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::
89 * Start of New Year Procedure::
94 * GDB Observers:: @value{GDBN} Currently available observers
95 * GNU Free Documentation License:: The license for this documentation
101 @chapter Requirements
102 @cindex requirements for @value{GDBN}
104 Before diving into the internals, you should understand the formal
105 requirements and other expectations for @value{GDBN}. Although some
106 of these may seem obvious, there have been proposals for @value{GDBN}
107 that have run counter to these requirements.
109 First of all, @value{GDBN} is a debugger. It's not designed to be a
110 front panel for embedded systems. It's not a text editor. It's not a
111 shell. It's not a programming environment.
113 @value{GDBN} is an interactive tool. Although a batch mode is
114 available, @value{GDBN}'s primary role is to interact with a human
117 @value{GDBN} should be responsive to the user. A programmer hot on
118 the trail of a nasty bug, and operating under a looming deadline, is
119 going to be very impatient of everything, including the response time
120 to debugger commands.
122 @value{GDBN} should be relatively permissive, such as for expressions.
123 While the compiler should be picky (or have the option to be made
124 picky), since source code lives for a long time usually, the
125 programmer doing debugging shouldn't be spending time figuring out to
126 mollify the debugger.
128 @value{GDBN} will be called upon to deal with really large programs.
129 Executable sizes of 50 to 100 megabytes occur regularly, and we've
130 heard reports of programs approaching 1 gigabyte in size.
132 @value{GDBN} should be able to run everywhere. No other debugger is
133 available for even half as many configurations as @value{GDBN}
137 @node Overall Structure
139 @chapter Overall Structure
141 @value{GDBN} consists of three major subsystems: user interface,
142 symbol handling (the @dfn{symbol side}), and target system handling (the
145 The user interface consists of several actual interfaces, plus
148 The symbol side consists of object file readers, debugging info
149 interpreters, symbol table management, source language expression
150 parsing, type and value printing.
152 The target side consists of execution control, stack frame analysis, and
153 physical target manipulation.
155 The target side/symbol side division is not formal, and there are a
156 number of exceptions. For instance, core file support involves symbolic
157 elements (the basic core file reader is in BFD) and target elements (it
158 supplies the contents of memory and the values of registers). Instead,
159 this division is useful for understanding how the minor subsystems
162 @section The Symbol Side
164 The symbolic side of @value{GDBN} can be thought of as ``everything
165 you can do in @value{GDBN} without having a live program running''.
166 For instance, you can look at the types of variables, and evaluate
167 many kinds of expressions.
169 @section The Target Side
171 The target side of @value{GDBN} is the ``bits and bytes manipulator''.
172 Although it may make reference to symbolic info here and there, most
173 of the target side will run with only a stripped executable
174 available---or even no executable at all, in remote debugging cases.
176 Operations such as disassembly, stack frame crawls, and register
177 display, are able to work with no symbolic info at all. In some cases,
178 such as disassembly, @value{GDBN} will use symbolic info to present addresses
179 relative to symbols rather than as raw numbers, but it will work either
182 @section Configurations
186 @dfn{Host} refers to attributes of the system where @value{GDBN} runs.
187 @dfn{Target} refers to the system where the program being debugged
188 executes. In most cases they are the same machine, in which case a
189 third type of @dfn{Native} attributes come into play.
191 Defines and include files needed to build on the host are host support.
192 Examples are tty support, system defined types, host byte order, host
195 Defines and information needed to handle the target format are target
196 dependent. Examples are the stack frame format, instruction set,
197 breakpoint instruction, registers, and how to set up and tear down the stack
200 Information that is only needed when the host and target are the same,
201 is native dependent. One example is Unix child process support; if the
202 host and target are not the same, doing a fork to start the target
203 process is a bad idea. The various macros needed for finding the
204 registers in the @code{upage}, running @code{ptrace}, and such are all
205 in the native-dependent files.
207 Another example of native-dependent code is support for features that
208 are really part of the target environment, but which require
209 @code{#include} files that are only available on the host system. Core
210 file handling and @code{setjmp} handling are two common cases.
212 When you want to make @value{GDBN} work ``native'' on a particular machine, you
213 have to include all three kinds of information.
221 @value{GDBN} uses a number of debugging-specific algorithms. They are
222 often not very complicated, but get lost in the thicket of special
223 cases and real-world issues. This chapter describes the basic
224 algorithms and mentions some of the specific target definitions that
230 @cindex call stack frame
231 A frame is a construct that @value{GDBN} uses to keep track of calling
232 and called functions.
234 @findex create_new_frame
236 @code{FRAME_FP} in the machine description has no meaning to the
237 machine-independent part of @value{GDBN}, except that it is used when
238 setting up a new frame from scratch, as follows:
241 create_new_frame (read_register (DEPRECATED_FP_REGNUM), read_pc ()));
244 @cindex frame pointer register
245 Other than that, all the meaning imparted to @code{DEPRECATED_FP_REGNUM}
246 is imparted by the machine-dependent code. So,
247 @code{DEPRECATED_FP_REGNUM} can have any value that is convenient for
248 the code that creates new frames. (@code{create_new_frame} calls
249 @code{DEPRECATED_INIT_EXTRA_FRAME_INFO} if it is defined; that is where
250 you should use the @code{DEPRECATED_FP_REGNUM} value, if your frames are
254 Given a @value{GDBN} frame, define @code{DEPRECATED_FRAME_CHAIN} to
255 determine the address of the calling function's frame. This will be
256 used to create a new @value{GDBN} frame struct, and then
257 @code{DEPRECATED_INIT_EXTRA_FRAME_INFO} and
258 @code{DEPRECATED_INIT_FRAME_PC} will be called for the new frame.
260 @section Breakpoint Handling
263 In general, a breakpoint is a user-designated location in the program
264 where the user wants to regain control if program execution ever reaches
267 There are two main ways to implement breakpoints; either as ``hardware''
268 breakpoints or as ``software'' breakpoints.
270 @cindex hardware breakpoints
271 @cindex program counter
272 Hardware breakpoints are sometimes available as a builtin debugging
273 features with some chips. Typically these work by having dedicated
274 register into which the breakpoint address may be stored. If the PC
275 (shorthand for @dfn{program counter})
276 ever matches a value in a breakpoint registers, the CPU raises an
277 exception and reports it to @value{GDBN}.
279 Another possibility is when an emulator is in use; many emulators
280 include circuitry that watches the address lines coming out from the
281 processor, and force it to stop if the address matches a breakpoint's
284 A third possibility is that the target already has the ability to do
285 breakpoints somehow; for instance, a ROM monitor may do its own
286 software breakpoints. So although these are not literally ``hardware
287 breakpoints'', from @value{GDBN}'s point of view they work the same;
288 @value{GDBN} need not do anything more than set the breakpoint and wait
289 for something to happen.
291 Since they depend on hardware resources, hardware breakpoints may be
292 limited in number; when the user asks for more, @value{GDBN} will
293 start trying to set software breakpoints. (On some architectures,
294 notably the 32-bit x86 platforms, @value{GDBN} cannot always know
295 whether there's enough hardware resources to insert all the hardware
296 breakpoints and watchpoints. On those platforms, @value{GDBN} prints
297 an error message only when the program being debugged is continued.)
299 @cindex software breakpoints
300 Software breakpoints require @value{GDBN} to do somewhat more work.
301 The basic theory is that @value{GDBN} will replace a program
302 instruction with a trap, illegal divide, or some other instruction
303 that will cause an exception, and then when it's encountered,
304 @value{GDBN} will take the exception and stop the program. When the
305 user says to continue, @value{GDBN} will restore the original
306 instruction, single-step, re-insert the trap, and continue on.
308 Since it literally overwrites the program being tested, the program area
309 must be writable, so this technique won't work on programs in ROM. It
310 can also distort the behavior of programs that examine themselves,
311 although such a situation would be highly unusual.
313 Also, the software breakpoint instruction should be the smallest size of
314 instruction, so it doesn't overwrite an instruction that might be a jump
315 target, and cause disaster when the program jumps into the middle of the
316 breakpoint instruction. (Strictly speaking, the breakpoint must be no
317 larger than the smallest interval between instructions that may be jump
318 targets; perhaps there is an architecture where only even-numbered
319 instructions may jumped to.) Note that it's possible for an instruction
320 set not to have any instructions usable for a software breakpoint,
321 although in practice only the ARC has failed to define such an
325 The basic definition of the software breakpoint is the macro
328 Basic breakpoint object handling is in @file{breakpoint.c}. However,
329 much of the interesting breakpoint action is in @file{infrun.c}.
331 @section Single Stepping
333 @section Signal Handling
335 @section Thread Handling
337 @section Inferior Function Calls
339 @section Longjmp Support
341 @cindex @code{longjmp} debugging
342 @value{GDBN} has support for figuring out that the target is doing a
343 @code{longjmp} and for stopping at the target of the jump, if we are
344 stepping. This is done with a few specialized internal breakpoints,
345 which are visible in the output of the @samp{maint info breakpoint}
348 @findex GET_LONGJMP_TARGET
349 To make this work, you need to define a macro called
350 @code{GET_LONGJMP_TARGET}, which will examine the @code{jmp_buf}
351 structure and extract the longjmp target address. Since @code{jmp_buf}
352 is target specific, you will need to define it in the appropriate
353 @file{tm-@var{target}.h} file. Look in @file{tm-sun4os4.h} and
354 @file{sparc-tdep.c} for examples of how to do this.
359 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
360 breakpoints}) which break when data is accessed rather than when some
361 instruction is executed. When you have data which changes without
362 your knowing what code does that, watchpoints are the silver bullet to
363 hunt down and kill such bugs.
365 @cindex hardware watchpoints
366 @cindex software watchpoints
367 Watchpoints can be either hardware-assisted or not; the latter type is
368 known as ``software watchpoints.'' @value{GDBN} always uses
369 hardware-assisted watchpoints if they are available, and falls back on
370 software watchpoints otherwise. Typical situations where @value{GDBN}
371 will use software watchpoints are:
375 The watched memory region is too large for the underlying hardware
376 watchpoint support. For example, each x86 debug register can watch up
377 to 4 bytes of memory, so trying to watch data structures whose size is
378 more than 16 bytes will cause @value{GDBN} to use software
382 The value of the expression to be watched depends on data held in
383 registers (as opposed to memory).
386 Too many different watchpoints requested. (On some architectures,
387 this situation is impossible to detect until the debugged program is
388 resumed.) Note that x86 debug registers are used both for hardware
389 breakpoints and for watchpoints, so setting too many hardware
390 breakpoints might cause watchpoint insertion to fail.
393 No hardware-assisted watchpoints provided by the target
397 Software watchpoints are very slow, since @value{GDBN} needs to
398 single-step the program being debugged and test the value of the
399 watched expression(s) after each instruction. The rest of this
400 section is mostly irrelevant for software watchpoints.
402 When the inferior stops, @value{GDBN} tries to establish, among other
403 possible reasons, whether it stopped due to a watchpoint being hit.
404 For a data-write watchpoint, it does so by evaluating, for each
405 watchpoint, the expression whose value is being watched, and testing
406 whether the watched value has changed. For data-read and data-access
407 watchpoints, @value{GDBN} needs the target to supply a primitive that
408 returns the address of the data that was accessed or read (see the
409 description of @code{target_stopped_data_address} below): if this
410 primitive returns a valid address, @value{GDBN} infers that a
411 watchpoint triggered if it watches an expression whose evaluation uses
414 @value{GDBN} uses several macros and primitives to support hardware
418 @findex TARGET_HAS_HARDWARE_WATCHPOINTS
419 @item TARGET_HAS_HARDWARE_WATCHPOINTS
420 If defined, the target supports hardware watchpoints.
422 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
423 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
424 Return the number of hardware watchpoints of type @var{type} that are
425 possible to be set. The value is positive if @var{count} watchpoints
426 of this type can be set, zero if setting watchpoints of this type is
427 not supported, and negative if @var{count} is more than the maximum
428 number of watchpoints of type @var{type} that can be set. @var{other}
429 is non-zero if other types of watchpoints are currently enabled (there
430 are architectures which cannot set watchpoints of different types at
433 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
434 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
435 Return non-zero if hardware watchpoints can be used to watch a region
436 whose address is @var{addr} and whose length in bytes is @var{len}.
438 @findex TARGET_REGION_SIZE_OK_FOR_HW_WATCHPOINT
439 @item TARGET_REGION_SIZE_OK_FOR_HW_WATCHPOINT (@var{size})
440 Return non-zero if hardware watchpoints can be used to watch a region
441 whose size is @var{size}. @value{GDBN} only uses this macro as a
442 fall-back, in case @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is not
445 @cindex insert or remove hardware watchpoint
446 @findex target_insert_watchpoint
447 @findex target_remove_watchpoint
448 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
449 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
450 Insert or remove a hardware watchpoint starting at @var{addr}, for
451 @var{len} bytes. @var{type} is the watchpoint type, one of the
452 possible values of the enumerated data type @code{target_hw_bp_type},
453 defined by @file{breakpoint.h} as follows:
456 enum target_hw_bp_type
458 hw_write = 0, /* Common (write) HW watchpoint */
459 hw_read = 1, /* Read HW watchpoint */
460 hw_access = 2, /* Access (read or write) HW watchpoint */
461 hw_execute = 3 /* Execute HW breakpoint */
466 These two macros should return 0 for success, non-zero for failure.
468 @cindex insert or remove hardware breakpoint
469 @findex target_remove_hw_breakpoint
470 @findex target_insert_hw_breakpoint
471 @item target_remove_hw_breakpoint (@var{addr}, @var{shadow})
472 @itemx target_insert_hw_breakpoint (@var{addr}, @var{shadow})
473 Insert or remove a hardware-assisted breakpoint at address @var{addr}.
474 Returns zero for success, non-zero for failure. @var{shadow} is the
475 real contents of the byte where the breakpoint has been inserted; it
476 is generally not valid when hardware breakpoints are used, but since
477 no other code touches these values, the implementations of the above
478 two macros can use them for their internal purposes.
480 @findex target_stopped_data_address
481 @item target_stopped_data_address (@var{addr_p})
482 If the inferior has some watchpoint that triggered, place the address
483 associated with the watchpoint at the location pointed to by
484 @var{addr_p} and return non-zero. Otherwise, return zero. Note that
485 this primitive is used by @value{GDBN} only on targets that support
486 data-read or data-access type watchpoints, so targets that have
487 support only for data-write watchpoints need not implement these
490 @findex HAVE_STEPPABLE_WATCHPOINT
491 @item HAVE_STEPPABLE_WATCHPOINT
492 If defined to a non-zero value, it is not necessary to disable a
493 watchpoint to step over it.
495 @findex HAVE_NONSTEPPABLE_WATCHPOINT
496 @item HAVE_NONSTEPPABLE_WATCHPOINT
497 If defined to a non-zero value, @value{GDBN} should disable a
498 watchpoint to step the inferior over it.
500 @findex HAVE_CONTINUABLE_WATCHPOINT
501 @item HAVE_CONTINUABLE_WATCHPOINT
502 If defined to a non-zero value, it is possible to continue the
503 inferior after a watchpoint has been hit.
505 @findex CANNOT_STEP_HW_WATCHPOINTS
506 @item CANNOT_STEP_HW_WATCHPOINTS
507 If this is defined to a non-zero value, @value{GDBN} will remove all
508 watchpoints before stepping the inferior.
510 @findex STOPPED_BY_WATCHPOINT
511 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
512 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
513 the type @code{struct target_waitstatus}, defined by @file{target.h}.
514 Normally, this macro is defined to invoke the function pointed to by
515 the @code{to_stopped_by_watchpoint} member of the structure (of the
516 type @code{target_ops}, defined on @file{target.h}) that describes the
517 target-specific operations; @code{to_stopped_by_watchpoint} ignores
518 the @var{wait_status} argument.
520 @value{GDBN} does not require the non-zero value returned by
521 @code{STOPPED_BY_WATCHPOINT} to be 100% correct, so if a target cannot
522 determine for sure whether the inferior stopped due to a watchpoint,
523 it could return non-zero ``just in case''.
526 @subsection x86 Watchpoints
527 @cindex x86 debug registers
528 @cindex watchpoints, on x86
530 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
531 registers designed to facilitate debugging. @value{GDBN} provides a
532 generic library of functions that x86-based ports can use to implement
533 support for watchpoints and hardware-assisted breakpoints. This
534 subsection documents the x86 watchpoint facilities in @value{GDBN}.
536 To use the generic x86 watchpoint support, a port should do the
540 @findex I386_USE_GENERIC_WATCHPOINTS
542 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
543 target-dependent headers.
546 Include the @file{config/i386/nm-i386.h} header file @emph{after}
547 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
550 Add @file{i386-nat.o} to the value of the Make variable
551 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}) or
552 @code{TDEPFILES} (@pxref{Target Architecture Definition, TDEPFILES}).
555 Provide implementations for the @code{I386_DR_LOW_*} macros described
556 below. Typically, each macro should call a target-specific function
557 which does the real work.
560 The x86 watchpoint support works by maintaining mirror images of the
561 debug registers. Values are copied between the mirror images and the
562 real debug registers via a set of macros which each target needs to
566 @findex I386_DR_LOW_SET_CONTROL
567 @item I386_DR_LOW_SET_CONTROL (@var{val})
568 Set the Debug Control (DR7) register to the value @var{val}.
570 @findex I386_DR_LOW_SET_ADDR
571 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
572 Put the address @var{addr} into the debug register number @var{idx}.
574 @findex I386_DR_LOW_RESET_ADDR
575 @item I386_DR_LOW_RESET_ADDR (@var{idx})
576 Reset (i.e.@: zero out) the address stored in the debug register
579 @findex I386_DR_LOW_GET_STATUS
580 @item I386_DR_LOW_GET_STATUS
581 Return the value of the Debug Status (DR6) register. This value is
582 used immediately after it is returned by
583 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
587 For each one of the 4 debug registers (whose indices are from 0 to 3)
588 that store addresses, a reference count is maintained by @value{GDBN},
589 to allow sharing of debug registers by several watchpoints. This
590 allows users to define several watchpoints that watch the same
591 expression, but with different conditions and/or commands, without
592 wasting debug registers which are in short supply. @value{GDBN}
593 maintains the reference counts internally, targets don't have to do
594 anything to use this feature.
596 The x86 debug registers can each watch a region that is 1, 2, or 4
597 bytes long. The ia32 architecture requires that each watched region
598 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
599 region on 4-byte boundary. However, the x86 watchpoint support in
600 @value{GDBN} can watch unaligned regions and regions larger than 4
601 bytes (up to 16 bytes) by allocating several debug registers to watch
602 a single region. This allocation of several registers per a watched
603 region is also done automatically without target code intervention.
605 The generic x86 watchpoint support provides the following API for the
606 @value{GDBN}'s application code:
609 @findex i386_region_ok_for_watchpoint
610 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
611 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
612 this function. It counts the number of debug registers required to
613 watch a given region, and returns a non-zero value if that number is
614 less than 4, the number of debug registers available to x86
617 @findex i386_stopped_data_address
618 @item i386_stopped_data_address (@var{addr_p})
620 @code{target_stopped_data_address} is set to call this function.
622 function examines the breakpoint condition bits in the DR6 Debug
623 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
624 macro, and returns the address associated with the first bit that is
627 @findex i386_stopped_by_watchpoint
628 @item i386_stopped_by_watchpoint (void)
629 The macro @code{STOPPED_BY_WATCHPOINT}
630 is set to call this function. The
631 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
632 function examines the breakpoint condition bits in the DR6 Debug
633 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
634 macro, and returns true if any bit is set. Otherwise, false is
637 @findex i386_insert_watchpoint
638 @findex i386_remove_watchpoint
639 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
640 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
641 Insert or remove a watchpoint. The macros
642 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
643 are set to call these functions. @code{i386_insert_watchpoint} first
644 looks for a debug register which is already set to watch the same
645 region for the same access types; if found, it just increments the
646 reference count of that debug register, thus implementing debug
647 register sharing between watchpoints. If no such register is found,
648 the function looks for a vacant debug register, sets its mirrored
649 value to @var{addr}, sets the mirrored value of DR7 Debug Control
650 register as appropriate for the @var{len} and @var{type} parameters,
651 and then passes the new values of the debug register and DR7 to the
652 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
653 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
654 required to cover the given region, the above process is repeated for
657 @code{i386_remove_watchpoint} does the opposite: it resets the address
658 in the mirrored value of the debug register and its read/write and
659 length bits in the mirrored value of DR7, then passes these new
660 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
661 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
662 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
663 decrements the reference count, and only calls
664 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
665 the count goes to zero.
667 @findex i386_insert_hw_breakpoint
668 @findex i386_remove_hw_breakpoint
669 @item i386_insert_hw_breakpoint (@var{addr}, @var{shadow}
670 @itemx i386_remove_hw_breakpoint (@var{addr}, @var{shadow})
671 These functions insert and remove hardware-assisted breakpoints. The
672 macros @code{target_insert_hw_breakpoint} and
673 @code{target_remove_hw_breakpoint} are set to call these functions.
674 These functions work like @code{i386_insert_watchpoint} and
675 @code{i386_remove_watchpoint}, respectively, except that they set up
676 the debug registers to watch instruction execution, and each
677 hardware-assisted breakpoint always requires exactly one debug
680 @findex i386_stopped_by_hwbp
681 @item i386_stopped_by_hwbp (void)
682 This function returns non-zero if the inferior has some watchpoint or
683 hardware breakpoint that triggered. It works like
684 @code{i386_stopped_data_address}, except that it doesn't record the
685 address whose watchpoint triggered.
687 @findex i386_cleanup_dregs
688 @item i386_cleanup_dregs (void)
689 This function clears all the reference counts, addresses, and control
690 bits in the mirror images of the debug registers. It doesn't affect
691 the actual debug registers in the inferior process.
698 x86 processors support setting watchpoints on I/O reads or writes.
699 However, since no target supports this (as of March 2001), and since
700 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
701 watchpoints, this feature is not yet available to @value{GDBN} running
705 x86 processors can enable watchpoints locally, for the current task
706 only, or globally, for all the tasks. For each debug register,
707 there's a bit in the DR7 Debug Control register that determines
708 whether the associated address is watched locally or globally. The
709 current implementation of x86 watchpoint support in @value{GDBN}
710 always sets watchpoints to be locally enabled, since global
711 watchpoints might interfere with the underlying OS and are probably
712 unavailable in many platforms.
715 @section Observing changes in @value{GDBN} internals
716 @cindex observer pattern interface
717 @cindex notifications about changes in internals
719 In order to function properly, several modules need to be notified when
720 some changes occur in the @value{GDBN} internals. Traditionally, these
721 modules have relied on several paradigms, the most common ones being
722 hooks and gdb-events. Unfortunately, none of these paradigms was
723 versatile enough to become the standard notification mechanism in
724 @value{GDBN}. The fact that they only supported one ``client'' was also
727 A new paradigm, based on the Observer pattern of the @cite{Design
728 Patterns} book, has therefore been implemented. The goal was to provide
729 a new interface overcoming the issues with the notification mechanisms
730 previously available. This new interface needed to be strongly typed,
731 easy to extend, and versatile enough to be used as the standard
732 interface when adding new notifications.
734 See @ref{GDB Observers} for a brief description of the observers
735 currently implemented in GDB. The rationale for the current
736 implementation is also briefly discussed.
740 @chapter User Interface
742 @value{GDBN} has several user interfaces. Although the command-line interface
743 is the most common and most familiar, there are others.
745 @section Command Interpreter
747 @cindex command interpreter
749 The command interpreter in @value{GDBN} is fairly simple. It is designed to
750 allow for the set of commands to be augmented dynamically, and also
751 has a recursive subcommand capability, where the first argument to
752 a command may itself direct a lookup on a different command list.
754 For instance, the @samp{set} command just starts a lookup on the
755 @code{setlist} command list, while @samp{set thread} recurses
756 to the @code{set_thread_cmd_list}.
760 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
761 the main command list, and should be used for those commands. The usual
762 place to add commands is in the @code{_initialize_@var{xyz}} routines at
763 the ends of most source files.
765 @findex add_setshow_cmd
766 @findex add_setshow_cmd_full
767 To add paired @samp{set} and @samp{show} commands, use
768 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
769 a slightly simpler interface which is useful when you don't need to
770 further modify the new command structures, while the latter returns
771 the new command structures for manipulation.
773 @cindex deprecating commands
774 @findex deprecate_cmd
775 Before removing commands from the command set it is a good idea to
776 deprecate them for some time. Use @code{deprecate_cmd} on commands or
777 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
778 @code{struct cmd_list_element} as it's first argument. You can use the
779 return value from @code{add_com} or @code{add_cmd} to deprecate the
780 command immediately after it is created.
782 The first time a command is used the user will be warned and offered a
783 replacement (if one exists). Note that the replacement string passed to
784 @code{deprecate_cmd} should be the full name of the command, i.e. the
785 entire string the user should type at the command line.
787 @section UI-Independent Output---the @code{ui_out} Functions
788 @c This section is based on the documentation written by Fernando
789 @c Nasser <fnasser@redhat.com>.
791 @cindex @code{ui_out} functions
792 The @code{ui_out} functions present an abstraction level for the
793 @value{GDBN} output code. They hide the specifics of different user
794 interfaces supported by @value{GDBN}, and thus free the programmer
795 from the need to write several versions of the same code, one each for
796 every UI, to produce output.
798 @subsection Overview and Terminology
800 In general, execution of each @value{GDBN} command produces some sort
801 of output, and can even generate an input request.
803 Output can be generated for the following purposes:
807 to display a @emph{result} of an operation;
810 to convey @emph{info} or produce side-effects of a requested
814 to provide a @emph{notification} of an asynchronous event (including
815 progress indication of a prolonged asynchronous operation);
818 to display @emph{error messages} (including warnings);
821 to show @emph{debug data};
824 to @emph{query} or prompt a user for input (a special case).
828 This section mainly concentrates on how to build result output,
829 although some of it also applies to other kinds of output.
831 Generation of output that displays the results of an operation
832 involves one or more of the following:
836 output of the actual data
839 formatting the output as appropriate for console output, to make it
840 easily readable by humans
843 machine oriented formatting--a more terse formatting to allow for easy
844 parsing by programs which read @value{GDBN}'s output
847 annotation, whose purpose is to help legacy GUIs to identify interesting
851 The @code{ui_out} routines take care of the first three aspects.
852 Annotations are provided by separate annotation routines. Note that use
853 of annotations for an interface between a GUI and @value{GDBN} is
856 Output can be in the form of a single item, which we call a @dfn{field};
857 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
858 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
859 header and a body. In a BNF-like form:
862 @item <table> @expansion{}
863 @code{<header> <body>}
864 @item <header> @expansion{}
865 @code{@{ <column> @}}
866 @item <column> @expansion{}
867 @code{<width> <alignment> <title>}
868 @item <body> @expansion{}
873 @subsection General Conventions
875 Most @code{ui_out} routines are of type @code{void}, the exceptions are
876 @code{ui_out_stream_new} (which returns a pointer to the newly created
877 object) and the @code{make_cleanup} routines.
879 The first parameter is always the @code{ui_out} vector object, a pointer
880 to a @code{struct ui_out}.
882 The @var{format} parameter is like in @code{printf} family of functions.
883 When it is present, there must also be a variable list of arguments
884 sufficient used to satisfy the @code{%} specifiers in the supplied
887 When a character string argument is not used in a @code{ui_out} function
888 call, a @code{NULL} pointer has to be supplied instead.
891 @subsection Table, Tuple and List Functions
893 @cindex list output functions
894 @cindex table output functions
895 @cindex tuple output functions
896 This section introduces @code{ui_out} routines for building lists,
897 tuples and tables. The routines to output the actual data items
898 (fields) are presented in the next section.
900 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
901 containing information about an object; a @dfn{list} is a sequence of
902 fields where each field describes an identical object.
904 Use the @dfn{table} functions when your output consists of a list of
905 rows (tuples) and the console output should include a heading. Use this
906 even when you are listing just one object but you still want the header.
908 @cindex nesting level in @code{ui_out} functions
909 Tables can not be nested. Tuples and lists can be nested up to a
910 maximum of five levels.
912 The overall structure of the table output code is something like this:
927 Here is the description of table-, tuple- and list-related @code{ui_out}
930 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
931 The function @code{ui_out_table_begin} marks the beginning of the output
932 of a table. It should always be called before any other @code{ui_out}
933 function for a given table. @var{nbrofcols} is the number of columns in
934 the table. @var{nr_rows} is the number of rows in the table.
935 @var{tblid} is an optional string identifying the table. The string
936 pointed to by @var{tblid} is copied by the implementation of
937 @code{ui_out_table_begin}, so the application can free the string if it
940 The companion function @code{ui_out_table_end}, described below, marks
941 the end of the table's output.
944 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
945 @code{ui_out_table_header} provides the header information for a single
946 table column. You call this function several times, one each for every
947 column of the table, after @code{ui_out_table_begin}, but before
948 @code{ui_out_table_body}.
950 The value of @var{width} gives the column width in characters. The
951 value of @var{alignment} is one of @code{left}, @code{center}, and
952 @code{right}, and it specifies how to align the header: left-justify,
953 center, or right-justify it. @var{colhdr} points to a string that
954 specifies the column header; the implementation copies that string, so
955 column header strings in @code{malloc}ed storage can be freed after the
959 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
960 This function delimits the table header from the table body.
963 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
964 This function signals the end of a table's output. It should be called
965 after the table body has been produced by the list and field output
968 There should be exactly one call to @code{ui_out_table_end} for each
969 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
970 will signal an internal error.
973 The output of the tuples that represent the table rows must follow the
974 call to @code{ui_out_table_body} and precede the call to
975 @code{ui_out_table_end}. You build a tuple by calling
976 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
977 calls to functions which actually output fields between them.
979 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
980 This function marks the beginning of a tuple output. @var{id} points
981 to an optional string that identifies the tuple; it is copied by the
982 implementation, and so strings in @code{malloc}ed storage can be freed
986 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
987 This function signals an end of a tuple output. There should be exactly
988 one call to @code{ui_out_tuple_end} for each call to
989 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
993 @deftypefun struct cleanup *make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
994 This function first opens the tuple and then establishes a cleanup
995 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
996 and correct implementation of the non-portable@footnote{The function
997 cast is not portable ISO C.} code sequence:
999 struct cleanup *old_cleanup;
1000 ui_out_tuple_begin (uiout, "...");
1001 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
1006 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
1007 This function marks the beginning of a list output. @var{id} points to
1008 an optional string that identifies the list; it is copied by the
1009 implementation, and so strings in @code{malloc}ed storage can be freed
1013 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
1014 This function signals an end of a list output. There should be exactly
1015 one call to @code{ui_out_list_end} for each call to
1016 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1020 @deftypefun struct cleanup *make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1021 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1022 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
1023 that will close the list.list.
1026 @subsection Item Output Functions
1028 @cindex item output functions
1029 @cindex field output functions
1031 The functions described below produce output for the actual data
1032 items, or fields, which contain information about the object.
1034 Choose the appropriate function accordingly to your particular needs.
1036 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1037 This is the most general output function. It produces the
1038 representation of the data in the variable-length argument list
1039 according to formatting specifications in @var{format}, a
1040 @code{printf}-like format string. The optional argument @var{fldname}
1041 supplies the name of the field. The data items themselves are
1042 supplied as additional arguments after @var{format}.
1044 This generic function should be used only when it is not possible to
1045 use one of the specialized versions (see below).
1048 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1049 This function outputs a value of an @code{int} variable. It uses the
1050 @code{"%d"} output conversion specification. @var{fldname} specifies
1051 the name of the field.
1054 @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})
1055 This function outputs a value of an @code{int} variable. It differs from
1056 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1057 @var{fldname} specifies
1058 the name of the field.
1061 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1062 This function outputs an address.
1065 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1066 This function outputs a string using the @code{"%s"} conversion
1070 Sometimes, there's a need to compose your output piece by piece using
1071 functions that operate on a stream, such as @code{value_print} or
1072 @code{fprintf_symbol_filtered}. These functions accept an argument of
1073 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1074 used to store the data stream used for the output. When you use one
1075 of these functions, you need a way to pass their results stored in a
1076 @code{ui_file} object to the @code{ui_out} functions. To this end,
1077 you first create a @code{ui_stream} object by calling
1078 @code{ui_out_stream_new}, pass the @code{stream} member of that
1079 @code{ui_stream} object to @code{value_print} and similar functions,
1080 and finally call @code{ui_out_field_stream} to output the field you
1081 constructed. When the @code{ui_stream} object is no longer needed,
1082 you should destroy it and free its memory by calling
1083 @code{ui_out_stream_delete}.
1085 @deftypefun struct ui_stream *ui_out_stream_new (struct ui_out *@var{uiout})
1086 This function creates a new @code{ui_stream} object which uses the
1087 same output methods as the @code{ui_out} object whose pointer is
1088 passed in @var{uiout}. It returns a pointer to the newly created
1089 @code{ui_stream} object.
1092 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1093 This functions destroys a @code{ui_stream} object specified by
1097 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1098 This function consumes all the data accumulated in
1099 @code{streambuf->stream} and outputs it like
1100 @code{ui_out_field_string} does. After a call to
1101 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1102 the stream is still valid and may be used for producing more fields.
1105 @strong{Important:} If there is any chance that your code could bail
1106 out before completing output generation and reaching the point where
1107 @code{ui_out_stream_delete} is called, it is necessary to set up a
1108 cleanup, to avoid leaking memory and other resources. Here's a
1109 skeleton code to do that:
1112 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1113 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1118 If the function already has the old cleanup chain set (for other kinds
1119 of cleanups), you just have to add your cleanup to it:
1122 mybuf = ui_out_stream_new (uiout);
1123 make_cleanup (ui_out_stream_delete, mybuf);
1126 Note that with cleanups in place, you should not call
1127 @code{ui_out_stream_delete} directly, or you would attempt to free the
1130 @subsection Utility Output Functions
1132 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1133 This function skips a field in a table. Use it if you have to leave
1134 an empty field without disrupting the table alignment. The argument
1135 @var{fldname} specifies a name for the (missing) filed.
1138 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1139 This function outputs the text in @var{string} in a way that makes it
1140 easy to be read by humans. For example, the console implementation of
1141 this method filters the text through a built-in pager, to prevent it
1142 from scrolling off the visible portion of the screen.
1144 Use this function for printing relatively long chunks of text around
1145 the actual field data: the text it produces is not aligned according
1146 to the table's format. Use @code{ui_out_field_string} to output a
1147 string field, and use @code{ui_out_message}, described below, to
1148 output short messages.
1151 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1152 This function outputs @var{nspaces} spaces. It is handy to align the
1153 text produced by @code{ui_out_text} with the rest of the table or
1157 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1158 This function produces a formatted message, provided that the current
1159 verbosity level is at least as large as given by @var{verbosity}. The
1160 current verbosity level is specified by the user with the @samp{set
1161 verbositylevel} command.@footnote{As of this writing (April 2001),
1162 setting verbosity level is not yet implemented, and is always returned
1163 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1164 argument more than zero will cause the message to never be printed.}
1167 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1168 This function gives the console output filter (a paging filter) a hint
1169 of where to break lines which are too long. Ignored for all other
1170 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1171 be printed to indent the wrapped text on the next line; it must remain
1172 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1173 explicit newline is produced by one of the other functions. If
1174 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1177 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1178 This function flushes whatever output has been accumulated so far, if
1179 the UI buffers output.
1183 @subsection Examples of Use of @code{ui_out} functions
1185 @cindex using @code{ui_out} functions
1186 @cindex @code{ui_out} functions, usage examples
1187 This section gives some practical examples of using the @code{ui_out}
1188 functions to generalize the old console-oriented code in
1189 @value{GDBN}. The examples all come from functions defined on the
1190 @file{breakpoints.c} file.
1192 This example, from the @code{breakpoint_1} function, shows how to
1195 The original code was:
1198 if (!found_a_breakpoint++)
1200 annotate_breakpoints_headers ();
1203 printf_filtered ("Num ");
1205 printf_filtered ("Type ");
1207 printf_filtered ("Disp ");
1209 printf_filtered ("Enb ");
1213 printf_filtered ("Address ");
1216 printf_filtered ("What\n");
1218 annotate_breakpoints_table ();
1222 Here's the new version:
1225 nr_printable_breakpoints = @dots{};
1228 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1230 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1232 if (nr_printable_breakpoints > 0)
1233 annotate_breakpoints_headers ();
1234 if (nr_printable_breakpoints > 0)
1236 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1237 if (nr_printable_breakpoints > 0)
1239 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1240 if (nr_printable_breakpoints > 0)
1242 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1243 if (nr_printable_breakpoints > 0)
1245 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1248 if (nr_printable_breakpoints > 0)
1250 if (TARGET_ADDR_BIT <= 32)
1251 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1253 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1255 if (nr_printable_breakpoints > 0)
1257 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1258 ui_out_table_body (uiout);
1259 if (nr_printable_breakpoints > 0)
1260 annotate_breakpoints_table ();
1263 This example, from the @code{print_one_breakpoint} function, shows how
1264 to produce the actual data for the table whose structure was defined
1265 in the above example. The original code was:
1270 printf_filtered ("%-3d ", b->number);
1272 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1273 || ((int) b->type != bptypes[(int) b->type].type))
1274 internal_error ("bptypes table does not describe type #%d.",
1276 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1278 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1280 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1284 This is the new version:
1288 ui_out_tuple_begin (uiout, "bkpt");
1290 ui_out_field_int (uiout, "number", b->number);
1292 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1293 || ((int) b->type != bptypes[(int) b->type].type))
1294 internal_error ("bptypes table does not describe type #%d.",
1296 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1298 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1300 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1304 This example, also from @code{print_one_breakpoint}, shows how to
1305 produce a complicated output field using the @code{print_expression}
1306 functions which requires a stream to be passed. It also shows how to
1307 automate stream destruction with cleanups. The original code was:
1311 print_expression (b->exp, gdb_stdout);
1317 struct ui_stream *stb = ui_out_stream_new (uiout);
1318 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1321 print_expression (b->exp, stb->stream);
1322 ui_out_field_stream (uiout, "what", local_stream);
1325 This example, also from @code{print_one_breakpoint}, shows how to use
1326 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1331 if (b->dll_pathname == NULL)
1332 printf_filtered ("<any library> ");
1334 printf_filtered ("library \"%s\" ", b->dll_pathname);
1341 if (b->dll_pathname == NULL)
1343 ui_out_field_string (uiout, "what", "<any library>");
1344 ui_out_spaces (uiout, 1);
1348 ui_out_text (uiout, "library \"");
1349 ui_out_field_string (uiout, "what", b->dll_pathname);
1350 ui_out_text (uiout, "\" ");
1354 The following example from @code{print_one_breakpoint} shows how to
1355 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1360 if (b->forked_inferior_pid != 0)
1361 printf_filtered ("process %d ", b->forked_inferior_pid);
1368 if (b->forked_inferior_pid != 0)
1370 ui_out_text (uiout, "process ");
1371 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1372 ui_out_spaces (uiout, 1);
1376 Here's an example of using @code{ui_out_field_string}. The original
1381 if (b->exec_pathname != NULL)
1382 printf_filtered ("program \"%s\" ", b->exec_pathname);
1389 if (b->exec_pathname != NULL)
1391 ui_out_text (uiout, "program \"");
1392 ui_out_field_string (uiout, "what", b->exec_pathname);
1393 ui_out_text (uiout, "\" ");
1397 Finally, here's an example of printing an address. The original code:
1401 printf_filtered ("%s ",
1402 hex_string_custom ((unsigned long) b->address, 8));
1409 ui_out_field_core_addr (uiout, "Address", b->address);
1413 @section Console Printing
1422 @cindex @code{libgdb}
1423 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1424 to provide an API to @value{GDBN}'s functionality.
1427 @cindex @code{libgdb}
1428 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1429 better able to support graphical and other environments.
1431 Since @code{libgdb} development is on-going, its architecture is still
1432 evolving. The following components have so far been identified:
1436 Observer - @file{gdb-events.h}.
1438 Builder - @file{ui-out.h}
1440 Event Loop - @file{event-loop.h}
1442 Library - @file{gdb.h}
1445 The model that ties these components together is described below.
1447 @section The @code{libgdb} Model
1449 A client of @code{libgdb} interacts with the library in two ways.
1453 As an observer (using @file{gdb-events}) receiving notifications from
1454 @code{libgdb} of any internal state changes (break point changes, run
1457 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1458 obtain various status values from @value{GDBN}.
1461 Since @code{libgdb} could have multiple clients (e.g. a GUI supporting
1462 the existing @value{GDBN} CLI), those clients must co-operate when
1463 controlling @code{libgdb}. In particular, a client must ensure that
1464 @code{libgdb} is idle (i.e. no other client is using @code{libgdb})
1465 before responding to a @file{gdb-event} by making a query.
1467 @section CLI support
1469 At present @value{GDBN}'s CLI is very much entangled in with the core of
1470 @code{libgdb}. Consequently, a client wishing to include the CLI in
1471 their interface needs to carefully co-ordinate its own and the CLI's
1474 It is suggested that the client set @code{libgdb} up to be bi-modal
1475 (alternate between CLI and client query modes). The notes below sketch
1480 The client registers itself as an observer of @code{libgdb}.
1482 The client create and install @code{cli-out} builder using its own
1483 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1484 @code{gdb_stdout} streams.
1486 The client creates a separate custom @code{ui-out} builder that is only
1487 used while making direct queries to @code{libgdb}.
1490 When the client receives input intended for the CLI, it simply passes it
1491 along. Since the @code{cli-out} builder is installed by default, all
1492 the CLI output in response to that command is routed (pronounced rooted)
1493 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1494 At the same time, the client is kept abreast of internal changes by
1495 virtue of being a @code{libgdb} observer.
1497 The only restriction on the client is that it must wait until
1498 @code{libgdb} becomes idle before initiating any queries (using the
1499 client's custom builder).
1501 @section @code{libgdb} components
1503 @subheading Observer - @file{gdb-events.h}
1504 @file{gdb-events} provides the client with a very raw mechanism that can
1505 be used to implement an observer. At present it only allows for one
1506 observer and that observer must, internally, handle the need to delay
1507 the processing of any event notifications until after @code{libgdb} has
1508 finished the current command.
1510 @subheading Builder - @file{ui-out.h}
1511 @file{ui-out} provides the infrastructure necessary for a client to
1512 create a builder. That builder is then passed down to @code{libgdb}
1513 when doing any queries.
1515 @subheading Event Loop - @file{event-loop.h}
1516 @c There could be an entire section on the event-loop
1517 @file{event-loop}, currently non-re-entrant, provides a simple event
1518 loop. A client would need to either plug its self into this loop or,
1519 implement a new event-loop that GDB would use.
1521 The event-loop will eventually be made re-entrant. This is so that
1522 @value{GDBN} can better handle the problem of some commands blocking
1523 instead of returning.
1525 @subheading Library - @file{gdb.h}
1526 @file{libgdb} is the most obvious component of this system. It provides
1527 the query interface. Each function is parameterized by a @code{ui-out}
1528 builder. The result of the query is constructed using that builder
1529 before the query function returns.
1531 @node Symbol Handling
1533 @chapter Symbol Handling
1535 Symbols are a key part of @value{GDBN}'s operation. Symbols include variables,
1536 functions, and types.
1538 @section Symbol Reading
1540 @cindex symbol reading
1541 @cindex reading of symbols
1542 @cindex symbol files
1543 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
1544 file is the file containing the program which @value{GDBN} is
1545 debugging. @value{GDBN} can be directed to use a different file for
1546 symbols (with the @samp{symbol-file} command), and it can also read
1547 more symbols via the @samp{add-file} and @samp{load} commands, or while
1548 reading symbols from shared libraries.
1550 @findex find_sym_fns
1551 Symbol files are initially opened by code in @file{symfile.c} using
1552 the BFD library (@pxref{Support Libraries}). BFD identifies the type
1553 of the file by examining its header. @code{find_sym_fns} then uses
1554 this identification to locate a set of symbol-reading functions.
1556 @findex add_symtab_fns
1557 @cindex @code{sym_fns} structure
1558 @cindex adding a symbol-reading module
1559 Symbol-reading modules identify themselves to @value{GDBN} by calling
1560 @code{add_symtab_fns} during their module initialization. The argument
1561 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
1562 name (or name prefix) of the symbol format, the length of the prefix,
1563 and pointers to four functions. These functions are called at various
1564 times to process symbol files whose identification matches the specified
1567 The functions supplied by each module are:
1570 @item @var{xyz}_symfile_init(struct sym_fns *sf)
1572 @cindex secondary symbol file
1573 Called from @code{symbol_file_add} when we are about to read a new
1574 symbol file. This function should clean up any internal state (possibly
1575 resulting from half-read previous files, for example) and prepare to
1576 read a new symbol file. Note that the symbol file which we are reading
1577 might be a new ``main'' symbol file, or might be a secondary symbol file
1578 whose symbols are being added to the existing symbol table.
1580 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
1581 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
1582 new symbol file being read. Its @code{private} field has been zeroed,
1583 and can be modified as desired. Typically, a struct of private
1584 information will be @code{malloc}'d, and a pointer to it will be placed
1585 in the @code{private} field.
1587 There is no result from @code{@var{xyz}_symfile_init}, but it can call
1588 @code{error} if it detects an unavoidable problem.
1590 @item @var{xyz}_new_init()
1592 Called from @code{symbol_file_add} when discarding existing symbols.
1593 This function needs only handle the symbol-reading module's internal
1594 state; the symbol table data structures visible to the rest of
1595 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
1596 arguments and no result. It may be called after
1597 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
1598 may be called alone if all symbols are simply being discarded.
1600 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
1602 Called from @code{symbol_file_add} to actually read the symbols from a
1603 symbol-file into a set of psymtabs or symtabs.
1605 @code{sf} points to the @code{struct sym_fns} originally passed to
1606 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
1607 the offset between the file's specified start address and its true
1608 address in memory. @code{mainline} is 1 if this is the main symbol
1609 table being read, and 0 if a secondary symbol file (e.g. shared library
1610 or dynamically loaded file) is being read.@refill
1613 In addition, if a symbol-reading module creates psymtabs when
1614 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
1615 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
1616 from any point in the @value{GDBN} symbol-handling code.
1619 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
1621 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
1622 the psymtab has not already been read in and had its @code{pst->symtab}
1623 pointer set. The argument is the psymtab to be fleshed-out into a
1624 symtab. Upon return, @code{pst->readin} should have been set to 1, and
1625 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
1626 zero if there were no symbols in that part of the symbol file.
1629 @section Partial Symbol Tables
1631 @value{GDBN} has three types of symbol tables:
1634 @cindex full symbol table
1637 Full symbol tables (@dfn{symtabs}). These contain the main
1638 information about symbols and addresses.
1642 Partial symbol tables (@dfn{psymtabs}). These contain enough
1643 information to know when to read the corresponding part of the full
1646 @cindex minimal symbol table
1649 Minimal symbol tables (@dfn{msymtabs}). These contain information
1650 gleaned from non-debugging symbols.
1653 @cindex partial symbol table
1654 This section describes partial symbol tables.
1656 A psymtab is constructed by doing a very quick pass over an executable
1657 file's debugging information. Small amounts of information are
1658 extracted---enough to identify which parts of the symbol table will
1659 need to be re-read and fully digested later, when the user needs the
1660 information. The speed of this pass causes @value{GDBN} to start up very
1661 quickly. Later, as the detailed rereading occurs, it occurs in small
1662 pieces, at various times, and the delay therefrom is mostly invisible to
1664 @c (@xref{Symbol Reading}.)
1666 The symbols that show up in a file's psymtab should be, roughly, those
1667 visible to the debugger's user when the program is not running code from
1668 that file. These include external symbols and types, static symbols and
1669 types, and @code{enum} values declared at file scope.
1671 The psymtab also contains the range of instruction addresses that the
1672 full symbol table would represent.
1674 @cindex finding a symbol
1675 @cindex symbol lookup
1676 The idea is that there are only two ways for the user (or much of the
1677 code in the debugger) to reference a symbol:
1680 @findex find_pc_function
1681 @findex find_pc_line
1683 By its address (e.g. execution stops at some address which is inside a
1684 function in this file). The address will be noticed to be in the
1685 range of this psymtab, and the full symtab will be read in.
1686 @code{find_pc_function}, @code{find_pc_line}, and other
1687 @code{find_pc_@dots{}} functions handle this.
1689 @cindex lookup_symbol
1692 (e.g. the user asks to print a variable, or set a breakpoint on a
1693 function). Global names and file-scope names will be found in the
1694 psymtab, which will cause the symtab to be pulled in. Local names will
1695 have to be qualified by a global name, or a file-scope name, in which
1696 case we will have already read in the symtab as we evaluated the
1697 qualifier. Or, a local symbol can be referenced when we are ``in'' a
1698 local scope, in which case the first case applies. @code{lookup_symbol}
1699 does most of the work here.
1702 The only reason that psymtabs exist is to cause a symtab to be read in
1703 at the right moment. Any symbol that can be elided from a psymtab,
1704 while still causing that to happen, should not appear in it. Since
1705 psymtabs don't have the idea of scope, you can't put local symbols in
1706 them anyway. Psymtabs don't have the idea of the type of a symbol,
1707 either, so types need not appear, unless they will be referenced by
1710 It is a bug for @value{GDBN} to behave one way when only a psymtab has
1711 been read, and another way if the corresponding symtab has been read
1712 in. Such bugs are typically caused by a psymtab that does not contain
1713 all the visible symbols, or which has the wrong instruction address
1716 The psymtab for a particular section of a symbol file (objfile) could be
1717 thrown away after the symtab has been read in. The symtab should always
1718 be searched before the psymtab, so the psymtab will never be used (in a
1719 bug-free environment). Currently, psymtabs are allocated on an obstack,
1720 and all the psymbols themselves are allocated in a pair of large arrays
1721 on an obstack, so there is little to be gained by trying to free them
1722 unless you want to do a lot more work.
1726 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
1728 @cindex fundamental types
1729 These are the fundamental types that @value{GDBN} uses internally. Fundamental
1730 types from the various debugging formats (stabs, ELF, etc) are mapped
1731 into one of these. They are basically a union of all fundamental types
1732 that @value{GDBN} knows about for all the languages that @value{GDBN}
1735 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
1738 Each time @value{GDBN} builds an internal type, it marks it with one
1739 of these types. The type may be a fundamental type, such as
1740 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
1741 which is a pointer to another type. Typically, several @code{FT_*}
1742 types map to one @code{TYPE_CODE_*} type, and are distinguished by
1743 other members of the type struct, such as whether the type is signed
1744 or unsigned, and how many bits it uses.
1746 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
1748 These are instances of type structs that roughly correspond to
1749 fundamental types and are created as global types for @value{GDBN} to
1750 use for various ugly historical reasons. We eventually want to
1751 eliminate these. Note for example that @code{builtin_type_int}
1752 initialized in @file{gdbtypes.c} is basically the same as a
1753 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
1754 an @code{FT_INTEGER} fundamental type. The difference is that the
1755 @code{builtin_type} is not associated with any particular objfile, and
1756 only one instance exists, while @file{c-lang.c} builds as many
1757 @code{TYPE_CODE_INT} types as needed, with each one associated with
1758 some particular objfile.
1760 @section Object File Formats
1761 @cindex object file formats
1765 @cindex @code{a.out} format
1766 The @code{a.out} format is the original file format for Unix. It
1767 consists of three sections: @code{text}, @code{data}, and @code{bss},
1768 which are for program code, initialized data, and uninitialized data,
1771 The @code{a.out} format is so simple that it doesn't have any reserved
1772 place for debugging information. (Hey, the original Unix hackers used
1773 @samp{adb}, which is a machine-language debugger!) The only debugging
1774 format for @code{a.out} is stabs, which is encoded as a set of normal
1775 symbols with distinctive attributes.
1777 The basic @code{a.out} reader is in @file{dbxread.c}.
1782 The COFF format was introduced with System V Release 3 (SVR3) Unix.
1783 COFF files may have multiple sections, each prefixed by a header. The
1784 number of sections is limited.
1786 The COFF specification includes support for debugging. Although this
1787 was a step forward, the debugging information was woefully limited. For
1788 instance, it was not possible to represent code that came from an
1791 The COFF reader is in @file{coffread.c}.
1795 @cindex ECOFF format
1796 ECOFF is an extended COFF originally introduced for Mips and Alpha
1799 The basic ECOFF reader is in @file{mipsread.c}.
1803 @cindex XCOFF format
1804 The IBM RS/6000 running AIX uses an object file format called XCOFF.
1805 The COFF sections, symbols, and line numbers are used, but debugging
1806 symbols are @code{dbx}-style stabs whose strings are located in the
1807 @code{.debug} section (rather than the string table). For more
1808 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
1810 The shared library scheme has a clean interface for figuring out what
1811 shared libraries are in use, but the catch is that everything which
1812 refers to addresses (symbol tables and breakpoints at least) needs to be
1813 relocated for both shared libraries and the main executable. At least
1814 using the standard mechanism this can only be done once the program has
1815 been run (or the core file has been read).
1819 @cindex PE-COFF format
1820 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
1821 executables. PE is basically COFF with additional headers.
1823 While BFD includes special PE support, @value{GDBN} needs only the basic
1829 The ELF format came with System V Release 4 (SVR4) Unix. ELF is similar
1830 to COFF in being organized into a number of sections, but it removes
1831 many of COFF's limitations.
1833 The basic ELF reader is in @file{elfread.c}.
1838 SOM is HP's object file and debug format (not to be confused with IBM's
1839 SOM, which is a cross-language ABI).
1841 The SOM reader is in @file{hpread.c}.
1843 @subsection Other File Formats
1845 @cindex Netware Loadable Module format
1846 Other file formats that have been supported by @value{GDBN} include Netware
1847 Loadable Modules (@file{nlmread.c}).
1849 @section Debugging File Formats
1851 This section describes characteristics of debugging information that
1852 are independent of the object file format.
1856 @cindex stabs debugging info
1857 @code{stabs} started out as special symbols within the @code{a.out}
1858 format. Since then, it has been encapsulated into other file
1859 formats, such as COFF and ELF.
1861 While @file{dbxread.c} does some of the basic stab processing,
1862 including for encapsulated versions, @file{stabsread.c} does
1867 @cindex COFF debugging info
1868 The basic COFF definition includes debugging information. The level
1869 of support is minimal and non-extensible, and is not often used.
1871 @subsection Mips debug (Third Eye)
1873 @cindex ECOFF debugging info
1874 ECOFF includes a definition of a special debug format.
1876 The file @file{mdebugread.c} implements reading for this format.
1880 @cindex DWARF 1 debugging info
1881 DWARF 1 is a debugging format that was originally designed to be
1882 used with ELF in SVR4 systems.
1887 @c If defined, these are the producer strings in a DWARF 1 file. All of
1888 @c these have reasonable defaults already.
1890 The DWARF 1 reader is in @file{dwarfread.c}.
1894 @cindex DWARF 2 debugging info
1895 DWARF 2 is an improved but incompatible version of DWARF 1.
1897 The DWARF 2 reader is in @file{dwarf2read.c}.
1901 @cindex SOM debugging info
1902 Like COFF, the SOM definition includes debugging information.
1904 @section Adding a New Symbol Reader to @value{GDBN}
1906 @cindex adding debugging info reader
1907 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
1908 there is probably little to be done.
1910 If you need to add a new object file format, you must first add it to
1911 BFD. This is beyond the scope of this document.
1913 You must then arrange for the BFD code to provide access to the
1914 debugging symbols. Generally @value{GDBN} will have to call swapping routines
1915 from BFD and a few other BFD internal routines to locate the debugging
1916 information. As much as possible, @value{GDBN} should not depend on the BFD
1917 internal data structures.
1919 For some targets (e.g., COFF), there is a special transfer vector used
1920 to call swapping routines, since the external data structures on various
1921 platforms have different sizes and layouts. Specialized routines that
1922 will only ever be implemented by one object file format may be called
1923 directly. This interface should be described in a file
1924 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
1927 @node Language Support
1929 @chapter Language Support
1931 @cindex language support
1932 @value{GDBN}'s language support is mainly driven by the symbol reader,
1933 although it is possible for the user to set the source language
1936 @value{GDBN} chooses the source language by looking at the extension
1937 of the file recorded in the debug info; @file{.c} means C, @file{.f}
1938 means Fortran, etc. It may also use a special-purpose language
1939 identifier if the debug format supports it, like with DWARF.
1941 @section Adding a Source Language to @value{GDBN}
1943 @cindex adding source language
1944 To add other languages to @value{GDBN}'s expression parser, follow the
1948 @item Create the expression parser.
1950 @cindex expression parser
1951 This should reside in a file @file{@var{lang}-exp.y}. Routines for
1952 building parsed expressions into a @code{union exp_element} list are in
1955 @cindex language parser
1956 Since we can't depend upon everyone having Bison, and YACC produces
1957 parsers that define a bunch of global names, the following lines
1958 @strong{must} be included at the top of the YACC parser, to prevent the
1959 various parsers from defining the same global names:
1962 #define yyparse @var{lang}_parse
1963 #define yylex @var{lang}_lex
1964 #define yyerror @var{lang}_error
1965 #define yylval @var{lang}_lval
1966 #define yychar @var{lang}_char
1967 #define yydebug @var{lang}_debug
1968 #define yypact @var{lang}_pact
1969 #define yyr1 @var{lang}_r1
1970 #define yyr2 @var{lang}_r2
1971 #define yydef @var{lang}_def
1972 #define yychk @var{lang}_chk
1973 #define yypgo @var{lang}_pgo
1974 #define yyact @var{lang}_act
1975 #define yyexca @var{lang}_exca
1976 #define yyerrflag @var{lang}_errflag
1977 #define yynerrs @var{lang}_nerrs
1980 At the bottom of your parser, define a @code{struct language_defn} and
1981 initialize it with the right values for your language. Define an
1982 @code{initialize_@var{lang}} routine and have it call
1983 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
1984 that your language exists. You'll need some other supporting variables
1985 and functions, which will be used via pointers from your
1986 @code{@var{lang}_language_defn}. See the declaration of @code{struct
1987 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
1988 for more information.
1990 @item Add any evaluation routines, if necessary
1992 @cindex expression evaluation routines
1993 @findex evaluate_subexp
1994 @findex prefixify_subexp
1995 @findex length_of_subexp
1996 If you need new opcodes (that represent the operations of the language),
1997 add them to the enumerated type in @file{expression.h}. Add support
1998 code for these operations in the @code{evaluate_subexp} function
1999 defined in the file @file{eval.c}. Add cases
2000 for new opcodes in two functions from @file{parse.c}:
2001 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
2002 the number of @code{exp_element}s that a given operation takes up.
2004 @item Update some existing code
2006 Add an enumerated identifier for your language to the enumerated type
2007 @code{enum language} in @file{defs.h}.
2009 Update the routines in @file{language.c} so your language is included.
2010 These routines include type predicates and such, which (in some cases)
2011 are language dependent. If your language does not appear in the switch
2012 statement, an error is reported.
2014 @vindex current_language
2015 Also included in @file{language.c} is the code that updates the variable
2016 @code{current_language}, and the routines that translate the
2017 @code{language_@var{lang}} enumerated identifier into a printable
2020 @findex _initialize_language
2021 Update the function @code{_initialize_language} to include your
2022 language. This function picks the default language upon startup, so is
2023 dependent upon which languages that @value{GDBN} is built for.
2025 @findex allocate_symtab
2026 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2027 code so that the language of each symtab (source file) is set properly.
2028 This is used to determine the language to use at each stack frame level.
2029 Currently, the language is set based upon the extension of the source
2030 file. If the language can be better inferred from the symbol
2031 information, please set the language of the symtab in the symbol-reading
2034 @findex print_subexp
2035 @findex op_print_tab
2036 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2037 expression opcodes you have added to @file{expression.h}. Also, add the
2038 printed representations of your operators to @code{op_print_tab}.
2040 @item Add a place of call
2043 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2044 @code{parse_exp_1} (defined in @file{parse.c}).
2046 @item Use macros to trim code
2048 @cindex trimming language-dependent code
2049 The user has the option of building @value{GDBN} for some or all of the
2050 languages. If the user decides to build @value{GDBN} for the language
2051 @var{lang}, then every file dependent on @file{language.h} will have the
2052 macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to
2053 leave out large routines that the user won't need if he or she is not
2054 using your language.
2056 Note that you do not need to do this in your YACC parser, since if @value{GDBN}
2057 is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the
2058 compiled form of your parser) is not linked into @value{GDBN} at all.
2060 See the file @file{configure.in} for how @value{GDBN} is configured
2061 for different languages.
2063 @item Edit @file{Makefile.in}
2065 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2066 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2067 not get linked in, or, worse yet, it may not get @code{tar}red into the
2072 @node Host Definition
2074 @chapter Host Definition
2076 With the advent of Autoconf, it's rarely necessary to have host
2077 definition machinery anymore. The following information is provided,
2078 mainly, as an historical reference.
2080 @section Adding a New Host
2082 @cindex adding a new host
2083 @cindex host, adding
2084 @value{GDBN}'s host configuration support normally happens via Autoconf.
2085 New host-specific definitions should not be needed. Older hosts
2086 @value{GDBN} still use the host-specific definitions and files listed
2087 below, but these mostly exist for historical reasons, and will
2088 eventually disappear.
2091 @item gdb/config/@var{arch}/@var{xyz}.mh
2092 This file once contained both host and native configuration information
2093 (@pxref{Native Debugging}) for the machine @var{xyz}. The host
2094 configuration information is now handed by Autoconf.
2096 Host configuration information included a definition of
2097 @code{XM_FILE=xm-@var{xyz}.h} and possibly definitions for @code{CC},
2098 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2099 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2101 New host only configurations do not need this file.
2103 @item gdb/config/@var{arch}/xm-@var{xyz}.h
2104 This file once contained definitions and includes required when hosting
2105 gdb on machine @var{xyz}. Those definitions and includes are now
2106 handled by Autoconf.
2108 New host and native configurations do not need this file.
2110 @emph{Maintainer's note: Some hosts continue to use the @file{xm-xyz.h}
2111 file to define the macros @var{HOST_FLOAT_FORMAT},
2112 @var{HOST_DOUBLE_FORMAT} and @var{HOST_LONG_DOUBLE_FORMAT}. That code
2113 also needs to be replaced with either an Autoconf or run-time test.}
2117 @subheading Generic Host Support Files
2119 @cindex generic host support
2120 There are some ``generic'' versions of routines that can be used by
2121 various systems. These can be customized in various ways by macros
2122 defined in your @file{xm-@var{xyz}.h} file. If these routines work for
2123 the @var{xyz} host, you can just include the generic file's name (with
2124 @samp{.o}, not @samp{.c}) in @code{XDEPFILES}.
2126 Otherwise, if your machine needs custom support routines, you will need
2127 to write routines that perform the same functions as the generic file.
2128 Put them into @code{@var{xyz}-xdep.c}, and put @code{@var{xyz}-xdep.o}
2129 into @code{XDEPFILES}.
2132 @cindex remote debugging support
2133 @cindex serial line support
2135 This contains serial line support for Unix systems. This is always
2136 included, via the makefile variable @code{SER_HARDWIRE}; override this
2137 variable in the @file{.mh} file to avoid it.
2140 This contains serial line support for 32-bit programs running under DOS,
2141 using the DJGPP (a.k.a.@: GO32) execution environment.
2143 @cindex TCP remote support
2145 This contains generic TCP support using sockets.
2148 @section Host Conditionals
2150 When @value{GDBN} is configured and compiled, various macros are
2151 defined or left undefined, to control compilation based on the
2152 attributes of the host system. These macros and their meanings (or if
2153 the meaning is not documented here, then one of the source files where
2154 they are used is indicated) are:
2157 @item @value{GDBN}INIT_FILENAME
2158 The default name of @value{GDBN}'s initialization file (normally
2162 This macro is deprecated.
2164 @item SIGWINCH_HANDLER
2165 If your host defines @code{SIGWINCH}, you can define this to be the name
2166 of a function to be called if @code{SIGWINCH} is received.
2168 @item SIGWINCH_HANDLER_BODY
2169 Define this to expand into code that will define the function named by
2170 the expansion of @code{SIGWINCH_HANDLER}.
2172 @item ALIGN_STACK_ON_STARTUP
2173 @cindex stack alignment
2174 Define this if your system is of a sort that will crash in
2175 @code{tgetent} if the stack happens not to be longword-aligned when
2176 @code{main} is called. This is a rare situation, but is known to occur
2177 on several different types of systems.
2179 @item CRLF_SOURCE_FILES
2180 @cindex DOS text files
2181 Define this if host files use @code{\r\n} rather than @code{\n} as a
2182 line terminator. This will cause source file listings to omit @code{\r}
2183 characters when printing and it will allow @code{\r\n} line endings of files
2184 which are ``sourced'' by gdb. It must be possible to open files in binary
2185 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2187 @item DEFAULT_PROMPT
2189 The default value of the prompt string (normally @code{"(gdb) "}).
2192 @cindex terminal device
2193 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2196 Define this if binary files are opened the same way as text files.
2200 In some cases, use the system call @code{mmap} for reading symbol
2201 tables. For some machines this allows for sharing and quick updates.
2204 Define this if the host system has @code{termio.h}.
2211 Values for host-side constants.
2214 Substitute for isatty, if not available.
2217 This is the longest integer type available on the host. If not defined,
2218 it will default to @code{long long} or @code{long}, depending on
2219 @code{CC_HAS_LONG_LONG}.
2221 @item CC_HAS_LONG_LONG
2222 @cindex @code{long long} data type
2223 Define this if the host C compiler supports @code{long long}. This is set
2224 by the @code{configure} script.
2226 @item PRINTF_HAS_LONG_LONG
2227 Define this if the host can handle printing of long long integers via
2228 the printf format conversion specifier @code{ll}. This is set by the
2229 @code{configure} script.
2231 @item HAVE_LONG_DOUBLE
2232 Define this if the host C compiler supports @code{long double}. This is
2233 set by the @code{configure} script.
2235 @item PRINTF_HAS_LONG_DOUBLE
2236 Define this if the host can handle printing of long double float-point
2237 numbers via the printf format conversion specifier @code{Lg}. This is
2238 set by the @code{configure} script.
2240 @item SCANF_HAS_LONG_DOUBLE
2241 Define this if the host can handle the parsing of long double
2242 float-point numbers via the scanf format conversion specifier
2243 @code{Lg}. This is set by the @code{configure} script.
2245 @item LSEEK_NOT_LINEAR
2246 Define this if @code{lseek (n)} does not necessarily move to byte number
2247 @code{n} in the file. This is only used when reading source files. It
2248 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2251 This macro is used as the argument to @code{lseek} (or, most commonly,
2252 @code{bfd_seek}). FIXME, should be replaced by SEEK_SET instead,
2253 which is the POSIX equivalent.
2256 If defined, this should be one or more tokens, such as @code{volatile},
2257 that can be used in both the declaration and definition of functions to
2258 indicate that they never return. The default is already set correctly
2259 if compiling with GCC. This will almost never need to be defined.
2262 If defined, this should be one or more tokens, such as
2263 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2264 of functions to indicate that they never return. The default is already
2265 set correctly if compiling with GCC. This will almost never need to be
2270 Define these to appropriate value for the system @code{lseek}, if not already
2274 This is the signal for stopping @value{GDBN}. Defaults to
2275 @code{SIGTSTP}. (Only redefined for the Convex.)
2278 Means that System V (prior to SVR4) include files are in use. (FIXME:
2279 This symbol is abused in @file{infrun.c}, @file{regex.c}, and
2280 @file{utils.c} for other things, at the moment.)
2283 Define this to help placate @code{lint} in some situations.
2286 Define this to override the defaults of @code{__volatile__} or
2291 @node Target Architecture Definition
2293 @chapter Target Architecture Definition
2295 @cindex target architecture definition
2296 @value{GDBN}'s target architecture defines what sort of
2297 machine-language programs @value{GDBN} can work with, and how it works
2300 The target architecture object is implemented as the C structure
2301 @code{struct gdbarch *}. The structure, and its methods, are generated
2302 using the Bourne shell script @file{gdbarch.sh}.
2304 @section Operating System ABI Variant Handling
2305 @cindex OS ABI variants
2307 @value{GDBN} provides a mechanism for handling variations in OS
2308 ABIs. An OS ABI variant may have influence over any number of
2309 variables in the target architecture definition. There are two major
2310 components in the OS ABI mechanism: sniffers and handlers.
2312 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2313 (the architecture may be wildcarded) in an attempt to determine the
2314 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2315 to be @dfn{generic}, while sniffers for a specific architecture are
2316 considered to be @dfn{specific}. A match from a specific sniffer
2317 overrides a match from a generic sniffer. Multiple sniffers for an
2318 architecture/flavour may exist, in order to differentiate between two
2319 different operating systems which use the same basic file format. The
2320 OS ABI framework provides a generic sniffer for ELF-format files which
2321 examines the @code{EI_OSABI} field of the ELF header, as well as note
2322 sections known to be used by several operating systems.
2324 @cindex fine-tuning @code{gdbarch} structure
2325 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2326 selected OS ABI. There may be only one handler for a given OS ABI
2327 for each BFD architecture.
2329 The following OS ABI variants are defined in @file{osabi.h}:
2333 @findex GDB_OSABI_UNKNOWN
2334 @item GDB_OSABI_UNKNOWN
2335 The ABI of the inferior is unknown. The default @code{gdbarch}
2336 settings for the architecture will be used.
2338 @findex GDB_OSABI_SVR4
2339 @item GDB_OSABI_SVR4
2340 UNIX System V Release 4
2342 @findex GDB_OSABI_HURD
2343 @item GDB_OSABI_HURD
2344 GNU using the Hurd kernel
2346 @findex GDB_OSABI_SOLARIS
2347 @item GDB_OSABI_SOLARIS
2350 @findex GDB_OSABI_OSF1
2351 @item GDB_OSABI_OSF1
2352 OSF/1, including Digital UNIX and Compaq Tru64 UNIX
2354 @findex GDB_OSABI_LINUX
2355 @item GDB_OSABI_LINUX
2356 GNU using the Linux kernel
2358 @findex GDB_OSABI_FREEBSD_AOUT
2359 @item GDB_OSABI_FREEBSD_AOUT
2360 FreeBSD using the a.out executable format
2362 @findex GDB_OSABI_FREEBSD_ELF
2363 @item GDB_OSABI_FREEBSD_ELF
2364 FreeBSD using the ELF executable format
2366 @findex GDB_OSABI_NETBSD_AOUT
2367 @item GDB_OSABI_NETBSD_AOUT
2368 NetBSD using the a.out executable format
2370 @findex GDB_OSABI_NETBSD_ELF
2371 @item GDB_OSABI_NETBSD_ELF
2372 NetBSD using the ELF executable format
2374 @findex GDB_OSABI_WINCE
2375 @item GDB_OSABI_WINCE
2378 @findex GDB_OSABI_GO32
2379 @item GDB_OSABI_GO32
2382 @findex GDB_OSABI_NETWARE
2383 @item GDB_OSABI_NETWARE
2386 @findex GDB_OSABI_ARM_EABI_V1
2387 @item GDB_OSABI_ARM_EABI_V1
2388 ARM Embedded ABI version 1
2390 @findex GDB_OSABI_ARM_EABI_V2
2391 @item GDB_OSABI_ARM_EABI_V2
2392 ARM Embedded ABI version 2
2394 @findex GDB_OSABI_ARM_APCS
2395 @item GDB_OSABI_ARM_APCS
2396 Generic ARM Procedure Call Standard
2400 Here are the functions that make up the OS ABI framework:
2402 @deftypefun const char *gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2403 Return the name of the OS ABI corresponding to @var{osabi}.
2406 @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}))
2407 Register the OS ABI handler specified by @var{init_osabi} for the
2408 architecture, machine type and OS ABI specified by @var{arch},
2409 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2410 machine type, which implies the architecture's default machine type,
2414 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2415 Register the OS ABI file sniffer specified by @var{sniffer} for the
2416 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2417 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2418 be generic, and is allowed to examine @var{flavour}-flavoured files for
2422 @deftypefun enum gdb_osabi gdbarch_lookup_osabi (bfd *@var{abfd})
2423 Examine the file described by @var{abfd} to determine its OS ABI.
2424 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2428 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2429 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2430 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2431 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2432 architecture, a warning will be issued and the debugging session will continue
2433 with the defaults already established for @var{gdbarch}.
2436 @section Registers and Memory
2438 @value{GDBN}'s model of the target machine is rather simple.
2439 @value{GDBN} assumes the machine includes a bank of registers and a
2440 block of memory. Each register may have a different size.
2442 @value{GDBN} does not have a magical way to match up with the
2443 compiler's idea of which registers are which; however, it is critical
2444 that they do match up accurately. The only way to make this work is
2445 to get accurate information about the order that the compiler uses,
2446 and to reflect that in the @code{REGISTER_NAME} and related macros.
2448 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
2450 @section Pointers Are Not Always Addresses
2451 @cindex pointer representation
2452 @cindex address representation
2453 @cindex word-addressed machines
2454 @cindex separate data and code address spaces
2455 @cindex spaces, separate data and code address
2456 @cindex address spaces, separate data and code
2457 @cindex code pointers, word-addressed
2458 @cindex converting between pointers and addresses
2459 @cindex D10V addresses
2461 On almost all 32-bit architectures, the representation of a pointer is
2462 indistinguishable from the representation of some fixed-length number
2463 whose value is the byte address of the object pointed to. On such
2464 machines, the words ``pointer'' and ``address'' can be used interchangeably.
2465 However, architectures with smaller word sizes are often cramped for
2466 address space, so they may choose a pointer representation that breaks this
2467 identity, and allows a larger code address space.
2469 For example, the Renesas D10V is a 16-bit VLIW processor whose
2470 instructions are 32 bits long@footnote{Some D10V instructions are
2471 actually pairs of 16-bit sub-instructions. However, since you can't
2472 jump into the middle of such a pair, code addresses can only refer to
2473 full 32 bit instructions, which is what matters in this explanation.}.
2474 If the D10V used ordinary byte addresses to refer to code locations,
2475 then the processor would only be able to address 64kb of instructions.
2476 However, since instructions must be aligned on four-byte boundaries, the
2477 low two bits of any valid instruction's byte address are always
2478 zero---byte addresses waste two bits. So instead of byte addresses,
2479 the D10V uses word addresses---byte addresses shifted right two bits---to
2480 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
2483 However, this means that code pointers and data pointers have different
2484 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
2485 @code{0xC020} when used as a data address, but refers to byte address
2486 @code{0x30080} when used as a code address.
2488 (The D10V also uses separate code and data address spaces, which also
2489 affects the correspondence between pointers and addresses, but we're
2490 going to ignore that here; this example is already too long.)
2492 To cope with architectures like this---the D10V is not the only
2493 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
2494 byte numbers, and @dfn{pointers}, which are the target's representation
2495 of an address of a particular type of data. In the example above,
2496 @code{0xC020} is the pointer, which refers to one of the addresses
2497 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
2498 @value{GDBN} provides functions for turning a pointer into an address
2499 and vice versa, in the appropriate way for the current architecture.
2501 Unfortunately, since addresses and pointers are identical on almost all
2502 processors, this distinction tends to bit-rot pretty quickly. Thus,
2503 each time you port @value{GDBN} to an architecture which does
2504 distinguish between pointers and addresses, you'll probably need to
2505 clean up some architecture-independent code.
2507 Here are functions which convert between pointers and addresses:
2509 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
2510 Treat the bytes at @var{buf} as a pointer or reference of type
2511 @var{type}, and return the address it represents, in a manner
2512 appropriate for the current architecture. This yields an address
2513 @value{GDBN} can use to read target memory, disassemble, etc. Note that
2514 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2517 For example, if the current architecture is the Intel x86, this function
2518 extracts a little-endian integer of the appropriate length from
2519 @var{buf} and returns it. However, if the current architecture is the
2520 D10V, this function will return a 16-bit integer extracted from
2521 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
2523 If @var{type} is not a pointer or reference type, then this function
2524 will signal an internal error.
2527 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
2528 Store the address @var{addr} in @var{buf}, in the proper format for a
2529 pointer of type @var{type} in the current architecture. Note that
2530 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2533 For example, if the current architecture is the Intel x86, this function
2534 stores @var{addr} unmodified as a little-endian integer of the
2535 appropriate length in @var{buf}. However, if the current architecture
2536 is the D10V, this function divides @var{addr} by four if @var{type} is
2537 a pointer to a function, and then stores it in @var{buf}.
2539 If @var{type} is not a pointer or reference type, then this function
2540 will signal an internal error.
2543 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
2544 Assuming that @var{val} is a pointer, return the address it represents,
2545 as appropriate for the current architecture.
2547 This function actually works on integral values, as well as pointers.
2548 For pointers, it performs architecture-specific conversions as
2549 described above for @code{extract_typed_address}.
2552 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
2553 Create and return a value representing a pointer of type @var{type} to
2554 the address @var{addr}, as appropriate for the current architecture.
2555 This function performs architecture-specific conversions as described
2556 above for @code{store_typed_address}.
2559 Here are some macros which architectures can define to indicate the
2560 relationship between pointers and addresses. These have default
2561 definitions, appropriate for architectures on which all pointers are
2562 simple unsigned byte addresses.
2564 @deftypefn {Target Macro} CORE_ADDR POINTER_TO_ADDRESS (struct type *@var{type}, char *@var{buf})
2565 Assume that @var{buf} holds a pointer of type @var{type}, in the
2566 appropriate format for the current architecture. Return the byte
2567 address the pointer refers to.
2569 This function may safely assume that @var{type} is either a pointer or a
2570 C@t{++} reference type.
2573 @deftypefn {Target Macro} void ADDRESS_TO_POINTER (struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
2574 Store in @var{buf} a pointer of type @var{type} representing the address
2575 @var{addr}, in the appropriate format for the current architecture.
2577 This function may safely assume that @var{type} is either a pointer or a
2578 C@t{++} reference type.
2581 @section Address Classes
2582 @cindex address classes
2583 @cindex DW_AT_byte_size
2584 @cindex DW_AT_address_class
2586 Sometimes information about different kinds of addresses is available
2587 via the debug information. For example, some programming environments
2588 define addresses of several different sizes. If the debug information
2589 distinguishes these kinds of address classes through either the size
2590 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
2591 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
2592 following macros should be defined in order to disambiguate these
2593 types within @value{GDBN} as well as provide the added information to
2594 a @value{GDBN} user when printing type expressions.
2596 @deftypefn {Target Macro} int ADDRESS_CLASS_TYPE_FLAGS (int @var{byte_size}, int @var{dwarf2_addr_class})
2597 Returns the type flags needed to construct a pointer type whose size
2598 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
2599 This function is normally called from within a symbol reader. See
2600 @file{dwarf2read.c}.
2603 @deftypefn {Target Macro} char *ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (int @var{type_flags})
2604 Given the type flags representing an address class qualifier, return
2607 @deftypefn {Target Macro} int ADDRESS_CLASS_NAME_to_TYPE_FLAGS (int @var{name}, int *var{type_flags_ptr})
2608 Given an address qualifier name, set the @code{int} refererenced by @var{type_flags_ptr} to the type flags
2609 for that address class qualifier.
2612 Since the need for address classes is rather rare, none of
2613 the address class macros defined by default. Predicate
2614 macros are provided to detect when they are defined.
2616 Consider a hypothetical architecture in which addresses are normally
2617 32-bits wide, but 16-bit addresses are also supported. Furthermore,
2618 suppose that the @w{DWARF 2} information for this architecture simply
2619 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
2620 of these "short" pointers. The following functions could be defined
2621 to implement the address class macros:
2624 somearch_address_class_type_flags (int byte_size,
2625 int dwarf2_addr_class)
2628 return TYPE_FLAG_ADDRESS_CLASS_1;
2634 somearch_address_class_type_flags_to_name (int type_flags)
2636 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
2643 somearch_address_class_name_to_type_flags (char *name,
2644 int *type_flags_ptr)
2646 if (strcmp (name, "short") == 0)
2648 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
2656 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
2657 to indicate the presence of one of these "short" pointers. E.g, if
2658 the debug information indicates that @code{short_ptr_var} is one of these
2659 short pointers, @value{GDBN} might show the following behavior:
2662 (gdb) ptype short_ptr_var
2663 type = int * @@short
2667 @section Raw and Virtual Register Representations
2668 @cindex raw register representation
2669 @cindex virtual register representation
2670 @cindex representations, raw and virtual registers
2672 @emph{Maintainer note: This section is pretty much obsolete. The
2673 functionality described here has largely been replaced by
2674 pseudo-registers and the mechanisms described in @ref{Target
2675 Architecture Definition, , Using Different Register and Memory Data
2676 Representations}. See also @uref{http://www.gnu.org/software/gdb/bugs/,
2677 Bug Tracking Database} and
2678 @uref{http://sources.redhat.com/gdb/current/ari/, ARI Index} for more
2679 up-to-date information.}
2681 Some architectures use one representation for a value when it lives in a
2682 register, but use a different representation when it lives in memory.
2683 In @value{GDBN}'s terminology, the @dfn{raw} representation is the one used in
2684 the target registers, and the @dfn{virtual} representation is the one
2685 used in memory, and within @value{GDBN} @code{struct value} objects.
2687 @emph{Maintainer note: Notice that the same mechanism is being used to
2688 both convert a register to a @code{struct value} and alternative
2691 For almost all data types on almost all architectures, the virtual and
2692 raw representations are identical, and no special handling is needed.
2693 However, they do occasionally differ. For example:
2697 The x86 architecture supports an 80-bit @code{long double} type. However, when
2698 we store those values in memory, they occupy twelve bytes: the
2699 floating-point number occupies the first ten, and the final two bytes
2700 are unused. This keeps the values aligned on four-byte boundaries,
2701 allowing more efficient access. Thus, the x86 80-bit floating-point
2702 type is the raw representation, and the twelve-byte loosely-packed
2703 arrangement is the virtual representation.
2706 Some 64-bit MIPS targets present 32-bit registers to @value{GDBN} as 64-bit
2707 registers, with garbage in their upper bits. @value{GDBN} ignores the top 32
2708 bits. Thus, the 64-bit form, with garbage in the upper 32 bits, is the
2709 raw representation, and the trimmed 32-bit representation is the
2710 virtual representation.
2713 In general, the raw representation is determined by the architecture, or
2714 @value{GDBN}'s interface to the architecture, while the virtual representation
2715 can be chosen for @value{GDBN}'s convenience. @value{GDBN}'s register file,
2716 @code{registers}, holds the register contents in raw format, and the
2717 @value{GDBN} remote protocol transmits register values in raw format.
2719 Your architecture may define the following macros to request
2720 conversions between the raw and virtual format:
2722 @deftypefn {Target Macro} int REGISTER_CONVERTIBLE (int @var{reg})
2723 Return non-zero if register number @var{reg}'s value needs different raw
2724 and virtual formats.
2726 You should not use @code{REGISTER_CONVERT_TO_VIRTUAL} for a register
2727 unless this macro returns a non-zero value for that register.
2730 @deftypefn {Target Macro} int DEPRECATED_REGISTER_RAW_SIZE (int @var{reg})
2731 The size of register number @var{reg}'s raw value. This is the number
2732 of bytes the register will occupy in @code{registers}, or in a @value{GDBN}
2733 remote protocol packet.
2736 @deftypefn {Target Macro} int DEPRECATED_REGISTER_VIRTUAL_SIZE (int @var{reg})
2737 The size of register number @var{reg}'s value, in its virtual format.
2738 This is the size a @code{struct value}'s buffer will have, holding that
2742 @deftypefn {Target Macro} struct type *DEPRECATED_REGISTER_VIRTUAL_TYPE (int @var{reg})
2743 This is the type of the virtual representation of register number
2744 @var{reg}. Note that there is no need for a macro giving a type for the
2745 register's raw form; once the register's value has been obtained, @value{GDBN}
2746 always uses the virtual form.
2749 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_VIRTUAL (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
2750 Convert the value of register number @var{reg} to @var{type}, which
2751 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2752 at @var{from} holds the register's value in raw format; the macro should
2753 convert the value to virtual format, and place it at @var{to}.
2755 Note that @code{REGISTER_CONVERT_TO_VIRTUAL} and
2756 @code{REGISTER_CONVERT_TO_RAW} take their @var{reg} and @var{type}
2757 arguments in different orders.
2759 You should only use @code{REGISTER_CONVERT_TO_VIRTUAL} with registers
2760 for which the @code{REGISTER_CONVERTIBLE} macro returns a non-zero
2764 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_RAW (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
2765 Convert the value of register number @var{reg} to @var{type}, which
2766 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2767 at @var{from} holds the register's value in raw format; the macro should
2768 convert the value to virtual format, and place it at @var{to}.
2770 Note that REGISTER_CONVERT_TO_VIRTUAL and REGISTER_CONVERT_TO_RAW take
2771 their @var{reg} and @var{type} arguments in different orders.
2775 @section Using Different Register and Memory Data Representations
2776 @cindex register representation
2777 @cindex memory representation
2778 @cindex representations, register and memory
2779 @cindex register data formats, converting
2780 @cindex @code{struct value}, converting register contents to
2782 @emph{Maintainer's note: The way GDB manipulates registers is undergoing
2783 significant change. Many of the macros and functions refered to in this
2784 section are likely to be subject to further revision. See
2785 @uref{http://sources.redhat.com/gdb/current/ari/, A.R. Index} and
2786 @uref{http://www.gnu.org/software/gdb/bugs, Bug Tracking Database} for
2787 further information. cagney/2002-05-06.}
2789 Some architectures can represent a data object in a register using a
2790 form that is different to the objects more normal memory representation.
2796 The Alpha architecture can represent 32 bit integer values in
2797 floating-point registers.
2800 The x86 architecture supports 80-bit floating-point registers. The
2801 @code{long double} data type occupies 96 bits in memory but only 80 bits
2802 when stored in a register.
2806 In general, the register representation of a data type is determined by
2807 the architecture, or @value{GDBN}'s interface to the architecture, while
2808 the memory representation is determined by the Application Binary
2811 For almost all data types on almost all architectures, the two
2812 representations are identical, and no special handling is needed.
2813 However, they do occasionally differ. Your architecture may define the
2814 following macros to request conversions between the register and memory
2815 representations of a data type:
2817 @deftypefn {Target Macro} int CONVERT_REGISTER_P (int @var{reg})
2818 Return non-zero if the representation of a data value stored in this
2819 register may be different to the representation of that same data value
2820 when stored in memory.
2822 When non-zero, the macros @code{REGISTER_TO_VALUE} and
2823 @code{VALUE_TO_REGISTER} are used to perform any necessary conversion.
2826 @deftypefn {Target Macro} void REGISTER_TO_VALUE (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
2827 Convert the value of register number @var{reg} to a data object of type
2828 @var{type}. The buffer at @var{from} holds the register's value in raw
2829 format; the converted value should be placed in the buffer at @var{to}.
2831 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
2832 their @var{reg} and @var{type} arguments in different orders.
2834 You should only use @code{REGISTER_TO_VALUE} with registers for which
2835 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
2838 @deftypefn {Target Macro} void VALUE_TO_REGISTER (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
2839 Convert a data value of type @var{type} to register number @var{reg}'
2842 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
2843 their @var{reg} and @var{type} arguments in different orders.
2845 You should only use @code{VALUE_TO_REGISTER} with registers for which
2846 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
2849 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_TYPE (int @var{regnum}, struct type *@var{type}, char *@var{buf})
2850 See @file{mips-tdep.c}. It does not do what you want.
2854 @section Frame Interpretation
2856 @section Inferior Call Setup
2858 @section Compiler Characteristics
2860 @section Target Conditionals
2862 This section describes the macros that you can use to define the target
2867 @item ADDR_BITS_REMOVE (addr)
2868 @findex ADDR_BITS_REMOVE
2869 If a raw machine instruction address includes any bits that are not
2870 really part of the address, then define this macro to expand into an
2871 expression that zeroes those bits in @var{addr}. This is only used for
2872 addresses of instructions, and even then not in all contexts.
2874 For example, the two low-order bits of the PC on the Hewlett-Packard PA
2875 2.0 architecture contain the privilege level of the corresponding
2876 instruction. Since instructions must always be aligned on four-byte
2877 boundaries, the processor masks out these bits to generate the actual
2878 address of the instruction. ADDR_BITS_REMOVE should filter out these
2879 bits with an expression such as @code{((addr) & ~3)}.
2881 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS (@var{name}, @var{type_flags_ptr})
2882 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS
2883 If @var{name} is a valid address class qualifier name, set the @code{int}
2884 referenced by @var{type_flags_ptr} to the mask representing the qualifier
2885 and return 1. If @var{name} is not a valid address class qualifier name,
2888 The value for @var{type_flags_ptr} should be one of
2889 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
2890 possibly some combination of these values or'd together.
2891 @xref{Target Architecture Definition, , Address Classes}.
2893 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P ()
2894 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P
2895 Predicate which indicates whether @code{ADDRESS_CLASS_NAME_TO_TYPE_FLAGS}
2898 @item ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
2899 @findex ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
2900 Given a pointers byte size (as described by the debug information) and
2901 the possible @code{DW_AT_address_class} value, return the type flags
2902 used by @value{GDBN} to represent this address class. The value
2903 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
2904 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
2905 values or'd together.
2906 @xref{Target Architecture Definition, , Address Classes}.
2908 @item ADDRESS_CLASS_TYPE_FLAGS_P ()
2909 @findex ADDRESS_CLASS_TYPE_FLAGS_P
2910 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS} has
2913 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (@var{type_flags})
2914 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME
2915 Return the name of the address class qualifier associated with the type
2916 flags given by @var{type_flags}.
2918 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P ()
2919 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P
2920 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS_TO_NAME} has
2922 @xref{Target Architecture Definition, , Address Classes}.
2924 @item ADDRESS_TO_POINTER (@var{type}, @var{buf}, @var{addr})
2925 @findex ADDRESS_TO_POINTER
2926 Store in @var{buf} a pointer of type @var{type} representing the address
2927 @var{addr}, in the appropriate format for the current architecture.
2928 This macro may safely assume that @var{type} is either a pointer or a
2929 C@t{++} reference type.
2930 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
2932 @item BELIEVE_PCC_PROMOTION
2933 @findex BELIEVE_PCC_PROMOTION
2934 Define if the compiler promotes a @code{short} or @code{char}
2935 parameter to an @code{int}, but still reports the parameter as its
2936 original type, rather than the promoted type.
2938 @item BITS_BIG_ENDIAN
2939 @findex BITS_BIG_ENDIAN
2940 Define this if the numbering of bits in the targets does @strong{not} match the
2941 endianness of the target byte order. A value of 1 means that the bits
2942 are numbered in a big-endian bit order, 0 means little-endian.
2946 This is the character array initializer for the bit pattern to put into
2947 memory where a breakpoint is set. Although it's common to use a trap
2948 instruction for a breakpoint, it's not required; for instance, the bit
2949 pattern could be an invalid instruction. The breakpoint must be no
2950 longer than the shortest instruction of the architecture.
2952 @code{BREAKPOINT} has been deprecated in favor of
2953 @code{BREAKPOINT_FROM_PC}.
2955 @item BIG_BREAKPOINT
2956 @itemx LITTLE_BREAKPOINT
2957 @findex LITTLE_BREAKPOINT
2958 @findex BIG_BREAKPOINT
2959 Similar to BREAKPOINT, but used for bi-endian targets.
2961 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
2962 favor of @code{BREAKPOINT_FROM_PC}.
2964 @item DEPRECATED_REMOTE_BREAKPOINT
2965 @itemx DEPRECATED_LITTLE_REMOTE_BREAKPOINT
2966 @itemx DEPRECATED_BIG_REMOTE_BREAKPOINT
2967 @findex DEPRECATED_BIG_REMOTE_BREAKPOINT
2968 @findex DEPRECATED_LITTLE_REMOTE_BREAKPOINT
2969 @findex DEPRECATED_REMOTE_BREAKPOINT
2970 Specify the breakpoint instruction sequence for a remote target.
2971 @code{DEPRECATED_REMOTE_BREAKPOINT},
2972 @code{DEPRECATED_BIG_REMOTE_BREAKPOINT} and
2973 @code{DEPRECATED_LITTLE_REMOTE_BREAKPOINT} have been deprecated in
2974 favor of @code{BREAKPOINT_FROM_PC} (@pxref{BREAKPOINT_FROM_PC}).
2976 @item BREAKPOINT_FROM_PC (@var{pcptr}, @var{lenptr})
2977 @findex BREAKPOINT_FROM_PC
2978 @anchor{BREAKPOINT_FROM_PC} Use the program counter to determine the
2979 contents and size of a breakpoint instruction. It returns a pointer to
2980 a string of bytes that encode a breakpoint instruction, stores the
2981 length of the string to @code{*@var{lenptr}}, and adjusts the program
2982 counter (if necessary) to point to the actual memory location where the
2983 breakpoint should be inserted.
2985 Although it is common to use a trap instruction for a breakpoint, it's
2986 not required; for instance, the bit pattern could be an invalid
2987 instruction. The breakpoint must be no longer than the shortest
2988 instruction of the architecture.
2990 Replaces all the other @var{BREAKPOINT} macros.
2992 @item MEMORY_INSERT_BREAKPOINT (@var{addr}, @var{contents_cache})
2993 @itemx MEMORY_REMOVE_BREAKPOINT (@var{addr}, @var{contents_cache})
2994 @findex MEMORY_REMOVE_BREAKPOINT
2995 @findex MEMORY_INSERT_BREAKPOINT
2996 Insert or remove memory based breakpoints. Reasonable defaults
2997 (@code{default_memory_insert_breakpoint} and
2998 @code{default_memory_remove_breakpoint} respectively) have been
2999 provided so that it is not necessary to define these for most
3000 architectures. Architectures which may want to define
3001 @code{MEMORY_INSERT_BREAKPOINT} and @code{MEMORY_REMOVE_BREAKPOINT} will
3002 likely have instructions that are oddly sized or are not stored in a
3003 conventional manner.
3005 It may also be desirable (from an efficiency standpoint) to define
3006 custom breakpoint insertion and removal routines if
3007 @code{BREAKPOINT_FROM_PC} needs to read the target's memory for some
3010 @item ADJUST_BREAKPOINT_ADDRESS (@var{address})
3011 @findex ADJUST_BREAKPOINT_ADDRESS
3012 @cindex breakpoint address adjusted
3013 Given an address at which a breakpoint is desired, return a breakpoint
3014 address adjusted to account for architectural constraints on
3015 breakpoint placement. This method is not needed by most targets.
3017 The FR-V target (see @file{frv-tdep.c}) requires this method.
3018 The FR-V is a VLIW architecture in which a number of RISC-like
3019 instructions are grouped (packed) together into an aggregate
3020 instruction or instruction bundle. When the processor executes
3021 one of these bundles, the component instructions are executed
3024 In the course of optimization, the compiler may group instructions
3025 from distinct source statements into the same bundle. The line number
3026 information associated with one of the latter statements will likely
3027 refer to some instruction other than the first one in the bundle. So,
3028 if the user attempts to place a breakpoint on one of these latter
3029 statements, @value{GDBN} must be careful to @emph{not} place the break
3030 instruction on any instruction other than the first one in the bundle.
3031 (Remember though that the instructions within a bundle execute
3032 in parallel, so the @emph{first} instruction is the instruction
3033 at the lowest address and has nothing to do with execution order.)
3035 The FR-V's @code{ADJUST_BREAKPOINT_ADDRESS} method will adjust a
3036 breakpoint's address by scanning backwards for the beginning of
3037 the bundle, returning the address of the bundle.
3039 Since the adjustment of a breakpoint may significantly alter a user's
3040 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
3041 is initially set and each time that that breakpoint is hit.
3043 @item CALL_DUMMY_LOCATION
3044 @findex CALL_DUMMY_LOCATION
3045 See the file @file{inferior.h}.
3047 This method has been replaced by @code{push_dummy_code}
3048 (@pxref{push_dummy_code}).
3050 @item CANNOT_FETCH_REGISTER (@var{regno})
3051 @findex CANNOT_FETCH_REGISTER
3052 A C expression that should be nonzero if @var{regno} cannot be fetched
3053 from an inferior process. This is only relevant if
3054 @code{FETCH_INFERIOR_REGISTERS} is not defined.
3056 @item CANNOT_STORE_REGISTER (@var{regno})
3057 @findex CANNOT_STORE_REGISTER
3058 A C expression that should be nonzero if @var{regno} should not be
3059 written to the target. This is often the case for program counters,
3060 status words, and other special registers. If this is not defined,
3061 @value{GDBN} will assume that all registers may be written.
3063 @item int CONVERT_REGISTER_P(@var{regnum})
3064 @findex CONVERT_REGISTER_P
3065 Return non-zero if register @var{regnum} can represent data values in a
3067 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3069 @item DECR_PC_AFTER_BREAK
3070 @findex DECR_PC_AFTER_BREAK
3071 Define this to be the amount by which to decrement the PC after the
3072 program encounters a breakpoint. This is often the number of bytes in
3073 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
3075 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
3076 @findex DISABLE_UNSETTABLE_BREAK
3077 If defined, this should evaluate to 1 if @var{addr} is in a shared
3078 library in which breakpoints cannot be set and so should be disabled.
3080 @item PRINT_FLOAT_INFO()
3081 @findex PRINT_FLOAT_INFO
3082 If defined, then the @samp{info float} command will print information about
3083 the processor's floating point unit.
3085 @item print_registers_info (@var{gdbarch}, @var{frame}, @var{regnum}, @var{all})
3086 @findex print_registers_info
3087 If defined, pretty print the value of the register @var{regnum} for the
3088 specified @var{frame}. If the value of @var{regnum} is -1, pretty print
3089 either all registers (@var{all} is non zero) or a select subset of
3090 registers (@var{all} is zero).
3092 The default method prints one register per line, and if @var{all} is
3093 zero omits floating-point registers.
3095 @item PRINT_VECTOR_INFO()
3096 @findex PRINT_VECTOR_INFO
3097 If defined, then the @samp{info vector} command will call this function
3098 to print information about the processor's vector unit.
3100 By default, the @samp{info vector} command will print all vector
3101 registers (the register's type having the vector attribute).
3103 @item DWARF_REG_TO_REGNUM
3104 @findex DWARF_REG_TO_REGNUM
3105 Convert DWARF register number into @value{GDBN} regnum. If not defined,
3106 no conversion will be performed.
3108 @item DWARF2_REG_TO_REGNUM
3109 @findex DWARF2_REG_TO_REGNUM
3110 Convert DWARF2 register number into @value{GDBN} regnum. If not
3111 defined, no conversion will be performed.
3113 @item ECOFF_REG_TO_REGNUM
3114 @findex ECOFF_REG_TO_REGNUM
3115 Convert ECOFF register number into @value{GDBN} regnum. If not defined,
3116 no conversion will be performed.
3118 @item END_OF_TEXT_DEFAULT
3119 @findex END_OF_TEXT_DEFAULT
3120 This is an expression that should designate the end of the text section.
3123 @item EXTRACT_RETURN_VALUE(@var{type}, @var{regbuf}, @var{valbuf})
3124 @findex EXTRACT_RETURN_VALUE
3125 Define this to extract a function's return value of type @var{type} from
3126 the raw register state @var{regbuf} and copy that, in virtual format,
3129 This method has been deprecated in favour of @code{gdbarch_return_value}
3130 (@pxref{gdbarch_return_value}).
3132 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS(@var{regbuf})
3133 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS
3134 @anchor{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}
3135 When defined, extract from the array @var{regbuf} (containing the raw
3136 register state) the @code{CORE_ADDR} at which a function should return
3137 its structure value.
3139 @xref{gdbarch_return_value}.
3141 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P()
3142 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P
3143 Predicate for @code{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}.
3145 @item DEPRECATED_FP_REGNUM
3146 @findex DEPRECATED_FP_REGNUM
3147 If the virtual frame pointer is kept in a register, then define this
3148 macro to be the number (greater than or equal to zero) of that register.
3150 This should only need to be defined if @code{DEPRECATED_TARGET_READ_FP}
3153 @item DEPRECATED_FRAMELESS_FUNCTION_INVOCATION(@var{fi})
3154 @findex DEPRECATED_FRAMELESS_FUNCTION_INVOCATION
3155 Define this to an expression that returns 1 if the function invocation
3156 represented by @var{fi} does not have a stack frame associated with it.
3159 @item frame_align (@var{address})
3160 @anchor{frame_align}
3162 Define this to adjust @var{address} so that it meets the alignment
3163 requirements for the start of a new stack frame. A stack frame's
3164 alignment requirements are typically stronger than a target processors
3165 stack alignment requirements (@pxref{DEPRECATED_STACK_ALIGN}).
3167 This function is used to ensure that, when creating a dummy frame, both
3168 the initial stack pointer and (if needed) the address of the return
3169 value are correctly aligned.
3171 Unlike @code{DEPRECATED_STACK_ALIGN}, this function always adjusts the
3172 address in the direction of stack growth.
3174 By default, no frame based stack alignment is performed.
3176 @item int frame_red_zone_size
3178 The number of bytes, beyond the innermost-stack-address, reserved by the
3179 @sc{abi}. A function is permitted to use this scratch area (instead of
3180 allocating extra stack space).
3182 When performing an inferior function call, to ensure that it does not
3183 modify this area, @value{GDBN} adjusts the innermost-stack-address by
3184 @var{frame_red_zone_size} bytes before pushing parameters onto the
3187 By default, zero bytes are allocated. The value must be aligned
3188 (@pxref{frame_align}).
3190 The @sc{amd64} (nee x86-64) @sc{abi} documentation refers to the
3191 @emph{red zone} when describing this scratch area.
3194 @item DEPRECATED_FRAME_CHAIN(@var{frame})
3195 @findex DEPRECATED_FRAME_CHAIN
3196 Given @var{frame}, return a pointer to the calling frame.
3198 @item DEPRECATED_FRAME_CHAIN_VALID(@var{chain}, @var{thisframe})
3199 @findex DEPRECATED_FRAME_CHAIN_VALID
3200 Define this to be an expression that returns zero if the given frame is an
3201 outermost frame, with no caller, and nonzero otherwise. Most normal
3202 situations can be handled without defining this macro, including @code{NULL}
3203 chain pointers, dummy frames, and frames whose PC values are inside the
3204 startup file (e.g.@: @file{crt0.o}), inside @code{main}, or inside
3207 @item DEPRECATED_FRAME_INIT_SAVED_REGS(@var{frame})
3208 @findex DEPRECATED_FRAME_INIT_SAVED_REGS
3209 See @file{frame.h}. Determines the address of all registers in the
3210 current stack frame storing each in @code{frame->saved_regs}. Space for
3211 @code{frame->saved_regs} shall be allocated by
3212 @code{DEPRECATED_FRAME_INIT_SAVED_REGS} using
3213 @code{frame_saved_regs_zalloc}.
3215 @code{FRAME_FIND_SAVED_REGS} is deprecated.
3217 @item FRAME_NUM_ARGS (@var{fi})
3218 @findex FRAME_NUM_ARGS
3219 For the frame described by @var{fi} return the number of arguments that
3220 are being passed. If the number of arguments is not known, return
3223 @item DEPRECATED_FRAME_SAVED_PC(@var{frame})
3224 @findex DEPRECATED_FRAME_SAVED_PC
3225 @anchor{DEPRECATED_FRAME_SAVED_PC} Given @var{frame}, return the pc
3226 saved there. This is the return address.
3228 This method is deprecated. @xref{unwind_pc}.
3230 @item CORE_ADDR unwind_pc (struct frame_info *@var{this_frame})
3232 @anchor{unwind_pc} Return the instruction address, in @var{this_frame}'s
3233 caller, at which execution will resume after @var{this_frame} returns.
3234 This is commonly refered to as the return address.
3236 The implementation, which must be frame agnostic (work with any frame),
3237 is typically no more than:
3241 frame_unwind_unsigned_register (this_frame, D10V_PC_REGNUM, &pc);
3242 return d10v_make_iaddr (pc);
3246 @xref{DEPRECATED_FRAME_SAVED_PC}, which this method replaces.
3248 @item CORE_ADDR unwind_sp (struct frame_info *@var{this_frame})
3250 @anchor{unwind_sp} Return the frame's inner most stack address. This is
3251 commonly refered to as the frame's @dfn{stack pointer}.
3253 The implementation, which must be frame agnostic (work with any frame),
3254 is typically no more than:
3258 frame_unwind_unsigned_register (this_frame, D10V_SP_REGNUM, &sp);
3259 return d10v_make_daddr (sp);
3263 @xref{TARGET_READ_SP}, which this method replaces.
3265 @item FUNCTION_EPILOGUE_SIZE
3266 @findex FUNCTION_EPILOGUE_SIZE
3267 For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the
3268 function end symbol is 0. For such targets, you must define
3269 @code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size of a
3270 function's epilogue.
3272 @item DEPRECATED_FUNCTION_START_OFFSET
3273 @findex DEPRECATED_FUNCTION_START_OFFSET
3274 An integer, giving the offset in bytes from a function's address (as
3275 used in the values of symbols, function pointers, etc.), and the
3276 function's first genuine instruction.
3278 This is zero on almost all machines: the function's address is usually
3279 the address of its first instruction. However, on the VAX, for
3280 example, each function starts with two bytes containing a bitmask
3281 indicating which registers to save upon entry to the function. The
3282 VAX @code{call} instructions check this value, and save the
3283 appropriate registers automatically. Thus, since the offset from the
3284 function's address to its first instruction is two bytes,
3285 @code{DEPRECATED_FUNCTION_START_OFFSET} would be 2 on the VAX.
3287 @item GCC_COMPILED_FLAG_SYMBOL
3288 @itemx GCC2_COMPILED_FLAG_SYMBOL
3289 @findex GCC2_COMPILED_FLAG_SYMBOL
3290 @findex GCC_COMPILED_FLAG_SYMBOL
3291 If defined, these are the names of the symbols that @value{GDBN} will
3292 look for to detect that GCC compiled the file. The default symbols
3293 are @code{gcc_compiled.} and @code{gcc2_compiled.},
3294 respectively. (Currently only defined for the Delta 68.)
3296 @item @value{GDBN}_MULTI_ARCH
3297 @findex @value{GDBN}_MULTI_ARCH
3298 If defined and non-zero, enables support for multiple architectures
3299 within @value{GDBN}.
3301 This support can be enabled at two levels. At level one, only
3302 definitions for previously undefined macros are provided; at level two,
3303 a multi-arch definition of all architecture dependent macros will be
3306 @item @value{GDBN}_TARGET_IS_HPPA
3307 @findex @value{GDBN}_TARGET_IS_HPPA
3308 This determines whether horrible kludge code in @file{dbxread.c} and
3309 @file{partial-stab.h} is used to mangle multiple-symbol-table files from
3310 HPPA's. This should all be ripped out, and a scheme like @file{elfread.c}
3313 @item GET_LONGJMP_TARGET
3314 @findex GET_LONGJMP_TARGET
3315 For most machines, this is a target-dependent parameter. On the
3316 DECstation and the Iris, this is a native-dependent parameter, since
3317 the header file @file{setjmp.h} is needed to define it.
3319 This macro determines the target PC address that @code{longjmp} will jump to,
3320 assuming that we have just stopped at a @code{longjmp} breakpoint. It takes a
3321 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3322 pointer. It examines the current state of the machine as needed.
3324 @item DEPRECATED_GET_SAVED_REGISTER
3325 @findex DEPRECATED_GET_SAVED_REGISTER
3326 Define this if you need to supply your own definition for the function
3327 @code{DEPRECATED_GET_SAVED_REGISTER}.
3329 @item DEPRECATED_IBM6000_TARGET
3330 @findex DEPRECATED_IBM6000_TARGET
3331 Shows that we are configured for an IBM RS/6000 system. This
3332 conditional should be eliminated (FIXME) and replaced by
3333 feature-specific macros. It was introduced in a haste and we are
3334 repenting at leisure.
3336 @item I386_USE_GENERIC_WATCHPOINTS
3337 An x86-based target can define this to use the generic x86 watchpoint
3338 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3340 @item SYMBOLS_CAN_START_WITH_DOLLAR
3341 @findex SYMBOLS_CAN_START_WITH_DOLLAR
3342 Some systems have routines whose names start with @samp{$}. Giving this
3343 macro a non-zero value tells @value{GDBN}'s expression parser to check for such
3344 routines when parsing tokens that begin with @samp{$}.
3346 On HP-UX, certain system routines (millicode) have names beginning with
3347 @samp{$} or @samp{$$}. For example, @code{$$dyncall} is a millicode
3348 routine that handles inter-space procedure calls on PA-RISC.
3350 @item DEPRECATED_INIT_EXTRA_FRAME_INFO (@var{fromleaf}, @var{frame})
3351 @findex DEPRECATED_INIT_EXTRA_FRAME_INFO
3352 If additional information about the frame is required this should be
3353 stored in @code{frame->extra_info}. Space for @code{frame->extra_info}
3354 is allocated using @code{frame_extra_info_zalloc}.
3356 @item DEPRECATED_INIT_FRAME_PC (@var{fromleaf}, @var{prev})
3357 @findex DEPRECATED_INIT_FRAME_PC
3358 This is a C statement that sets the pc of the frame pointed to by
3359 @var{prev}. [By default...]
3361 @item INNER_THAN (@var{lhs}, @var{rhs})
3363 Returns non-zero if stack address @var{lhs} is inner than (nearer to the
3364 stack top) stack address @var{rhs}. Define this as @code{lhs < rhs} if
3365 the target's stack grows downward in memory, or @code{lhs > rsh} if the
3368 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{pc})
3369 @findex gdbarch_in_function_epilogue_p
3370 Returns non-zero if the given @var{pc} is in the epilogue of a function.
3371 The epilogue of a function is defined as the part of a function where
3372 the stack frame of the function already has been destroyed up to the
3373 final `return from function call' instruction.
3375 @item DEPRECATED_SIGTRAMP_START (@var{pc})
3376 @findex DEPRECATED_SIGTRAMP_START
3377 @itemx DEPRECATED_SIGTRAMP_END (@var{pc})
3378 @findex DEPRECATED_SIGTRAMP_END
3379 Define these to be the start and end address of the @code{sigtramp} for the
3380 given @var{pc}. On machines where the address is just a compile time
3381 constant, the macro expansion will typically just ignore the supplied
3384 @item IN_SOLIB_CALL_TRAMPOLINE (@var{pc}, @var{name})
3385 @findex IN_SOLIB_CALL_TRAMPOLINE
3386 Define this to evaluate to nonzero if the program is stopped in the
3387 trampoline that connects to a shared library.
3389 @item IN_SOLIB_RETURN_TRAMPOLINE (@var{pc}, @var{name})
3390 @findex IN_SOLIB_RETURN_TRAMPOLINE
3391 Define this to evaluate to nonzero if the program is stopped in the
3392 trampoline that returns from a shared library.
3394 @item IN_SOLIB_DYNSYM_RESOLVE_CODE (@var{pc})
3395 @findex IN_SOLIB_DYNSYM_RESOLVE_CODE
3396 Define this to evaluate to nonzero if the program is stopped in the
3399 @item SKIP_SOLIB_RESOLVER (@var{pc})
3400 @findex SKIP_SOLIB_RESOLVER
3401 Define this to evaluate to the (nonzero) address at which execution
3402 should continue to get past the dynamic linker's symbol resolution
3403 function. A zero value indicates that it is not important or necessary
3404 to set a breakpoint to get through the dynamic linker and that single
3405 stepping will suffice.
3407 @item INTEGER_TO_ADDRESS (@var{type}, @var{buf})
3408 @findex INTEGER_TO_ADDRESS
3409 @cindex converting integers to addresses
3410 Define this when the architecture needs to handle non-pointer to address
3411 conversions specially. Converts that value to an address according to
3412 the current architectures conventions.
3414 @emph{Pragmatics: When the user copies a well defined expression from
3415 their source code and passes it, as a parameter, to @value{GDBN}'s
3416 @code{print} command, they should get the same value as would have been
3417 computed by the target program. Any deviation from this rule can cause
3418 major confusion and annoyance, and needs to be justified carefully. In
3419 other words, @value{GDBN} doesn't really have the freedom to do these
3420 conversions in clever and useful ways. It has, however, been pointed
3421 out that users aren't complaining about how @value{GDBN} casts integers
3422 to pointers; they are complaining that they can't take an address from a
3423 disassembly listing and give it to @code{x/i}. Adding an architecture
3424 method like @code{INTEGER_TO_ADDRESS} certainly makes it possible for
3425 @value{GDBN} to ``get it right'' in all circumstances.}
3427 @xref{Target Architecture Definition, , Pointers Are Not Always
3430 @item NO_HIF_SUPPORT
3431 @findex NO_HIF_SUPPORT
3432 (Specific to the a29k.)
3434 @item POINTER_TO_ADDRESS (@var{type}, @var{buf})
3435 @findex POINTER_TO_ADDRESS
3436 Assume that @var{buf} holds a pointer of type @var{type}, in the
3437 appropriate format for the current architecture. Return the byte
3438 address the pointer refers to.
3439 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3441 @item REGISTER_CONVERTIBLE (@var{reg})
3442 @findex REGISTER_CONVERTIBLE
3443 Return non-zero if @var{reg} uses different raw and virtual formats.
3444 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3446 @item REGISTER_TO_VALUE(@var{regnum}, @var{type}, @var{from}, @var{to})
3447 @findex REGISTER_TO_VALUE
3448 Convert the raw contents of register @var{regnum} into a value of type
3450 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3452 @item DEPRECATED_REGISTER_RAW_SIZE (@var{reg})
3453 @findex DEPRECATED_REGISTER_RAW_SIZE
3454 Return the raw size of @var{reg}; defaults to the size of the register's
3456 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3458 @item register_reggroup_p (@var{gdbarch}, @var{regnum}, @var{reggroup})
3459 @findex register_reggroup_p
3460 @cindex register groups
3461 Return non-zero if register @var{regnum} is a member of the register
3462 group @var{reggroup}.
3464 By default, registers are grouped as follows:
3467 @item float_reggroup
3468 Any register with a valid name and a floating-point type.
3469 @item vector_reggroup
3470 Any register with a valid name and a vector type.
3471 @item general_reggroup
3472 Any register with a valid name and a type other than vector or
3473 floating-point. @samp{float_reggroup}.
3475 @itemx restore_reggroup
3477 Any register with a valid name.
3480 @item DEPRECATED_REGISTER_VIRTUAL_SIZE (@var{reg})
3481 @findex DEPRECATED_REGISTER_VIRTUAL_SIZE
3482 Return the virtual size of @var{reg}; defaults to the size of the
3483 register's virtual type.
3484 Return the virtual size of @var{reg}.
3485 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3487 @item DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})
3488 @findex REGISTER_VIRTUAL_TYPE
3489 Return the virtual type of @var{reg}.
3490 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3492 @item struct type *register_type (@var{gdbarch}, @var{reg})
3493 @findex register_type
3494 If defined, return the type of register @var{reg}. This function
3495 superseeds @code{DEPRECATED_REGISTER_VIRTUAL_TYPE}. @xref{Target Architecture
3496 Definition, , Raw and Virtual Register Representations}.
3498 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
3499 @findex REGISTER_CONVERT_TO_VIRTUAL
3500 Convert the value of register @var{reg} from its raw form to its virtual
3502 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3504 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
3505 @findex REGISTER_CONVERT_TO_RAW
3506 Convert the value of register @var{reg} from its virtual form to its raw
3508 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3510 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
3511 @findex regset_from_core_section
3512 Return the appropriate register set for a core file section with name
3513 @var{sect_name} and size @var{sect_size}.
3515 @item SOFTWARE_SINGLE_STEP_P()
3516 @findex SOFTWARE_SINGLE_STEP_P
3517 Define this as 1 if the target does not have a hardware single-step
3518 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
3520 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breapoints_p})
3521 @findex SOFTWARE_SINGLE_STEP
3522 A function that inserts or removes (depending on
3523 @var{insert_breapoints_p}) breakpoints at each possible destinations of
3524 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
3527 @item SOFUN_ADDRESS_MAYBE_MISSING
3528 @findex SOFUN_ADDRESS_MAYBE_MISSING
3529 Somebody clever observed that, the more actual addresses you have in the
3530 debug information, the more time the linker has to spend relocating
3531 them. So whenever there's some other way the debugger could find the
3532 address it needs, you should omit it from the debug info, to make
3535 @code{SOFUN_ADDRESS_MAYBE_MISSING} indicates that a particular set of
3536 hacks of this sort are in use, affecting @code{N_SO} and @code{N_FUN}
3537 entries in stabs-format debugging information. @code{N_SO} stabs mark
3538 the beginning and ending addresses of compilation units in the text
3539 segment. @code{N_FUN} stabs mark the starts and ends of functions.
3541 @code{SOFUN_ADDRESS_MAYBE_MISSING} means two things:
3545 @code{N_FUN} stabs have an address of zero. Instead, you should find the
3546 addresses where the function starts by taking the function name from
3547 the stab, and then looking that up in the minsyms (the
3548 linker/assembler symbol table). In other words, the stab has the
3549 name, and the linker/assembler symbol table is the only place that carries
3553 @code{N_SO} stabs have an address of zero, too. You just look at the
3554 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab,
3555 and guess the starting and ending addresses of the compilation unit from
3559 @item PC_LOAD_SEGMENT
3560 @findex PC_LOAD_SEGMENT
3561 If defined, print information about the load segment for the program
3562 counter. (Defined only for the RS/6000.)
3566 If the program counter is kept in a register, then define this macro to
3567 be the number (greater than or equal to zero) of that register.
3569 This should only need to be defined if @code{TARGET_READ_PC} and
3570 @code{TARGET_WRITE_PC} are not defined.
3573 @findex PARM_BOUNDARY
3574 If non-zero, round arguments to a boundary of this many bits before
3575 pushing them on the stack.
3577 @item stabs_argument_has_addr (@var{gdbarch}, @var{type})
3578 @findex stabs_argument_has_addr
3579 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3580 @anchor{stabs_argument_has_addr} Define this to return nonzero if a
3581 function argument of type @var{type} is passed by reference instead of
3584 This method replaces @code{DEPRECATED_REG_STRUCT_HAS_ADDR}
3585 (@pxref{DEPRECATED_REG_STRUCT_HAS_ADDR}).
3587 @item PROCESS_LINENUMBER_HOOK
3588 @findex PROCESS_LINENUMBER_HOOK
3589 A hook defined for XCOFF reading.
3591 @item PROLOGUE_FIRSTLINE_OVERLAP
3592 @findex PROLOGUE_FIRSTLINE_OVERLAP
3593 (Only used in unsupported Convex configuration.)
3597 If defined, this is the number of the processor status register. (This
3598 definition is only used in generic code when parsing "$ps".)
3600 @item DEPRECATED_POP_FRAME
3601 @findex DEPRECATED_POP_FRAME
3603 If defined, used by @code{frame_pop} to remove a stack frame. This
3604 method has been superseeded by generic code.
3606 @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})
3607 @findex push_dummy_call
3608 @findex DEPRECATED_PUSH_ARGUMENTS.
3609 @anchor{push_dummy_call} Define this to push the dummy frame's call to
3610 the inferior function onto the stack. In addition to pushing
3611 @var{nargs}, the code should push @var{struct_addr} (when
3612 @var{struct_return}), and the return address (@var{bp_addr}).
3614 @var{function} is a pointer to a @code{struct value}; on architectures that use
3615 function descriptors, this contains the function descriptor value.
3617 Returns the updated top-of-stack pointer.
3619 This method replaces @code{DEPRECATED_PUSH_ARGUMENTS}.
3621 @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})
3622 @findex push_dummy_code
3623 @anchor{push_dummy_code} Given a stack based call dummy, push the
3624 instruction sequence (including space for a breakpoint) to which the
3625 called function should return.
3627 Set @var{bp_addr} to the address at which the breakpoint instruction
3628 should be inserted, @var{real_pc} to the resume address when starting
3629 the call sequence, and return the updated inner-most stack address.
3631 By default, the stack is grown sufficient to hold a frame-aligned
3632 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
3633 reserved for that breakpoint, and @var{real_pc} set to @var{funaddr}.
3635 This method replaces @code{CALL_DUMMY_LOCATION},
3636 @code{DEPRECATED_REGISTER_SIZE}.
3638 @item REGISTER_NAME(@var{i})
3639 @findex REGISTER_NAME
3640 Return the name of register @var{i} as a string. May return @code{NULL}
3641 or @code{NUL} to indicate that register @var{i} is not valid.
3643 @item DEPRECATED_REG_STRUCT_HAS_ADDR (@var{gcc_p}, @var{type})
3644 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3645 @anchor{DEPRECATED_REG_STRUCT_HAS_ADDR}Define this to return 1 if the
3646 given type will be passed by pointer rather than directly.
3648 This method has been replaced by @code{stabs_argument_has_addr}
3649 (@pxref{stabs_argument_has_addr}).
3651 @item SAVE_DUMMY_FRAME_TOS (@var{sp})
3652 @findex SAVE_DUMMY_FRAME_TOS
3653 @anchor{SAVE_DUMMY_FRAME_TOS} Used in @samp{call_function_by_hand} to
3654 notify the target dependent code of the top-of-stack value that will be
3655 passed to the the inferior code. This is the value of the @code{SP}
3656 after both the dummy frame and space for parameters/results have been
3657 allocated on the stack. @xref{unwind_dummy_id}.
3659 @item SDB_REG_TO_REGNUM
3660 @findex SDB_REG_TO_REGNUM
3661 Define this to convert sdb register numbers into @value{GDBN} regnums. If not
3662 defined, no conversion will be done.
3664 @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})
3665 @findex gdbarch_return_value
3666 @anchor{gdbarch_return_value} Given a function with a return-value of
3667 type @var{rettype}, return which return-value convention that function
3670 @value{GDBN} currently recognizes two function return-value conventions:
3671 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
3672 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
3673 value is found in memory and the address of that memory location is
3674 passed in as the function's first parameter.
3676 If the register convention is being used, and @var{writebuf} is
3677 non-@code{NULL}, also copy the return-value in @var{writebuf} into
3680 If the register convention is being used, and @var{readbuf} is
3681 non-@code{NULL}, also copy the return value from @var{regcache} into
3682 @var{readbuf} (@var{regcache} contains a copy of the registers from the
3683 just returned function).
3685 @xref{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}, for a description of how
3686 return-values that use the struct convention are handled.
3688 @emph{Maintainer note: This method replaces separate predicate, extract,
3689 store methods. By having only one method, the logic needed to determine
3690 the return-value convention need only be implemented in one place. If
3691 @value{GDBN} were written in an @sc{oo} language, this method would
3692 instead return an object that knew how to perform the register
3693 return-value extract and store.}
3695 @emph{Maintainer note: This method does not take a @var{gcc_p}
3696 parameter, and such a parameter should not be added. If an architecture
3697 that requires per-compiler or per-function information be identified,
3698 then the replacement of @var{rettype} with @code{struct value}
3699 @var{function} should be persued.}
3701 @emph{Maintainer note: The @var{regcache} parameter limits this methods
3702 to the inner most frame. While replacing @var{regcache} with a
3703 @code{struct frame_info} @var{frame} parameter would remove that
3704 limitation there has yet to be a demonstrated need for such a change.}
3706 @item SKIP_PERMANENT_BREAKPOINT
3707 @findex SKIP_PERMANENT_BREAKPOINT
3708 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
3709 steps over a breakpoint by removing it, stepping one instruction, and
3710 re-inserting the breakpoint. However, permanent breakpoints are
3711 hardwired into the inferior, and can't be removed, so this strategy
3712 doesn't work. Calling @code{SKIP_PERMANENT_BREAKPOINT} adjusts the processor's
3713 state so that execution will resume just after the breakpoint. This
3714 macro does the right thing even when the breakpoint is in the delay slot
3715 of a branch or jump.
3717 @item SKIP_PROLOGUE (@var{pc})
3718 @findex SKIP_PROLOGUE
3719 A C expression that returns the address of the ``real'' code beyond the
3720 function entry prologue found at @var{pc}.
3722 @item SKIP_TRAMPOLINE_CODE (@var{pc})
3723 @findex SKIP_TRAMPOLINE_CODE
3724 If the target machine has trampoline code that sits between callers and
3725 the functions being called, then define this macro to return a new PC
3726 that is at the start of the real function.
3730 If the stack-pointer is kept in a register, then define this macro to be
3731 the number (greater than or equal to zero) of that register, or -1 if
3732 there is no such register.
3734 @item STAB_REG_TO_REGNUM
3735 @findex STAB_REG_TO_REGNUM
3736 Define this to convert stab register numbers (as gotten from `r'
3737 declarations) into @value{GDBN} regnums. If not defined, no conversion will be
3740 @item DEPRECATED_STACK_ALIGN (@var{addr})
3741 @anchor{DEPRECATED_STACK_ALIGN}
3742 @findex DEPRECATED_STACK_ALIGN
3743 Define this to increase @var{addr} so that it meets the alignment
3744 requirements for the processor's stack.
3746 Unlike @ref{frame_align}, this function always adjusts @var{addr}
3749 By default, no stack alignment is performed.
3751 @item STEP_SKIPS_DELAY (@var{addr})
3752 @findex STEP_SKIPS_DELAY
3753 Define this to return true if the address is of an instruction with a
3754 delay slot. If a breakpoint has been placed in the instruction's delay
3755 slot, @value{GDBN} will single-step over that instruction before resuming
3756 normally. Currently only defined for the Mips.
3758 @item STORE_RETURN_VALUE (@var{type}, @var{regcache}, @var{valbuf})
3759 @findex STORE_RETURN_VALUE
3760 A C expression that writes the function return value, found in
3761 @var{valbuf}, into the @var{regcache}. @var{type} is the type of the
3762 value that is to be returned.
3764 This method has been deprecated in favour of @code{gdbarch_return_value}
3765 (@pxref{gdbarch_return_value}).
3767 @item SYMBOL_RELOADING_DEFAULT
3768 @findex SYMBOL_RELOADING_DEFAULT
3769 The default value of the ``symbol-reloading'' variable. (Never defined in
3772 @item TARGET_CHAR_BIT
3773 @findex TARGET_CHAR_BIT
3774 Number of bits in a char; defaults to 8.
3776 @item TARGET_CHAR_SIGNED
3777 @findex TARGET_CHAR_SIGNED
3778 Non-zero if @code{char} is normally signed on this architecture; zero if
3779 it should be unsigned.
3781 The ISO C standard requires the compiler to treat @code{char} as
3782 equivalent to either @code{signed char} or @code{unsigned char}; any
3783 character in the standard execution set is supposed to be positive.
3784 Most compilers treat @code{char} as signed, but @code{char} is unsigned
3785 on the IBM S/390, RS6000, and PowerPC targets.
3787 @item TARGET_COMPLEX_BIT
3788 @findex TARGET_COMPLEX_BIT
3789 Number of bits in a complex number; defaults to @code{2 * TARGET_FLOAT_BIT}.
3791 At present this macro is not used.
3793 @item TARGET_DOUBLE_BIT
3794 @findex TARGET_DOUBLE_BIT
3795 Number of bits in a double float; defaults to @code{8 * TARGET_CHAR_BIT}.
3797 @item TARGET_DOUBLE_COMPLEX_BIT
3798 @findex TARGET_DOUBLE_COMPLEX_BIT
3799 Number of bits in a double complex; defaults to @code{2 * TARGET_DOUBLE_BIT}.
3801 At present this macro is not used.
3803 @item TARGET_FLOAT_BIT
3804 @findex TARGET_FLOAT_BIT
3805 Number of bits in a float; defaults to @code{4 * TARGET_CHAR_BIT}.
3807 @item TARGET_INT_BIT
3808 @findex TARGET_INT_BIT
3809 Number of bits in an integer; defaults to @code{4 * TARGET_CHAR_BIT}.
3811 @item TARGET_LONG_BIT
3812 @findex TARGET_LONG_BIT
3813 Number of bits in a long integer; defaults to @code{4 * TARGET_CHAR_BIT}.
3815 @item TARGET_LONG_DOUBLE_BIT
3816 @findex TARGET_LONG_DOUBLE_BIT
3817 Number of bits in a long double float;
3818 defaults to @code{2 * TARGET_DOUBLE_BIT}.
3820 @item TARGET_LONG_LONG_BIT
3821 @findex TARGET_LONG_LONG_BIT
3822 Number of bits in a long long integer; defaults to @code{2 * TARGET_LONG_BIT}.
3824 @item TARGET_PTR_BIT
3825 @findex TARGET_PTR_BIT
3826 Number of bits in a pointer; defaults to @code{TARGET_INT_BIT}.
3828 @item TARGET_SHORT_BIT
3829 @findex TARGET_SHORT_BIT
3830 Number of bits in a short integer; defaults to @code{2 * TARGET_CHAR_BIT}.
3832 @item TARGET_READ_PC
3833 @findex TARGET_READ_PC
3834 @itemx TARGET_WRITE_PC (@var{val}, @var{pid})
3835 @findex TARGET_WRITE_PC
3836 @anchor{TARGET_WRITE_PC}
3837 @itemx TARGET_READ_SP
3838 @findex TARGET_READ_SP
3839 @itemx TARGET_READ_FP
3840 @findex TARGET_READ_FP
3845 @anchor{TARGET_READ_SP} These change the behavior of @code{read_pc},
3846 @code{write_pc}, and @code{read_sp}. For most targets, these may be
3847 left undefined. @value{GDBN} will call the read and write register
3848 functions with the relevant @code{_REGNUM} argument.
3850 These macros are useful when a target keeps one of these registers in a
3851 hard to get at place; for example, part in a segment register and part
3852 in an ordinary register.
3854 @xref{unwind_sp}, which replaces @code{TARGET_READ_SP}.
3856 @item TARGET_VIRTUAL_FRAME_POINTER(@var{pc}, @var{regp}, @var{offsetp})
3857 @findex TARGET_VIRTUAL_FRAME_POINTER
3858 Returns a @code{(register, offset)} pair representing the virtual frame
3859 pointer in use at the code address @var{pc}. If virtual frame pointers
3860 are not used, a default definition simply returns
3861 @code{DEPRECATED_FP_REGNUM}, with an offset of zero.
3863 @item TARGET_HAS_HARDWARE_WATCHPOINTS
3864 If non-zero, the target has support for hardware-assisted
3865 watchpoints. @xref{Algorithms, watchpoints}, for more details and
3866 other related macros.
3868 @item TARGET_PRINT_INSN (@var{addr}, @var{info})
3869 @findex TARGET_PRINT_INSN
3870 This is the function used by @value{GDBN} to print an assembly
3871 instruction. It prints the instruction at address @var{addr} in
3872 debugged memory and returns the length of the instruction, in bytes. If
3873 a target doesn't define its own printing routine, it defaults to an
3874 accessor function for the global pointer
3875 @code{deprecated_tm_print_insn}. This usually points to a function in
3876 the @code{opcodes} library (@pxref{Support Libraries, ,Opcodes}).
3877 @var{info} is a structure (of type @code{disassemble_info}) defined in
3878 @file{include/dis-asm.h} used to pass information to the instruction
3881 @item struct frame_id unwind_dummy_id (struct frame_info *@var{frame})
3882 @findex unwind_dummy_id
3883 @anchor{unwind_dummy_id} Given @var{frame} return a @code{struct
3884 frame_id} that uniquely identifies an inferior function call's dummy
3885 frame. The value returned must match the dummy frame stack value
3886 previously saved using @code{SAVE_DUMMY_FRAME_TOS}.
3887 @xref{SAVE_DUMMY_FRAME_TOS}.
3889 @item DEPRECATED_USE_STRUCT_CONVENTION (@var{gcc_p}, @var{type})
3890 @findex DEPRECATED_USE_STRUCT_CONVENTION
3891 If defined, this must be an expression that is nonzero if a value of the
3892 given @var{type} being returned from a function must have space
3893 allocated for it on the stack. @var{gcc_p} is true if the function
3894 being considered is known to have been compiled by GCC; this is helpful
3895 for systems where GCC is known to use different calling convention than
3898 This method has been deprecated in favour of @code{gdbarch_return_value}
3899 (@pxref{gdbarch_return_value}).
3901 @item VALUE_TO_REGISTER(@var{type}, @var{regnum}, @var{from}, @var{to})
3902 @findex VALUE_TO_REGISTER
3903 Convert a value of type @var{type} into the raw contents of register
3905 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3907 @item VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
3908 @findex VARIABLES_INSIDE_BLOCK
3909 For dbx-style debugging information, if the compiler puts variable
3910 declarations inside LBRAC/RBRAC blocks, this should be defined to be
3911 nonzero. @var{desc} is the value of @code{n_desc} from the
3912 @code{N_RBRAC} symbol, and @var{gcc_p} is true if @value{GDBN} has noticed the
3913 presence of either the @code{GCC_COMPILED_SYMBOL} or the
3914 @code{GCC2_COMPILED_SYMBOL}. By default, this is 0.
3916 @item OS9K_VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
3917 @findex OS9K_VARIABLES_INSIDE_BLOCK
3918 Similarly, for OS/9000. Defaults to 1.
3921 Motorola M68K target conditionals.
3925 Define this to be the 4-bit location of the breakpoint trap vector. If
3926 not defined, it will default to @code{0xf}.
3928 @item REMOTE_BPT_VECTOR
3929 Defaults to @code{1}.
3931 @item NAME_OF_MALLOC
3932 @findex NAME_OF_MALLOC
3933 A string containing the name of the function to call in order to
3934 allocate some memory in the inferior. The default value is "malloc".
3938 @section Adding a New Target
3940 @cindex adding a target
3941 The following files add a target to @value{GDBN}:
3945 @item gdb/config/@var{arch}/@var{ttt}.mt
3946 Contains a Makefile fragment specific to this target. Specifies what
3947 object files are needed for target @var{ttt}, by defining
3948 @samp{TDEPFILES=@dots{}} and @samp{TDEPLIBS=@dots{}}. Also specifies
3949 the header file which describes @var{ttt}, by defining @samp{TM_FILE=
3952 You can also define @samp{TM_CFLAGS}, @samp{TM_CLIBS}, @samp{TM_CDEPS},
3953 but these are now deprecated, replaced by autoconf, and may go away in
3954 future versions of @value{GDBN}.
3956 @item gdb/@var{ttt}-tdep.c
3957 Contains any miscellaneous code required for this target machine. On
3958 some machines it doesn't exist at all. Sometimes the macros in
3959 @file{tm-@var{ttt}.h} become very complicated, so they are implemented
3960 as functions here instead, and the macro is simply defined to call the
3961 function. This is vastly preferable, since it is easier to understand
3964 @item gdb/@var{arch}-tdep.c
3965 @itemx gdb/@var{arch}-tdep.h
3966 This often exists to describe the basic layout of the target machine's
3967 processor chip (registers, stack, etc.). If used, it is included by
3968 @file{@var{ttt}-tdep.h}. It can be shared among many targets that use
3971 @item gdb/config/@var{arch}/tm-@var{ttt}.h
3972 (@file{tm.h} is a link to this file, created by @code{configure}). Contains
3973 macro definitions about the target machine's registers, stack frame
3974 format and instructions.
3976 New targets do not need this file and should not create it.
3978 @item gdb/config/@var{arch}/tm-@var{arch}.h
3979 This often exists to describe the basic layout of the target machine's
3980 processor chip (registers, stack, etc.). If used, it is included by
3981 @file{tm-@var{ttt}.h}. It can be shared among many targets that use the
3984 New targets do not need this file and should not create it.
3988 If you are adding a new operating system for an existing CPU chip, add a
3989 @file{config/tm-@var{os}.h} file that describes the operating system
3990 facilities that are unusual (extra symbol table info; the breakpoint
3991 instruction needed; etc.). Then write a @file{@var{arch}/tm-@var{os}.h}
3992 that just @code{#include}s @file{tm-@var{arch}.h} and
3993 @file{config/tm-@var{os}.h}.
3996 @section Converting an existing Target Architecture to Multi-arch
3997 @cindex converting targets to multi-arch
3999 This section describes the current accepted best practice for converting
4000 an existing target architecture to the multi-arch framework.
4002 The process consists of generating, testing, posting and committing a
4003 sequence of patches. Each patch must contain a single change, for
4009 Directly convert a group of functions into macros (the conversion does
4010 not change the behavior of any of the functions).
4013 Replace a non-multi-arch with a multi-arch mechanism (e.g.,
4017 Enable multi-arch level one.
4020 Delete one or more files.
4025 There isn't a size limit on a patch, however, a developer is strongly
4026 encouraged to keep the patch size down.
4028 Since each patch is well defined, and since each change has been tested
4029 and shows no regressions, the patches are considered @emph{fairly}
4030 obvious. Such patches, when submitted by developers listed in the
4031 @file{MAINTAINERS} file, do not need approval. Occasional steps in the
4032 process may be more complicated and less clear. The developer is
4033 expected to use their judgment and is encouraged to seek advice as
4036 @subsection Preparation
4038 The first step is to establish control. Build (with @option{-Werror}
4039 enabled) and test the target so that there is a baseline against which
4040 the debugger can be compared.
4042 At no stage can the test results regress or @value{GDBN} stop compiling
4043 with @option{-Werror}.
4045 @subsection Add the multi-arch initialization code
4047 The objective of this step is to establish the basic multi-arch
4048 framework. It involves
4053 The addition of a @code{@var{arch}_gdbarch_init} function@footnote{The
4054 above is from the original example and uses K&R C. @value{GDBN}
4055 has since converted to ISO C but lets ignore that.} that creates
4058 static struct gdbarch *
4059 d10v_gdbarch_init (info, arches)
4060 struct gdbarch_info info;
4061 struct gdbarch_list *arches;
4063 struct gdbarch *gdbarch;
4064 /* there is only one d10v architecture */
4066 return arches->gdbarch;
4067 gdbarch = gdbarch_alloc (&info, NULL);
4075 A per-architecture dump function to print any architecture specific
4079 mips_dump_tdep (struct gdbarch *current_gdbarch,
4080 struct ui_file *file)
4082 @dots{} code to print architecture specific info @dots{}
4087 A change to @code{_initialize_@var{arch}_tdep} to register this new
4091 _initialize_mips_tdep (void)
4093 gdbarch_register (bfd_arch_mips, mips_gdbarch_init,
4098 Add the macro @code{GDB_MULTI_ARCH}, defined as 0 (zero), to the file@*
4099 @file{config/@var{arch}/tm-@var{arch}.h}.
4103 @subsection Update multi-arch incompatible mechanisms
4105 Some mechanisms do not work with multi-arch. They include:
4108 @item FRAME_FIND_SAVED_REGS
4109 Replaced with @code{DEPRECATED_FRAME_INIT_SAVED_REGS}
4113 At this stage you could also consider converting the macros into
4116 @subsection Prepare for multi-arch level to one
4118 Temporally set @code{GDB_MULTI_ARCH} to @code{GDB_MULTI_ARCH_PARTIAL}
4119 and then build and start @value{GDBN} (the change should not be
4120 committed). @value{GDBN} may not build, and once built, it may die with
4121 an internal error listing the architecture methods that must be
4124 Fix any build problems (patch(es)).
4126 Convert all the architecture methods listed, which are only macros, into
4127 functions (patch(es)).
4129 Update @code{@var{arch}_gdbarch_init} to set all the missing
4130 architecture methods and wrap the corresponding macros in @code{#if
4131 !GDB_MULTI_ARCH} (patch(es)).
4133 @subsection Set multi-arch level one
4135 Change the value of @code{GDB_MULTI_ARCH} to GDB_MULTI_ARCH_PARTIAL (a
4138 Any problems with throwing ``the switch'' should have been fixed
4141 @subsection Convert remaining macros
4143 Suggest converting macros into functions (and setting the corresponding
4144 architecture method) in small batches.
4146 @subsection Set multi-arch level to two
4148 This should go smoothly.
4150 @subsection Delete the TM file
4152 The @file{tm-@var{arch}.h} can be deleted. @file{@var{arch}.mt} and
4153 @file{configure.in} updated.
4156 @node Target Vector Definition
4158 @chapter Target Vector Definition
4159 @cindex target vector
4161 The target vector defines the interface between @value{GDBN}'s
4162 abstract handling of target systems, and the nitty-gritty code that
4163 actually exercises control over a process or a serial port.
4164 @value{GDBN} includes some 30-40 different target vectors; however,
4165 each configuration of @value{GDBN} includes only a few of them.
4167 @section File Targets
4169 Both executables and core files have target vectors.
4171 @section Standard Protocol and Remote Stubs
4173 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code
4174 that runs in the target system. @value{GDBN} provides several sample
4175 @dfn{stubs} that can be integrated into target programs or operating
4176 systems for this purpose; they are named @file{*-stub.c}.
4178 The @value{GDBN} user's manual describes how to put such a stub into
4179 your target code. What follows is a discussion of integrating the
4180 SPARC stub into a complicated operating system (rather than a simple
4181 program), by Stu Grossman, the author of this stub.
4183 The trap handling code in the stub assumes the following upon entry to
4188 %l1 and %l2 contain pc and npc respectively at the time of the trap;
4194 you are in the correct trap window.
4197 As long as your trap handler can guarantee those conditions, then there
4198 is no reason why you shouldn't be able to ``share'' traps with the stub.
4199 The stub has no requirement that it be jumped to directly from the
4200 hardware trap vector. That is why it calls @code{exceptionHandler()},
4201 which is provided by the external environment. For instance, this could
4202 set up the hardware traps to actually execute code which calls the stub
4203 first, and then transfers to its own trap handler.
4205 For the most point, there probably won't be much of an issue with
4206 ``sharing'' traps, as the traps we use are usually not used by the kernel,
4207 and often indicate unrecoverable error conditions. Anyway, this is all
4208 controlled by a table, and is trivial to modify. The most important
4209 trap for us is for @code{ta 1}. Without that, we can't single step or
4210 do breakpoints. Everything else is unnecessary for the proper operation
4211 of the debugger/stub.
4213 From reading the stub, it's probably not obvious how breakpoints work.
4214 They are simply done by deposit/examine operations from @value{GDBN}.
4216 @section ROM Monitor Interface
4218 @section Custom Protocols
4220 @section Transport Layer
4222 @section Builtin Simulator
4225 @node Native Debugging
4227 @chapter Native Debugging
4228 @cindex native debugging
4230 Several files control @value{GDBN}'s configuration for native support:
4234 @item gdb/config/@var{arch}/@var{xyz}.mh
4235 Specifies Makefile fragments needed by a @emph{native} configuration on
4236 machine @var{xyz}. In particular, this lists the required
4237 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
4238 Also specifies the header file which describes native support on
4239 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
4240 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
4241 @samp{NAT_CDEPS}, etc.; see @file{Makefile.in}.
4243 @emph{Maintainer's note: The @file{.mh} suffix is because this file
4244 originally contained @file{Makefile} fragments for hosting @value{GDBN}
4245 on machine @var{xyz}. While the file is no longer used for this
4246 purpose, the @file{.mh} suffix remains. Perhaps someone will
4247 eventually rename these fragments so that they have a @file{.mn}
4250 @item gdb/config/@var{arch}/nm-@var{xyz}.h
4251 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
4252 macro definitions describing the native system environment, such as
4253 child process control and core file support.
4255 @item gdb/@var{xyz}-nat.c
4256 Contains any miscellaneous C code required for this native support of
4257 this machine. On some machines it doesn't exist at all.
4260 There are some ``generic'' versions of routines that can be used by
4261 various systems. These can be customized in various ways by macros
4262 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
4263 the @var{xyz} host, you can just include the generic file's name (with
4264 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
4266 Otherwise, if your machine needs custom support routines, you will need
4267 to write routines that perform the same functions as the generic file.
4268 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
4269 into @code{NATDEPFILES}.
4273 This contains the @emph{target_ops vector} that supports Unix child
4274 processes on systems which use ptrace and wait to control the child.
4277 This contains the @emph{target_ops vector} that supports Unix child
4278 processes on systems which use /proc to control the child.
4281 This does the low-level grunge that uses Unix system calls to do a ``fork
4282 and exec'' to start up a child process.
4285 This is the low level interface to inferior processes for systems using
4286 the Unix @code{ptrace} call in a vanilla way.
4289 @section Native core file Support
4290 @cindex native core files
4293 @findex fetch_core_registers
4294 @item core-aout.c::fetch_core_registers()
4295 Support for reading registers out of a core file. This routine calls
4296 @code{register_addr()}, see below. Now that BFD is used to read core
4297 files, virtually all machines should use @code{core-aout.c}, and should
4298 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
4299 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
4301 @item core-aout.c::register_addr()
4302 If your @code{nm-@var{xyz}.h} file defines the macro
4303 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
4304 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
4305 register number @code{regno}. @code{blockend} is the offset within the
4306 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
4307 @file{core-aout.c} will define the @code{register_addr()} function and
4308 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
4309 you are using the standard @code{fetch_core_registers()}, you will need
4310 to define your own version of @code{register_addr()}, put it into your
4311 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
4312 the @code{NATDEPFILES} list. If you have your own
4313 @code{fetch_core_registers()}, you may not need a separate
4314 @code{register_addr()}. Many custom @code{fetch_core_registers()}
4315 implementations simply locate the registers themselves.@refill
4318 When making @value{GDBN} run native on a new operating system, to make it
4319 possible to debug core files, you will need to either write specific
4320 code for parsing your OS's core files, or customize
4321 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
4322 machine uses to define the struct of registers that is accessible
4323 (possibly in the u-area) in a core file (rather than
4324 @file{machine/reg.h}), and an include file that defines whatever header
4325 exists on a core file (e.g. the u-area or a @code{struct core}). Then
4326 modify @code{trad_unix_core_file_p} to use these values to set up the
4327 section information for the data segment, stack segment, any other
4328 segments in the core file (perhaps shared library contents or control
4329 information), ``registers'' segment, and if there are two discontiguous
4330 sets of registers (e.g. integer and float), the ``reg2'' segment. This
4331 section information basically delimits areas in the core file in a
4332 standard way, which the section-reading routines in BFD know how to seek
4335 Then back in @value{GDBN}, you need a matching routine called
4336 @code{fetch_core_registers}. If you can use the generic one, it's in
4337 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
4338 It will be passed a char pointer to the entire ``registers'' segment,
4339 its length, and a zero; or a char pointer to the entire ``regs2''
4340 segment, its length, and a 2. The routine should suck out the supplied
4341 register values and install them into @value{GDBN}'s ``registers'' array.
4343 If your system uses @file{/proc} to control processes, and uses ELF
4344 format core files, then you may be able to use the same routines for
4345 reading the registers out of processes and out of core files.
4353 @section shared libraries
4355 @section Native Conditionals
4356 @cindex native conditionals
4358 When @value{GDBN} is configured and compiled, various macros are
4359 defined or left undefined, to control compilation when the host and
4360 target systems are the same. These macros should be defined (or left
4361 undefined) in @file{nm-@var{system}.h}.
4365 @item CHILD_PREPARE_TO_STORE
4366 @findex CHILD_PREPARE_TO_STORE
4367 If the machine stores all registers at once in the child process, then
4368 define this to ensure that all values are correct. This usually entails
4369 a read from the child.
4371 [Note that this is incorrectly defined in @file{xm-@var{system}.h} files
4374 @item FETCH_INFERIOR_REGISTERS
4375 @findex FETCH_INFERIOR_REGISTERS
4376 Define this if the native-dependent code will provide its own routines
4377 @code{fetch_inferior_registers} and @code{store_inferior_registers} in
4378 @file{@var{host}-nat.c}. If this symbol is @emph{not} defined, and
4379 @file{infptrace.c} is included in this configuration, the default
4380 routines in @file{infptrace.c} are used for these functions.
4384 This macro is normally defined to be the number of the first floating
4385 point register, if the machine has such registers. As such, it would
4386 appear only in target-specific code. However, @file{/proc} support uses this
4387 to decide whether floats are in use on this target.
4389 @item GET_LONGJMP_TARGET
4390 @findex GET_LONGJMP_TARGET
4391 For most machines, this is a target-dependent parameter. On the
4392 DECstation and the Iris, this is a native-dependent parameter, since
4393 @file{setjmp.h} is needed to define it.
4395 This macro determines the target PC address that @code{longjmp} will jump to,
4396 assuming that we have just stopped at a longjmp breakpoint. It takes a
4397 @code{CORE_ADDR *} as argument, and stores the target PC value through this
4398 pointer. It examines the current state of the machine as needed.
4400 @item I386_USE_GENERIC_WATCHPOINTS
4401 An x86-based machine can define this to use the generic x86 watchpoint
4402 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4405 @findex KERNEL_U_ADDR
4406 Define this to the address of the @code{u} structure (the ``user
4407 struct'', also known as the ``u-page'') in kernel virtual memory. @value{GDBN}
4408 needs to know this so that it can subtract this address from absolute
4409 addresses in the upage, that are obtained via ptrace or from core files.
4410 On systems that don't need this value, set it to zero.
4412 @item KERNEL_U_ADDR_HPUX
4413 @findex KERNEL_U_ADDR_HPUX
4414 Define this to cause @value{GDBN} to determine the address of @code{u} at
4415 runtime, by using HP-style @code{nlist} on the kernel's image in the
4418 @item ONE_PROCESS_WRITETEXT
4419 @findex ONE_PROCESS_WRITETEXT
4420 Define this to be able to, when a breakpoint insertion fails, warn the
4421 user that another process may be running with the same executable.
4424 @findex PROC_NAME_FMT
4425 Defines the format for the name of a @file{/proc} device. Should be
4426 defined in @file{nm.h} @emph{only} in order to override the default
4427 definition in @file{procfs.c}.
4429 @item PTRACE_ARG3_TYPE
4430 @findex PTRACE_ARG3_TYPE
4431 The type of the third argument to the @code{ptrace} system call, if it
4432 exists and is different from @code{int}.
4434 @item REGISTER_U_ADDR
4435 @findex REGISTER_U_ADDR
4436 Defines the offset of the registers in the ``u area''.
4438 @item SHELL_COMMAND_CONCAT
4439 @findex SHELL_COMMAND_CONCAT
4440 If defined, is a string to prefix on the shell command used to start the
4445 If defined, this is the name of the shell to use to run the inferior.
4446 Defaults to @code{"/bin/sh"}.
4448 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
4450 Define this to expand into an expression that will cause the symbols in
4451 @var{filename} to be added to @value{GDBN}'s symbol table. If
4452 @var{readsyms} is zero symbols are not read but any necessary low level
4453 processing for @var{filename} is still done.
4455 @item SOLIB_CREATE_INFERIOR_HOOK
4456 @findex SOLIB_CREATE_INFERIOR_HOOK
4457 Define this to expand into any shared-library-relocation code that you
4458 want to be run just after the child process has been forked.
4460 @item START_INFERIOR_TRAPS_EXPECTED
4461 @findex START_INFERIOR_TRAPS_EXPECTED
4462 When starting an inferior, @value{GDBN} normally expects to trap
4464 the shell execs, and once when the program itself execs. If the actual
4465 number of traps is something other than 2, then define this macro to
4466 expand into the number expected.
4470 This determines whether small routines in @file{*-tdep.c}, which
4471 translate register values between @value{GDBN}'s internal
4472 representation and the @file{/proc} representation, are compiled.
4475 @findex U_REGS_OFFSET
4476 This is the offset of the registers in the upage. It need only be
4477 defined if the generic ptrace register access routines in
4478 @file{infptrace.c} are being used (that is, @file{infptrace.c} is
4479 configured in, and @code{FETCH_INFERIOR_REGISTERS} is not defined). If
4480 the default value from @file{infptrace.c} is good enough, leave it
4483 The default value means that u.u_ar0 @emph{points to} the location of
4484 the registers. I'm guessing that @code{#define U_REGS_OFFSET 0} means
4485 that @code{u.u_ar0} @emph{is} the location of the registers.
4489 See @file{objfiles.c}.
4492 @findex DEBUG_PTRACE
4493 Define this to debug @code{ptrace} calls.
4497 @node Support Libraries
4499 @chapter Support Libraries
4504 BFD provides support for @value{GDBN} in several ways:
4507 @item identifying executable and core files
4508 BFD will identify a variety of file types, including a.out, coff, and
4509 several variants thereof, as well as several kinds of core files.
4511 @item access to sections of files
4512 BFD parses the file headers to determine the names, virtual addresses,
4513 sizes, and file locations of all the various named sections in files
4514 (such as the text section or the data section). @value{GDBN} simply
4515 calls BFD to read or write section @var{x} at byte offset @var{y} for
4518 @item specialized core file support
4519 BFD provides routines to determine the failing command name stored in a
4520 core file, the signal with which the program failed, and whether a core
4521 file matches (i.e.@: could be a core dump of) a particular executable
4524 @item locating the symbol information
4525 @value{GDBN} uses an internal interface of BFD to determine where to find the
4526 symbol information in an executable file or symbol-file. @value{GDBN} itself
4527 handles the reading of symbols, since BFD does not ``understand'' debug
4528 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
4533 @cindex opcodes library
4535 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
4536 library because it's also used in binutils, for @file{objdump}).
4543 @cindex @code{libiberty} library
4545 The @code{libiberty} library provides a set of functions and features
4546 that integrate and improve on functionality found in modern operating
4547 systems. Broadly speaking, such features can be divided into three
4548 groups: supplemental functions (functions that may be missing in some
4549 environments and operating systems), replacement functions (providing
4550 a uniform and easier to use interface for commonly used standard
4551 functions), and extensions (which provide additional functionality
4552 beyond standard functions).
4554 @value{GDBN} uses various features provided by the @code{libiberty}
4555 library, for instance the C@t{++} demangler, the @acronym{IEEE}
4556 floating format support functions, the input options parser
4557 @samp{getopt}, the @samp{obstack} extension, and other functions.
4559 @subsection @code{obstacks} in @value{GDBN}
4560 @cindex @code{obstacks}
4562 The obstack mechanism provides a convenient way to allocate and free
4563 chunks of memory. Each obstack is a pool of memory that is managed
4564 like a stack. Objects (of any nature, size and alignment) are
4565 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
4566 @code{libiberty}'s documenatation for a more detailed explanation of
4569 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
4570 object files. There is an obstack associated with each internal
4571 representation of an object file. Lots of things get allocated on
4572 these @code{obstacks}: dictionary entries, blocks, blockvectors,
4573 symbols, minimal symbols, types, vectors of fundamental types, class
4574 fields of types, object files section lists, object files section
4575 offets lists, line tables, symbol tables, partial symbol tables,
4576 string tables, symbol table private data, macros tables, debug
4577 information sections and entries, import and export lists (som),
4578 unwind information (hppa), dwarf2 location expressions data. Plus
4579 various strings such as directory names strings, debug format strings,
4582 An essential and convenient property of all data on @code{obstacks} is
4583 that memory for it gets allocated (with @code{obstack_alloc}) at
4584 various times during a debugging sesssion, but it is released all at
4585 once using the @code{obstack_free} function. The @code{obstack_free}
4586 function takes a pointer to where in the stack it must start the
4587 deletion from (much like the cleanup chains have a pointer to where to
4588 start the cleanups). Because of the stack like structure of the
4589 @code{obstacks}, this allows to free only a top portion of the
4590 obstack. There are a few instances in @value{GDBN} where such thing
4591 happens. Calls to @code{obstack_free} are done after some local data
4592 is allocated to the obstack. Only the local data is deleted from the
4593 obstack. Of course this assumes that nothing between the
4594 @code{obstack_alloc} and the @code{obstack_free} allocates anything
4595 else on the same obstack. For this reason it is best and safest to
4596 use temporary @code{obstacks}.
4598 Releasing the whole obstack is also not safe per se. It is safe only
4599 under the condition that we know the @code{obstacks} memory is no
4600 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
4601 when we get rid of the whole objfile(s), for instance upon reading a
4605 @cindex regular expressions library
4616 @item SIGN_EXTEND_CHAR
4618 @item SWITCH_ENUM_BUG
4633 This chapter covers topics that are lower-level than the major
4634 algorithms of @value{GDBN}.
4639 Cleanups are a structured way to deal with things that need to be done
4642 When your code does something (e.g., @code{xmalloc} some memory, or
4643 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
4644 the memory or @code{close} the file), it can make a cleanup. The
4645 cleanup will be done at some future point: when the command is finished
4646 and control returns to the top level; when an error occurs and the stack
4647 is unwound; or when your code decides it's time to explicitly perform
4648 cleanups. Alternatively you can elect to discard the cleanups you
4654 @item struct cleanup *@var{old_chain};
4655 Declare a variable which will hold a cleanup chain handle.
4657 @findex make_cleanup
4658 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
4659 Make a cleanup which will cause @var{function} to be called with
4660 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
4661 handle that can later be passed to @code{do_cleanups} or
4662 @code{discard_cleanups}. Unless you are going to call
4663 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
4664 from @code{make_cleanup}.
4667 @item do_cleanups (@var{old_chain});
4668 Do all cleanups added to the chain since the corresponding
4669 @code{make_cleanup} call was made.
4671 @findex discard_cleanups
4672 @item discard_cleanups (@var{old_chain});
4673 Same as @code{do_cleanups} except that it just removes the cleanups from
4674 the chain and does not call the specified functions.
4677 Cleanups are implemented as a chain. The handle returned by
4678 @code{make_cleanups} includes the cleanup passed to the call and any
4679 later cleanups appended to the chain (but not yet discarded or
4683 make_cleanup (a, 0);
4685 struct cleanup *old = make_cleanup (b, 0);
4693 will call @code{c()} and @code{b()} but will not call @code{a()}. The
4694 cleanup that calls @code{a()} will remain in the cleanup chain, and will
4695 be done later unless otherwise discarded.@refill
4697 Your function should explicitly do or discard the cleanups it creates.
4698 Failing to do this leads to non-deterministic behavior since the caller
4699 will arbitrarily do or discard your functions cleanups. This need leads
4700 to two common cleanup styles.
4702 The first style is try/finally. Before it exits, your code-block calls
4703 @code{do_cleanups} with the old cleanup chain and thus ensures that your
4704 code-block's cleanups are always performed. For instance, the following
4705 code-segment avoids a memory leak problem (even when @code{error} is
4706 called and a forced stack unwind occurs) by ensuring that the
4707 @code{xfree} will always be called:
4710 struct cleanup *old = make_cleanup (null_cleanup, 0);
4711 data = xmalloc (sizeof blah);
4712 make_cleanup (xfree, data);
4717 The second style is try/except. Before it exits, your code-block calls
4718 @code{discard_cleanups} with the old cleanup chain and thus ensures that
4719 any created cleanups are not performed. For instance, the following
4720 code segment, ensures that the file will be closed but only if there is
4724 FILE *file = fopen ("afile", "r");
4725 struct cleanup *old = make_cleanup (close_file, file);
4727 discard_cleanups (old);
4731 Some functions, e.g. @code{fputs_filtered()} or @code{error()}, specify
4732 that they ``should not be called when cleanups are not in place''. This
4733 means that any actions you need to reverse in the case of an error or
4734 interruption must be on the cleanup chain before you call these
4735 functions, since they might never return to your code (they
4736 @samp{longjmp} instead).
4738 @section Per-architecture module data
4739 @cindex per-architecture module data
4740 @cindex multi-arch data
4741 @cindex data-pointer, per-architecture/per-module
4743 The multi-arch framework includes a mechanism for adding module
4744 specific per-architecture data-pointers to the @code{struct gdbarch}
4745 architecture object.
4747 A module registers one or more per-architecture data-pointers using:
4749 @deftypefun struct gdbarch_data *gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
4750 @var{pre_init} is used to, on-demand, allocate an initial value for a
4751 per-architecture data-pointer using the architecture's obstack (passed
4752 in as a parameter). Since @var{pre_init} can be called during
4753 architecture creation, it is not parameterized with the architecture.
4754 and must not call modules that use per-architecture data.
4757 @deftypefun struct gdbarch_data *gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
4758 @var{post_init} is used to obtain an initial value for a
4759 per-architecture data-pointer @emph{after}. Since @var{post_init} is
4760 always called after architecture creation, it both receives the fully
4761 initialized architecture and is free to call modules that use
4762 per-architecture data (care needs to be taken to ensure that those
4763 other modules do not try to call back to this module as that will
4764 create in cycles in the initialization call graph).
4767 These functions return a @code{struct gdbarch_data} that is used to
4768 identify the per-architecture data-pointer added for that module.
4770 The per-architecture data-pointer is accessed using the function:
4772 @deftypefun void *gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
4773 Given the architecture @var{arch} and module data handle
4774 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
4775 or @code{gdbarch_data_register_post_init}), this function returns the
4776 current value of the per-architecture data-pointer. If the data
4777 pointer is @code{NULL}, it is first initialized by calling the
4778 corresponding @var{pre_init} or @var{post_init} method.
4781 The examples below assume the following definitions:
4784 struct nozel @{ int total; @};
4785 static struct gdbarch_data *nozel_handle;
4788 A module can extend the architecture vector, adding additional
4789 per-architecture data, using the @var{pre_init} method. The module's
4790 per-architecture data is then initialized during architecture
4793 In the below, the module's per-architecture @emph{nozel} is added. An
4794 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
4795 from @code{gdbarch_init}.
4799 nozel_pre_init (struct obstack *obstack)
4801 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
4808 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
4810 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
4811 data->total = nozel;
4815 A module can on-demand create architecture dependant data structures
4816 using @code{post_init}.
4818 In the below, the nozel's total is computed on-demand by
4819 @code{nozel_post_init} using information obtained from the
4824 nozel_post_init (struct gdbarch *gdbarch)
4826 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
4827 nozel->total = gdbarch@dots{} (gdbarch);
4834 nozel_total (struct gdbarch *gdbarch)
4836 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
4841 @section Wrapping Output Lines
4842 @cindex line wrap in output
4845 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
4846 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
4847 added in places that would be good breaking points. The utility
4848 routines will take care of actually wrapping if the line width is
4851 The argument to @code{wrap_here} is an indentation string which is
4852 printed @emph{only} if the line breaks there. This argument is saved
4853 away and used later. It must remain valid until the next call to
4854 @code{wrap_here} or until a newline has been printed through the
4855 @code{*_filtered} functions. Don't pass in a local variable and then
4858 It is usually best to call @code{wrap_here} after printing a comma or
4859 space. If you call it before printing a space, make sure that your
4860 indentation properly accounts for the leading space that will print if
4861 the line wraps there.
4863 Any function or set of functions that produce filtered output must
4864 finish by printing a newline, to flush the wrap buffer, before switching
4865 to unfiltered (@code{printf}) output. Symbol reading routines that
4866 print warnings are a good example.
4868 @section @value{GDBN} Coding Standards
4869 @cindex coding standards
4871 @value{GDBN} follows the GNU coding standards, as described in
4872 @file{etc/standards.texi}. This file is also available for anonymous
4873 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
4874 of the standard; in general, when the GNU standard recommends a practice
4875 but does not require it, @value{GDBN} requires it.
4877 @value{GDBN} follows an additional set of coding standards specific to
4878 @value{GDBN}, as described in the following sections.
4883 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
4886 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
4889 @subsection Memory Management
4891 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
4892 @code{calloc}, @code{free} and @code{asprintf}.
4894 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
4895 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
4896 these functions do not return when the memory pool is empty. Instead,
4897 they unwind the stack using cleanups. These functions return
4898 @code{NULL} when requested to allocate a chunk of memory of size zero.
4900 @emph{Pragmatics: By using these functions, the need to check every
4901 memory allocation is removed. These functions provide portable
4904 @value{GDBN} does not use the function @code{free}.
4906 @value{GDBN} uses the function @code{xfree} to return memory to the
4907 memory pool. Consistent with ISO-C, this function ignores a request to
4908 free a @code{NULL} pointer.
4910 @emph{Pragmatics: On some systems @code{free} fails when passed a
4911 @code{NULL} pointer.}
4913 @value{GDBN} can use the non-portable function @code{alloca} for the
4914 allocation of small temporary values (such as strings).
4916 @emph{Pragmatics: This function is very non-portable. Some systems
4917 restrict the memory being allocated to no more than a few kilobytes.}
4919 @value{GDBN} uses the string function @code{xstrdup} and the print
4920 function @code{xstrprintf}.
4922 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
4923 functions such as @code{sprintf} are very prone to buffer overflow
4927 @subsection Compiler Warnings
4928 @cindex compiler warnings
4930 With few exceptions, developers should include the configuration option
4931 @samp{--enable-gdb-build-warnings=,-Werror} when building @value{GDBN}.
4932 The exceptions are listed in the file @file{gdb/MAINTAINERS}.
4934 This option causes @value{GDBN} (when built using GCC) to be compiled
4935 with a carefully selected list of compiler warning flags. Any warnings
4936 from those flags being treated as errors.
4938 The current list of warning flags includes:
4942 Since @value{GDBN} coding standard requires all functions to be declared
4943 using a prototype, the flag has the side effect of ensuring that
4944 prototyped functions are always visible with out resorting to
4945 @samp{-Wstrict-prototypes}.
4948 Such code often appears to work except on instruction set architectures
4949 that use register windows.
4956 @itemx -Wformat-nonliteral
4957 Since @value{GDBN} uses the @code{format printf} attribute on all
4958 @code{printf} like functions these check not just @code{printf} calls
4959 but also calls to functions such as @code{fprintf_unfiltered}.
4962 This warning includes uses of the assignment operator within an
4963 @code{if} statement.
4965 @item -Wpointer-arith
4967 @item -Wuninitialized
4969 @item -Wunused-label
4970 This warning has the additional benefit of detecting the absence of the
4971 @code{case} reserved word in a switch statement:
4973 enum @{ FD_SCHEDULED, NOTHING_SCHEDULED @} sched;
4986 @item -Wunused-function
4989 @emph{Pragmatics: Due to the way that @value{GDBN} is implemented most
4990 functions have unused parameters. Consequently the warning
4991 @samp{-Wunused-parameter} is precluded from the list. The macro
4992 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
4993 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
4994 is being used. The options @samp{-Wall} and @samp{-Wunused} are also
4995 precluded because they both include @samp{-Wunused-parameter}.}
4997 @emph{Pragmatics: @value{GDBN} has not simply accepted the warnings
4998 enabled by @samp{-Wall -Werror -W...}. Instead it is selecting warnings
4999 when and where their benefits can be demonstrated.}
5001 @subsection Formatting
5003 @cindex source code formatting
5004 The standard GNU recommendations for formatting must be followed
5007 A function declaration should not have its name in column zero. A
5008 function definition should have its name in column zero.
5012 static void foo (void);
5020 @emph{Pragmatics: This simplifies scripting. Function definitions can
5021 be found using @samp{^function-name}.}
5023 There must be a space between a function or macro name and the opening
5024 parenthesis of its argument list (except for macro definitions, as
5025 required by C). There must not be a space after an open paren/bracket
5026 or before a close paren/bracket.
5028 While additional whitespace is generally helpful for reading, do not use
5029 more than one blank line to separate blocks, and avoid adding whitespace
5030 after the end of a program line (as of 1/99, some 600 lines had
5031 whitespace after the semicolon). Excess whitespace causes difficulties
5032 for @code{diff} and @code{patch} utilities.
5034 Pointers are declared using the traditional K&R C style:
5048 @subsection Comments
5050 @cindex comment formatting
5051 The standard GNU requirements on comments must be followed strictly.
5053 Block comments must appear in the following form, with no @code{/*}- or
5054 @code{*/}-only lines, and no leading @code{*}:
5057 /* Wait for control to return from inferior to debugger. If inferior
5058 gets a signal, we may decide to start it up again instead of
5059 returning. That is why there is a loop in this function. When
5060 this function actually returns it means the inferior should be left
5061 stopped and @value{GDBN} should read more commands. */
5064 (Note that this format is encouraged by Emacs; tabbing for a multi-line
5065 comment works correctly, and @kbd{M-q} fills the block consistently.)
5067 Put a blank line between the block comments preceding function or
5068 variable definitions, and the definition itself.
5070 In general, put function-body comments on lines by themselves, rather
5071 than trying to fit them into the 20 characters left at the end of a
5072 line, since either the comment or the code will inevitably get longer
5073 than will fit, and then somebody will have to move it anyhow.
5077 @cindex C data types
5078 Code must not depend on the sizes of C data types, the format of the
5079 host's floating point numbers, the alignment of anything, or the order
5080 of evaluation of expressions.
5082 @cindex function usage
5083 Use functions freely. There are only a handful of compute-bound areas
5084 in @value{GDBN} that might be affected by the overhead of a function
5085 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
5086 limited by the target interface (whether serial line or system call).
5088 However, use functions with moderation. A thousand one-line functions
5089 are just as hard to understand as a single thousand-line function.
5091 @emph{Macros are bad, M'kay.}
5092 (But if you have to use a macro, make sure that the macro arguments are
5093 protected with parentheses.)
5097 Declarations like @samp{struct foo *} should be used in preference to
5098 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
5101 @subsection Function Prototypes
5102 @cindex function prototypes
5104 Prototypes must be used when both @emph{declaring} and @emph{defining}
5105 a function. Prototypes for @value{GDBN} functions must include both the
5106 argument type and name, with the name matching that used in the actual
5107 function definition.
5109 All external functions should have a declaration in a header file that
5110 callers include, except for @code{_initialize_*} functions, which must
5111 be external so that @file{init.c} construction works, but shouldn't be
5112 visible to random source files.
5114 Where a source file needs a forward declaration of a static function,
5115 that declaration must appear in a block near the top of the source file.
5118 @subsection Internal Error Recovery
5120 During its execution, @value{GDBN} can encounter two types of errors.
5121 User errors and internal errors. User errors include not only a user
5122 entering an incorrect command but also problems arising from corrupt
5123 object files and system errors when interacting with the target.
5124 Internal errors include situations where @value{GDBN} has detected, at
5125 run time, a corrupt or erroneous situation.
5127 When reporting an internal error, @value{GDBN} uses
5128 @code{internal_error} and @code{gdb_assert}.
5130 @value{GDBN} must not call @code{abort} or @code{assert}.
5132 @emph{Pragmatics: There is no @code{internal_warning} function. Either
5133 the code detected a user error, recovered from it and issued a
5134 @code{warning} or the code failed to correctly recover from the user
5135 error and issued an @code{internal_error}.}
5137 @subsection File Names
5139 Any file used when building the core of @value{GDBN} must be in lower
5140 case. Any file used when building the core of @value{GDBN} must be 8.3
5141 unique. These requirements apply to both source and generated files.
5143 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
5144 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
5145 is introduced to the build process both @file{Makefile.in} and
5146 @file{configure.in} need to be modified accordingly. Compare the
5147 convoluted conversion process needed to transform @file{COPYING} into
5148 @file{copying.c} with the conversion needed to transform
5149 @file{version.in} into @file{version.c}.}
5151 Any file non 8.3 compliant file (that is not used when building the core
5152 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
5154 @emph{Pragmatics: This is clearly a compromise.}
5156 When @value{GDBN} has a local version of a system header file (ex
5157 @file{string.h}) the file name based on the POSIX header prefixed with
5158 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
5159 independent: they should use only macros defined by @file{configure},
5160 the compiler, or the host; they should include only system headers; they
5161 should refer only to system types. They may be shared between multiple
5162 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
5164 For other files @samp{-} is used as the separator.
5167 @subsection Include Files
5169 A @file{.c} file should include @file{defs.h} first.
5171 A @file{.c} file should directly include the @code{.h} file of every
5172 declaration and/or definition it directly refers to. It cannot rely on
5175 A @file{.h} file should directly include the @code{.h} file of every
5176 declaration and/or definition it directly refers to. It cannot rely on
5177 indirect inclusion. Exception: The file @file{defs.h} does not need to
5178 be directly included.
5180 An external declaration should only appear in one include file.
5182 An external declaration should never appear in a @code{.c} file.
5183 Exception: a declaration for the @code{_initialize} function that
5184 pacifies @option{-Wmissing-declaration}.
5186 A @code{typedef} definition should only appear in one include file.
5188 An opaque @code{struct} declaration can appear in multiple @file{.h}
5189 files. Where possible, a @file{.h} file should use an opaque
5190 @code{struct} declaration instead of an include.
5192 All @file{.h} files should be wrapped in:
5195 #ifndef INCLUDE_FILE_NAME_H
5196 #define INCLUDE_FILE_NAME_H
5202 @subsection Clean Design and Portable Implementation
5205 In addition to getting the syntax right, there's the little question of
5206 semantics. Some things are done in certain ways in @value{GDBN} because long
5207 experience has shown that the more obvious ways caused various kinds of
5210 @cindex assumptions about targets
5211 You can't assume the byte order of anything that comes from a target
5212 (including @var{value}s, object files, and instructions). Such things
5213 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
5214 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
5215 such as @code{bfd_get_32}.
5217 You can't assume that you know what interface is being used to talk to
5218 the target system. All references to the target must go through the
5219 current @code{target_ops} vector.
5221 You can't assume that the host and target machines are the same machine
5222 (except in the ``native'' support modules). In particular, you can't
5223 assume that the target machine's header files will be available on the
5224 host machine. Target code must bring along its own header files --
5225 written from scratch or explicitly donated by their owner, to avoid
5229 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
5230 to write the code portably than to conditionalize it for various
5233 @cindex system dependencies
5234 New @code{#ifdef}'s which test for specific compilers or manufacturers
5235 or operating systems are unacceptable. All @code{#ifdef}'s should test
5236 for features. The information about which configurations contain which
5237 features should be segregated into the configuration files. Experience
5238 has proven far too often that a feature unique to one particular system
5239 often creeps into other systems; and that a conditional based on some
5240 predefined macro for your current system will become worthless over
5241 time, as new versions of your system come out that behave differently
5242 with regard to this feature.
5244 Adding code that handles specific architectures, operating systems,
5245 target interfaces, or hosts, is not acceptable in generic code.
5247 @cindex portable file name handling
5248 @cindex file names, portability
5249 One particularly notorious area where system dependencies tend to
5250 creep in is handling of file names. The mainline @value{GDBN} code
5251 assumes Posix semantics of file names: absolute file names begin with
5252 a forward slash @file{/}, slashes are used to separate leading
5253 directories, case-sensitive file names. These assumptions are not
5254 necessarily true on non-Posix systems such as MS-Windows. To avoid
5255 system-dependent code where you need to take apart or construct a file
5256 name, use the following portable macros:
5259 @findex HAVE_DOS_BASED_FILE_SYSTEM
5260 @item HAVE_DOS_BASED_FILE_SYSTEM
5261 This preprocessing symbol is defined to a non-zero value on hosts
5262 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
5263 symbol to write conditional code which should only be compiled for
5266 @findex IS_DIR_SEPARATOR
5267 @item IS_DIR_SEPARATOR (@var{c})
5268 Evaluates to a non-zero value if @var{c} is a directory separator
5269 character. On Unix and GNU/Linux systems, only a slash @file{/} is
5270 such a character, but on Windows, both @file{/} and @file{\} will
5273 @findex IS_ABSOLUTE_PATH
5274 @item IS_ABSOLUTE_PATH (@var{file})
5275 Evaluates to a non-zero value if @var{file} is an absolute file name.
5276 For Unix and GNU/Linux hosts, a name which begins with a slash
5277 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
5278 @file{x:\bar} are also absolute file names.
5280 @findex FILENAME_CMP
5281 @item FILENAME_CMP (@var{f1}, @var{f2})
5282 Calls a function which compares file names @var{f1} and @var{f2} as
5283 appropriate for the underlying host filesystem. For Posix systems,
5284 this simply calls @code{strcmp}; on case-insensitive filesystems it
5285 will call @code{strcasecmp} instead.
5287 @findex DIRNAME_SEPARATOR
5288 @item DIRNAME_SEPARATOR
5289 Evaluates to a character which separates directories in
5290 @code{PATH}-style lists, typically held in environment variables.
5291 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
5293 @findex SLASH_STRING
5295 This evaluates to a constant string you should use to produce an
5296 absolute filename from leading directories and the file's basename.
5297 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
5298 @code{"\\"} for some Windows-based ports.
5301 In addition to using these macros, be sure to use portable library
5302 functions whenever possible. For example, to extract a directory or a
5303 basename part from a file name, use the @code{dirname} and
5304 @code{basename} library functions (available in @code{libiberty} for
5305 platforms which don't provide them), instead of searching for a slash
5306 with @code{strrchr}.
5308 Another way to generalize @value{GDBN} along a particular interface is with an
5309 attribute struct. For example, @value{GDBN} has been generalized to handle
5310 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
5311 by defining the @code{target_ops} structure and having a current target (as
5312 well as a stack of targets below it, for memory references). Whenever
5313 something needs to be done that depends on which remote interface we are
5314 using, a flag in the current target_ops structure is tested (e.g.,
5315 @code{target_has_stack}), or a function is called through a pointer in the
5316 current target_ops structure. In this way, when a new remote interface
5317 is added, only one module needs to be touched---the one that actually
5318 implements the new remote interface. Other examples of
5319 attribute-structs are BFD access to multiple kinds of object file
5320 formats, or @value{GDBN}'s access to multiple source languages.
5322 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
5323 the code interfacing between @code{ptrace} and the rest of
5324 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
5325 something was very painful. In @value{GDBN} 4.x, these have all been
5326 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
5327 with variations between systems the same way any system-independent
5328 file would (hooks, @code{#if defined}, etc.), and machines which are
5329 radically different don't need to use @file{infptrace.c} at all.
5331 All debugging code must be controllable using the @samp{set debug
5332 @var{module}} command. Do not use @code{printf} to print trace
5333 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
5334 @code{#ifdef DEBUG}.
5339 @chapter Porting @value{GDBN}
5340 @cindex porting to new machines
5342 Most of the work in making @value{GDBN} compile on a new machine is in
5343 specifying the configuration of the machine. This is done in a
5344 dizzying variety of header files and configuration scripts, which we
5345 hope to make more sensible soon. Let's say your new host is called an
5346 @var{xyz} (e.g., @samp{sun4}), and its full three-part configuration
5347 name is @code{@var{arch}-@var{xvend}-@var{xos}} (e.g.,
5348 @samp{sparc-sun-sunos4}). In particular:
5352 In the top level directory, edit @file{config.sub} and add @var{arch},
5353 @var{xvend}, and @var{xos} to the lists of supported architectures,
5354 vendors, and operating systems near the bottom of the file. Also, add
5355 @var{xyz} as an alias that maps to
5356 @code{@var{arch}-@var{xvend}-@var{xos}}. You can test your changes by
5360 ./config.sub @var{xyz}
5367 ./config.sub @code{@var{arch}-@var{xvend}-@var{xos}}
5371 which should both respond with @code{@var{arch}-@var{xvend}-@var{xos}}
5372 and no error messages.
5375 You need to port BFD, if that hasn't been done already. Porting BFD is
5376 beyond the scope of this manual.
5379 To configure @value{GDBN} itself, edit @file{gdb/configure.host} to recognize
5380 your system and set @code{gdb_host} to @var{xyz}, and (unless your
5381 desired target is already available) also edit @file{gdb/configure.tgt},
5382 setting @code{gdb_target} to something appropriate (for instance,
5385 @emph{Maintainer's note: Work in progress. The file
5386 @file{gdb/configure.host} originally needed to be modified when either a
5387 new native target or a new host machine was being added to @value{GDBN}.
5388 Recent changes have removed this requirement. The file now only needs
5389 to be modified when adding a new native configuration. This will likely
5390 changed again in the future.}
5393 Finally, you'll need to specify and define @value{GDBN}'s host-, native-, and
5394 target-dependent @file{.h} and @file{.c} files used for your
5398 @node Versions and Branches
5399 @chapter Versions and Branches
5403 @value{GDBN}'s version is determined by the file
5404 @file{gdb/version.in} and takes one of the following forms:
5407 @item @var{major}.@var{minor}
5408 @itemx @var{major}.@var{minor}.@var{patchlevel}
5409 an official release (e.g., 6.2 or 6.2.1)
5410 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}
5411 a snapshot taken at @var{YYYY}-@var{MM}-@var{DD}-gmt (e.g.,
5412 6.1.50.20020302, 6.1.90.20020304, or 6.1.0.20020308)
5413 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}-cvs
5414 a @sc{cvs} check out drawn on @var{YYYY}-@var{MM}-@var{DD} (e.g.,
5415 6.1.50.20020302-cvs, 6.1.90.20020304-cvs, or 6.1.0.20020308-cvs)
5416 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD} (@var{vendor})
5417 a vendor specific release of @value{GDBN}, that while based on@*
5418 @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD},
5419 may include additional changes
5422 @value{GDBN}'s mainline uses the @var{major} and @var{minor} version
5423 numbers from the most recent release branch, with a @var{patchlevel}
5424 of 50. At the time each new release branch is created, the mainline's
5425 @var{major} and @var{minor} version numbers are updated.
5427 @value{GDBN}'s release branch is similar. When the branch is cut, the
5428 @var{patchlevel} is changed from 50 to 90. As draft releases are
5429 drawn from the branch, the @var{patchlevel} is incremented. Once the
5430 first release (@var{major}.@var{minor}) has been made, the
5431 @var{patchlevel} is set to 0 and updates have an incremented
5434 For snapshots, and @sc{cvs} check outs, it is also possible to
5435 identify the @sc{cvs} origin:
5438 @item @var{major}.@var{minor}.50.@var{YYYY}@var{MM}@var{DD}
5439 drawn from the @sc{head} of mainline @sc{cvs} (e.g., 6.1.50.20020302)
5440 @item @var{major}.@var{minor}.90.@var{YYYY}@var{MM}@var{DD}
5441 @itemx @var{major}.@var{minor}.91.@var{YYYY}@var{MM}@var{DD} @dots{}
5442 drawn from a release branch prior to the release (e.g.,
5444 @item @var{major}.@var{minor}.0.@var{YYYY}@var{MM}@var{DD}
5445 @itemx @var{major}.@var{minor}.1.@var{YYYY}@var{MM}@var{DD} @dots{}
5446 drawn from a release branch after the release (e.g., 6.2.0.20020308)
5449 If the previous @value{GDBN} version is 6.1 and the current version is
5450 6.2, then, substituting 6 for @var{major} and 1 or 2 for @var{minor},
5451 here's an illustration of a typical sequence:
5458 +--------------------------.
5461 6.2.50.20020303-cvs 6.1.90 (draft #1)
5463 6.2.50.20020304-cvs 6.1.90.20020304-cvs
5465 6.2.50.20020305-cvs 6.1.91 (draft #2)
5467 6.2.50.20020306-cvs 6.1.91.20020306-cvs
5469 6.2.50.20020307-cvs 6.2 (release)
5471 6.2.50.20020308-cvs 6.2.0.20020308-cvs
5473 6.2.50.20020309-cvs 6.2.1 (update)
5475 6.2.50.20020310-cvs <branch closed>
5479 +--------------------------.
5482 6.3.50.20020312-cvs 6.2.90 (draft #1)
5486 @section Release Branches
5487 @cindex Release Branches
5489 @value{GDBN} draws a release series (6.2, 6.2.1, @dots{}) from a
5490 single release branch, and identifies that branch using the @sc{cvs}
5494 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-branchpoint
5495 gdb_@var{major}_@var{minor}-branch
5496 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-release
5499 @emph{Pragmatics: To help identify the date at which a branch or
5500 release is made, both the branchpoint and release tags include the
5501 date that they are cut (@var{YYYY}@var{MM}@var{DD}) in the tag. The
5502 branch tag, denoting the head of the branch, does not need this.}
5504 @section Vendor Branches
5505 @cindex vendor branches
5507 To avoid version conflicts, vendors are expected to modify the file
5508 @file{gdb/version.in} to include a vendor unique alphabetic identifier
5509 (an official @value{GDBN} release never uses alphabetic characters in
5510 its version identifer). E.g., @samp{6.2widgit2}, or @samp{6.2 (Widgit
5513 @section Experimental Branches
5514 @cindex experimental branches
5516 @subsection Guidelines
5518 @value{GDBN} permits the creation of branches, cut from the @sc{cvs}
5519 repository, for experimental development. Branches make it possible
5520 for developers to share preliminary work, and maintainers to examine
5521 significant new developments.
5523 The following are a set of guidelines for creating such branches:
5527 @item a branch has an owner
5528 The owner can set further policy for a branch, but may not change the
5529 ground rules. In particular, they can set a policy for commits (be it
5530 adding more reviewers or deciding who can commit).
5532 @item all commits are posted
5533 All changes committed to a branch shall also be posted to
5534 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} patches
5535 mailing list}. While commentary on such changes are encouraged, people
5536 should remember that the changes only apply to a branch.
5538 @item all commits are covered by an assignment
5539 This ensures that all changes belong to the Free Software Foundation,
5540 and avoids the possibility that the branch may become contaminated.
5542 @item a branch is focused
5543 A focused branch has a single objective or goal, and does not contain
5544 unnecessary or irrelevant changes. Cleanups, where identified, being
5545 be pushed into the mainline as soon as possible.
5547 @item a branch tracks mainline
5548 This keeps the level of divergence under control. It also keeps the
5549 pressure on developers to push cleanups and other stuff into the
5552 @item a branch shall contain the entire @value{GDBN} module
5553 The @value{GDBN} module @code{gdb} should be specified when creating a
5554 branch (branches of individual files should be avoided). @xref{Tags}.
5556 @item a branch shall be branded using @file{version.in}
5557 The file @file{gdb/version.in} shall be modified so that it identifies
5558 the branch @var{owner} and branch @var{name}, e.g.,
5559 @samp{6.2.50.20030303_owner_name} or @samp{6.2 (Owner Name)}.
5566 To simplify the identification of @value{GDBN} branches, the following
5567 branch tagging convention is strongly recommended:
5571 @item @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5572 @itemx @var{owner}_@var{name}-@var{YYYYMMDD}-branch
5573 The branch point and corresponding branch tag. @var{YYYYMMDD} is the
5574 date that the branch was created. A branch is created using the
5575 sequence: @anchor{experimental branch tags}
5577 cvs rtag @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint gdb
5578 cvs rtag -b -r @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint \
5579 @var{owner}_@var{name}-@var{YYYYMMDD}-branch gdb
5582 @item @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5583 The tagged point, on the mainline, that was used when merging the branch
5584 on @var{yyyymmdd}. To merge in all changes since the branch was cut,
5585 use a command sequence like:
5587 cvs rtag @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint gdb
5589 -j@var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5590 -j@var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5593 Similar sequences can be used to just merge in changes since the last
5599 For further information on @sc{cvs}, see
5600 @uref{http://www.gnu.org/software/cvs/, Concurrent Versions System}.
5602 @node Start of New Year Procedure
5603 @chapter Start of New Year Procedure
5604 @cindex new year procedure
5606 At the start of each new year, the following actions should be performed:
5610 Rotate the ChangeLog file
5612 The current @file{ChangeLog} file should be renamed into
5613 @file{ChangeLog-YYYY} where YYYY is the year that has just passed.
5614 A new @file{ChangeLog} file should be created, and its contents should
5615 contain a reference to the previous ChangeLog. The following should
5616 also be preserved at the end of the new ChangeLog, in order to provide
5617 the appropriate settings when editing this file with Emacs:
5623 version-control: never
5628 Update the copyright year in the startup message
5630 Update the copyright year in file @file{top.c}, function
5631 @code{print_gdb_version}.
5636 @chapter Releasing @value{GDBN}
5637 @cindex making a new release of gdb
5639 @section Branch Commit Policy
5641 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
5642 5.1 and 5.2 all used the below:
5646 The @file{gdb/MAINTAINERS} file still holds.
5648 Don't fix something on the branch unless/until it is also fixed in the
5649 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
5650 file is better than committing a hack.
5652 When considering a patch for the branch, suggested criteria include:
5653 Does it fix a build? Does it fix the sequence @kbd{break main; run}
5654 when debugging a static binary?
5656 The further a change is from the core of @value{GDBN}, the less likely
5657 the change will worry anyone (e.g., target specific code).
5659 Only post a proposal to change the core of @value{GDBN} after you've
5660 sent individual bribes to all the people listed in the
5661 @file{MAINTAINERS} file @t{;-)}
5664 @emph{Pragmatics: Provided updates are restricted to non-core
5665 functionality there is little chance that a broken change will be fatal.
5666 This means that changes such as adding a new architectures or (within
5667 reason) support for a new host are considered acceptable.}
5670 @section Obsoleting code
5672 Before anything else, poke the other developers (and around the source
5673 code) to see if there is anything that can be removed from @value{GDBN}
5674 (an old target, an unused file).
5676 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
5677 line. Doing this means that it is easy to identify something that has
5678 been obsoleted when greping through the sources.
5680 The process is done in stages --- this is mainly to ensure that the
5681 wider @value{GDBN} community has a reasonable opportunity to respond.
5682 Remember, everything on the Internet takes a week.
5686 Post the proposal on @email{gdb@@sources.redhat.com, the GDB mailing
5687 list} Creating a bug report to track the task's state, is also highly
5692 Post the proposal on @email{gdb-announce@@sources.redhat.com, the GDB
5693 Announcement mailing list}.
5697 Go through and edit all relevant files and lines so that they are
5698 prefixed with the word @code{OBSOLETE}.
5700 Wait until the next GDB version, containing this obsolete code, has been
5703 Remove the obsolete code.
5707 @emph{Maintainer note: While removing old code is regrettable it is
5708 hopefully better for @value{GDBN}'s long term development. Firstly it
5709 helps the developers by removing code that is either no longer relevant
5710 or simply wrong. Secondly since it removes any history associated with
5711 the file (effectively clearing the slate) the developer has a much freer
5712 hand when it comes to fixing broken files.}
5716 @section Before the Branch
5718 The most important objective at this stage is to find and fix simple
5719 changes that become a pain to track once the branch is created. For
5720 instance, configuration problems that stop @value{GDBN} from even
5721 building. If you can't get the problem fixed, document it in the
5722 @file{gdb/PROBLEMS} file.
5724 @subheading Prompt for @file{gdb/NEWS}
5726 People always forget. Send a post reminding them but also if you know
5727 something interesting happened add it yourself. The @code{schedule}
5728 script will mention this in its e-mail.
5730 @subheading Review @file{gdb/README}
5732 Grab one of the nightly snapshots and then walk through the
5733 @file{gdb/README} looking for anything that can be improved. The
5734 @code{schedule} script will mention this in its e-mail.
5736 @subheading Refresh any imported files.
5738 A number of files are taken from external repositories. They include:
5742 @file{texinfo/texinfo.tex}
5744 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
5747 @file{etc/standards.texi}, @file{etc/make-stds.texi}
5750 @subheading Check the ARI
5752 @uref{http://sources.redhat.com/gdb/ari,,A.R.I.} is an @code{awk} script
5753 (Awk Regression Index ;-) that checks for a number of errors and coding
5754 conventions. The checks include things like using @code{malloc} instead
5755 of @code{xmalloc} and file naming problems. There shouldn't be any
5758 @subsection Review the bug data base
5760 Close anything obviously fixed.
5762 @subsection Check all cross targets build
5764 The targets are listed in @file{gdb/MAINTAINERS}.
5767 @section Cut the Branch
5769 @subheading Create the branch
5774 $ V=`echo $v | sed 's/\./_/g'`
5775 $ D=`date -u +%Y-%m-%d`
5778 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
5779 -D $D-gmt gdb_$V-$D-branchpoint insight+dejagnu
5780 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag
5781 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight+dejagnu
5784 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
5785 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight+dejagnu
5786 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
5787 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight+dejagnu
5795 by using @kbd{-D YYYY-MM-DD-gmt} the branch is forced to an exact
5798 the trunk is first taged so that the branch point can easily be found
5800 Insight (which includes GDB) and dejagnu are all tagged at the same time
5802 @file{version.in} gets bumped to avoid version number conflicts
5804 the reading of @file{.cvsrc} is disabled using @file{-f}
5807 @subheading Update @file{version.in}
5812 $ V=`echo $v | sed 's/\./_/g'`
5816 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src co \
5817 -r gdb_$V-branch src/gdb/version.in
5818 cvs -f -d :ext:sources.redhat.com:/cvs/src co
5819 -r gdb_5_2-branch src/gdb/version.in
5821 U src/gdb/version.in
5823 $ echo $u.90-0000-00-00-cvs > version.in
5825 5.1.90-0000-00-00-cvs
5826 $ cvs -f commit version.in
5831 @file{0000-00-00} is used as a date to pump prime the version.in update
5834 @file{.90} and the previous branch version are used as fairly arbitrary
5835 initial branch version number
5839 @subheading Update the web and news pages
5843 @subheading Tweak cron to track the new branch
5845 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
5846 This file needs to be updated so that:
5850 a daily timestamp is added to the file @file{version.in}
5852 the new branch is included in the snapshot process
5856 See the file @file{gdbadmin/cron/README} for how to install the updated
5859 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
5860 any changes. That file is copied to both the branch/ and current/
5861 snapshot directories.
5864 @subheading Update the NEWS and README files
5866 The @file{NEWS} file needs to be updated so that on the branch it refers
5867 to @emph{changes in the current release} while on the trunk it also
5868 refers to @emph{changes since the current release}.
5870 The @file{README} file needs to be updated so that it refers to the
5873 @subheading Post the branch info
5875 Send an announcement to the mailing lists:
5879 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
5881 @email{gdb@@sources.redhat.com, GDB Discsussion mailing list} and
5882 @email{gdb-testers@@sources.redhat.com, GDB Discsussion mailing list}
5885 @emph{Pragmatics: The branch creation is sent to the announce list to
5886 ensure that people people not subscribed to the higher volume discussion
5889 The announcement should include:
5895 how to check out the branch using CVS
5897 the date/number of weeks until the release
5899 the branch commit policy
5903 @section Stabilize the branch
5905 Something goes here.
5907 @section Create a Release
5909 The process of creating and then making available a release is broken
5910 down into a number of stages. The first part addresses the technical
5911 process of creating a releasable tar ball. The later stages address the
5912 process of releasing that tar ball.
5914 When making a release candidate just the first section is needed.
5916 @subsection Create a release candidate
5918 The objective at this stage is to create a set of tar balls that can be
5919 made available as a formal release (or as a less formal release
5922 @subsubheading Freeze the branch
5924 Send out an e-mail notifying everyone that the branch is frozen to
5925 @email{gdb-patches@@sources.redhat.com}.
5927 @subsubheading Establish a few defaults.
5932 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
5934 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
5938 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
5940 /home/gdbadmin/bin/autoconf
5949 Check the @code{autoconf} version carefully. You want to be using the
5950 version taken from the @file{binutils} snapshot directory, which can be
5951 found at @uref{ftp://sources.redhat.com/pub/binutils/}. It is very
5952 unlikely that a system installed version of @code{autoconf} (e.g.,
5953 @file{/usr/bin/autoconf}) is correct.
5956 @subsubheading Check out the relevant modules:
5959 $ for m in gdb insight dejagnu
5961 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
5971 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
5972 any confusion between what is written here and what your local
5973 @code{cvs} really does.
5976 @subsubheading Update relevant files.
5982 Major releases get their comments added as part of the mainline. Minor
5983 releases should probably mention any significant bugs that were fixed.
5985 Don't forget to include the @file{ChangeLog} entry.
5988 $ emacs gdb/src/gdb/NEWS
5993 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
5994 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
5999 You'll need to update:
6011 $ emacs gdb/src/gdb/README
6016 $ cp gdb/src/gdb/README insight/src/gdb/README
6017 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6020 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
6021 before the initial branch was cut so just a simple substitute is needed
6024 @emph{Maintainer note: Other projects generate @file{README} and
6025 @file{INSTALL} from the core documentation. This might be worth
6028 @item gdb/version.in
6031 $ echo $v > gdb/src/gdb/version.in
6032 $ cat gdb/src/gdb/version.in
6034 $ emacs gdb/src/gdb/version.in
6037 ... Bump to version ...
6039 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
6040 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6043 @item dejagnu/src/dejagnu/configure.in
6045 Dejagnu is more complicated. The version number is a parameter to
6046 @code{AM_INIT_AUTOMAKE}. Tweak it to read something like gdb-5.1.91.
6048 Don't forget to re-generate @file{configure}.
6050 Don't forget to include a @file{ChangeLog} entry.
6053 $ emacs dejagnu/src/dejagnu/configure.in
6058 $ ( cd dejagnu/src/dejagnu && autoconf )
6063 @subsubheading Do the dirty work
6065 This is identical to the process used to create the daily snapshot.
6068 $ for m in gdb insight
6070 ( cd $m/src && gmake -f src-release $m.tar )
6072 $ ( m=dejagnu; cd $m/src && gmake -f src-release $m.tar.bz2 )
6075 If the top level source directory does not have @file{src-release}
6076 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
6079 $ for m in gdb insight
6081 ( cd $m/src && gmake -f Makefile.in $m.tar )
6083 $ ( m=dejagnu; cd $m/src && gmake -f Makefile.in $m.tar.bz2 )
6086 @subsubheading Check the source files
6088 You're looking for files that have mysteriously disappeared.
6089 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
6090 for the @file{version.in} update @kbd{cronjob}.
6093 $ ( cd gdb/src && cvs -f -q -n update )
6097 @dots{} lots of generated files @dots{}
6102 @dots{} lots of generated files @dots{}
6107 @emph{Don't worry about the @file{gdb.info-??} or
6108 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
6109 was also generated only something strange with CVS means that they
6110 didn't get supressed). Fixing it would be nice though.}
6112 @subsubheading Create compressed versions of the release
6118 dejagnu/ dejagnu-gdb-5.2.tar.bz2 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
6119 $ for m in gdb insight
6121 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
6122 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
6132 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
6133 in that mode, @code{gzip} does not know the name of the file and, hence,
6134 can not include it in the compressed file. This is also why the release
6135 process runs @code{tar} and @code{bzip2} as separate passes.
6138 @subsection Sanity check the tar ball
6140 Pick a popular machine (Solaris/PPC?) and try the build on that.
6143 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
6148 $ ./gdb/gdb ./gdb/gdb
6152 Breakpoint 1 at 0x80732bc: file main.c, line 734.
6154 Starting program: /tmp/gdb-5.2/gdb/gdb
6156 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
6157 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
6159 $1 = @{argc = 136426532, argv = 0x821b7f0@}
6163 @subsection Make a release candidate available
6165 If this is a release candidate then the only remaining steps are:
6169 Commit @file{version.in} and @file{ChangeLog}
6171 Tweak @file{version.in} (and @file{ChangeLog} to read
6172 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
6173 process can restart.
6175 Make the release candidate available in
6176 @uref{ftp://sources.redhat.com/pub/gdb/snapshots/branch}
6178 Notify the relevant mailing lists ( @email{gdb@@sources.redhat.com} and
6179 @email{gdb-testers@@sources.redhat.com} that the candidate is available.
6182 @subsection Make a formal release available
6184 (And you thought all that was required was to post an e-mail.)
6186 @subsubheading Install on sware
6188 Copy the new files to both the release and the old release directory:
6191 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
6192 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
6196 Clean up the releases directory so that only the most recent releases
6197 are available (e.g. keep 5.2 and 5.2.1 but remove 5.1):
6200 $ cd ~ftp/pub/gdb/releases
6205 Update the file @file{README} and @file{.message} in the releases
6212 $ ln README .message
6215 @subsubheading Update the web pages.
6219 @item htdocs/download/ANNOUNCEMENT
6220 This file, which is posted as the official announcement, includes:
6223 General announcement
6225 News. If making an @var{M}.@var{N}.1 release, retain the news from
6226 earlier @var{M}.@var{N} release.
6231 @item htdocs/index.html
6232 @itemx htdocs/news/index.html
6233 @itemx htdocs/download/index.html
6234 These files include:
6237 announcement of the most recent release
6239 news entry (remember to update both the top level and the news directory).
6241 These pages also need to be regenerate using @code{index.sh}.
6243 @item download/onlinedocs/
6244 You need to find the magic command that is used to generate the online
6245 docs from the @file{.tar.bz2}. The best way is to look in the output
6246 from one of the nightly @code{cron} jobs and then just edit accordingly.
6250 $ ~/ss/update-web-docs \
6251 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6253 /www/sourceware/htdocs/gdb/download/onlinedocs \
6258 Just like the online documentation. Something like:
6261 $ /bin/sh ~/ss/update-web-ari \
6262 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6264 /www/sourceware/htdocs/gdb/download/ari \
6270 @subsubheading Shadow the pages onto gnu
6272 Something goes here.
6275 @subsubheading Install the @value{GDBN} tar ball on GNU
6277 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
6278 @file{~ftp/gnu/gdb}.
6280 @subsubheading Make the @file{ANNOUNCEMENT}
6282 Post the @file{ANNOUNCEMENT} file you created above to:
6286 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6288 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
6289 day or so to let things get out)
6291 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
6296 The release is out but you're still not finished.
6298 @subsubheading Commit outstanding changes
6300 In particular you'll need to commit any changes to:
6304 @file{gdb/ChangeLog}
6306 @file{gdb/version.in}
6313 @subsubheading Tag the release
6318 $ d=`date -u +%Y-%m-%d`
6321 $ ( cd insight/src/gdb && cvs -f -q update )
6322 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
6325 Insight is used since that contains more of the release than
6326 @value{GDBN} (@code{dejagnu} doesn't get tagged but I think we can live
6329 @subsubheading Mention the release on the trunk
6331 Just put something in the @file{ChangeLog} so that the trunk also
6332 indicates when the release was made.
6334 @subsubheading Restart @file{gdb/version.in}
6336 If @file{gdb/version.in} does not contain an ISO date such as
6337 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
6338 committed all the release changes it can be set to
6339 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
6340 is important - it affects the snapshot process).
6342 Don't forget the @file{ChangeLog}.
6344 @subsubheading Merge into trunk
6346 The files committed to the branch may also need changes merged into the
6349 @subsubheading Revise the release schedule
6351 Post a revised release schedule to @email{gdb@@sources.redhat.com, GDB
6352 Discussion List} with an updated announcement. The schedule can be
6353 generated by running:
6356 $ ~/ss/schedule `date +%s` schedule
6360 The first parameter is approximate date/time in seconds (from the epoch)
6361 of the most recent release.
6363 Also update the schedule @code{cronjob}.
6365 @section Post release
6367 Remove any @code{OBSOLETE} code.
6374 The testsuite is an important component of the @value{GDBN} package.
6375 While it is always worthwhile to encourage user testing, in practice
6376 this is rarely sufficient; users typically use only a small subset of
6377 the available commands, and it has proven all too common for a change
6378 to cause a significant regression that went unnoticed for some time.
6380 The @value{GDBN} testsuite uses the DejaGNU testing framework.
6381 DejaGNU is built using @code{Tcl} and @code{expect}. The tests
6382 themselves are calls to various @code{Tcl} procs; the framework runs all the
6383 procs and summarizes the passes and fails.
6385 @section Using the Testsuite
6387 @cindex running the test suite
6388 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
6389 testsuite's objdir) and type @code{make check}. This just sets up some
6390 environment variables and invokes DejaGNU's @code{runtest} script. While
6391 the testsuite is running, you'll get mentions of which test file is in use,
6392 and a mention of any unexpected passes or fails. When the testsuite is
6393 finished, you'll get a summary that looks like this:
6398 # of expected passes 6016
6399 # of unexpected failures 58
6400 # of unexpected successes 5
6401 # of expected failures 183
6402 # of unresolved testcases 3
6403 # of untested testcases 5
6406 The ideal test run consists of expected passes only; however, reality
6407 conspires to keep us from this ideal. Unexpected failures indicate
6408 real problems, whether in @value{GDBN} or in the testsuite. Expected
6409 failures are still failures, but ones which have been decided are too
6410 hard to deal with at the time; for instance, a test case might work
6411 everywhere except on AIX, and there is no prospect of the AIX case
6412 being fixed in the near future. Expected failures should not be added
6413 lightly, since you may be masking serious bugs in @value{GDBN}.
6414 Unexpected successes are expected fails that are passing for some
6415 reason, while unresolved and untested cases often indicate some minor
6416 catastrophe, such as the compiler being unable to deal with a test
6419 When making any significant change to @value{GDBN}, you should run the
6420 testsuite before and after the change, to confirm that there are no
6421 regressions. Note that truly complete testing would require that you
6422 run the testsuite with all supported configurations and a variety of
6423 compilers; however this is more than really necessary. In many cases
6424 testing with a single configuration is sufficient. Other useful
6425 options are to test one big-endian (Sparc) and one little-endian (x86)
6426 host, a cross config with a builtin simulator (powerpc-eabi,
6427 mips-elf), or a 64-bit host (Alpha).
6429 If you add new functionality to @value{GDBN}, please consider adding
6430 tests for it as well; this way future @value{GDBN} hackers can detect
6431 and fix their changes that break the functionality you added.
6432 Similarly, if you fix a bug that was not previously reported as a test
6433 failure, please add a test case for it. Some cases are extremely
6434 difficult to test, such as code that handles host OS failures or bugs
6435 in particular versions of compilers, and it's OK not to try to write
6436 tests for all of those.
6438 DejaGNU supports separate build, host, and target machines. However,
6439 some @value{GDBN} test scripts do not work if the build machine and
6440 the host machine are not the same. In such an environment, these scripts
6441 will give a result of ``UNRESOLVED'', like this:
6444 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
6447 @section Testsuite Organization
6449 @cindex test suite organization
6450 The testsuite is entirely contained in @file{gdb/testsuite}. While the
6451 testsuite includes some makefiles and configury, these are very minimal,
6452 and used for little besides cleaning up, since the tests themselves
6453 handle the compilation of the programs that @value{GDBN} will run. The file
6454 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
6455 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
6456 configuration-specific files, typically used for special-purpose
6457 definitions of procs like @code{gdb_load} and @code{gdb_start}.
6459 The tests themselves are to be found in @file{testsuite/gdb.*} and
6460 subdirectories of those. The names of the test files must always end
6461 with @file{.exp}. DejaGNU collects the test files by wildcarding
6462 in the test directories, so both subdirectories and individual files
6463 get chosen and run in alphabetical order.
6465 The following table lists the main types of subdirectories and what they
6466 are for. Since DejaGNU finds test files no matter where they are
6467 located, and since each test file sets up its own compilation and
6468 execution environment, this organization is simply for convenience and
6473 This is the base testsuite. The tests in it should apply to all
6474 configurations of @value{GDBN} (but generic native-only tests may live here).
6475 The test programs should be in the subset of C that is valid K&R,
6476 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
6479 @item gdb.@var{lang}
6480 Language-specific tests for any language @var{lang} besides C. Examples are
6481 @file{gdb.cp} and @file{gdb.java}.
6483 @item gdb.@var{platform}
6484 Non-portable tests. The tests are specific to a specific configuration
6485 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
6488 @item gdb.@var{compiler}
6489 Tests specific to a particular compiler. As of this writing (June
6490 1999), there aren't currently any groups of tests in this category that
6491 couldn't just as sensibly be made platform-specific, but one could
6492 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
6495 @item gdb.@var{subsystem}
6496 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
6497 instance, @file{gdb.disasm} exercises various disassemblers, while
6498 @file{gdb.stabs} tests pathways through the stabs symbol reader.
6501 @section Writing Tests
6502 @cindex writing tests
6504 In many areas, the @value{GDBN} tests are already quite comprehensive; you
6505 should be able to copy existing tests to handle new cases.
6507 You should try to use @code{gdb_test} whenever possible, since it
6508 includes cases to handle all the unexpected errors that might happen.
6509 However, it doesn't cost anything to add new test procedures; for
6510 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
6511 calls @code{gdb_test} multiple times.
6513 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
6514 necessary, such as when @value{GDBN} has several valid responses to a command.
6516 The source language programs do @emph{not} need to be in a consistent
6517 style. Since @value{GDBN} is used to debug programs written in many different
6518 styles, it's worth having a mix of styles in the testsuite; for
6519 instance, some @value{GDBN} bugs involving the display of source lines would
6520 never manifest themselves if the programs used GNU coding style
6527 Check the @file{README} file, it often has useful information that does not
6528 appear anywhere else in the directory.
6531 * Getting Started:: Getting started working on @value{GDBN}
6532 * Debugging GDB:: Debugging @value{GDBN} with itself
6535 @node Getting Started,,, Hints
6537 @section Getting Started
6539 @value{GDBN} is a large and complicated program, and if you first starting to
6540 work on it, it can be hard to know where to start. Fortunately, if you
6541 know how to go about it, there are ways to figure out what is going on.
6543 This manual, the @value{GDBN} Internals manual, has information which applies
6544 generally to many parts of @value{GDBN}.
6546 Information about particular functions or data structures are located in
6547 comments with those functions or data structures. If you run across a
6548 function or a global variable which does not have a comment correctly
6549 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
6550 free to submit a bug report, with a suggested comment if you can figure
6551 out what the comment should say. If you find a comment which is
6552 actually wrong, be especially sure to report that.
6554 Comments explaining the function of macros defined in host, target, or
6555 native dependent files can be in several places. Sometimes they are
6556 repeated every place the macro is defined. Sometimes they are where the
6557 macro is used. Sometimes there is a header file which supplies a
6558 default definition of the macro, and the comment is there. This manual
6559 also documents all the available macros.
6560 @c (@pxref{Host Conditionals}, @pxref{Target
6561 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
6564 Start with the header files. Once you have some idea of how
6565 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
6566 @file{gdbtypes.h}), you will find it much easier to understand the
6567 code which uses and creates those symbol tables.
6569 You may wish to process the information you are getting somehow, to
6570 enhance your understanding of it. Summarize it, translate it to another
6571 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
6572 the code to predict what a test case would do and write the test case
6573 and verify your prediction, etc. If you are reading code and your eyes
6574 are starting to glaze over, this is a sign you need to use a more active
6577 Once you have a part of @value{GDBN} to start with, you can find more
6578 specifically the part you are looking for by stepping through each
6579 function with the @code{next} command. Do not use @code{step} or you
6580 will quickly get distracted; when the function you are stepping through
6581 calls another function try only to get a big-picture understanding
6582 (perhaps using the comment at the beginning of the function being
6583 called) of what it does. This way you can identify which of the
6584 functions being called by the function you are stepping through is the
6585 one which you are interested in. You may need to examine the data
6586 structures generated at each stage, with reference to the comments in
6587 the header files explaining what the data structures are supposed to
6590 Of course, this same technique can be used if you are just reading the
6591 code, rather than actually stepping through it. The same general
6592 principle applies---when the code you are looking at calls something
6593 else, just try to understand generally what the code being called does,
6594 rather than worrying about all its details.
6596 @cindex command implementation
6597 A good place to start when tracking down some particular area is with
6598 a command which invokes that feature. Suppose you want to know how
6599 single-stepping works. As a @value{GDBN} user, you know that the
6600 @code{step} command invokes single-stepping. The command is invoked
6601 via command tables (see @file{command.h}); by convention the function
6602 which actually performs the command is formed by taking the name of
6603 the command and adding @samp{_command}, or in the case of an
6604 @code{info} subcommand, @samp{_info}. For example, the @code{step}
6605 command invokes the @code{step_command} function and the @code{info
6606 display} command invokes @code{display_info}. When this convention is
6607 not followed, you might have to use @code{grep} or @kbd{M-x
6608 tags-search} in emacs, or run @value{GDBN} on itself and set a
6609 breakpoint in @code{execute_command}.
6611 @cindex @code{bug-gdb} mailing list
6612 If all of the above fail, it may be appropriate to ask for information
6613 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
6614 wondering if anyone could give me some tips about understanding
6615 @value{GDBN}''---if we had some magic secret we would put it in this manual.
6616 Suggestions for improving the manual are always welcome, of course.
6618 @node Debugging GDB,,,Hints
6620 @section Debugging @value{GDBN} with itself
6621 @cindex debugging @value{GDBN}
6623 If @value{GDBN} is limping on your machine, this is the preferred way to get it
6624 fully functional. Be warned that in some ancient Unix systems, like
6625 Ultrix 4.2, a program can't be running in one process while it is being
6626 debugged in another. Rather than typing the command @kbd{@w{./gdb
6627 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
6628 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
6630 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
6631 @file{.gdbinit} file that sets up some simple things to make debugging
6632 gdb easier. The @code{info} command, when executed without a subcommand
6633 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
6634 gdb. See @file{.gdbinit} for details.
6636 If you use emacs, you will probably want to do a @code{make TAGS} after
6637 you configure your distribution; this will put the machine dependent
6638 routines for your local machine where they will be accessed first by
6641 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
6642 have run @code{fixincludes} if you are compiling with gcc.
6644 @section Submitting Patches
6646 @cindex submitting patches
6647 Thanks for thinking of offering your changes back to the community of
6648 @value{GDBN} users. In general we like to get well designed enhancements.
6649 Thanks also for checking in advance about the best way to transfer the
6652 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
6653 This manual summarizes what we believe to be clean design for @value{GDBN}.
6655 If the maintainers don't have time to put the patch in when it arrives,
6656 or if there is any question about a patch, it goes into a large queue
6657 with everyone else's patches and bug reports.
6659 @cindex legal papers for code contributions
6660 The legal issue is that to incorporate substantial changes requires a
6661 copyright assignment from you and/or your employer, granting ownership
6662 of the changes to the Free Software Foundation. You can get the
6663 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
6664 and asking for it. We recommend that people write in "All programs
6665 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
6666 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
6668 contributed with only one piece of legalese pushed through the
6669 bureaucracy and filed with the FSF. We can't start merging changes until
6670 this paperwork is received by the FSF (their rules, which we follow
6671 since we maintain it for them).
6673 Technically, the easiest way to receive changes is to receive each
6674 feature as a small context diff or unidiff, suitable for @code{patch}.
6675 Each message sent to me should include the changes to C code and
6676 header files for a single feature, plus @file{ChangeLog} entries for
6677 each directory where files were modified, and diffs for any changes
6678 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
6679 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
6680 single feature, they can be split down into multiple messages.
6682 In this way, if we read and like the feature, we can add it to the
6683 sources with a single patch command, do some testing, and check it in.
6684 If you leave out the @file{ChangeLog}, we have to write one. If you leave
6685 out the doc, we have to puzzle out what needs documenting. Etc., etc.
6687 The reason to send each change in a separate message is that we will not
6688 install some of the changes. They'll be returned to you with questions
6689 or comments. If we're doing our job correctly, the message back to you
6690 will say what you have to fix in order to make the change acceptable.
6691 The reason to have separate messages for separate features is so that
6692 the acceptable changes can be installed while one or more changes are
6693 being reworked. If multiple features are sent in a single message, we
6694 tend to not put in the effort to sort out the acceptable changes from
6695 the unacceptable, so none of the features get installed until all are
6698 If this sounds painful or authoritarian, well, it is. But we get a lot
6699 of bug reports and a lot of patches, and many of them don't get
6700 installed because we don't have the time to finish the job that the bug
6701 reporter or the contributor could have done. Patches that arrive
6702 complete, working, and well designed, tend to get installed on the day
6703 they arrive. The others go into a queue and get installed as time
6704 permits, which, since the maintainers have many demands to meet, may not
6705 be for quite some time.
6707 Please send patches directly to
6708 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} maintainers}.
6710 @section Obsolete Conditionals
6711 @cindex obsolete code
6713 Fragments of old code in @value{GDBN} sometimes reference or set the following
6714 configuration macros. They should not be used by new code, and old uses
6715 should be removed as those parts of the debugger are otherwise touched.
6718 @item STACK_END_ADDR
6719 This macro used to define where the end of the stack appeared, for use
6720 in interpreting core file formats that don't record this address in the
6721 core file itself. This information is now configured in BFD, and @value{GDBN}
6722 gets the info portably from there. The values in @value{GDBN}'s configuration
6723 files should be moved into BFD configuration files (if needed there),
6724 and deleted from all of @value{GDBN}'s config files.
6726 Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR
6727 is so old that it has never been converted to use BFD. Now that's old!
6731 @include observer.texi