable to schedule the insns that
6670 are involved in the dependence too close to one another. The parameters
6671 to this hook are as follows: The first parameter @var{_dep} is the dependence
6672 being evaluated. The second parameter @var{cost} is the cost of the
6673 dependence as estimated by the scheduler, and the third
6674 parameter @var{distance} is the distance in cycles between the two insns.
6675 The hook returns @code{true} if considering the distance between the two
6676 insns the dependence between them is considered costly by the target,
6677 and @code{false} otherwise.
6679 Defining this hook can be useful in multiple-issue out-of-order machines,
6680 where (a) it's practically hopeless to predict the actual data/resource
6681 delays, however: (b) there's a better chance to predict the actual grouping
6682 that will be formed, and (c) correctly emulating the grouping can be very
6683 important. In such targets one may want to allow issuing dependent insns
6684 closer to one another---i.e., closer than the dependence distance; however,
6685 not in cases of ``costly dependences'', which this hooks allows to define.
6688 @hook TARGET_SCHED_H_I_D_EXTENDED
6689 This hook is called by the insn scheduler after emitting a new instruction to
6690 the instruction stream. The hook notifies a target backend to extend its
6691 per instruction data structures.
6694 @hook TARGET_SCHED_ALLOC_SCHED_CONTEXT
6695 Return a pointer to a store large enough to hold target scheduling context.
6698 @hook TARGET_SCHED_INIT_SCHED_CONTEXT
6699 Initialize store pointed to by @var{tc} to hold target scheduling context.
6700 It @var{clean_p} is true then initialize @var{tc} as if scheduler is at the
6701 beginning of the block. Otherwise, copy the current context into @var{tc}.
6704 @hook TARGET_SCHED_SET_SCHED_CONTEXT
6705 Copy target scheduling context pointed to by @var{tc} to the current context.
6708 @hook TARGET_SCHED_CLEAR_SCHED_CONTEXT
6709 Deallocate internal data in target scheduling context pointed to by @var{tc}.
6712 @hook TARGET_SCHED_FREE_SCHED_CONTEXT
6713 Deallocate a store for target scheduling context pointed to by @var{tc}.
6716 @hook TARGET_SCHED_SPECULATE_INSN
6717 This hook is called by the insn scheduler when @var{insn} has only
6718 speculative dependencies and therefore can be scheduled speculatively.
6719 The hook is used to check if the pattern of @var{insn} has a speculative
6720 version and, in case of successful check, to generate that speculative
6721 pattern. The hook should return 1, if the instruction has a speculative form,
6722 or @minus{}1, if it doesn't. @var{request} describes the type of requested
6723 speculation. If the return value equals 1 then @var{new_pat} is assigned
6724 the generated speculative pattern.
6727 @hook TARGET_SCHED_NEEDS_BLOCK_P
6728 This hook is called by the insn scheduler during generation of recovery code
6729 for @var{insn}. It should return @code{true}, if the corresponding check
6730 instruction should branch to recovery code, or @code{false} otherwise.
6733 @hook TARGET_SCHED_GEN_SPEC_CHECK
6734 This hook is called by the insn scheduler to generate a pattern for recovery
6735 check instruction. If @var{mutate_p} is zero, then @var{insn} is a
6736 speculative instruction for which the check should be generated.
6737 @var{label} is either a label of a basic block, where recovery code should
6738 be emitted, or a null pointer, when requested check doesn't branch to
6739 recovery code (a simple check). If @var{mutate_p} is nonzero, then
6740 a pattern for a branchy check corresponding to a simple check denoted by
6741 @var{insn} should be generated. In this case @var{label} can't be null.
6744 @hook TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC
6745 This hook is used as a workaround for
6746 @samp{TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD} not being
6747 called on the first instruction of the ready list. The hook is used to
6748 discard speculative instructions that stand first in the ready list from
6749 being scheduled on the current cycle. If the hook returns @code{false},
6750 @var{insn} will not be chosen to be issued.
6751 For non-speculative instructions,
6752 the hook should always return @code{true}. For example, in the ia64 backend
6753 the hook is used to cancel data speculative insns when the ALAT table
6757 @hook TARGET_SCHED_SET_SCHED_FLAGS
6758 This hook is used by the insn scheduler to find out what features should be
6760 The structure *@var{spec_info} should be filled in by the target.
6761 The structure describes speculation types that can be used in the scheduler.
6764 @hook TARGET_SCHED_SMS_RES_MII
6765 This hook is called by the swing modulo scheduler to calculate a
6766 resource-based lower bound which is based on the resources available in
6767 the machine and the resources required by each instruction. The target
6768 backend can use @var{g} to calculate such bound. A very simple lower
6769 bound will be used in case this hook is not implemented: the total number
6770 of instructions divided by the issue rate.
6773 @hook TARGET_SCHED_DISPATCH
6774 This hook is called by Haifa Scheduler. It returns true if dispatch scheduling
6775 is supported in hardware and the condition specified in the parameter is true.
6778 @hook TARGET_SCHED_DISPATCH_DO
6779 This hook is called by Haifa Scheduler. It performs the operation specified
6780 in its second parameter.
6783 @hook TARGET_SCHED_EXPOSED_PIPELINE
6785 @hook TARGET_SCHED_REASSOCIATION_WIDTH
6788 @section Dividing the Output into Sections (Texts, Data, @dots{})
6789 @c the above section title is WAY too long. maybe cut the part between
6790 @c the (...)? --mew 10feb93
6792 An object file is divided into sections containing different types of
6793 data. In the most common case, there are three sections: the @dfn{text
6794 section}, which holds instructions and read-only data; the @dfn{data
6795 section}, which holds initialized writable data; and the @dfn{bss
6796 section}, which holds uninitialized data. Some systems have other kinds
6799 @file{varasm.c} provides several well-known sections, such as
6800 @code{text_section}, @code{data_section} and @code{bss_section}.
6801 The normal way of controlling a @code{@var{foo}_section} variable
6802 is to define the associated @code{@var{FOO}_SECTION_ASM_OP} macro,
6803 as described below. The macros are only read once, when @file{varasm.c}
6804 initializes itself, so their values must be run-time constants.
6805 They may however depend on command-line flags.
6807 @emph{Note:} Some run-time files, such @file{crtstuff.c}, also make
6808 use of the @code{@var{FOO}_SECTION_ASM_OP} macros, and expect them
6809 to be string literals.
6811 Some assemblers require a different string to be written every time a
6812 section is selected. If your assembler falls into this category, you
6813 should define the @code{TARGET_ASM_INIT_SECTIONS} hook and use
6814 @code{get_unnamed_section} to set up the sections.
6816 You must always create a @code{text_section}, either by defining
6817 @code{TEXT_SECTION_ASM_OP} or by initializing @code{text_section}
6818 in @code{TARGET_ASM_INIT_SECTIONS}. The same is true of
6819 @code{data_section} and @code{DATA_SECTION_ASM_OP}. If you do not
6820 create a distinct @code{readonly_data_section}, the default is to
6821 reuse @code{text_section}.
6823 All the other @file{varasm.c} sections are optional, and are null
6824 if the target does not provide them.
6826 @defmac TEXT_SECTION_ASM_OP
6827 A C expression whose value is a string, including spacing, containing the
6828 assembler operation that should precede instructions and read-only data.
6829 Normally @code{"\t.text"} is right.
6832 @defmac HOT_TEXT_SECTION_NAME
6833 If defined, a C string constant for the name of the section containing most
6834 frequently executed functions of the program. If not defined, GCC will provide
6835 a default definition if the target supports named sections.
6838 @defmac UNLIKELY_EXECUTED_TEXT_SECTION_NAME
6839 If defined, a C string constant for the name of the section containing unlikely
6840 executed functions in the program.
6843 @defmac DATA_SECTION_ASM_OP
6844 A C expression whose value is a string, including spacing, containing the
6845 assembler operation to identify the following data as writable initialized
6846 data. Normally @code{"\t.data"} is right.
6849 @defmac SDATA_SECTION_ASM_OP
6850 If defined, a C expression whose value is a string, including spacing,
6851 containing the assembler operation to identify the following data as
6852 initialized, writable small data.
6855 @defmac READONLY_DATA_SECTION_ASM_OP
6856 A C expression whose value is a string, including spacing, containing the
6857 assembler operation to identify the following data as read-only initialized
6861 @defmac BSS_SECTION_ASM_OP
6862 If defined, a C expression whose value is a string, including spacing,
6863 containing the assembler operation to identify the following data as
6864 uninitialized global data. If not defined, and
6865 @code{ASM_OUTPUT_ALIGNED_BSS} not defined,
6866 uninitialized global data will be output in the data section if
6867 @option{-fno-common} is passed, otherwise @code{ASM_OUTPUT_COMMON} will be
6871 @defmac SBSS_SECTION_ASM_OP
6872 If defined, a C expression whose value is a string, including spacing,
6873 containing the assembler operation to identify the following data as
6874 uninitialized, writable small data.
6877 @defmac TLS_COMMON_ASM_OP
6878 If defined, a C expression whose value is a string containing the
6879 assembler operation to identify the following data as thread-local
6880 common data. The default is @code{".tls_common"}.
6883 @defmac TLS_SECTION_ASM_FLAG
6884 If defined, a C expression whose value is a character constant
6885 containing the flag used to mark a section as a TLS section. The
6886 default is @code{'T'}.
6889 @defmac INIT_SECTION_ASM_OP
6890 If defined, a C expression whose value is a string, including spacing,
6891 containing the assembler operation to identify the following data as
6892 initialization code. If not defined, GCC will assume such a section does
6893 not exist. This section has no corresponding @code{init_section}
6894 variable; it is used entirely in runtime code.
6897 @defmac FINI_SECTION_ASM_OP
6898 If defined, a C expression whose value is a string, including spacing,
6899 containing the assembler operation to identify the following data as
6900 finalization code. If not defined, GCC will assume such a section does
6901 not exist. This section has no corresponding @code{fini_section}
6902 variable; it is used entirely in runtime code.
6905 @defmac INIT_ARRAY_SECTION_ASM_OP
6906 If defined, a C expression whose value is a string, including spacing,
6907 containing the assembler operation to identify the following data as
6908 part of the @code{.init_array} (or equivalent) section. If not
6909 defined, GCC will assume such a section does not exist. Do not define
6910 both this macro and @code{INIT_SECTION_ASM_OP}.
6913 @defmac FINI_ARRAY_SECTION_ASM_OP
6914 If defined, a C expression whose value is a string, including spacing,
6915 containing the assembler operation to identify the following data as
6916 part of the @code{.fini_array} (or equivalent) section. If not
6917 defined, GCC will assume such a section does not exist. Do not define
6918 both this macro and @code{FINI_SECTION_ASM_OP}.
6921 @defmac CRT_CALL_STATIC_FUNCTION (@var{section_op}, @var{function})
6922 If defined, an ASM statement that switches to a different section
6923 via @var{section_op}, calls @var{function}, and switches back to
6924 the text section. This is used in @file{crtstuff.c} if
6925 @code{INIT_SECTION_ASM_OP} or @code{FINI_SECTION_ASM_OP} to calls
6926 to initialization and finalization functions from the init and fini
6927 sections. By default, this macro uses a simple function call. Some
6928 ports need hand-crafted assembly code to avoid dependencies on
6929 registers initialized in the function prologue or to ensure that
6930 constant pools don't end up too far way in the text section.
6933 @defmac TARGET_LIBGCC_SDATA_SECTION
6934 If defined, a string which names the section into which small
6935 variables defined in crtstuff and libgcc should go. This is useful
6936 when the target has options for optimizing access to small data, and
6937 you want the crtstuff and libgcc routines to be conservative in what
6938 they expect of your application yet liberal in what your application
6939 expects. For example, for targets with a @code{.sdata} section (like
6940 MIPS), you could compile crtstuff with @code{-G 0} so that it doesn't
6941 require small data support from your application, but use this macro
6942 to put small data into @code{.sdata} so that your application can
6943 access these variables whether it uses small data or not.
6946 @defmac FORCE_CODE_SECTION_ALIGN
6947 If defined, an ASM statement that aligns a code section to some
6948 arbitrary boundary. This is used to force all fragments of the
6949 @code{.init} and @code{.fini} sections to have to same alignment
6950 and thus prevent the linker from having to add any padding.
6953 @defmac JUMP_TABLES_IN_TEXT_SECTION
6954 Define this macro to be an expression with a nonzero value if jump
6955 tables (for @code{tablejump} insns) should be output in the text
6956 section, along with the assembler instructions. Otherwise, the
6957 readonly data section is used.
6959 This macro is irrelevant if there is no separate readonly data section.
6962 @hook TARGET_ASM_INIT_SECTIONS
6963 Define this hook if you need to do something special to set up the
6964 @file{varasm.c} sections, or if your target has some special sections
6965 of its own that you need to create.
6967 GCC calls this hook after processing the command line, but before writing
6968 any assembly code, and before calling any of the section-returning hooks
6972 @hook TARGET_ASM_RELOC_RW_MASK
6973 Return a mask describing how relocations should be treated when
6974 selecting sections. Bit 1 should be set if global relocations
6975 should be placed in a read-write section; bit 0 should be set if
6976 local relocations should be placed in a read-write section.
6978 The default version of this function returns 3 when @option{-fpic}
6979 is in effect, and 0 otherwise. The hook is typically redefined
6980 when the target cannot support (some kinds of) dynamic relocations
6981 in read-only sections even in executables.
6984 @hook TARGET_ASM_SELECT_SECTION
6985 Return the section into which @var{exp} should be placed. You can
6986 assume that @var{exp} is either a @code{VAR_DECL} node or a constant of
6987 some sort. @var{reloc} indicates whether the initial value of @var{exp}
6988 requires link-time relocations. Bit 0 is set when variable contains
6989 local relocations only, while bit 1 is set for global relocations.
6990 @var{align} is the constant alignment in bits.
6992 The default version of this function takes care of putting read-only
6993 variables in @code{readonly_data_section}.
6995 See also @var{USE_SELECT_SECTION_FOR_FUNCTIONS}.
6998 @defmac USE_SELECT_SECTION_FOR_FUNCTIONS
6999 Define this macro if you wish TARGET_ASM_SELECT_SECTION to be called
7000 for @code{FUNCTION_DECL}s as well as for variables and constants.
7002 In the case of a @code{FUNCTION_DECL}, @var{reloc} will be zero if the
7003 function has been determined to be likely to be called, and nonzero if
7004 it is unlikely to be called.
7007 @hook TARGET_ASM_UNIQUE_SECTION
7008 Build up a unique section name, expressed as a @code{STRING_CST} node,
7009 and assign it to @samp{DECL_SECTION_NAME (@var{decl})}.
7010 As with @code{TARGET_ASM_SELECT_SECTION}, @var{reloc} indicates whether
7011 the initial value of @var{exp} requires link-time relocations.
7013 The default version of this function appends the symbol name to the
7014 ELF section name that would normally be used for the symbol. For
7015 example, the function @code{foo} would be placed in @code{.text.foo}.
7016 Whatever the actual target object format, this is often good enough.
7019 @hook TARGET_ASM_FUNCTION_RODATA_SECTION
7020 Return the readonly data section associated with
7021 @samp{DECL_SECTION_NAME (@var{decl})}.
7022 The default version of this function selects @code{.gnu.linkonce.r.name} if
7023 the function's section is @code{.gnu.linkonce.t.name}, @code{.rodata.name}
7024 if function is in @code{.text.name}, and the normal readonly-data section
7028 @hook TARGET_ASM_MERGEABLE_RODATA_PREFIX
7030 @hook TARGET_ASM_TM_CLONE_TABLE_SECTION
7032 @hook TARGET_ASM_SELECT_RTX_SECTION
7033 Return the section into which a constant @var{x}, of mode @var{mode},
7034 should be placed. You can assume that @var{x} is some kind of
7035 constant in RTL@. The argument @var{mode} is redundant except in the
7036 case of a @code{const_int} rtx. @var{align} is the constant alignment
7039 The default version of this function takes care of putting symbolic
7040 constants in @code{flag_pic} mode in @code{data_section} and everything
7041 else in @code{readonly_data_section}.
7044 @hook TARGET_MANGLE_DECL_ASSEMBLER_NAME
7045 Define this hook if you need to postprocess the assembler name generated
7046 by target-independent code. The @var{id} provided to this hook will be
7047 the computed name (e.g., the macro @code{DECL_NAME} of the @var{decl} in C,
7048 or the mangled name of the @var{decl} in C++). The return value of the
7049 hook is an @code{IDENTIFIER_NODE} for the appropriate mangled name on
7050 your target system. The default implementation of this hook just
7051 returns the @var{id} provided.
7054 @hook TARGET_ENCODE_SECTION_INFO
7055 Define this hook if references to a symbol or a constant must be
7056 treated differently depending on something about the variable or
7057 function named by the symbol (such as what section it is in).
7059 The hook is executed immediately after rtl has been created for
7060 @var{decl}, which may be a variable or function declaration or
7061 an entry in the constant pool. In either case, @var{rtl} is the
7062 rtl in question. Do @emph{not} use @code{DECL_RTL (@var{decl})}
7063 in this hook; that field may not have been initialized yet.
7065 In the case of a constant, it is safe to assume that the rtl is
7066 a @code{mem} whose address is a @code{symbol_ref}. Most decls
7067 will also have this form, but that is not guaranteed. Global
7068 register variables, for instance, will have a @code{reg} for their
7069 rtl. (Normally the right thing to do with such unusual rtl is
7072 The @var{new_decl_p} argument will be true if this is the first time
7073 that @code{TARGET_ENCODE_SECTION_INFO} has been invoked on this decl. It will
7074 be false for subsequent invocations, which will happen for duplicate
7075 declarations. Whether or not anything must be done for the duplicate
7076 declaration depends on whether the hook examines @code{DECL_ATTRIBUTES}.
7077 @var{new_decl_p} is always true when the hook is called for a constant.
7079 @cindex @code{SYMBOL_REF_FLAG}, in @code{TARGET_ENCODE_SECTION_INFO}
7080 The usual thing for this hook to do is to record flags in the
7081 @code{symbol_ref}, using @code{SYMBOL_REF_FLAG} or @code{SYMBOL_REF_FLAGS}.
7082 Historically, the name string was modified if it was necessary to
7083 encode more than one bit of information, but this practice is now
7084 discouraged; use @code{SYMBOL_REF_FLAGS}.
7086 The default definition of this hook, @code{default_encode_section_info}
7087 in @file{varasm.c}, sets a number of commonly-useful bits in
7088 @code{SYMBOL_REF_FLAGS}. Check whether the default does what you need
7089 before overriding it.
7092 @hook TARGET_STRIP_NAME_ENCODING
7093 Decode @var{name} and return the real name part, sans
7094 the characters that @code{TARGET_ENCODE_SECTION_INFO}
7098 @hook TARGET_IN_SMALL_DATA_P
7099 Returns true if @var{exp} should be placed into a ``small data'' section.
7100 The default version of this hook always returns false.
7103 @hook TARGET_HAVE_SRODATA_SECTION
7104 Contains the value true if the target places read-only
7105 ``small data'' into a separate section. The default value is false.
7108 @hook TARGET_PROFILE_BEFORE_PROLOGUE
7110 @hook TARGET_BINDS_LOCAL_P
7111 Returns true if @var{exp} names an object for which name resolution
7112 rules must resolve to the current ``module'' (dynamic shared library
7113 or executable image).
7115 The default version of this hook implements the name resolution rules
7116 for ELF, which has a looser model of global name binding than other
7117 currently supported object file formats.
7120 @hook TARGET_HAVE_TLS
7121 Contains the value true if the target supports thread-local storage.
7122 The default value is false.
7127 @section Position Independent Code
7128 @cindex position independent code
7131 This section describes macros that help implement generation of position
7132 independent code. Simply defining these macros is not enough to
7133 generate valid PIC; you must also add support to the hook
7134 @code{TARGET_LEGITIMATE_ADDRESS_P} and to the macro
7135 @code{PRINT_OPERAND_ADDRESS}, as well as @code{LEGITIMIZE_ADDRESS}. You
7136 must modify the definition of @samp{movsi} to do something appropriate
7137 when the source operand contains a symbolic address. You may also
7138 need to alter the handling of switch statements so that they use
7140 @c i rearranged the order of the macros above to try to force one of
7141 @c them to the next line, to eliminate an overfull hbox. --mew 10feb93
7143 @defmac PIC_OFFSET_TABLE_REGNUM
7144 The register number of the register used to address a table of static
7145 data addresses in memory. In some cases this register is defined by a
7146 processor's ``application binary interface'' (ABI)@. When this macro
7147 is defined, RTL is generated for this register once, as with the stack
7148 pointer and frame pointer registers. If this macro is not defined, it
7149 is up to the machine-dependent files to allocate such a register (if
7150 necessary). Note that this register must be fixed when in use (e.g.@:
7151 when @code{flag_pic} is true).
7154 @defmac PIC_OFFSET_TABLE_REG_CALL_CLOBBERED
7155 A C expression that is nonzero if the register defined by
7156 @code{PIC_OFFSET_TABLE_REGNUM} is clobbered by calls. If not defined,
7157 the default is zero. Do not define
7158 this macro if @code{PIC_OFFSET_TABLE_REGNUM} is not defined.
7161 @defmac LEGITIMATE_PIC_OPERAND_P (@var{x})
7162 A C expression that is nonzero if @var{x} is a legitimate immediate
7163 operand on the target machine when generating position independent code.
7164 You can assume that @var{x} satisfies @code{CONSTANT_P}, so you need not
7165 check this. You can also assume @var{flag_pic} is true, so you need not
7166 check it either. You need not define this macro if all constants
7167 (including @code{SYMBOL_REF}) can be immediate operands when generating
7168 position independent code.
7171 @node Assembler Format
7172 @section Defining the Output Assembler Language
7174 This section describes macros whose principal purpose is to describe how
7175 to write instructions in assembler language---rather than what the
7179 * File Framework:: Structural information for the assembler file.
7180 * Data Output:: Output of constants (numbers, strings, addresses).
7181 * Uninitialized Data:: Output of uninitialized variables.
7182 * Label Output:: Output and generation of labels.
7183 * Initialization:: General principles of initialization
7184 and termination routines.
7185 * Macros for Initialization::
7186 Specific macros that control the handling of
7187 initialization and termination routines.
7188 * Instruction Output:: Output of actual instructions.
7189 * Dispatch Tables:: Output of jump tables.
7190 * Exception Region Output:: Output of exception region code.
7191 * Alignment Output:: Pseudo ops for alignment and skipping data.
7194 @node File Framework
7195 @subsection The Overall Framework of an Assembler File
7196 @cindex assembler format
7197 @cindex output of assembler code
7199 @c prevent bad page break with this line
7200 This describes the overall framework of an assembly file.
7202 @findex default_file_start
7203 @hook TARGET_ASM_FILE_START
7204 Output to @code{asm_out_file} any text which the assembler expects to
7205 find at the beginning of a file. The default behavior is controlled
7206 by two flags, documented below. Unless your target's assembler is
7207 quite unusual, if you override the default, you should call
7208 @code{default_file_start} at some point in your target hook. This
7209 lets other target files rely on these variables.
7212 @hook TARGET_ASM_FILE_START_APP_OFF
7213 If this flag is true, the text of the macro @code{ASM_APP_OFF} will be
7214 printed as the very first line in the assembly file, unless
7215 @option{-fverbose-asm} is in effect. (If that macro has been defined
7216 to the empty string, this variable has no effect.) With the normal
7217 definition of @code{ASM_APP_OFF}, the effect is to notify the GNU
7218 assembler that it need not bother stripping comments or extra
7219 whitespace from its input. This allows it to work a bit faster.
7221 The default is false. You should not set it to true unless you have
7222 verified that your port does not generate any extra whitespace or
7223 comments that will cause GAS to issue errors in NO_APP mode.
7226 @hook TARGET_ASM_FILE_START_FILE_DIRECTIVE
7227 If this flag is true, @code{output_file_directive} will be called
7228 for the primary source file, immediately after printing
7229 @code{ASM_APP_OFF} (if that is enabled). Most ELF assemblers expect
7230 this to be done. The default is false.
7233 @hook TARGET_ASM_FILE_END
7234 Output to @code{asm_out_file} any text which the assembler expects
7235 to find at the end of a file. The default is to output nothing.
7238 @deftypefun void file_end_indicate_exec_stack ()
7239 Some systems use a common convention, the @samp{.note.GNU-stack}
7240 special section, to indicate whether or not an object file relies on
7241 the stack being executable. If your system uses this convention, you
7242 should define @code{TARGET_ASM_FILE_END} to this function. If you
7243 need to do other things in that hook, have your hook function call
7247 @hook TARGET_ASM_LTO_START
7248 Output to @code{asm_out_file} any text which the assembler expects
7249 to find at the start of an LTO section. The default is to output
7253 @hook TARGET_ASM_LTO_END
7254 Output to @code{asm_out_file} any text which the assembler expects
7255 to find at the end of an LTO section. The default is to output
7259 @hook TARGET_ASM_CODE_END
7260 Output to @code{asm_out_file} any text which is needed before emitting
7261 unwind info and debug info at the end of a file. Some targets emit
7262 here PIC setup thunks that cannot be emitted at the end of file,
7263 because they couldn't have unwind info then. The default is to output
7267 @defmac ASM_COMMENT_START
7268 A C string constant describing how to begin a comment in the target
7269 assembler language. The compiler assumes that the comment will end at
7270 the end of the line.
7274 A C string constant for text to be output before each @code{asm}
7275 statement or group of consecutive ones. Normally this is
7276 @code{"#APP"}, which is a comment that has no effect on most
7277 assemblers but tells the GNU assembler that it must check the lines
7278 that follow for all valid assembler constructs.
7282 A C string constant for text to be output after each @code{asm}
7283 statement or group of consecutive ones. Normally this is
7284 @code{"#NO_APP"}, which tells the GNU assembler to resume making the
7285 time-saving assumptions that are valid for ordinary compiler output.
7288 @defmac ASM_OUTPUT_SOURCE_FILENAME (@var{stream}, @var{name})
7289 A C statement to output COFF information or DWARF debugging information
7290 which indicates that filename @var{name} is the current source file to
7291 the stdio stream @var{stream}.
7293 This macro need not be defined if the standard form of output
7294 for the file format in use is appropriate.
7297 @hook TARGET_ASM_OUTPUT_SOURCE_FILENAME
7299 @defmac OUTPUT_QUOTED_STRING (@var{stream}, @var{string})
7300 A C statement to output the string @var{string} to the stdio stream
7301 @var{stream}. If you do not call the function @code{output_quoted_string}
7302 in your config files, GCC will only call it to output filenames to
7303 the assembler source. So you can use it to canonicalize the format
7304 of the filename using this macro.
7307 @defmac ASM_OUTPUT_IDENT (@var{stream}, @var{string})
7308 A C statement to output something to the assembler file to handle a
7309 @samp{#ident} directive containing the text @var{string}. If this
7310 macro is not defined, nothing is output for a @samp{#ident} directive.
7313 @hook TARGET_ASM_NAMED_SECTION
7314 Output assembly directives to switch to section @var{name}. The section
7315 should have attributes as specified by @var{flags}, which is a bit mask
7316 of the @code{SECTION_*} flags defined in @file{output.h}. If @var{decl}
7317 is non-NULL, it is the @code{VAR_DECL} or @code{FUNCTION_DECL} with which
7318 this section is associated.
7321 @hook TARGET_ASM_FUNCTION_SECTION
7322 Return preferred text (sub)section for function @var{decl}.
7323 Main purpose of this function is to separate cold, normal and hot
7324 functions. @var{startup} is true when function is known to be used only
7325 at startup (from static constructors or it is @code{main()}).
7326 @var{exit} is true when function is known to be used only at exit
7327 (from static destructors).
7328 Return NULL if function should go to default text section.
7331 @hook TARGET_ASM_FUNCTION_SWITCHED_TEXT_SECTIONS
7333 @hook TARGET_HAVE_NAMED_SECTIONS
7334 This flag is true if the target supports @code{TARGET_ASM_NAMED_SECTION}.
7335 It must not be modified by command-line option processing.
7338 @anchor{TARGET_HAVE_SWITCHABLE_BSS_SECTIONS}
7339 @hook TARGET_HAVE_SWITCHABLE_BSS_SECTIONS
7340 This flag is true if we can create zeroed data by switching to a BSS
7341 section and then using @code{ASM_OUTPUT_SKIP} to allocate the space.
7342 This is true on most ELF targets.
7345 @hook TARGET_SECTION_TYPE_FLAGS
7346 Choose a set of section attributes for use by @code{TARGET_ASM_NAMED_SECTION}
7347 based on a variable or function decl, a section name, and whether or not the
7348 declaration's initializer may contain runtime relocations. @var{decl} may be
7349 null, in which case read-write data should be assumed.
7351 The default version of this function handles choosing code vs data,
7352 read-only vs read-write data, and @code{flag_pic}. You should only
7353 need to override this if your target has special flags that might be
7354 set via @code{__attribute__}.
7357 @hook TARGET_ASM_RECORD_GCC_SWITCHES
7358 Provides the target with the ability to record the gcc command line
7359 switches that have been passed to the compiler, and options that are
7360 enabled. The @var{type} argument specifies what is being recorded.
7361 It can take the following values:
7364 @item SWITCH_TYPE_PASSED
7365 @var{text} is a command line switch that has been set by the user.
7367 @item SWITCH_TYPE_ENABLED
7368 @var{text} is an option which has been enabled. This might be as a
7369 direct result of a command line switch, or because it is enabled by
7370 default or because it has been enabled as a side effect of a different
7371 command line switch. For example, the @option{-O2} switch enables
7372 various different individual optimization passes.
7374 @item SWITCH_TYPE_DESCRIPTIVE
7375 @var{text} is either NULL or some descriptive text which should be
7376 ignored. If @var{text} is NULL then it is being used to warn the
7377 target hook that either recording is starting or ending. The first
7378 time @var{type} is SWITCH_TYPE_DESCRIPTIVE and @var{text} is NULL, the
7379 warning is for start up and the second time the warning is for
7380 wind down. This feature is to allow the target hook to make any
7381 necessary preparations before it starts to record switches and to
7382 perform any necessary tidying up after it has finished recording
7385 @item SWITCH_TYPE_LINE_START
7386 This option can be ignored by this target hook.
7388 @item SWITCH_TYPE_LINE_END
7389 This option can be ignored by this target hook.
7392 The hook's return value must be zero. Other return values may be
7393 supported in the future.
7395 By default this hook is set to NULL, but an example implementation is
7396 provided for ELF based targets. Called @var{elf_record_gcc_switches},
7397 it records the switches as ASCII text inside a new, string mergeable
7398 section in the assembler output file. The name of the new section is
7399 provided by the @code{TARGET_ASM_RECORD_GCC_SWITCHES_SECTION} target
7403 @hook TARGET_ASM_RECORD_GCC_SWITCHES_SECTION
7404 This is the name of the section that will be created by the example
7405 ELF implementation of the @code{TARGET_ASM_RECORD_GCC_SWITCHES} target
7411 @subsection Output of Data
7414 @hook TARGET_ASM_BYTE_OP
7415 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_HI_OP
7416 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_SI_OP
7417 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_DI_OP
7418 @deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_TI_OP
7419 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_HI_OP
7420 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_SI_OP
7421 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_DI_OP
7422 @deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_TI_OP
7423 These hooks specify assembly directives for creating certain kinds
7424 of integer object. The @code{TARGET_ASM_BYTE_OP} directive creates a
7425 byte-sized object, the @code{TARGET_ASM_ALIGNED_HI_OP} one creates an
7426 aligned two-byte object, and so on. Any of the hooks may be
7427 @code{NULL}, indicating that no suitable directive is available.
7429 The compiler will print these strings at the start of a new line,
7430 followed immediately by the object's initial value. In most cases,
7431 the string should contain a tab, a pseudo-op, and then another tab.
7434 @hook TARGET_ASM_INTEGER
7435 The @code{assemble_integer} function uses this hook to output an
7436 integer object. @var{x} is the object's value, @var{size} is its size
7437 in bytes and @var{aligned_p} indicates whether it is aligned. The
7438 function should return @code{true} if it was able to output the
7439 object. If it returns false, @code{assemble_integer} will try to
7440 split the object into smaller parts.
7442 The default implementation of this hook will use the
7443 @code{TARGET_ASM_BYTE_OP} family of strings, returning @code{false}
7444 when the relevant string is @code{NULL}.
7447 @hook TARGET_ASM_OUTPUT_ADDR_CONST_EXTRA
7448 A target hook to recognize @var{rtx} patterns that @code{output_addr_const}
7449 can't deal with, and output assembly code to @var{file} corresponding to
7450 the pattern @var{x}. This may be used to allow machine-dependent
7451 @code{UNSPEC}s to appear within constants.
7453 If target hook fails to recognize a pattern, it must return @code{false},
7454 so that a standard error message is printed. If it prints an error message
7455 itself, by calling, for example, @code{output_operand_lossage}, it may just
7459 @defmac ASM_OUTPUT_ASCII (@var{stream}, @var{ptr}, @var{len})
7460 A C statement to output to the stdio stream @var{stream} an assembler
7461 instruction to assemble a string constant containing the @var{len}
7462 bytes at @var{ptr}. @var{ptr} will be a C expression of type
7463 @code{char *} and @var{len} a C expression of type @code{int}.
7465 If the assembler has a @code{.ascii} pseudo-op as found in the
7466 Berkeley Unix assembler, do not define the macro
7467 @code{ASM_OUTPUT_ASCII}.
7470 @defmac ASM_OUTPUT_FDESC (@var{stream}, @var{decl}, @var{n})
7471 A C statement to output word @var{n} of a function descriptor for
7472 @var{decl}. This must be defined if @code{TARGET_VTABLE_USES_DESCRIPTORS}
7473 is defined, and is otherwise unused.
7476 @defmac CONSTANT_POOL_BEFORE_FUNCTION
7477 You may define this macro as a C expression. You should define the
7478 expression to have a nonzero value if GCC should output the constant
7479 pool for a function before the code for the function, or a zero value if
7480 GCC should output the constant pool after the function. If you do
7481 not define this macro, the usual case, GCC will output the constant
7482 pool before the function.
7485 @defmac ASM_OUTPUT_POOL_PROLOGUE (@var{file}, @var{funname}, @var{fundecl}, @var{size})
7486 A C statement to output assembler commands to define the start of the
7487 constant pool for a function. @var{funname} is a string giving
7488 the name of the function. Should the return type of the function
7489 be required, it can be obtained via @var{fundecl}. @var{size}
7490 is the size, in bytes, of the constant pool that will be written
7491 immediately after this call.
7493 If no constant-pool prefix is required, the usual case, this macro need
7497 @defmac ASM_OUTPUT_SPECIAL_POOL_ENTRY (@var{file}, @var{x}, @var{mode}, @var{align}, @var{labelno}, @var{jumpto})
7498 A C statement (with or without semicolon) to output a constant in the
7499 constant pool, if it needs special treatment. (This macro need not do
7500 anything for RTL expressions that can be output normally.)
7502 The argument @var{file} is the standard I/O stream to output the
7503 assembler code on. @var{x} is the RTL expression for the constant to
7504 output, and @var{mode} is the machine mode (in case @var{x} is a
7505 @samp{const_int}). @var{align} is the required alignment for the value
7506 @var{x}; you should output an assembler directive to force this much
7509 The argument @var{labelno} is a number to use in an internal label for
7510 the address of this pool entry. The definition of this macro is
7511 responsible for outputting the label definition at the proper place.
7512 Here is how to do this:
7515 @code{(*targetm.asm_out.internal_label)} (@var{file}, "LC", @var{labelno});
7518 When you output a pool entry specially, you should end with a
7519 @code{goto} to the label @var{jumpto}. This will prevent the same pool
7520 entry from being output a second time in the usual manner.
7522 You need not define this macro if it would do nothing.
7525 @defmac ASM_OUTPUT_POOL_EPILOGUE (@var{file} @var{funname} @var{fundecl} @var{size})
7526 A C statement to output assembler commands to at the end of the constant
7527 pool for a function. @var{funname} is a string giving the name of the
7528 function. Should the return type of the function be required, you can
7529 obtain it via @var{fundecl}. @var{size} is the size, in bytes, of the
7530 constant pool that GCC wrote immediately before this call.
7532 If no constant-pool epilogue is required, the usual case, you need not
7536 @defmac IS_ASM_LOGICAL_LINE_SEPARATOR (@var{C}, @var{STR})
7537 Define this macro as a C expression which is nonzero if @var{C} is
7538 used as a logical line separator by the assembler. @var{STR} points
7539 to the position in the string where @var{C} was found; this can be used if
7540 a line separator uses multiple characters.
7542 If you do not define this macro, the default is that only
7543 the character @samp{;} is treated as a logical line separator.
7546 @hook TARGET_ASM_OPEN_PAREN
7547 These target hooks are C string constants, describing the syntax in the
7548 assembler for grouping arithmetic expressions. If not overridden, they
7549 default to normal parentheses, which is correct for most assemblers.
7552 These macros are provided by @file{real.h} for writing the definitions
7553 of @code{ASM_OUTPUT_DOUBLE} and the like:
7555 @defmac REAL_VALUE_TO_TARGET_SINGLE (@var{x}, @var{l})
7556 @defmacx REAL_VALUE_TO_TARGET_DOUBLE (@var{x}, @var{l})
7557 @defmacx REAL_VALUE_TO_TARGET_LONG_DOUBLE (@var{x}, @var{l})
7558 @defmacx REAL_VALUE_TO_TARGET_DECIMAL32 (@var{x}, @var{l})
7559 @defmacx REAL_VALUE_TO_TARGET_DECIMAL64 (@var{x}, @var{l})
7560 @defmacx REAL_VALUE_TO_TARGET_DECIMAL128 (@var{x}, @var{l})
7561 These translate @var{x}, of type @code{REAL_VALUE_TYPE}, to the
7562 target's floating point representation, and store its bit pattern in
7563 the variable @var{l}. For @code{REAL_VALUE_TO_TARGET_SINGLE} and
7564 @code{REAL_VALUE_TO_TARGET_DECIMAL32}, this variable should be a
7565 simple @code{long int}. For the others, it should be an array of
7566 @code{long int}. The number of elements in this array is determined
7567 by the size of the desired target floating point data type: 32 bits of
7568 it go in each @code{long int} array element. Each array element holds
7569 32 bits of the result, even if @code{long int} is wider than 32 bits
7570 on the host machine.
7572 The array element values are designed so that you can print them out
7573 using @code{fprintf} in the order they should appear in the target
7577 @node Uninitialized Data
7578 @subsection Output of Uninitialized Variables
7580 Each of the macros in this section is used to do the whole job of
7581 outputting a single uninitialized variable.
7583 @defmac ASM_OUTPUT_COMMON (@var{stream}, @var{name}, @var{size}, @var{rounded})
7584 A C statement (sans semicolon) to output to the stdio stream
7585 @var{stream} the assembler definition of a common-label named
7586 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
7587 is the size rounded up to whatever alignment the caller wants. It is
7588 possible that @var{size} may be zero, for instance if a struct with no
7589 other member than a zero-length array is defined. In this case, the
7590 backend must output a symbol definition that allocates at least one
7591 byte, both so that the address of the resulting object does not compare
7592 equal to any other, and because some object formats cannot even express
7593 the concept of a zero-sized common symbol, as that is how they represent
7594 an ordinary undefined external.
7596 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
7597 output the name itself; before and after that, output the additional
7598 assembler syntax for defining the name, and a newline.
7600 This macro controls how the assembler definitions of uninitialized
7601 common global variables are output.
7604 @defmac ASM_OUTPUT_ALIGNED_COMMON (@var{stream}, @var{name}, @var{size}, @var{alignment})
7605 Like @code{ASM_OUTPUT_COMMON} except takes the required alignment as a
7606 separate, explicit argument. If you define this macro, it is used in
7607 place of @code{ASM_OUTPUT_COMMON}, and gives you more flexibility in
7608 handling the required alignment of the variable. The alignment is specified
7609 as the number of bits.
7612 @defmac ASM_OUTPUT_ALIGNED_DECL_COMMON (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
7613 Like @code{ASM_OUTPUT_ALIGNED_COMMON} except that @var{decl} of the
7614 variable to be output, if there is one, or @code{NULL_TREE} if there
7615 is no corresponding variable. If you define this macro, GCC will use it
7616 in place of both @code{ASM_OUTPUT_COMMON} and
7617 @code{ASM_OUTPUT_ALIGNED_COMMON}. Define this macro when you need to see
7618 the variable's decl in order to chose what to output.
7621 @defmac ASM_OUTPUT_ALIGNED_BSS (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
7622 A C statement (sans semicolon) to output to the stdio stream
7623 @var{stream} the assembler definition of uninitialized global @var{decl} named
7624 @var{name} whose size is @var{size} bytes. The variable @var{alignment}
7625 is the alignment specified as the number of bits.
7627 Try to use function @code{asm_output_aligned_bss} defined in file
7628 @file{varasm.c} when defining this macro. If unable, use the expression
7629 @code{assemble_name (@var{stream}, @var{name})} to output the name itself;
7630 before and after that, output the additional assembler syntax for defining
7631 the name, and a newline.
7633 There are two ways of handling global BSS@. One is to define this macro.
7634 The other is to have @code{TARGET_ASM_SELECT_SECTION} return a
7635 switchable BSS section (@pxref{TARGET_HAVE_SWITCHABLE_BSS_SECTIONS}).
7636 You do not need to do both.
7638 Some languages do not have @code{common} data, and require a
7639 non-common form of global BSS in order to handle uninitialized globals
7640 efficiently. C++ is one example of this. However, if the target does
7641 not support global BSS, the front end may choose to make globals
7642 common in order to save space in the object file.
7645 @defmac ASM_OUTPUT_LOCAL (@var{stream}, @var{name}, @var{size}, @var{rounded})
7646 A C statement (sans semicolon) to output to the stdio stream
7647 @var{stream} the assembler definition of a local-common-label named
7648 @var{name} whose size is @var{size} bytes. The variable @var{rounded}
7649 is the size rounded up to whatever alignment the caller wants.
7651 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
7652 output the name itself; before and after that, output the additional
7653 assembler syntax for defining the name, and a newline.
7655 This macro controls how the assembler definitions of uninitialized
7656 static variables are output.
7659 @defmac ASM_OUTPUT_ALIGNED_LOCAL (@var{stream}, @var{name}, @var{size}, @var{alignment})
7660 Like @code{ASM_OUTPUT_LOCAL} except takes the required alignment as a
7661 separate, explicit argument. If you define this macro, it is used in
7662 place of @code{ASM_OUTPUT_LOCAL}, and gives you more flexibility in
7663 handling the required alignment of the variable. The alignment is specified
7664 as the number of bits.
7667 @defmac ASM_OUTPUT_ALIGNED_DECL_LOCAL (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
7668 Like @code{ASM_OUTPUT_ALIGNED_DECL} except that @var{decl} of the
7669 variable to be output, if there is one, or @code{NULL_TREE} if there
7670 is no corresponding variable. If you define this macro, GCC will use it
7671 in place of both @code{ASM_OUTPUT_DECL} and
7672 @code{ASM_OUTPUT_ALIGNED_DECL}. Define this macro when you need to see
7673 the variable's decl in order to chose what to output.
7677 @subsection Output and Generation of Labels
7679 @c prevent bad page break with this line
7680 This is about outputting labels.
7682 @findex assemble_name
7683 @defmac ASM_OUTPUT_LABEL (@var{stream}, @var{name})
7684 A C statement (sans semicolon) to output to the stdio stream
7685 @var{stream} the assembler definition of a label named @var{name}.
7686 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
7687 output the name itself; before and after that, output the additional
7688 assembler syntax for defining the name, and a newline. A default
7689 definition of this macro is provided which is correct for most systems.
7692 @defmac ASM_OUTPUT_FUNCTION_LABEL (@var{stream}, @var{name}, @var{decl})
7693 A C statement (sans semicolon) to output to the stdio stream
7694 @var{stream} the assembler definition of a label named @var{name} of
7696 Use the expression @code{assemble_name (@var{stream}, @var{name})} to
7697 output the name itself; before and after that, output the additional
7698 assembler syntax for defining the name, and a newline. A default
7699 definition of this macro is provided which is correct for most systems.
7701 If this macro is not defined, then the function name is defined in the
7702 usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
7705 @findex assemble_name_raw
7706 @defmac ASM_OUTPUT_INTERNAL_LABEL (@var{stream}, @var{name})
7707 Identical to @code{ASM_OUTPUT_LABEL}, except that @var{name} is known
7708 to refer to a compiler-generated label. The default definition uses
7709 @code{assemble_name_raw}, which is like @code{assemble_name} except
7710 that it is more efficient.
7714 A C string containing the appropriate assembler directive to specify the
7715 size of a symbol, without any arguments. On systems that use ELF, the
7716 default (in @file{config/elfos.h}) is @samp{"\t.size\t"}; on other
7717 systems, the default is not to define this macro.
7719 Define this macro only if it is correct to use the default definitions
7720 of @code{ASM_OUTPUT_SIZE_DIRECTIVE} and @code{ASM_OUTPUT_MEASURED_SIZE}
7721 for your system. If you need your own custom definitions of those
7722 macros, or if you do not need explicit symbol sizes at all, do not
7726 @defmac ASM_OUTPUT_SIZE_DIRECTIVE (@var{stream}, @var{name}, @var{size})
7727 A C statement (sans semicolon) to output to the stdio stream
7728 @var{stream} a directive telling the assembler that the size of the
7729 symbol @var{name} is @var{size}. @var{size} is a @code{HOST_WIDE_INT}.
7730 If you define @code{SIZE_ASM_OP}, a default definition of this macro is
7734 @defmac ASM_OUTPUT_MEASURED_SIZE (@var{stream}, @var{name})
7735 A C statement (sans semicolon) to output to the stdio stream
7736 @var{stream} a directive telling the assembler to calculate the size of
7737 the symbol @var{name} by subtracting its address from the current
7740 If you define @code{SIZE_ASM_OP}, a default definition of this macro is
7741 provided. The default assumes that the assembler recognizes a special
7742 @samp{.} symbol as referring to the current address, and can calculate
7743 the difference between this and another symbol. If your assembler does
7744 not recognize @samp{.} or cannot do calculations with it, you will need
7745 to redefine @code{ASM_OUTPUT_MEASURED_SIZE} to use some other technique.
7749 A C string containing the appropriate assembler directive to specify the
7750 type of a symbol, without any arguments. On systems that use ELF, the
7751 default (in @file{config/elfos.h}) is @samp{"\t.type\t"}; on other
7752 systems, the default is not to define this macro.
7754 Define this macro only if it is correct to use the default definition of
7755 @code{ASM_OUTPUT_TYPE_DIRECTIVE} for your system. If you need your own
7756 custom definition of this macro, or if you do not need explicit symbol
7757 types at all, do not define this macro.
7760 @defmac TYPE_OPERAND_FMT
7761 A C string which specifies (using @code{printf} syntax) the format of
7762 the second operand to @code{TYPE_ASM_OP}. On systems that use ELF, the
7763 default (in @file{config/elfos.h}) is @samp{"@@%s"}; on other systems,
7764 the default is not to define this macro.
7766 Define this macro only if it is correct to use the default definition of
7767 @code{ASM_OUTPUT_TYPE_DIRECTIVE} for your system. If you need your own
7768 custom definition of this macro, or if you do not need explicit symbol
7769 types at all, do not define this macro.
7772 @defmac ASM_OUTPUT_TYPE_DIRECTIVE (@var{stream}, @var{type})
7773 A C statement (sans semicolon) to output to the stdio stream
7774 @var{stream} a directive telling the assembler that the type of the
7775 symbol @var{name} is @var{type}. @var{type} is a C string; currently,
7776 that string is always either @samp{"function"} or @samp{"object"}, but
7777 you should not count on this.
7779 If you define @code{TYPE_ASM_OP} and @code{TYPE_OPERAND_FMT}, a default
7780 definition of this macro is provided.
7783 @defmac ASM_DECLARE_FUNCTION_NAME (@var{stream}, @var{name}, @var{decl})
7784 A C statement (sans semicolon) to output to the stdio stream
7785 @var{stream} any text necessary for declaring the name @var{name} of a
7786 function which is being defined. This macro is responsible for
7787 outputting the label definition (perhaps using
7788 @code{ASM_OUTPUT_FUNCTION_LABEL}). The argument @var{decl} is the
7789 @code{FUNCTION_DECL} tree node representing the function.
7791 If this macro is not defined, then the function name is defined in the
7792 usual manner as a label (by means of @code{ASM_OUTPUT_FUNCTION_LABEL}).
7794 You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} in the definition
7798 @defmac ASM_DECLARE_FUNCTION_SIZE (@var{stream}, @var{name}, @var{decl})
7799 A C statement (sans semicolon) to output to the stdio stream
7800 @var{stream} any text necessary for declaring the size of a function
7801 which is being defined. The argument @var{name} is the name of the
7802 function. The argument @var{decl} is the @code{FUNCTION_DECL} tree node
7803 representing the function.
7805 If this macro is not defined, then the function size is not defined.
7807 You may wish to use @code{ASM_OUTPUT_MEASURED_SIZE} in the definition
7811 @defmac ASM_DECLARE_OBJECT_NAME (@var{stream}, @var{name}, @var{decl})
7812 A C statement (sans semicolon) to output to the stdio stream
7813 @var{stream} any text necessary for declaring the name @var{name} of an
7814 initialized variable which is being defined. This macro must output the
7815 label definition (perhaps using @code{ASM_OUTPUT_LABEL}). The argument
7816 @var{decl} is the @code{VAR_DECL} tree node representing the variable.
7818 If this macro is not defined, then the variable name is defined in the
7819 usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
7821 You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} and/or
7822 @code{ASM_OUTPUT_SIZE_DIRECTIVE} in the definition of this macro.
7825 @hook TARGET_ASM_DECLARE_CONSTANT_NAME
7826 A target hook to output to the stdio stream @var{file} any text necessary
7827 for declaring the name @var{name} of a constant which is being defined. This
7828 target hook is responsible for outputting the label definition (perhaps using
7829 @code{assemble_label}). The argument @var{exp} is the value of the constant,
7830 and @var{size} is the size of the constant in bytes. The @var{name}
7831 will be an internal label.
7833 The default version of this target hook, define the @var{name} in the
7834 usual manner as a label (by means of @code{assemble_label}).
7836 You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} in this target hook.
7839 @defmac ASM_DECLARE_REGISTER_GLOBAL (@var{stream}, @var{decl}, @var{regno}, @var{name})
7840 A C statement (sans semicolon) to output to the stdio stream
7841 @var{stream} any text necessary for claiming a register @var{regno}
7842 for a global variable @var{decl} with name @var{name}.
7844 If you don't define this macro, that is equivalent to defining it to do
7848 @defmac ASM_FINISH_DECLARE_OBJECT (@var{stream}, @var{decl}, @var{toplevel}, @var{atend})
7849 A C statement (sans semicolon) to finish up declaring a variable name
7850 once the compiler has processed its initializer fully and thus has had a
7851 chance to determine the size of an array when controlled by an
7852 initializer. This is used on systems where it's necessary to declare
7853 something about the size of the object.
7855 If you don't define this macro, that is equivalent to defining it to do
7858 You may wish to use @code{ASM_OUTPUT_SIZE_DIRECTIVE} and/or
7859 @code{ASM_OUTPUT_MEASURED_SIZE} in the definition of this macro.
7862 @hook TARGET_ASM_GLOBALIZE_LABEL
7863 This target hook is a function to output to the stdio stream
7864 @var{stream} some commands that will make the label @var{name} global;
7865 that is, available for reference from other files.
7867 The default implementation relies on a proper definition of
7868 @code{GLOBAL_ASM_OP}.
7871 @hook TARGET_ASM_GLOBALIZE_DECL_NAME
7872 This target hook is a function to output to the stdio stream
7873 @var{stream} some commands that will make the name associated with @var{decl}
7874 global; that is, available for reference from other files.
7876 The default implementation uses the TARGET_ASM_GLOBALIZE_LABEL target hook.
7879 @defmac ASM_WEAKEN_LABEL (@var{stream}, @var{name})
7880 A C statement (sans semicolon) to output to the stdio stream
7881 @var{stream} some commands that will make the label @var{name} weak;
7882 that is, available for reference from other files but only used if
7883 no other definition is available. Use the expression
7884 @code{assemble_name (@var{stream}, @var{name})} to output the name
7885 itself; before and after that, output the additional assembler syntax
7886 for making that name weak, and a newline.
7888 If you don't define this macro or @code{ASM_WEAKEN_DECL}, GCC will not
7889 support weak symbols and you should not define the @code{SUPPORTS_WEAK}
7893 @defmac ASM_WEAKEN_DECL (@var{stream}, @var{decl}, @var{name}, @var{value})
7894 Combines (and replaces) the function of @code{ASM_WEAKEN_LABEL} and
7895 @code{ASM_OUTPUT_WEAK_ALIAS}, allowing access to the associated function
7896 or variable decl. If @var{value} is not @code{NULL}, this C statement
7897 should output to the stdio stream @var{stream} assembler code which
7898 defines (equates) the weak symbol @var{name} to have the value
7899 @var{value}. If @var{value} is @code{NULL}, it should output commands
7900 to make @var{name} weak.
7903 @defmac ASM_OUTPUT_WEAKREF (@var{stream}, @var{decl}, @var{name}, @var{value})
7904 Outputs a directive that enables @var{name} to be used to refer to
7905 symbol @var{value} with weak-symbol semantics. @code{decl} is the
7906 declaration of @code{name}.
7909 @defmac SUPPORTS_WEAK
7910 A preprocessor constant expression which evaluates to true if the target
7911 supports weak symbols.
7913 If you don't define this macro, @file{defaults.h} provides a default
7914 definition. If either @code{ASM_WEAKEN_LABEL} or @code{ASM_WEAKEN_DECL}
7915 is defined, the default definition is @samp{1}; otherwise, it is @samp{0}.
7918 @defmac TARGET_SUPPORTS_WEAK
7919 A C expression which evaluates to true if the target supports weak symbols.
7921 If you don't define this macro, @file{defaults.h} provides a default
7922 definition. The default definition is @samp{(SUPPORTS_WEAK)}. Define
7923 this macro if you want to control weak symbol support with a compiler
7924 flag such as @option{-melf}.
7927 @defmac MAKE_DECL_ONE_ONLY (@var{decl})
7928 A C statement (sans semicolon) to mark @var{decl} to be emitted as a
7929 public symbol such that extra copies in multiple translation units will
7930 be discarded by the linker. Define this macro if your object file
7931 format provides support for this concept, such as the @samp{COMDAT}
7932 section flags in the Microsoft Windows PE/COFF format, and this support
7933 requires changes to @var{decl}, such as putting it in a separate section.
7936 @defmac SUPPORTS_ONE_ONLY
7937 A C expression which evaluates to true if the target supports one-only
7940 If you don't define this macro, @file{varasm.c} provides a default
7941 definition. If @code{MAKE_DECL_ONE_ONLY} is defined, the default
7942 definition is @samp{1}; otherwise, it is @samp{0}. Define this macro if
7943 you want to control one-only symbol support with a compiler flag, or if
7944 setting the @code{DECL_ONE_ONLY} flag is enough to mark a declaration to
7945 be emitted as one-only.
7948 @hook TARGET_ASM_ASSEMBLE_VISIBILITY
7949 This target hook is a function to output to @var{asm_out_file} some
7950 commands that will make the symbol(s) associated with @var{decl} have
7951 hidden, protected or internal visibility as specified by @var{visibility}.
7954 @defmac TARGET_WEAK_NOT_IN_ARCHIVE_TOC
7955 A C expression that evaluates to true if the target's linker expects
7956 that weak symbols do not appear in a static archive's table of contents.
7957 The default is @code{0}.
7959 Leaving weak symbols out of an archive's table of contents means that,
7960 if a symbol will only have a definition in one translation unit and
7961 will have undefined references from other translation units, that
7962 symbol should not be weak. Defining this macro to be nonzero will
7963 thus have the effect that certain symbols that would normally be weak
7964 (explicit template instantiations, and vtables for polymorphic classes
7965 with noninline key methods) will instead be nonweak.
7967 The C++ ABI requires this macro to be zero. Define this macro for
7968 targets where full C++ ABI compliance is impossible and where linker
7969 restrictions require weak symbols to be left out of a static archive's
7973 @defmac ASM_OUTPUT_EXTERNAL (@var{stream}, @var{decl}, @var{name})
7974 A C statement (sans semicolon) to output to the stdio stream
7975 @var{stream} any text necessary for declaring the name of an external
7976 symbol named @var{name} which is referenced in this compilation but
7977 not defined. The value of @var{decl} is the tree node for the
7980 This macro need not be defined if it does not need to output anything.
7981 The GNU assembler and most Unix assemblers don't require anything.
7984 @hook TARGET_ASM_EXTERNAL_LIBCALL
7985 This target hook is a function to output to @var{asm_out_file} an assembler
7986 pseudo-op to declare a library function name external. The name of the
7987 library function is given by @var{symref}, which is a @code{symbol_ref}.
7990 @hook TARGET_ASM_MARK_DECL_PRESERVED
7991 This target hook is a function to output to @var{asm_out_file} an assembler
7992 directive to annotate @var{symbol} as used. The Darwin target uses the
7993 .no_dead_code_strip directive.
7996 @defmac ASM_OUTPUT_LABELREF (@var{stream}, @var{name})
7997 A C statement (sans semicolon) to output to the stdio stream
7998 @var{stream} a reference in assembler syntax to a label named
7999 @var{name}. This should add @samp{_} to the front of the name, if that
8000 is customary on your operating system, as it is in most Berkeley Unix
8001 systems. This macro is used in @code{assemble_name}.
8004 @hook TARGET_MANGLE_ASSEMBLER_NAME
8006 @defmac ASM_OUTPUT_SYMBOL_REF (@var{stream}, @var{sym})
8007 A C statement (sans semicolon) to output a reference to
8008 @code{SYMBOL_REF} @var{sym}. If not defined, @code{assemble_name}
8009 will be used to output the name of the symbol. This macro may be used
8010 to modify the way a symbol is referenced depending on information
8011 encoded by @code{TARGET_ENCODE_SECTION_INFO}.
8014 @defmac ASM_OUTPUT_LABEL_REF (@var{stream}, @var{buf})
8015 A C statement (sans semicolon) to output a reference to @var{buf}, the
8016 result of @code{ASM_GENERATE_INTERNAL_LABEL}. If not defined,
8017 @code{assemble_name} will be used to output the name of the symbol.
8018 This macro is not used by @code{output_asm_label}, or the @code{%l}
8019 specifier that calls it; the intention is that this macro should be set
8020 when it is necessary to output a label differently when its address is
8024 @hook TARGET_ASM_INTERNAL_LABEL
8025 A function to output to the stdio stream @var{stream} a label whose
8026 name is made from the string @var{prefix} and the number @var{labelno}.
8028 It is absolutely essential that these labels be distinct from the labels
8029 used for user-level functions and variables. Otherwise, certain programs
8030 will have name conflicts with internal labels.
8032 It is desirable to exclude internal labels from the symbol table of the
8033 object file. Most assemblers have a naming convention for labels that
8034 should be excluded; on many systems, the letter @samp{L} at the
8035 beginning of a label has this effect. You should find out what
8036 convention your system uses, and follow it.
8038 The default version of this function utilizes @code{ASM_GENERATE_INTERNAL_LABEL}.
8041 @defmac ASM_OUTPUT_DEBUG_LABEL (@var{stream}, @var{prefix}, @var{num})
8042 A C statement to output to the stdio stream @var{stream} a debug info
8043 label whose name is made from the string @var{prefix} and the number
8044 @var{num}. This is useful for VLIW targets, where debug info labels
8045 may need to be treated differently than branch target labels. On some
8046 systems, branch target labels must be at the beginning of instruction
8047 bundles, but debug info labels can occur in the middle of instruction
8050 If this macro is not defined, then @code{(*targetm.asm_out.internal_label)} will be
8054 @defmac ASM_GENERATE_INTERNAL_LABEL (@var{string}, @var{prefix}, @var{num})
8055 A C statement to store into the string @var{string} a label whose name
8056 is made from the string @var{prefix} and the number @var{num}.
8058 This string, when output subsequently by @code{assemble_name}, should
8059 produce the output that @code{(*targetm.asm_out.internal_label)} would produce
8060 with the same @var{prefix} and @var{num}.
8062 If the string begins with @samp{*}, then @code{assemble_name} will
8063 output the rest of the string unchanged. It is often convenient for
8064 @code{ASM_GENERATE_INTERNAL_LABEL} to use @samp{*} in this way. If the
8065 string doesn't start with @samp{*}, then @code{ASM_OUTPUT_LABELREF} gets
8066 to output the string, and may change it. (Of course,
8067 @code{ASM_OUTPUT_LABELREF} is also part of your machine description, so
8068 you should know what it does on your machine.)
8071 @defmac ASM_FORMAT_PRIVATE_NAME (@var{outvar}, @var{name}, @var{number})
8072 A C expression to assign to @var{outvar} (which is a variable of type
8073 @code{char *}) a newly allocated string made from the string
8074 @var{name} and the number @var{number}, with some suitable punctuation
8075 added. Use @code{alloca} to get space for the string.
8077 The string will be used as an argument to @code{ASM_OUTPUT_LABELREF} to
8078 produce an assembler label for an internal static variable whose name is
8079 @var{name}. Therefore, the string must be such as to result in valid
8080 assembler code. The argument @var{number} is different each time this
8081 macro is executed; it prevents conflicts between similarly-named
8082 internal static variables in different scopes.
8084 Ideally this string should not be a valid C identifier, to prevent any
8085 conflict with the user's own symbols. Most assemblers allow periods
8086 or percent signs in assembler symbols; putting at least one of these
8087 between the name and the number will suffice.
8089 If this macro is not defined, a default definition will be provided
8090 which is correct for most systems.
8093 @defmac ASM_OUTPUT_DEF (@var{stream}, @var{name}, @var{value})
8094 A C statement to output to the stdio stream @var{stream} assembler code
8095 which defines (equates) the symbol @var{name} to have the value @var{value}.
8098 If @code{SET_ASM_OP} is defined, a default definition is provided which is
8099 correct for most systems.
8102 @defmac ASM_OUTPUT_DEF_FROM_DECLS (@var{stream}, @var{decl_of_name}, @var{decl_of_value})
8103 A C statement to output to the stdio stream @var{stream} assembler code
8104 which defines (equates) the symbol whose tree node is @var{decl_of_name}
8105 to have the value of the tree node @var{decl_of_value}. This macro will
8106 be used in preference to @samp{ASM_OUTPUT_DEF} if it is defined and if
8107 the tree nodes are available.
8110 If @code{SET_ASM_OP} is defined, a default definition is provided which is
8111 correct for most systems.
8114 @defmac TARGET_DEFERRED_OUTPUT_DEFS (@var{decl_of_name}, @var{decl_of_value})
8115 A C statement that evaluates to true if the assembler code which defines
8116 (equates) the symbol whose tree node is @var{decl_of_name} to have the value
8117 of the tree node @var{decl_of_value} should be emitted near the end of the
8118 current compilation unit. The default is to not defer output of defines.
8119 This macro affects defines output by @samp{ASM_OUTPUT_DEF} and
8120 @samp{ASM_OUTPUT_DEF_FROM_DECLS}.
8123 @defmac ASM_OUTPUT_WEAK_ALIAS (@var{stream}, @var{name}, @var{value})
8124 A C statement to output to the stdio stream @var{stream} assembler code
8125 which defines (equates) the weak symbol @var{name} to have the value
8126 @var{value}. If @var{value} is @code{NULL}, it defines @var{name} as
8127 an undefined weak symbol.
8129 Define this macro if the target only supports weak aliases; define
8130 @code{ASM_OUTPUT_DEF} instead if possible.
8133 @defmac OBJC_GEN_METHOD_LABEL (@var{buf}, @var{is_inst}, @var{class_name}, @var{cat_name}, @var{sel_name})
8134 Define this macro to override the default assembler names used for
8135 Objective-C methods.
8137 The default name is a unique method number followed by the name of the
8138 class (e.g.@: @samp{_1_Foo}). For methods in categories, the name of
8139 the category is also included in the assembler name (e.g.@:
8142 These names are safe on most systems, but make debugging difficult since
8143 the method's selector is not present in the name. Therefore, particular
8144 systems define other ways of computing names.
8146 @var{buf} is an expression of type @code{char *} which gives you a
8147 buffer in which to store the name; its length is as long as
8148 @var{class_name}, @var{cat_name} and @var{sel_name} put together, plus
8149 50 characters extra.
8151 The argument @var{is_inst} specifies whether the method is an instance
8152 method or a class method; @var{class_name} is the name of the class;
8153 @var{cat_name} is the name of the category (or @code{NULL} if the method is not
8154 in a category); and @var{sel_name} is the name of the selector.
8156 On systems where the assembler can handle quoted names, you can use this
8157 macro to provide more human-readable names.
8160 @defmac ASM_DECLARE_CLASS_REFERENCE (@var{stream}, @var{name})
8161 A C statement (sans semicolon) to output to the stdio stream
8162 @var{stream} commands to declare that the label @var{name} is an
8163 Objective-C class reference. This is only needed for targets whose
8164 linkers have special support for NeXT-style runtimes.
8167 @defmac ASM_DECLARE_UNRESOLVED_REFERENCE (@var{stream}, @var{name})
8168 A C statement (sans semicolon) to output to the stdio stream
8169 @var{stream} commands to declare that the label @var{name} is an
8170 unresolved Objective-C class reference. This is only needed for targets
8171 whose linkers have special support for NeXT-style runtimes.
8174 @node Initialization
8175 @subsection How Initialization Functions Are Handled
8176 @cindex initialization routines
8177 @cindex termination routines
8178 @cindex constructors, output of
8179 @cindex destructors, output of
8181 The compiled code for certain languages includes @dfn{constructors}
8182 (also called @dfn{initialization routines})---functions to initialize
8183 data in the program when the program is started. These functions need
8184 to be called before the program is ``started''---that is to say, before
8185 @code{main} is called.
8187 Compiling some languages generates @dfn{destructors} (also called
8188 @dfn{termination routines}) that should be called when the program
8191 To make the initialization and termination functions work, the compiler
8192 must output something in the assembler code to cause those functions to
8193 be called at the appropriate time. When you port the compiler to a new
8194 system, you need to specify how to do this.
8196 There are two major ways that GCC currently supports the execution of
8197 initialization and termination functions. Each way has two variants.
8198 Much of the structure is common to all four variations.
8200 @findex __CTOR_LIST__
8201 @findex __DTOR_LIST__
8202 The linker must build two lists of these functions---a list of
8203 initialization functions, called @code{__CTOR_LIST__}, and a list of
8204 termination functions, called @code{__DTOR_LIST__}.
8206 Each list always begins with an ignored function pointer (which may hold
8207 0, @minus{}1, or a count of the function pointers after it, depending on
8208 the environment). This is followed by a series of zero or more function
8209 pointers to constructors (or destructors), followed by a function
8210 pointer containing zero.
8212 Depending on the operating system and its executable file format, either
8213 @file{crtstuff.c} or @file{libgcc2.c} traverses these lists at startup
8214 time and exit time. Constructors are called in reverse order of the
8215 list; destructors in forward order.
8217 The best way to handle static constructors works only for object file
8218 formats which provide arbitrarily-named sections. A section is set
8219 aside for a list of constructors, and another for a list of destructors.
8220 Traditionally these are called @samp{.ctors} and @samp{.dtors}. Each
8221 object file that defines an initialization function also puts a word in
8222 the constructor section to point to that function. The linker
8223 accumulates all these words into one contiguous @samp{.ctors} section.
8224 Termination functions are handled similarly.
8226 This method will be chosen as the default by @file{target-def.h} if
8227 @code{TARGET_ASM_NAMED_SECTION} is defined. A target that does not
8228 support arbitrary sections, but does support special designated
8229 constructor and destructor sections may define @code{CTORS_SECTION_ASM_OP}
8230 and @code{DTORS_SECTION_ASM_OP} to achieve the same effect.
8232 When arbitrary sections are available, there are two variants, depending
8233 upon how the code in @file{crtstuff.c} is called. On systems that
8234 support a @dfn{.init} section which is executed at program startup,
8235 parts of @file{crtstuff.c} are compiled into that section. The
8236 program is linked by the @command{gcc} driver like this:
8239 ld -o @var{output_file} crti.o crtbegin.o @dots{} -lgcc crtend.o crtn.o
8242 The prologue of a function (@code{__init}) appears in the @code{.init}
8243 section of @file{crti.o}; the epilogue appears in @file{crtn.o}. Likewise
8244 for the function @code{__fini} in the @dfn{.fini} section. Normally these
8245 files are provided by the operating system or by the GNU C library, but
8246 are provided by GCC for a few targets.
8248 The objects @file{crtbegin.o} and @file{crtend.o} are (for most targets)
8249 compiled from @file{crtstuff.c}. They contain, among other things, code
8250 fragments within the @code{.init} and @code{.fini} sections that branch
8251 to routines in the @code{.text} section. The linker will pull all parts
8252 of a section together, which results in a complete @code{__init} function
8253 that invokes the routines we need at startup.
8255 To use this variant, you must define the @code{INIT_SECTION_ASM_OP}
8258 If no init section is available, when GCC compiles any function called
8259 @code{main} (or more accurately, any function designated as a program
8260 entry point by the language front end calling @code{expand_main_function}),
8261 it inserts a procedure call to @code{__main} as the first executable code
8262 after the function prologue. The @code{__main} function is defined
8263 in @file{libgcc2.c} and runs the global constructors.
8265 In file formats that don't support arbitrary sections, there are again
8266 two variants. In the simplest variant, the GNU linker (GNU @code{ld})
8267 and an `a.out' format must be used. In this case,
8268 @code{TARGET_ASM_CONSTRUCTOR} is defined to produce a @code{.stabs}
8269 entry of type @samp{N_SETT}, referencing the name @code{__CTOR_LIST__},
8270 and with the address of the void function containing the initialization
8271 code as its value. The GNU linker recognizes this as a request to add
8272 the value to a @dfn{set}; the values are accumulated, and are eventually
8273 placed in the executable as a vector in the format described above, with
8274 a leading (ignored) count and a trailing zero element.
8275 @code{TARGET_ASM_DESTRUCTOR} is handled similarly. Since no init
8276 section is available, the absence of @code{INIT_SECTION_ASM_OP} causes
8277 the compilation of @code{main} to call @code{__main} as above, starting
8278 the initialization process.
8280 The last variant uses neither arbitrary sections nor the GNU linker.
8281 This is preferable when you want to do dynamic linking and when using
8282 file formats which the GNU linker does not support, such as `ECOFF'@. In
8283 this case, @code{TARGET_HAVE_CTORS_DTORS} is false, initialization and
8284 termination functions are recognized simply by their names. This requires
8285 an extra program in the linkage step, called @command{collect2}. This program
8286 pretends to be the linker, for use with GCC; it does its job by running
8287 the ordinary linker, but also arranges to include the vectors of
8288 initialization and termination functions. These functions are called
8289 via @code{__main} as described above. In order to use this method,
8290 @code{use_collect2} must be defined in the target in @file{config.gcc}.
8293 The following section describes the specific macros that control and
8294 customize the handling of initialization and termination functions.
8297 @node Macros for Initialization
8298 @subsection Macros Controlling Initialization Routines
8300 Here are the macros that control how the compiler handles initialization
8301 and termination functions:
8303 @defmac INIT_SECTION_ASM_OP
8304 If defined, a C string constant, including spacing, for the assembler
8305 operation to identify the following data as initialization code. If not
8306 defined, GCC will assume such a section does not exist. When you are
8307 using special sections for initialization and termination functions, this
8308 macro also controls how @file{crtstuff.c} and @file{libgcc2.c} arrange to
8309 run the initialization functions.
8312 @defmac HAS_INIT_SECTION
8313 If defined, @code{main} will not call @code{__main} as described above.
8314 This macro should be defined for systems that control start-up code
8315 on a symbol-by-symbol basis, such as OSF/1, and should not
8316 be defined explicitly for systems that support @code{INIT_SECTION_ASM_OP}.
8319 @defmac LD_INIT_SWITCH
8320 If defined, a C string constant for a switch that tells the linker that
8321 the following symbol is an initialization routine.
8324 @defmac LD_FINI_SWITCH
8325 If defined, a C string constant for a switch that tells the linker that
8326 the following symbol is a finalization routine.
8329 @defmac COLLECT_SHARED_INIT_FUNC (@var{stream}, @var{func})
8330 If defined, a C statement that will write a function that can be
8331 automatically called when a shared library is loaded. The function
8332 should call @var{func}, which takes no arguments. If not defined, and
8333 the object format requires an explicit initialization function, then a
8334 function called @code{_GLOBAL__DI} will be generated.
8336 This function and the following one are used by collect2 when linking a
8337 shared library that needs constructors or destructors, or has DWARF2
8338 exception tables embedded in the code.
8341 @defmac COLLECT_SHARED_FINI_FUNC (@var{stream}, @var{func})
8342 If defined, a C statement that will write a function that can be
8343 automatically called when a shared library is unloaded. The function
8344 should call @var{func}, which takes no arguments. If not defined, and
8345 the object format requires an explicit finalization function, then a
8346 function called @code{_GLOBAL__DD} will be generated.
8349 @defmac INVOKE__main
8350 If defined, @code{main} will call @code{__main} despite the presence of
8351 @code{INIT_SECTION_ASM_OP}. This macro should be defined for systems
8352 where the init section is not actually run automatically, but is still
8353 useful for collecting the lists of constructors and destructors.
8356 @defmac SUPPORTS_INIT_PRIORITY
8357 If nonzero, the C++ @code{init_priority} attribute is supported and the
8358 compiler should emit instructions to control the order of initialization
8359 of objects. If zero, the compiler will issue an error message upon
8360 encountering an @code{init_priority} attribute.
8363 @hook TARGET_HAVE_CTORS_DTORS
8364 This value is true if the target supports some ``native'' method of
8365 collecting constructors and destructors to be run at startup and exit.
8366 It is false if we must use @command{collect2}.
8369 @hook TARGET_ASM_CONSTRUCTOR
8370 If defined, a function that outputs assembler code to arrange to call
8371 the function referenced by @var{symbol} at initialization time.
8373 Assume that @var{symbol} is a @code{SYMBOL_REF} for a function taking
8374 no arguments and with no return value. If the target supports initialization
8375 priorities, @var{priority} is a value between 0 and @code{MAX_INIT_PRIORITY};
8376 otherwise it must be @code{DEFAULT_INIT_PRIORITY}.
8378 If this macro is not defined by the target, a suitable default will
8379 be chosen if (1) the target supports arbitrary section names, (2) the
8380 target defines @code{CTORS_SECTION_ASM_OP}, or (3) @code{USE_COLLECT2}
8384 @hook TARGET_ASM_DESTRUCTOR
8385 This is like @code{TARGET_ASM_CONSTRUCTOR} but used for termination
8386 functions rather than initialization functions.
8389 If @code{TARGET_HAVE_CTORS_DTORS} is true, the initialization routine
8390 generated for the generated object file will have static linkage.
8392 If your system uses @command{collect2} as the means of processing
8393 constructors, then that program normally uses @command{nm} to scan
8394 an object file for constructor functions to be called.
8396 On certain kinds of systems, you can define this macro to make
8397 @command{collect2} work faster (and, in some cases, make it work at all):
8399 @defmac OBJECT_FORMAT_COFF
8400 Define this macro if the system uses COFF (Common Object File Format)
8401 object files, so that @command{collect2} can assume this format and scan
8402 object files directly for dynamic constructor/destructor functions.
8404 This macro is effective only in a native compiler; @command{collect2} as
8405 part of a cross compiler always uses @command{nm} for the target machine.
8408 @defmac REAL_NM_FILE_NAME
8409 Define this macro as a C string constant containing the file name to use
8410 to execute @command{nm}. The default is to search the path normally for
8415 @command{collect2} calls @command{nm} to scan object files for static
8416 constructors and destructors and LTO info. By default, @option{-n} is
8417 passed. Define @code{NM_FLAGS} to a C string constant if other options
8418 are needed to get the same output format as GNU @command{nm -n}
8422 If your system supports shared libraries and has a program to list the
8423 dynamic dependencies of a given library or executable, you can define
8424 these macros to enable support for running initialization and
8425 termination functions in shared libraries:
8428 Define this macro to a C string constant containing the name of the program
8429 which lists dynamic dependencies, like @command{ldd} under SunOS 4.
8432 @defmac PARSE_LDD_OUTPUT (@var{ptr})
8433 Define this macro to be C code that extracts filenames from the output
8434 of the program denoted by @code{LDD_SUFFIX}. @var{ptr} is a variable
8435 of type @code{char *} that points to the beginning of a line of output
8436 from @code{LDD_SUFFIX}. If the line lists a dynamic dependency, the
8437 code must advance @var{ptr} to the beginning of the filename on that
8438 line. Otherwise, it must set @var{ptr} to @code{NULL}.
8441 @defmac SHLIB_SUFFIX
8442 Define this macro to a C string constant containing the default shared
8443 library extension of the target (e.g., @samp{".so"}). @command{collect2}
8444 strips version information after this suffix when generating global
8445 constructor and destructor names. This define is only needed on targets
8446 that use @command{collect2} to process constructors and destructors.
8449 @node Instruction Output
8450 @subsection Output of Assembler Instructions
8452 @c prevent bad page break with this line
8453 This describes assembler instruction output.
8455 @defmac REGISTER_NAMES
8456 A C initializer containing the assembler's names for the machine
8457 registers, each one as a C string constant. This is what translates
8458 register numbers in the compiler into assembler language.
8461 @defmac ADDITIONAL_REGISTER_NAMES
8462 If defined, a C initializer for an array of structures containing a name
8463 and a register number. This macro defines additional names for hard
8464 registers, thus allowing the @code{asm} option in declarations to refer
8465 to registers using alternate names.
8468 @defmac OVERLAPPING_REGISTER_NAMES
8469 If defined, a C initializer for an array of structures containing a
8470 name, a register number and a count of the number of consecutive
8471 machine registers the name overlaps. This macro defines additional
8472 names for hard registers, thus allowing the @code{asm} option in
8473 declarations to refer to registers using alternate names. Unlike
8474 @code{ADDITIONAL_REGISTER_NAMES}, this macro should be used when the
8475 register name implies multiple underlying registers.
8477 This macro should be used when it is important that a clobber in an
8478 @code{asm} statement clobbers all the underlying values implied by the
8479 register name. For example, on ARM, clobbering the double-precision
8480 VFP register ``d0'' implies clobbering both single-precision registers
8484 @defmac ASM_OUTPUT_OPCODE (@var{stream}, @var{ptr})
8485 Define this macro if you are using an unusual assembler that
8486 requires different names for the machine instructions.
8488 The definition is a C statement or statements which output an
8489 assembler instruction opcode to the stdio stream @var{stream}. The
8490 macro-operand @var{ptr} is a variable of type @code{char *} which
8491 points to the opcode name in its ``internal'' form---the form that is
8492 written in the machine description. The definition should output the
8493 opcode name to @var{stream}, performing any translation you desire, and
8494 increment the variable @var{ptr} to point at the end of the opcode
8495 so that it will not be output twice.
8497 In fact, your macro definition may process less than the entire opcode
8498 name, or more than the opcode name; but if you want to process text
8499 that includes @samp{%}-sequences to substitute operands, you must take
8500 care of the substitution yourself. Just be sure to increment
8501 @var{ptr} over whatever text should not be output normally.
8503 @findex recog_data.operand
8504 If you need to look at the operand values, they can be found as the
8505 elements of @code{recog_data.operand}.
8507 If the macro definition does nothing, the instruction is output
8511 @defmac FINAL_PRESCAN_INSN (@var{insn}, @var{opvec}, @var{noperands})
8512 If defined, a C statement to be executed just prior to the output of
8513 assembler code for @var{insn}, to modify the extracted operands so
8514 they will be output differently.
8516 Here the argument @var{opvec} is the vector containing the operands
8517 extracted from @var{insn}, and @var{noperands} is the number of
8518 elements of the vector which contain meaningful data for this insn.
8519 The contents of this vector are what will be used to convert the insn
8520 template into assembler code, so you can change the assembler output
8521 by changing the contents of the vector.
8523 This macro is useful when various assembler syntaxes share a single
8524 file of instruction patterns; by defining this macro differently, you
8525 can cause a large class of instructions to be output differently (such
8526 as with rearranged operands). Naturally, variations in assembler
8527 syntax affecting individual insn patterns ought to be handled by
8528 writing conditional output routines in those patterns.
8530 If this macro is not defined, it is equivalent to a null statement.
8533 @hook TARGET_ASM_FINAL_POSTSCAN_INSN
8534 If defined, this target hook is a function which is executed just after the
8535 output of assembler code for @var{insn}, to change the mode of the assembler
8538 Here the argument @var{opvec} is the vector containing the operands
8539 extracted from @var{insn}, and @var{noperands} is the number of
8540 elements of the vector which contain meaningful data for this insn.
8541 The contents of this vector are what was used to convert the insn
8542 template into assembler code, so you can change the assembler mode
8543 by checking the contents of the vector.
8546 @defmac PRINT_OPERAND (@var{stream}, @var{x}, @var{code})
8547 A C compound statement to output to stdio stream @var{stream} the
8548 assembler syntax for an instruction operand @var{x}. @var{x} is an
8551 @var{code} is a value that can be used to specify one of several ways
8552 of printing the operand. It is used when identical operands must be
8553 printed differently depending on the context. @var{code} comes from
8554 the @samp{%} specification that was used to request printing of the
8555 operand. If the specification was just @samp{%@var{digit}} then
8556 @var{code} is 0; if the specification was @samp{%@var{ltr}
8557 @var{digit}} then @var{code} is the ASCII code for @var{ltr}.
8560 If @var{x} is a register, this macro should print the register's name.
8561 The names can be found in an array @code{reg_names} whose type is
8562 @code{char *[]}. @code{reg_names} is initialized from
8563 @code{REGISTER_NAMES}.
8565 When the machine description has a specification @samp{%@var{punct}}
8566 (a @samp{%} followed by a punctuation character), this macro is called
8567 with a null pointer for @var{x} and the punctuation character for
8571 @defmac PRINT_OPERAND_PUNCT_VALID_P (@var{code})
8572 A C expression which evaluates to true if @var{code} is a valid
8573 punctuation character for use in the @code{PRINT_OPERAND} macro. If
8574 @code{PRINT_OPERAND_PUNCT_VALID_P} is not defined, it means that no
8575 punctuation characters (except for the standard one, @samp{%}) are used
8579 @defmac PRINT_OPERAND_ADDRESS (@var{stream}, @var{x})
8580 A C compound statement to output to stdio stream @var{stream} the
8581 assembler syntax for an instruction operand that is a memory reference
8582 whose address is @var{x}. @var{x} is an RTL expression.
8584 @cindex @code{TARGET_ENCODE_SECTION_INFO} usage
8585 On some machines, the syntax for a symbolic address depends on the
8586 section that the address refers to. On these machines, define the hook
8587 @code{TARGET_ENCODE_SECTION_INFO} to store the information into the
8588 @code{symbol_ref}, and then check for it here. @xref{Assembler
8592 @findex dbr_sequence_length
8593 @defmac DBR_OUTPUT_SEQEND (@var{file})
8594 A C statement, to be executed after all slot-filler instructions have
8595 been output. If necessary, call @code{dbr_sequence_length} to
8596 determine the number of slots filled in a sequence (zero if not
8597 currently outputting a sequence), to decide how many no-ops to output,
8600 Don't define this macro if it has nothing to do, but it is helpful in
8601 reading assembly output if the extent of the delay sequence is made
8602 explicit (e.g.@: with white space).
8605 @findex final_sequence
8606 Note that output routines for instructions with delay slots must be
8607 prepared to deal with not being output as part of a sequence
8608 (i.e.@: when the scheduling pass is not run, or when no slot fillers could be
8609 found.) The variable @code{final_sequence} is null when not
8610 processing a sequence, otherwise it contains the @code{sequence} rtx
8614 @defmac REGISTER_PREFIX
8615 @defmacx LOCAL_LABEL_PREFIX
8616 @defmacx USER_LABEL_PREFIX
8617 @defmacx IMMEDIATE_PREFIX
8618 If defined, C string expressions to be used for the @samp{%R}, @samp{%L},
8619 @samp{%U}, and @samp{%I} options of @code{asm_fprintf} (see
8620 @file{final.c}). These are useful when a single @file{md} file must
8621 support multiple assembler formats. In that case, the various @file{tm.h}
8622 files can define these macros differently.
8625 @defmac ASM_FPRINTF_EXTENSIONS (@var{file}, @var{argptr}, @var{format})
8626 If defined this macro should expand to a series of @code{case}
8627 statements which will be parsed inside the @code{switch} statement of
8628 the @code{asm_fprintf} function. This allows targets to define extra
8629 printf formats which may useful when generating their assembler
8630 statements. Note that uppercase letters are reserved for future
8631 generic extensions to asm_fprintf, and so are not available to target
8632 specific code. The output file is given by the parameter @var{file}.
8633 The varargs input pointer is @var{argptr} and the rest of the format
8634 string, starting the character after the one that is being switched
8635 upon, is pointed to by @var{format}.
8638 @defmac ASSEMBLER_DIALECT
8639 If your target supports multiple dialects of assembler language (such as
8640 different opcodes), define this macro as a C expression that gives the
8641 numeric index of the assembler language dialect to use, with zero as the
8644 If this macro is defined, you may use constructs of the form
8646 @samp{@{option0|option1|option2@dots{}@}}
8649 in the output templates of patterns (@pxref{Output Template}) or in the
8650 first argument of @code{asm_fprintf}. This construct outputs
8651 @samp{option0}, @samp{option1}, @samp{option2}, etc., if the value of
8652 @code{ASSEMBLER_DIALECT} is zero, one, two, etc. Any special characters
8653 within these strings retain their usual meaning. If there are fewer
8654 alternatives within the braces than the value of
8655 @code{ASSEMBLER_DIALECT}, the construct outputs nothing.
8657 If you do not define this macro, the characters @samp{@{}, @samp{|} and
8658 @samp{@}} do not have any special meaning when used in templates or
8659 operands to @code{asm_fprintf}.
8661 Define the macros @code{REGISTER_PREFIX}, @code{LOCAL_LABEL_PREFIX},
8662 @code{USER_LABEL_PREFIX} and @code{IMMEDIATE_PREFIX} if you can express
8663 the variations in assembler language syntax with that mechanism. Define
8664 @code{ASSEMBLER_DIALECT} and use the @samp{@{option0|option1@}} syntax
8665 if the syntax variant are larger and involve such things as different
8666 opcodes or operand order.
8669 @defmac ASM_OUTPUT_REG_PUSH (@var{stream}, @var{regno})
8670 A C expression to output to @var{stream} some assembler code
8671 which will push hard register number @var{regno} onto the stack.
8672 The code need not be optimal, since this macro is used only when
8676 @defmac ASM_OUTPUT_REG_POP (@var{stream}, @var{regno})
8677 A C expression to output to @var{stream} some assembler code
8678 which will pop hard register number @var{regno} off of the stack.
8679 The code need not be optimal, since this macro is used only when
8683 @node Dispatch Tables
8684 @subsection Output of Dispatch Tables
8686 @c prevent bad page break with this line
8687 This concerns dispatch tables.
8689 @cindex dispatch table
8690 @defmac ASM_OUTPUT_ADDR_DIFF_ELT (@var{stream}, @var{body}, @var{value}, @var{rel})
8691 A C statement to output to the stdio stream @var{stream} an assembler
8692 pseudo-instruction to generate a difference between two labels.
8693 @var{value} and @var{rel} are the numbers of two internal labels. The
8694 definitions of these labels are output using
8695 @code{(*targetm.asm_out.internal_label)}, and they must be printed in the same
8696 way here. For example,
8699 fprintf (@var{stream}, "\t.word L%d-L%d\n",
8700 @var{value}, @var{rel})
8703 You must provide this macro on machines where the addresses in a
8704 dispatch table are relative to the table's own address. If defined, GCC
8705 will also use this macro on all machines when producing PIC@.
8706 @var{body} is the body of the @code{ADDR_DIFF_VEC}; it is provided so that the
8707 mode and flags can be read.
8710 @defmac ASM_OUTPUT_ADDR_VEC_ELT (@var{stream}, @var{value})
8711 This macro should be provided on machines where the addresses
8712 in a dispatch table are absolute.
8714 The definition should be a C statement to output to the stdio stream
8715 @var{stream} an assembler pseudo-instruction to generate a reference to
8716 a label. @var{value} is the number of an internal label whose
8717 definition is output using @code{(*targetm.asm_out.internal_label)}.
8721 fprintf (@var{stream}, "\t.word L%d\n", @var{value})
8725 @defmac ASM_OUTPUT_CASE_LABEL (@var{stream}, @var{prefix}, @var{num}, @var{table})
8726 Define this if the label before a jump-table needs to be output
8727 specially. The first three arguments are the same as for
8728 @code{(*targetm.asm_out.internal_label)}; the fourth argument is the
8729 jump-table which follows (a @code{jump_insn} containing an
8730 @code{addr_vec} or @code{addr_diff_vec}).
8732 This feature is used on system V to output a @code{swbeg} statement
8735 If this macro is not defined, these labels are output with
8736 @code{(*targetm.asm_out.internal_label)}.
8739 @defmac ASM_OUTPUT_CASE_END (@var{stream}, @var{num}, @var{table})
8740 Define this if something special must be output at the end of a
8741 jump-table. The definition should be a C statement to be executed
8742 after the assembler code for the table is written. It should write
8743 the appropriate code to stdio stream @var{stream}. The argument
8744 @var{table} is the jump-table insn, and @var{num} is the label-number
8745 of the preceding label.
8747 If this macro is not defined, nothing special is output at the end of
8751 @hook TARGET_ASM_EMIT_UNWIND_LABEL
8752 This target hook emits a label at the beginning of each FDE@. It
8753 should be defined on targets where FDEs need special labels, and it
8754 should write the appropriate label, for the FDE associated with the
8755 function declaration @var{decl}, to the stdio stream @var{stream}.
8756 The third argument, @var{for_eh}, is a boolean: true if this is for an
8757 exception table. The fourth argument, @var{empty}, is a boolean:
8758 true if this is a placeholder label for an omitted FDE@.
8760 The default is that FDEs are not given nonlocal labels.
8763 @hook TARGET_ASM_EMIT_EXCEPT_TABLE_LABEL
8764 This target hook emits a label at the beginning of the exception table.
8765 It should be defined on targets where it is desirable for the table
8766 to be broken up according to function.
8768 The default is that no label is emitted.
8771 @hook TARGET_ASM_EMIT_EXCEPT_PERSONALITY
8773 @hook TARGET_ASM_UNWIND_EMIT
8774 This target hook emits assembly directives required to unwind the
8775 given instruction. This is only used when @code{TARGET_EXCEPT_UNWIND_INFO}
8776 returns @code{UI_TARGET}.
8779 @hook TARGET_ASM_UNWIND_EMIT_BEFORE_INSN
8781 @node Exception Region Output
8782 @subsection Assembler Commands for Exception Regions
8784 @c prevent bad page break with this line
8786 This describes commands marking the start and the end of an exception
8789 @defmac EH_FRAME_SECTION_NAME
8790 If defined, a C string constant for the name of the section containing
8791 exception handling frame unwind information. If not defined, GCC will
8792 provide a default definition if the target supports named sections.
8793 @file{crtstuff.c} uses this macro to switch to the appropriate section.
8795 You should define this symbol if your target supports DWARF 2 frame
8796 unwind information and the default definition does not work.
8799 @defmac EH_FRAME_IN_DATA_SECTION
8800 If defined, DWARF 2 frame unwind information will be placed in the
8801 data section even though the target supports named sections. This
8802 might be necessary, for instance, if the system linker does garbage
8803 collection and sections cannot be marked as not to be collected.
8805 Do not define this macro unless @code{TARGET_ASM_NAMED_SECTION} is
8809 @defmac EH_TABLES_CAN_BE_READ_ONLY
8810 Define this macro to 1 if your target is such that no frame unwind
8811 information encoding used with non-PIC code will ever require a
8812 runtime relocation, but the linker may not support merging read-only
8813 and read-write sections into a single read-write section.
8816 @defmac MASK_RETURN_ADDR
8817 An rtx used to mask the return address found via @code{RETURN_ADDR_RTX}, so
8818 that it does not contain any extraneous set bits in it.
8821 @defmac DWARF2_UNWIND_INFO
8822 Define this macro to 0 if your target supports DWARF 2 frame unwind
8823 information, but it does not yet work with exception handling.
8824 Otherwise, if your target supports this information (if it defines
8825 @code{INCOMING_RETURN_ADDR_RTX} and either @code{UNALIGNED_INT_ASM_OP}
8826 or @code{OBJECT_FORMAT_ELF}), GCC will provide a default definition of 1.
8829 @hook TARGET_EXCEPT_UNWIND_INFO
8830 This hook defines the mechanism that will be used for exception handling
8831 by the target. If the target has ABI specified unwind tables, the hook
8832 should return @code{UI_TARGET}. If the target is to use the
8833 @code{setjmp}/@code{longjmp}-based exception handling scheme, the hook
8834 should return @code{UI_SJLJ}. If the target supports DWARF 2 frame unwind
8835 information, the hook should return @code{UI_DWARF2}.
8837 A target may, if exceptions are disabled, choose to return @code{UI_NONE}.
8838 This may end up simplifying other parts of target-specific code. The
8839 default implementation of this hook never returns @code{UI_NONE}.
8841 Note that the value returned by this hook should be constant. It should
8842 not depend on anything except the command-line switches described by
8843 @var{opts}. In particular, the
8844 setting @code{UI_SJLJ} must be fixed at compiler start-up as C pre-processor
8845 macros and builtin functions related to exception handling are set up
8846 depending on this setting.
8848 The default implementation of the hook first honors the
8849 @option{--enable-sjlj-exceptions} configure option, then
8850 @code{DWARF2_UNWIND_INFO}, and finally defaults to @code{UI_SJLJ}. If
8851 @code{DWARF2_UNWIND_INFO} depends on command-line options, the target
8852 must define this hook so that @var{opts} is used correctly.
8855 @hook TARGET_UNWIND_TABLES_DEFAULT
8856 This variable should be set to @code{true} if the target ABI requires unwinding
8857 tables even when exceptions are not used. It must not be modified by
8858 command-line option processing.
8861 @defmac DONT_USE_BUILTIN_SETJMP
8862 Define this macro to 1 if the @code{setjmp}/@code{longjmp}-based scheme
8863 should use the @code{setjmp}/@code{longjmp} functions from the C library
8864 instead of the @code{__builtin_setjmp}/@code{__builtin_longjmp} machinery.
8867 @defmac DWARF_CIE_DATA_ALIGNMENT
8868 This macro need only be defined if the target might save registers in the
8869 function prologue at an offset to the stack pointer that is not aligned to
8870 @code{UNITS_PER_WORD}. The definition should be the negative minimum
8871 alignment if @code{STACK_GROWS_DOWNWARD} is defined, and the positive
8872 minimum alignment otherwise. @xref{SDB and DWARF}. Only applicable if
8873 the target supports DWARF 2 frame unwind information.
8876 @hook TARGET_TERMINATE_DW2_EH_FRAME_INFO
8877 Contains the value true if the target should add a zero word onto the
8878 end of a Dwarf-2 frame info section when used for exception handling.
8879 Default value is false if @code{EH_FRAME_SECTION_NAME} is defined, and
8883 @hook TARGET_DWARF_REGISTER_SPAN
8884 Given a register, this hook should return a parallel of registers to
8885 represent where to find the register pieces. Define this hook if the
8886 register and its mode are represented in Dwarf in non-contiguous
8887 locations, or if the register should be represented in more than one
8888 register in Dwarf. Otherwise, this hook should return @code{NULL_RTX}.
8889 If not defined, the default is to return @code{NULL_RTX}.
8892 @hook TARGET_INIT_DWARF_REG_SIZES_EXTRA
8893 If some registers are represented in Dwarf-2 unwind information in
8894 multiple pieces, define this hook to fill in information about the
8895 sizes of those pieces in the table used by the unwinder at runtime.
8896 It will be called by @code{expand_builtin_init_dwarf_reg_sizes} after
8897 filling in a single size corresponding to each hard register;
8898 @var{address} is the address of the table.
8901 @hook TARGET_ASM_TTYPE
8902 This hook is used to output a reference from a frame unwinding table to
8903 the type_info object identified by @var{sym}. It should return @code{true}
8904 if the reference was output. Returning @code{false} will cause the
8905 reference to be output using the normal Dwarf2 routines.
8908 @hook TARGET_ARM_EABI_UNWINDER
8909 This flag should be set to @code{true} on targets that use an ARM EABI
8910 based unwinding library, and @code{false} on other targets. This effects
8911 the format of unwinding tables, and how the unwinder in entered after
8912 running a cleanup. The default is @code{false}.
8915 @node Alignment Output
8916 @subsection Assembler Commands for Alignment
8918 @c prevent bad page break with this line
8919 This describes commands for alignment.
8921 @defmac JUMP_ALIGN (@var{label})
8922 The alignment (log base 2) to put in front of @var{label}, which is
8923 a common destination of jumps and has no fallthru incoming edge.
8925 This macro need not be defined if you don't want any special alignment
8926 to be done at such a time. Most machine descriptions do not currently
8929 Unless it's necessary to inspect the @var{label} parameter, it is better
8930 to set the variable @var{align_jumps} in the target's
8931 @code{TARGET_OPTION_OVERRIDE}. Otherwise, you should try to honor the user's
8932 selection in @var{align_jumps} in a @code{JUMP_ALIGN} implementation.
8935 @hook TARGET_ASM_JUMP_ALIGN_MAX_SKIP
8936 The maximum number of bytes to skip before @var{label} when applying
8937 @code{JUMP_ALIGN}. This works only if
8938 @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
8941 @defmac LABEL_ALIGN_AFTER_BARRIER (@var{label})
8942 The alignment (log base 2) to put in front of @var{label}, which follows
8945 This macro need not be defined if you don't want any special alignment
8946 to be done at such a time. Most machine descriptions do not currently
8950 @hook TARGET_ASM_LABEL_ALIGN_AFTER_BARRIER_MAX_SKIP
8951 The maximum number of bytes to skip before @var{label} when applying
8952 @code{LABEL_ALIGN_AFTER_BARRIER}. This works only if
8953 @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
8956 @defmac LOOP_ALIGN (@var{label})
8957 The alignment (log base 2) to put in front of @var{label}, which follows
8958 a @code{NOTE_INSN_LOOP_BEG} note.
8960 This macro need not be defined if you don't want any special alignment
8961 to be done at such a time. Most machine descriptions do not currently
8964 Unless it's necessary to inspect the @var{label} parameter, it is better
8965 to set the variable @code{align_loops} in the target's
8966 @code{TARGET_OPTION_OVERRIDE}. Otherwise, you should try to honor the user's
8967 selection in @code{align_loops} in a @code{LOOP_ALIGN} implementation.
8970 @hook TARGET_ASM_LOOP_ALIGN_MAX_SKIP
8971 The maximum number of bytes to skip when applying @code{LOOP_ALIGN} to
8972 @var{label}. This works only if @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is
8976 @defmac LABEL_ALIGN (@var{label})
8977 The alignment (log base 2) to put in front of @var{label}.
8978 If @code{LABEL_ALIGN_AFTER_BARRIER} / @code{LOOP_ALIGN} specify a different alignment,
8979 the maximum of the specified values is used.
8981 Unless it's necessary to inspect the @var{label} parameter, it is better
8982 to set the variable @code{align_labels} in the target's
8983 @code{TARGET_OPTION_OVERRIDE}. Otherwise, you should try to honor the user's
8984 selection in @code{align_labels} in a @code{LABEL_ALIGN} implementation.
8987 @hook TARGET_ASM_LABEL_ALIGN_MAX_SKIP
8988 The maximum number of bytes to skip when applying @code{LABEL_ALIGN}
8989 to @var{label}. This works only if @code{ASM_OUTPUT_MAX_SKIP_ALIGN}
8993 @defmac ASM_OUTPUT_SKIP (@var{stream}, @var{nbytes})
8994 A C statement to output to the stdio stream @var{stream} an assembler
8995 instruction to advance the location counter by @var{nbytes} bytes.
8996 Those bytes should be zero when loaded. @var{nbytes} will be a C
8997 expression of type @code{unsigned HOST_WIDE_INT}.
9000 @defmac ASM_NO_SKIP_IN_TEXT
9001 Define this macro if @code{ASM_OUTPUT_SKIP} should not be used in the
9002 text section because it fails to put zeros in the bytes that are skipped.
9003 This is true on many Unix systems, where the pseudo--op to skip bytes
9004 produces no-op instructions rather than zeros when used in the text
9008 @defmac ASM_OUTPUT_ALIGN (@var{stream}, @var{power})
9009 A C statement to output to the stdio stream @var{stream} an assembler
9010 command to advance the location counter to a multiple of 2 to the
9011 @var{power} bytes. @var{power} will be a C expression of type @code{int}.
9014 @defmac ASM_OUTPUT_ALIGN_WITH_NOP (@var{stream}, @var{power})
9015 Like @code{ASM_OUTPUT_ALIGN}, except that the ``nop'' instruction is used
9016 for padding, if necessary.
9019 @defmac ASM_OUTPUT_MAX_SKIP_ALIGN (@var{stream}, @var{power}, @var{max_skip})
9020 A C statement to output to the stdio stream @var{stream} an assembler
9021 command to advance the location counter to a multiple of 2 to the
9022 @var{power} bytes, but only if @var{max_skip} or fewer bytes are needed to
9023 satisfy the alignment request. @var{power} and @var{max_skip} will be
9024 a C expression of type @code{int}.
9028 @node Debugging Info
9029 @section Controlling Debugging Information Format
9031 @c prevent bad page break with this line
9032 This describes how to specify debugging information.
9035 * All Debuggers:: Macros that affect all debugging formats uniformly.
9036 * DBX Options:: Macros enabling specific options in DBX format.
9037 * DBX Hooks:: Hook macros for varying DBX format.
9038 * File Names and DBX:: Macros controlling output of file names in DBX format.
9039 * SDB and DWARF:: Macros for SDB (COFF) and DWARF formats.
9040 * VMS Debug:: Macros for VMS debug format.
9044 @subsection Macros Affecting All Debugging Formats
9046 @c prevent bad page break with this line
9047 These macros affect all debugging formats.
9049 @defmac DBX_REGISTER_NUMBER (@var{regno})
9050 A C expression that returns the DBX register number for the compiler
9051 register number @var{regno}. In the default macro provided, the value
9052 of this expression will be @var{regno} itself. But sometimes there are
9053 some registers that the compiler knows about and DBX does not, or vice
9054 versa. In such cases, some register may need to have one number in the
9055 compiler and another for DBX@.
9057 If two registers have consecutive numbers inside GCC, and they can be
9058 used as a pair to hold a multiword value, then they @emph{must} have
9059 consecutive numbers after renumbering with @code{DBX_REGISTER_NUMBER}.
9060 Otherwise, debuggers will be unable to access such a pair, because they
9061 expect register pairs to be consecutive in their own numbering scheme.
9063 If you find yourself defining @code{DBX_REGISTER_NUMBER} in way that
9064 does not preserve register pairs, then what you must do instead is
9065 redefine the actual register numbering scheme.
9068 @defmac DEBUGGER_AUTO_OFFSET (@var{x})
9069 A C expression that returns the integer offset value for an automatic
9070 variable having address @var{x} (an RTL expression). The default
9071 computation assumes that @var{x} is based on the frame-pointer and
9072 gives the offset from the frame-pointer. This is required for targets
9073 that produce debugging output for DBX or COFF-style debugging output
9074 for SDB and allow the frame-pointer to be eliminated when the
9075 @option{-g} options is used.
9078 @defmac DEBUGGER_ARG_OFFSET (@var{offset}, @var{x})
9079 A C expression that returns the integer offset value for an argument
9080 having address @var{x} (an RTL expression). The nominal offset is
9084 @defmac PREFERRED_DEBUGGING_TYPE
9085 A C expression that returns the type of debugging output GCC should
9086 produce when the user specifies just @option{-g}. Define
9087 this if you have arranged for GCC to support more than one format of
9088 debugging output. Currently, the allowable values are @code{DBX_DEBUG},
9089 @code{SDB_DEBUG}, @code{DWARF_DEBUG}, @code{DWARF2_DEBUG},
9090 @code{XCOFF_DEBUG}, @code{VMS_DEBUG}, and @code{VMS_AND_DWARF2_DEBUG}.
9092 When the user specifies @option{-ggdb}, GCC normally also uses the
9093 value of this macro to select the debugging output format, but with two
9094 exceptions. If @code{DWARF2_DEBUGGING_INFO} is defined, GCC uses the
9095 value @code{DWARF2_DEBUG}. Otherwise, if @code{DBX_DEBUGGING_INFO} is
9096 defined, GCC uses @code{DBX_DEBUG}.
9098 The value of this macro only affects the default debugging output; the
9099 user can always get a specific type of output by using @option{-gstabs},
9100 @option{-gcoff}, @option{-gdwarf-2}, @option{-gxcoff}, or @option{-gvms}.
9104 @subsection Specific Options for DBX Output
9106 @c prevent bad page break with this line
9107 These are specific options for DBX output.
9109 @defmac DBX_DEBUGGING_INFO
9110 Define this macro if GCC should produce debugging output for DBX
9111 in response to the @option{-g} option.
9114 @defmac XCOFF_DEBUGGING_INFO
9115 Define this macro if GCC should produce XCOFF format debugging output
9116 in response to the @option{-g} option. This is a variant of DBX format.
9119 @defmac DEFAULT_GDB_EXTENSIONS
9120 Define this macro to control whether GCC should by default generate
9121 GDB's extended version of DBX debugging information (assuming DBX-format
9122 debugging information is enabled at all). If you don't define the
9123 macro, the default is 1: always generate the extended information
9124 if there is any occasion to.
9127 @defmac DEBUG_SYMS_TEXT
9128 Define this macro if all @code{.stabs} commands should be output while
9129 in the text section.
9132 @defmac ASM_STABS_OP
9133 A C string constant, including spacing, naming the assembler pseudo op to
9134 use instead of @code{"\t.stabs\t"} to define an ordinary debugging symbol.
9135 If you don't define this macro, @code{"\t.stabs\t"} is used. This macro
9136 applies only to DBX debugging information format.
9139 @defmac ASM_STABD_OP
9140 A C string constant, including spacing, naming the assembler pseudo op to
9141 use instead of @code{"\t.stabd\t"} to define a debugging symbol whose
9142 value is the current location. If you don't define this macro,
9143 @code{"\t.stabd\t"} is used. This macro applies only to DBX debugging
9147 @defmac ASM_STABN_OP
9148 A C string constant, including spacing, naming the assembler pseudo op to
9149 use instead of @code{"\t.stabn\t"} to define a debugging symbol with no
9150 name. If you don't define this macro, @code{"\t.stabn\t"} is used. This
9151 macro applies only to DBX debugging information format.
9154 @defmac DBX_NO_XREFS
9155 Define this macro if DBX on your system does not support the construct
9156 @samp{xs@var{tagname}}. On some systems, this construct is used to
9157 describe a forward reference to a structure named @var{tagname}.
9158 On other systems, this construct is not supported at all.
9161 @defmac DBX_CONTIN_LENGTH
9162 A symbol name in DBX-format debugging information is normally
9163 continued (split into two separate @code{.stabs} directives) when it
9164 exceeds a certain length (by default, 80 characters). On some
9165 operating systems, DBX requires this splitting; on others, splitting
9166 must not be done. You can inhibit splitting by defining this macro
9167 with the value zero. You can override the default splitting-length by
9168 defining this macro as an expression for the length you desire.
9171 @defmac DBX_CONTIN_CHAR
9172 Normally continuation is indicated by adding a @samp{\} character to
9173 the end of a @code{.stabs} string when a continuation follows. To use
9174 a different character instead, define this macro as a character
9175 constant for the character you want to use. Do not define this macro
9176 if backslash is correct for your system.
9179 @defmac DBX_STATIC_STAB_DATA_SECTION
9180 Define this macro if it is necessary to go to the data section before
9181 outputting the @samp{.stabs} pseudo-op for a non-global static
9185 @defmac DBX_TYPE_DECL_STABS_CODE
9186 The value to use in the ``code'' field of the @code{.stabs} directive
9187 for a typedef. The default is @code{N_LSYM}.
9190 @defmac DBX_STATIC_CONST_VAR_CODE
9191 The value to use in the ``code'' field of the @code{.stabs} directive
9192 for a static variable located in the text section. DBX format does not
9193 provide any ``right'' way to do this. The default is @code{N_FUN}.
9196 @defmac DBX_REGPARM_STABS_CODE
9197 The value to use in the ``code'' field of the @code{.stabs} directive
9198 for a parameter passed in registers. DBX format does not provide any
9199 ``right'' way to do this. The default is @code{N_RSYM}.
9202 @defmac DBX_REGPARM_STABS_LETTER
9203 The letter to use in DBX symbol data to identify a symbol as a parameter
9204 passed in registers. DBX format does not customarily provide any way to
9205 do this. The default is @code{'P'}.
9208 @defmac DBX_FUNCTION_FIRST
9209 Define this macro if the DBX information for a function and its
9210 arguments should precede the assembler code for the function. Normally,
9211 in DBX format, the debugging information entirely follows the assembler
9215 @defmac DBX_BLOCKS_FUNCTION_RELATIVE
9216 Define this macro, with value 1, if the value of a symbol describing
9217 the scope of a block (@code{N_LBRAC} or @code{N_RBRAC}) should be
9218 relative to the start of the enclosing function. Normally, GCC uses
9219 an absolute address.
9222 @defmac DBX_LINES_FUNCTION_RELATIVE
9223 Define this macro, with value 1, if the value of a symbol indicating
9224 the current line number (@code{N_SLINE}) should be relative to the
9225 start of the enclosing function. Normally, GCC uses an absolute address.
9228 @defmac DBX_USE_BINCL
9229 Define this macro if GCC should generate @code{N_BINCL} and
9230 @code{N_EINCL} stabs for included header files, as on Sun systems. This
9231 macro also directs GCC to output a type number as a pair of a file
9232 number and a type number within the file. Normally, GCC does not
9233 generate @code{N_BINCL} or @code{N_EINCL} stabs, and it outputs a single
9234 number for a type number.
9238 @subsection Open-Ended Hooks for DBX Format
9240 @c prevent bad page break with this line
9241 These are hooks for DBX format.
9243 @defmac DBX_OUTPUT_LBRAC (@var{stream}, @var{name})
9244 Define this macro to say how to output to @var{stream} the debugging
9245 information for the start of a scope level for variable names. The
9246 argument @var{name} is the name of an assembler symbol (for use with
9247 @code{assemble_name}) whose value is the address where the scope begins.
9250 @defmac DBX_OUTPUT_RBRAC (@var{stream}, @var{name})
9251 Like @code{DBX_OUTPUT_LBRAC}, but for the end of a scope level.
9254 @defmac DBX_OUTPUT_NFUN (@var{stream}, @var{lscope_label}, @var{decl})
9255 Define this macro if the target machine requires special handling to
9256 output an @code{N_FUN} entry for the function @var{decl}.
9259 @defmac DBX_OUTPUT_SOURCE_LINE (@var{stream}, @var{line}, @var{counter})
9260 A C statement to output DBX debugging information before code for line
9261 number @var{line} of the current source file to the stdio stream
9262 @var{stream}. @var{counter} is the number of time the macro was
9263 invoked, including the current invocation; it is intended to generate
9264 unique labels in the assembly output.
9266 This macro should not be defined if the default output is correct, or
9267 if it can be made correct by defining @code{DBX_LINES_FUNCTION_RELATIVE}.
9270 @defmac NO_DBX_FUNCTION_END
9271 Some stabs encapsulation formats (in particular ECOFF), cannot handle the
9272 @code{.stabs "",N_FUN,,0,0,Lscope-function-1} gdb dbx extension construct.
9273 On those machines, define this macro to turn this feature off without
9274 disturbing the rest of the gdb extensions.
9277 @defmac NO_DBX_BNSYM_ENSYM
9278 Some assemblers cannot handle the @code{.stabd BNSYM/ENSYM,0,0} gdb dbx
9279 extension construct. On those machines, define this macro to turn this
9280 feature off without disturbing the rest of the gdb extensions.
9283 @node File Names and DBX
9284 @subsection File Names in DBX Format
9286 @c prevent bad page break with this line
9287 This describes file names in DBX format.
9289 @defmac DBX_OUTPUT_MAIN_SOURCE_FILENAME (@var{stream}, @var{name})
9290 A C statement to output DBX debugging information to the stdio stream
9291 @var{stream}, which indicates that file @var{name} is the main source
9292 file---the file specified as the input file for compilation.
9293 This macro is called only once, at the beginning of compilation.
9295 This macro need not be defined if the standard form of output
9296 for DBX debugging information is appropriate.
9298 It may be necessary to refer to a label equal to the beginning of the
9299 text section. You can use @samp{assemble_name (stream, ltext_label_name)}
9300 to do so. If you do this, you must also set the variable
9301 @var{used_ltext_label_name} to @code{true}.
9304 @defmac NO_DBX_MAIN_SOURCE_DIRECTORY
9305 Define this macro, with value 1, if GCC should not emit an indication
9306 of the current directory for compilation and current source language at
9307 the beginning of the file.
9310 @defmac NO_DBX_GCC_MARKER
9311 Define this macro, with value 1, if GCC should not emit an indication
9312 that this object file was compiled by GCC@. The default is to emit
9313 an @code{N_OPT} stab at the beginning of every source file, with
9314 @samp{gcc2_compiled.} for the string and value 0.
9317 @defmac DBX_OUTPUT_MAIN_SOURCE_FILE_END (@var{stream}, @var{name})
9318 A C statement to output DBX debugging information at the end of
9319 compilation of the main source file @var{name}. Output should be
9320 written to the stdio stream @var{stream}.
9322 If you don't define this macro, nothing special is output at the end
9323 of compilation, which is correct for most machines.
9326 @defmac DBX_OUTPUT_NULL_N_SO_AT_MAIN_SOURCE_FILE_END
9327 Define this macro @emph{instead of} defining
9328 @code{DBX_OUTPUT_MAIN_SOURCE_FILE_END}, if what needs to be output at
9329 the end of compilation is an @code{N_SO} stab with an empty string,
9330 whose value is the highest absolute text address in the file.
9335 @subsection Macros for SDB and DWARF Output
9337 @c prevent bad page break with this line
9338 Here are macros for SDB and DWARF output.
9340 @defmac SDB_DEBUGGING_INFO
9341 Define this macro if GCC should produce COFF-style debugging output
9342 for SDB in response to the @option{-g} option.
9345 @defmac DWARF2_DEBUGGING_INFO
9346 Define this macro if GCC should produce dwarf version 2 format
9347 debugging output in response to the @option{-g} option.
9349 @hook TARGET_DWARF_CALLING_CONVENTION
9350 Define this to enable the dwarf attribute @code{DW_AT_calling_convention} to
9351 be emitted for each function. Instead of an integer return the enum
9352 value for the @code{DW_CC_} tag.
9355 To support optional call frame debugging information, you must also
9356 define @code{INCOMING_RETURN_ADDR_RTX} and either set
9357 @code{RTX_FRAME_RELATED_P} on the prologue insns if you use RTL for the
9358 prologue, or call @code{dwarf2out_def_cfa} and @code{dwarf2out_reg_save}
9359 as appropriate from @code{TARGET_ASM_FUNCTION_PROLOGUE} if you don't.
9362 @defmac DWARF2_FRAME_INFO
9363 Define this macro to a nonzero value if GCC should always output
9364 Dwarf 2 frame information. If @code{TARGET_EXCEPT_UNWIND_INFO}
9365 (@pxref{Exception Region Output}) returns @code{UI_DWARF2}, and
9366 exceptions are enabled, GCC will output this information not matter
9367 how you define @code{DWARF2_FRAME_INFO}.
9370 @hook TARGET_DEBUG_UNWIND_INFO
9371 This hook defines the mechanism that will be used for describing frame
9372 unwind information to the debugger. Normally the hook will return
9373 @code{UI_DWARF2} if DWARF 2 debug information is enabled, and
9374 return @code{UI_NONE} otherwise.
9376 A target may return @code{UI_DWARF2} even when DWARF 2 debug information
9377 is disabled in order to always output DWARF 2 frame information.
9379 A target may return @code{UI_TARGET} if it has ABI specified unwind tables.
9380 This will suppress generation of the normal debug frame unwind information.
9383 @defmac DWARF2_ASM_LINE_DEBUG_INFO
9384 Define this macro to be a nonzero value if the assembler can generate Dwarf 2
9385 line debug info sections. This will result in much more compact line number
9386 tables, and hence is desirable if it works.
9389 @hook TARGET_WANT_DEBUG_PUB_SECTIONS
9391 @hook TARGET_DELAY_SCHED2
9393 @hook TARGET_DELAY_VARTRACK
9395 @defmac ASM_OUTPUT_DWARF_DELTA (@var{stream}, @var{size}, @var{label1}, @var{label2})
9396 A C statement to issue assembly directives that create a difference
9397 @var{lab1} minus @var{lab2}, using an integer of the given @var{size}.
9400 @defmac ASM_OUTPUT_DWARF_VMS_DELTA (@var{stream}, @var{size}, @var{label1}, @var{label2})
9401 A C statement to issue assembly directives that create a difference
9402 between the two given labels in system defined units, e.g. instruction
9403 slots on IA64 VMS, using an integer of the given size.
9406 @defmac ASM_OUTPUT_DWARF_OFFSET (@var{stream}, @var{size}, @var{label}, @var{section})
9407 A C statement to issue assembly directives that create a
9408 section-relative reference to the given @var{label}, using an integer of the
9409 given @var{size}. The label is known to be defined in the given @var{section}.
9412 @defmac ASM_OUTPUT_DWARF_PCREL (@var{stream}, @var{size}, @var{label})
9413 A C statement to issue assembly directives that create a self-relative
9414 reference to the given @var{label}, using an integer of the given @var{size}.
9417 @defmac ASM_OUTPUT_DWARF_TABLE_REF (@var{label})
9418 A C statement to issue assembly directives that create a reference to
9419 the DWARF table identifier @var{label} from the current section. This
9420 is used on some systems to avoid garbage collecting a DWARF table which
9421 is referenced by a function.
9424 @hook TARGET_ASM_OUTPUT_DWARF_DTPREL
9425 If defined, this target hook is a function which outputs a DTP-relative
9426 reference to the given TLS symbol of the specified size.
9429 @defmac PUT_SDB_@dots{}
9430 Define these macros to override the assembler syntax for the special
9431 SDB assembler directives. See @file{sdbout.c} for a list of these
9432 macros and their arguments. If the standard syntax is used, you need
9433 not define them yourself.
9437 Some assemblers do not support a semicolon as a delimiter, even between
9438 SDB assembler directives. In that case, define this macro to be the
9439 delimiter to use (usually @samp{\n}). It is not necessary to define
9440 a new set of @code{PUT_SDB_@var{op}} macros if this is the only change
9444 @defmac SDB_ALLOW_UNKNOWN_REFERENCES
9445 Define this macro to allow references to unknown structure,
9446 union, or enumeration tags to be emitted. Standard COFF does not
9447 allow handling of unknown references, MIPS ECOFF has support for
9451 @defmac SDB_ALLOW_FORWARD_REFERENCES
9452 Define this macro to allow references to structure, union, or
9453 enumeration tags that have not yet been seen to be handled. Some
9454 assemblers choke if forward tags are used, while some require it.
9457 @defmac SDB_OUTPUT_SOURCE_LINE (@var{stream}, @var{line})
9458 A C statement to output SDB debugging information before code for line
9459 number @var{line} of the current source file to the stdio stream
9460 @var{stream}. The default is to emit an @code{.ln} directive.
9465 @subsection Macros for VMS Debug Format
9467 @c prevent bad page break with this line
9468 Here are macros for VMS debug format.
9470 @defmac VMS_DEBUGGING_INFO
9471 Define this macro if GCC should produce debugging output for VMS
9472 in response to the @option{-g} option. The default behavior for VMS
9473 is to generate minimal debug info for a traceback in the absence of
9474 @option{-g} unless explicitly overridden with @option{-g0}. This
9475 behavior is controlled by @code{TARGET_OPTION_OPTIMIZATION} and
9476 @code{TARGET_OPTION_OVERRIDE}.
9479 @node Floating Point
9480 @section Cross Compilation and Floating Point
9481 @cindex cross compilation and floating point
9482 @cindex floating point and cross compilation
9484 While all modern machines use twos-complement representation for integers,
9485 there are a variety of representations for floating point numbers. This
9486 means that in a cross-compiler the representation of floating point numbers
9487 in the compiled program may be different from that used in the machine
9488 doing the compilation.
9490 Because different representation systems may offer different amounts of
9491 range and precision, all floating point constants must be represented in
9492 the target machine's format. Therefore, the cross compiler cannot
9493 safely use the host machine's floating point arithmetic; it must emulate
9494 the target's arithmetic. To ensure consistency, GCC always uses
9495 emulation to work with floating point values, even when the host and
9496 target floating point formats are identical.
9498 The following macros are provided by @file{real.h} for the compiler to
9499 use. All parts of the compiler which generate or optimize
9500 floating-point calculations must use these macros. They may evaluate
9501 their operands more than once, so operands must not have side effects.
9503 @defmac REAL_VALUE_TYPE
9504 The C data type to be used to hold a floating point value in the target
9505 machine's format. Typically this is a @code{struct} containing an
9506 array of @code{HOST_WIDE_INT}, but all code should treat it as an opaque
9510 @deftypefn Macro int REAL_VALUES_EQUAL (REAL_VALUE_TYPE @var{x}, REAL_VALUE_TYPE @var{y})
9511 Compares for equality the two values, @var{x} and @var{y}. If the target
9512 floating point format supports negative zeroes and/or NaNs,
9513 @samp{REAL_VALUES_EQUAL (-0.0, 0.0)} is true, and
9514 @samp{REAL_VALUES_EQUAL (NaN, NaN)} is false.
9517 @deftypefn Macro int REAL_VALUES_LESS (REAL_VALUE_TYPE @var{x}, REAL_VALUE_TYPE @var{y})
9518 Tests whether @var{x} is less than @var{y}.
9521 @deftypefn Macro HOST_WIDE_INT REAL_VALUE_FIX (REAL_VALUE_TYPE @var{x})
9522 Truncates @var{x} to a signed integer, rounding toward zero.
9525 @deftypefn Macro {unsigned HOST_WIDE_INT} REAL_VALUE_UNSIGNED_FIX (REAL_VALUE_TYPE @var{x})
9526 Truncates @var{x} to an unsigned integer, rounding toward zero. If
9527 @var{x} is negative, returns zero.
9530 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_ATOF (const char *@var{string}, enum machine_mode @var{mode})
9531 Converts @var{string} into a floating point number in the target machine's
9532 representation for mode @var{mode}. This routine can handle both
9533 decimal and hexadecimal floating point constants, using the syntax
9534 defined by the C language for both.
9537 @deftypefn Macro int REAL_VALUE_NEGATIVE (REAL_VALUE_TYPE @var{x})
9538 Returns 1 if @var{x} is negative (including negative zero), 0 otherwise.
9541 @deftypefn Macro int REAL_VALUE_ISINF (REAL_VALUE_TYPE @var{x})
9542 Determines whether @var{x} represents infinity (positive or negative).
9545 @deftypefn Macro int REAL_VALUE_ISNAN (REAL_VALUE_TYPE @var{x})
9546 Determines whether @var{x} represents a ``NaN'' (not-a-number).
9549 @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})
9550 Calculates an arithmetic operation on the two floating point values
9551 @var{x} and @var{y}, storing the result in @var{output} (which must be a
9554 The operation to be performed is specified by @var{code}. Only the
9555 following codes are supported: @code{PLUS_EXPR}, @code{MINUS_EXPR},
9556 @code{MULT_EXPR}, @code{RDIV_EXPR}, @code{MAX_EXPR}, @code{MIN_EXPR}.
9558 If @code{REAL_ARITHMETIC} is asked to evaluate division by zero and the
9559 target's floating point format cannot represent infinity, it will call
9560 @code{abort}. Callers should check for this situation first, using
9561 @code{MODE_HAS_INFINITIES}. @xref{Storage Layout}.
9564 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_NEGATE (REAL_VALUE_TYPE @var{x})
9565 Returns the negative of the floating point value @var{x}.
9568 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_ABS (REAL_VALUE_TYPE @var{x})
9569 Returns the absolute value of @var{x}.
9572 @deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_TRUNCATE (REAL_VALUE_TYPE @var{mode}, enum machine_mode @var{x})
9573 Truncates the floating point value @var{x} to fit in @var{mode}. The
9574 return value is still a full-size @code{REAL_VALUE_TYPE}, but it has an
9575 appropriate bit pattern to be output as a floating constant whose
9576 precision accords with mode @var{mode}.
9579 @deftypefn Macro void REAL_VALUE_TO_INT (HOST_WIDE_INT @var{low}, HOST_WIDE_INT @var{high}, REAL_VALUE_TYPE @var{x})
9580 Converts a floating point value @var{x} into a double-precision integer
9581 which is then stored into @var{low} and @var{high}. If the value is not
9582 integral, it is truncated.
9585 @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})
9586 Converts a double-precision integer found in @var{low} and @var{high},
9587 into a floating point value which is then stored into @var{x}. The
9588 value is truncated to fit in mode @var{mode}.
9591 @node Mode Switching
9592 @section Mode Switching Instructions
9593 @cindex mode switching
9594 The following macros control mode switching optimizations:
9596 @defmac OPTIMIZE_MODE_SWITCHING (@var{entity})
9597 Define this macro if the port needs extra instructions inserted for mode
9598 switching in an optimizing compilation.
9600 For an example, the SH4 can perform both single and double precision
9601 floating point operations, but to perform a single precision operation,
9602 the FPSCR PR bit has to be cleared, while for a double precision
9603 operation, this bit has to be set. Changing the PR bit requires a general
9604 purpose register as a scratch register, hence these FPSCR sets have to
9605 be inserted before reload, i.e.@: you can't put this into instruction emitting
9606 or @code{TARGET_MACHINE_DEPENDENT_REORG}.
9608 You can have multiple entities that are mode-switched, and select at run time
9609 which entities actually need it. @code{OPTIMIZE_MODE_SWITCHING} should
9610 return nonzero for any @var{entity} that needs mode-switching.
9611 If you define this macro, you also have to define
9612 @code{NUM_MODES_FOR_MODE_SWITCHING}, @code{MODE_NEEDED},
9613 @code{MODE_PRIORITY_TO_MODE} and @code{EMIT_MODE_SET}.
9614 @code{MODE_AFTER}, @code{MODE_ENTRY}, and @code{MODE_EXIT}
9618 @defmac NUM_MODES_FOR_MODE_SWITCHING
9619 If you define @code{OPTIMIZE_MODE_SWITCHING}, you have to define this as
9620 initializer for an array of integers. Each initializer element
9621 N refers to an entity that needs mode switching, and specifies the number
9622 of different modes that might need to be set for this entity.
9623 The position of the initializer in the initializer---starting counting at
9624 zero---determines the integer that is used to refer to the mode-switched
9626 In macros that take mode arguments / yield a mode result, modes are
9627 represented as numbers 0 @dots{} N @minus{} 1. N is used to specify that no mode
9628 switch is needed / supplied.
9631 @defmac MODE_NEEDED (@var{entity}, @var{insn})
9632 @var{entity} is an integer specifying a mode-switched entity. If
9633 @code{OPTIMIZE_MODE_SWITCHING} is defined, you must define this macro to
9634 return an integer value not larger than the corresponding element in
9635 @code{NUM_MODES_FOR_MODE_SWITCHING}, to denote the mode that @var{entity} must
9636 be switched into prior to the execution of @var{insn}.
9639 @defmac MODE_AFTER (@var{mode}, @var{insn})
9640 If this macro is defined, it is evaluated for every @var{insn} during
9641 mode switching. It determines the mode that an insn results in (if
9642 different from the incoming mode).
9645 @defmac MODE_ENTRY (@var{entity})
9646 If this macro is defined, it is evaluated for every @var{entity} that needs
9647 mode switching. It should evaluate to an integer, which is a mode that
9648 @var{entity} is assumed to be switched to at function entry. If @code{MODE_ENTRY}
9649 is defined then @code{MODE_EXIT} must be defined.
9652 @defmac MODE_EXIT (@var{entity})
9653 If this macro is defined, it is evaluated for every @var{entity} that needs
9654 mode switching. It should evaluate to an integer, which is a mode that
9655 @var{entity} is assumed to be switched to at function exit. If @code{MODE_EXIT}
9656 is defined then @code{MODE_ENTRY} must be defined.
9659 @defmac MODE_PRIORITY_TO_MODE (@var{entity}, @var{n})
9660 This macro specifies the order in which modes for @var{entity} are processed.
9661 0 is the highest priority, @code{NUM_MODES_FOR_MODE_SWITCHING[@var{entity}] - 1} the
9662 lowest. The value of the macro should be an integer designating a mode
9663 for @var{entity}. For any fixed @var{entity}, @code{mode_priority_to_mode}
9664 (@var{entity}, @var{n}) shall be a bijection in 0 @dots{}
9665 @code{num_modes_for_mode_switching[@var{entity}] - 1}.
9668 @defmac EMIT_MODE_SET (@var{entity}, @var{mode}, @var{hard_regs_live})
9669 Generate one or more insns to set @var{entity} to @var{mode}.
9670 @var{hard_reg_live} is the set of hard registers live at the point where
9671 the insn(s) are to be inserted.
9674 @node Target Attributes
9675 @section Defining target-specific uses of @code{__attribute__}
9676 @cindex target attributes
9677 @cindex machine attributes
9678 @cindex attributes, target-specific
9680 Target-specific attributes may be defined for functions, data and types.
9681 These are described using the following target hooks; they also need to
9682 be documented in @file{extend.texi}.
9684 @hook TARGET_ATTRIBUTE_TABLE
9685 If defined, this target hook points to an array of @samp{struct
9686 attribute_spec} (defined in @file{tree.h}) specifying the machine
9687 specific attributes for this target and some of the restrictions on the
9688 entities to which these attributes are applied and the arguments they
9692 @hook TARGET_ATTRIBUTE_TAKES_IDENTIFIER_P
9693 If defined, this target hook is a function which returns true if the
9694 machine-specific attribute named @var{name} expects an identifier
9695 given as its first argument to be passed on as a plain identifier, not
9696 subjected to name lookup. If this is not defined, the default is
9697 false for all machine-specific attributes.
9700 @hook TARGET_COMP_TYPE_ATTRIBUTES
9701 If defined, this target hook is a function which returns zero if the attributes on
9702 @var{type1} and @var{type2} are incompatible, one if they are compatible,
9703 and two if they are nearly compatible (which causes a warning to be
9704 generated). If this is not defined, machine-specific attributes are
9705 supposed always to be compatible.
9708 @hook TARGET_SET_DEFAULT_TYPE_ATTRIBUTES
9709 If defined, this target hook is a function which assigns default attributes to
9710 the newly defined @var{type}.
9713 @hook TARGET_MERGE_TYPE_ATTRIBUTES
9714 Define this target hook if the merging of type attributes needs special
9715 handling. If defined, the result is a list of the combined
9716 @code{TYPE_ATTRIBUTES} of @var{type1} and @var{type2}. It is assumed
9717 that @code{comptypes} has already been called and returned 1. This
9718 function may call @code{merge_attributes} to handle machine-independent
9722 @hook TARGET_MERGE_DECL_ATTRIBUTES
9723 Define this target hook if the merging of decl attributes needs special
9724 handling. If defined, the result is a list of the combined
9725 @code{DECL_ATTRIBUTES} of @var{olddecl} and @var{newdecl}.
9726 @var{newdecl} is a duplicate declaration of @var{olddecl}. Examples of
9727 when this is needed are when one attribute overrides another, or when an
9728 attribute is nullified by a subsequent definition. This function may
9729 call @code{merge_attributes} to handle machine-independent merging.
9731 @findex TARGET_DLLIMPORT_DECL_ATTRIBUTES
9732 If the only target-specific handling you require is @samp{dllimport}
9733 for Microsoft Windows targets, you should define the macro
9734 @code{TARGET_DLLIMPORT_DECL_ATTRIBUTES} to @code{1}. The compiler
9735 will then define a function called
9736 @code{merge_dllimport_decl_attributes} which can then be defined as
9737 the expansion of @code{TARGET_MERGE_DECL_ATTRIBUTES}. You can also
9738 add @code{handle_dll_attribute} in the attribute table for your port
9739 to perform initial processing of the @samp{dllimport} and
9740 @samp{dllexport} attributes. This is done in @file{i386/cygwin.h} and
9741 @file{i386/i386.c}, for example.
9744 @hook TARGET_VALID_DLLIMPORT_ATTRIBUTE_P
9746 @defmac TARGET_DECLSPEC
9747 Define this macro to a nonzero value if you want to treat
9748 @code{__declspec(X)} as equivalent to @code{__attribute((X))}. By
9749 default, this behavior is enabled only for targets that define
9750 @code{TARGET_DLLIMPORT_DECL_ATTRIBUTES}. The current implementation
9751 of @code{__declspec} is via a built-in macro, but you should not rely
9752 on this implementation detail.
9755 @hook TARGET_INSERT_ATTRIBUTES
9756 Define this target hook if you want to be able to add attributes to a decl
9757 when it is being created. This is normally useful for back ends which
9758 wish to implement a pragma by using the attributes which correspond to
9759 the pragma's effect. The @var{node} argument is the decl which is being
9760 created. The @var{attr_ptr} argument is a pointer to the attribute list
9761 for this decl. The list itself should not be modified, since it may be
9762 shared with other decls, but attributes may be chained on the head of
9763 the list and @code{*@var{attr_ptr}} modified to point to the new
9764 attributes, or a copy of the list may be made if further changes are
9768 @hook TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P
9770 This target hook returns @code{true} if it is ok to inline @var{fndecl}
9771 into the current function, despite its having target-specific
9772 attributes, @code{false} otherwise. By default, if a function has a
9773 target specific attribute attached to it, it will not be inlined.
9776 @hook TARGET_OPTION_VALID_ATTRIBUTE_P
9777 This hook is called to parse the @code{attribute(option("..."))}, and
9778 it allows the function to set different target machine compile time
9779 options for the current function that might be different than the
9780 options specified on the command line. The hook should return
9781 @code{true} if the options are valid.
9783 The hook should set the @var{DECL_FUNCTION_SPECIFIC_TARGET} field in
9784 the function declaration to hold a pointer to a target specific
9785 @var{struct cl_target_option} structure.
9788 @hook TARGET_OPTION_SAVE
9789 This hook is called to save any additional target specific information
9790 in the @var{struct cl_target_option} structure for function specific
9792 @xref{Option file format}.
9795 @hook TARGET_OPTION_RESTORE
9796 This hook is called to restore any additional target specific
9797 information in the @var{struct cl_target_option} structure for
9798 function specific options.
9801 @hook TARGET_OPTION_PRINT
9802 This hook is called to print any additional target specific
9803 information in the @var{struct cl_target_option} structure for
9804 function specific options.
9807 @hook TARGET_OPTION_PRAGMA_PARSE
9808 This target hook parses the options for @code{#pragma GCC option} to
9809 set the machine specific options for functions that occur later in the
9810 input stream. The options should be the same as handled by the
9811 @code{TARGET_OPTION_VALID_ATTRIBUTE_P} hook.
9814 @hook TARGET_OPTION_OVERRIDE
9815 Sometimes certain combinations of command options do not make sense on
9816 a particular target machine. You can override the hook
9817 @code{TARGET_OPTION_OVERRIDE} to take account of this. This hooks is called
9818 once just after all the command options have been parsed.
9820 Don't use this hook to turn on various extra optimizations for
9821 @option{-O}. That is what @code{TARGET_OPTION_OPTIMIZATION} is for.
9823 If you need to do something whenever the optimization level is
9824 changed via the optimize attribute or pragma, see
9825 @code{TARGET_OVERRIDE_OPTIONS_AFTER_CHANGE}
9828 @hook TARGET_CAN_INLINE_P
9829 This target hook returns @code{false} if the @var{caller} function
9830 cannot inline @var{callee}, based on target specific information. By
9831 default, inlining is not allowed if the callee function has function
9832 specific target options and the caller does not use the same options.
9836 @section Emulating TLS
9837 @cindex Emulated TLS
9839 For targets whose psABI does not provide Thread Local Storage via
9840 specific relocations and instruction sequences, an emulation layer is
9841 used. A set of target hooks allows this emulation layer to be
9842 configured for the requirements of a particular target. For instance
9843 the psABI may in fact specify TLS support in terms of an emulation
9846 The emulation layer works by creating a control object for every TLS
9847 object. To access the TLS object, a lookup function is provided
9848 which, when given the address of the control object, will return the
9849 address of the current thread's instance of the TLS object.
9851 @hook TARGET_EMUTLS_GET_ADDRESS
9852 Contains the name of the helper function that uses a TLS control
9853 object to locate a TLS instance. The default causes libgcc's
9854 emulated TLS helper function to be used.
9857 @hook TARGET_EMUTLS_REGISTER_COMMON
9858 Contains the name of the helper function that should be used at
9859 program startup to register TLS objects that are implicitly
9860 initialized to zero. If this is @code{NULL}, all TLS objects will
9861 have explicit initializers. The default causes libgcc's emulated TLS
9862 registration function to be used.
9865 @hook TARGET_EMUTLS_VAR_SECTION
9866 Contains the name of the section in which TLS control variables should
9867 be placed. The default of @code{NULL} allows these to be placed in
9871 @hook TARGET_EMUTLS_TMPL_SECTION
9872 Contains the name of the section in which TLS initializers should be
9873 placed. The default of @code{NULL} allows these to be placed in any
9877 @hook TARGET_EMUTLS_VAR_PREFIX
9878 Contains the prefix to be prepended to TLS control variable names.
9879 The default of @code{NULL} uses a target-specific prefix.
9882 @hook TARGET_EMUTLS_TMPL_PREFIX
9883 Contains the prefix to be prepended to TLS initializer objects. The
9884 default of @code{NULL} uses a target-specific prefix.
9887 @hook TARGET_EMUTLS_VAR_FIELDS
9888 Specifies a function that generates the FIELD_DECLs for a TLS control
9889 object type. @var{type} is the RECORD_TYPE the fields are for and
9890 @var{name} should be filled with the structure tag, if the default of
9891 @code{__emutls_object} is unsuitable. The default creates a type suitable
9892 for libgcc's emulated TLS function.
9895 @hook TARGET_EMUTLS_VAR_INIT
9896 Specifies a function that generates the CONSTRUCTOR to initialize a
9897 TLS control object. @var{var} is the TLS control object, @var{decl}
9898 is the TLS object and @var{tmpl_addr} is the address of the
9899 initializer. The default initializes libgcc's emulated TLS control object.
9902 @hook TARGET_EMUTLS_VAR_ALIGN_FIXED
9903 Specifies whether the alignment of TLS control variable objects is
9904 fixed and should not be increased as some backends may do to optimize
9905 single objects. The default is false.
9908 @hook TARGET_EMUTLS_DEBUG_FORM_TLS_ADDRESS
9909 Specifies whether a DWARF @code{DW_OP_form_tls_address} location descriptor
9910 may be used to describe emulated TLS control objects.
9913 @node MIPS Coprocessors
9914 @section Defining coprocessor specifics for MIPS targets.
9915 @cindex MIPS coprocessor-definition macros
9917 The MIPS specification allows MIPS implementations to have as many as 4
9918 coprocessors, each with as many as 32 private registers. GCC supports
9919 accessing these registers and transferring values between the registers
9920 and memory using asm-ized variables. For example:
9923 register unsigned int cp0count asm ("c0r1");
9929 (``c0r1'' is the default name of register 1 in coprocessor 0; alternate
9930 names may be added as described below, or the default names may be
9931 overridden entirely in @code{SUBTARGET_CONDITIONAL_REGISTER_USAGE}.)
9933 Coprocessor registers are assumed to be epilogue-used; sets to them will
9934 be preserved even if it does not appear that the register is used again
9935 later in the function.
9937 Another note: according to the MIPS spec, coprocessor 1 (if present) is
9938 the FPU@. One accesses COP1 registers through standard mips
9939 floating-point support; they are not included in this mechanism.
9941 There is one macro used in defining the MIPS coprocessor interface which
9942 you may want to override in subtargets; it is described below.
9944 @defmac ALL_COP_ADDITIONAL_REGISTER_NAMES
9945 A comma-separated list (with leading comma) of pairs describing the
9946 alternate names of coprocessor registers. The format of each entry should be
9948 @{ @var{alternatename}, @var{register_number}@}
9954 @section Parameters for Precompiled Header Validity Checking
9955 @cindex parameters, precompiled headers
9957 @hook TARGET_GET_PCH_VALIDITY
9958 This hook returns a pointer to the data needed by
9959 @code{TARGET_PCH_VALID_P} and sets
9960 @samp{*@var{sz}} to the size of the data in bytes.
9963 @hook TARGET_PCH_VALID_P
9964 This hook checks whether the options used to create a PCH file are
9965 compatible with the current settings. It returns @code{NULL}
9966 if so and a suitable error message if not. Error messages will
9967 be presented to the user and must be localized using @samp{_(@var{msg})}.
9969 @var{data} is the data that was returned by @code{TARGET_GET_PCH_VALIDITY}
9970 when the PCH file was created and @var{sz} is the size of that data in bytes.
9971 It's safe to assume that the data was created by the same version of the
9972 compiler, so no format checking is needed.
9974 The default definition of @code{default_pch_valid_p} should be
9975 suitable for most targets.
9978 @hook TARGET_CHECK_PCH_TARGET_FLAGS
9979 If this hook is nonnull, the default implementation of
9980 @code{TARGET_PCH_VALID_P} will use it to check for compatible values
9981 of @code{target_flags}. @var{pch_flags} specifies the value that
9982 @code{target_flags} had when the PCH file was created. The return
9983 value is the same as for @code{TARGET_PCH_VALID_P}.
9986 @hook TARGET_PREPARE_PCH_SAVE
9989 @section C++ ABI parameters
9990 @cindex parameters, c++ abi
9992 @hook TARGET_CXX_GUARD_TYPE
9993 Define this hook to override the integer type used for guard variables.
9994 These are used to implement one-time construction of static objects. The
9995 default is long_long_integer_type_node.
9998 @hook TARGET_CXX_GUARD_MASK_BIT
9999 This hook determines how guard variables are used. It should return
10000 @code{false} (the default) if the first byte should be used. A return value of
10001 @code{true} indicates that only the least significant bit should be used.
10004 @hook TARGET_CXX_GET_COOKIE_SIZE
10005 This hook returns the size of the cookie to use when allocating an array
10006 whose elements have the indicated @var{type}. Assumes that it is already
10007 known that a cookie is needed. The default is
10008 @code{max(sizeof (size_t), alignof(type))}, as defined in section 2.7 of the
10009 IA64/Generic C++ ABI@.
10012 @hook TARGET_CXX_COOKIE_HAS_SIZE
10013 This hook should return @code{true} if the element size should be stored in
10014 array cookies. The default is to return @code{false}.
10017 @hook TARGET_CXX_IMPORT_EXPORT_CLASS
10018 If defined by a backend this hook allows the decision made to export
10019 class @var{type} to be overruled. Upon entry @var{import_export}
10020 will contain 1 if the class is going to be exported, @minus{}1 if it is going
10021 to be imported and 0 otherwise. This function should return the
10022 modified value and perform any other actions necessary to support the
10023 backend's targeted operating system.
10026 @hook TARGET_CXX_CDTOR_RETURNS_THIS
10027 This hook should return @code{true} if constructors and destructors return
10028 the address of the object created/destroyed. The default is to return
10032 @hook TARGET_CXX_KEY_METHOD_MAY_BE_INLINE
10033 This hook returns true if the key method for a class (i.e., the method
10034 which, if defined in the current translation unit, causes the virtual
10035 table to be emitted) may be an inline function. Under the standard
10036 Itanium C++ ABI the key method may be an inline function so long as
10037 the function is not declared inline in the class definition. Under
10038 some variants of the ABI, an inline function can never be the key
10039 method. The default is to return @code{true}.
10042 @hook TARGET_CXX_DETERMINE_CLASS_DATA_VISIBILITY
10044 @hook TARGET_CXX_CLASS_DATA_ALWAYS_COMDAT
10045 This hook returns true (the default) if virtual tables and other
10046 similar implicit class data objects are always COMDAT if they have
10047 external linkage. If this hook returns false, then class data for
10048 classes whose virtual table will be emitted in only one translation
10049 unit will not be COMDAT.
10052 @hook TARGET_CXX_LIBRARY_RTTI_COMDAT
10053 This hook returns true (the default) if the RTTI information for
10054 the basic types which is defined in the C++ runtime should always
10055 be COMDAT, false if it should not be COMDAT.
10058 @hook TARGET_CXX_USE_AEABI_ATEXIT
10059 This hook returns true if @code{__aeabi_atexit} (as defined by the ARM EABI)
10060 should be used to register static destructors when @option{-fuse-cxa-atexit}
10061 is in effect. The default is to return false to use @code{__cxa_atexit}.
10064 @hook TARGET_CXX_USE_ATEXIT_FOR_CXA_ATEXIT
10065 This hook returns true if the target @code{atexit} function can be used
10066 in the same manner as @code{__cxa_atexit} to register C++ static
10067 destructors. This requires that @code{atexit}-registered functions in
10068 shared libraries are run in the correct order when the libraries are
10069 unloaded. The default is to return false.
10072 @hook TARGET_CXX_ADJUST_CLASS_AT_DEFINITION
10074 @node Named Address Spaces
10075 @section Adding support for named address spaces
10076 @cindex named address spaces
10078 The draft technical report of the ISO/IEC JTC1 S22 WG14 N1275
10079 standards committee, @cite{Programming Languages - C - Extensions to
10080 support embedded processors}, specifies a syntax for embedded
10081 processors to specify alternate address spaces. You can configure a
10082 GCC port to support section 5.1 of the draft report to add support for
10083 address spaces other than the default address space. These address
10084 spaces are new keywords that are similar to the @code{volatile} and
10085 @code{const} type attributes.
10087 Pointers to named address spaces can have a different size than
10088 pointers to the generic address space.
10090 For example, the SPU port uses the @code{__ea} address space to refer
10091 to memory in the host processor, rather than memory local to the SPU
10092 processor. Access to memory in the @code{__ea} address space involves
10093 issuing DMA operations to move data between the host processor and the
10094 local processor memory address space. Pointers in the @code{__ea}
10095 address space are either 32 bits or 64 bits based on the
10096 @option{-mea32} or @option{-mea64} switches (native SPU pointers are
10099 Internally, address spaces are represented as a small integer in the
10100 range 0 to 15 with address space 0 being reserved for the generic
10103 To register a named address space qualifier keyword with the C front end,
10104 the target may call the @code{c_register_addr_space} routine. For example,
10105 the SPU port uses the following to declare @code{__ea} as the keyword for
10106 named address space #1:
10108 #define ADDR_SPACE_EA 1
10109 c_register_addr_space ("__ea", ADDR_SPACE_EA);
10112 @hook TARGET_ADDR_SPACE_POINTER_MODE
10113 Define this to return the machine mode to use for pointers to
10114 @var{address_space} if the target supports named address spaces.
10115 The default version of this hook returns @code{ptr_mode} for the
10116 generic address space only.
10119 @hook TARGET_ADDR_SPACE_ADDRESS_MODE
10120 Define this to return the machine mode to use for addresses in
10121 @var{address_space} if the target supports named address spaces.
10122 The default version of this hook returns @code{Pmode} for the
10123 generic address space only.
10126 @hook TARGET_ADDR_SPACE_VALID_POINTER_MODE
10127 Define this to return nonzero if the port can handle pointers
10128 with machine mode @var{mode} to address space @var{as}. This target
10129 hook is the same as the @code{TARGET_VALID_POINTER_MODE} target hook,
10130 except that it includes explicit named address space support. The default
10131 version of this hook returns true for the modes returned by either the
10132 @code{TARGET_ADDR_SPACE_POINTER_MODE} or @code{TARGET_ADDR_SPACE_ADDRESS_MODE}
10133 target hooks for the given address space.
10136 @hook TARGET_ADDR_SPACE_LEGITIMATE_ADDRESS_P
10137 Define this to return true if @var{exp} is a valid address for mode
10138 @var{mode} in the named address space @var{as}. The @var{strict}
10139 parameter says whether strict addressing is in effect after reload has
10140 finished. This target hook is the same as the
10141 @code{TARGET_LEGITIMATE_ADDRESS_P} target hook, except that it includes
10142 explicit named address space support.
10145 @hook TARGET_ADDR_SPACE_LEGITIMIZE_ADDRESS
10146 Define this to modify an invalid address @var{x} to be a valid address
10147 with mode @var{mode} in the named address space @var{as}. This target
10148 hook is the same as the @code{TARGET_LEGITIMIZE_ADDRESS} target hook,
10149 except that it includes explicit named address space support.
10152 @hook TARGET_ADDR_SPACE_SUBSET_P
10153 Define this to return whether the @var{subset} named address space is
10154 contained within the @var{superset} named address space. Pointers to
10155 a named address space that is a subset of another named address space
10156 will be converted automatically without a cast if used together in
10157 arithmetic operations. Pointers to a superset address space can be
10158 converted to pointers to a subset address space via explicit casts.
10161 @hook TARGET_ADDR_SPACE_CONVERT
10162 Define this to convert the pointer expression represented by the RTL
10163 @var{op} with type @var{from_type} that points to a named address
10164 space to a new pointer expression with type @var{to_type} that points
10165 to a different named address space. When this hook it called, it is
10166 guaranteed that one of the two address spaces is a subset of the other,
10167 as determined by the @code{TARGET_ADDR_SPACE_SUBSET_P} target hook.
10171 @section Miscellaneous Parameters
10172 @cindex parameters, miscellaneous
10174 @c prevent bad page break with this line
10175 Here are several miscellaneous parameters.
10177 @defmac HAS_LONG_COND_BRANCH
10178 Define this boolean macro to indicate whether or not your architecture
10179 has conditional branches that can span all of memory. It is used in
10180 conjunction with an optimization that partitions hot and cold basic
10181 blocks into separate sections of the executable. If this macro is
10182 set to false, gcc will convert any conditional branches that attempt
10183 to cross between sections into unconditional branches or indirect jumps.
10186 @defmac HAS_LONG_UNCOND_BRANCH
10187 Define this boolean macro to indicate whether or not your architecture
10188 has unconditional branches that can span all of memory. It is used in
10189 conjunction with an optimization that partitions hot and cold basic
10190 blocks into separate sections of the executable. If this macro is
10191 set to false, gcc will convert any unconditional branches that attempt
10192 to cross between sections into indirect jumps.
10195 @defmac CASE_VECTOR_MODE
10196 An alias for a machine mode name. This is the machine mode that
10197 elements of a jump-table should have.
10200 @defmac CASE_VECTOR_SHORTEN_MODE (@var{min_offset}, @var{max_offset}, @var{body})
10201 Optional: return the preferred mode for an @code{addr_diff_vec}
10202 when the minimum and maximum offset are known. If you define this,
10203 it enables extra code in branch shortening to deal with @code{addr_diff_vec}.
10204 To make this work, you also have to define @code{INSN_ALIGN} and
10205 make the alignment for @code{addr_diff_vec} explicit.
10206 The @var{body} argument is provided so that the offset_unsigned and scale
10207 flags can be updated.
10210 @defmac CASE_VECTOR_PC_RELATIVE
10211 Define this macro to be a C expression to indicate when jump-tables
10212 should contain relative addresses. You need not define this macro if
10213 jump-tables never contain relative addresses, or jump-tables should
10214 contain relative addresses only when @option{-fPIC} or @option{-fPIC}
10218 @hook TARGET_CASE_VALUES_THRESHOLD
10219 This function return the smallest number of different values for which it
10220 is best to use a jump-table instead of a tree of conditional branches.
10221 The default is four for machines with a @code{casesi} instruction and
10222 five otherwise. This is best for most machines.
10225 @defmac CASE_USE_BIT_TESTS
10226 Define this macro to be a C expression to indicate whether C switch
10227 statements may be implemented by a sequence of bit tests. This is
10228 advantageous on processors that can efficiently implement left shift
10229 of 1 by the number of bits held in a register, but inappropriate on
10230 targets that would require a loop. By default, this macro returns
10231 @code{true} if the target defines an @code{ashlsi3} pattern, and
10232 @code{false} otherwise.
10235 @defmac WORD_REGISTER_OPERATIONS
10236 Define this macro if operations between registers with integral mode
10237 smaller than a word are always performed on the entire register.
10238 Most RISC machines have this property and most CISC machines do not.
10241 @defmac LOAD_EXTEND_OP (@var{mem_mode})
10242 Define this macro to be a C expression indicating when insns that read
10243 memory in @var{mem_mode}, an integral mode narrower than a word, set the
10244 bits outside of @var{mem_mode} to be either the sign-extension or the
10245 zero-extension of the data read. Return @code{SIGN_EXTEND} for values
10246 of @var{mem_mode} for which the
10247 insn sign-extends, @code{ZERO_EXTEND} for which it zero-extends, and
10248 @code{UNKNOWN} for other modes.
10250 This macro is not called with @var{mem_mode} non-integral or with a width
10251 greater than or equal to @code{BITS_PER_WORD}, so you may return any
10252 value in this case. Do not define this macro if it would always return
10253 @code{UNKNOWN}. On machines where this macro is defined, you will normally
10254 define it as the constant @code{SIGN_EXTEND} or @code{ZERO_EXTEND}.
10256 You may return a non-@code{UNKNOWN} value even if for some hard registers
10257 the sign extension is not performed, if for the @code{REGNO_REG_CLASS}
10258 of these hard registers @code{CANNOT_CHANGE_MODE_CLASS} returns nonzero
10259 when the @var{from} mode is @var{mem_mode} and the @var{to} mode is any
10260 integral mode larger than this but not larger than @code{word_mode}.
10262 You must return @code{UNKNOWN} if for some hard registers that allow this
10263 mode, @code{CANNOT_CHANGE_MODE_CLASS} says that they cannot change to
10264 @code{word_mode}, but that they can change to another integral mode that
10265 is larger then @var{mem_mode} but still smaller than @code{word_mode}.
10268 @defmac SHORT_IMMEDIATES_SIGN_EXTEND
10269 Define this macro if loading short immediate values into registers sign
10273 @defmac FIXUNS_TRUNC_LIKE_FIX_TRUNC
10274 Define this macro if the same instructions that convert a floating
10275 point number to a signed fixed point number also convert validly to an
10279 @hook TARGET_MIN_DIVISIONS_FOR_RECIP_MUL
10280 When @option{-ffast-math} is in effect, GCC tries to optimize
10281 divisions by the same divisor, by turning them into multiplications by
10282 the reciprocal. This target hook specifies the minimum number of divisions
10283 that should be there for GCC to perform the optimization for a variable
10284 of mode @var{mode}. The default implementation returns 3 if the machine
10285 has an instruction for the division, and 2 if it does not.
10289 The maximum number of bytes that a single instruction can move quickly
10290 between memory and registers or between two memory locations.
10293 @defmac MAX_MOVE_MAX
10294 The maximum number of bytes that a single instruction can move quickly
10295 between memory and registers or between two memory locations. If this
10296 is undefined, the default is @code{MOVE_MAX}. Otherwise, it is the
10297 constant value that is the largest value that @code{MOVE_MAX} can have
10301 @defmac SHIFT_COUNT_TRUNCATED
10302 A C expression that is nonzero if on this machine the number of bits
10303 actually used for the count of a shift operation is equal to the number
10304 of bits needed to represent the size of the object being shifted. When
10305 this macro is nonzero, the compiler will assume that it is safe to omit
10306 a sign-extend, zero-extend, and certain bitwise `and' instructions that
10307 truncates the count of a shift operation. On machines that have
10308 instructions that act on bit-fields at variable positions, which may
10309 include `bit test' instructions, a nonzero @code{SHIFT_COUNT_TRUNCATED}
10310 also enables deletion of truncations of the values that serve as
10311 arguments to bit-field instructions.
10313 If both types of instructions truncate the count (for shifts) and
10314 position (for bit-field operations), or if no variable-position bit-field
10315 instructions exist, you should define this macro.
10317 However, on some machines, such as the 80386 and the 680x0, truncation
10318 only applies to shift operations and not the (real or pretended)
10319 bit-field operations. Define @code{SHIFT_COUNT_TRUNCATED} to be zero on
10320 such machines. Instead, add patterns to the @file{md} file that include
10321 the implied truncation of the shift instructions.
10323 You need not define this macro if it would always have the value of zero.
10326 @anchor{TARGET_SHIFT_TRUNCATION_MASK}
10327 @hook TARGET_SHIFT_TRUNCATION_MASK
10328 This function describes how the standard shift patterns for @var{mode}
10329 deal with shifts by negative amounts or by more than the width of the mode.
10330 @xref{shift patterns}.
10332 On many machines, the shift patterns will apply a mask @var{m} to the
10333 shift count, meaning that a fixed-width shift of @var{x} by @var{y} is
10334 equivalent to an arbitrary-width shift of @var{x} by @var{y & m}. If
10335 this is true for mode @var{mode}, the function should return @var{m},
10336 otherwise it should return 0. A return value of 0 indicates that no
10337 particular behavior is guaranteed.
10339 Note that, unlike @code{SHIFT_COUNT_TRUNCATED}, this function does
10340 @emph{not} apply to general shift rtxes; it applies only to instructions
10341 that are generated by the named shift patterns.
10343 The default implementation of this function returns
10344 @code{GET_MODE_BITSIZE (@var{mode}) - 1} if @code{SHIFT_COUNT_TRUNCATED}
10345 and 0 otherwise. This definition is always safe, but if
10346 @code{SHIFT_COUNT_TRUNCATED} is false, and some shift patterns
10347 nevertheless truncate the shift count, you may get better code
10351 @defmac TRULY_NOOP_TRUNCATION (@var{outprec}, @var{inprec})
10352 A C expression which is nonzero if on this machine it is safe to
10353 ``convert'' an integer of @var{inprec} bits to one of @var{outprec}
10354 bits (where @var{outprec} is smaller than @var{inprec}) by merely
10355 operating on it as if it had only @var{outprec} bits.
10357 On many machines, this expression can be 1.
10359 @c rearranged this, removed the phrase "it is reported that". this was
10360 @c to fix an overfull hbox. --mew 10feb93
10361 When @code{TRULY_NOOP_TRUNCATION} returns 1 for a pair of sizes for
10362 modes for which @code{MODES_TIEABLE_P} is 0, suboptimal code can result.
10363 If this is the case, making @code{TRULY_NOOP_TRUNCATION} return 0 in
10364 such cases may improve things.
10367 @hook TARGET_MODE_REP_EXTENDED
10368 The representation of an integral mode can be such that the values
10369 are always extended to a wider integral mode. Return
10370 @code{SIGN_EXTEND} if values of @var{mode} are represented in
10371 sign-extended form to @var{rep_mode}. Return @code{UNKNOWN}
10372 otherwise. (Currently, none of the targets use zero-extended
10373 representation this way so unlike @code{LOAD_EXTEND_OP},
10374 @code{TARGET_MODE_REP_EXTENDED} is expected to return either
10375 @code{SIGN_EXTEND} or @code{UNKNOWN}. Also no target extends
10376 @var{mode} to @var{rep_mode} so that @var{rep_mode} is not the next
10377 widest integral mode and currently we take advantage of this fact.)
10379 Similarly to @code{LOAD_EXTEND_OP} you may return a non-@code{UNKNOWN}
10380 value even if the extension is not performed on certain hard registers
10381 as long as for the @code{REGNO_REG_CLASS} of these hard registers
10382 @code{CANNOT_CHANGE_MODE_CLASS} returns nonzero.
10384 Note that @code{TARGET_MODE_REP_EXTENDED} and @code{LOAD_EXTEND_OP}
10385 describe two related properties. If you define
10386 @code{TARGET_MODE_REP_EXTENDED (mode, word_mode)} you probably also want
10387 to define @code{LOAD_EXTEND_OP (mode)} to return the same type of
10390 In order to enforce the representation of @code{mode},
10391 @code{TRULY_NOOP_TRUNCATION} should return false when truncating to
10395 @defmac STORE_FLAG_VALUE
10396 A C expression describing the value returned by a comparison operator
10397 with an integral mode and stored by a store-flag instruction
10398 (@samp{cstore@var{mode}4}) when the condition is true. This description must
10399 apply to @emph{all} the @samp{cstore@var{mode}4} patterns and all the
10400 comparison operators whose results have a @code{MODE_INT} mode.
10402 A value of 1 or @minus{}1 means that the instruction implementing the
10403 comparison operator returns exactly 1 or @minus{}1 when the comparison is true
10404 and 0 when the comparison is false. Otherwise, the value indicates
10405 which bits of the result are guaranteed to be 1 when the comparison is
10406 true. This value is interpreted in the mode of the comparison
10407 operation, which is given by the mode of the first operand in the
10408 @samp{cstore@var{mode}4} pattern. Either the low bit or the sign bit of
10409 @code{STORE_FLAG_VALUE} be on. Presently, only those bits are used by
10412 If @code{STORE_FLAG_VALUE} is neither 1 or @minus{}1, the compiler will
10413 generate code that depends only on the specified bits. It can also
10414 replace comparison operators with equivalent operations if they cause
10415 the required bits to be set, even if the remaining bits are undefined.
10416 For example, on a machine whose comparison operators return an
10417 @code{SImode} value and where @code{STORE_FLAG_VALUE} is defined as
10418 @samp{0x80000000}, saying that just the sign bit is relevant, the
10422 (ne:SI (and:SI @var{x} (const_int @var{power-of-2})) (const_int 0))
10426 can be converted to
10429 (ashift:SI @var{x} (const_int @var{n}))
10433 where @var{n} is the appropriate shift count to move the bit being
10434 tested into the sign bit.
10436 There is no way to describe a machine that always sets the low-order bit
10437 for a true value, but does not guarantee the value of any other bits,
10438 but we do not know of any machine that has such an instruction. If you
10439 are trying to port GCC to such a machine, include an instruction to
10440 perform a logical-and of the result with 1 in the pattern for the
10441 comparison operators and let us know at @email{gcc@@gcc.gnu.org}.
10443 Often, a machine will have multiple instructions that obtain a value
10444 from a comparison (or the condition codes). Here are rules to guide the
10445 choice of value for @code{STORE_FLAG_VALUE}, and hence the instructions
10450 Use the shortest sequence that yields a valid definition for
10451 @code{STORE_FLAG_VALUE}. It is more efficient for the compiler to
10452 ``normalize'' the value (convert it to, e.g., 1 or 0) than for the
10453 comparison operators to do so because there may be opportunities to
10454 combine the normalization with other operations.
10457 For equal-length sequences, use a value of 1 or @minus{}1, with @minus{}1 being
10458 slightly preferred on machines with expensive jumps and 1 preferred on
10462 As a second choice, choose a value of @samp{0x80000001} if instructions
10463 exist that set both the sign and low-order bits but do not define the
10467 Otherwise, use a value of @samp{0x80000000}.
10470 Many machines can produce both the value chosen for
10471 @code{STORE_FLAG_VALUE} and its negation in the same number of
10472 instructions. On those machines, you should also define a pattern for
10473 those cases, e.g., one matching
10476 (set @var{A} (neg:@var{m} (ne:@var{m} @var{B} @var{C})))
10479 Some machines can also perform @code{and} or @code{plus} operations on
10480 condition code values with less instructions than the corresponding
10481 @samp{cstore@var{mode}4} insn followed by @code{and} or @code{plus}. On those
10482 machines, define the appropriate patterns. Use the names @code{incscc}
10483 and @code{decscc}, respectively, for the patterns which perform
10484 @code{plus} or @code{minus} operations on condition code values. See
10485 @file{rs6000.md} for some examples. The GNU Superoptimizer can be used to
10486 find such instruction sequences on other machines.
10488 If this macro is not defined, the default value, 1, is used. You need
10489 not define @code{STORE_FLAG_VALUE} if the machine has no store-flag
10490 instructions, or if the value generated by these instructions is 1.
10493 @defmac FLOAT_STORE_FLAG_VALUE (@var{mode})
10494 A C expression that gives a nonzero @code{REAL_VALUE_TYPE} value that is
10495 returned when comparison operators with floating-point results are true.
10496 Define this macro on machines that have comparison operations that return
10497 floating-point values. If there are no such operations, do not define
10501 @defmac VECTOR_STORE_FLAG_VALUE (@var{mode})
10502 A C expression that gives a rtx representing the nonzero true element
10503 for vector comparisons. The returned rtx should be valid for the inner
10504 mode of @var{mode} which is guaranteed to be a vector mode. Define
10505 this macro on machines that have vector comparison operations that
10506 return a vector result. If there are no such operations, do not define
10507 this macro. Typically, this macro is defined as @code{const1_rtx} or
10508 @code{constm1_rtx}. This macro may return @code{NULL_RTX} to prevent
10509 the compiler optimizing such vector comparison operations for the
10513 @defmac CLZ_DEFINED_VALUE_AT_ZERO (@var{mode}, @var{value})
10514 @defmacx CTZ_DEFINED_VALUE_AT_ZERO (@var{mode}, @var{value})
10515 A C expression that indicates whether the architecture defines a value
10516 for @code{clz} or @code{ctz} with a zero operand.
10517 A result of @code{0} indicates the value is undefined.
10518 If the value is defined for only the RTL expression, the macro should
10519 evaluate to @code{1}; if the value applies also to the corresponding optab
10520 entry (which is normally the case if it expands directly into
10521 the corresponding RTL), then the macro should evaluate to @code{2}.
10522 In the cases where the value is defined, @var{value} should be set to
10525 If this macro is not defined, the value of @code{clz} or
10526 @code{ctz} at zero is assumed to be undefined.
10528 This macro must be defined if the target's expansion for @code{ffs}
10529 relies on a particular value to get correct results. Otherwise it
10530 is not necessary, though it may be used to optimize some corner cases, and
10531 to provide a default expansion for the @code{ffs} optab.
10533 Note that regardless of this macro the ``definedness'' of @code{clz}
10534 and @code{ctz} at zero do @emph{not} extend to the builtin functions
10535 visible to the user. Thus one may be free to adjust the value at will
10536 to match the target expansion of these operations without fear of
10541 An alias for the machine mode for pointers. On most machines, define
10542 this to be the integer mode corresponding to the width of a hardware
10543 pointer; @code{SImode} on 32-bit machine or @code{DImode} on 64-bit machines.
10544 On some machines you must define this to be one of the partial integer
10545 modes, such as @code{PSImode}.
10547 The width of @code{Pmode} must be at least as large as the value of
10548 @code{POINTER_SIZE}. If it is not equal, you must define the macro
10549 @code{POINTERS_EXTEND_UNSIGNED} to specify how pointers are extended
10553 @defmac FUNCTION_MODE
10554 An alias for the machine mode used for memory references to functions
10555 being called, in @code{call} RTL expressions. On most CISC machines,
10556 where an instruction can begin at any byte address, this should be
10557 @code{QImode}. On most RISC machines, where all instructions have fixed
10558 size and alignment, this should be a mode with the same size and alignment
10559 as the machine instruction words - typically @code{SImode} or @code{HImode}.
10562 @defmac STDC_0_IN_SYSTEM_HEADERS
10563 In normal operation, the preprocessor expands @code{__STDC__} to the
10564 constant 1, to signify that GCC conforms to ISO Standard C@. On some
10565 hosts, like Solaris, the system compiler uses a different convention,
10566 where @code{__STDC__} is normally 0, but is 1 if the user specifies
10567 strict conformance to the C Standard.
10569 Defining @code{STDC_0_IN_SYSTEM_HEADERS} makes GNU CPP follows the host
10570 convention when processing system header files, but when processing user
10571 files @code{__STDC__} will always expand to 1.
10574 @defmac NO_IMPLICIT_EXTERN_C
10575 Define this macro if the system header files support C++ as well as C@.
10576 This macro inhibits the usual method of using system header files in
10577 C++, which is to pretend that the file's contents are enclosed in
10578 @samp{extern "C" @{@dots{}@}}.
10583 @defmac REGISTER_TARGET_PRAGMAS ()
10584 Define this macro if you want to implement any target-specific pragmas.
10585 If defined, it is a C expression which makes a series of calls to
10586 @code{c_register_pragma} or @code{c_register_pragma_with_expansion}
10587 for each pragma. The macro may also do any
10588 setup required for the pragmas.
10590 The primary reason to define this macro is to provide compatibility with
10591 other compilers for the same target. In general, we discourage
10592 definition of target-specific pragmas for GCC@.
10594 If the pragma can be implemented by attributes then you should consider
10595 defining the target hook @samp{TARGET_INSERT_ATTRIBUTES} as well.
10597 Preprocessor macros that appear on pragma lines are not expanded. All
10598 @samp{#pragma} directives that do not match any registered pragma are
10599 silently ignored, unless the user specifies @option{-Wunknown-pragmas}.
10602 @deftypefun void c_register_pragma (const char *@var{space}, const char *@var{name}, void (*@var{callback}) (struct cpp_reader *))
10603 @deftypefunx void c_register_pragma_with_expansion (const char *@var{space}, const char *@var{name}, void (*@var{callback}) (struct cpp_reader *))
10605 Each call to @code{c_register_pragma} or
10606 @code{c_register_pragma_with_expansion} establishes one pragma. The
10607 @var{callback} routine will be called when the preprocessor encounters a
10611 #pragma [@var{space}] @var{name} @dots{}
10614 @var{space} is the case-sensitive namespace of the pragma, or
10615 @code{NULL} to put the pragma in the global namespace. The callback
10616 routine receives @var{pfile} as its first argument, which can be passed
10617 on to cpplib's functions if necessary. You can lex tokens after the
10618 @var{name} by calling @code{pragma_lex}. Tokens that are not read by the
10619 callback will be silently ignored. The end of the line is indicated by
10620 a token of type @code{CPP_EOF}. Macro expansion occurs on the
10621 arguments of pragmas registered with
10622 @code{c_register_pragma_with_expansion} but not on the arguments of
10623 pragmas registered with @code{c_register_pragma}.
10625 Note that the use of @code{pragma_lex} is specific to the C and C++
10626 compilers. It will not work in the Java or Fortran compilers, or any
10627 other language compilers for that matter. Thus if @code{pragma_lex} is going
10628 to be called from target-specific code, it must only be done so when
10629 building the C and C++ compilers. This can be done by defining the
10630 variables @code{c_target_objs} and @code{cxx_target_objs} in the
10631 target entry in the @file{config.gcc} file. These variables should name
10632 the target-specific, language-specific object file which contains the
10633 code that uses @code{pragma_lex}. Note it will also be necessary to add a
10634 rule to the makefile fragment pointed to by @code{tmake_file} that shows
10635 how to build this object file.
10638 @defmac HANDLE_PRAGMA_PACK_WITH_EXPANSION
10639 Define this macro if macros should be expanded in the
10640 arguments of @samp{#pragma pack}.
10643 @hook TARGET_HANDLE_PRAGMA_EXTERN_PREFIX
10645 @defmac TARGET_DEFAULT_PACK_STRUCT
10646 If your target requires a structure packing default other than 0 (meaning
10647 the machine default), define this macro to the necessary value (in bytes).
10648 This must be a value that would also be valid to use with
10649 @samp{#pragma pack()} (that is, a small power of two).
10652 @defmac DOLLARS_IN_IDENTIFIERS
10653 Define this macro to control use of the character @samp{$} in
10654 identifier names for the C family of languages. 0 means @samp{$} is
10655 not allowed by default; 1 means it is allowed. 1 is the default;
10656 there is no need to define this macro in that case.
10659 @defmac NO_DOLLAR_IN_LABEL
10660 Define this macro if the assembler does not accept the character
10661 @samp{$} in label names. By default constructors and destructors in
10662 G++ have @samp{$} in the identifiers. If this macro is defined,
10663 @samp{.} is used instead.
10666 @defmac NO_DOT_IN_LABEL
10667 Define this macro if the assembler does not accept the character
10668 @samp{.} in label names. By default constructors and destructors in G++
10669 have names that use @samp{.}. If this macro is defined, these names
10670 are rewritten to avoid @samp{.}.
10673 @defmac INSN_SETS_ARE_DELAYED (@var{insn})
10674 Define this macro as a C expression that is nonzero if it is safe for the
10675 delay slot scheduler to place instructions in the delay slot of @var{insn},
10676 even if they appear to use a resource set or clobbered in @var{insn}.
10677 @var{insn} is always a @code{jump_insn} or an @code{insn}; GCC knows that
10678 every @code{call_insn} has this behavior. On machines where some @code{insn}
10679 or @code{jump_insn} is really a function call and hence has this behavior,
10680 you should define this macro.
10682 You need not define this macro if it would always return zero.
10685 @defmac INSN_REFERENCES_ARE_DELAYED (@var{insn})
10686 Define this macro as a C expression that is nonzero if it is safe for the
10687 delay slot scheduler to place instructions in the delay slot of @var{insn},
10688 even if they appear to set or clobber a resource referenced in @var{insn}.
10689 @var{insn} is always a @code{jump_insn} or an @code{insn}. On machines where
10690 some @code{insn} or @code{jump_insn} is really a function call and its operands
10691 are registers whose use is actually in the subroutine it calls, you should
10692 define this macro. Doing so allows the delay slot scheduler to move
10693 instructions which copy arguments into the argument registers into the delay
10694 slot of @var{insn}.
10696 You need not define this macro if it would always return zero.
10699 @defmac MULTIPLE_SYMBOL_SPACES
10700 Define this macro as a C expression that is nonzero if, in some cases,
10701 global symbols from one translation unit may not be bound to undefined
10702 symbols in another translation unit without user intervention. For
10703 instance, under Microsoft Windows symbols must be explicitly imported
10704 from shared libraries (DLLs).
10706 You need not define this macro if it would always evaluate to zero.
10709 @hook TARGET_MD_ASM_CLOBBERS
10710 This target hook should add to @var{clobbers} @code{STRING_CST} trees for
10711 any hard regs the port wishes to automatically clobber for an asm.
10712 It should return the result of the last @code{tree_cons} used to add a
10713 clobber. The @var{outputs}, @var{inputs} and @var{clobber} lists are the
10714 corresponding parameters to the asm and may be inspected to avoid
10715 clobbering a register that is an input or output of the asm. You can use
10716 @code{tree_overlaps_hard_reg_set}, declared in @file{tree.h}, to test
10717 for overlap with regards to asm-declared registers.
10720 @defmac MATH_LIBRARY
10721 Define this macro as a C string constant for the linker argument to link
10722 in the system math library, minus the initial @samp{"-l"}, or
10723 @samp{""} if the target does not have a
10724 separate math library.
10726 You need only define this macro if the default of @samp{"m"} is wrong.
10729 @defmac LIBRARY_PATH_ENV
10730 Define this macro as a C string constant for the environment variable that
10731 specifies where the linker should look for libraries.
10733 You need only define this macro if the default of @samp{"LIBRARY_PATH"}
10737 @defmac TARGET_POSIX_IO
10738 Define this macro if the target supports the following POSIX@ file
10739 functions, access, mkdir and file locking with fcntl / F_SETLKW@.
10740 Defining @code{TARGET_POSIX_IO} will enable the test coverage code
10741 to use file locking when exiting a program, which avoids race conditions
10742 if the program has forked. It will also create directories at run-time
10743 for cross-profiling.
10746 @defmac MAX_CONDITIONAL_EXECUTE
10748 A C expression for the maximum number of instructions to execute via
10749 conditional execution instructions instead of a branch. A value of
10750 @code{BRANCH_COST}+1 is the default if the machine does not use cc0, and
10751 1 if it does use cc0.
10754 @defmac IFCVT_MODIFY_TESTS (@var{ce_info}, @var{true_expr}, @var{false_expr})
10755 Used if the target needs to perform machine-dependent modifications on the
10756 conditionals used for turning basic blocks into conditionally executed code.
10757 @var{ce_info} points to a data structure, @code{struct ce_if_block}, which
10758 contains information about the currently processed blocks. @var{true_expr}
10759 and @var{false_expr} are the tests that are used for converting the
10760 then-block and the else-block, respectively. Set either @var{true_expr} or
10761 @var{false_expr} to a null pointer if the tests cannot be converted.
10764 @defmac IFCVT_MODIFY_MULTIPLE_TESTS (@var{ce_info}, @var{bb}, @var{true_expr}, @var{false_expr})
10765 Like @code{IFCVT_MODIFY_TESTS}, but used when converting more complicated
10766 if-statements into conditions combined by @code{and} and @code{or} operations.
10767 @var{bb} contains the basic block that contains the test that is currently
10768 being processed and about to be turned into a condition.
10771 @defmac IFCVT_MODIFY_INSN (@var{ce_info}, @var{pattern}, @var{insn})
10772 A C expression to modify the @var{PATTERN} of an @var{INSN} that is to
10773 be converted to conditional execution format. @var{ce_info} points to
10774 a data structure, @code{struct ce_if_block}, which contains information
10775 about the currently processed blocks.
10778 @defmac IFCVT_MODIFY_FINAL (@var{ce_info})
10779 A C expression to perform any final machine dependent modifications in
10780 converting code to conditional execution. The involved basic blocks
10781 can be found in the @code{struct ce_if_block} structure that is pointed
10782 to by @var{ce_info}.
10785 @defmac IFCVT_MODIFY_CANCEL (@var{ce_info})
10786 A C expression to cancel any machine dependent modifications in
10787 converting code to conditional execution. The involved basic blocks
10788 can be found in the @code{struct ce_if_block} structure that is pointed
10789 to by @var{ce_info}.
10792 @defmac IFCVT_INIT_EXTRA_FIELDS (@var{ce_info})
10793 A C expression to initialize any extra fields in a @code{struct ce_if_block}
10794 structure, which are defined by the @code{IFCVT_EXTRA_FIELDS} macro.
10797 @defmac IFCVT_EXTRA_FIELDS
10798 If defined, it should expand to a set of field declarations that will be
10799 added to the @code{struct ce_if_block} structure. These should be initialized
10800 by the @code{IFCVT_INIT_EXTRA_FIELDS} macro.
10803 @hook TARGET_MACHINE_DEPENDENT_REORG
10804 If non-null, this hook performs a target-specific pass over the
10805 instruction stream. The compiler will run it at all optimization levels,
10806 just before the point at which it normally does delayed-branch scheduling.
10808 The exact purpose of the hook varies from target to target. Some use
10809 it to do transformations that are necessary for correctness, such as
10810 laying out in-function constant pools or avoiding hardware hazards.
10811 Others use it as an opportunity to do some machine-dependent optimizations.
10813 You need not implement the hook if it has nothing to do. The default
10814 definition is null.
10817 @hook TARGET_INIT_BUILTINS
10818 Define this hook if you have any machine-specific built-in functions
10819 that need to be defined. It should be a function that performs the
10822 Machine specific built-in functions can be useful to expand special machine
10823 instructions that would otherwise not normally be generated because
10824 they have no equivalent in the source language (for example, SIMD vector
10825 instructions or prefetch instructions).
10827 To create a built-in function, call the function
10828 @code{lang_hooks.builtin_function}
10829 which is defined by the language front end. You can use any type nodes set
10830 up by @code{build_common_tree_nodes};
10831 only language front ends that use those two functions will call
10832 @samp{TARGET_INIT_BUILTINS}.
10835 @hook TARGET_BUILTIN_DECL
10836 Define this hook if you have any machine-specific built-in functions
10837 that need to be defined. It should be a function that returns the
10838 builtin function declaration for the builtin function code @var{code}.
10839 If there is no such builtin and it cannot be initialized at this time
10840 if @var{initialize_p} is true the function should return @code{NULL_TREE}.
10841 If @var{code} is out of range the function should return
10842 @code{error_mark_node}.
10845 @hook TARGET_EXPAND_BUILTIN
10847 Expand a call to a machine specific built-in function that was set up by
10848 @samp{TARGET_INIT_BUILTINS}. @var{exp} is the expression for the
10849 function call; the result should go to @var{target} if that is
10850 convenient, and have mode @var{mode} if that is convenient.
10851 @var{subtarget} may be used as the target for computing one of
10852 @var{exp}'s operands. @var{ignore} is nonzero if the value is to be
10853 ignored. This function should return the result of the call to the
10857 @hook TARGET_RESOLVE_OVERLOADED_BUILTIN
10858 Select a replacement for a machine specific built-in function that
10859 was set up by @samp{TARGET_INIT_BUILTINS}. This is done
10860 @emph{before} regular type checking, and so allows the target to
10861 implement a crude form of function overloading. @var{fndecl} is the
10862 declaration of the built-in function. @var{arglist} is the list of
10863 arguments passed to the built-in function. The result is a
10864 complete expression that implements the operation, usually
10865 another @code{CALL_EXPR}.
10866 @var{arglist} really has type @samp{VEC(tree,gc)*}
10869 @hook TARGET_FOLD_BUILTIN
10870 Fold a call to a machine specific built-in function that was set up by
10871 @samp{TARGET_INIT_BUILTINS}. @var{fndecl} is the declaration of the
10872 built-in function. @var{n_args} is the number of arguments passed to
10873 the function; the arguments themselves are pointed to by @var{argp}.
10874 The result is another tree containing a simplified expression for the
10875 call's result. If @var{ignore} is true the value will be ignored.
10878 @hook TARGET_INVALID_WITHIN_DOLOOP
10880 Take an instruction in @var{insn} and return NULL if it is valid within a
10881 low-overhead loop, otherwise return a string explaining why doloop
10882 could not be applied.
10884 Many targets use special registers for low-overhead looping. For any
10885 instruction that clobbers these this function should return a string indicating
10886 the reason why the doloop could not be applied.
10887 By default, the RTL loop optimizer does not use a present doloop pattern for
10888 loops containing function calls or branch on table instructions.
10891 @defmac MD_CAN_REDIRECT_BRANCH (@var{branch1}, @var{branch2})
10893 Take a branch insn in @var{branch1} and another in @var{branch2}.
10894 Return true if redirecting @var{branch1} to the destination of
10895 @var{branch2} is possible.
10897 On some targets, branches may have a limited range. Optimizing the
10898 filling of delay slots can result in branches being redirected, and this
10899 may in turn cause a branch offset to overflow.
10902 @hook TARGET_COMMUTATIVE_P
10903 This target hook returns @code{true} if @var{x} is considered to be commutative.
10904 Usually, this is just COMMUTATIVE_P (@var{x}), but the HP PA doesn't consider
10905 PLUS to be commutative inside a MEM@. @var{outer_code} is the rtx code
10906 of the enclosing rtl, if known, otherwise it is UNKNOWN.
10909 @hook TARGET_ALLOCATE_INITIAL_VALUE
10911 When the initial value of a hard register has been copied in a pseudo
10912 register, it is often not necessary to actually allocate another register
10913 to this pseudo register, because the original hard register or a stack slot
10914 it has been saved into can be used. @code{TARGET_ALLOCATE_INITIAL_VALUE}
10915 is called at the start of register allocation once for each hard register
10916 that had its initial value copied by using
10917 @code{get_func_hard_reg_initial_val} or @code{get_hard_reg_initial_val}.
10918 Possible values are @code{NULL_RTX}, if you don't want
10919 to do any special allocation, a @code{REG} rtx---that would typically be
10920 the hard register itself, if it is known not to be clobbered---or a
10922 If you are returning a @code{MEM}, this is only a hint for the allocator;
10923 it might decide to use another register anyways.
10924 You may use @code{current_function_leaf_function} in the hook, functions
10925 that use @code{REG_N_SETS}, to determine if the hard
10926 register in question will not be clobbered.
10927 The default value of this hook is @code{NULL}, which disables any special
10931 @hook TARGET_UNSPEC_MAY_TRAP_P
10932 This target hook returns nonzero if @var{x}, an @code{unspec} or
10933 @code{unspec_volatile} operation, might cause a trap. Targets can use
10934 this hook to enhance precision of analysis for @code{unspec} and
10935 @code{unspec_volatile} operations. You may call @code{may_trap_p_1}
10936 to analyze inner elements of @var{x} in which case @var{flags} should be
10940 @hook TARGET_SET_CURRENT_FUNCTION
10941 The compiler invokes this hook whenever it changes its current function
10942 context (@code{cfun}). You can define this function if
10943 the back end needs to perform any initialization or reset actions on a
10944 per-function basis. For example, it may be used to implement function
10945 attributes that affect register usage or code generation patterns.
10946 The argument @var{decl} is the declaration for the new function context,
10947 and may be null to indicate that the compiler has left a function context
10948 and is returning to processing at the top level.
10949 The default hook function does nothing.
10951 GCC sets @code{cfun} to a dummy function context during initialization of
10952 some parts of the back end. The hook function is not invoked in this
10953 situation; you need not worry about the hook being invoked recursively,
10954 or when the back end is in a partially-initialized state.
10955 @code{cfun} might be @code{NULL} to indicate processing at top level,
10956 outside of any function scope.
10959 @defmac TARGET_OBJECT_SUFFIX
10960 Define this macro to be a C string representing the suffix for object
10961 files on your target machine. If you do not define this macro, GCC will
10962 use @samp{.o} as the suffix for object files.
10965 @defmac TARGET_EXECUTABLE_SUFFIX
10966 Define this macro to be a C string representing the suffix to be
10967 automatically added to executable files on your target machine. If you
10968 do not define this macro, GCC will use the null string as the suffix for
10972 @defmac COLLECT_EXPORT_LIST
10973 If defined, @code{collect2} will scan the individual object files
10974 specified on its command line and create an export list for the linker.
10975 Define this macro for systems like AIX, where the linker discards
10976 object files that are not referenced from @code{main} and uses export
10980 @defmac MODIFY_JNI_METHOD_CALL (@var{mdecl})
10981 Define this macro to a C expression representing a variant of the
10982 method call @var{mdecl}, if Java Native Interface (JNI) methods
10983 must be invoked differently from other methods on your target.
10984 For example, on 32-bit Microsoft Windows, JNI methods must be invoked using
10985 the @code{stdcall} calling convention and this macro is then
10986 defined as this expression:
10989 build_type_attribute_variant (@var{mdecl},
10991 (get_identifier ("stdcall"),
10996 @hook TARGET_CANNOT_MODIFY_JUMPS_P
10997 This target hook returns @code{true} past the point in which new jump
10998 instructions could be created. On machines that require a register for
10999 every jump such as the SHmedia ISA of SH5, this point would typically be
11000 reload, so this target hook should be defined to a function such as:
11004 cannot_modify_jumps_past_reload_p ()
11006 return (reload_completed || reload_in_progress);
11011 @hook TARGET_BRANCH_TARGET_REGISTER_CLASS
11012 This target hook returns a register class for which branch target register
11013 optimizations should be applied. All registers in this class should be
11014 usable interchangeably. After reload, registers in this class will be
11015 re-allocated and loads will be hoisted out of loops and be subjected
11016 to inter-block scheduling.
11019 @hook TARGET_BRANCH_TARGET_REGISTER_CALLEE_SAVED
11020 Branch target register optimization will by default exclude callee-saved
11022 that are not already live during the current function; if this target hook
11023 returns true, they will be included. The target code must than make sure
11024 that all target registers in the class returned by
11025 @samp{TARGET_BRANCH_TARGET_REGISTER_CLASS} that might need saving are
11026 saved. @var{after_prologue_epilogue_gen} indicates if prologues and
11027 epilogues have already been generated. Note, even if you only return
11028 true when @var{after_prologue_epilogue_gen} is false, you still are likely
11029 to have to make special provisions in @code{INITIAL_ELIMINATION_OFFSET}
11030 to reserve space for caller-saved target registers.
11033 @hook TARGET_HAVE_CONDITIONAL_EXECUTION
11034 This target hook returns true if the target supports conditional execution.
11035 This target hook is required only when the target has several different
11036 modes and they have different conditional execution capability, such as ARM.
11039 @hook TARGET_LOOP_UNROLL_ADJUST
11040 This target hook returns a new value for the number of times @var{loop}
11041 should be unrolled. The parameter @var{nunroll} is the number of times
11042 the loop is to be unrolled. The parameter @var{loop} is a pointer to
11043 the loop, which is going to be checked for unrolling. This target hook
11044 is required only when the target has special constraints like maximum
11045 number of memory accesses.
11048 @defmac POWI_MAX_MULTS
11049 If defined, this macro is interpreted as a signed integer C expression
11050 that specifies the maximum number of floating point multiplications
11051 that should be emitted when expanding exponentiation by an integer
11052 constant inline. When this value is defined, exponentiation requiring
11053 more than this number of multiplications is implemented by calling the
11054 system library's @code{pow}, @code{powf} or @code{powl} routines.
11055 The default value places no upper bound on the multiplication count.
11058 @deftypefn Macro void TARGET_EXTRA_INCLUDES (const char *@var{sysroot}, const char *@var{iprefix}, int @var{stdinc})
11059 This target hook should register any extra include files for the
11060 target. The parameter @var{stdinc} indicates if normal include files
11061 are present. The parameter @var{sysroot} is the system root directory.
11062 The parameter @var{iprefix} is the prefix for the gcc directory.
11065 @deftypefn Macro void TARGET_EXTRA_PRE_INCLUDES (const char *@var{sysroot}, const char *@var{iprefix}, int @var{stdinc})
11066 This target hook should register any extra include files for the
11067 target before any standard headers. The parameter @var{stdinc}
11068 indicates if normal include files are present. The parameter
11069 @var{sysroot} is the system root directory. The parameter
11070 @var{iprefix} is the prefix for the gcc directory.
11073 @deftypefn Macro void TARGET_OPTF (char *@var{path})
11074 This target hook should register special include paths for the target.
11075 The parameter @var{path} is the include to register. On Darwin
11076 systems, this is used for Framework includes, which have semantics
11077 that are different from @option{-I}.
11080 @defmac bool TARGET_USE_LOCAL_THUNK_ALIAS_P (tree @var{fndecl})
11081 This target macro returns @code{true} if it is safe to use a local alias
11082 for a virtual function @var{fndecl} when constructing thunks,
11083 @code{false} otherwise. By default, the macro returns @code{true} for all
11084 functions, if a target supports aliases (i.e.@: defines
11085 @code{ASM_OUTPUT_DEF}), @code{false} otherwise,
11088 @defmac TARGET_FORMAT_TYPES
11089 If defined, this macro is the name of a global variable containing
11090 target-specific format checking information for the @option{-Wformat}
11091 option. The default is to have no target-specific format checks.
11094 @defmac TARGET_N_FORMAT_TYPES
11095 If defined, this macro is the number of entries in
11096 @code{TARGET_FORMAT_TYPES}.
11099 @defmac TARGET_OVERRIDES_FORMAT_ATTRIBUTES
11100 If defined, this macro is the name of a global variable containing
11101 target-specific format overrides for the @option{-Wformat} option. The
11102 default is to have no target-specific format overrides. If defined,
11103 @code{TARGET_FORMAT_TYPES} must be defined, too.
11106 @defmac TARGET_OVERRIDES_FORMAT_ATTRIBUTES_COUNT
11107 If defined, this macro specifies the number of entries in
11108 @code{TARGET_OVERRIDES_FORMAT_ATTRIBUTES}.
11111 @defmac TARGET_OVERRIDES_FORMAT_INIT
11112 If defined, this macro specifies the optional initialization
11113 routine for target specific customizations of the system printf
11114 and scanf formatter settings.
11117 @hook TARGET_RELAXED_ORDERING
11118 If set to @code{true}, means that the target's memory model does not
11119 guarantee that loads which do not depend on one another will access
11120 main memory in the order of the instruction stream; if ordering is
11121 important, an explicit memory barrier must be used. This is true of
11122 many recent processors which implement a policy of ``relaxed,''
11123 ``weak,'' or ``release'' memory consistency, such as Alpha, PowerPC,
11124 and ia64. The default is @code{false}.
11127 @hook TARGET_INVALID_ARG_FOR_UNPROTOTYPED_FN
11128 If defined, this macro returns the diagnostic message when it is
11129 illegal to pass argument @var{val} to function @var{funcdecl}
11130 with prototype @var{typelist}.
11133 @hook TARGET_INVALID_CONVERSION
11134 If defined, this macro returns the diagnostic message when it is
11135 invalid to convert from @var{fromtype} to @var{totype}, or @code{NULL}
11136 if validity should be determined by the front end.
11139 @hook TARGET_INVALID_UNARY_OP
11140 If defined, this macro returns the diagnostic message when it is
11141 invalid to apply operation @var{op} (where unary plus is denoted by
11142 @code{CONVERT_EXPR}) to an operand of type @var{type}, or @code{NULL}
11143 if validity should be determined by the front end.
11146 @hook TARGET_INVALID_BINARY_OP
11147 If defined, this macro returns the diagnostic message when it is
11148 invalid to apply operation @var{op} to operands of types @var{type1}
11149 and @var{type2}, or @code{NULL} if validity should be determined by
11153 @hook TARGET_INVALID_PARAMETER_TYPE
11154 If defined, this macro returns the diagnostic message when it is
11155 invalid for functions to include parameters of type @var{type},
11156 or @code{NULL} if validity should be determined by
11157 the front end. This is currently used only by the C and C++ front ends.
11160 @hook TARGET_INVALID_RETURN_TYPE
11161 If defined, this macro returns the diagnostic message when it is
11162 invalid for functions to have return type @var{type},
11163 or @code{NULL} if validity should be determined by
11164 the front end. This is currently used only by the C and C++ front ends.
11167 @hook TARGET_PROMOTED_TYPE
11168 If defined, this target hook returns the type to which values of
11169 @var{type} should be promoted when they appear in expressions,
11170 analogous to the integer promotions, or @code{NULL_TREE} to use the
11171 front end's normal promotion rules. This hook is useful when there are
11172 target-specific types with special promotion rules.
11173 This is currently used only by the C and C++ front ends.
11176 @hook TARGET_CONVERT_TO_TYPE
11177 If defined, this hook returns the result of converting @var{expr} to
11178 @var{type}. It should return the converted expression,
11179 or @code{NULL_TREE} to apply the front end's normal conversion rules.
11180 This hook is useful when there are target-specific types with special
11182 This is currently used only by the C and C++ front ends.
11185 @defmac TARGET_USE_JCR_SECTION
11186 This macro determines whether to use the JCR section to register Java
11187 classes. By default, TARGET_USE_JCR_SECTION is defined to 1 if both
11188 SUPPORTS_WEAK and TARGET_HAVE_NAMED_SECTIONS are true, else 0.
11192 This macro determines the size of the objective C jump buffer for the
11193 NeXT runtime. By default, OBJC_JBLEN is defined to an innocuous value.
11196 @defmac LIBGCC2_UNWIND_ATTRIBUTE
11197 Define this macro if any target-specific attributes need to be attached
11198 to the functions in @file{libgcc} that provide low-level support for
11199 call stack unwinding. It is used in declarations in @file{unwind-generic.h}
11200 and the associated definitions of those functions.
11203 @hook TARGET_UPDATE_STACK_BOUNDARY
11204 Define this macro to update the current function stack boundary if
11208 @hook TARGET_GET_DRAP_RTX
11209 This hook should return an rtx for Dynamic Realign Argument Pointer (DRAP) if a
11210 different argument pointer register is needed to access the function's
11211 argument list due to stack realignment. Return @code{NULL} if no DRAP
11215 @hook TARGET_ALLOCATE_STACK_SLOTS_FOR_ARGS
11216 When optimization is disabled, this hook indicates whether or not
11217 arguments should be allocated to stack slots. Normally, GCC allocates
11218 stacks slots for arguments when not optimizing in order to make
11219 debugging easier. However, when a function is declared with
11220 @code{__attribute__((naked))}, there is no stack frame, and the compiler
11221 cannot safely move arguments from the registers in which they are passed
11222 to the stack. Therefore, this hook should return true in general, but
11223 false for naked functions. The default implementation always returns true.
11226 @hook TARGET_CONST_ANCHOR
11227 On some architectures it can take multiple instructions to synthesize
11228 a constant. If there is another constant already in a register that
11229 is close enough in value then it is preferable that the new constant
11230 is computed from this register using immediate addition or
11231 subtraction. We accomplish this through CSE. Besides the value of
11232 the constant we also add a lower and an upper constant anchor to the
11233 available expressions. These are then queried when encountering new
11234 constants. The anchors are computed by rounding the constant up and
11235 down to a multiple of the value of @code{TARGET_CONST_ANCHOR}.
11236 @code{TARGET_CONST_ANCHOR} should be the maximum positive value
11237 accepted by immediate-add plus one. We currently assume that the
11238 value of @code{TARGET_CONST_ANCHOR} is a power of 2. For example, on
11239 MIPS, where add-immediate takes a 16-bit signed value,
11240 @code{TARGET_CONST_ANCHOR} is set to @samp{0x8000}. The default value
11241 is zero, which disables this optimization. @end deftypevr
11243 @hook TARGET_ATOMIC_TEST_AND_SET_TRUEVAL