1 \input texinfo @c -*- texinfo -*-
2 @setfilename gdbint.info
4 @dircategory Software development
6 * Gdb-Internals: (gdbint). The GNU debugger's internals.
10 This file documents the internals of the GNU debugger @value{GDBN}.
11 Copyright 1990, 1991, 1992, 1993, 1994, 1996, 1998, 1999, 2000, 2001,
12 2002, 2003, 2004, 2005
13 Free Software Foundation, Inc.
14 Contributed by Cygnus Solutions. Written by John Gilmore.
15 Second Edition by Stan Shebs.
17 Permission is granted to copy, distribute and/or modify this document
18 under the terms of the GNU Free Documentation License, Version 1.1 or
19 any later version published by the Free Software Foundation; with no
20 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
21 Texts. A copy of the license is included in the section entitled ``GNU
22 Free Documentation License''.
25 @setchapternewpage off
26 @settitle @value{GDBN} Internals
32 @title @value{GDBN} Internals
33 @subtitle{A guide to the internals of the GNU debugger}
35 @author Cygnus Solutions
36 @author Second Edition:
38 @author Cygnus Solutions
41 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
42 \xdef\manvers{\$Revision$} % For use in headers, footers too
44 \hfill Cygnus Solutions\par
46 \hfill \TeX{}info \texinfoversion\par
50 @vskip 0pt plus 1filll
51 Copyright @copyright{} 1990,1991,1992,1993,1994,1996,1998,1999,2000,2001,
52 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
54 Permission is granted to copy, distribute and/or modify this document
55 under the terms of the GNU Free Documentation License, Version 1.1 or
56 any later version published by the Free Software Foundation; with no
57 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
58 Texts. A copy of the license is included in the section entitled ``GNU
59 Free Documentation License''.
65 @c Perhaps this should be the title of the document (but only for info,
66 @c not for TeX). Existing GNU manuals seem inconsistent on this point.
67 @top Scope of this Document
69 This document documents the internals of the GNU debugger, @value{GDBN}. It
70 includes description of @value{GDBN}'s key algorithms and operations, as well
71 as the mechanisms that adapt @value{GDBN} to specific hosts and targets.
82 * Target Architecture Definition::
83 * Target Vector Definition::
88 * Versions and Branches::
93 * GDB Observers:: @value{GDBN} Currently available observers
94 * GNU Free Documentation License:: The license for this documentation
100 @chapter Requirements
101 @cindex requirements for @value{GDBN}
103 Before diving into the internals, you should understand the formal
104 requirements and other expectations for @value{GDBN}. Although some
105 of these may seem obvious, there have been proposals for @value{GDBN}
106 that have run counter to these requirements.
108 First of all, @value{GDBN} is a debugger. It's not designed to be a
109 front panel for embedded systems. It's not a text editor. It's not a
110 shell. It's not a programming environment.
112 @value{GDBN} is an interactive tool. Although a batch mode is
113 available, @value{GDBN}'s primary role is to interact with a human
116 @value{GDBN} should be responsive to the user. A programmer hot on
117 the trail of a nasty bug, and operating under a looming deadline, is
118 going to be very impatient of everything, including the response time
119 to debugger commands.
121 @value{GDBN} should be relatively permissive, such as for expressions.
122 While the compiler should be picky (or have the option to be made
123 picky), since source code lives for a long time usually, the
124 programmer doing debugging shouldn't be spending time figuring out to
125 mollify the debugger.
127 @value{GDBN} will be called upon to deal with really large programs.
128 Executable sizes of 50 to 100 megabytes occur regularly, and we've
129 heard reports of programs approaching 1 gigabyte in size.
131 @value{GDBN} should be able to run everywhere. No other debugger is
132 available for even half as many configurations as @value{GDBN}
136 @node Overall Structure
138 @chapter Overall Structure
140 @value{GDBN} consists of three major subsystems: user interface,
141 symbol handling (the @dfn{symbol side}), and target system handling (the
144 The user interface consists of several actual interfaces, plus
147 The symbol side consists of object file readers, debugging info
148 interpreters, symbol table management, source language expression
149 parsing, type and value printing.
151 The target side consists of execution control, stack frame analysis, and
152 physical target manipulation.
154 The target side/symbol side division is not formal, and there are a
155 number of exceptions. For instance, core file support involves symbolic
156 elements (the basic core file reader is in BFD) and target elements (it
157 supplies the contents of memory and the values of registers). Instead,
158 this division is useful for understanding how the minor subsystems
161 @section The Symbol Side
163 The symbolic side of @value{GDBN} can be thought of as ``everything
164 you can do in @value{GDBN} without having a live program running''.
165 For instance, you can look at the types of variables, and evaluate
166 many kinds of expressions.
168 @section The Target Side
170 The target side of @value{GDBN} is the ``bits and bytes manipulator''.
171 Although it may make reference to symbolic info here and there, most
172 of the target side will run with only a stripped executable
173 available---or even no executable at all, in remote debugging cases.
175 Operations such as disassembly, stack frame crawls, and register
176 display, are able to work with no symbolic info at all. In some cases,
177 such as disassembly, @value{GDBN} will use symbolic info to present addresses
178 relative to symbols rather than as raw numbers, but it will work either
181 @section Configurations
185 @dfn{Host} refers to attributes of the system where @value{GDBN} runs.
186 @dfn{Target} refers to the system where the program being debugged
187 executes. In most cases they are the same machine, in which case a
188 third type of @dfn{Native} attributes come into play.
190 Defines and include files needed to build on the host are host support.
191 Examples are tty support, system defined types, host byte order, host
194 Defines and information needed to handle the target format are target
195 dependent. Examples are the stack frame format, instruction set,
196 breakpoint instruction, registers, and how to set up and tear down the stack
199 Information that is only needed when the host and target are the same,
200 is native dependent. One example is Unix child process support; if the
201 host and target are not the same, doing a fork to start the target
202 process is a bad idea. The various macros needed for finding the
203 registers in the @code{upage}, running @code{ptrace}, and such are all
204 in the native-dependent files.
206 Another example of native-dependent code is support for features that
207 are really part of the target environment, but which require
208 @code{#include} files that are only available on the host system. Core
209 file handling and @code{setjmp} handling are two common cases.
211 When you want to make @value{GDBN} work ``native'' on a particular machine, you
212 have to include all three kinds of information.
220 @value{GDBN} uses a number of debugging-specific algorithms. They are
221 often not very complicated, but get lost in the thicket of special
222 cases and real-world issues. This chapter describes the basic
223 algorithms and mentions some of the specific target definitions that
229 @cindex call stack frame
230 A frame is a construct that @value{GDBN} uses to keep track of calling
231 and called functions.
233 @findex create_new_frame
235 @code{FRAME_FP} in the machine description has no meaning to the
236 machine-independent part of @value{GDBN}, except that it is used when
237 setting up a new frame from scratch, as follows:
240 create_new_frame (read_register (DEPRECATED_FP_REGNUM), read_pc ()));
243 @cindex frame pointer register
244 Other than that, all the meaning imparted to @code{DEPRECATED_FP_REGNUM}
245 is imparted by the machine-dependent code. So,
246 @code{DEPRECATED_FP_REGNUM} can have any value that is convenient for
247 the code that creates new frames. (@code{create_new_frame} calls
248 @code{DEPRECATED_INIT_EXTRA_FRAME_INFO} if it is defined; that is where
249 you should use the @code{DEPRECATED_FP_REGNUM} value, if your frames are
253 Given a @value{GDBN} frame, define @code{DEPRECATED_FRAME_CHAIN} to
254 determine the address of the calling function's frame. This will be
255 used to create a new @value{GDBN} frame struct, and then
256 @code{DEPRECATED_INIT_EXTRA_FRAME_INFO} and
257 @code{DEPRECATED_INIT_FRAME_PC} will be called for the new frame.
259 @section Breakpoint Handling
262 In general, a breakpoint is a user-designated location in the program
263 where the user wants to regain control if program execution ever reaches
266 There are two main ways to implement breakpoints; either as ``hardware''
267 breakpoints or as ``software'' breakpoints.
269 @cindex hardware breakpoints
270 @cindex program counter
271 Hardware breakpoints are sometimes available as a builtin debugging
272 features with some chips. Typically these work by having dedicated
273 register into which the breakpoint address may be stored. If the PC
274 (shorthand for @dfn{program counter})
275 ever matches a value in a breakpoint registers, the CPU raises an
276 exception and reports it to @value{GDBN}.
278 Another possibility is when an emulator is in use; many emulators
279 include circuitry that watches the address lines coming out from the
280 processor, and force it to stop if the address matches a breakpoint's
283 A third possibility is that the target already has the ability to do
284 breakpoints somehow; for instance, a ROM monitor may do its own
285 software breakpoints. So although these are not literally ``hardware
286 breakpoints'', from @value{GDBN}'s point of view they work the same;
287 @value{GDBN} need not do anything more than set the breakpoint and wait
288 for something to happen.
290 Since they depend on hardware resources, hardware breakpoints may be
291 limited in number; when the user asks for more, @value{GDBN} will
292 start trying to set software breakpoints. (On some architectures,
293 notably the 32-bit x86 platforms, @value{GDBN} cannot always know
294 whether there's enough hardware resources to insert all the hardware
295 breakpoints and watchpoints. On those platforms, @value{GDBN} prints
296 an error message only when the program being debugged is continued.)
298 @cindex software breakpoints
299 Software breakpoints require @value{GDBN} to do somewhat more work.
300 The basic theory is that @value{GDBN} will replace a program
301 instruction with a trap, illegal divide, or some other instruction
302 that will cause an exception, and then when it's encountered,
303 @value{GDBN} will take the exception and stop the program. When the
304 user says to continue, @value{GDBN} will restore the original
305 instruction, single-step, re-insert the trap, and continue on.
307 Since it literally overwrites the program being tested, the program area
308 must be writable, so this technique won't work on programs in ROM. It
309 can also distort the behavior of programs that examine themselves,
310 although such a situation would be highly unusual.
312 Also, the software breakpoint instruction should be the smallest size of
313 instruction, so it doesn't overwrite an instruction that might be a jump
314 target, and cause disaster when the program jumps into the middle of the
315 breakpoint instruction. (Strictly speaking, the breakpoint must be no
316 larger than the smallest interval between instructions that may be jump
317 targets; perhaps there is an architecture where only even-numbered
318 instructions may jumped to.) Note that it's possible for an instruction
319 set not to have any instructions usable for a software breakpoint,
320 although in practice only the ARC has failed to define such an
324 The basic definition of the software breakpoint is the macro
327 Basic breakpoint object handling is in @file{breakpoint.c}. However,
328 much of the interesting breakpoint action is in @file{infrun.c}.
330 @section Single Stepping
332 @section Signal Handling
334 @section Thread Handling
336 @section Inferior Function Calls
338 @section Longjmp Support
340 @cindex @code{longjmp} debugging
341 @value{GDBN} has support for figuring out that the target is doing a
342 @code{longjmp} and for stopping at the target of the jump, if we are
343 stepping. This is done with a few specialized internal breakpoints,
344 which are visible in the output of the @samp{maint info breakpoint}
347 @findex GET_LONGJMP_TARGET
348 To make this work, you need to define a macro called
349 @code{GET_LONGJMP_TARGET}, which will examine the @code{jmp_buf}
350 structure and extract the longjmp target address. Since @code{jmp_buf}
351 is target specific, you will need to define it in the appropriate
352 @file{tm-@var{target}.h} file. Look in @file{tm-sun4os4.h} and
353 @file{sparc-tdep.c} for examples of how to do this.
358 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
359 breakpoints}) which break when data is accessed rather than when some
360 instruction is executed. When you have data which changes without
361 your knowing what code does that, watchpoints are the silver bullet to
362 hunt down and kill such bugs.
364 @cindex hardware watchpoints
365 @cindex software watchpoints
366 Watchpoints can be either hardware-assisted or not; the latter type is
367 known as ``software watchpoints.'' @value{GDBN} always uses
368 hardware-assisted watchpoints if they are available, and falls back on
369 software watchpoints otherwise. Typical situations where @value{GDBN}
370 will use software watchpoints are:
374 The watched memory region is too large for the underlying hardware
375 watchpoint support. For example, each x86 debug register can watch up
376 to 4 bytes of memory, so trying to watch data structures whose size is
377 more than 16 bytes will cause @value{GDBN} to use software
381 The value of the expression to be watched depends on data held in
382 registers (as opposed to memory).
385 Too many different watchpoints requested. (On some architectures,
386 this situation is impossible to detect until the debugged program is
387 resumed.) Note that x86 debug registers are used both for hardware
388 breakpoints and for watchpoints, so setting too many hardware
389 breakpoints might cause watchpoint insertion to fail.
392 No hardware-assisted watchpoints provided by the target
396 Software watchpoints are very slow, since @value{GDBN} needs to
397 single-step the program being debugged and test the value of the
398 watched expression(s) after each instruction. The rest of this
399 section is mostly irrelevant for software watchpoints.
401 When the inferior stops, @value{GDBN} tries to establish, among other
402 possible reasons, whether it stopped due to a watchpoint being hit.
403 For a data-write watchpoint, it does so by evaluating, for each
404 watchpoint, the expression whose value is being watched, and testing
405 whether the watched value has changed. For data-read and data-access
406 watchpoints, @value{GDBN} needs the target to supply a primitive that
407 returns the address of the data that was accessed or read (see the
408 description of @code{target_stopped_data_address} below): if this
409 primitive returns a valid address, @value{GDBN} infers that a
410 watchpoint triggered if it watches an expression whose evaluation uses
413 @value{GDBN} uses several macros and primitives to support hardware
417 @findex TARGET_HAS_HARDWARE_WATCHPOINTS
418 @item TARGET_HAS_HARDWARE_WATCHPOINTS
419 If defined, the target supports hardware watchpoints.
421 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
422 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
423 Return the number of hardware watchpoints of type @var{type} that are
424 possible to be set. The value is positive if @var{count} watchpoints
425 of this type can be set, zero if setting watchpoints of this type is
426 not supported, and negative if @var{count} is more than the maximum
427 number of watchpoints of type @var{type} that can be set. @var{other}
428 is non-zero if other types of watchpoints are currently enabled (there
429 are architectures which cannot set watchpoints of different types at
432 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
433 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
434 Return non-zero if hardware watchpoints can be used to watch a region
435 whose address is @var{addr} and whose length in bytes is @var{len}.
437 @findex TARGET_REGION_SIZE_OK_FOR_HW_WATCHPOINT
438 @item TARGET_REGION_SIZE_OK_FOR_HW_WATCHPOINT (@var{size})
439 Return non-zero if hardware watchpoints can be used to watch a region
440 whose size is @var{size}. @value{GDBN} only uses this macro as a
441 fall-back, in case @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is not
444 @cindex insert or remove hardware watchpoint
445 @findex target_insert_watchpoint
446 @findex target_remove_watchpoint
447 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
448 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
449 Insert or remove a hardware watchpoint starting at @var{addr}, for
450 @var{len} bytes. @var{type} is the watchpoint type, one of the
451 possible values of the enumerated data type @code{target_hw_bp_type},
452 defined by @file{breakpoint.h} as follows:
455 enum target_hw_bp_type
457 hw_write = 0, /* Common (write) HW watchpoint */
458 hw_read = 1, /* Read HW watchpoint */
459 hw_access = 2, /* Access (read or write) HW watchpoint */
460 hw_execute = 3 /* Execute HW breakpoint */
465 These two macros should return 0 for success, non-zero for failure.
467 @cindex insert or remove hardware breakpoint
468 @findex target_remove_hw_breakpoint
469 @findex target_insert_hw_breakpoint
470 @item target_remove_hw_breakpoint (@var{addr}, @var{shadow})
471 @itemx target_insert_hw_breakpoint (@var{addr}, @var{shadow})
472 Insert or remove a hardware-assisted breakpoint at address @var{addr}.
473 Returns zero for success, non-zero for failure. @var{shadow} is the
474 real contents of the byte where the breakpoint has been inserted; it
475 is generally not valid when hardware breakpoints are used, but since
476 no other code touches these values, the implementations of the above
477 two macros can use them for their internal purposes.
479 @findex target_stopped_data_address
480 @item target_stopped_data_address (@var{addr_p})
481 If the inferior has some watchpoint that triggered, place the address
482 associated with the watchpoint at the location pointed to by
483 @var{addr_p} and return non-zero. Otherwise, return zero. Note that
484 this primitive is used by @value{GDBN} only on targets that support
485 data-read or data-access type watchpoints, so targets that have
486 support only for data-write watchpoints need not implement these
489 @findex HAVE_STEPPABLE_WATCHPOINT
490 @item HAVE_STEPPABLE_WATCHPOINT
491 If defined to a non-zero value, it is not necessary to disable a
492 watchpoint to step over it.
494 @findex HAVE_NONSTEPPABLE_WATCHPOINT
495 @item HAVE_NONSTEPPABLE_WATCHPOINT
496 If defined to a non-zero value, @value{GDBN} should disable a
497 watchpoint to step the inferior over it.
499 @findex HAVE_CONTINUABLE_WATCHPOINT
500 @item HAVE_CONTINUABLE_WATCHPOINT
501 If defined to a non-zero value, it is possible to continue the
502 inferior after a watchpoint has been hit.
504 @findex CANNOT_STEP_HW_WATCHPOINTS
505 @item CANNOT_STEP_HW_WATCHPOINTS
506 If this is defined to a non-zero value, @value{GDBN} will remove all
507 watchpoints before stepping the inferior.
509 @findex STOPPED_BY_WATCHPOINT
510 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
511 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
512 the type @code{struct target_waitstatus}, defined by @file{target.h}.
513 Normally, this macro is defined to invoke the function pointed to by
514 the @code{to_stopped_by_watchpoint} member of the structure (of the
515 type @code{target_ops}, defined on @file{target.h}) that describes the
516 target-specific operations; @code{to_stopped_by_watchpoint} ignores
517 the @var{wait_status} argument.
519 @value{GDBN} does not require the non-zero value returned by
520 @code{STOPPED_BY_WATCHPOINT} to be 100% correct, so if a target cannot
521 determine for sure whether the inferior stopped due to a watchpoint,
522 it could return non-zero ``just in case''.
525 @subsection x86 Watchpoints
526 @cindex x86 debug registers
527 @cindex watchpoints, on x86
529 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
530 registers designed to facilitate debugging. @value{GDBN} provides a
531 generic library of functions that x86-based ports can use to implement
532 support for watchpoints and hardware-assisted breakpoints. This
533 subsection documents the x86 watchpoint facilities in @value{GDBN}.
535 To use the generic x86 watchpoint support, a port should do the
539 @findex I386_USE_GENERIC_WATCHPOINTS
541 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
542 target-dependent headers.
545 Include the @file{config/i386/nm-i386.h} header file @emph{after}
546 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
549 Add @file{i386-nat.o} to the value of the Make variable
550 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}) or
551 @code{TDEPFILES} (@pxref{Target Architecture Definition, TDEPFILES}).
554 Provide implementations for the @code{I386_DR_LOW_*} macros described
555 below. Typically, each macro should call a target-specific function
556 which does the real work.
559 The x86 watchpoint support works by maintaining mirror images of the
560 debug registers. Values are copied between the mirror images and the
561 real debug registers via a set of macros which each target needs to
565 @findex I386_DR_LOW_SET_CONTROL
566 @item I386_DR_LOW_SET_CONTROL (@var{val})
567 Set the Debug Control (DR7) register to the value @var{val}.
569 @findex I386_DR_LOW_SET_ADDR
570 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
571 Put the address @var{addr} into the debug register number @var{idx}.
573 @findex I386_DR_LOW_RESET_ADDR
574 @item I386_DR_LOW_RESET_ADDR (@var{idx})
575 Reset (i.e.@: zero out) the address stored in the debug register
578 @findex I386_DR_LOW_GET_STATUS
579 @item I386_DR_LOW_GET_STATUS
580 Return the value of the Debug Status (DR6) register. This value is
581 used immediately after it is returned by
582 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
586 For each one of the 4 debug registers (whose indices are from 0 to 3)
587 that store addresses, a reference count is maintained by @value{GDBN},
588 to allow sharing of debug registers by several watchpoints. This
589 allows users to define several watchpoints that watch the same
590 expression, but with different conditions and/or commands, without
591 wasting debug registers which are in short supply. @value{GDBN}
592 maintains the reference counts internally, targets don't have to do
593 anything to use this feature.
595 The x86 debug registers can each watch a region that is 1, 2, or 4
596 bytes long. The ia32 architecture requires that each watched region
597 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
598 region on 4-byte boundary. However, the x86 watchpoint support in
599 @value{GDBN} can watch unaligned regions and regions larger than 4
600 bytes (up to 16 bytes) by allocating several debug registers to watch
601 a single region. This allocation of several registers per a watched
602 region is also done automatically without target code intervention.
604 The generic x86 watchpoint support provides the following API for the
605 @value{GDBN}'s application code:
608 @findex i386_region_ok_for_watchpoint
609 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
610 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
611 this function. It counts the number of debug registers required to
612 watch a given region, and returns a non-zero value if that number is
613 less than 4, the number of debug registers available to x86
616 @findex i386_stopped_data_address
617 @item i386_stopped_data_address (@var{addr_p})
619 @code{target_stopped_data_address} is set to call this function.
621 function examines the breakpoint condition bits in the DR6 Debug
622 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
623 macro, and returns the address associated with the first bit that is
626 @findex i386_stopped_by_watchpoint
627 @item i386_stopped_by_watchpoint (void)
628 The macro @code{STOPPED_BY_WATCHPOINT}
629 is set to call this function. The
630 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
631 function examines the breakpoint condition bits in the DR6 Debug
632 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
633 macro, and returns true if any bit is set. Otherwise, false is
636 @findex i386_insert_watchpoint
637 @findex i386_remove_watchpoint
638 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
639 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
640 Insert or remove a watchpoint. The macros
641 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
642 are set to call these functions. @code{i386_insert_watchpoint} first
643 looks for a debug register which is already set to watch the same
644 region for the same access types; if found, it just increments the
645 reference count of that debug register, thus implementing debug
646 register sharing between watchpoints. If no such register is found,
647 the function looks for a vacant debug register, sets its mirrored
648 value to @var{addr}, sets the mirrored value of DR7 Debug Control
649 register as appropriate for the @var{len} and @var{type} parameters,
650 and then passes the new values of the debug register and DR7 to the
651 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
652 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
653 required to cover the given region, the above process is repeated for
656 @code{i386_remove_watchpoint} does the opposite: it resets the address
657 in the mirrored value of the debug register and its read/write and
658 length bits in the mirrored value of DR7, then passes these new
659 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
660 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
661 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
662 decrements the reference count, and only calls
663 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
664 the count goes to zero.
666 @findex i386_insert_hw_breakpoint
667 @findex i386_remove_hw_breakpoint
668 @item i386_insert_hw_breakpoint (@var{addr}, @var{shadow}
669 @itemx i386_remove_hw_breakpoint (@var{addr}, @var{shadow})
670 These functions insert and remove hardware-assisted breakpoints. The
671 macros @code{target_insert_hw_breakpoint} and
672 @code{target_remove_hw_breakpoint} are set to call these functions.
673 These functions work like @code{i386_insert_watchpoint} and
674 @code{i386_remove_watchpoint}, respectively, except that they set up
675 the debug registers to watch instruction execution, and each
676 hardware-assisted breakpoint always requires exactly one debug
679 @findex i386_stopped_by_hwbp
680 @item i386_stopped_by_hwbp (void)
681 This function returns non-zero if the inferior has some watchpoint or
682 hardware breakpoint that triggered. It works like
683 @code{i386_stopped_data_address}, except that it doesn't record the
684 address whose watchpoint triggered.
686 @findex i386_cleanup_dregs
687 @item i386_cleanup_dregs (void)
688 This function clears all the reference counts, addresses, and control
689 bits in the mirror images of the debug registers. It doesn't affect
690 the actual debug registers in the inferior process.
697 x86 processors support setting watchpoints on I/O reads or writes.
698 However, since no target supports this (as of March 2001), and since
699 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
700 watchpoints, this feature is not yet available to @value{GDBN} running
704 x86 processors can enable watchpoints locally, for the current task
705 only, or globally, for all the tasks. For each debug register,
706 there's a bit in the DR7 Debug Control register that determines
707 whether the associated address is watched locally or globally. The
708 current implementation of x86 watchpoint support in @value{GDBN}
709 always sets watchpoints to be locally enabled, since global
710 watchpoints might interfere with the underlying OS and are probably
711 unavailable in many platforms.
714 @section Observing changes in @value{GDBN} internals
715 @cindex observer pattern interface
716 @cindex notifications about changes in internals
718 In order to function properly, several modules need to be notified when
719 some changes occur in the @value{GDBN} internals. Traditionally, these
720 modules have relied on several paradigms, the most common ones being
721 hooks and gdb-events. Unfortunately, none of these paradigms was
722 versatile enough to become the standard notification mechanism in
723 @value{GDBN}. The fact that they only supported one ``client'' was also
726 A new paradigm, based on the Observer pattern of the @cite{Design
727 Patterns} book, has therefore been implemented. The goal was to provide
728 a new interface overcoming the issues with the notification mechanisms
729 previously available. This new interface needed to be strongly typed,
730 easy to extend, and versatile enough to be used as the standard
731 interface when adding new notifications.
733 See @ref{GDB Observers} for a brief description of the observers
734 currently implemented in GDB. The rationale for the current
735 implementation is also briefly discussed.
739 @chapter User Interface
741 @value{GDBN} has several user interfaces. Although the command-line interface
742 is the most common and most familiar, there are others.
744 @section Command Interpreter
746 @cindex command interpreter
748 The command interpreter in @value{GDBN} is fairly simple. It is designed to
749 allow for the set of commands to be augmented dynamically, and also
750 has a recursive subcommand capability, where the first argument to
751 a command may itself direct a lookup on a different command list.
753 For instance, the @samp{set} command just starts a lookup on the
754 @code{setlist} command list, while @samp{set thread} recurses
755 to the @code{set_thread_cmd_list}.
759 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
760 the main command list, and should be used for those commands. The usual
761 place to add commands is in the @code{_initialize_@var{xyz}} routines at
762 the ends of most source files.
764 @findex add_setshow_cmd
765 @findex add_setshow_cmd_full
766 To add paired @samp{set} and @samp{show} commands, use
767 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
768 a slightly simpler interface which is useful when you don't need to
769 further modify the new command structures, while the latter returns
770 the new command structures for manipulation.
772 @cindex deprecating commands
773 @findex deprecate_cmd
774 Before removing commands from the command set it is a good idea to
775 deprecate them for some time. Use @code{deprecate_cmd} on commands or
776 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
777 @code{struct cmd_list_element} as it's first argument. You can use the
778 return value from @code{add_com} or @code{add_cmd} to deprecate the
779 command immediately after it is created.
781 The first time a command is used the user will be warned and offered a
782 replacement (if one exists). Note that the replacement string passed to
783 @code{deprecate_cmd} should be the full name of the command, i.e. the
784 entire string the user should type at the command line.
786 @section UI-Independent Output---the @code{ui_out} Functions
787 @c This section is based on the documentation written by Fernando
788 @c Nasser <fnasser@redhat.com>.
790 @cindex @code{ui_out} functions
791 The @code{ui_out} functions present an abstraction level for the
792 @value{GDBN} output code. They hide the specifics of different user
793 interfaces supported by @value{GDBN}, and thus free the programmer
794 from the need to write several versions of the same code, one each for
795 every UI, to produce output.
797 @subsection Overview and Terminology
799 In general, execution of each @value{GDBN} command produces some sort
800 of output, and can even generate an input request.
802 Output can be generated for the following purposes:
806 to display a @emph{result} of an operation;
809 to convey @emph{info} or produce side-effects of a requested
813 to provide a @emph{notification} of an asynchronous event (including
814 progress indication of a prolonged asynchronous operation);
817 to display @emph{error messages} (including warnings);
820 to show @emph{debug data};
823 to @emph{query} or prompt a user for input (a special case).
827 This section mainly concentrates on how to build result output,
828 although some of it also applies to other kinds of output.
830 Generation of output that displays the results of an operation
831 involves one or more of the following:
835 output of the actual data
838 formatting the output as appropriate for console output, to make it
839 easily readable by humans
842 machine oriented formatting--a more terse formatting to allow for easy
843 parsing by programs which read @value{GDBN}'s output
846 annotation, whose purpose is to help legacy GUIs to identify interesting
850 The @code{ui_out} routines take care of the first three aspects.
851 Annotations are provided by separate annotation routines. Note that use
852 of annotations for an interface between a GUI and @value{GDBN} is
855 Output can be in the form of a single item, which we call a @dfn{field};
856 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
857 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
858 header and a body. In a BNF-like form:
861 @item <table> @expansion{}
862 @code{<header> <body>}
863 @item <header> @expansion{}
864 @code{@{ <column> @}}
865 @item <column> @expansion{}
866 @code{<width> <alignment> <title>}
867 @item <body> @expansion{}
872 @subsection General Conventions
874 Most @code{ui_out} routines are of type @code{void}, the exceptions are
875 @code{ui_out_stream_new} (which returns a pointer to the newly created
876 object) and the @code{make_cleanup} routines.
878 The first parameter is always the @code{ui_out} vector object, a pointer
879 to a @code{struct ui_out}.
881 The @var{format} parameter is like in @code{printf} family of functions.
882 When it is present, there must also be a variable list of arguments
883 sufficient used to satisfy the @code{%} specifiers in the supplied
886 When a character string argument is not used in a @code{ui_out} function
887 call, a @code{NULL} pointer has to be supplied instead.
890 @subsection Table, Tuple and List Functions
892 @cindex list output functions
893 @cindex table output functions
894 @cindex tuple output functions
895 This section introduces @code{ui_out} routines for building lists,
896 tuples and tables. The routines to output the actual data items
897 (fields) are presented in the next section.
899 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
900 containing information about an object; a @dfn{list} is a sequence of
901 fields where each field describes an identical object.
903 Use the @dfn{table} functions when your output consists of a list of
904 rows (tuples) and the console output should include a heading. Use this
905 even when you are listing just one object but you still want the header.
907 @cindex nesting level in @code{ui_out} functions
908 Tables can not be nested. Tuples and lists can be nested up to a
909 maximum of five levels.
911 The overall structure of the table output code is something like this:
926 Here is the description of table-, tuple- and list-related @code{ui_out}
929 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
930 The function @code{ui_out_table_begin} marks the beginning of the output
931 of a table. It should always be called before any other @code{ui_out}
932 function for a given table. @var{nbrofcols} is the number of columns in
933 the table. @var{nr_rows} is the number of rows in the table.
934 @var{tblid} is an optional string identifying the table. The string
935 pointed to by @var{tblid} is copied by the implementation of
936 @code{ui_out_table_begin}, so the application can free the string if it
939 The companion function @code{ui_out_table_end}, described below, marks
940 the end of the table's output.
943 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
944 @code{ui_out_table_header} provides the header information for a single
945 table column. You call this function several times, one each for every
946 column of the table, after @code{ui_out_table_begin}, but before
947 @code{ui_out_table_body}.
949 The value of @var{width} gives the column width in characters. The
950 value of @var{alignment} is one of @code{left}, @code{center}, and
951 @code{right}, and it specifies how to align the header: left-justify,
952 center, or right-justify it. @var{colhdr} points to a string that
953 specifies the column header; the implementation copies that string, so
954 column header strings in @code{malloc}ed storage can be freed after the
958 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
959 This function delimits the table header from the table body.
962 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
963 This function signals the end of a table's output. It should be called
964 after the table body has been produced by the list and field output
967 There should be exactly one call to @code{ui_out_table_end} for each
968 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
969 will signal an internal error.
972 The output of the tuples that represent the table rows must follow the
973 call to @code{ui_out_table_body} and precede the call to
974 @code{ui_out_table_end}. You build a tuple by calling
975 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
976 calls to functions which actually output fields between them.
978 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
979 This function marks the beginning of a tuple output. @var{id} points
980 to an optional string that identifies the tuple; it is copied by the
981 implementation, and so strings in @code{malloc}ed storage can be freed
985 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
986 This function signals an end of a tuple output. There should be exactly
987 one call to @code{ui_out_tuple_end} for each call to
988 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
992 @deftypefun struct cleanup *make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
993 This function first opens the tuple and then establishes a cleanup
994 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
995 and correct implementation of the non-portable@footnote{The function
996 cast is not portable ISO C.} code sequence:
998 struct cleanup *old_cleanup;
999 ui_out_tuple_begin (uiout, "...");
1000 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
1005 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
1006 This function marks the beginning of a list output. @var{id} points to
1007 an optional string that identifies the list; it is copied by the
1008 implementation, and so strings in @code{malloc}ed storage can be freed
1012 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
1013 This function signals an end of a list output. There should be exactly
1014 one call to @code{ui_out_list_end} for each call to
1015 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1019 @deftypefun struct cleanup *make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1020 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1021 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
1022 that will close the list.list.
1025 @subsection Item Output Functions
1027 @cindex item output functions
1028 @cindex field output functions
1030 The functions described below produce output for the actual data
1031 items, or fields, which contain information about the object.
1033 Choose the appropriate function accordingly to your particular needs.
1035 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1036 This is the most general output function. It produces the
1037 representation of the data in the variable-length argument list
1038 according to formatting specifications in @var{format}, a
1039 @code{printf}-like format string. The optional argument @var{fldname}
1040 supplies the name of the field. The data items themselves are
1041 supplied as additional arguments after @var{format}.
1043 This generic function should be used only when it is not possible to
1044 use one of the specialized versions (see below).
1047 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1048 This function outputs a value of an @code{int} variable. It uses the
1049 @code{"%d"} output conversion specification. @var{fldname} specifies
1050 the name of the field.
1053 @deftypefun void ui_out_field_fmt_int (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{fldname}, int @var{value})
1054 This function outputs a value of an @code{int} variable. It differs from
1055 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1056 @var{fldname} specifies
1057 the name of the field.
1060 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1061 This function outputs an address.
1064 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1065 This function outputs a string using the @code{"%s"} conversion
1069 Sometimes, there's a need to compose your output piece by piece using
1070 functions that operate on a stream, such as @code{value_print} or
1071 @code{fprintf_symbol_filtered}. These functions accept an argument of
1072 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1073 used to store the data stream used for the output. When you use one
1074 of these functions, you need a way to pass their results stored in a
1075 @code{ui_file} object to the @code{ui_out} functions. To this end,
1076 you first create a @code{ui_stream} object by calling
1077 @code{ui_out_stream_new}, pass the @code{stream} member of that
1078 @code{ui_stream} object to @code{value_print} and similar functions,
1079 and finally call @code{ui_out_field_stream} to output the field you
1080 constructed. When the @code{ui_stream} object is no longer needed,
1081 you should destroy it and free its memory by calling
1082 @code{ui_out_stream_delete}.
1084 @deftypefun struct ui_stream *ui_out_stream_new (struct ui_out *@var{uiout})
1085 This function creates a new @code{ui_stream} object which uses the
1086 same output methods as the @code{ui_out} object whose pointer is
1087 passed in @var{uiout}. It returns a pointer to the newly created
1088 @code{ui_stream} object.
1091 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1092 This functions destroys a @code{ui_stream} object specified by
1096 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1097 This function consumes all the data accumulated in
1098 @code{streambuf->stream} and outputs it like
1099 @code{ui_out_field_string} does. After a call to
1100 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1101 the stream is still valid and may be used for producing more fields.
1104 @strong{Important:} If there is any chance that your code could bail
1105 out before completing output generation and reaching the point where
1106 @code{ui_out_stream_delete} is called, it is necessary to set up a
1107 cleanup, to avoid leaking memory and other resources. Here's a
1108 skeleton code to do that:
1111 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1112 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1117 If the function already has the old cleanup chain set (for other kinds
1118 of cleanups), you just have to add your cleanup to it:
1121 mybuf = ui_out_stream_new (uiout);
1122 make_cleanup (ui_out_stream_delete, mybuf);
1125 Note that with cleanups in place, you should not call
1126 @code{ui_out_stream_delete} directly, or you would attempt to free the
1129 @subsection Utility Output Functions
1131 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1132 This function skips a field in a table. Use it if you have to leave
1133 an empty field without disrupting the table alignment. The argument
1134 @var{fldname} specifies a name for the (missing) filed.
1137 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1138 This function outputs the text in @var{string} in a way that makes it
1139 easy to be read by humans. For example, the console implementation of
1140 this method filters the text through a built-in pager, to prevent it
1141 from scrolling off the visible portion of the screen.
1143 Use this function for printing relatively long chunks of text around
1144 the actual field data: the text it produces is not aligned according
1145 to the table's format. Use @code{ui_out_field_string} to output a
1146 string field, and use @code{ui_out_message}, described below, to
1147 output short messages.
1150 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1151 This function outputs @var{nspaces} spaces. It is handy to align the
1152 text produced by @code{ui_out_text} with the rest of the table or
1156 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1157 This function produces a formatted message, provided that the current
1158 verbosity level is at least as large as given by @var{verbosity}. The
1159 current verbosity level is specified by the user with the @samp{set
1160 verbositylevel} command.@footnote{As of this writing (April 2001),
1161 setting verbosity level is not yet implemented, and is always returned
1162 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1163 argument more than zero will cause the message to never be printed.}
1166 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1167 This function gives the console output filter (a paging filter) a hint
1168 of where to break lines which are too long. Ignored for all other
1169 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1170 be printed to indent the wrapped text on the next line; it must remain
1171 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1172 explicit newline is produced by one of the other functions. If
1173 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1176 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1177 This function flushes whatever output has been accumulated so far, if
1178 the UI buffers output.
1182 @subsection Examples of Use of @code{ui_out} functions
1184 @cindex using @code{ui_out} functions
1185 @cindex @code{ui_out} functions, usage examples
1186 This section gives some practical examples of using the @code{ui_out}
1187 functions to generalize the old console-oriented code in
1188 @value{GDBN}. The examples all come from functions defined on the
1189 @file{breakpoints.c} file.
1191 This example, from the @code{breakpoint_1} function, shows how to
1194 The original code was:
1197 if (!found_a_breakpoint++)
1199 annotate_breakpoints_headers ();
1202 printf_filtered ("Num ");
1204 printf_filtered ("Type ");
1206 printf_filtered ("Disp ");
1208 printf_filtered ("Enb ");
1212 printf_filtered ("Address ");
1215 printf_filtered ("What\n");
1217 annotate_breakpoints_table ();
1221 Here's the new version:
1224 nr_printable_breakpoints = @dots{};
1227 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1229 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1231 if (nr_printable_breakpoints > 0)
1232 annotate_breakpoints_headers ();
1233 if (nr_printable_breakpoints > 0)
1235 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1236 if (nr_printable_breakpoints > 0)
1238 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1239 if (nr_printable_breakpoints > 0)
1241 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1242 if (nr_printable_breakpoints > 0)
1244 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1247 if (nr_printable_breakpoints > 0)
1249 if (TARGET_ADDR_BIT <= 32)
1250 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1252 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1254 if (nr_printable_breakpoints > 0)
1256 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1257 ui_out_table_body (uiout);
1258 if (nr_printable_breakpoints > 0)
1259 annotate_breakpoints_table ();
1262 This example, from the @code{print_one_breakpoint} function, shows how
1263 to produce the actual data for the table whose structure was defined
1264 in the above example. The original code was:
1269 printf_filtered ("%-3d ", b->number);
1271 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1272 || ((int) b->type != bptypes[(int) b->type].type))
1273 internal_error ("bptypes table does not describe type #%d.",
1275 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1277 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1279 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1283 This is the new version:
1287 ui_out_tuple_begin (uiout, "bkpt");
1289 ui_out_field_int (uiout, "number", b->number);
1291 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1292 || ((int) b->type != bptypes[(int) b->type].type))
1293 internal_error ("bptypes table does not describe type #%d.",
1295 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1297 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1299 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1303 This example, also from @code{print_one_breakpoint}, shows how to
1304 produce a complicated output field using the @code{print_expression}
1305 functions which requires a stream to be passed. It also shows how to
1306 automate stream destruction with cleanups. The original code was:
1310 print_expression (b->exp, gdb_stdout);
1316 struct ui_stream *stb = ui_out_stream_new (uiout);
1317 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1320 print_expression (b->exp, stb->stream);
1321 ui_out_field_stream (uiout, "what", local_stream);
1324 This example, also from @code{print_one_breakpoint}, shows how to use
1325 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1330 if (b->dll_pathname == NULL)
1331 printf_filtered ("<any library> ");
1333 printf_filtered ("library \"%s\" ", b->dll_pathname);
1340 if (b->dll_pathname == NULL)
1342 ui_out_field_string (uiout, "what", "<any library>");
1343 ui_out_spaces (uiout, 1);
1347 ui_out_text (uiout, "library \"");
1348 ui_out_field_string (uiout, "what", b->dll_pathname);
1349 ui_out_text (uiout, "\" ");
1353 The following example from @code{print_one_breakpoint} shows how to
1354 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1359 if (b->forked_inferior_pid != 0)
1360 printf_filtered ("process %d ", b->forked_inferior_pid);
1367 if (b->forked_inferior_pid != 0)
1369 ui_out_text (uiout, "process ");
1370 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1371 ui_out_spaces (uiout, 1);
1375 Here's an example of using @code{ui_out_field_string}. The original
1380 if (b->exec_pathname != NULL)
1381 printf_filtered ("program \"%s\" ", b->exec_pathname);
1388 if (b->exec_pathname != NULL)
1390 ui_out_text (uiout, "program \"");
1391 ui_out_field_string (uiout, "what", b->exec_pathname);
1392 ui_out_text (uiout, "\" ");
1396 Finally, here's an example of printing an address. The original code:
1400 printf_filtered ("%s ",
1401 hex_string_custom ((unsigned long) b->address, 8));
1408 ui_out_field_core_addr (uiout, "Address", b->address);
1412 @section Console Printing
1421 @cindex @code{libgdb}
1422 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1423 to provide an API to @value{GDBN}'s functionality.
1426 @cindex @code{libgdb}
1427 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1428 better able to support graphical and other environments.
1430 Since @code{libgdb} development is on-going, its architecture is still
1431 evolving. The following components have so far been identified:
1435 Observer - @file{gdb-events.h}.
1437 Builder - @file{ui-out.h}
1439 Event Loop - @file{event-loop.h}
1441 Library - @file{gdb.h}
1444 The model that ties these components together is described below.
1446 @section The @code{libgdb} Model
1448 A client of @code{libgdb} interacts with the library in two ways.
1452 As an observer (using @file{gdb-events}) receiving notifications from
1453 @code{libgdb} of any internal state changes (break point changes, run
1456 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1457 obtain various status values from @value{GDBN}.
1460 Since @code{libgdb} could have multiple clients (e.g. a GUI supporting
1461 the existing @value{GDBN} CLI), those clients must co-operate when
1462 controlling @code{libgdb}. In particular, a client must ensure that
1463 @code{libgdb} is idle (i.e. no other client is using @code{libgdb})
1464 before responding to a @file{gdb-event} by making a query.
1466 @section CLI support
1468 At present @value{GDBN}'s CLI is very much entangled in with the core of
1469 @code{libgdb}. Consequently, a client wishing to include the CLI in
1470 their interface needs to carefully co-ordinate its own and the CLI's
1473 It is suggested that the client set @code{libgdb} up to be bi-modal
1474 (alternate between CLI and client query modes). The notes below sketch
1479 The client registers itself as an observer of @code{libgdb}.
1481 The client create and install @code{cli-out} builder using its own
1482 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1483 @code{gdb_stdout} streams.
1485 The client creates a separate custom @code{ui-out} builder that is only
1486 used while making direct queries to @code{libgdb}.
1489 When the client receives input intended for the CLI, it simply passes it
1490 along. Since the @code{cli-out} builder is installed by default, all
1491 the CLI output in response to that command is routed (pronounced rooted)
1492 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1493 At the same time, the client is kept abreast of internal changes by
1494 virtue of being a @code{libgdb} observer.
1496 The only restriction on the client is that it must wait until
1497 @code{libgdb} becomes idle before initiating any queries (using the
1498 client's custom builder).
1500 @section @code{libgdb} components
1502 @subheading Observer - @file{gdb-events.h}
1503 @file{gdb-events} provides the client with a very raw mechanism that can
1504 be used to implement an observer. At present it only allows for one
1505 observer and that observer must, internally, handle the need to delay
1506 the processing of any event notifications until after @code{libgdb} has
1507 finished the current command.
1509 @subheading Builder - @file{ui-out.h}
1510 @file{ui-out} provides the infrastructure necessary for a client to
1511 create a builder. That builder is then passed down to @code{libgdb}
1512 when doing any queries.
1514 @subheading Event Loop - @file{event-loop.h}
1515 @c There could be an entire section on the event-loop
1516 @file{event-loop}, currently non-re-entrant, provides a simple event
1517 loop. A client would need to either plug its self into this loop or,
1518 implement a new event-loop that GDB would use.
1520 The event-loop will eventually be made re-entrant. This is so that
1521 @value{GDBN} can better handle the problem of some commands blocking
1522 instead of returning.
1524 @subheading Library - @file{gdb.h}
1525 @file{libgdb} is the most obvious component of this system. It provides
1526 the query interface. Each function is parameterized by a @code{ui-out}
1527 builder. The result of the query is constructed using that builder
1528 before the query function returns.
1530 @node Symbol Handling
1532 @chapter Symbol Handling
1534 Symbols are a key part of @value{GDBN}'s operation. Symbols include variables,
1535 functions, and types.
1537 @section Symbol Reading
1539 @cindex symbol reading
1540 @cindex reading of symbols
1541 @cindex symbol files
1542 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
1543 file is the file containing the program which @value{GDBN} is
1544 debugging. @value{GDBN} can be directed to use a different file for
1545 symbols (with the @samp{symbol-file} command), and it can also read
1546 more symbols via the @samp{add-file} and @samp{load} commands, or while
1547 reading symbols from shared libraries.
1549 @findex find_sym_fns
1550 Symbol files are initially opened by code in @file{symfile.c} using
1551 the BFD library (@pxref{Support Libraries}). BFD identifies the type
1552 of the file by examining its header. @code{find_sym_fns} then uses
1553 this identification to locate a set of symbol-reading functions.
1555 @findex add_symtab_fns
1556 @cindex @code{sym_fns} structure
1557 @cindex adding a symbol-reading module
1558 Symbol-reading modules identify themselves to @value{GDBN} by calling
1559 @code{add_symtab_fns} during their module initialization. The argument
1560 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
1561 name (or name prefix) of the symbol format, the length of the prefix,
1562 and pointers to four functions. These functions are called at various
1563 times to process symbol files whose identification matches the specified
1566 The functions supplied by each module are:
1569 @item @var{xyz}_symfile_init(struct sym_fns *sf)
1571 @cindex secondary symbol file
1572 Called from @code{symbol_file_add} when we are about to read a new
1573 symbol file. This function should clean up any internal state (possibly
1574 resulting from half-read previous files, for example) and prepare to
1575 read a new symbol file. Note that the symbol file which we are reading
1576 might be a new ``main'' symbol file, or might be a secondary symbol file
1577 whose symbols are being added to the existing symbol table.
1579 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
1580 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
1581 new symbol file being read. Its @code{private} field has been zeroed,
1582 and can be modified as desired. Typically, a struct of private
1583 information will be @code{malloc}'d, and a pointer to it will be placed
1584 in the @code{private} field.
1586 There is no result from @code{@var{xyz}_symfile_init}, but it can call
1587 @code{error} if it detects an unavoidable problem.
1589 @item @var{xyz}_new_init()
1591 Called from @code{symbol_file_add} when discarding existing symbols.
1592 This function needs only handle the symbol-reading module's internal
1593 state; the symbol table data structures visible to the rest of
1594 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
1595 arguments and no result. It may be called after
1596 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
1597 may be called alone if all symbols are simply being discarded.
1599 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
1601 Called from @code{symbol_file_add} to actually read the symbols from a
1602 symbol-file into a set of psymtabs or symtabs.
1604 @code{sf} points to the @code{struct sym_fns} originally passed to
1605 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
1606 the offset between the file's specified start address and its true
1607 address in memory. @code{mainline} is 1 if this is the main symbol
1608 table being read, and 0 if a secondary symbol file (e.g. shared library
1609 or dynamically loaded file) is being read.@refill
1612 In addition, if a symbol-reading module creates psymtabs when
1613 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
1614 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
1615 from any point in the @value{GDBN} symbol-handling code.
1618 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
1620 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
1621 the psymtab has not already been read in and had its @code{pst->symtab}
1622 pointer set. The argument is the psymtab to be fleshed-out into a
1623 symtab. Upon return, @code{pst->readin} should have been set to 1, and
1624 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
1625 zero if there were no symbols in that part of the symbol file.
1628 @section Partial Symbol Tables
1630 @value{GDBN} has three types of symbol tables:
1633 @cindex full symbol table
1636 Full symbol tables (@dfn{symtabs}). These contain the main
1637 information about symbols and addresses.
1641 Partial symbol tables (@dfn{psymtabs}). These contain enough
1642 information to know when to read the corresponding part of the full
1645 @cindex minimal symbol table
1648 Minimal symbol tables (@dfn{msymtabs}). These contain information
1649 gleaned from non-debugging symbols.
1652 @cindex partial symbol table
1653 This section describes partial symbol tables.
1655 A psymtab is constructed by doing a very quick pass over an executable
1656 file's debugging information. Small amounts of information are
1657 extracted---enough to identify which parts of the symbol table will
1658 need to be re-read and fully digested later, when the user needs the
1659 information. The speed of this pass causes @value{GDBN} to start up very
1660 quickly. Later, as the detailed rereading occurs, it occurs in small
1661 pieces, at various times, and the delay therefrom is mostly invisible to
1663 @c (@xref{Symbol Reading}.)
1665 The symbols that show up in a file's psymtab should be, roughly, those
1666 visible to the debugger's user when the program is not running code from
1667 that file. These include external symbols and types, static symbols and
1668 types, and @code{enum} values declared at file scope.
1670 The psymtab also contains the range of instruction addresses that the
1671 full symbol table would represent.
1673 @cindex finding a symbol
1674 @cindex symbol lookup
1675 The idea is that there are only two ways for the user (or much of the
1676 code in the debugger) to reference a symbol:
1679 @findex find_pc_function
1680 @findex find_pc_line
1682 By its address (e.g. execution stops at some address which is inside a
1683 function in this file). The address will be noticed to be in the
1684 range of this psymtab, and the full symtab will be read in.
1685 @code{find_pc_function}, @code{find_pc_line}, and other
1686 @code{find_pc_@dots{}} functions handle this.
1688 @cindex lookup_symbol
1691 (e.g. the user asks to print a variable, or set a breakpoint on a
1692 function). Global names and file-scope names will be found in the
1693 psymtab, which will cause the symtab to be pulled in. Local names will
1694 have to be qualified by a global name, or a file-scope name, in which
1695 case we will have already read in the symtab as we evaluated the
1696 qualifier. Or, a local symbol can be referenced when we are ``in'' a
1697 local scope, in which case the first case applies. @code{lookup_symbol}
1698 does most of the work here.
1701 The only reason that psymtabs exist is to cause a symtab to be read in
1702 at the right moment. Any symbol that can be elided from a psymtab,
1703 while still causing that to happen, should not appear in it. Since
1704 psymtabs don't have the idea of scope, you can't put local symbols in
1705 them anyway. Psymtabs don't have the idea of the type of a symbol,
1706 either, so types need not appear, unless they will be referenced by
1709 It is a bug for @value{GDBN} to behave one way when only a psymtab has
1710 been read, and another way if the corresponding symtab has been read
1711 in. Such bugs are typically caused by a psymtab that does not contain
1712 all the visible symbols, or which has the wrong instruction address
1715 The psymtab for a particular section of a symbol file (objfile) could be
1716 thrown away after the symtab has been read in. The symtab should always
1717 be searched before the psymtab, so the psymtab will never be used (in a
1718 bug-free environment). Currently, psymtabs are allocated on an obstack,
1719 and all the psymbols themselves are allocated in a pair of large arrays
1720 on an obstack, so there is little to be gained by trying to free them
1721 unless you want to do a lot more work.
1725 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
1727 @cindex fundamental types
1728 These are the fundamental types that @value{GDBN} uses internally. Fundamental
1729 types from the various debugging formats (stabs, ELF, etc) are mapped
1730 into one of these. They are basically a union of all fundamental types
1731 that @value{GDBN} knows about for all the languages that @value{GDBN}
1734 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
1737 Each time @value{GDBN} builds an internal type, it marks it with one
1738 of these types. The type may be a fundamental type, such as
1739 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
1740 which is a pointer to another type. Typically, several @code{FT_*}
1741 types map to one @code{TYPE_CODE_*} type, and are distinguished by
1742 other members of the type struct, such as whether the type is signed
1743 or unsigned, and how many bits it uses.
1745 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
1747 These are instances of type structs that roughly correspond to
1748 fundamental types and are created as global types for @value{GDBN} to
1749 use for various ugly historical reasons. We eventually want to
1750 eliminate these. Note for example that @code{builtin_type_int}
1751 initialized in @file{gdbtypes.c} is basically the same as a
1752 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
1753 an @code{FT_INTEGER} fundamental type. The difference is that the
1754 @code{builtin_type} is not associated with any particular objfile, and
1755 only one instance exists, while @file{c-lang.c} builds as many
1756 @code{TYPE_CODE_INT} types as needed, with each one associated with
1757 some particular objfile.
1759 @section Object File Formats
1760 @cindex object file formats
1764 @cindex @code{a.out} format
1765 The @code{a.out} format is the original file format for Unix. It
1766 consists of three sections: @code{text}, @code{data}, and @code{bss},
1767 which are for program code, initialized data, and uninitialized data,
1770 The @code{a.out} format is so simple that it doesn't have any reserved
1771 place for debugging information. (Hey, the original Unix hackers used
1772 @samp{adb}, which is a machine-language debugger!) The only debugging
1773 format for @code{a.out} is stabs, which is encoded as a set of normal
1774 symbols with distinctive attributes.
1776 The basic @code{a.out} reader is in @file{dbxread.c}.
1781 The COFF format was introduced with System V Release 3 (SVR3) Unix.
1782 COFF files may have multiple sections, each prefixed by a header. The
1783 number of sections is limited.
1785 The COFF specification includes support for debugging. Although this
1786 was a step forward, the debugging information was woefully limited. For
1787 instance, it was not possible to represent code that came from an
1790 The COFF reader is in @file{coffread.c}.
1794 @cindex ECOFF format
1795 ECOFF is an extended COFF originally introduced for Mips and Alpha
1798 The basic ECOFF reader is in @file{mipsread.c}.
1802 @cindex XCOFF format
1803 The IBM RS/6000 running AIX uses an object file format called XCOFF.
1804 The COFF sections, symbols, and line numbers are used, but debugging
1805 symbols are @code{dbx}-style stabs whose strings are located in the
1806 @code{.debug} section (rather than the string table). For more
1807 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
1809 The shared library scheme has a clean interface for figuring out what
1810 shared libraries are in use, but the catch is that everything which
1811 refers to addresses (symbol tables and breakpoints at least) needs to be
1812 relocated for both shared libraries and the main executable. At least
1813 using the standard mechanism this can only be done once the program has
1814 been run (or the core file has been read).
1818 @cindex PE-COFF format
1819 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
1820 executables. PE is basically COFF with additional headers.
1822 While BFD includes special PE support, @value{GDBN} needs only the basic
1828 The ELF format came with System V Release 4 (SVR4) Unix. ELF is similar
1829 to COFF in being organized into a number of sections, but it removes
1830 many of COFF's limitations.
1832 The basic ELF reader is in @file{elfread.c}.
1837 SOM is HP's object file and debug format (not to be confused with IBM's
1838 SOM, which is a cross-language ABI).
1840 The SOM reader is in @file{hpread.c}.
1842 @subsection Other File Formats
1844 @cindex Netware Loadable Module format
1845 Other file formats that have been supported by @value{GDBN} include Netware
1846 Loadable Modules (@file{nlmread.c}).
1848 @section Debugging File Formats
1850 This section describes characteristics of debugging information that
1851 are independent of the object file format.
1855 @cindex stabs debugging info
1856 @code{stabs} started out as special symbols within the @code{a.out}
1857 format. Since then, it has been encapsulated into other file
1858 formats, such as COFF and ELF.
1860 While @file{dbxread.c} does some of the basic stab processing,
1861 including for encapsulated versions, @file{stabsread.c} does
1866 @cindex COFF debugging info
1867 The basic COFF definition includes debugging information. The level
1868 of support is minimal and non-extensible, and is not often used.
1870 @subsection Mips debug (Third Eye)
1872 @cindex ECOFF debugging info
1873 ECOFF includes a definition of a special debug format.
1875 The file @file{mdebugread.c} implements reading for this format.
1879 @cindex DWARF 1 debugging info
1880 DWARF 1 is a debugging format that was originally designed to be
1881 used with ELF in SVR4 systems.
1886 @c If defined, these are the producer strings in a DWARF 1 file. All of
1887 @c these have reasonable defaults already.
1889 The DWARF 1 reader is in @file{dwarfread.c}.
1893 @cindex DWARF 2 debugging info
1894 DWARF 2 is an improved but incompatible version of DWARF 1.
1896 The DWARF 2 reader is in @file{dwarf2read.c}.
1900 @cindex SOM debugging info
1901 Like COFF, the SOM definition includes debugging information.
1903 @section Adding a New Symbol Reader to @value{GDBN}
1905 @cindex adding debugging info reader
1906 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
1907 there is probably little to be done.
1909 If you need to add a new object file format, you must first add it to
1910 BFD. This is beyond the scope of this document.
1912 You must then arrange for the BFD code to provide access to the
1913 debugging symbols. Generally @value{GDBN} will have to call swapping routines
1914 from BFD and a few other BFD internal routines to locate the debugging
1915 information. As much as possible, @value{GDBN} should not depend on the BFD
1916 internal data structures.
1918 For some targets (e.g., COFF), there is a special transfer vector used
1919 to call swapping routines, since the external data structures on various
1920 platforms have different sizes and layouts. Specialized routines that
1921 will only ever be implemented by one object file format may be called
1922 directly. This interface should be described in a file
1923 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
1926 @node Language Support
1928 @chapter Language Support
1930 @cindex language support
1931 @value{GDBN}'s language support is mainly driven by the symbol reader,
1932 although it is possible for the user to set the source language
1935 @value{GDBN} chooses the source language by looking at the extension
1936 of the file recorded in the debug info; @file{.c} means C, @file{.f}
1937 means Fortran, etc. It may also use a special-purpose language
1938 identifier if the debug format supports it, like with DWARF.
1940 @section Adding a Source Language to @value{GDBN}
1942 @cindex adding source language
1943 To add other languages to @value{GDBN}'s expression parser, follow the
1947 @item Create the expression parser.
1949 @cindex expression parser
1950 This should reside in a file @file{@var{lang}-exp.y}. Routines for
1951 building parsed expressions into a @code{union exp_element} list are in
1954 @cindex language parser
1955 Since we can't depend upon everyone having Bison, and YACC produces
1956 parsers that define a bunch of global names, the following lines
1957 @strong{must} be included at the top of the YACC parser, to prevent the
1958 various parsers from defining the same global names:
1961 #define yyparse @var{lang}_parse
1962 #define yylex @var{lang}_lex
1963 #define yyerror @var{lang}_error
1964 #define yylval @var{lang}_lval
1965 #define yychar @var{lang}_char
1966 #define yydebug @var{lang}_debug
1967 #define yypact @var{lang}_pact
1968 #define yyr1 @var{lang}_r1
1969 #define yyr2 @var{lang}_r2
1970 #define yydef @var{lang}_def
1971 #define yychk @var{lang}_chk
1972 #define yypgo @var{lang}_pgo
1973 #define yyact @var{lang}_act
1974 #define yyexca @var{lang}_exca
1975 #define yyerrflag @var{lang}_errflag
1976 #define yynerrs @var{lang}_nerrs
1979 At the bottom of your parser, define a @code{struct language_defn} and
1980 initialize it with the right values for your language. Define an
1981 @code{initialize_@var{lang}} routine and have it call
1982 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
1983 that your language exists. You'll need some other supporting variables
1984 and functions, which will be used via pointers from your
1985 @code{@var{lang}_language_defn}. See the declaration of @code{struct
1986 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
1987 for more information.
1989 @item Add any evaluation routines, if necessary
1991 @cindex expression evaluation routines
1992 @findex evaluate_subexp
1993 @findex prefixify_subexp
1994 @findex length_of_subexp
1995 If you need new opcodes (that represent the operations of the language),
1996 add them to the enumerated type in @file{expression.h}. Add support
1997 code for these operations in the @code{evaluate_subexp} function
1998 defined in the file @file{eval.c}. Add cases
1999 for new opcodes in two functions from @file{parse.c}:
2000 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
2001 the number of @code{exp_element}s that a given operation takes up.
2003 @item Update some existing code
2005 Add an enumerated identifier for your language to the enumerated type
2006 @code{enum language} in @file{defs.h}.
2008 Update the routines in @file{language.c} so your language is included.
2009 These routines include type predicates and such, which (in some cases)
2010 are language dependent. If your language does not appear in the switch
2011 statement, an error is reported.
2013 @vindex current_language
2014 Also included in @file{language.c} is the code that updates the variable
2015 @code{current_language}, and the routines that translate the
2016 @code{language_@var{lang}} enumerated identifier into a printable
2019 @findex _initialize_language
2020 Update the function @code{_initialize_language} to include your
2021 language. This function picks the default language upon startup, so is
2022 dependent upon which languages that @value{GDBN} is built for.
2024 @findex allocate_symtab
2025 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2026 code so that the language of each symtab (source file) is set properly.
2027 This is used to determine the language to use at each stack frame level.
2028 Currently, the language is set based upon the extension of the source
2029 file. If the language can be better inferred from the symbol
2030 information, please set the language of the symtab in the symbol-reading
2033 @findex print_subexp
2034 @findex op_print_tab
2035 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2036 expression opcodes you have added to @file{expression.h}. Also, add the
2037 printed representations of your operators to @code{op_print_tab}.
2039 @item Add a place of call
2042 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2043 @code{parse_exp_1} (defined in @file{parse.c}).
2045 @item Use macros to trim code
2047 @cindex trimming language-dependent code
2048 The user has the option of building @value{GDBN} for some or all of the
2049 languages. If the user decides to build @value{GDBN} for the language
2050 @var{lang}, then every file dependent on @file{language.h} will have the
2051 macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to
2052 leave out large routines that the user won't need if he or she is not
2053 using your language.
2055 Note that you do not need to do this in your YACC parser, since if @value{GDBN}
2056 is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the
2057 compiled form of your parser) is not linked into @value{GDBN} at all.
2059 See the file @file{configure.in} for how @value{GDBN} is configured
2060 for different languages.
2062 @item Edit @file{Makefile.in}
2064 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2065 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2066 not get linked in, or, worse yet, it may not get @code{tar}red into the
2071 @node Host Definition
2073 @chapter Host Definition
2075 With the advent of Autoconf, it's rarely necessary to have host
2076 definition machinery anymore. The following information is provided,
2077 mainly, as an historical reference.
2079 @section Adding a New Host
2081 @cindex adding a new host
2082 @cindex host, adding
2083 @value{GDBN}'s host configuration support normally happens via Autoconf.
2084 New host-specific definitions should not be needed. Older hosts
2085 @value{GDBN} still use the host-specific definitions and files listed
2086 below, but these mostly exist for historical reasons, and will
2087 eventually disappear.
2090 @item gdb/config/@var{arch}/@var{xyz}.mh
2091 This file once contained both host and native configuration information
2092 (@pxref{Native Debugging}) for the machine @var{xyz}. The host
2093 configuration information is now handed by Autoconf.
2095 Host configuration information included a definition of
2096 @code{XM_FILE=xm-@var{xyz}.h} and possibly definitions for @code{CC},
2097 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2098 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2100 New host only configurations do not need this file.
2102 @item gdb/config/@var{arch}/xm-@var{xyz}.h
2103 This file once contained definitions and includes required when hosting
2104 gdb on machine @var{xyz}. Those definitions and includes are now
2105 handled by Autoconf.
2107 New host and native configurations do not need this file.
2109 @emph{Maintainer's note: Some hosts continue to use the @file{xm-xyz.h}
2110 file to define the macros @var{HOST_FLOAT_FORMAT},
2111 @var{HOST_DOUBLE_FORMAT} and @var{HOST_LONG_DOUBLE_FORMAT}. That code
2112 also needs to be replaced with either an Autoconf or run-time test.}
2116 @subheading Generic Host Support Files
2118 @cindex generic host support
2119 There are some ``generic'' versions of routines that can be used by
2120 various systems. These can be customized in various ways by macros
2121 defined in your @file{xm-@var{xyz}.h} file. If these routines work for
2122 the @var{xyz} host, you can just include the generic file's name (with
2123 @samp{.o}, not @samp{.c}) in @code{XDEPFILES}.
2125 Otherwise, if your machine needs custom support routines, you will need
2126 to write routines that perform the same functions as the generic file.
2127 Put them into @code{@var{xyz}-xdep.c}, and put @code{@var{xyz}-xdep.o}
2128 into @code{XDEPFILES}.
2131 @cindex remote debugging support
2132 @cindex serial line support
2134 This contains serial line support for Unix systems. This is always
2135 included, via the makefile variable @code{SER_HARDWIRE}; override this
2136 variable in the @file{.mh} file to avoid it.
2139 This contains serial line support for 32-bit programs running under DOS,
2140 using the DJGPP (a.k.a.@: GO32) execution environment.
2142 @cindex TCP remote support
2144 This contains generic TCP support using sockets.
2147 @section Host Conditionals
2149 When @value{GDBN} is configured and compiled, various macros are
2150 defined or left undefined, to control compilation based on the
2151 attributes of the host system. These macros and their meanings (or if
2152 the meaning is not documented here, then one of the source files where
2153 they are used is indicated) are:
2156 @item @value{GDBN}INIT_FILENAME
2157 The default name of @value{GDBN}'s initialization file (normally
2161 This macro is deprecated.
2163 @item SIGWINCH_HANDLER
2164 If your host defines @code{SIGWINCH}, you can define this to be the name
2165 of a function to be called if @code{SIGWINCH} is received.
2167 @item SIGWINCH_HANDLER_BODY
2168 Define this to expand into code that will define the function named by
2169 the expansion of @code{SIGWINCH_HANDLER}.
2171 @item ALIGN_STACK_ON_STARTUP
2172 @cindex stack alignment
2173 Define this if your system is of a sort that will crash in
2174 @code{tgetent} if the stack happens not to be longword-aligned when
2175 @code{main} is called. This is a rare situation, but is known to occur
2176 on several different types of systems.
2178 @item CRLF_SOURCE_FILES
2179 @cindex DOS text files
2180 Define this if host files use @code{\r\n} rather than @code{\n} as a
2181 line terminator. This will cause source file listings to omit @code{\r}
2182 characters when printing and it will allow @code{\r\n} line endings of files
2183 which are ``sourced'' by gdb. It must be possible to open files in binary
2184 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2186 @item DEFAULT_PROMPT
2188 The default value of the prompt string (normally @code{"(gdb) "}).
2191 @cindex terminal device
2192 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2195 Define this if binary files are opened the same way as text files.
2199 In some cases, use the system call @code{mmap} for reading symbol
2200 tables. For some machines this allows for sharing and quick updates.
2203 Define this if the host system has @code{termio.h}.
2210 Values for host-side constants.
2213 Substitute for isatty, if not available.
2216 This is the longest integer type available on the host. If not defined,
2217 it will default to @code{long long} or @code{long}, depending on
2218 @code{CC_HAS_LONG_LONG}.
2220 @item CC_HAS_LONG_LONG
2221 @cindex @code{long long} data type
2222 Define this if the host C compiler supports @code{long long}. This is set
2223 by the @code{configure} script.
2225 @item PRINTF_HAS_LONG_LONG
2226 Define this if the host can handle printing of long long integers via
2227 the printf format conversion specifier @code{ll}. This is set by the
2228 @code{configure} script.
2230 @item HAVE_LONG_DOUBLE
2231 Define this if the host C compiler supports @code{long double}. This is
2232 set by the @code{configure} script.
2234 @item PRINTF_HAS_LONG_DOUBLE
2235 Define this if the host can handle printing of long double float-point
2236 numbers via the printf format conversion specifier @code{Lg}. This is
2237 set by the @code{configure} script.
2239 @item SCANF_HAS_LONG_DOUBLE
2240 Define this if the host can handle the parsing of long double
2241 float-point numbers via the scanf format conversion specifier
2242 @code{Lg}. This is set by the @code{configure} script.
2244 @item LSEEK_NOT_LINEAR
2245 Define this if @code{lseek (n)} does not necessarily move to byte number
2246 @code{n} in the file. This is only used when reading source files. It
2247 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2250 This macro is used as the argument to @code{lseek} (or, most commonly,
2251 @code{bfd_seek}). FIXME, should be replaced by SEEK_SET instead,
2252 which is the POSIX equivalent.
2255 If defined, this should be one or more tokens, such as @code{volatile},
2256 that can be used in both the declaration and definition of functions to
2257 indicate that they never return. The default is already set correctly
2258 if compiling with GCC. This will almost never need to be defined.
2261 If defined, this should be one or more tokens, such as
2262 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2263 of functions to indicate that they never return. The default is already
2264 set correctly if compiling with GCC. This will almost never need to be
2269 Define these to appropriate value for the system @code{lseek}, if not already
2273 This is the signal for stopping @value{GDBN}. Defaults to
2274 @code{SIGTSTP}. (Only redefined for the Convex.)
2277 Means that System V (prior to SVR4) include files are in use. (FIXME:
2278 This symbol is abused in @file{infrun.c}, @file{regex.c}, and
2279 @file{utils.c} for other things, at the moment.)
2282 Define this to help placate @code{lint} in some situations.
2285 Define this to override the defaults of @code{__volatile__} or
2290 @node Target Architecture Definition
2292 @chapter Target Architecture Definition
2294 @cindex target architecture definition
2295 @value{GDBN}'s target architecture defines what sort of
2296 machine-language programs @value{GDBN} can work with, and how it works
2299 The target architecture object is implemented as the C structure
2300 @code{struct gdbarch *}. The structure, and its methods, are generated
2301 using the Bourne shell script @file{gdbarch.sh}.
2303 @section Operating System ABI Variant Handling
2304 @cindex OS ABI variants
2306 @value{GDBN} provides a mechanism for handling variations in OS
2307 ABIs. An OS ABI variant may have influence over any number of
2308 variables in the target architecture definition. There are two major
2309 components in the OS ABI mechanism: sniffers and handlers.
2311 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2312 (the architecture may be wildcarded) in an attempt to determine the
2313 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2314 to be @dfn{generic}, while sniffers for a specific architecture are
2315 considered to be @dfn{specific}. A match from a specific sniffer
2316 overrides a match from a generic sniffer. Multiple sniffers for an
2317 architecture/flavour may exist, in order to differentiate between two
2318 different operating systems which use the same basic file format. The
2319 OS ABI framework provides a generic sniffer for ELF-format files which
2320 examines the @code{EI_OSABI} field of the ELF header, as well as note
2321 sections known to be used by several operating systems.
2323 @cindex fine-tuning @code{gdbarch} structure
2324 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2325 selected OS ABI. There may be only one handler for a given OS ABI
2326 for each BFD architecture.
2328 The following OS ABI variants are defined in @file{osabi.h}:
2332 @findex GDB_OSABI_UNKNOWN
2333 @item GDB_OSABI_UNKNOWN
2334 The ABI of the inferior is unknown. The default @code{gdbarch}
2335 settings for the architecture will be used.
2337 @findex GDB_OSABI_SVR4
2338 @item GDB_OSABI_SVR4
2339 UNIX System V Release 4
2341 @findex GDB_OSABI_HURD
2342 @item GDB_OSABI_HURD
2343 GNU using the Hurd kernel
2345 @findex GDB_OSABI_SOLARIS
2346 @item GDB_OSABI_SOLARIS
2349 @findex GDB_OSABI_OSF1
2350 @item GDB_OSABI_OSF1
2351 OSF/1, including Digital UNIX and Compaq Tru64 UNIX
2353 @findex GDB_OSABI_LINUX
2354 @item GDB_OSABI_LINUX
2355 GNU using the Linux kernel
2357 @findex GDB_OSABI_FREEBSD_AOUT
2358 @item GDB_OSABI_FREEBSD_AOUT
2359 FreeBSD using the a.out executable format
2361 @findex GDB_OSABI_FREEBSD_ELF
2362 @item GDB_OSABI_FREEBSD_ELF
2363 FreeBSD using the ELF executable format
2365 @findex GDB_OSABI_NETBSD_AOUT
2366 @item GDB_OSABI_NETBSD_AOUT
2367 NetBSD using the a.out executable format
2369 @findex GDB_OSABI_NETBSD_ELF
2370 @item GDB_OSABI_NETBSD_ELF
2371 NetBSD using the ELF executable format
2373 @findex GDB_OSABI_WINCE
2374 @item GDB_OSABI_WINCE
2377 @findex GDB_OSABI_GO32
2378 @item GDB_OSABI_GO32
2381 @findex GDB_OSABI_NETWARE
2382 @item GDB_OSABI_NETWARE
2385 @findex GDB_OSABI_ARM_EABI_V1
2386 @item GDB_OSABI_ARM_EABI_V1
2387 ARM Embedded ABI version 1
2389 @findex GDB_OSABI_ARM_EABI_V2
2390 @item GDB_OSABI_ARM_EABI_V2
2391 ARM Embedded ABI version 2
2393 @findex GDB_OSABI_ARM_APCS
2394 @item GDB_OSABI_ARM_APCS
2395 Generic ARM Procedure Call Standard
2399 Here are the functions that make up the OS ABI framework:
2401 @deftypefun const char *gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2402 Return the name of the OS ABI corresponding to @var{osabi}.
2405 @deftypefun void gdbarch_register_osabi (enum bfd_architecture @var{arch}, unsigned long @var{machine}, enum gdb_osabi @var{osabi}, void (*@var{init_osabi})(struct gdbarch_info @var{info}, struct gdbarch *@var{gdbarch}))
2406 Register the OS ABI handler specified by @var{init_osabi} for the
2407 architecture, machine type and OS ABI specified by @var{arch},
2408 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2409 machine type, which implies the architecture's default machine type,
2413 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2414 Register the OS ABI file sniffer specified by @var{sniffer} for the
2415 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2416 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2417 be generic, and is allowed to examine @var{flavour}-flavoured files for
2421 @deftypefun enum gdb_osabi gdbarch_lookup_osabi (bfd *@var{abfd})
2422 Examine the file described by @var{abfd} to determine its OS ABI.
2423 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2427 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2428 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2429 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2430 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2431 architecture, a warning will be issued and the debugging session will continue
2432 with the defaults already established for @var{gdbarch}.
2435 @section Registers and Memory
2437 @value{GDBN}'s model of the target machine is rather simple.
2438 @value{GDBN} assumes the machine includes a bank of registers and a
2439 block of memory. Each register may have a different size.
2441 @value{GDBN} does not have a magical way to match up with the
2442 compiler's idea of which registers are which; however, it is critical
2443 that they do match up accurately. The only way to make this work is
2444 to get accurate information about the order that the compiler uses,
2445 and to reflect that in the @code{REGISTER_NAME} and related macros.
2447 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
2449 @section Pointers Are Not Always Addresses
2450 @cindex pointer representation
2451 @cindex address representation
2452 @cindex word-addressed machines
2453 @cindex separate data and code address spaces
2454 @cindex spaces, separate data and code address
2455 @cindex address spaces, separate data and code
2456 @cindex code pointers, word-addressed
2457 @cindex converting between pointers and addresses
2458 @cindex D10V addresses
2460 On almost all 32-bit architectures, the representation of a pointer is
2461 indistinguishable from the representation of some fixed-length number
2462 whose value is the byte address of the object pointed to. On such
2463 machines, the words ``pointer'' and ``address'' can be used interchangeably.
2464 However, architectures with smaller word sizes are often cramped for
2465 address space, so they may choose a pointer representation that breaks this
2466 identity, and allows a larger code address space.
2468 For example, the Renesas D10V is a 16-bit VLIW processor whose
2469 instructions are 32 bits long@footnote{Some D10V instructions are
2470 actually pairs of 16-bit sub-instructions. However, since you can't
2471 jump into the middle of such a pair, code addresses can only refer to
2472 full 32 bit instructions, which is what matters in this explanation.}.
2473 If the D10V used ordinary byte addresses to refer to code locations,
2474 then the processor would only be able to address 64kb of instructions.
2475 However, since instructions must be aligned on four-byte boundaries, the
2476 low two bits of any valid instruction's byte address are always
2477 zero---byte addresses waste two bits. So instead of byte addresses,
2478 the D10V uses word addresses---byte addresses shifted right two bits---to
2479 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
2482 However, this means that code pointers and data pointers have different
2483 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
2484 @code{0xC020} when used as a data address, but refers to byte address
2485 @code{0x30080} when used as a code address.
2487 (The D10V also uses separate code and data address spaces, which also
2488 affects the correspondence between pointers and addresses, but we're
2489 going to ignore that here; this example is already too long.)
2491 To cope with architectures like this---the D10V is not the only
2492 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
2493 byte numbers, and @dfn{pointers}, which are the target's representation
2494 of an address of a particular type of data. In the example above,
2495 @code{0xC020} is the pointer, which refers to one of the addresses
2496 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
2497 @value{GDBN} provides functions for turning a pointer into an address
2498 and vice versa, in the appropriate way for the current architecture.
2500 Unfortunately, since addresses and pointers are identical on almost all
2501 processors, this distinction tends to bit-rot pretty quickly. Thus,
2502 each time you port @value{GDBN} to an architecture which does
2503 distinguish between pointers and addresses, you'll probably need to
2504 clean up some architecture-independent code.
2506 Here are functions which convert between pointers and addresses:
2508 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
2509 Treat the bytes at @var{buf} as a pointer or reference of type
2510 @var{type}, and return the address it represents, in a manner
2511 appropriate for the current architecture. This yields an address
2512 @value{GDBN} can use to read target memory, disassemble, etc. Note that
2513 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2516 For example, if the current architecture is the Intel x86, this function
2517 extracts a little-endian integer of the appropriate length from
2518 @var{buf} and returns it. However, if the current architecture is the
2519 D10V, this function will return a 16-bit integer extracted from
2520 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
2522 If @var{type} is not a pointer or reference type, then this function
2523 will signal an internal error.
2526 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
2527 Store the address @var{addr} in @var{buf}, in the proper format for a
2528 pointer of type @var{type} in the current architecture. Note that
2529 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2532 For example, if the current architecture is the Intel x86, this function
2533 stores @var{addr} unmodified as a little-endian integer of the
2534 appropriate length in @var{buf}. However, if the current architecture
2535 is the D10V, this function divides @var{addr} by four if @var{type} is
2536 a pointer to a function, and then stores it in @var{buf}.
2538 If @var{type} is not a pointer or reference type, then this function
2539 will signal an internal error.
2542 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
2543 Assuming that @var{val} is a pointer, return the address it represents,
2544 as appropriate for the current architecture.
2546 This function actually works on integral values, as well as pointers.
2547 For pointers, it performs architecture-specific conversions as
2548 described above for @code{extract_typed_address}.
2551 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
2552 Create and return a value representing a pointer of type @var{type} to
2553 the address @var{addr}, as appropriate for the current architecture.
2554 This function performs architecture-specific conversions as described
2555 above for @code{store_typed_address}.
2558 Here are some macros which architectures can define to indicate the
2559 relationship between pointers and addresses. These have default
2560 definitions, appropriate for architectures on which all pointers are
2561 simple unsigned byte addresses.
2563 @deftypefn {Target Macro} CORE_ADDR POINTER_TO_ADDRESS (struct type *@var{type}, char *@var{buf})
2564 Assume that @var{buf} holds a pointer of type @var{type}, in the
2565 appropriate format for the current architecture. Return the byte
2566 address the pointer refers to.
2568 This function may safely assume that @var{type} is either a pointer or a
2569 C@t{++} reference type.
2572 @deftypefn {Target Macro} void ADDRESS_TO_POINTER (struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
2573 Store in @var{buf} a pointer of type @var{type} representing the address
2574 @var{addr}, in the appropriate format for the current architecture.
2576 This function may safely assume that @var{type} is either a pointer or a
2577 C@t{++} reference type.
2580 @section Address Classes
2581 @cindex address classes
2582 @cindex DW_AT_byte_size
2583 @cindex DW_AT_address_class
2585 Sometimes information about different kinds of addresses is available
2586 via the debug information. For example, some programming environments
2587 define addresses of several different sizes. If the debug information
2588 distinguishes these kinds of address classes through either the size
2589 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
2590 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
2591 following macros should be defined in order to disambiguate these
2592 types within @value{GDBN} as well as provide the added information to
2593 a @value{GDBN} user when printing type expressions.
2595 @deftypefn {Target Macro} int ADDRESS_CLASS_TYPE_FLAGS (int @var{byte_size}, int @var{dwarf2_addr_class})
2596 Returns the type flags needed to construct a pointer type whose size
2597 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
2598 This function is normally called from within a symbol reader. See
2599 @file{dwarf2read.c}.
2602 @deftypefn {Target Macro} char *ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (int @var{type_flags})
2603 Given the type flags representing an address class qualifier, return
2606 @deftypefn {Target Macro} int ADDRESS_CLASS_NAME_to_TYPE_FLAGS (int @var{name}, int *var{type_flags_ptr})
2607 Given an address qualifier name, set the @code{int} refererenced by @var{type_flags_ptr} to the type flags
2608 for that address class qualifier.
2611 Since the need for address classes is rather rare, none of
2612 the address class macros defined by default. Predicate
2613 macros are provided to detect when they are defined.
2615 Consider a hypothetical architecture in which addresses are normally
2616 32-bits wide, but 16-bit addresses are also supported. Furthermore,
2617 suppose that the @w{DWARF 2} information for this architecture simply
2618 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
2619 of these "short" pointers. The following functions could be defined
2620 to implement the address class macros:
2623 somearch_address_class_type_flags (int byte_size,
2624 int dwarf2_addr_class)
2627 return TYPE_FLAG_ADDRESS_CLASS_1;
2633 somearch_address_class_type_flags_to_name (int type_flags)
2635 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
2642 somearch_address_class_name_to_type_flags (char *name,
2643 int *type_flags_ptr)
2645 if (strcmp (name, "short") == 0)
2647 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
2655 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
2656 to indicate the presence of one of these "short" pointers. E.g, if
2657 the debug information indicates that @code{short_ptr_var} is one of these
2658 short pointers, @value{GDBN} might show the following behavior:
2661 (gdb) ptype short_ptr_var
2662 type = int * @@short
2666 @section Raw and Virtual Register Representations
2667 @cindex raw register representation
2668 @cindex virtual register representation
2669 @cindex representations, raw and virtual registers
2671 @emph{Maintainer note: This section is pretty much obsolete. The
2672 functionality described here has largely been replaced by
2673 pseudo-registers and the mechanisms described in @ref{Target
2674 Architecture Definition, , Using Different Register and Memory Data
2675 Representations}. See also @uref{http://www.gnu.org/software/gdb/bugs/,
2676 Bug Tracking Database} and
2677 @uref{http://sources.redhat.com/gdb/current/ari/, ARI Index} for more
2678 up-to-date information.}
2680 Some architectures use one representation for a value when it lives in a
2681 register, but use a different representation when it lives in memory.
2682 In @value{GDBN}'s terminology, the @dfn{raw} representation is the one used in
2683 the target registers, and the @dfn{virtual} representation is the one
2684 used in memory, and within @value{GDBN} @code{struct value} objects.
2686 @emph{Maintainer note: Notice that the same mechanism is being used to
2687 both convert a register to a @code{struct value} and alternative
2690 For almost all data types on almost all architectures, the virtual and
2691 raw representations are identical, and no special handling is needed.
2692 However, they do occasionally differ. For example:
2696 The x86 architecture supports an 80-bit @code{long double} type. However, when
2697 we store those values in memory, they occupy twelve bytes: the
2698 floating-point number occupies the first ten, and the final two bytes
2699 are unused. This keeps the values aligned on four-byte boundaries,
2700 allowing more efficient access. Thus, the x86 80-bit floating-point
2701 type is the raw representation, and the twelve-byte loosely-packed
2702 arrangement is the virtual representation.
2705 Some 64-bit MIPS targets present 32-bit registers to @value{GDBN} as 64-bit
2706 registers, with garbage in their upper bits. @value{GDBN} ignores the top 32
2707 bits. Thus, the 64-bit form, with garbage in the upper 32 bits, is the
2708 raw representation, and the trimmed 32-bit representation is the
2709 virtual representation.
2712 In general, the raw representation is determined by the architecture, or
2713 @value{GDBN}'s interface to the architecture, while the virtual representation
2714 can be chosen for @value{GDBN}'s convenience. @value{GDBN}'s register file,
2715 @code{registers}, holds the register contents in raw format, and the
2716 @value{GDBN} remote protocol transmits register values in raw format.
2718 Your architecture may define the following macros to request
2719 conversions between the raw and virtual format:
2721 @deftypefn {Target Macro} int REGISTER_CONVERTIBLE (int @var{reg})
2722 Return non-zero if register number @var{reg}'s value needs different raw
2723 and virtual formats.
2725 You should not use @code{REGISTER_CONVERT_TO_VIRTUAL} for a register
2726 unless this macro returns a non-zero value for that register.
2729 @deftypefn {Target Macro} int DEPRECATED_REGISTER_RAW_SIZE (int @var{reg})
2730 The size of register number @var{reg}'s raw value. This is the number
2731 of bytes the register will occupy in @code{registers}, or in a @value{GDBN}
2732 remote protocol packet.
2735 @deftypefn {Target Macro} int DEPRECATED_REGISTER_VIRTUAL_SIZE (int @var{reg})
2736 The size of register number @var{reg}'s value, in its virtual format.
2737 This is the size a @code{struct value}'s buffer will have, holding that
2741 @deftypefn {Target Macro} struct type *DEPRECATED_REGISTER_VIRTUAL_TYPE (int @var{reg})
2742 This is the type of the virtual representation of register number
2743 @var{reg}. Note that there is no need for a macro giving a type for the
2744 register's raw form; once the register's value has been obtained, @value{GDBN}
2745 always uses the virtual form.
2748 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_VIRTUAL (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
2749 Convert the value of register number @var{reg} to @var{type}, which
2750 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2751 at @var{from} holds the register's value in raw format; the macro should
2752 convert the value to virtual format, and place it at @var{to}.
2754 Note that @code{REGISTER_CONVERT_TO_VIRTUAL} and
2755 @code{REGISTER_CONVERT_TO_RAW} take their @var{reg} and @var{type}
2756 arguments in different orders.
2758 You should only use @code{REGISTER_CONVERT_TO_VIRTUAL} with registers
2759 for which the @code{REGISTER_CONVERTIBLE} macro returns a non-zero
2763 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_RAW (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
2764 Convert the value of register number @var{reg} to @var{type}, which
2765 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2766 at @var{from} holds the register's value in raw format; the macro should
2767 convert the value to virtual format, and place it at @var{to}.
2769 Note that REGISTER_CONVERT_TO_VIRTUAL and REGISTER_CONVERT_TO_RAW take
2770 their @var{reg} and @var{type} arguments in different orders.
2774 @section Using Different Register and Memory Data Representations
2775 @cindex register representation
2776 @cindex memory representation
2777 @cindex representations, register and memory
2778 @cindex register data formats, converting
2779 @cindex @code{struct value}, converting register contents to
2781 @emph{Maintainer's note: The way GDB manipulates registers is undergoing
2782 significant change. Many of the macros and functions refered to in this
2783 section are likely to be subject to further revision. See
2784 @uref{http://sources.redhat.com/gdb/current/ari/, A.R. Index} and
2785 @uref{http://www.gnu.org/software/gdb/bugs, Bug Tracking Database} for
2786 further information. cagney/2002-05-06.}
2788 Some architectures can represent a data object in a register using a
2789 form that is different to the objects more normal memory representation.
2795 The Alpha architecture can represent 32 bit integer values in
2796 floating-point registers.
2799 The x86 architecture supports 80-bit floating-point registers. The
2800 @code{long double} data type occupies 96 bits in memory but only 80 bits
2801 when stored in a register.
2805 In general, the register representation of a data type is determined by
2806 the architecture, or @value{GDBN}'s interface to the architecture, while
2807 the memory representation is determined by the Application Binary
2810 For almost all data types on almost all architectures, the two
2811 representations are identical, and no special handling is needed.
2812 However, they do occasionally differ. Your architecture may define the
2813 following macros to request conversions between the register and memory
2814 representations of a data type:
2816 @deftypefn {Target Macro} int CONVERT_REGISTER_P (int @var{reg})
2817 Return non-zero if the representation of a data value stored in this
2818 register may be different to the representation of that same data value
2819 when stored in memory.
2821 When non-zero, the macros @code{REGISTER_TO_VALUE} and
2822 @code{VALUE_TO_REGISTER} are used to perform any necessary conversion.
2825 @deftypefn {Target Macro} void REGISTER_TO_VALUE (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
2826 Convert the value of register number @var{reg} to a data object of type
2827 @var{type}. The buffer at @var{from} holds the register's value in raw
2828 format; the converted value should be placed in the buffer at @var{to}.
2830 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
2831 their @var{reg} and @var{type} arguments in different orders.
2833 You should only use @code{REGISTER_TO_VALUE} with registers for which
2834 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
2837 @deftypefn {Target Macro} void VALUE_TO_REGISTER (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
2838 Convert a data value of type @var{type} to register number @var{reg}'
2841 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
2842 their @var{reg} and @var{type} arguments in different orders.
2844 You should only use @code{VALUE_TO_REGISTER} with registers for which
2845 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
2848 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_TYPE (int @var{regnum}, struct type *@var{type}, char *@var{buf})
2849 See @file{mips-tdep.c}. It does not do what you want.
2853 @section Frame Interpretation
2855 @section Inferior Call Setup
2857 @section Compiler Characteristics
2859 @section Target Conditionals
2861 This section describes the macros that you can use to define the target
2866 @item ADDR_BITS_REMOVE (addr)
2867 @findex ADDR_BITS_REMOVE
2868 If a raw machine instruction address includes any bits that are not
2869 really part of the address, then define this macro to expand into an
2870 expression that zeroes those bits in @var{addr}. This is only used for
2871 addresses of instructions, and even then not in all contexts.
2873 For example, the two low-order bits of the PC on the Hewlett-Packard PA
2874 2.0 architecture contain the privilege level of the corresponding
2875 instruction. Since instructions must always be aligned on four-byte
2876 boundaries, the processor masks out these bits to generate the actual
2877 address of the instruction. ADDR_BITS_REMOVE should filter out these
2878 bits with an expression such as @code{((addr) & ~3)}.
2880 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS (@var{name}, @var{type_flags_ptr})
2881 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS
2882 If @var{name} is a valid address class qualifier name, set the @code{int}
2883 referenced by @var{type_flags_ptr} to the mask representing the qualifier
2884 and return 1. If @var{name} is not a valid address class qualifier name,
2887 The value for @var{type_flags_ptr} should be one of
2888 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
2889 possibly some combination of these values or'd together.
2890 @xref{Target Architecture Definition, , Address Classes}.
2892 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P ()
2893 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P
2894 Predicate which indicates whether @code{ADDRESS_CLASS_NAME_TO_TYPE_FLAGS}
2897 @item ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
2898 @findex ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
2899 Given a pointers byte size (as described by the debug information) and
2900 the possible @code{DW_AT_address_class} value, return the type flags
2901 used by @value{GDBN} to represent this address class. The value
2902 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
2903 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
2904 values or'd together.
2905 @xref{Target Architecture Definition, , Address Classes}.
2907 @item ADDRESS_CLASS_TYPE_FLAGS_P ()
2908 @findex ADDRESS_CLASS_TYPE_FLAGS_P
2909 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS} has
2912 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (@var{type_flags})
2913 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME
2914 Return the name of the address class qualifier associated with the type
2915 flags given by @var{type_flags}.
2917 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P ()
2918 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P
2919 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS_TO_NAME} has
2921 @xref{Target Architecture Definition, , Address Classes}.
2923 @item ADDRESS_TO_POINTER (@var{type}, @var{buf}, @var{addr})
2924 @findex ADDRESS_TO_POINTER
2925 Store in @var{buf} a pointer of type @var{type} representing the address
2926 @var{addr}, in the appropriate format for the current architecture.
2927 This macro may safely assume that @var{type} is either a pointer or a
2928 C@t{++} reference type.
2929 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
2931 @item BELIEVE_PCC_PROMOTION
2932 @findex BELIEVE_PCC_PROMOTION
2933 Define if the compiler promotes a @code{short} or @code{char}
2934 parameter to an @code{int}, but still reports the parameter as its
2935 original type, rather than the promoted type.
2937 @item BITS_BIG_ENDIAN
2938 @findex BITS_BIG_ENDIAN
2939 Define this if the numbering of bits in the targets does @strong{not} match the
2940 endianness of the target byte order. A value of 1 means that the bits
2941 are numbered in a big-endian bit order, 0 means little-endian.
2945 This is the character array initializer for the bit pattern to put into
2946 memory where a breakpoint is set. Although it's common to use a trap
2947 instruction for a breakpoint, it's not required; for instance, the bit
2948 pattern could be an invalid instruction. The breakpoint must be no
2949 longer than the shortest instruction of the architecture.
2951 @code{BREAKPOINT} has been deprecated in favor of
2952 @code{BREAKPOINT_FROM_PC}.
2954 @item BIG_BREAKPOINT
2955 @itemx LITTLE_BREAKPOINT
2956 @findex LITTLE_BREAKPOINT
2957 @findex BIG_BREAKPOINT
2958 Similar to BREAKPOINT, but used for bi-endian targets.
2960 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
2961 favor of @code{BREAKPOINT_FROM_PC}.
2963 @item DEPRECATED_REMOTE_BREAKPOINT
2964 @itemx DEPRECATED_LITTLE_REMOTE_BREAKPOINT
2965 @itemx DEPRECATED_BIG_REMOTE_BREAKPOINT
2966 @findex DEPRECATED_BIG_REMOTE_BREAKPOINT
2967 @findex DEPRECATED_LITTLE_REMOTE_BREAKPOINT
2968 @findex DEPRECATED_REMOTE_BREAKPOINT
2969 Specify the breakpoint instruction sequence for a remote target.
2970 @code{DEPRECATED_REMOTE_BREAKPOINT},
2971 @code{DEPRECATED_BIG_REMOTE_BREAKPOINT} and
2972 @code{DEPRECATED_LITTLE_REMOTE_BREAKPOINT} have been deprecated in
2973 favor of @code{BREAKPOINT_FROM_PC} (@pxref{BREAKPOINT_FROM_PC}).
2975 @item BREAKPOINT_FROM_PC (@var{pcptr}, @var{lenptr})
2976 @findex BREAKPOINT_FROM_PC
2977 @anchor{BREAKPOINT_FROM_PC} Use the program counter to determine the
2978 contents and size of a breakpoint instruction. It returns a pointer to
2979 a string of bytes that encode a breakpoint instruction, stores the
2980 length of the string to @code{*@var{lenptr}}, and adjusts the program
2981 counter (if necessary) to point to the actual memory location where the
2982 breakpoint should be inserted.
2984 Although it is common to use a trap instruction for a breakpoint, it's
2985 not required; for instance, the bit pattern could be an invalid
2986 instruction. The breakpoint must be no longer than the shortest
2987 instruction of the architecture.
2989 Replaces all the other @var{BREAKPOINT} macros.
2991 @item MEMORY_INSERT_BREAKPOINT (@var{addr}, @var{contents_cache})
2992 @itemx MEMORY_REMOVE_BREAKPOINT (@var{addr}, @var{contents_cache})
2993 @findex MEMORY_REMOVE_BREAKPOINT
2994 @findex MEMORY_INSERT_BREAKPOINT
2995 Insert or remove memory based breakpoints. Reasonable defaults
2996 (@code{default_memory_insert_breakpoint} and
2997 @code{default_memory_remove_breakpoint} respectively) have been
2998 provided so that it is not necessary to define these for most
2999 architectures. Architectures which may want to define
3000 @code{MEMORY_INSERT_BREAKPOINT} and @code{MEMORY_REMOVE_BREAKPOINT} will
3001 likely have instructions that are oddly sized or are not stored in a
3002 conventional manner.
3004 It may also be desirable (from an efficiency standpoint) to define
3005 custom breakpoint insertion and removal routines if
3006 @code{BREAKPOINT_FROM_PC} needs to read the target's memory for some
3009 @item ADJUST_BREAKPOINT_ADDRESS (@var{address})
3010 @findex ADJUST_BREAKPOINT_ADDRESS
3011 @cindex breakpoint address adjusted
3012 Given an address at which a breakpoint is desired, return a breakpoint
3013 address adjusted to account for architectural constraints on
3014 breakpoint placement. This method is not needed by most targets.
3016 The FR-V target (see @file{frv-tdep.c}) requires this method.
3017 The FR-V is a VLIW architecture in which a number of RISC-like
3018 instructions are grouped (packed) together into an aggregate
3019 instruction or instruction bundle. When the processor executes
3020 one of these bundles, the component instructions are executed
3023 In the course of optimization, the compiler may group instructions
3024 from distinct source statements into the same bundle. The line number
3025 information associated with one of the latter statements will likely
3026 refer to some instruction other than the first one in the bundle. So,
3027 if the user attempts to place a breakpoint on one of these latter
3028 statements, @value{GDBN} must be careful to @emph{not} place the break
3029 instruction on any instruction other than the first one in the bundle.
3030 (Remember though that the instructions within a bundle execute
3031 in parallel, so the @emph{first} instruction is the instruction
3032 at the lowest address and has nothing to do with execution order.)
3034 The FR-V's @code{ADJUST_BREAKPOINT_ADDRESS} method will adjust a
3035 breakpoint's address by scanning backwards for the beginning of
3036 the bundle, returning the address of the bundle.
3038 Since the adjustment of a breakpoint may significantly alter a user's
3039 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
3040 is initially set and each time that that breakpoint is hit.
3042 @item CALL_DUMMY_LOCATION
3043 @findex CALL_DUMMY_LOCATION
3044 See the file @file{inferior.h}.
3046 This method has been replaced by @code{push_dummy_code}
3047 (@pxref{push_dummy_code}).
3049 @item CANNOT_FETCH_REGISTER (@var{regno})
3050 @findex CANNOT_FETCH_REGISTER
3051 A C expression that should be nonzero if @var{regno} cannot be fetched
3052 from an inferior process. This is only relevant if
3053 @code{FETCH_INFERIOR_REGISTERS} is not defined.
3055 @item CANNOT_STORE_REGISTER (@var{regno})
3056 @findex CANNOT_STORE_REGISTER
3057 A C expression that should be nonzero if @var{regno} should not be
3058 written to the target. This is often the case for program counters,
3059 status words, and other special registers. If this is not defined,
3060 @value{GDBN} will assume that all registers may be written.
3062 @item int CONVERT_REGISTER_P(@var{regnum})
3063 @findex CONVERT_REGISTER_P
3064 Return non-zero if register @var{regnum} can represent data values in a
3066 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3068 @item DECR_PC_AFTER_BREAK
3069 @findex DECR_PC_AFTER_BREAK
3070 Define this to be the amount by which to decrement the PC after the
3071 program encounters a breakpoint. This is often the number of bytes in
3072 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
3074 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
3075 @findex DISABLE_UNSETTABLE_BREAK
3076 If defined, this should evaluate to 1 if @var{addr} is in a shared
3077 library in which breakpoints cannot be set and so should be disabled.
3079 @item PRINT_FLOAT_INFO()
3080 @findex PRINT_FLOAT_INFO
3081 If defined, then the @samp{info float} command will print information about
3082 the processor's floating point unit.
3084 @item print_registers_info (@var{gdbarch}, @var{frame}, @var{regnum}, @var{all})
3085 @findex print_registers_info
3086 If defined, pretty print the value of the register @var{regnum} for the
3087 specified @var{frame}. If the value of @var{regnum} is -1, pretty print
3088 either all registers (@var{all} is non zero) or a select subset of
3089 registers (@var{all} is zero).
3091 The default method prints one register per line, and if @var{all} is
3092 zero omits floating-point registers.
3094 @item PRINT_VECTOR_INFO()
3095 @findex PRINT_VECTOR_INFO
3096 If defined, then the @samp{info vector} command will call this function
3097 to print information about the processor's vector unit.
3099 By default, the @samp{info vector} command will print all vector
3100 registers (the register's type having the vector attribute).
3102 @item DWARF_REG_TO_REGNUM
3103 @findex DWARF_REG_TO_REGNUM
3104 Convert DWARF register number into @value{GDBN} regnum. If not defined,
3105 no conversion will be performed.
3107 @item DWARF2_REG_TO_REGNUM
3108 @findex DWARF2_REG_TO_REGNUM
3109 Convert DWARF2 register number into @value{GDBN} regnum. If not
3110 defined, no conversion will be performed.
3112 @item ECOFF_REG_TO_REGNUM
3113 @findex ECOFF_REG_TO_REGNUM
3114 Convert ECOFF register number into @value{GDBN} regnum. If not defined,
3115 no conversion will be performed.
3117 @item END_OF_TEXT_DEFAULT
3118 @findex END_OF_TEXT_DEFAULT
3119 This is an expression that should designate the end of the text section.
3122 @item EXTRACT_RETURN_VALUE(@var{type}, @var{regbuf}, @var{valbuf})
3123 @findex EXTRACT_RETURN_VALUE
3124 Define this to extract a function's return value of type @var{type} from
3125 the raw register state @var{regbuf} and copy that, in virtual format,
3128 This method has been deprecated in favour of @code{gdbarch_return_value}
3129 (@pxref{gdbarch_return_value}).
3131 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS(@var{regbuf})
3132 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS
3133 @anchor{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}
3134 When defined, extract from the array @var{regbuf} (containing the raw
3135 register state) the @code{CORE_ADDR} at which a function should return
3136 its structure value.
3138 @xref{gdbarch_return_value}.
3140 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P()
3141 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P
3142 Predicate for @code{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}.
3144 @item DEPRECATED_FP_REGNUM
3145 @findex DEPRECATED_FP_REGNUM
3146 If the virtual frame pointer is kept in a register, then define this
3147 macro to be the number (greater than or equal to zero) of that register.
3149 This should only need to be defined if @code{DEPRECATED_TARGET_READ_FP}
3152 @item DEPRECATED_FRAMELESS_FUNCTION_INVOCATION(@var{fi})
3153 @findex DEPRECATED_FRAMELESS_FUNCTION_INVOCATION
3154 Define this to an expression that returns 1 if the function invocation
3155 represented by @var{fi} does not have a stack frame associated with it.
3158 @item frame_align (@var{address})
3159 @anchor{frame_align}
3161 Define this to adjust @var{address} so that it meets the alignment
3162 requirements for the start of a new stack frame. A stack frame's
3163 alignment requirements are typically stronger than a target processors
3164 stack alignment requirements (@pxref{DEPRECATED_STACK_ALIGN}).
3166 This function is used to ensure that, when creating a dummy frame, both
3167 the initial stack pointer and (if needed) the address of the return
3168 value are correctly aligned.
3170 Unlike @code{DEPRECATED_STACK_ALIGN}, this function always adjusts the
3171 address in the direction of stack growth.
3173 By default, no frame based stack alignment is performed.
3175 @item int frame_red_zone_size
3177 The number of bytes, beyond the innermost-stack-address, reserved by the
3178 @sc{abi}. A function is permitted to use this scratch area (instead of
3179 allocating extra stack space).
3181 When performing an inferior function call, to ensure that it does not
3182 modify this area, @value{GDBN} adjusts the innermost-stack-address by
3183 @var{frame_red_zone_size} bytes before pushing parameters onto the
3186 By default, zero bytes are allocated. The value must be aligned
3187 (@pxref{frame_align}).
3189 The @sc{amd64} (nee x86-64) @sc{abi} documentation refers to the
3190 @emph{red zone} when describing this scratch area.
3193 @item DEPRECATED_FRAME_CHAIN(@var{frame})
3194 @findex DEPRECATED_FRAME_CHAIN
3195 Given @var{frame}, return a pointer to the calling frame.
3197 @item DEPRECATED_FRAME_CHAIN_VALID(@var{chain}, @var{thisframe})
3198 @findex DEPRECATED_FRAME_CHAIN_VALID
3199 Define this to be an expression that returns zero if the given frame is an
3200 outermost frame, with no caller, and nonzero otherwise. Most normal
3201 situations can be handled without defining this macro, including @code{NULL}
3202 chain pointers, dummy frames, and frames whose PC values are inside the
3203 startup file (e.g.@: @file{crt0.o}), inside @code{main}, or inside
3206 @item DEPRECATED_FRAME_INIT_SAVED_REGS(@var{frame})
3207 @findex DEPRECATED_FRAME_INIT_SAVED_REGS
3208 See @file{frame.h}. Determines the address of all registers in the
3209 current stack frame storing each in @code{frame->saved_regs}. Space for
3210 @code{frame->saved_regs} shall be allocated by
3211 @code{DEPRECATED_FRAME_INIT_SAVED_REGS} using
3212 @code{frame_saved_regs_zalloc}.
3214 @code{FRAME_FIND_SAVED_REGS} is deprecated.
3216 @item FRAME_NUM_ARGS (@var{fi})
3217 @findex FRAME_NUM_ARGS
3218 For the frame described by @var{fi} return the number of arguments that
3219 are being passed. If the number of arguments is not known, return
3222 @item DEPRECATED_FRAME_SAVED_PC(@var{frame})
3223 @findex DEPRECATED_FRAME_SAVED_PC
3224 @anchor{DEPRECATED_FRAME_SAVED_PC} Given @var{frame}, return the pc
3225 saved there. This is the return address.
3227 This method is deprecated. @xref{unwind_pc}.
3229 @item CORE_ADDR unwind_pc (struct frame_info *@var{this_frame})
3231 @anchor{unwind_pc} Return the instruction address, in @var{this_frame}'s
3232 caller, at which execution will resume after @var{this_frame} returns.
3233 This is commonly refered to as the return address.
3235 The implementation, which must be frame agnostic (work with any frame),
3236 is typically no more than:
3240 frame_unwind_unsigned_register (this_frame, D10V_PC_REGNUM, &pc);
3241 return d10v_make_iaddr (pc);
3245 @xref{DEPRECATED_FRAME_SAVED_PC}, which this method replaces.
3247 @item CORE_ADDR unwind_sp (struct frame_info *@var{this_frame})
3249 @anchor{unwind_sp} Return the frame's inner most stack address. This is
3250 commonly refered to as the frame's @dfn{stack pointer}.
3252 The implementation, which must be frame agnostic (work with any frame),
3253 is typically no more than:
3257 frame_unwind_unsigned_register (this_frame, D10V_SP_REGNUM, &sp);
3258 return d10v_make_daddr (sp);
3262 @xref{TARGET_READ_SP}, which this method replaces.
3264 @item FUNCTION_EPILOGUE_SIZE
3265 @findex FUNCTION_EPILOGUE_SIZE
3266 For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the
3267 function end symbol is 0. For such targets, you must define
3268 @code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size of a
3269 function's epilogue.
3271 @item DEPRECATED_FUNCTION_START_OFFSET
3272 @findex DEPRECATED_FUNCTION_START_OFFSET
3273 An integer, giving the offset in bytes from a function's address (as
3274 used in the values of symbols, function pointers, etc.), and the
3275 function's first genuine instruction.
3277 This is zero on almost all machines: the function's address is usually
3278 the address of its first instruction. However, on the VAX, for
3279 example, each function starts with two bytes containing a bitmask
3280 indicating which registers to save upon entry to the function. The
3281 VAX @code{call} instructions check this value, and save the
3282 appropriate registers automatically. Thus, since the offset from the
3283 function's address to its first instruction is two bytes,
3284 @code{DEPRECATED_FUNCTION_START_OFFSET} would be 2 on the VAX.
3286 @item GCC_COMPILED_FLAG_SYMBOL
3287 @itemx GCC2_COMPILED_FLAG_SYMBOL
3288 @findex GCC2_COMPILED_FLAG_SYMBOL
3289 @findex GCC_COMPILED_FLAG_SYMBOL
3290 If defined, these are the names of the symbols that @value{GDBN} will
3291 look for to detect that GCC compiled the file. The default symbols
3292 are @code{gcc_compiled.} and @code{gcc2_compiled.},
3293 respectively. (Currently only defined for the Delta 68.)
3295 @item @value{GDBN}_MULTI_ARCH
3296 @findex @value{GDBN}_MULTI_ARCH
3297 If defined and non-zero, enables support for multiple architectures
3298 within @value{GDBN}.
3300 This support can be enabled at two levels. At level one, only
3301 definitions for previously undefined macros are provided; at level two,
3302 a multi-arch definition of all architecture dependent macros will be
3305 @item @value{GDBN}_TARGET_IS_HPPA
3306 @findex @value{GDBN}_TARGET_IS_HPPA
3307 This determines whether horrible kludge code in @file{dbxread.c} and
3308 @file{partial-stab.h} is used to mangle multiple-symbol-table files from
3309 HPPA's. This should all be ripped out, and a scheme like @file{elfread.c}
3312 @item GET_LONGJMP_TARGET
3313 @findex GET_LONGJMP_TARGET
3314 For most machines, this is a target-dependent parameter. On the
3315 DECstation and the Iris, this is a native-dependent parameter, since
3316 the header file @file{setjmp.h} is needed to define it.
3318 This macro determines the target PC address that @code{longjmp} will jump to,
3319 assuming that we have just stopped at a @code{longjmp} breakpoint. It takes a
3320 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3321 pointer. It examines the current state of the machine as needed.
3323 @item DEPRECATED_GET_SAVED_REGISTER
3324 @findex DEPRECATED_GET_SAVED_REGISTER
3325 Define this if you need to supply your own definition for the function
3326 @code{DEPRECATED_GET_SAVED_REGISTER}.
3328 @item DEPRECATED_IBM6000_TARGET
3329 @findex DEPRECATED_IBM6000_TARGET
3330 Shows that we are configured for an IBM RS/6000 system. This
3331 conditional should be eliminated (FIXME) and replaced by
3332 feature-specific macros. It was introduced in a haste and we are
3333 repenting at leisure.
3335 @item I386_USE_GENERIC_WATCHPOINTS
3336 An x86-based target can define this to use the generic x86 watchpoint
3337 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3339 @item SYMBOLS_CAN_START_WITH_DOLLAR
3340 @findex SYMBOLS_CAN_START_WITH_DOLLAR
3341 Some systems have routines whose names start with @samp{$}. Giving this
3342 macro a non-zero value tells @value{GDBN}'s expression parser to check for such
3343 routines when parsing tokens that begin with @samp{$}.
3345 On HP-UX, certain system routines (millicode) have names beginning with
3346 @samp{$} or @samp{$$}. For example, @code{$$dyncall} is a millicode
3347 routine that handles inter-space procedure calls on PA-RISC.
3349 @item DEPRECATED_INIT_EXTRA_FRAME_INFO (@var{fromleaf}, @var{frame})
3350 @findex DEPRECATED_INIT_EXTRA_FRAME_INFO
3351 If additional information about the frame is required this should be
3352 stored in @code{frame->extra_info}. Space for @code{frame->extra_info}
3353 is allocated using @code{frame_extra_info_zalloc}.
3355 @item DEPRECATED_INIT_FRAME_PC (@var{fromleaf}, @var{prev})
3356 @findex DEPRECATED_INIT_FRAME_PC
3357 This is a C statement that sets the pc of the frame pointed to by
3358 @var{prev}. [By default...]
3360 @item INNER_THAN (@var{lhs}, @var{rhs})
3362 Returns non-zero if stack address @var{lhs} is inner than (nearer to the
3363 stack top) stack address @var{rhs}. Define this as @code{lhs < rhs} if
3364 the target's stack grows downward in memory, or @code{lhs > rsh} if the
3367 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{pc})
3368 @findex gdbarch_in_function_epilogue_p
3369 Returns non-zero if the given @var{pc} is in the epilogue of a function.
3370 The epilogue of a function is defined as the part of a function where
3371 the stack frame of the function already has been destroyed up to the
3372 final `return from function call' instruction.
3374 @item DEPRECATED_SIGTRAMP_START (@var{pc})
3375 @findex DEPRECATED_SIGTRAMP_START
3376 @itemx DEPRECATED_SIGTRAMP_END (@var{pc})
3377 @findex DEPRECATED_SIGTRAMP_END
3378 Define these to be the start and end address of the @code{sigtramp} for the
3379 given @var{pc}. On machines where the address is just a compile time
3380 constant, the macro expansion will typically just ignore the supplied
3383 @item IN_SOLIB_CALL_TRAMPOLINE (@var{pc}, @var{name})
3384 @findex IN_SOLIB_CALL_TRAMPOLINE
3385 Define this to evaluate to nonzero if the program is stopped in the
3386 trampoline that connects to a shared library.
3388 @item IN_SOLIB_RETURN_TRAMPOLINE (@var{pc}, @var{name})
3389 @findex IN_SOLIB_RETURN_TRAMPOLINE
3390 Define this to evaluate to nonzero if the program is stopped in the
3391 trampoline that returns from a shared library.
3393 @item IN_SOLIB_DYNSYM_RESOLVE_CODE (@var{pc})
3394 @findex IN_SOLIB_DYNSYM_RESOLVE_CODE
3395 Define this to evaluate to nonzero if the program is stopped in the
3398 @item SKIP_SOLIB_RESOLVER (@var{pc})
3399 @findex SKIP_SOLIB_RESOLVER
3400 Define this to evaluate to the (nonzero) address at which execution
3401 should continue to get past the dynamic linker's symbol resolution
3402 function. A zero value indicates that it is not important or necessary
3403 to set a breakpoint to get through the dynamic linker and that single
3404 stepping will suffice.
3406 @item INTEGER_TO_ADDRESS (@var{type}, @var{buf})
3407 @findex INTEGER_TO_ADDRESS
3408 @cindex converting integers to addresses
3409 Define this when the architecture needs to handle non-pointer to address
3410 conversions specially. Converts that value to an address according to
3411 the current architectures conventions.
3413 @emph{Pragmatics: When the user copies a well defined expression from
3414 their source code and passes it, as a parameter, to @value{GDBN}'s
3415 @code{print} command, they should get the same value as would have been
3416 computed by the target program. Any deviation from this rule can cause
3417 major confusion and annoyance, and needs to be justified carefully. In
3418 other words, @value{GDBN} doesn't really have the freedom to do these
3419 conversions in clever and useful ways. It has, however, been pointed
3420 out that users aren't complaining about how @value{GDBN} casts integers
3421 to pointers; they are complaining that they can't take an address from a
3422 disassembly listing and give it to @code{x/i}. Adding an architecture
3423 method like @code{INTEGER_TO_ADDRESS} certainly makes it possible for
3424 @value{GDBN} to ``get it right'' in all circumstances.}
3426 @xref{Target Architecture Definition, , Pointers Are Not Always
3429 @item NO_HIF_SUPPORT
3430 @findex NO_HIF_SUPPORT
3431 (Specific to the a29k.)
3433 @item POINTER_TO_ADDRESS (@var{type}, @var{buf})
3434 @findex POINTER_TO_ADDRESS
3435 Assume that @var{buf} holds a pointer of type @var{type}, in the
3436 appropriate format for the current architecture. Return the byte
3437 address the pointer refers to.
3438 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3440 @item REGISTER_CONVERTIBLE (@var{reg})
3441 @findex REGISTER_CONVERTIBLE
3442 Return non-zero if @var{reg} uses different raw and virtual formats.
3443 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3445 @item REGISTER_TO_VALUE(@var{regnum}, @var{type}, @var{from}, @var{to})
3446 @findex REGISTER_TO_VALUE
3447 Convert the raw contents of register @var{regnum} into a value of type
3449 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3451 @item DEPRECATED_REGISTER_RAW_SIZE (@var{reg})
3452 @findex DEPRECATED_REGISTER_RAW_SIZE
3453 Return the raw size of @var{reg}; defaults to the size of the register's
3455 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3457 @item register_reggroup_p (@var{gdbarch}, @var{regnum}, @var{reggroup})
3458 @findex register_reggroup_p
3459 @cindex register groups
3460 Return non-zero if register @var{regnum} is a member of the register
3461 group @var{reggroup}.
3463 By default, registers are grouped as follows:
3466 @item float_reggroup
3467 Any register with a valid name and a floating-point type.
3468 @item vector_reggroup
3469 Any register with a valid name and a vector type.
3470 @item general_reggroup
3471 Any register with a valid name and a type other than vector or
3472 floating-point. @samp{float_reggroup}.
3474 @itemx restore_reggroup
3476 Any register with a valid name.
3479 @item DEPRECATED_REGISTER_VIRTUAL_SIZE (@var{reg})
3480 @findex DEPRECATED_REGISTER_VIRTUAL_SIZE
3481 Return the virtual size of @var{reg}; defaults to the size of the
3482 register's virtual type.
3483 Return the virtual size of @var{reg}.
3484 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3486 @item DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})
3487 @findex REGISTER_VIRTUAL_TYPE
3488 Return the virtual type of @var{reg}.
3489 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3491 @item struct type *register_type (@var{gdbarch}, @var{reg})
3492 @findex register_type
3493 If defined, return the type of register @var{reg}. This function
3494 superseeds @code{DEPRECATED_REGISTER_VIRTUAL_TYPE}. @xref{Target Architecture
3495 Definition, , Raw and Virtual Register Representations}.
3497 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
3498 @findex REGISTER_CONVERT_TO_VIRTUAL
3499 Convert the value of register @var{reg} from its raw form to its virtual
3501 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3503 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
3504 @findex REGISTER_CONVERT_TO_RAW
3505 Convert the value of register @var{reg} from its virtual form to its raw
3507 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3509 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
3510 @findex regset_from_core_section
3511 Return the appropriate register set for a core file section with name
3512 @var{sect_name} and size @var{sect_size}.
3514 @item SOFTWARE_SINGLE_STEP_P()
3515 @findex SOFTWARE_SINGLE_STEP_P
3516 Define this as 1 if the target does not have a hardware single-step
3517 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
3519 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breapoints_p})
3520 @findex SOFTWARE_SINGLE_STEP
3521 A function that inserts or removes (depending on
3522 @var{insert_breapoints_p}) breakpoints at each possible destinations of
3523 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
3526 @item SOFUN_ADDRESS_MAYBE_MISSING
3527 @findex SOFUN_ADDRESS_MAYBE_MISSING
3528 Somebody clever observed that, the more actual addresses you have in the
3529 debug information, the more time the linker has to spend relocating
3530 them. So whenever there's some other way the debugger could find the
3531 address it needs, you should omit it from the debug info, to make
3534 @code{SOFUN_ADDRESS_MAYBE_MISSING} indicates that a particular set of
3535 hacks of this sort are in use, affecting @code{N_SO} and @code{N_FUN}
3536 entries in stabs-format debugging information. @code{N_SO} stabs mark
3537 the beginning and ending addresses of compilation units in the text
3538 segment. @code{N_FUN} stabs mark the starts and ends of functions.
3540 @code{SOFUN_ADDRESS_MAYBE_MISSING} means two things:
3544 @code{N_FUN} stabs have an address of zero. Instead, you should find the
3545 addresses where the function starts by taking the function name from
3546 the stab, and then looking that up in the minsyms (the
3547 linker/assembler symbol table). In other words, the stab has the
3548 name, and the linker/assembler symbol table is the only place that carries
3552 @code{N_SO} stabs have an address of zero, too. You just look at the
3553 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab,
3554 and guess the starting and ending addresses of the compilation unit from
3558 @item PC_LOAD_SEGMENT
3559 @findex PC_LOAD_SEGMENT
3560 If defined, print information about the load segment for the program
3561 counter. (Defined only for the RS/6000.)
3565 If the program counter is kept in a register, then define this macro to
3566 be the number (greater than or equal to zero) of that register.
3568 This should only need to be defined if @code{TARGET_READ_PC} and
3569 @code{TARGET_WRITE_PC} are not defined.
3572 @findex PARM_BOUNDARY
3573 If non-zero, round arguments to a boundary of this many bits before
3574 pushing them on the stack.
3576 @item stabs_argument_has_addr (@var{gdbarch}, @var{type})
3577 @findex stabs_argument_has_addr
3578 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3579 @anchor{stabs_argument_has_addr} Define this to return nonzero if a
3580 function argument of type @var{type} is passed by reference instead of
3583 This method replaces @code{DEPRECATED_REG_STRUCT_HAS_ADDR}
3584 (@pxref{DEPRECATED_REG_STRUCT_HAS_ADDR}).
3586 @item PROCESS_LINENUMBER_HOOK
3587 @findex PROCESS_LINENUMBER_HOOK
3588 A hook defined for XCOFF reading.
3590 @item PROLOGUE_FIRSTLINE_OVERLAP
3591 @findex PROLOGUE_FIRSTLINE_OVERLAP
3592 (Only used in unsupported Convex configuration.)
3596 If defined, this is the number of the processor status register. (This
3597 definition is only used in generic code when parsing "$ps".)
3599 @item DEPRECATED_POP_FRAME
3600 @findex DEPRECATED_POP_FRAME
3602 If defined, used by @code{frame_pop} to remove a stack frame. This
3603 method has been superseeded by generic code.
3605 @item push_dummy_call (@var{gdbarch}, @var{function}, @var{regcache}, @var{pc_addr}, @var{nargs}, @var{args}, @var{sp}, @var{struct_return}, @var{struct_addr})
3606 @findex push_dummy_call
3607 @findex DEPRECATED_PUSH_ARGUMENTS.
3608 @anchor{push_dummy_call} Define this to push the dummy frame's call to
3609 the inferior function onto the stack. In addition to pushing
3610 @var{nargs}, the code should push @var{struct_addr} (when
3611 @var{struct_return}), and the return address (@var{bp_addr}).
3613 @var{function} is a pointer to a @code{struct value}; on architectures that use
3614 function descriptors, this contains the function descriptor value.
3616 Returns the updated top-of-stack pointer.
3618 This method replaces @code{DEPRECATED_PUSH_ARGUMENTS}.
3620 @item CORE_ADDR push_dummy_code (@var{gdbarch}, @var{sp}, @var{funaddr}, @var{using_gcc}, @var{args}, @var{nargs}, @var{value_type}, @var{real_pc}, @var{bp_addr})
3621 @findex push_dummy_code
3622 @anchor{push_dummy_code} Given a stack based call dummy, push the
3623 instruction sequence (including space for a breakpoint) to which the
3624 called function should return.
3626 Set @var{bp_addr} to the address at which the breakpoint instruction
3627 should be inserted, @var{real_pc} to the resume address when starting
3628 the call sequence, and return the updated inner-most stack address.
3630 By default, the stack is grown sufficient to hold a frame-aligned
3631 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
3632 reserved for that breakpoint, and @var{real_pc} set to @var{funaddr}.
3634 This method replaces @code{CALL_DUMMY_LOCATION},
3635 @code{DEPRECATED_REGISTER_SIZE}.
3637 @item REGISTER_NAME(@var{i})
3638 @findex REGISTER_NAME
3639 Return the name of register @var{i} as a string. May return @code{NULL}
3640 or @code{NUL} to indicate that register @var{i} is not valid.
3642 @item DEPRECATED_REG_STRUCT_HAS_ADDR (@var{gcc_p}, @var{type})
3643 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3644 @anchor{DEPRECATED_REG_STRUCT_HAS_ADDR}Define this to return 1 if the
3645 given type will be passed by pointer rather than directly.
3647 This method has been replaced by @code{stabs_argument_has_addr}
3648 (@pxref{stabs_argument_has_addr}).
3650 @item SAVE_DUMMY_FRAME_TOS (@var{sp})
3651 @findex SAVE_DUMMY_FRAME_TOS
3652 @anchor{SAVE_DUMMY_FRAME_TOS} Used in @samp{call_function_by_hand} to
3653 notify the target dependent code of the top-of-stack value that will be
3654 passed to the the inferior code. This is the value of the @code{SP}
3655 after both the dummy frame and space for parameters/results have been
3656 allocated on the stack. @xref{unwind_dummy_id}.
3658 @item SDB_REG_TO_REGNUM
3659 @findex SDB_REG_TO_REGNUM
3660 Define this to convert sdb register numbers into @value{GDBN} regnums. If not
3661 defined, no conversion will be done.
3663 @item enum return_value_convention gdbarch_return_value (struct gdbarch *@var{gdbarch}, struct type *@var{valtype}, struct regcache *@var{regcache}, void *@var{readbuf}, const void *@var{writebuf})
3664 @findex gdbarch_return_value
3665 @anchor{gdbarch_return_value} Given a function with a return-value of
3666 type @var{rettype}, return which return-value convention that function
3669 @value{GDBN} currently recognizes two function return-value conventions:
3670 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
3671 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
3672 value is found in memory and the address of that memory location is
3673 passed in as the function's first parameter.
3675 If the register convention is being used, and @var{writebuf} is
3676 non-@code{NULL}, also copy the return-value in @var{writebuf} into
3679 If the register convention is being used, and @var{readbuf} is
3680 non-@code{NULL}, also copy the return value from @var{regcache} into
3681 @var{readbuf} (@var{regcache} contains a copy of the registers from the
3682 just returned function).
3684 @xref{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}, for a description of how
3685 return-values that use the struct convention are handled.
3687 @emph{Maintainer note: This method replaces separate predicate, extract,
3688 store methods. By having only one method, the logic needed to determine
3689 the return-value convention need only be implemented in one place. If
3690 @value{GDBN} were written in an @sc{oo} language, this method would
3691 instead return an object that knew how to perform the register
3692 return-value extract and store.}
3694 @emph{Maintainer note: This method does not take a @var{gcc_p}
3695 parameter, and such a parameter should not be added. If an architecture
3696 that requires per-compiler or per-function information be identified,
3697 then the replacement of @var{rettype} with @code{struct value}
3698 @var{function} should be persued.}
3700 @emph{Maintainer note: The @var{regcache} parameter limits this methods
3701 to the inner most frame. While replacing @var{regcache} with a
3702 @code{struct frame_info} @var{frame} parameter would remove that
3703 limitation there has yet to be a demonstrated need for such a change.}
3705 @item SKIP_PERMANENT_BREAKPOINT
3706 @findex SKIP_PERMANENT_BREAKPOINT
3707 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
3708 steps over a breakpoint by removing it, stepping one instruction, and
3709 re-inserting the breakpoint. However, permanent breakpoints are
3710 hardwired into the inferior, and can't be removed, so this strategy
3711 doesn't work. Calling @code{SKIP_PERMANENT_BREAKPOINT} adjusts the processor's
3712 state so that execution will resume just after the breakpoint. This
3713 macro does the right thing even when the breakpoint is in the delay slot
3714 of a branch or jump.
3716 @item SKIP_PROLOGUE (@var{pc})
3717 @findex SKIP_PROLOGUE
3718 A C expression that returns the address of the ``real'' code beyond the
3719 function entry prologue found at @var{pc}.
3721 @item SKIP_TRAMPOLINE_CODE (@var{pc})
3722 @findex SKIP_TRAMPOLINE_CODE
3723 If the target machine has trampoline code that sits between callers and
3724 the functions being called, then define this macro to return a new PC
3725 that is at the start of the real function.
3729 If the stack-pointer is kept in a register, then define this macro to be
3730 the number (greater than or equal to zero) of that register, or -1 if
3731 there is no such register.
3733 @item STAB_REG_TO_REGNUM
3734 @findex STAB_REG_TO_REGNUM
3735 Define this to convert stab register numbers (as gotten from `r'
3736 declarations) into @value{GDBN} regnums. If not defined, no conversion will be
3739 @item DEPRECATED_STACK_ALIGN (@var{addr})
3740 @anchor{DEPRECATED_STACK_ALIGN}
3741 @findex DEPRECATED_STACK_ALIGN
3742 Define this to increase @var{addr} so that it meets the alignment
3743 requirements for the processor's stack.
3745 Unlike @ref{frame_align}, this function always adjusts @var{addr}
3748 By default, no stack alignment is performed.
3750 @item STEP_SKIPS_DELAY (@var{addr})
3751 @findex STEP_SKIPS_DELAY
3752 Define this to return true if the address is of an instruction with a
3753 delay slot. If a breakpoint has been placed in the instruction's delay
3754 slot, @value{GDBN} will single-step over that instruction before resuming
3755 normally. Currently only defined for the Mips.
3757 @item STORE_RETURN_VALUE (@var{type}, @var{regcache}, @var{valbuf})
3758 @findex STORE_RETURN_VALUE
3759 A C expression that writes the function return value, found in
3760 @var{valbuf}, into the @var{regcache}. @var{type} is the type of the
3761 value that is to be returned.
3763 This method has been deprecated in favour of @code{gdbarch_return_value}
3764 (@pxref{gdbarch_return_value}).
3766 @item SYMBOL_RELOADING_DEFAULT
3767 @findex SYMBOL_RELOADING_DEFAULT
3768 The default value of the ``symbol-reloading'' variable. (Never defined in
3771 @item TARGET_CHAR_BIT
3772 @findex TARGET_CHAR_BIT
3773 Number of bits in a char; defaults to 8.
3775 @item TARGET_CHAR_SIGNED
3776 @findex TARGET_CHAR_SIGNED
3777 Non-zero if @code{char} is normally signed on this architecture; zero if
3778 it should be unsigned.
3780 The ISO C standard requires the compiler to treat @code{char} as
3781 equivalent to either @code{signed char} or @code{unsigned char}; any
3782 character in the standard execution set is supposed to be positive.
3783 Most compilers treat @code{char} as signed, but @code{char} is unsigned
3784 on the IBM S/390, RS6000, and PowerPC targets.
3786 @item TARGET_COMPLEX_BIT
3787 @findex TARGET_COMPLEX_BIT
3788 Number of bits in a complex number; defaults to @code{2 * TARGET_FLOAT_BIT}.
3790 At present this macro is not used.
3792 @item TARGET_DOUBLE_BIT
3793 @findex TARGET_DOUBLE_BIT
3794 Number of bits in a double float; defaults to @code{8 * TARGET_CHAR_BIT}.
3796 @item TARGET_DOUBLE_COMPLEX_BIT
3797 @findex TARGET_DOUBLE_COMPLEX_BIT
3798 Number of bits in a double complex; defaults to @code{2 * TARGET_DOUBLE_BIT}.
3800 At present this macro is not used.
3802 @item TARGET_FLOAT_BIT
3803 @findex TARGET_FLOAT_BIT
3804 Number of bits in a float; defaults to @code{4 * TARGET_CHAR_BIT}.
3806 @item TARGET_INT_BIT
3807 @findex TARGET_INT_BIT
3808 Number of bits in an integer; defaults to @code{4 * TARGET_CHAR_BIT}.
3810 @item TARGET_LONG_BIT
3811 @findex TARGET_LONG_BIT
3812 Number of bits in a long integer; defaults to @code{4 * TARGET_CHAR_BIT}.
3814 @item TARGET_LONG_DOUBLE_BIT
3815 @findex TARGET_LONG_DOUBLE_BIT
3816 Number of bits in a long double float;
3817 defaults to @code{2 * TARGET_DOUBLE_BIT}.
3819 @item TARGET_LONG_LONG_BIT
3820 @findex TARGET_LONG_LONG_BIT
3821 Number of bits in a long long integer; defaults to @code{2 * TARGET_LONG_BIT}.
3823 @item TARGET_PTR_BIT
3824 @findex TARGET_PTR_BIT
3825 Number of bits in a pointer; defaults to @code{TARGET_INT_BIT}.
3827 @item TARGET_SHORT_BIT
3828 @findex TARGET_SHORT_BIT
3829 Number of bits in a short integer; defaults to @code{2 * TARGET_CHAR_BIT}.
3831 @item TARGET_READ_PC
3832 @findex TARGET_READ_PC
3833 @itemx TARGET_WRITE_PC (@var{val}, @var{pid})
3834 @findex TARGET_WRITE_PC
3835 @anchor{TARGET_WRITE_PC}
3836 @itemx TARGET_READ_SP
3837 @findex TARGET_READ_SP
3838 @itemx TARGET_READ_FP
3839 @findex TARGET_READ_FP
3844 @anchor{TARGET_READ_SP} These change the behavior of @code{read_pc},
3845 @code{write_pc}, and @code{read_sp}. For most targets, these may be
3846 left undefined. @value{GDBN} will call the read and write register
3847 functions with the relevant @code{_REGNUM} argument.
3849 These macros are useful when a target keeps one of these registers in a
3850 hard to get at place; for example, part in a segment register and part
3851 in an ordinary register.
3853 @xref{unwind_sp}, which replaces @code{TARGET_READ_SP}.
3855 @item TARGET_VIRTUAL_FRAME_POINTER(@var{pc}, @var{regp}, @var{offsetp})
3856 @findex TARGET_VIRTUAL_FRAME_POINTER
3857 Returns a @code{(register, offset)} pair representing the virtual frame
3858 pointer in use at the code address @var{pc}. If virtual frame pointers
3859 are not used, a default definition simply returns
3860 @code{DEPRECATED_FP_REGNUM}, with an offset of zero.
3862 @item TARGET_HAS_HARDWARE_WATCHPOINTS
3863 If non-zero, the target has support for hardware-assisted
3864 watchpoints. @xref{Algorithms, watchpoints}, for more details and
3865 other related macros.
3867 @item TARGET_PRINT_INSN (@var{addr}, @var{info})
3868 @findex TARGET_PRINT_INSN
3869 This is the function used by @value{GDBN} to print an assembly
3870 instruction. It prints the instruction at address @var{addr} in
3871 debugged memory and returns the length of the instruction, in bytes. If
3872 a target doesn't define its own printing routine, it defaults to an
3873 accessor function for the global pointer
3874 @code{deprecated_tm_print_insn}. This usually points to a function in
3875 the @code{opcodes} library (@pxref{Support Libraries, ,Opcodes}).
3876 @var{info} is a structure (of type @code{disassemble_info}) defined in
3877 @file{include/dis-asm.h} used to pass information to the instruction
3880 @item struct frame_id unwind_dummy_id (struct frame_info *@var{frame})
3881 @findex unwind_dummy_id
3882 @anchor{unwind_dummy_id} Given @var{frame} return a @code{struct
3883 frame_id} that uniquely identifies an inferior function call's dummy
3884 frame. The value returned must match the dummy frame stack value
3885 previously saved using @code{SAVE_DUMMY_FRAME_TOS}.
3886 @xref{SAVE_DUMMY_FRAME_TOS}.
3888 @item DEPRECATED_USE_STRUCT_CONVENTION (@var{gcc_p}, @var{type})
3889 @findex DEPRECATED_USE_STRUCT_CONVENTION
3890 If defined, this must be an expression that is nonzero if a value of the
3891 given @var{type} being returned from a function must have space
3892 allocated for it on the stack. @var{gcc_p} is true if the function
3893 being considered is known to have been compiled by GCC; this is helpful
3894 for systems where GCC is known to use different calling convention than
3897 This method has been deprecated in favour of @code{gdbarch_return_value}
3898 (@pxref{gdbarch_return_value}).
3900 @item VALUE_TO_REGISTER(@var{type}, @var{regnum}, @var{from}, @var{to})
3901 @findex VALUE_TO_REGISTER
3902 Convert a value of type @var{type} into the raw contents of register
3904 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3906 @item VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
3907 @findex VARIABLES_INSIDE_BLOCK
3908 For dbx-style debugging information, if the compiler puts variable
3909 declarations inside LBRAC/RBRAC blocks, this should be defined to be
3910 nonzero. @var{desc} is the value of @code{n_desc} from the
3911 @code{N_RBRAC} symbol, and @var{gcc_p} is true if @value{GDBN} has noticed the
3912 presence of either the @code{GCC_COMPILED_SYMBOL} or the
3913 @code{GCC2_COMPILED_SYMBOL}. By default, this is 0.
3915 @item OS9K_VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
3916 @findex OS9K_VARIABLES_INSIDE_BLOCK
3917 Similarly, for OS/9000. Defaults to 1.
3920 Motorola M68K target conditionals.
3924 Define this to be the 4-bit location of the breakpoint trap vector. If
3925 not defined, it will default to @code{0xf}.
3927 @item REMOTE_BPT_VECTOR
3928 Defaults to @code{1}.
3930 @item NAME_OF_MALLOC
3931 @findex NAME_OF_MALLOC
3932 A string containing the name of the function to call in order to
3933 allocate some memory in the inferior. The default value is "malloc".
3937 @section Adding a New Target
3939 @cindex adding a target
3940 The following files add a target to @value{GDBN}:
3944 @item gdb/config/@var{arch}/@var{ttt}.mt
3945 Contains a Makefile fragment specific to this target. Specifies what
3946 object files are needed for target @var{ttt}, by defining
3947 @samp{TDEPFILES=@dots{}} and @samp{TDEPLIBS=@dots{}}. Also specifies
3948 the header file which describes @var{ttt}, by defining @samp{TM_FILE=
3951 You can also define @samp{TM_CFLAGS}, @samp{TM_CLIBS}, @samp{TM_CDEPS},
3952 but these are now deprecated, replaced by autoconf, and may go away in
3953 future versions of @value{GDBN}.
3955 @item gdb/@var{ttt}-tdep.c
3956 Contains any miscellaneous code required for this target machine. On
3957 some machines it doesn't exist at all. Sometimes the macros in
3958 @file{tm-@var{ttt}.h} become very complicated, so they are implemented
3959 as functions here instead, and the macro is simply defined to call the
3960 function. This is vastly preferable, since it is easier to understand
3963 @item gdb/@var{arch}-tdep.c
3964 @itemx gdb/@var{arch}-tdep.h
3965 This often exists to describe the basic layout of the target machine's
3966 processor chip (registers, stack, etc.). If used, it is included by
3967 @file{@var{ttt}-tdep.h}. It can be shared among many targets that use
3970 @item gdb/config/@var{arch}/tm-@var{ttt}.h
3971 (@file{tm.h} is a link to this file, created by @code{configure}). Contains
3972 macro definitions about the target machine's registers, stack frame
3973 format and instructions.
3975 New targets do not need this file and should not create it.
3977 @item gdb/config/@var{arch}/tm-@var{arch}.h
3978 This often exists to describe the basic layout of the target machine's
3979 processor chip (registers, stack, etc.). If used, it is included by
3980 @file{tm-@var{ttt}.h}. It can be shared among many targets that use the
3983 New targets do not need this file and should not create it.
3987 If you are adding a new operating system for an existing CPU chip, add a
3988 @file{config/tm-@var{os}.h} file that describes the operating system
3989 facilities that are unusual (extra symbol table info; the breakpoint
3990 instruction needed; etc.). Then write a @file{@var{arch}/tm-@var{os}.h}
3991 that just @code{#include}s @file{tm-@var{arch}.h} and
3992 @file{config/tm-@var{os}.h}.
3995 @section Converting an existing Target Architecture to Multi-arch
3996 @cindex converting targets to multi-arch
3998 This section describes the current accepted best practice for converting
3999 an existing target architecture to the multi-arch framework.
4001 The process consists of generating, testing, posting and committing a
4002 sequence of patches. Each patch must contain a single change, for
4008 Directly convert a group of functions into macros (the conversion does
4009 not change the behavior of any of the functions).
4012 Replace a non-multi-arch with a multi-arch mechanism (e.g.,
4016 Enable multi-arch level one.
4019 Delete one or more files.
4024 There isn't a size limit on a patch, however, a developer is strongly
4025 encouraged to keep the patch size down.
4027 Since each patch is well defined, and since each change has been tested
4028 and shows no regressions, the patches are considered @emph{fairly}
4029 obvious. Such patches, when submitted by developers listed in the
4030 @file{MAINTAINERS} file, do not need approval. Occasional steps in the
4031 process may be more complicated and less clear. The developer is
4032 expected to use their judgment and is encouraged to seek advice as
4035 @subsection Preparation
4037 The first step is to establish control. Build (with @option{-Werror}
4038 enabled) and test the target so that there is a baseline against which
4039 the debugger can be compared.
4041 At no stage can the test results regress or @value{GDBN} stop compiling
4042 with @option{-Werror}.
4044 @subsection Add the multi-arch initialization code
4046 The objective of this step is to establish the basic multi-arch
4047 framework. It involves
4052 The addition of a @code{@var{arch}_gdbarch_init} function@footnote{The
4053 above is from the original example and uses K&R C. @value{GDBN}
4054 has since converted to ISO C but lets ignore that.} that creates
4057 static struct gdbarch *
4058 d10v_gdbarch_init (info, arches)
4059 struct gdbarch_info info;
4060 struct gdbarch_list *arches;
4062 struct gdbarch *gdbarch;
4063 /* there is only one d10v architecture */
4065 return arches->gdbarch;
4066 gdbarch = gdbarch_alloc (&info, NULL);
4074 A per-architecture dump function to print any architecture specific
4078 mips_dump_tdep (struct gdbarch *current_gdbarch,
4079 struct ui_file *file)
4081 @dots{} code to print architecture specific info @dots{}
4086 A change to @code{_initialize_@var{arch}_tdep} to register this new
4090 _initialize_mips_tdep (void)
4092 gdbarch_register (bfd_arch_mips, mips_gdbarch_init,
4097 Add the macro @code{GDB_MULTI_ARCH}, defined as 0 (zero), to the file@*
4098 @file{config/@var{arch}/tm-@var{arch}.h}.
4102 @subsection Update multi-arch incompatible mechanisms
4104 Some mechanisms do not work with multi-arch. They include:
4107 @item FRAME_FIND_SAVED_REGS
4108 Replaced with @code{DEPRECATED_FRAME_INIT_SAVED_REGS}
4112 At this stage you could also consider converting the macros into
4115 @subsection Prepare for multi-arch level to one
4117 Temporally set @code{GDB_MULTI_ARCH} to @code{GDB_MULTI_ARCH_PARTIAL}
4118 and then build and start @value{GDBN} (the change should not be
4119 committed). @value{GDBN} may not build, and once built, it may die with
4120 an internal error listing the architecture methods that must be
4123 Fix any build problems (patch(es)).
4125 Convert all the architecture methods listed, which are only macros, into
4126 functions (patch(es)).
4128 Update @code{@var{arch}_gdbarch_init} to set all the missing
4129 architecture methods and wrap the corresponding macros in @code{#if
4130 !GDB_MULTI_ARCH} (patch(es)).
4132 @subsection Set multi-arch level one
4134 Change the value of @code{GDB_MULTI_ARCH} to GDB_MULTI_ARCH_PARTIAL (a
4137 Any problems with throwing ``the switch'' should have been fixed
4140 @subsection Convert remaining macros
4142 Suggest converting macros into functions (and setting the corresponding
4143 architecture method) in small batches.
4145 @subsection Set multi-arch level to two
4147 This should go smoothly.
4149 @subsection Delete the TM file
4151 The @file{tm-@var{arch}.h} can be deleted. @file{@var{arch}.mt} and
4152 @file{configure.in} updated.
4155 @node Target Vector Definition
4157 @chapter Target Vector Definition
4158 @cindex target vector
4160 The target vector defines the interface between @value{GDBN}'s
4161 abstract handling of target systems, and the nitty-gritty code that
4162 actually exercises control over a process or a serial port.
4163 @value{GDBN} includes some 30-40 different target vectors; however,
4164 each configuration of @value{GDBN} includes only a few of them.
4166 @section File Targets
4168 Both executables and core files have target vectors.
4170 @section Standard Protocol and Remote Stubs
4172 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code
4173 that runs in the target system. @value{GDBN} provides several sample
4174 @dfn{stubs} that can be integrated into target programs or operating
4175 systems for this purpose; they are named @file{*-stub.c}.
4177 The @value{GDBN} user's manual describes how to put such a stub into
4178 your target code. What follows is a discussion of integrating the
4179 SPARC stub into a complicated operating system (rather than a simple
4180 program), by Stu Grossman, the author of this stub.
4182 The trap handling code in the stub assumes the following upon entry to
4187 %l1 and %l2 contain pc and npc respectively at the time of the trap;
4193 you are in the correct trap window.
4196 As long as your trap handler can guarantee those conditions, then there
4197 is no reason why you shouldn't be able to ``share'' traps with the stub.
4198 The stub has no requirement that it be jumped to directly from the
4199 hardware trap vector. That is why it calls @code{exceptionHandler()},
4200 which is provided by the external environment. For instance, this could
4201 set up the hardware traps to actually execute code which calls the stub
4202 first, and then transfers to its own trap handler.
4204 For the most point, there probably won't be much of an issue with
4205 ``sharing'' traps, as the traps we use are usually not used by the kernel,
4206 and often indicate unrecoverable error conditions. Anyway, this is all
4207 controlled by a table, and is trivial to modify. The most important
4208 trap for us is for @code{ta 1}. Without that, we can't single step or
4209 do breakpoints. Everything else is unnecessary for the proper operation
4210 of the debugger/stub.
4212 From reading the stub, it's probably not obvious how breakpoints work.
4213 They are simply done by deposit/examine operations from @value{GDBN}.
4215 @section ROM Monitor Interface
4217 @section Custom Protocols
4219 @section Transport Layer
4221 @section Builtin Simulator
4224 @node Native Debugging
4226 @chapter Native Debugging
4227 @cindex native debugging
4229 Several files control @value{GDBN}'s configuration for native support:
4233 @item gdb/config/@var{arch}/@var{xyz}.mh
4234 Specifies Makefile fragments needed by a @emph{native} configuration on
4235 machine @var{xyz}. In particular, this lists the required
4236 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
4237 Also specifies the header file which describes native support on
4238 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
4239 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
4240 @samp{NAT_CDEPS}, etc.; see @file{Makefile.in}.
4242 @emph{Maintainer's note: The @file{.mh} suffix is because this file
4243 originally contained @file{Makefile} fragments for hosting @value{GDBN}
4244 on machine @var{xyz}. While the file is no longer used for this
4245 purpose, the @file{.mh} suffix remains. Perhaps someone will
4246 eventually rename these fragments so that they have a @file{.mn}
4249 @item gdb/config/@var{arch}/nm-@var{xyz}.h
4250 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
4251 macro definitions describing the native system environment, such as
4252 child process control and core file support.
4254 @item gdb/@var{xyz}-nat.c
4255 Contains any miscellaneous C code required for this native support of
4256 this machine. On some machines it doesn't exist at all.
4259 There are some ``generic'' versions of routines that can be used by
4260 various systems. These can be customized in various ways by macros
4261 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
4262 the @var{xyz} host, you can just include the generic file's name (with
4263 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
4265 Otherwise, if your machine needs custom support routines, you will need
4266 to write routines that perform the same functions as the generic file.
4267 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
4268 into @code{NATDEPFILES}.
4272 This contains the @emph{target_ops vector} that supports Unix child
4273 processes on systems which use ptrace and wait to control the child.
4276 This contains the @emph{target_ops vector} that supports Unix child
4277 processes on systems which use /proc to control the child.
4280 This does the low-level grunge that uses Unix system calls to do a ``fork
4281 and exec'' to start up a child process.
4284 This is the low level interface to inferior processes for systems using
4285 the Unix @code{ptrace} call in a vanilla way.
4288 @section Native core file Support
4289 @cindex native core files
4292 @findex fetch_core_registers
4293 @item core-aout.c::fetch_core_registers()
4294 Support for reading registers out of a core file. This routine calls
4295 @code{register_addr()}, see below. Now that BFD is used to read core
4296 files, virtually all machines should use @code{core-aout.c}, and should
4297 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
4298 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
4300 @item core-aout.c::register_addr()
4301 If your @code{nm-@var{xyz}.h} file defines the macro
4302 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
4303 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
4304 register number @code{regno}. @code{blockend} is the offset within the
4305 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
4306 @file{core-aout.c} will define the @code{register_addr()} function and
4307 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
4308 you are using the standard @code{fetch_core_registers()}, you will need
4309 to define your own version of @code{register_addr()}, put it into your
4310 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
4311 the @code{NATDEPFILES} list. If you have your own
4312 @code{fetch_core_registers()}, you may not need a separate
4313 @code{register_addr()}. Many custom @code{fetch_core_registers()}
4314 implementations simply locate the registers themselves.@refill
4317 When making @value{GDBN} run native on a new operating system, to make it
4318 possible to debug core files, you will need to either write specific
4319 code for parsing your OS's core files, or customize
4320 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
4321 machine uses to define the struct of registers that is accessible
4322 (possibly in the u-area) in a core file (rather than
4323 @file{machine/reg.h}), and an include file that defines whatever header
4324 exists on a core file (e.g. the u-area or a @code{struct core}). Then
4325 modify @code{trad_unix_core_file_p} to use these values to set up the
4326 section information for the data segment, stack segment, any other
4327 segments in the core file (perhaps shared library contents or control
4328 information), ``registers'' segment, and if there are two discontiguous
4329 sets of registers (e.g. integer and float), the ``reg2'' segment. This
4330 section information basically delimits areas in the core file in a
4331 standard way, which the section-reading routines in BFD know how to seek
4334 Then back in @value{GDBN}, you need a matching routine called
4335 @code{fetch_core_registers}. If you can use the generic one, it's in
4336 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
4337 It will be passed a char pointer to the entire ``registers'' segment,
4338 its length, and a zero; or a char pointer to the entire ``regs2''
4339 segment, its length, and a 2. The routine should suck out the supplied
4340 register values and install them into @value{GDBN}'s ``registers'' array.
4342 If your system uses @file{/proc} to control processes, and uses ELF
4343 format core files, then you may be able to use the same routines for
4344 reading the registers out of processes and out of core files.
4352 @section shared libraries
4354 @section Native Conditionals
4355 @cindex native conditionals
4357 When @value{GDBN} is configured and compiled, various macros are
4358 defined or left undefined, to control compilation when the host and
4359 target systems are the same. These macros should be defined (or left
4360 undefined) in @file{nm-@var{system}.h}.
4364 @item CHILD_PREPARE_TO_STORE
4365 @findex CHILD_PREPARE_TO_STORE
4366 If the machine stores all registers at once in the child process, then
4367 define this to ensure that all values are correct. This usually entails
4368 a read from the child.
4370 [Note that this is incorrectly defined in @file{xm-@var{system}.h} files
4373 @item FETCH_INFERIOR_REGISTERS
4374 @findex FETCH_INFERIOR_REGISTERS
4375 Define this if the native-dependent code will provide its own routines
4376 @code{fetch_inferior_registers} and @code{store_inferior_registers} in
4377 @file{@var{host}-nat.c}. If this symbol is @emph{not} defined, and
4378 @file{infptrace.c} is included in this configuration, the default
4379 routines in @file{infptrace.c} are used for these functions.
4383 This macro is normally defined to be the number of the first floating
4384 point register, if the machine has such registers. As such, it would
4385 appear only in target-specific code. However, @file{/proc} support uses this
4386 to decide whether floats are in use on this target.
4388 @item GET_LONGJMP_TARGET
4389 @findex GET_LONGJMP_TARGET
4390 For most machines, this is a target-dependent parameter. On the
4391 DECstation and the Iris, this is a native-dependent parameter, since
4392 @file{setjmp.h} is needed to define it.
4394 This macro determines the target PC address that @code{longjmp} will jump to,
4395 assuming that we have just stopped at a longjmp breakpoint. It takes a
4396 @code{CORE_ADDR *} as argument, and stores the target PC value through this
4397 pointer. It examines the current state of the machine as needed.
4399 @item I386_USE_GENERIC_WATCHPOINTS
4400 An x86-based machine can define this to use the generic x86 watchpoint
4401 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4404 @findex KERNEL_U_ADDR
4405 Define this to the address of the @code{u} structure (the ``user
4406 struct'', also known as the ``u-page'') in kernel virtual memory. @value{GDBN}
4407 needs to know this so that it can subtract this address from absolute
4408 addresses in the upage, that are obtained via ptrace or from core files.
4409 On systems that don't need this value, set it to zero.
4411 @item KERNEL_U_ADDR_HPUX
4412 @findex KERNEL_U_ADDR_HPUX
4413 Define this to cause @value{GDBN} to determine the address of @code{u} at
4414 runtime, by using HP-style @code{nlist} on the kernel's image in the
4417 @item ONE_PROCESS_WRITETEXT
4418 @findex ONE_PROCESS_WRITETEXT
4419 Define this to be able to, when a breakpoint insertion fails, warn the
4420 user that another process may be running with the same executable.
4423 @findex PROC_NAME_FMT
4424 Defines the format for the name of a @file{/proc} device. Should be
4425 defined in @file{nm.h} @emph{only} in order to override the default
4426 definition in @file{procfs.c}.
4428 @item PTRACE_ARG3_TYPE
4429 @findex PTRACE_ARG3_TYPE
4430 The type of the third argument to the @code{ptrace} system call, if it
4431 exists and is different from @code{int}.
4433 @item REGISTER_U_ADDR
4434 @findex REGISTER_U_ADDR
4435 Defines the offset of the registers in the ``u area''.
4437 @item SHELL_COMMAND_CONCAT
4438 @findex SHELL_COMMAND_CONCAT
4439 If defined, is a string to prefix on the shell command used to start the
4444 If defined, this is the name of the shell to use to run the inferior.
4445 Defaults to @code{"/bin/sh"}.
4447 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
4449 Define this to expand into an expression that will cause the symbols in
4450 @var{filename} to be added to @value{GDBN}'s symbol table. If
4451 @var{readsyms} is zero symbols are not read but any necessary low level
4452 processing for @var{filename} is still done.
4454 @item SOLIB_CREATE_INFERIOR_HOOK
4455 @findex SOLIB_CREATE_INFERIOR_HOOK
4456 Define this to expand into any shared-library-relocation code that you
4457 want to be run just after the child process has been forked.
4459 @item START_INFERIOR_TRAPS_EXPECTED
4460 @findex START_INFERIOR_TRAPS_EXPECTED
4461 When starting an inferior, @value{GDBN} normally expects to trap
4463 the shell execs, and once when the program itself execs. If the actual
4464 number of traps is something other than 2, then define this macro to
4465 expand into the number expected.
4469 This determines whether small routines in @file{*-tdep.c}, which
4470 translate register values between @value{GDBN}'s internal
4471 representation and the @file{/proc} representation, are compiled.
4474 @findex U_REGS_OFFSET
4475 This is the offset of the registers in the upage. It need only be
4476 defined if the generic ptrace register access routines in
4477 @file{infptrace.c} are being used (that is, @file{infptrace.c} is
4478 configured in, and @code{FETCH_INFERIOR_REGISTERS} is not defined). If
4479 the default value from @file{infptrace.c} is good enough, leave it
4482 The default value means that u.u_ar0 @emph{points to} the location of
4483 the registers. I'm guessing that @code{#define U_REGS_OFFSET 0} means
4484 that @code{u.u_ar0} @emph{is} the location of the registers.
4488 See @file{objfiles.c}.
4491 @findex DEBUG_PTRACE
4492 Define this to debug @code{ptrace} calls.
4496 @node Support Libraries
4498 @chapter Support Libraries
4503 BFD provides support for @value{GDBN} in several ways:
4506 @item identifying executable and core files
4507 BFD will identify a variety of file types, including a.out, coff, and
4508 several variants thereof, as well as several kinds of core files.
4510 @item access to sections of files
4511 BFD parses the file headers to determine the names, virtual addresses,
4512 sizes, and file locations of all the various named sections in files
4513 (such as the text section or the data section). @value{GDBN} simply
4514 calls BFD to read or write section @var{x} at byte offset @var{y} for
4517 @item specialized core file support
4518 BFD provides routines to determine the failing command name stored in a
4519 core file, the signal with which the program failed, and whether a core
4520 file matches (i.e.@: could be a core dump of) a particular executable
4523 @item locating the symbol information
4524 @value{GDBN} uses an internal interface of BFD to determine where to find the
4525 symbol information in an executable file or symbol-file. @value{GDBN} itself
4526 handles the reading of symbols, since BFD does not ``understand'' debug
4527 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
4532 @cindex opcodes library
4534 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
4535 library because it's also used in binutils, for @file{objdump}).
4542 @cindex @code{libiberty} library
4544 The @code{libiberty} library provides a set of functions and features
4545 that integrate and improve on functionality found in modern operating
4546 systems. Broadly speaking, such features can be divided into three
4547 groups: supplemental functions (functions that may be missing in some
4548 environments and operating systems), replacement functions (providing
4549 a uniform and easier to use interface for commonly used standard
4550 functions), and extensions (which provide additional functionality
4551 beyond standard functions).
4553 @value{GDBN} uses various features provided by the @code{libiberty}
4554 library, for instance the C@t{++} demangler, the @acronym{IEEE}
4555 floating format support functions, the input options parser
4556 @samp{getopt}, the @samp{obstack} extension, and other functions.
4558 @subsection @code{obstacks} in @value{GDBN}
4559 @cindex @code{obstacks}
4561 The obstack mechanism provides a convenient way to allocate and free
4562 chunks of memory. Each obstack is a pool of memory that is managed
4563 like a stack. Objects (of any nature, size and alignment) are
4564 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
4565 @code{libiberty}'s documenatation for a more detailed explanation of
4568 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
4569 object files. There is an obstack associated with each internal
4570 representation of an object file. Lots of things get allocated on
4571 these @code{obstacks}: dictionary entries, blocks, blockvectors,
4572 symbols, minimal symbols, types, vectors of fundamental types, class
4573 fields of types, object files section lists, object files section
4574 offets lists, line tables, symbol tables, partial symbol tables,
4575 string tables, symbol table private data, macros tables, debug
4576 information sections and entries, import and export lists (som),
4577 unwind information (hppa), dwarf2 location expressions data. Plus
4578 various strings such as directory names strings, debug format strings,
4581 An essential and convenient property of all data on @code{obstacks} is
4582 that memory for it gets allocated (with @code{obstack_alloc}) at
4583 various times during a debugging sesssion, but it is released all at
4584 once using the @code{obstack_free} function. The @code{obstack_free}
4585 function takes a pointer to where in the stack it must start the
4586 deletion from (much like the cleanup chains have a pointer to where to
4587 start the cleanups). Because of the stack like structure of the
4588 @code{obstacks}, this allows to free only a top portion of the
4589 obstack. There are a few instances in @value{GDBN} where such thing
4590 happens. Calls to @code{obstack_free} are done after some local data
4591 is allocated to the obstack. Only the local data is deleted from the
4592 obstack. Of course this assumes that nothing between the
4593 @code{obstack_alloc} and the @code{obstack_free} allocates anything
4594 else on the same obstack. For this reason it is best and safest to
4595 use temporary @code{obstacks}.
4597 Releasing the whole obstack is also not safe per se. It is safe only
4598 under the condition that we know the @code{obstacks} memory is no
4599 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
4600 when we get rid of the whole objfile(s), for instance upon reading a
4604 @cindex regular expressions library
4615 @item SIGN_EXTEND_CHAR
4617 @item SWITCH_ENUM_BUG
4632 This chapter covers topics that are lower-level than the major
4633 algorithms of @value{GDBN}.
4638 Cleanups are a structured way to deal with things that need to be done
4641 When your code does something (e.g., @code{xmalloc} some memory, or
4642 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
4643 the memory or @code{close} the file), it can make a cleanup. The
4644 cleanup will be done at some future point: when the command is finished
4645 and control returns to the top level; when an error occurs and the stack
4646 is unwound; or when your code decides it's time to explicitly perform
4647 cleanups. Alternatively you can elect to discard the cleanups you
4653 @item struct cleanup *@var{old_chain};
4654 Declare a variable which will hold a cleanup chain handle.
4656 @findex make_cleanup
4657 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
4658 Make a cleanup which will cause @var{function} to be called with
4659 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
4660 handle that can later be passed to @code{do_cleanups} or
4661 @code{discard_cleanups}. Unless you are going to call
4662 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
4663 from @code{make_cleanup}.
4666 @item do_cleanups (@var{old_chain});
4667 Do all cleanups added to the chain since the corresponding
4668 @code{make_cleanup} call was made.
4670 @findex discard_cleanups
4671 @item discard_cleanups (@var{old_chain});
4672 Same as @code{do_cleanups} except that it just removes the cleanups from
4673 the chain and does not call the specified functions.
4676 Cleanups are implemented as a chain. The handle returned by
4677 @code{make_cleanups} includes the cleanup passed to the call and any
4678 later cleanups appended to the chain (but not yet discarded or
4682 make_cleanup (a, 0);
4684 struct cleanup *old = make_cleanup (b, 0);
4692 will call @code{c()} and @code{b()} but will not call @code{a()}. The
4693 cleanup that calls @code{a()} will remain in the cleanup chain, and will
4694 be done later unless otherwise discarded.@refill
4696 Your function should explicitly do or discard the cleanups it creates.
4697 Failing to do this leads to non-deterministic behavior since the caller
4698 will arbitrarily do or discard your functions cleanups. This need leads
4699 to two common cleanup styles.
4701 The first style is try/finally. Before it exits, your code-block calls
4702 @code{do_cleanups} with the old cleanup chain and thus ensures that your
4703 code-block's cleanups are always performed. For instance, the following
4704 code-segment avoids a memory leak problem (even when @code{error} is
4705 called and a forced stack unwind occurs) by ensuring that the
4706 @code{xfree} will always be called:
4709 struct cleanup *old = make_cleanup (null_cleanup, 0);
4710 data = xmalloc (sizeof blah);
4711 make_cleanup (xfree, data);
4716 The second style is try/except. Before it exits, your code-block calls
4717 @code{discard_cleanups} with the old cleanup chain and thus ensures that
4718 any created cleanups are not performed. For instance, the following
4719 code segment, ensures that the file will be closed but only if there is
4723 FILE *file = fopen ("afile", "r");
4724 struct cleanup *old = make_cleanup (close_file, file);
4726 discard_cleanups (old);
4730 Some functions, e.g. @code{fputs_filtered()} or @code{error()}, specify
4731 that they ``should not be called when cleanups are not in place''. This
4732 means that any actions you need to reverse in the case of an error or
4733 interruption must be on the cleanup chain before you call these
4734 functions, since they might never return to your code (they
4735 @samp{longjmp} instead).
4737 @section Per-architecture module data
4738 @cindex per-architecture module data
4739 @cindex multi-arch data
4740 @cindex data-pointer, per-architecture/per-module
4742 The multi-arch framework includes a mechanism for adding module
4743 specific per-architecture data-pointers to the @code{struct gdbarch}
4744 architecture object.
4746 A module registers one or more per-architecture data-pointers using:
4748 @deftypefun struct gdbarch_data *gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
4749 @var{pre_init} is used to, on-demand, allocate an initial value for a
4750 per-architecture data-pointer using the architecture's obstack (passed
4751 in as a parameter). Since @var{pre_init} can be called during
4752 architecture creation, it is not parameterized with the architecture.
4753 and must not call modules that use per-architecture data.
4756 @deftypefun struct gdbarch_data *gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
4757 @var{post_init} is used to obtain an initial value for a
4758 per-architecture data-pointer @emph{after}. Since @var{post_init} is
4759 always called after architecture creation, it both receives the fully
4760 initialized architecture and is free to call modules that use
4761 per-architecture data (care needs to be taken to ensure that those
4762 other modules do not try to call back to this module as that will
4763 create in cycles in the initialization call graph).
4766 These functions return a @code{struct gdbarch_data} that is used to
4767 identify the per-architecture data-pointer added for that module.
4769 The per-architecture data-pointer is accessed using the function:
4771 @deftypefun void *gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
4772 Given the architecture @var{arch} and module data handle
4773 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
4774 or @code{gdbarch_data_register_post_init}), this function returns the
4775 current value of the per-architecture data-pointer. If the data
4776 pointer is @code{NULL}, it is first initialized by calling the
4777 corresponding @var{pre_init} or @var{post_init} method.
4780 The examples below assume the following definitions:
4783 struct nozel @{ int total; @};
4784 static struct gdbarch_data *nozel_handle;
4787 A module can extend the architecture vector, adding additional
4788 per-architecture data, using the @var{pre_init} method. The module's
4789 per-architecture data is then initialized during architecture
4792 In the below, the module's per-architecture @emph{nozel} is added. An
4793 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
4794 from @code{gdbarch_init}.
4798 nozel_pre_init (struct obstack *obstack)
4800 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
4807 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
4809 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
4810 data->total = nozel;
4814 A module can on-demand create architecture dependant data structures
4815 using @code{post_init}.
4817 In the below, the nozel's total is computed on-demand by
4818 @code{nozel_post_init} using information obtained from the
4823 nozel_post_init (struct gdbarch *gdbarch)
4825 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
4826 nozel->total = gdbarch@dots{} (gdbarch);
4833 nozel_total (struct gdbarch *gdbarch)
4835 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
4840 @section Wrapping Output Lines
4841 @cindex line wrap in output
4844 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
4845 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
4846 added in places that would be good breaking points. The utility
4847 routines will take care of actually wrapping if the line width is
4850 The argument to @code{wrap_here} is an indentation string which is
4851 printed @emph{only} if the line breaks there. This argument is saved
4852 away and used later. It must remain valid until the next call to
4853 @code{wrap_here} or until a newline has been printed through the
4854 @code{*_filtered} functions. Don't pass in a local variable and then
4857 It is usually best to call @code{wrap_here} after printing a comma or
4858 space. If you call it before printing a space, make sure that your
4859 indentation properly accounts for the leading space that will print if
4860 the line wraps there.
4862 Any function or set of functions that produce filtered output must
4863 finish by printing a newline, to flush the wrap buffer, before switching
4864 to unfiltered (@code{printf}) output. Symbol reading routines that
4865 print warnings are a good example.
4867 @section @value{GDBN} Coding Standards
4868 @cindex coding standards
4870 @value{GDBN} follows the GNU coding standards, as described in
4871 @file{etc/standards.texi}. This file is also available for anonymous
4872 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
4873 of the standard; in general, when the GNU standard recommends a practice
4874 but does not require it, @value{GDBN} requires it.
4876 @value{GDBN} follows an additional set of coding standards specific to
4877 @value{GDBN}, as described in the following sections.
4882 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
4885 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
4888 @subsection Memory Management
4890 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
4891 @code{calloc}, @code{free} and @code{asprintf}.
4893 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
4894 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
4895 these functions do not return when the memory pool is empty. Instead,
4896 they unwind the stack using cleanups. These functions return
4897 @code{NULL} when requested to allocate a chunk of memory of size zero.
4899 @emph{Pragmatics: By using these functions, the need to check every
4900 memory allocation is removed. These functions provide portable
4903 @value{GDBN} does not use the function @code{free}.
4905 @value{GDBN} uses the function @code{xfree} to return memory to the
4906 memory pool. Consistent with ISO-C, this function ignores a request to
4907 free a @code{NULL} pointer.
4909 @emph{Pragmatics: On some systems @code{free} fails when passed a
4910 @code{NULL} pointer.}
4912 @value{GDBN} can use the non-portable function @code{alloca} for the
4913 allocation of small temporary values (such as strings).
4915 @emph{Pragmatics: This function is very non-portable. Some systems
4916 restrict the memory being allocated to no more than a few kilobytes.}
4918 @value{GDBN} uses the string function @code{xstrdup} and the print
4919 function @code{xstrprintf}.
4921 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
4922 functions such as @code{sprintf} are very prone to buffer overflow
4926 @subsection Compiler Warnings
4927 @cindex compiler warnings
4929 With few exceptions, developers should include the configuration option
4930 @samp{--enable-gdb-build-warnings=,-Werror} when building @value{GDBN}.
4931 The exceptions are listed in the file @file{gdb/MAINTAINERS}.
4933 This option causes @value{GDBN} (when built using GCC) to be compiled
4934 with a carefully selected list of compiler warning flags. Any warnings
4935 from those flags being treated as errors.
4937 The current list of warning flags includes:
4941 Since @value{GDBN} coding standard requires all functions to be declared
4942 using a prototype, the flag has the side effect of ensuring that
4943 prototyped functions are always visible with out resorting to
4944 @samp{-Wstrict-prototypes}.
4947 Such code often appears to work except on instruction set architectures
4948 that use register windows.
4955 @itemx -Wformat-nonliteral
4956 Since @value{GDBN} uses the @code{format printf} attribute on all
4957 @code{printf} like functions these check not just @code{printf} calls
4958 but also calls to functions such as @code{fprintf_unfiltered}.
4961 This warning includes uses of the assignment operator within an
4962 @code{if} statement.
4964 @item -Wpointer-arith
4966 @item -Wuninitialized
4968 @item -Wunused-label
4969 This warning has the additional benefit of detecting the absence of the
4970 @code{case} reserved word in a switch statement:
4972 enum @{ FD_SCHEDULED, NOTHING_SCHEDULED @} sched;
4985 @item -Wunused-function
4988 @emph{Pragmatics: Due to the way that @value{GDBN} is implemented most
4989 functions have unused parameters. Consequently the warning
4990 @samp{-Wunused-parameter} is precluded from the list. The macro
4991 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
4992 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
4993 is being used. The options @samp{-Wall} and @samp{-Wunused} are also
4994 precluded because they both include @samp{-Wunused-parameter}.}
4996 @emph{Pragmatics: @value{GDBN} has not simply accepted the warnings
4997 enabled by @samp{-Wall -Werror -W...}. Instead it is selecting warnings
4998 when and where their benefits can be demonstrated.}
5000 @subsection Formatting
5002 @cindex source code formatting
5003 The standard GNU recommendations for formatting must be followed
5006 A function declaration should not have its name in column zero. A
5007 function definition should have its name in column zero.
5011 static void foo (void);
5019 @emph{Pragmatics: This simplifies scripting. Function definitions can
5020 be found using @samp{^function-name}.}
5022 There must be a space between a function or macro name and the opening
5023 parenthesis of its argument list (except for macro definitions, as
5024 required by C). There must not be a space after an open paren/bracket
5025 or before a close paren/bracket.
5027 While additional whitespace is generally helpful for reading, do not use
5028 more than one blank line to separate blocks, and avoid adding whitespace
5029 after the end of a program line (as of 1/99, some 600 lines had
5030 whitespace after the semicolon). Excess whitespace causes difficulties
5031 for @code{diff} and @code{patch} utilities.
5033 Pointers are declared using the traditional K&R C style:
5047 @subsection Comments
5049 @cindex comment formatting
5050 The standard GNU requirements on comments must be followed strictly.
5052 Block comments must appear in the following form, with no @code{/*}- or
5053 @code{*/}-only lines, and no leading @code{*}:
5056 /* Wait for control to return from inferior to debugger. If inferior
5057 gets a signal, we may decide to start it up again instead of
5058 returning. That is why there is a loop in this function. When
5059 this function actually returns it means the inferior should be left
5060 stopped and @value{GDBN} should read more commands. */
5063 (Note that this format is encouraged by Emacs; tabbing for a multi-line
5064 comment works correctly, and @kbd{M-q} fills the block consistently.)
5066 Put a blank line between the block comments preceding function or
5067 variable definitions, and the definition itself.
5069 In general, put function-body comments on lines by themselves, rather
5070 than trying to fit them into the 20 characters left at the end of a
5071 line, since either the comment or the code will inevitably get longer
5072 than will fit, and then somebody will have to move it anyhow.
5076 @cindex C data types
5077 Code must not depend on the sizes of C data types, the format of the
5078 host's floating point numbers, the alignment of anything, or the order
5079 of evaluation of expressions.
5081 @cindex function usage
5082 Use functions freely. There are only a handful of compute-bound areas
5083 in @value{GDBN} that might be affected by the overhead of a function
5084 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
5085 limited by the target interface (whether serial line or system call).
5087 However, use functions with moderation. A thousand one-line functions
5088 are just as hard to understand as a single thousand-line function.
5090 @emph{Macros are bad, M'kay.}
5091 (But if you have to use a macro, make sure that the macro arguments are
5092 protected with parentheses.)
5096 Declarations like @samp{struct foo *} should be used in preference to
5097 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
5100 @subsection Function Prototypes
5101 @cindex function prototypes
5103 Prototypes must be used when both @emph{declaring} and @emph{defining}
5104 a function. Prototypes for @value{GDBN} functions must include both the
5105 argument type and name, with the name matching that used in the actual
5106 function definition.
5108 All external functions should have a declaration in a header file that
5109 callers include, except for @code{_initialize_*} functions, which must
5110 be external so that @file{init.c} construction works, but shouldn't be
5111 visible to random source files.
5113 Where a source file needs a forward declaration of a static function,
5114 that declaration must appear in a block near the top of the source file.
5117 @subsection Internal Error Recovery
5119 During its execution, @value{GDBN} can encounter two types of errors.
5120 User errors and internal errors. User errors include not only a user
5121 entering an incorrect command but also problems arising from corrupt
5122 object files and system errors when interacting with the target.
5123 Internal errors include situations where @value{GDBN} has detected, at
5124 run time, a corrupt or erroneous situation.
5126 When reporting an internal error, @value{GDBN} uses
5127 @code{internal_error} and @code{gdb_assert}.
5129 @value{GDBN} must not call @code{abort} or @code{assert}.
5131 @emph{Pragmatics: There is no @code{internal_warning} function. Either
5132 the code detected a user error, recovered from it and issued a
5133 @code{warning} or the code failed to correctly recover from the user
5134 error and issued an @code{internal_error}.}
5136 @subsection File Names
5138 Any file used when building the core of @value{GDBN} must be in lower
5139 case. Any file used when building the core of @value{GDBN} must be 8.3
5140 unique. These requirements apply to both source and generated files.
5142 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
5143 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
5144 is introduced to the build process both @file{Makefile.in} and
5145 @file{configure.in} need to be modified accordingly. Compare the
5146 convoluted conversion process needed to transform @file{COPYING} into
5147 @file{copying.c} with the conversion needed to transform
5148 @file{version.in} into @file{version.c}.}
5150 Any file non 8.3 compliant file (that is not used when building the core
5151 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
5153 @emph{Pragmatics: This is clearly a compromise.}
5155 When @value{GDBN} has a local version of a system header file (ex
5156 @file{string.h}) the file name based on the POSIX header prefixed with
5157 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
5158 independent: they should use only macros defined by @file{configure},
5159 the compiler, or the host; they should include only system headers; they
5160 should refer only to system types. They may be shared between multiple
5161 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
5163 For other files @samp{-} is used as the separator.
5166 @subsection Include Files
5168 A @file{.c} file should include @file{defs.h} first.
5170 A @file{.c} file should directly include the @code{.h} file of every
5171 declaration and/or definition it directly refers to. It cannot rely on
5174 A @file{.h} file should directly include the @code{.h} file of every
5175 declaration and/or definition it directly refers to. It cannot rely on
5176 indirect inclusion. Exception: The file @file{defs.h} does not need to
5177 be directly included.
5179 An external declaration should only appear in one include file.
5181 An external declaration should never appear in a @code{.c} file.
5182 Exception: a declaration for the @code{_initialize} function that
5183 pacifies @option{-Wmissing-declaration}.
5185 A @code{typedef} definition should only appear in one include file.
5187 An opaque @code{struct} declaration can appear in multiple @file{.h}
5188 files. Where possible, a @file{.h} file should use an opaque
5189 @code{struct} declaration instead of an include.
5191 All @file{.h} files should be wrapped in:
5194 #ifndef INCLUDE_FILE_NAME_H
5195 #define INCLUDE_FILE_NAME_H
5201 @subsection Clean Design and Portable Implementation
5204 In addition to getting the syntax right, there's the little question of
5205 semantics. Some things are done in certain ways in @value{GDBN} because long
5206 experience has shown that the more obvious ways caused various kinds of
5209 @cindex assumptions about targets
5210 You can't assume the byte order of anything that comes from a target
5211 (including @var{value}s, object files, and instructions). Such things
5212 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
5213 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
5214 such as @code{bfd_get_32}.
5216 You can't assume that you know what interface is being used to talk to
5217 the target system. All references to the target must go through the
5218 current @code{target_ops} vector.
5220 You can't assume that the host and target machines are the same machine
5221 (except in the ``native'' support modules). In particular, you can't
5222 assume that the target machine's header files will be available on the
5223 host machine. Target code must bring along its own header files --
5224 written from scratch or explicitly donated by their owner, to avoid
5228 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
5229 to write the code portably than to conditionalize it for various
5232 @cindex system dependencies
5233 New @code{#ifdef}'s which test for specific compilers or manufacturers
5234 or operating systems are unacceptable. All @code{#ifdef}'s should test
5235 for features. The information about which configurations contain which
5236 features should be segregated into the configuration files. Experience
5237 has proven far too often that a feature unique to one particular system
5238 often creeps into other systems; and that a conditional based on some
5239 predefined macro for your current system will become worthless over
5240 time, as new versions of your system come out that behave differently
5241 with regard to this feature.
5243 Adding code that handles specific architectures, operating systems,
5244 target interfaces, or hosts, is not acceptable in generic code.
5246 @cindex portable file name handling
5247 @cindex file names, portability
5248 One particularly notorious area where system dependencies tend to
5249 creep in is handling of file names. The mainline @value{GDBN} code
5250 assumes Posix semantics of file names: absolute file names begin with
5251 a forward slash @file{/}, slashes are used to separate leading
5252 directories, case-sensitive file names. These assumptions are not
5253 necessarily true on non-Posix systems such as MS-Windows. To avoid
5254 system-dependent code where you need to take apart or construct a file
5255 name, use the following portable macros:
5258 @findex HAVE_DOS_BASED_FILE_SYSTEM
5259 @item HAVE_DOS_BASED_FILE_SYSTEM
5260 This preprocessing symbol is defined to a non-zero value on hosts
5261 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
5262 symbol to write conditional code which should only be compiled for
5265 @findex IS_DIR_SEPARATOR
5266 @item IS_DIR_SEPARATOR (@var{c})
5267 Evaluates to a non-zero value if @var{c} is a directory separator
5268 character. On Unix and GNU/Linux systems, only a slash @file{/} is
5269 such a character, but on Windows, both @file{/} and @file{\} will
5272 @findex IS_ABSOLUTE_PATH
5273 @item IS_ABSOLUTE_PATH (@var{file})
5274 Evaluates to a non-zero value if @var{file} is an absolute file name.
5275 For Unix and GNU/Linux hosts, a name which begins with a slash
5276 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
5277 @file{x:\bar} are also absolute file names.
5279 @findex FILENAME_CMP
5280 @item FILENAME_CMP (@var{f1}, @var{f2})
5281 Calls a function which compares file names @var{f1} and @var{f2} as
5282 appropriate for the underlying host filesystem. For Posix systems,
5283 this simply calls @code{strcmp}; on case-insensitive filesystems it
5284 will call @code{strcasecmp} instead.
5286 @findex DIRNAME_SEPARATOR
5287 @item DIRNAME_SEPARATOR
5288 Evaluates to a character which separates directories in
5289 @code{PATH}-style lists, typically held in environment variables.
5290 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
5292 @findex SLASH_STRING
5294 This evaluates to a constant string you should use to produce an
5295 absolute filename from leading directories and the file's basename.
5296 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
5297 @code{"\\"} for some Windows-based ports.
5300 In addition to using these macros, be sure to use portable library
5301 functions whenever possible. For example, to extract a directory or a
5302 basename part from a file name, use the @code{dirname} and
5303 @code{basename} library functions (available in @code{libiberty} for
5304 platforms which don't provide them), instead of searching for a slash
5305 with @code{strrchr}.
5307 Another way to generalize @value{GDBN} along a particular interface is with an
5308 attribute struct. For example, @value{GDBN} has been generalized to handle
5309 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
5310 by defining the @code{target_ops} structure and having a current target (as
5311 well as a stack of targets below it, for memory references). Whenever
5312 something needs to be done that depends on which remote interface we are
5313 using, a flag in the current target_ops structure is tested (e.g.,
5314 @code{target_has_stack}), or a function is called through a pointer in the
5315 current target_ops structure. In this way, when a new remote interface
5316 is added, only one module needs to be touched---the one that actually
5317 implements the new remote interface. Other examples of
5318 attribute-structs are BFD access to multiple kinds of object file
5319 formats, or @value{GDBN}'s access to multiple source languages.
5321 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
5322 the code interfacing between @code{ptrace} and the rest of
5323 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
5324 something was very painful. In @value{GDBN} 4.x, these have all been
5325 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
5326 with variations between systems the same way any system-independent
5327 file would (hooks, @code{#if defined}, etc.), and machines which are
5328 radically different don't need to use @file{infptrace.c} at all.
5330 All debugging code must be controllable using the @samp{set debug
5331 @var{module}} command. Do not use @code{printf} to print trace
5332 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
5333 @code{#ifdef DEBUG}.
5338 @chapter Porting @value{GDBN}
5339 @cindex porting to new machines
5341 Most of the work in making @value{GDBN} compile on a new machine is in
5342 specifying the configuration of the machine. This is done in a
5343 dizzying variety of header files and configuration scripts, which we
5344 hope to make more sensible soon. Let's say your new host is called an
5345 @var{xyz} (e.g., @samp{sun4}), and its full three-part configuration
5346 name is @code{@var{arch}-@var{xvend}-@var{xos}} (e.g.,
5347 @samp{sparc-sun-sunos4}). In particular:
5351 In the top level directory, edit @file{config.sub} and add @var{arch},
5352 @var{xvend}, and @var{xos} to the lists of supported architectures,
5353 vendors, and operating systems near the bottom of the file. Also, add
5354 @var{xyz} as an alias that maps to
5355 @code{@var{arch}-@var{xvend}-@var{xos}}. You can test your changes by
5359 ./config.sub @var{xyz}
5366 ./config.sub @code{@var{arch}-@var{xvend}-@var{xos}}
5370 which should both respond with @code{@var{arch}-@var{xvend}-@var{xos}}
5371 and no error messages.
5374 You need to port BFD, if that hasn't been done already. Porting BFD is
5375 beyond the scope of this manual.
5378 To configure @value{GDBN} itself, edit @file{gdb/configure.host} to recognize
5379 your system and set @code{gdb_host} to @var{xyz}, and (unless your
5380 desired target is already available) also edit @file{gdb/configure.tgt},
5381 setting @code{gdb_target} to something appropriate (for instance,
5384 @emph{Maintainer's note: Work in progress. The file
5385 @file{gdb/configure.host} originally needed to be modified when either a
5386 new native target or a new host machine was being added to @value{GDBN}.
5387 Recent changes have removed this requirement. The file now only needs
5388 to be modified when adding a new native configuration. This will likely
5389 changed again in the future.}
5392 Finally, you'll need to specify and define @value{GDBN}'s host-, native-, and
5393 target-dependent @file{.h} and @file{.c} files used for your
5397 @node Versions and Branches
5398 @chapter Versions and Branches
5402 @value{GDBN}'s version is determined by the file
5403 @file{gdb/version.in} and takes one of the following forms:
5406 @item @var{major}.@var{minor}
5407 @itemx @var{major}.@var{minor}.@var{patchlevel}
5408 an official release (e.g., 6.2 or 6.2.1)
5409 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}
5410 a snapshot taken at @var{YYYY}-@var{MM}-@var{DD}-gmt (e.g.,
5411 6.1.50.20020302, 6.1.90.20020304, or 6.1.0.20020308)
5412 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}-cvs
5413 a @sc{cvs} check out drawn on @var{YYYY}-@var{MM}-@var{DD} (e.g.,
5414 6.1.50.20020302-cvs, 6.1.90.20020304-cvs, or 6.1.0.20020308-cvs)
5415 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD} (@var{vendor})
5416 a vendor specific release of @value{GDBN}, that while based on@*
5417 @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD},
5418 may include additional changes
5421 @value{GDBN}'s mainline uses the @var{major} and @var{minor} version
5422 numbers from the most recent release branch, with a @var{patchlevel}
5423 of 50. At the time each new release branch is created, the mainline's
5424 @var{major} and @var{minor} version numbers are updated.
5426 @value{GDBN}'s release branch is similar. When the branch is cut, the
5427 @var{patchlevel} is changed from 50 to 90. As draft releases are
5428 drawn from the branch, the @var{patchlevel} is incremented. Once the
5429 first release (@var{major}.@var{minor}) has been made, the
5430 @var{patchlevel} is set to 0 and updates have an incremented
5433 For snapshots, and @sc{cvs} check outs, it is also possible to
5434 identify the @sc{cvs} origin:
5437 @item @var{major}.@var{minor}.50.@var{YYYY}@var{MM}@var{DD}
5438 drawn from the @sc{head} of mainline @sc{cvs} (e.g., 6.1.50.20020302)
5439 @item @var{major}.@var{minor}.90.@var{YYYY}@var{MM}@var{DD}
5440 @itemx @var{major}.@var{minor}.91.@var{YYYY}@var{MM}@var{DD} @dots{}
5441 drawn from a release branch prior to the release (e.g.,
5443 @item @var{major}.@var{minor}.0.@var{YYYY}@var{MM}@var{DD}
5444 @itemx @var{major}.@var{minor}.1.@var{YYYY}@var{MM}@var{DD} @dots{}
5445 drawn from a release branch after the release (e.g., 6.2.0.20020308)
5448 If the previous @value{GDBN} version is 6.1 and the current version is
5449 6.2, then, substituting 6 for @var{major} and 1 or 2 for @var{minor},
5450 here's an illustration of a typical sequence:
5457 +--------------------------.
5460 6.2.50.20020303-cvs 6.1.90 (draft #1)
5462 6.2.50.20020304-cvs 6.1.90.20020304-cvs
5464 6.2.50.20020305-cvs 6.1.91 (draft #2)
5466 6.2.50.20020306-cvs 6.1.91.20020306-cvs
5468 6.2.50.20020307-cvs 6.2 (release)
5470 6.2.50.20020308-cvs 6.2.0.20020308-cvs
5472 6.2.50.20020309-cvs 6.2.1 (update)
5474 6.2.50.20020310-cvs <branch closed>
5478 +--------------------------.
5481 6.3.50.20020312-cvs 6.2.90 (draft #1)
5485 @section Release Branches
5486 @cindex Release Branches
5488 @value{GDBN} draws a release series (6.2, 6.2.1, @dots{}) from a
5489 single release branch, and identifies that branch using the @sc{cvs}
5493 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-branchpoint
5494 gdb_@var{major}_@var{minor}-branch
5495 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-release
5498 @emph{Pragmatics: To help identify the date at which a branch or
5499 release is made, both the branchpoint and release tags include the
5500 date that they are cut (@var{YYYY}@var{MM}@var{DD}) in the tag. The
5501 branch tag, denoting the head of the branch, does not need this.}
5503 @section Vendor Branches
5504 @cindex vendor branches
5506 To avoid version conflicts, vendors are expected to modify the file
5507 @file{gdb/version.in} to include a vendor unique alphabetic identifier
5508 (an official @value{GDBN} release never uses alphabetic characters in
5509 its version identifer). E.g., @samp{6.2widgit2}, or @samp{6.2 (Widgit
5512 @section Experimental Branches
5513 @cindex experimental branches
5515 @subsection Guidelines
5517 @value{GDBN} permits the creation of branches, cut from the @sc{cvs}
5518 repository, for experimental development. Branches make it possible
5519 for developers to share preliminary work, and maintainers to examine
5520 significant new developments.
5522 The following are a set of guidelines for creating such branches:
5526 @item a branch has an owner
5527 The owner can set further policy for a branch, but may not change the
5528 ground rules. In particular, they can set a policy for commits (be it
5529 adding more reviewers or deciding who can commit).
5531 @item all commits are posted
5532 All changes committed to a branch shall also be posted to
5533 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} patches
5534 mailing list}. While commentary on such changes are encouraged, people
5535 should remember that the changes only apply to a branch.
5537 @item all commits are covered by an assignment
5538 This ensures that all changes belong to the Free Software Foundation,
5539 and avoids the possibility that the branch may become contaminated.
5541 @item a branch is focused
5542 A focused branch has a single objective or goal, and does not contain
5543 unnecessary or irrelevant changes. Cleanups, where identified, being
5544 be pushed into the mainline as soon as possible.
5546 @item a branch tracks mainline
5547 This keeps the level of divergence under control. It also keeps the
5548 pressure on developers to push cleanups and other stuff into the
5551 @item a branch shall contain the entire @value{GDBN} module
5552 The @value{GDBN} module @code{gdb} should be specified when creating a
5553 branch (branches of individual files should be avoided). @xref{Tags}.
5555 @item a branch shall be branded using @file{version.in}
5556 The file @file{gdb/version.in} shall be modified so that it identifies
5557 the branch @var{owner} and branch @var{name}, e.g.,
5558 @samp{6.2.50.20030303_owner_name} or @samp{6.2 (Owner Name)}.
5565 To simplify the identification of @value{GDBN} branches, the following
5566 branch tagging convention is strongly recommended:
5570 @item @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5571 @itemx @var{owner}_@var{name}-@var{YYYYMMDD}-branch
5572 The branch point and corresponding branch tag. @var{YYYYMMDD} is the
5573 date that the branch was created. A branch is created using the
5574 sequence: @anchor{experimental branch tags}
5576 cvs rtag @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint gdb
5577 cvs rtag -b -r @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint \
5578 @var{owner}_@var{name}-@var{YYYYMMDD}-branch gdb
5581 @item @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5582 The tagged point, on the mainline, that was used when merging the branch
5583 on @var{yyyymmdd}. To merge in all changes since the branch was cut,
5584 use a command sequence like:
5586 cvs rtag @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint gdb
5588 -j@var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5589 -j@var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5592 Similar sequences can be used to just merge in changes since the last
5598 For further information on @sc{cvs}, see
5599 @uref{http://www.gnu.org/software/cvs/, Concurrent Versions System}.
5603 @chapter Releasing @value{GDBN}
5604 @cindex making a new release of gdb
5606 @section Branch Commit Policy
5608 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
5609 5.1 and 5.2 all used the below:
5613 The @file{gdb/MAINTAINERS} file still holds.
5615 Don't fix something on the branch unless/until it is also fixed in the
5616 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
5617 file is better than committing a hack.
5619 When considering a patch for the branch, suggested criteria include:
5620 Does it fix a build? Does it fix the sequence @kbd{break main; run}
5621 when debugging a static binary?
5623 The further a change is from the core of @value{GDBN}, the less likely
5624 the change will worry anyone (e.g., target specific code).
5626 Only post a proposal to change the core of @value{GDBN} after you've
5627 sent individual bribes to all the people listed in the
5628 @file{MAINTAINERS} file @t{;-)}
5631 @emph{Pragmatics: Provided updates are restricted to non-core
5632 functionality there is little chance that a broken change will be fatal.
5633 This means that changes such as adding a new architectures or (within
5634 reason) support for a new host are considered acceptable.}
5637 @section Obsoleting code
5639 Before anything else, poke the other developers (and around the source
5640 code) to see if there is anything that can be removed from @value{GDBN}
5641 (an old target, an unused file).
5643 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
5644 line. Doing this means that it is easy to identify something that has
5645 been obsoleted when greping through the sources.
5647 The process is done in stages --- this is mainly to ensure that the
5648 wider @value{GDBN} community has a reasonable opportunity to respond.
5649 Remember, everything on the Internet takes a week.
5653 Post the proposal on @email{gdb@@sources.redhat.com, the GDB mailing
5654 list} Creating a bug report to track the task's state, is also highly
5659 Post the proposal on @email{gdb-announce@@sources.redhat.com, the GDB
5660 Announcement mailing list}.
5664 Go through and edit all relevant files and lines so that they are
5665 prefixed with the word @code{OBSOLETE}.
5667 Wait until the next GDB version, containing this obsolete code, has been
5670 Remove the obsolete code.
5674 @emph{Maintainer note: While removing old code is regrettable it is
5675 hopefully better for @value{GDBN}'s long term development. Firstly it
5676 helps the developers by removing code that is either no longer relevant
5677 or simply wrong. Secondly since it removes any history associated with
5678 the file (effectively clearing the slate) the developer has a much freer
5679 hand when it comes to fixing broken files.}
5683 @section Before the Branch
5685 The most important objective at this stage is to find and fix simple
5686 changes that become a pain to track once the branch is created. For
5687 instance, configuration problems that stop @value{GDBN} from even
5688 building. If you can't get the problem fixed, document it in the
5689 @file{gdb/PROBLEMS} file.
5691 @subheading Prompt for @file{gdb/NEWS}
5693 People always forget. Send a post reminding them but also if you know
5694 something interesting happened add it yourself. The @code{schedule}
5695 script will mention this in its e-mail.
5697 @subheading Review @file{gdb/README}
5699 Grab one of the nightly snapshots and then walk through the
5700 @file{gdb/README} looking for anything that can be improved. The
5701 @code{schedule} script will mention this in its e-mail.
5703 @subheading Refresh any imported files.
5705 A number of files are taken from external repositories. They include:
5709 @file{texinfo/texinfo.tex}
5711 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
5714 @file{etc/standards.texi}, @file{etc/make-stds.texi}
5717 @subheading Check the ARI
5719 @uref{http://sources.redhat.com/gdb/ari,,A.R.I.} is an @code{awk} script
5720 (Awk Regression Index ;-) that checks for a number of errors and coding
5721 conventions. The checks include things like using @code{malloc} instead
5722 of @code{xmalloc} and file naming problems. There shouldn't be any
5725 @subsection Review the bug data base
5727 Close anything obviously fixed.
5729 @subsection Check all cross targets build
5731 The targets are listed in @file{gdb/MAINTAINERS}.
5734 @section Cut the Branch
5736 @subheading Create the branch
5741 $ V=`echo $v | sed 's/\./_/g'`
5742 $ D=`date -u +%Y-%m-%d`
5745 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
5746 -D $D-gmt gdb_$V-$D-branchpoint insight+dejagnu
5747 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag
5748 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight+dejagnu
5751 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
5752 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight+dejagnu
5753 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
5754 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight+dejagnu
5762 by using @kbd{-D YYYY-MM-DD-gmt} the branch is forced to an exact
5765 the trunk is first taged so that the branch point can easily be found
5767 Insight (which includes GDB) and dejagnu are all tagged at the same time
5769 @file{version.in} gets bumped to avoid version number conflicts
5771 the reading of @file{.cvsrc} is disabled using @file{-f}
5774 @subheading Update @file{version.in}
5779 $ V=`echo $v | sed 's/\./_/g'`
5783 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src co \
5784 -r gdb_$V-branch src/gdb/version.in
5785 cvs -f -d :ext:sources.redhat.com:/cvs/src co
5786 -r gdb_5_2-branch src/gdb/version.in
5788 U src/gdb/version.in
5790 $ echo $u.90-0000-00-00-cvs > version.in
5792 5.1.90-0000-00-00-cvs
5793 $ cvs -f commit version.in
5798 @file{0000-00-00} is used as a date to pump prime the version.in update
5801 @file{.90} and the previous branch version are used as fairly arbitrary
5802 initial branch version number
5806 @subheading Update the web and news pages
5810 @subheading Tweak cron to track the new branch
5812 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
5813 This file needs to be updated so that:
5817 a daily timestamp is added to the file @file{version.in}
5819 the new branch is included in the snapshot process
5823 See the file @file{gdbadmin/cron/README} for how to install the updated
5826 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
5827 any changes. That file is copied to both the branch/ and current/
5828 snapshot directories.
5831 @subheading Update the NEWS and README files
5833 The @file{NEWS} file needs to be updated so that on the branch it refers
5834 to @emph{changes in the current release} while on the trunk it also
5835 refers to @emph{changes since the current release}.
5837 The @file{README} file needs to be updated so that it refers to the
5840 @subheading Post the branch info
5842 Send an announcement to the mailing lists:
5846 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
5848 @email{gdb@@sources.redhat.com, GDB Discsussion mailing list} and
5849 @email{gdb-testers@@sources.redhat.com, GDB Discsussion mailing list}
5852 @emph{Pragmatics: The branch creation is sent to the announce list to
5853 ensure that people people not subscribed to the higher volume discussion
5856 The announcement should include:
5862 how to check out the branch using CVS
5864 the date/number of weeks until the release
5866 the branch commit policy
5870 @section Stabilize the branch
5872 Something goes here.
5874 @section Create a Release
5876 The process of creating and then making available a release is broken
5877 down into a number of stages. The first part addresses the technical
5878 process of creating a releasable tar ball. The later stages address the
5879 process of releasing that tar ball.
5881 When making a release candidate just the first section is needed.
5883 @subsection Create a release candidate
5885 The objective at this stage is to create a set of tar balls that can be
5886 made available as a formal release (or as a less formal release
5889 @subsubheading Freeze the branch
5891 Send out an e-mail notifying everyone that the branch is frozen to
5892 @email{gdb-patches@@sources.redhat.com}.
5894 @subsubheading Establish a few defaults.
5899 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
5901 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
5905 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
5907 /home/gdbadmin/bin/autoconf
5916 Check the @code{autoconf} version carefully. You want to be using the
5917 version taken from the @file{binutils} snapshot directory, which can be
5918 found at @uref{ftp://sources.redhat.com/pub/binutils/}. It is very
5919 unlikely that a system installed version of @code{autoconf} (e.g.,
5920 @file{/usr/bin/autoconf}) is correct.
5923 @subsubheading Check out the relevant modules:
5926 $ for m in gdb insight dejagnu
5928 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
5938 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
5939 any confusion between what is written here and what your local
5940 @code{cvs} really does.
5943 @subsubheading Update relevant files.
5949 Major releases get their comments added as part of the mainline. Minor
5950 releases should probably mention any significant bugs that were fixed.
5952 Don't forget to include the @file{ChangeLog} entry.
5955 $ emacs gdb/src/gdb/NEWS
5960 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
5961 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
5966 You'll need to update:
5978 $ emacs gdb/src/gdb/README
5983 $ cp gdb/src/gdb/README insight/src/gdb/README
5984 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
5987 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
5988 before the initial branch was cut so just a simple substitute is needed
5991 @emph{Maintainer note: Other projects generate @file{README} and
5992 @file{INSTALL} from the core documentation. This might be worth
5995 @item gdb/version.in
5998 $ echo $v > gdb/src/gdb/version.in
5999 $ cat gdb/src/gdb/version.in
6001 $ emacs gdb/src/gdb/version.in
6004 ... Bump to version ...
6006 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
6007 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6010 @item dejagnu/src/dejagnu/configure.in
6012 Dejagnu is more complicated. The version number is a parameter to
6013 @code{AM_INIT_AUTOMAKE}. Tweak it to read something like gdb-5.1.91.
6015 Don't forget to re-generate @file{configure}.
6017 Don't forget to include a @file{ChangeLog} entry.
6020 $ emacs dejagnu/src/dejagnu/configure.in
6025 $ ( cd dejagnu/src/dejagnu && autoconf )
6030 @subsubheading Do the dirty work
6032 This is identical to the process used to create the daily snapshot.
6035 $ for m in gdb insight
6037 ( cd $m/src && gmake -f src-release $m.tar )
6039 $ ( m=dejagnu; cd $m/src && gmake -f src-release $m.tar.bz2 )
6042 If the top level source directory does not have @file{src-release}
6043 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
6046 $ for m in gdb insight
6048 ( cd $m/src && gmake -f Makefile.in $m.tar )
6050 $ ( m=dejagnu; cd $m/src && gmake -f Makefile.in $m.tar.bz2 )
6053 @subsubheading Check the source files
6055 You're looking for files that have mysteriously disappeared.
6056 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
6057 for the @file{version.in} update @kbd{cronjob}.
6060 $ ( cd gdb/src && cvs -f -q -n update )
6064 @dots{} lots of generated files @dots{}
6069 @dots{} lots of generated files @dots{}
6074 @emph{Don't worry about the @file{gdb.info-??} or
6075 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
6076 was also generated only something strange with CVS means that they
6077 didn't get supressed). Fixing it would be nice though.}
6079 @subsubheading Create compressed versions of the release
6085 dejagnu/ dejagnu-gdb-5.2.tar.bz2 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
6086 $ for m in gdb insight
6088 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
6089 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
6099 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
6100 in that mode, @code{gzip} does not know the name of the file and, hence,
6101 can not include it in the compressed file. This is also why the release
6102 process runs @code{tar} and @code{bzip2} as separate passes.
6105 @subsection Sanity check the tar ball
6107 Pick a popular machine (Solaris/PPC?) and try the build on that.
6110 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
6115 $ ./gdb/gdb ./gdb/gdb
6119 Breakpoint 1 at 0x80732bc: file main.c, line 734.
6121 Starting program: /tmp/gdb-5.2/gdb/gdb
6123 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
6124 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
6126 $1 = @{argc = 136426532, argv = 0x821b7f0@}
6130 @subsection Make a release candidate available
6132 If this is a release candidate then the only remaining steps are:
6136 Commit @file{version.in} and @file{ChangeLog}
6138 Tweak @file{version.in} (and @file{ChangeLog} to read
6139 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
6140 process can restart.
6142 Make the release candidate available in
6143 @uref{ftp://sources.redhat.com/pub/gdb/snapshots/branch}
6145 Notify the relevant mailing lists ( @email{gdb@@sources.redhat.com} and
6146 @email{gdb-testers@@sources.redhat.com} that the candidate is available.
6149 @subsection Make a formal release available
6151 (And you thought all that was required was to post an e-mail.)
6153 @subsubheading Install on sware
6155 Copy the new files to both the release and the old release directory:
6158 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
6159 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
6163 Clean up the releases directory so that only the most recent releases
6164 are available (e.g. keep 5.2 and 5.2.1 but remove 5.1):
6167 $ cd ~ftp/pub/gdb/releases
6172 Update the file @file{README} and @file{.message} in the releases
6179 $ ln README .message
6182 @subsubheading Update the web pages.
6186 @item htdocs/download/ANNOUNCEMENT
6187 This file, which is posted as the official announcement, includes:
6190 General announcement
6192 News. If making an @var{M}.@var{N}.1 release, retain the news from
6193 earlier @var{M}.@var{N} release.
6198 @item htdocs/index.html
6199 @itemx htdocs/news/index.html
6200 @itemx htdocs/download/index.html
6201 These files include:
6204 announcement of the most recent release
6206 news entry (remember to update both the top level and the news directory).
6208 These pages also need to be regenerate using @code{index.sh}.
6210 @item download/onlinedocs/
6211 You need to find the magic command that is used to generate the online
6212 docs from the @file{.tar.bz2}. The best way is to look in the output
6213 from one of the nightly @code{cron} jobs and then just edit accordingly.
6217 $ ~/ss/update-web-docs \
6218 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6220 /www/sourceware/htdocs/gdb/download/onlinedocs \
6225 Just like the online documentation. Something like:
6228 $ /bin/sh ~/ss/update-web-ari \
6229 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6231 /www/sourceware/htdocs/gdb/download/ari \
6237 @subsubheading Shadow the pages onto gnu
6239 Something goes here.
6242 @subsubheading Install the @value{GDBN} tar ball on GNU
6244 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
6245 @file{~ftp/gnu/gdb}.
6247 @subsubheading Make the @file{ANNOUNCEMENT}
6249 Post the @file{ANNOUNCEMENT} file you created above to:
6253 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6255 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
6256 day or so to let things get out)
6258 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
6263 The release is out but you're still not finished.
6265 @subsubheading Commit outstanding changes
6267 In particular you'll need to commit any changes to:
6271 @file{gdb/ChangeLog}
6273 @file{gdb/version.in}
6280 @subsubheading Tag the release
6285 $ d=`date -u +%Y-%m-%d`
6288 $ ( cd insight/src/gdb && cvs -f -q update )
6289 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
6292 Insight is used since that contains more of the release than
6293 @value{GDBN} (@code{dejagnu} doesn't get tagged but I think we can live
6296 @subsubheading Mention the release on the trunk
6298 Just put something in the @file{ChangeLog} so that the trunk also
6299 indicates when the release was made.
6301 @subsubheading Restart @file{gdb/version.in}
6303 If @file{gdb/version.in} does not contain an ISO date such as
6304 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
6305 committed all the release changes it can be set to
6306 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
6307 is important - it affects the snapshot process).
6309 Don't forget the @file{ChangeLog}.
6311 @subsubheading Merge into trunk
6313 The files committed to the branch may also need changes merged into the
6316 @subsubheading Revise the release schedule
6318 Post a revised release schedule to @email{gdb@@sources.redhat.com, GDB
6319 Discussion List} with an updated announcement. The schedule can be
6320 generated by running:
6323 $ ~/ss/schedule `date +%s` schedule
6327 The first parameter is approximate date/time in seconds (from the epoch)
6328 of the most recent release.
6330 Also update the schedule @code{cronjob}.
6332 @section Post release
6334 Remove any @code{OBSOLETE} code.
6341 The testsuite is an important component of the @value{GDBN} package.
6342 While it is always worthwhile to encourage user testing, in practice
6343 this is rarely sufficient; users typically use only a small subset of
6344 the available commands, and it has proven all too common for a change
6345 to cause a significant regression that went unnoticed for some time.
6347 The @value{GDBN} testsuite uses the DejaGNU testing framework.
6348 DejaGNU is built using @code{Tcl} and @code{expect}. The tests
6349 themselves are calls to various @code{Tcl} procs; the framework runs all the
6350 procs and summarizes the passes and fails.
6352 @section Using the Testsuite
6354 @cindex running the test suite
6355 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
6356 testsuite's objdir) and type @code{make check}. This just sets up some
6357 environment variables and invokes DejaGNU's @code{runtest} script. While
6358 the testsuite is running, you'll get mentions of which test file is in use,
6359 and a mention of any unexpected passes or fails. When the testsuite is
6360 finished, you'll get a summary that looks like this:
6365 # of expected passes 6016
6366 # of unexpected failures 58
6367 # of unexpected successes 5
6368 # of expected failures 183
6369 # of unresolved testcases 3
6370 # of untested testcases 5
6373 The ideal test run consists of expected passes only; however, reality
6374 conspires to keep us from this ideal. Unexpected failures indicate
6375 real problems, whether in @value{GDBN} or in the testsuite. Expected
6376 failures are still failures, but ones which have been decided are too
6377 hard to deal with at the time; for instance, a test case might work
6378 everywhere except on AIX, and there is no prospect of the AIX case
6379 being fixed in the near future. Expected failures should not be added
6380 lightly, since you may be masking serious bugs in @value{GDBN}.
6381 Unexpected successes are expected fails that are passing for some
6382 reason, while unresolved and untested cases often indicate some minor
6383 catastrophe, such as the compiler being unable to deal with a test
6386 When making any significant change to @value{GDBN}, you should run the
6387 testsuite before and after the change, to confirm that there are no
6388 regressions. Note that truly complete testing would require that you
6389 run the testsuite with all supported configurations and a variety of
6390 compilers; however this is more than really necessary. In many cases
6391 testing with a single configuration is sufficient. Other useful
6392 options are to test one big-endian (Sparc) and one little-endian (x86)
6393 host, a cross config with a builtin simulator (powerpc-eabi,
6394 mips-elf), or a 64-bit host (Alpha).
6396 If you add new functionality to @value{GDBN}, please consider adding
6397 tests for it as well; this way future @value{GDBN} hackers can detect
6398 and fix their changes that break the functionality you added.
6399 Similarly, if you fix a bug that was not previously reported as a test
6400 failure, please add a test case for it. Some cases are extremely
6401 difficult to test, such as code that handles host OS failures or bugs
6402 in particular versions of compilers, and it's OK not to try to write
6403 tests for all of those.
6405 DejaGNU supports separate build, host, and target machines. However,
6406 some @value{GDBN} test scripts do not work if the build machine and
6407 the host machine are not the same. In such an environment, these scripts
6408 will give a result of ``UNRESOLVED'', like this:
6411 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
6414 @section Testsuite Organization
6416 @cindex test suite organization
6417 The testsuite is entirely contained in @file{gdb/testsuite}. While the
6418 testsuite includes some makefiles and configury, these are very minimal,
6419 and used for little besides cleaning up, since the tests themselves
6420 handle the compilation of the programs that @value{GDBN} will run. The file
6421 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
6422 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
6423 configuration-specific files, typically used for special-purpose
6424 definitions of procs like @code{gdb_load} and @code{gdb_start}.
6426 The tests themselves are to be found in @file{testsuite/gdb.*} and
6427 subdirectories of those. The names of the test files must always end
6428 with @file{.exp}. DejaGNU collects the test files by wildcarding
6429 in the test directories, so both subdirectories and individual files
6430 get chosen and run in alphabetical order.
6432 The following table lists the main types of subdirectories and what they
6433 are for. Since DejaGNU finds test files no matter where they are
6434 located, and since each test file sets up its own compilation and
6435 execution environment, this organization is simply for convenience and
6440 This is the base testsuite. The tests in it should apply to all
6441 configurations of @value{GDBN} (but generic native-only tests may live here).
6442 The test programs should be in the subset of C that is valid K&R,
6443 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
6446 @item gdb.@var{lang}
6447 Language-specific tests for any language @var{lang} besides C. Examples are
6448 @file{gdb.cp} and @file{gdb.java}.
6450 @item gdb.@var{platform}
6451 Non-portable tests. The tests are specific to a specific configuration
6452 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
6455 @item gdb.@var{compiler}
6456 Tests specific to a particular compiler. As of this writing (June
6457 1999), there aren't currently any groups of tests in this category that
6458 couldn't just as sensibly be made platform-specific, but one could
6459 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
6462 @item gdb.@var{subsystem}
6463 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
6464 instance, @file{gdb.disasm} exercises various disassemblers, while
6465 @file{gdb.stabs} tests pathways through the stabs symbol reader.
6468 @section Writing Tests
6469 @cindex writing tests
6471 In many areas, the @value{GDBN} tests are already quite comprehensive; you
6472 should be able to copy existing tests to handle new cases.
6474 You should try to use @code{gdb_test} whenever possible, since it
6475 includes cases to handle all the unexpected errors that might happen.
6476 However, it doesn't cost anything to add new test procedures; for
6477 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
6478 calls @code{gdb_test} multiple times.
6480 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
6481 necessary, such as when @value{GDBN} has several valid responses to a command.
6483 The source language programs do @emph{not} need to be in a consistent
6484 style. Since @value{GDBN} is used to debug programs written in many different
6485 styles, it's worth having a mix of styles in the testsuite; for
6486 instance, some @value{GDBN} bugs involving the display of source lines would
6487 never manifest themselves if the programs used GNU coding style
6494 Check the @file{README} file, it often has useful information that does not
6495 appear anywhere else in the directory.
6498 * Getting Started:: Getting started working on @value{GDBN}
6499 * Debugging GDB:: Debugging @value{GDBN} with itself
6502 @node Getting Started,,, Hints
6504 @section Getting Started
6506 @value{GDBN} is a large and complicated program, and if you first starting to
6507 work on it, it can be hard to know where to start. Fortunately, if you
6508 know how to go about it, there are ways to figure out what is going on.
6510 This manual, the @value{GDBN} Internals manual, has information which applies
6511 generally to many parts of @value{GDBN}.
6513 Information about particular functions or data structures are located in
6514 comments with those functions or data structures. If you run across a
6515 function or a global variable which does not have a comment correctly
6516 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
6517 free to submit a bug report, with a suggested comment if you can figure
6518 out what the comment should say. If you find a comment which is
6519 actually wrong, be especially sure to report that.
6521 Comments explaining the function of macros defined in host, target, or
6522 native dependent files can be in several places. Sometimes they are
6523 repeated every place the macro is defined. Sometimes they are where the
6524 macro is used. Sometimes there is a header file which supplies a
6525 default definition of the macro, and the comment is there. This manual
6526 also documents all the available macros.
6527 @c (@pxref{Host Conditionals}, @pxref{Target
6528 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
6531 Start with the header files. Once you have some idea of how
6532 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
6533 @file{gdbtypes.h}), you will find it much easier to understand the
6534 code which uses and creates those symbol tables.
6536 You may wish to process the information you are getting somehow, to
6537 enhance your understanding of it. Summarize it, translate it to another
6538 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
6539 the code to predict what a test case would do and write the test case
6540 and verify your prediction, etc. If you are reading code and your eyes
6541 are starting to glaze over, this is a sign you need to use a more active
6544 Once you have a part of @value{GDBN} to start with, you can find more
6545 specifically the part you are looking for by stepping through each
6546 function with the @code{next} command. Do not use @code{step} or you
6547 will quickly get distracted; when the function you are stepping through
6548 calls another function try only to get a big-picture understanding
6549 (perhaps using the comment at the beginning of the function being
6550 called) of what it does. This way you can identify which of the
6551 functions being called by the function you are stepping through is the
6552 one which you are interested in. You may need to examine the data
6553 structures generated at each stage, with reference to the comments in
6554 the header files explaining what the data structures are supposed to
6557 Of course, this same technique can be used if you are just reading the
6558 code, rather than actually stepping through it. The same general
6559 principle applies---when the code you are looking at calls something
6560 else, just try to understand generally what the code being called does,
6561 rather than worrying about all its details.
6563 @cindex command implementation
6564 A good place to start when tracking down some particular area is with
6565 a command which invokes that feature. Suppose you want to know how
6566 single-stepping works. As a @value{GDBN} user, you know that the
6567 @code{step} command invokes single-stepping. The command is invoked
6568 via command tables (see @file{command.h}); by convention the function
6569 which actually performs the command is formed by taking the name of
6570 the command and adding @samp{_command}, or in the case of an
6571 @code{info} subcommand, @samp{_info}. For example, the @code{step}
6572 command invokes the @code{step_command} function and the @code{info
6573 display} command invokes @code{display_info}. When this convention is
6574 not followed, you might have to use @code{grep} or @kbd{M-x
6575 tags-search} in emacs, or run @value{GDBN} on itself and set a
6576 breakpoint in @code{execute_command}.
6578 @cindex @code{bug-gdb} mailing list
6579 If all of the above fail, it may be appropriate to ask for information
6580 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
6581 wondering if anyone could give me some tips about understanding
6582 @value{GDBN}''---if we had some magic secret we would put it in this manual.
6583 Suggestions for improving the manual are always welcome, of course.
6585 @node Debugging GDB,,,Hints
6587 @section Debugging @value{GDBN} with itself
6588 @cindex debugging @value{GDBN}
6590 If @value{GDBN} is limping on your machine, this is the preferred way to get it
6591 fully functional. Be warned that in some ancient Unix systems, like
6592 Ultrix 4.2, a program can't be running in one process while it is being
6593 debugged in another. Rather than typing the command @kbd{@w{./gdb
6594 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
6595 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
6597 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
6598 @file{.gdbinit} file that sets up some simple things to make debugging
6599 gdb easier. The @code{info} command, when executed without a subcommand
6600 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
6601 gdb. See @file{.gdbinit} for details.
6603 If you use emacs, you will probably want to do a @code{make TAGS} after
6604 you configure your distribution; this will put the machine dependent
6605 routines for your local machine where they will be accessed first by
6608 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
6609 have run @code{fixincludes} if you are compiling with gcc.
6611 @section Submitting Patches
6613 @cindex submitting patches
6614 Thanks for thinking of offering your changes back to the community of
6615 @value{GDBN} users. In general we like to get well designed enhancements.
6616 Thanks also for checking in advance about the best way to transfer the
6619 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
6620 This manual summarizes what we believe to be clean design for @value{GDBN}.
6622 If the maintainers don't have time to put the patch in when it arrives,
6623 or if there is any question about a patch, it goes into a large queue
6624 with everyone else's patches and bug reports.
6626 @cindex legal papers for code contributions
6627 The legal issue is that to incorporate substantial changes requires a
6628 copyright assignment from you and/or your employer, granting ownership
6629 of the changes to the Free Software Foundation. You can get the
6630 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
6631 and asking for it. We recommend that people write in "All programs
6632 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
6633 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
6635 contributed with only one piece of legalese pushed through the
6636 bureaucracy and filed with the FSF. We can't start merging changes until
6637 this paperwork is received by the FSF (their rules, which we follow
6638 since we maintain it for them).
6640 Technically, the easiest way to receive changes is to receive each
6641 feature as a small context diff or unidiff, suitable for @code{patch}.
6642 Each message sent to me should include the changes to C code and
6643 header files for a single feature, plus @file{ChangeLog} entries for
6644 each directory where files were modified, and diffs for any changes
6645 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
6646 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
6647 single feature, they can be split down into multiple messages.
6649 In this way, if we read and like the feature, we can add it to the
6650 sources with a single patch command, do some testing, and check it in.
6651 If you leave out the @file{ChangeLog}, we have to write one. If you leave
6652 out the doc, we have to puzzle out what needs documenting. Etc., etc.
6654 The reason to send each change in a separate message is that we will not
6655 install some of the changes. They'll be returned to you with questions
6656 or comments. If we're doing our job correctly, the message back to you
6657 will say what you have to fix in order to make the change acceptable.
6658 The reason to have separate messages for separate features is so that
6659 the acceptable changes can be installed while one or more changes are
6660 being reworked. If multiple features are sent in a single message, we
6661 tend to not put in the effort to sort out the acceptable changes from
6662 the unacceptable, so none of the features get installed until all are
6665 If this sounds painful or authoritarian, well, it is. But we get a lot
6666 of bug reports and a lot of patches, and many of them don't get
6667 installed because we don't have the time to finish the job that the bug
6668 reporter or the contributor could have done. Patches that arrive
6669 complete, working, and well designed, tend to get installed on the day
6670 they arrive. The others go into a queue and get installed as time
6671 permits, which, since the maintainers have many demands to meet, may not
6672 be for quite some time.
6674 Please send patches directly to
6675 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} maintainers}.
6677 @section Obsolete Conditionals
6678 @cindex obsolete code
6680 Fragments of old code in @value{GDBN} sometimes reference or set the following
6681 configuration macros. They should not be used by new code, and old uses
6682 should be removed as those parts of the debugger are otherwise touched.
6685 @item STACK_END_ADDR
6686 This macro used to define where the end of the stack appeared, for use
6687 in interpreting core file formats that don't record this address in the
6688 core file itself. This information is now configured in BFD, and @value{GDBN}
6689 gets the info portably from there. The values in @value{GDBN}'s configuration
6690 files should be moved into BFD configuration files (if needed there),
6691 and deleted from all of @value{GDBN}'s config files.
6693 Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR
6694 is so old that it has never been converted to use BFD. Now that's old!
6698 @include observer.texi