1 \input texinfo @c -*- texinfo -*-
2 @setfilename gdbint.info
4 @dircategory Programming & development tools.
6 * Gdb-Internals: (gdbint). The GNU debugger's internals.
10 This file documents the internals of the GNU debugger @value{GDBN}.
11 Copyright 1990,1991,1992,1993,1994,1996,1998,1999,2000,2001,2002
12 Free Software Foundation, Inc.
13 Contributed by Cygnus Solutions. Written by John Gilmore.
14 Second Edition by Stan Shebs.
16 Permission is granted to copy, distribute and/or modify this document
17 under the terms of the GNU Free Documentation License, Version 1.1 or
18 any later version published by the Free Software Foundation; with no
19 Invariant Sections, with the Front-Cover Texts being ``A GNU Manual,''
20 and with the Back-Cover Texts as in (a) below.
22 (a) The FSF's Back-Cover Text is: ``You have freedom to copy and modify
23 this GNU Manual, like GNU software. Copies published by the Free
24 Software Foundation raise funds for GNU development.''
27 @setchapternewpage off
28 @settitle @value{GDBN} Internals
34 @title @value{GDBN} Internals
35 @subtitle{A guide to the internals of the GNU debugger}
37 @author Cygnus Solutions
38 @author Second Edition:
40 @author Cygnus Solutions
43 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
44 \xdef\manvers{\$Revision$} % For use in headers, footers too
46 \hfill Cygnus Solutions\par
48 \hfill \TeX{}info \texinfoversion\par
52 @vskip 0pt plus 1filll
53 Copyright @copyright{} 1990,1991,1992,1993,1994,1996,1998,1999,2000,2001
54 Free Software Foundation, Inc.
56 Permission is granted to copy, distribute and/or modify this document
57 under the terms of the GNU Free Documentation License, Version 1.1 or
58 any later version published by the Free Software Foundation; with no
59 Invariant Sections, with the Front-Cover Texts being ``A GNU Manual,''
60 and with the Back-Cover Texts as in (a) below.
62 (a) The FSF's Back-Cover Text is: ``You have freedom to copy and modify
63 this GNU Manual, like GNU software. Copies published by the Free
64 Software Foundation raise funds for GNU development.''
67 @c TeX can handle the contents at the start but makeinfo 3.12 can not
73 @c Perhaps this should be the title of the document (but only for info,
74 @c not for TeX). Existing GNU manuals seem inconsistent on this point.
75 @top Scope of this Document
77 This document documents the internals of the GNU debugger, @value{GDBN}. It
78 includes description of @value{GDBN}'s key algorithms and operations, as well
79 as the mechanisms that adapt @value{GDBN} to specific hosts and targets.
90 * Target Architecture Definition::
91 * Target Vector Definition::
104 @chapter Requirements
105 @cindex requirements for @value{GDBN}
107 Before diving into the internals, you should understand the formal
108 requirements and other expectations for @value{GDBN}. Although some
109 of these may seem obvious, there have been proposals for @value{GDBN}
110 that have run counter to these requirements.
112 First of all, @value{GDBN} is a debugger. It's not designed to be a
113 front panel for embedded systems. It's not a text editor. It's not a
114 shell. It's not a programming environment.
116 @value{GDBN} is an interactive tool. Although a batch mode is
117 available, @value{GDBN}'s primary role is to interact with a human
120 @value{GDBN} should be responsive to the user. A programmer hot on
121 the trail of a nasty bug, and operating under a looming deadline, is
122 going to be very impatient of everything, including the response time
123 to debugger commands.
125 @value{GDBN} should be relatively permissive, such as for expressions.
126 While the compiler should be picky (or have the option to be made
127 picky), since source code lives for a long time usually, the
128 programmer doing debugging shouldn't be spending time figuring out to
129 mollify the debugger.
131 @value{GDBN} will be called upon to deal with really large programs.
132 Executable sizes of 50 to 100 megabytes occur regularly, and we've
133 heard reports of programs approaching 1 gigabyte in size.
135 @value{GDBN} should be able to run everywhere. No other debugger is
136 available for even half as many configurations as @value{GDBN}
140 @node Overall Structure
142 @chapter Overall Structure
144 @value{GDBN} consists of three major subsystems: user interface,
145 symbol handling (the @dfn{symbol side}), and target system handling (the
148 The user interface consists of several actual interfaces, plus
151 The symbol side consists of object file readers, debugging info
152 interpreters, symbol table management, source language expression
153 parsing, type and value printing.
155 The target side consists of execution control, stack frame analysis, and
156 physical target manipulation.
158 The target side/symbol side division is not formal, and there are a
159 number of exceptions. For instance, core file support involves symbolic
160 elements (the basic core file reader is in BFD) and target elements (it
161 supplies the contents of memory and the values of registers). Instead,
162 this division is useful for understanding how the minor subsystems
165 @section The Symbol Side
167 The symbolic side of @value{GDBN} can be thought of as ``everything
168 you can do in @value{GDBN} without having a live program running''.
169 For instance, you can look at the types of variables, and evaluate
170 many kinds of expressions.
172 @section The Target Side
174 The target side of @value{GDBN} is the ``bits and bytes manipulator''.
175 Although it may make reference to symbolic info here and there, most
176 of the target side will run with only a stripped executable
177 available---or even no executable at all, in remote debugging cases.
179 Operations such as disassembly, stack frame crawls, and register
180 display, are able to work with no symbolic info at all. In some cases,
181 such as disassembly, @value{GDBN} will use symbolic info to present addresses
182 relative to symbols rather than as raw numbers, but it will work either
185 @section Configurations
189 @dfn{Host} refers to attributes of the system where @value{GDBN} runs.
190 @dfn{Target} refers to the system where the program being debugged
191 executes. In most cases they are the same machine, in which case a
192 third type of @dfn{Native} attributes come into play.
194 Defines and include files needed to build on the host are host support.
195 Examples are tty support, system defined types, host byte order, host
198 Defines and information needed to handle the target format are target
199 dependent. Examples are the stack frame format, instruction set,
200 breakpoint instruction, registers, and how to set up and tear down the stack
203 Information that is only needed when the host and target are the same,
204 is native dependent. One example is Unix child process support; if the
205 host and target are not the same, doing a fork to start the target
206 process is a bad idea. The various macros needed for finding the
207 registers in the @code{upage}, running @code{ptrace}, and such are all
208 in the native-dependent files.
210 Another example of native-dependent code is support for features that
211 are really part of the target environment, but which require
212 @code{#include} files that are only available on the host system. Core
213 file handling and @code{setjmp} handling are two common cases.
215 When you want to make @value{GDBN} work ``native'' on a particular machine, you
216 have to include all three kinds of information.
224 @value{GDBN} uses a number of debugging-specific algorithms. They are
225 often not very complicated, but get lost in the thicket of special
226 cases and real-world issues. This chapter describes the basic
227 algorithms and mentions some of the specific target definitions that
233 @cindex call stack frame
234 A frame is a construct that @value{GDBN} uses to keep track of calling
235 and called functions.
237 @findex create_new_frame
239 @code{FRAME_FP} in the machine description has no meaning to the
240 machine-independent part of @value{GDBN}, except that it is used when
241 setting up a new frame from scratch, as follows:
244 create_new_frame (read_register (FP_REGNUM), read_pc ()));
247 @cindex frame pointer register
248 Other than that, all the meaning imparted to @code{FP_REGNUM} is
249 imparted by the machine-dependent code. So, @code{FP_REGNUM} can have
250 any value that is convenient for the code that creates new frames.
251 (@code{create_new_frame} calls @code{INIT_EXTRA_FRAME_INFO} if it is
252 defined; that is where you should use the @code{FP_REGNUM} value, if
253 your frames are nonstandard.)
256 Given a @value{GDBN} frame, define @code{FRAME_CHAIN} to determine the
257 address of the calling function's frame. This will be used to create
258 a new @value{GDBN} frame struct, and then @code{INIT_EXTRA_FRAME_INFO}
259 and @code{INIT_FRAME_PC} will be called for the new frame.
261 @section Breakpoint Handling
264 In general, a breakpoint is a user-designated location in the program
265 where the user wants to regain control if program execution ever reaches
268 There are two main ways to implement breakpoints; either as ``hardware''
269 breakpoints or as ``software'' breakpoints.
271 @cindex hardware breakpoints
272 @cindex program counter
273 Hardware breakpoints are sometimes available as a builtin debugging
274 features with some chips. Typically these work by having dedicated
275 register into which the breakpoint address may be stored. If the PC
276 (shorthand for @dfn{program counter})
277 ever matches a value in a breakpoint registers, the CPU raises an
278 exception and reports it to @value{GDBN}.
280 Another possibility is when an emulator is in use; many emulators
281 include circuitry that watches the address lines coming out from the
282 processor, and force it to stop if the address matches a breakpoint's
285 A third possibility is that the target already has the ability to do
286 breakpoints somehow; for instance, a ROM monitor may do its own
287 software breakpoints. So although these are not literally ``hardware
288 breakpoints'', from @value{GDBN}'s point of view they work the same;
289 @value{GDBN} need not do nothing more than set the breakpoint and wait
290 for something to happen.
292 Since they depend on hardware resources, hardware breakpoints may be
293 limited in number; when the user asks for more, @value{GDBN} will
294 start trying to set software breakpoints. (On some architectures,
295 notably the 32-bit x86 platforms, @value{GDBN} cannot alsways know
296 whether there's enough hardware resources to insert all the hardware
297 breakpoints and watchpoints. On those platforms, @value{GDBN} prints
298 an error message only when the program being debugged is continued.)
300 @cindex software breakpoints
301 Software breakpoints require @value{GDBN} to do somewhat more work.
302 The basic theory is that @value{GDBN} will replace a program
303 instruction with a trap, illegal divide, or some other instruction
304 that will cause an exception, and then when it's encountered,
305 @value{GDBN} will take the exception and stop the program. When the
306 user says to continue, @value{GDBN} will restore the original
307 instruction, single-step, re-insert the trap, and continue on.
309 Since it literally overwrites the program being tested, the program area
310 must be writable, so this technique won't work on programs in ROM. It
311 can also distort the behavior of programs that examine themselves,
312 although such a situation would be highly unusual.
314 Also, the software breakpoint instruction should be the smallest size of
315 instruction, so it doesn't overwrite an instruction that might be a jump
316 target, and cause disaster when the program jumps into the middle of the
317 breakpoint instruction. (Strictly speaking, the breakpoint must be no
318 larger than the smallest interval between instructions that may be jump
319 targets; perhaps there is an architecture where only even-numbered
320 instructions may jumped to.) Note that it's possible for an instruction
321 set not to have any instructions usable for a software breakpoint,
322 although in practice only the ARC has failed to define such an
326 The basic definition of the software breakpoint is the macro
329 Basic breakpoint object handling is in @file{breakpoint.c}. However,
330 much of the interesting breakpoint action is in @file{infrun.c}.
332 @section Single Stepping
334 @section Signal Handling
336 @section Thread Handling
338 @section Inferior Function Calls
340 @section Longjmp Support
342 @cindex @code{longjmp} debugging
343 @value{GDBN} has support for figuring out that the target is doing a
344 @code{longjmp} and for stopping at the target of the jump, if we are
345 stepping. This is done with a few specialized internal breakpoints,
346 which are visible in the output of the @samp{maint info breakpoint}
349 @findex GET_LONGJMP_TARGET
350 To make this work, you need to define a macro called
351 @code{GET_LONGJMP_TARGET}, which will examine the @code{jmp_buf}
352 structure and extract the longjmp target address. Since @code{jmp_buf}
353 is target specific, you will need to define it in the appropriate
354 @file{tm-@var{target}.h} file. Look in @file{tm-sun4os4.h} and
355 @file{sparc-tdep.c} for examples of how to do this.
360 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
361 breakpoints}) which break when data is accessed rather than when some
362 instruction is executed. When you have data which changes without
363 your knowing what code does that, watchpoints are the silver bullet to
364 hunt down and kill such bugs.
366 @cindex hardware watchpoints
367 @cindex software watchpoints
368 Watchpoints can be either hardware-assisted or not; the latter type is
369 known as ``software watchpoints.'' @value{GDBN} always uses
370 hardware-assisted watchpoints if they are available, and falls back on
371 software watchpoints otherwise. Typical situations where @value{GDBN}
372 will use software watchpoints are:
376 The watched memory region is too large for the underlying hardware
377 watchpoint support. For example, each x86 debug register can watch up
378 to 4 bytes of memory, so trying to watch data structures whose size is
379 more than 16 bytes will cause @value{GDBN} to use software
383 The value of the expression to be watched depends on data held in
384 registers (as opposed to memory).
387 Too many different watchpoints requested. (On some architectures,
388 this situation is impossible to detect until the debugged program is
389 resumed.) Note that x86 debug registers are used both for hardware
390 breakpoints and for watchpoints, so setting too many hardware
391 breakpoints might cause watchpoint insertion to fail.
394 No hardware-assisted watchpoints provided by the target
398 Software watchpoints are very slow, since @value{GDBN} needs to
399 single-step the program being debugged and test the value of the
400 watched expression(s) after each instruction. The rest of this
401 section is mostly irrelevant for software watchpoints.
403 @value{GDBN} uses several macros and primitives to support hardware
407 @findex TARGET_HAS_HARDWARE_WATCHPOINTS
408 @item TARGET_HAS_HARDWARE_WATCHPOINTS
409 If defined, the target supports hardware watchpoints.
411 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
412 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
413 Return the number of hardware watchpoints of type @var{type} that are
414 possible to be set. The value is positive if @var{count} watchpoints
415 of this type can be set, zero if setting watchpoints of this type is
416 not supported, and negative if @var{count} is more than the maximum
417 number of watchpoints of type @var{type} that can be set. @var{other}
418 is non-zero if other types of watchpoints are currently enabled (there
419 are architectures which cannot set watchpoints of different types at
422 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
423 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
424 Return non-zero if hardware watchpoints can be used to watch a region
425 whose address is @var{addr} and whose length in bytes is @var{len}.
427 @findex TARGET_REGION_SIZE_OK_FOR_HW_WATCHPOINT
428 @item TARGET_REGION_SIZE_OK_FOR_HW_WATCHPOINT (@var{size})
429 Return non-zero if hardware watchpoints can be used to watch a region
430 whose size is @var{size}. @value{GDBN} only uses this macro as a
431 fall-back, in case @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is not
434 @findex TARGET_DISABLE_HW_WATCHPOINTS
435 @item TARGET_DISABLE_HW_WATCHPOINTS (@var{pid})
436 Disables watchpoints in the process identified by @var{pid}. This is
437 used, e.g., on HP-UX which provides operations to disable and enable
438 the page-level memory protection that implements hardware watchpoints
441 @findex TARGET_ENABLE_HW_WATCHPOINTS
442 @item TARGET_ENABLE_HW_WATCHPOINTS (@var{pid})
443 Enables watchpoints in the process identified by @var{pid}. This is
444 used, e.g., on HP-UX which provides operations to disable and enable
445 the page-level memory protection that implements hardware watchpoints
448 @findex target_insert_watchpoint
449 @findex target_remove_watchpoint
450 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
451 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
452 Insert or remove a hardware watchpoint starting at @var{addr}, for
453 @var{len} bytes. @var{type} is the watchpoint type, one of the
454 possible values of the enumerated data type @code{target_hw_bp_type},
455 defined by @file{breakpoint.h} as follows:
458 enum target_hw_bp_type
460 hw_write = 0, /* Common (write) HW watchpoint */
461 hw_read = 1, /* Read HW watchpoint */
462 hw_access = 2, /* Access (read or write) HW watchpoint */
463 hw_execute = 3 /* Execute HW breakpoint */
468 These two macros should return 0 for success, non-zero for failure.
470 @cindex insert or remove hardware breakpoint
471 @findex target_remove_hw_breakpoint
472 @findex target_insert_hw_breakpoint
473 @item target_remove_hw_breakpoint (@var{addr}, @var{shadow})
474 @itemx target_insert_hw_breakpoint (@var{addr}, @var{shadow})
475 Insert or remove a hardware-assisted breakpoint at address @var{addr}.
476 Returns zero for success, non-zero for failure. @var{shadow} is the
477 real contents of the byte where the breakpoint has been inserted; it
478 is generally not valid when hardware breakpoints are used, but since
479 no other code touches these values, the implementations of the above
480 two macros can use them for their internal purposes.
482 @findex target_stopped_data_address
483 @item target_stopped_data_address ()
484 If the inferior has some watchpoint that triggered, return the address
485 associated with that watchpoint. Otherwise, return zero.
487 @findex DECR_PC_AFTER_HW_BREAK
488 @item DECR_PC_AFTER_HW_BREAK
489 If defined, @value{GDBN} decrements the program counter by the value
490 of @code{DECR_PC_AFTER_HW_BREAK} after a hardware break-point. This
491 overrides the value of @code{DECR_PC_AFTER_BREAK} when a breakpoint
492 that breaks is a hardware-assisted breakpoint.
494 @findex HAVE_STEPPABLE_WATCHPOINT
495 @item HAVE_STEPPABLE_WATCHPOINT
496 If defined to a non-zero value, it is not necessary to disable a
497 watchpoint to step over it.
499 @findex HAVE_NONSTEPPABLE_WATCHPOINT
500 @item HAVE_NONSTEPPABLE_WATCHPOINT
501 If defined to a non-zero value, @value{GDBN} should disable a
502 watchpoint to step the inferior over it.
504 @findex HAVE_CONTINUABLE_WATCHPOINT
505 @item HAVE_CONTINUABLE_WATCHPOINT
506 If defined to a non-zero value, it is possible to continue the
507 inferior after a watchpoint has been hit.
509 @findex CANNOT_STEP_HW_WATCHPOINTS
510 @item CANNOT_STEP_HW_WATCHPOINTS
511 If this is defined to a non-zero value, @value{GDBN} will remove all
512 watchpoints before stepping the inferior.
514 @findex STOPPED_BY_WATCHPOINT
515 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
516 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
517 the type @code{struct target_waitstatus}, defined by @file{target.h}.
520 @subsection x86 Watchpoints
521 @cindex x86 debug registers
522 @cindex watchpoints, on x86
524 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
525 registers designed to facilitate debugging. @value{GDBN} provides a
526 generic library of functions that x86-based ports can use to implement
527 support for watchpoints and hardware-assisted breakpoints. This
528 subsection documents the x86 watchpoint facilities in @value{GDBN}.
530 To use the generic x86 watchpoint support, a port should do the
534 @findex I386_USE_GENERIC_WATCHPOINTS
536 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
537 target-dependent headers.
540 Include the @file{config/i386/nm-i386.h} header file @emph{after}
541 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
544 Add @file{i386-nat.o} to the value of the Make variable
545 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}) or
546 @code{TDEPFILES} (@pxref{Target Architecture Definition, TDEPFILES}).
549 Provide implementations for the @code{I386_DR_LOW_*} macros described
550 below. Typically, each macro should call a target-specific function
551 which does the real work.
554 The x86 watchpoint support works by maintaining mirror images of the
555 debug registers. Values are copied between the mirror images and the
556 real debug registers via a set of macros which each target needs to
560 @findex I386_DR_LOW_SET_CONTROL
561 @item I386_DR_LOW_SET_CONTROL (@var{val})
562 Set the Debug Control (DR7) register to the value @var{val}.
564 @findex I386_DR_LOW_SET_ADDR
565 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
566 Put the address @var{addr} into the debug register number @var{idx}.
568 @findex I386_DR_LOW_RESET_ADDR
569 @item I386_DR_LOW_RESET_ADDR (@var{idx})
570 Reset (i.e.@: zero out) the address stored in the debug register
573 @findex I386_DR_LOW_GET_STATUS
574 @item I386_DR_LOW_GET_STATUS
575 Return the value of the Debug Status (DR6) register. This value is
576 used immediately after it is returned by
577 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
581 For each one of the 4 debug registers (whose indices are from 0 to 3)
582 that store addresses, a reference count is maintained by @value{GDBN},
583 to allow sharing of debug registers by several watchpoints. This
584 allows users to define several watchpoints that watch the same
585 expression, but with different conditions and/or commands, without
586 wasting debug registers which are in short supply. @value{GDBN}
587 maintains the reference counts internally, targets don't have to do
588 anything to use this feature.
590 The x86 debug registers can each watch a region that is 1, 2, or 4
591 bytes long. The ia32 architecture requires that each watched region
592 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
593 region on 4-byte boundary. However, the x86 watchpoint support in
594 @value{GDBN} can watch unaligned regions and regions larger than 4
595 bytes (up to 16 bytes) by allocating several debug registers to watch
596 a single region. This allocation of several registers per a watched
597 region is also done automatically without target code intervention.
599 The generic x86 watchpoint support provides the following API for the
600 @value{GDBN}'s application code:
603 @findex i386_region_ok_for_watchpoint
604 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
605 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
606 this function. It counts the number of debug registers required to
607 watch a given region, and returns a non-zero value if that number is
608 less than 4, the number of debug registers available to x86
611 @findex i386_stopped_data_address
612 @item i386_stopped_data_address (void)
613 The macros @code{STOPPED_BY_WATCHPOINT} and
614 @code{target_stopped_data_address} are set to call this function. The
615 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
616 function examines the breakpoint condition bits in the DR6 Debug
617 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
618 macro, and returns the address associated with the first bit that is
621 @findex i386_insert_watchpoint
622 @findex i386_remove_watchpoint
623 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
624 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
625 Insert or remove a watchpoint. The macros
626 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
627 are set to call these functions. @code{i386_insert_watchpoint} first
628 looks for a debug register which is already set to watch the same
629 region for the same access types; if found, it just increments the
630 reference count of that debug register, thus implementing debug
631 register sharing between watchpoints. If no such register is found,
632 the function looks for a vacant debug register, sets its mirrorred
633 value to @var{addr}, sets the mirrorred value of DR7 Debug Control
634 register as appropriate for the @var{len} and @var{type} parameters,
635 and then passes the new values of the debug register and DR7 to the
636 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
637 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
638 required to cover the given region, the above process is repeated for
641 @code{i386_remove_watchpoint} does the opposite: it resets the address
642 in the mirrorred value of the debug register and its read/write and
643 length bits in the mirrorred value of DR7, then passes these new
644 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
645 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
646 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
647 decrements the reference count, and only calls
648 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
649 the count goes to zero.
651 @findex i386_insert_hw_breakpoint
652 @findex i386_remove_hw_breakpoint
653 @item i386_insert_hw_breakpoint (@var{addr}, @var{shadow}
654 @itemx i386_remove_hw_breakpoint (@var{addr}, @var{shadow})
655 These functions insert and remove hardware-assisted breakpoints. The
656 macros @code{target_insert_hw_breakpoint} and
657 @code{target_remove_hw_breakpoint} are set to call these functions.
658 These functions work like @code{i386_insert_watchpoint} and
659 @code{i386_remove_watchpoint}, respectively, except that they set up
660 the debug registers to watch instruction execution, and each
661 hardware-assisted breakpoint always requires exactly one debug
664 @findex i386_stopped_by_hwbp
665 @item i386_stopped_by_hwbp (void)
666 This function returns non-zero if the inferior has some watchpoint or
667 hardware breakpoint that triggered. It works like
668 @code{i386_stopped_data_address}, except that it doesn't return the
669 address whose watchpoint triggered.
671 @findex i386_cleanup_dregs
672 @item i386_cleanup_dregs (void)
673 This function clears all the reference counts, addresses, and control
674 bits in the mirror images of the debug registers. It doesn't affect
675 the actual debug registers in the inferior process.
682 x86 processors support setting watchpoints on I/O reads or writes.
683 However, since no target supports this (as of March 2001), and since
684 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
685 watchpoints, this feature is not yet available to @value{GDBN} running
689 x86 processors can enable watchpoints locally, for the current task
690 only, or globally, for all the tasks. For each debug register,
691 there's a bit in the DR7 Debug Control register that determines
692 whether the associated address is watched locally or globally. The
693 current implementation of x86 watchpoint support in @value{GDBN}
694 always sets watchpoints to be locally enabled, since global
695 watchpoints might interfere with the underlying OS and are probably
696 unavailable in many platforms.
701 @chapter User Interface
703 @value{GDBN} has several user interfaces. Although the command-line interface
704 is the most common and most familiar, there are others.
706 @section Command Interpreter
708 @cindex command interpreter
710 The command interpreter in @value{GDBN} is fairly simple. It is designed to
711 allow for the set of commands to be augmented dynamically, and also
712 has a recursive subcommand capability, where the first argument to
713 a command may itself direct a lookup on a different command list.
715 For instance, the @samp{set} command just starts a lookup on the
716 @code{setlist} command list, while @samp{set thread} recurses
717 to the @code{set_thread_cmd_list}.
721 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
722 the main command list, and should be used for those commands. The usual
723 place to add commands is in the @code{_initialize_@var{xyz}} routines at
724 the ends of most source files.
726 @cindex deprecating commands
727 @findex deprecate_cmd
728 Before removing commands from the command set it is a good idea to
729 deprecate them for some time. Use @code{deprecate_cmd} on commands or
730 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
731 @code{struct cmd_list_element} as it's first argument. You can use the
732 return value from @code{add_com} or @code{add_cmd} to deprecate the
733 command immediately after it is created.
735 The first time a command is used the user will be warned and offered a
736 replacement (if one exists). Note that the replacement string passed to
737 @code{deprecate_cmd} should be the full name of the command, i.e. the
738 entire string the user should type at the command line.
740 @section UI-Independent Output---the @code{ui_out} Functions
741 @c This section is based on the documentation written by Fernando
742 @c Nasser <fnasser@redhat.com>.
744 @cindex @code{ui_out} functions
745 The @code{ui_out} functions present an abstraction level for the
746 @value{GDBN} output code. They hide the specifics of different user
747 interfaces supported by @value{GDBN}, and thus free the programmer
748 from the need to write several versions of the same code, one each for
749 every UI, to produce output.
751 @subsection Overview and Terminology
753 In general, execution of each @value{GDBN} command produces some sort
754 of output, and can even generate an input request.
756 Output can be generated for the following purposes:
760 to display a @emph{result} of an operation;
763 to convey @emph{info} or produce side-effects of a requested
767 to provide a @emph{notification} of an asynchronous event (including
768 progress indication of a prolonged asynchronous operation);
771 to display @emph{error messages} (including warnings);
774 to show @emph{debug data};
777 to @emph{query} or prompt a user for input (a special case).
781 This section mainly concentrates on how to build result output,
782 although some of it also applies to other kinds of output.
784 Generation of output that displays the results of an operation
785 involves one or more of the following:
789 output of the actual data
792 formatting the output as appropriate for console output, to make it
793 easily readable by humans
796 machine oriented formatting--a more terse formatting to allow for easy
797 parsing by programs which read @value{GDBN}'s output
800 annotation, whose purpose is to help legacy GUIs to identify interesting
804 The @code{ui_out} routines take care of the first three aspects.
805 Annotations are provided by separate annotation routines. Note that use
806 of annotations for an interface between a GUI and @value{GDBN} is
809 Output can be in the form of a single item, which we call a @dfn{field};
810 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
811 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
812 header and a body. In a BNF-like form:
815 @item <table> @expansion{}
816 @code{<header> <body>}
817 @item <header> @expansion{}
818 @code{@{ <column> @}}
819 @item <column> @expansion{}
820 @code{<width> <alignment> <title>}
821 @item <body> @expansion{}
826 @subsection General Conventions
828 Most @code{ui_out} routines are of type @code{void}, the exceptions are
829 @code{ui_out_stream_new} (which returns a pointer to the newly created
830 object) and the @code{make_cleanup} routines.
832 The first parameter is always the @code{ui_out} vector object, a pointer
833 to a @code{struct ui_out}.
835 The @var{format} parameter is like in @code{printf} family of functions.
836 When it is present, there must also be a variable list of arguments
837 sufficient used to satisfy the @code{%} specifiers in the supplied
840 When a character string argument is not used in a @code{ui_out} function
841 call, a @code{NULL} pointer has to be supplied instead.
844 @subsection Table, Tuple and List Functions
846 @cindex list output functions
847 @cindex table output functions
848 @cindex tuple output functions
849 This section introduces @code{ui_out} routines for building lists,
850 tuples and tables. The routines to output the actual data items
851 (fields) are presented in the next section.
853 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
854 containing information about an object; a @dfn{list} is a sequence of
855 fields where each field describes an identical object.
857 Use the @dfn{table} functions when your output consists of a list of
858 rows (tuples) and the console output should include a heading. Use this
859 even when you are listing just one object but you still want the header.
861 @cindex nesting level in @code{ui_out} functions
862 Tables can not be nested. Tuples and lists can be nested up to a
863 maximum of five levels.
865 The overall structure of the table output code is something like this:
880 Here is the description of table-, tuple- and list-related @code{ui_out}
883 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
884 The function @code{ui_out_table_begin} marks the beginning of the output
885 of a table. It should always be called before any other @code{ui_out}
886 function for a given table. @var{nbrofcols} is the number of columns in
887 the table. @var{nr_rows} is the number of rows in the table.
888 @var{tblid} is an optional string identifying the table. The string
889 pointed to by @var{tblid} is copied by the implementation of
890 @code{ui_out_table_begin}, so the application can free the string if it
893 The companion function @code{ui_out_table_end}, described below, marks
894 the end of the table's output.
897 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
898 @code{ui_out_table_header} provides the header information for a single
899 table column. You call this function several times, one each for every
900 column of the table, after @code{ui_out_table_begin}, but before
901 @code{ui_out_table_body}.
903 The value of @var{width} gives the column width in characters. The
904 value of @var{alignment} is one of @code{left}, @code{center}, and
905 @code{right}, and it specifies how to align the header: left-justify,
906 center, or right-justify it. @var{colhdr} points to a string that
907 specifies the column header; the implementation copies that string, so
908 column header strings in @code{malloc}ed storage can be freed after the
912 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
913 This function delimits the table header from the table body.
916 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
917 This function signals the end of a table's output. It should be called
918 after the table body has been produced by the list and field output
921 There should be exactly one call to @code{ui_out_table_end} for each
922 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
923 will signal an internal error.
926 The output of the tuples that represent the table rows must follow the
927 call to @code{ui_out_table_body} and precede the call to
928 @code{ui_out_table_end}. You build a tuple by calling
929 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
930 calls to functions which actually output fields between them.
932 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
933 This function marks the beginning of a tuple output. @var{id} points
934 to an optional string that identifies the tuple; it is copied by the
935 implementation, and so strings in @code{malloc}ed storage can be freed
939 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
940 This function signals an end of a tuple output. There should be exactly
941 one call to @code{ui_out_tuple_end} for each call to
942 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
946 @deftypefun struct cleanup *make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
947 This function first opens the tuple and then establishes a cleanup
948 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
949 and correct implementation of the non-portable@footnote{The function
950 cast is not portable ISO-C.} code sequence:
952 struct cleanup *old_cleanup;
953 ui_out_tuple_begin (uiout, "...");
954 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
959 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
960 This function marks the beginning of a list output. @var{id} points to
961 an optional string that identifies the list; it is copied by the
962 implementation, and so strings in @code{malloc}ed storage can be freed
966 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
967 This function signals an end of a list output. There should be exactly
968 one call to @code{ui_out_list_end} for each call to
969 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
973 @deftypefun struct cleanup *make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
974 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
975 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
976 that will close the list.list.
979 @subsection Item Output Functions
981 @cindex item output functions
982 @cindex field output functions
984 The functions described below produce output for the actual data
985 items, or fields, which contain information about the object.
987 Choose the appropriate function accordingly to your particular needs.
989 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
990 This is the most general output function. It produces the
991 representation of the data in the variable-length argument list
992 according to formatting specifications in @var{format}, a
993 @code{printf}-like format string. The optional argument @var{fldname}
994 supplies the name of the field. The data items themselves are
995 supplied as additional arguments after @var{format}.
997 This generic function should be used only when it is not possible to
998 use one of the specialized versions (see below).
1001 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1002 This function outputs a value of an @code{int} variable. It uses the
1003 @code{"%d"} output conversion specification. @var{fldname} specifies
1004 the name of the field.
1007 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1008 This function outputs an address.
1011 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1012 This function outputs a string using the @code{"%s"} conversion
1016 Sometimes, there's a need to compose your output piece by piece using
1017 functions that operate on a stream, such as @code{value_print} or
1018 @code{fprintf_symbol_filtered}. These functions accept an argument of
1019 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1020 used to store the data stream used for the output. When you use one
1021 of these functions, you need a way to pass their results stored in a
1022 @code{ui_file} object to the @code{ui_out} functions. To this end,
1023 you first create a @code{ui_stream} object by calling
1024 @code{ui_out_stream_new}, pass the @code{stream} member of that
1025 @code{ui_stream} object to @code{value_print} and similar functions,
1026 and finally call @code{ui_out_field_stream} to output the field you
1027 constructed. When the @code{ui_stream} object is no longer needed,
1028 you should destroy it and free its memory by calling
1029 @code{ui_out_stream_delete}.
1031 @deftypefun struct ui_stream *ui_out_stream_new (struct ui_out *@var{uiout})
1032 This function creates a new @code{ui_stream} object which uses the
1033 same output methods as the @code{ui_out} object whose pointer is
1034 passed in @var{uiout}. It returns a pointer to the newly created
1035 @code{ui_stream} object.
1038 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1039 This functions destroys a @code{ui_stream} object specified by
1043 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1044 This function consumes all the data accumulated in
1045 @code{streambuf->stream} and outputs it like
1046 @code{ui_out_field_string} does. After a call to
1047 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1048 the stream is still valid and may be used for producing more fields.
1051 @strong{Important:} If there is any chance that your code could bail
1052 out before completing output generation and reaching the point where
1053 @code{ui_out_stream_delete} is called, it is necessary to set up a
1054 cleanup, to avoid leaking memory and other resources. Here's a
1055 skeleton code to do that:
1058 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1059 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1064 If the function already has the old cleanup chain set (for other kinds
1065 of cleanups), you just have to add your cleanup to it:
1068 mybuf = ui_out_stream_new (uiout);
1069 make_cleanup (ui_out_stream_delete, mybuf);
1072 Note that with cleanups in place, you should not call
1073 @code{ui_out_stream_delete} directly, or you would attempt to free the
1076 @subsection Utility Output Functions
1078 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1079 This function skips a field in a table. Use it if you have to leave
1080 an empty field without disrupting the table alignment. The argument
1081 @var{fldname} specifies a name for the (missing) filed.
1084 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1085 This function outputs the text in @var{string} in a way that makes it
1086 easy to be read by humans. For example, the console implementation of
1087 this method filters the text through a built-in pager, to prevent it
1088 from scrolling off the visible portion of the screen.
1090 Use this function for printing relatively long chunks of text around
1091 the actual field data: the text it produces is not aligned according
1092 to the table's format. Use @code{ui_out_field_string} to output a
1093 string field, and use @code{ui_out_message}, described below, to
1094 output short messages.
1097 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1098 This function outputs @var{nspaces} spaces. It is handy to align the
1099 text produced by @code{ui_out_text} with the rest of the table or
1103 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1104 This function produces a formatted message, provided that the current
1105 verbosity level is at least as large as given by @var{verbosity}. The
1106 current verbosity level is specified by the user with the @samp{set
1107 verbositylevel} command.@footnote{As of this writing (April 2001),
1108 setting verbosity level is not yet implemented, and is always returned
1109 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1110 argument more than zero will cause the message to never be printed.}
1113 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1114 This function gives the console output filter (a paging filter) a hint
1115 of where to break lines which are too long. Ignored for all other
1116 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1117 be printed to indent the wrapped text on the next line; it must remain
1118 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1119 explicit newline is produced by one of the other functions. If
1120 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1123 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1124 This function flushes whatever output has been accumulated so far, if
1125 the UI buffers output.
1129 @subsection Examples of Use of @code{ui_out} functions
1131 @cindex using @code{ui_out} functions
1132 @cindex @code{ui_out} functions, usage examples
1133 This section gives some practical examples of using the @code{ui_out}
1134 functions to generalize the old console-oriented code in
1135 @value{GDBN}. The examples all come from functions defined on the
1136 @file{breakpoints.c} file.
1138 This example, from the @code{breakpoint_1} function, shows how to
1141 The original code was:
1144 if (!found_a_breakpoint++)
1146 annotate_breakpoints_headers ();
1149 printf_filtered ("Num ");
1151 printf_filtered ("Type ");
1153 printf_filtered ("Disp ");
1155 printf_filtered ("Enb ");
1159 printf_filtered ("Address ");
1162 printf_filtered ("What\n");
1164 annotate_breakpoints_table ();
1168 Here's the new version:
1171 nr_printable_breakpoints = @dots{};
1174 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1176 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1178 if (nr_printable_breakpoints > 0)
1179 annotate_breakpoints_headers ();
1180 if (nr_printable_breakpoints > 0)
1182 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1183 if (nr_printable_breakpoints > 0)
1185 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1186 if (nr_printable_breakpoints > 0)
1188 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1189 if (nr_printable_breakpoints > 0)
1191 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1194 if (nr_printable_breakpoints > 0)
1196 if (TARGET_ADDR_BIT <= 32)
1197 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1199 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1201 if (nr_printable_breakpoints > 0)
1203 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1204 ui_out_table_body (uiout);
1205 if (nr_printable_breakpoints > 0)
1206 annotate_breakpoints_table ();
1209 This example, from the @code{print_one_breakpoint} function, shows how
1210 to produce the actual data for the table whose structure was defined
1211 in the above example. The original code was:
1216 printf_filtered ("%-3d ", b->number);
1218 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1219 || ((int) b->type != bptypes[(int) b->type].type))
1220 internal_error ("bptypes table does not describe type #%d.",
1222 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1224 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1226 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1230 This is the new version:
1234 ui_out_tuple_begin (uiout, "bkpt");
1236 ui_out_field_int (uiout, "number", b->number);
1238 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1239 || ((int) b->type != bptypes[(int) b->type].type))
1240 internal_error ("bptypes table does not describe type #%d.",
1242 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1244 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1246 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1250 This example, also from @code{print_one_breakpoint}, shows how to
1251 produce a complicated output field using the @code{print_expression}
1252 functions which requires a stream to be passed. It also shows how to
1253 automate stream destruction with cleanups. The original code was:
1257 print_expression (b->exp, gdb_stdout);
1263 struct ui_stream *stb = ui_out_stream_new (uiout);
1264 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1267 print_expression (b->exp, stb->stream);
1268 ui_out_field_stream (uiout, "what", local_stream);
1271 This example, also from @code{print_one_breakpoint}, shows how to use
1272 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1277 if (b->dll_pathname == NULL)
1278 printf_filtered ("<any library> ");
1280 printf_filtered ("library \"%s\" ", b->dll_pathname);
1287 if (b->dll_pathname == NULL)
1289 ui_out_field_string (uiout, "what", "<any library>");
1290 ui_out_spaces (uiout, 1);
1294 ui_out_text (uiout, "library \"");
1295 ui_out_field_string (uiout, "what", b->dll_pathname);
1296 ui_out_text (uiout, "\" ");
1300 The following example from @code{print_one_breakpoint} shows how to
1301 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1306 if (b->forked_inferior_pid != 0)
1307 printf_filtered ("process %d ", b->forked_inferior_pid);
1314 if (b->forked_inferior_pid != 0)
1316 ui_out_text (uiout, "process ");
1317 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1318 ui_out_spaces (uiout, 1);
1322 Here's an example of using @code{ui_out_field_string}. The original
1327 if (b->exec_pathname != NULL)
1328 printf_filtered ("program \"%s\" ", b->exec_pathname);
1335 if (b->exec_pathname != NULL)
1337 ui_out_text (uiout, "program \"");
1338 ui_out_field_string (uiout, "what", b->exec_pathname);
1339 ui_out_text (uiout, "\" ");
1343 Finally, here's an example of printing an address. The original code:
1347 printf_filtered ("%s ",
1348 local_hex_string_custom ((unsigned long) b->address, "08l"));
1355 ui_out_field_core_addr (uiout, "Address", b->address);
1359 @section Console Printing
1368 @cindex @code{libgdb}
1369 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1370 to provide an API to @value{GDBN}'s functionality.
1373 @cindex @code{libgdb}
1374 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1375 better able to support graphical and other environments.
1377 Since @code{libgdb} development is on-going, its architecture is still
1378 evolving. The following components have so far been identified:
1382 Observer - @file{gdb-events.h}.
1384 Builder - @file{ui-out.h}
1386 Event Loop - @file{event-loop.h}
1388 Library - @file{gdb.h}
1391 The model that ties these components together is described below.
1393 @section The @code{libgdb} Model
1395 A client of @code{libgdb} interacts with the library in two ways.
1399 As an observer (using @file{gdb-events}) receiving notifications from
1400 @code{libgdb} of any internal state changes (break point changes, run
1403 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1404 obtain various status values from @value{GDBN}.
1407 Since @code{libgdb} could have multiple clients (e.g. a GUI supporting
1408 the existing @value{GDBN} CLI), those clients must co-operate when
1409 controlling @code{libgdb}. In particular, a client must ensure that
1410 @code{libgdb} is idle (i.e. no other client is using @code{libgdb})
1411 before responding to a @file{gdb-event} by making a query.
1413 @section CLI support
1415 At present @value{GDBN}'s CLI is very much entangled in with the core of
1416 @code{libgdb}. Consequently, a client wishing to include the CLI in
1417 their interface needs to carefully co-ordinate its own and the CLI's
1420 It is suggested that the client set @code{libgdb} up to be bi-modal
1421 (alternate between CLI and client query modes). The notes below sketch
1426 The client registers itself as an observer of @code{libgdb}.
1428 The client create and install @code{cli-out} builder using its own
1429 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1430 @code{gdb_stdout} streams.
1432 The client creates a separate custom @code{ui-out} builder that is only
1433 used while making direct queries to @code{libgdb}.
1436 When the client receives input intended for the CLI, it simply passes it
1437 along. Since the @code{cli-out} builder is installed by default, all
1438 the CLI output in response to that command is routed (pronounced rooted)
1439 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1440 At the same time, the client is kept abreast of internal changes by
1441 virtue of being a @code{libgdb} observer.
1443 The only restriction on the client is that it must wait until
1444 @code{libgdb} becomes idle before initiating any queries (using the
1445 client's custom builder).
1447 @section @code{libgdb} components
1449 @subheading Observer - @file{gdb-events.h}
1450 @file{gdb-events} provides the client with a very raw mechanism that can
1451 be used to implement an observer. At present it only allows for one
1452 observer and that observer must, internally, handle the need to delay
1453 the processing of any event notifications until after @code{libgdb} has
1454 finished the current command.
1456 @subheading Builder - @file{ui-out.h}
1457 @file{ui-out} provides the infrastructure necessary for a client to
1458 create a builder. That builder is then passed down to @code{libgdb}
1459 when doing any queries.
1461 @subheading Event Loop - @file{event-loop.h}
1462 @c There could be an entire section on the event-loop
1463 @file{event-loop}, currently non-re-entrant, provides a simple event
1464 loop. A client would need to either plug its self into this loop or,
1465 implement a new event-loop that GDB would use.
1467 The event-loop will eventually be made re-entrant. This is so that
1468 @value{GDB} can better handle the problem of some commands blocking
1469 instead of returning.
1471 @subheading Library - @file{gdb.h}
1472 @file{libgdb} is the most obvious component of this system. It provides
1473 the query interface. Each function is parameterized by a @code{ui-out}
1474 builder. The result of the query is constructed using that builder
1475 before the query function returns.
1477 @node Symbol Handling
1479 @chapter Symbol Handling
1481 Symbols are a key part of @value{GDBN}'s operation. Symbols include variables,
1482 functions, and types.
1484 @section Symbol Reading
1486 @cindex symbol reading
1487 @cindex reading of symbols
1488 @cindex symbol files
1489 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
1490 file is the file containing the program which @value{GDBN} is
1491 debugging. @value{GDBN} can be directed to use a different file for
1492 symbols (with the @samp{symbol-file} command), and it can also read
1493 more symbols via the @samp{add-file} and @samp{load} commands, or while
1494 reading symbols from shared libraries.
1496 @findex find_sym_fns
1497 Symbol files are initially opened by code in @file{symfile.c} using
1498 the BFD library (@pxref{Support Libraries}). BFD identifies the type
1499 of the file by examining its header. @code{find_sym_fns} then uses
1500 this identification to locate a set of symbol-reading functions.
1502 @findex add_symtab_fns
1503 @cindex @code{sym_fns} structure
1504 @cindex adding a symbol-reading module
1505 Symbol-reading modules identify themselves to @value{GDBN} by calling
1506 @code{add_symtab_fns} during their module initialization. The argument
1507 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
1508 name (or name prefix) of the symbol format, the length of the prefix,
1509 and pointers to four functions. These functions are called at various
1510 times to process symbol files whose identification matches the specified
1513 The functions supplied by each module are:
1516 @item @var{xyz}_symfile_init(struct sym_fns *sf)
1518 @cindex secondary symbol file
1519 Called from @code{symbol_file_add} when we are about to read a new
1520 symbol file. This function should clean up any internal state (possibly
1521 resulting from half-read previous files, for example) and prepare to
1522 read a new symbol file. Note that the symbol file which we are reading
1523 might be a new ``main'' symbol file, or might be a secondary symbol file
1524 whose symbols are being added to the existing symbol table.
1526 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
1527 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
1528 new symbol file being read. Its @code{private} field has been zeroed,
1529 and can be modified as desired. Typically, a struct of private
1530 information will be @code{malloc}'d, and a pointer to it will be placed
1531 in the @code{private} field.
1533 There is no result from @code{@var{xyz}_symfile_init}, but it can call
1534 @code{error} if it detects an unavoidable problem.
1536 @item @var{xyz}_new_init()
1538 Called from @code{symbol_file_add} when discarding existing symbols.
1539 This function needs only handle the symbol-reading module's internal
1540 state; the symbol table data structures visible to the rest of
1541 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
1542 arguments and no result. It may be called after
1543 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
1544 may be called alone if all symbols are simply being discarded.
1546 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
1548 Called from @code{symbol_file_add} to actually read the symbols from a
1549 symbol-file into a set of psymtabs or symtabs.
1551 @code{sf} points to the @code{struct sym_fns} originally passed to
1552 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
1553 the offset between the file's specified start address and its true
1554 address in memory. @code{mainline} is 1 if this is the main symbol
1555 table being read, and 0 if a secondary symbol file (e.g. shared library
1556 or dynamically loaded file) is being read.@refill
1559 In addition, if a symbol-reading module creates psymtabs when
1560 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
1561 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
1562 from any point in the @value{GDBN} symbol-handling code.
1565 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
1567 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
1568 the psymtab has not already been read in and had its @code{pst->symtab}
1569 pointer set. The argument is the psymtab to be fleshed-out into a
1570 symtab. Upon return, @code{pst->readin} should have been set to 1, and
1571 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
1572 zero if there were no symbols in that part of the symbol file.
1575 @section Partial Symbol Tables
1577 @value{GDBN} has three types of symbol tables:
1580 @cindex full symbol table
1583 Full symbol tables (@dfn{symtabs}). These contain the main
1584 information about symbols and addresses.
1588 Partial symbol tables (@dfn{psymtabs}). These contain enough
1589 information to know when to read the corresponding part of the full
1592 @cindex minimal symbol table
1595 Minimal symbol tables (@dfn{msymtabs}). These contain information
1596 gleaned from non-debugging symbols.
1599 @cindex partial symbol table
1600 This section describes partial symbol tables.
1602 A psymtab is constructed by doing a very quick pass over an executable
1603 file's debugging information. Small amounts of information are
1604 extracted---enough to identify which parts of the symbol table will
1605 need to be re-read and fully digested later, when the user needs the
1606 information. The speed of this pass causes @value{GDBN} to start up very
1607 quickly. Later, as the detailed rereading occurs, it occurs in small
1608 pieces, at various times, and the delay therefrom is mostly invisible to
1610 @c (@xref{Symbol Reading}.)
1612 The symbols that show up in a file's psymtab should be, roughly, those
1613 visible to the debugger's user when the program is not running code from
1614 that file. These include external symbols and types, static symbols and
1615 types, and @code{enum} values declared at file scope.
1617 The psymtab also contains the range of instruction addresses that the
1618 full symbol table would represent.
1620 @cindex finding a symbol
1621 @cindex symbol lookup
1622 The idea is that there are only two ways for the user (or much of the
1623 code in the debugger) to reference a symbol:
1626 @findex find_pc_function
1627 @findex find_pc_line
1629 By its address (e.g. execution stops at some address which is inside a
1630 function in this file). The address will be noticed to be in the
1631 range of this psymtab, and the full symtab will be read in.
1632 @code{find_pc_function}, @code{find_pc_line}, and other
1633 @code{find_pc_@dots{}} functions handle this.
1635 @cindex lookup_symbol
1638 (e.g. the user asks to print a variable, or set a breakpoint on a
1639 function). Global names and file-scope names will be found in the
1640 psymtab, which will cause the symtab to be pulled in. Local names will
1641 have to be qualified by a global name, or a file-scope name, in which
1642 case we will have already read in the symtab as we evaluated the
1643 qualifier. Or, a local symbol can be referenced when we are ``in'' a
1644 local scope, in which case the first case applies. @code{lookup_symbol}
1645 does most of the work here.
1648 The only reason that psymtabs exist is to cause a symtab to be read in
1649 at the right moment. Any symbol that can be elided from a psymtab,
1650 while still causing that to happen, should not appear in it. Since
1651 psymtabs don't have the idea of scope, you can't put local symbols in
1652 them anyway. Psymtabs don't have the idea of the type of a symbol,
1653 either, so types need not appear, unless they will be referenced by
1656 It is a bug for @value{GDBN} to behave one way when only a psymtab has
1657 been read, and another way if the corresponding symtab has been read
1658 in. Such bugs are typically caused by a psymtab that does not contain
1659 all the visible symbols, or which has the wrong instruction address
1662 The psymtab for a particular section of a symbol file (objfile) could be
1663 thrown away after the symtab has been read in. The symtab should always
1664 be searched before the psymtab, so the psymtab will never be used (in a
1665 bug-free environment). Currently, psymtabs are allocated on an obstack,
1666 and all the psymbols themselves are allocated in a pair of large arrays
1667 on an obstack, so there is little to be gained by trying to free them
1668 unless you want to do a lot more work.
1672 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
1674 @cindex fundamental types
1675 These are the fundamental types that @value{GDBN} uses internally. Fundamental
1676 types from the various debugging formats (stabs, ELF, etc) are mapped
1677 into one of these. They are basically a union of all fundamental types
1678 that @value{GDBN} knows about for all the languages that @value{GDBN}
1681 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
1684 Each time @value{GDBN} builds an internal type, it marks it with one
1685 of these types. The type may be a fundamental type, such as
1686 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
1687 which is a pointer to another type. Typically, several @code{FT_*}
1688 types map to one @code{TYPE_CODE_*} type, and are distinguished by
1689 other members of the type struct, such as whether the type is signed
1690 or unsigned, and how many bits it uses.
1692 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
1694 These are instances of type structs that roughly correspond to
1695 fundamental types and are created as global types for @value{GDBN} to
1696 use for various ugly historical reasons. We eventually want to
1697 eliminate these. Note for example that @code{builtin_type_int}
1698 initialized in @file{gdbtypes.c} is basically the same as a
1699 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
1700 an @code{FT_INTEGER} fundamental type. The difference is that the
1701 @code{builtin_type} is not associated with any particular objfile, and
1702 only one instance exists, while @file{c-lang.c} builds as many
1703 @code{TYPE_CODE_INT} types as needed, with each one associated with
1704 some particular objfile.
1706 @section Object File Formats
1707 @cindex object file formats
1711 @cindex @code{a.out} format
1712 The @code{a.out} format is the original file format for Unix. It
1713 consists of three sections: @code{text}, @code{data}, and @code{bss},
1714 which are for program code, initialized data, and uninitialized data,
1717 The @code{a.out} format is so simple that it doesn't have any reserved
1718 place for debugging information. (Hey, the original Unix hackers used
1719 @samp{adb}, which is a machine-language debugger!) The only debugging
1720 format for @code{a.out} is stabs, which is encoded as a set of normal
1721 symbols with distinctive attributes.
1723 The basic @code{a.out} reader is in @file{dbxread.c}.
1728 The COFF format was introduced with System V Release 3 (SVR3) Unix.
1729 COFF files may have multiple sections, each prefixed by a header. The
1730 number of sections is limited.
1732 The COFF specification includes support for debugging. Although this
1733 was a step forward, the debugging information was woefully limited. For
1734 instance, it was not possible to represent code that came from an
1737 The COFF reader is in @file{coffread.c}.
1741 @cindex ECOFF format
1742 ECOFF is an extended COFF originally introduced for Mips and Alpha
1745 The basic ECOFF reader is in @file{mipsread.c}.
1749 @cindex XCOFF format
1750 The IBM RS/6000 running AIX uses an object file format called XCOFF.
1751 The COFF sections, symbols, and line numbers are used, but debugging
1752 symbols are @code{dbx}-style stabs whose strings are located in the
1753 @code{.debug} section (rather than the string table). For more
1754 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
1756 The shared library scheme has a clean interface for figuring out what
1757 shared libraries are in use, but the catch is that everything which
1758 refers to addresses (symbol tables and breakpoints at least) needs to be
1759 relocated for both shared libraries and the main executable. At least
1760 using the standard mechanism this can only be done once the program has
1761 been run (or the core file has been read).
1765 @cindex PE-COFF format
1766 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
1767 executables. PE is basically COFF with additional headers.
1769 While BFD includes special PE support, @value{GDBN} needs only the basic
1775 The ELF format came with System V Release 4 (SVR4) Unix. ELF is similar
1776 to COFF in being organized into a number of sections, but it removes
1777 many of COFF's limitations.
1779 The basic ELF reader is in @file{elfread.c}.
1784 SOM is HP's object file and debug format (not to be confused with IBM's
1785 SOM, which is a cross-language ABI).
1787 The SOM reader is in @file{hpread.c}.
1789 @subsection Other File Formats
1791 @cindex Netware Loadable Module format
1792 Other file formats that have been supported by @value{GDBN} include Netware
1793 Loadable Modules (@file{nlmread.c}).
1795 @section Debugging File Formats
1797 This section describes characteristics of debugging information that
1798 are independent of the object file format.
1802 @cindex stabs debugging info
1803 @code{stabs} started out as special symbols within the @code{a.out}
1804 format. Since then, it has been encapsulated into other file
1805 formats, such as COFF and ELF.
1807 While @file{dbxread.c} does some of the basic stab processing,
1808 including for encapsulated versions, @file{stabsread.c} does
1813 @cindex COFF debugging info
1814 The basic COFF definition includes debugging information. The level
1815 of support is minimal and non-extensible, and is not often used.
1817 @subsection Mips debug (Third Eye)
1819 @cindex ECOFF debugging info
1820 ECOFF includes a definition of a special debug format.
1822 The file @file{mdebugread.c} implements reading for this format.
1826 @cindex DWARF 1 debugging info
1827 DWARF 1 is a debugging format that was originally designed to be
1828 used with ELF in SVR4 systems.
1834 @c If defined, these are the producer strings in a DWARF 1 file. All of
1835 @c these have reasonable defaults already.
1837 The DWARF 1 reader is in @file{dwarfread.c}.
1841 @cindex DWARF 2 debugging info
1842 DWARF 2 is an improved but incompatible version of DWARF 1.
1844 The DWARF 2 reader is in @file{dwarf2read.c}.
1848 @cindex SOM debugging info
1849 Like COFF, the SOM definition includes debugging information.
1851 @section Adding a New Symbol Reader to @value{GDBN}
1853 @cindex adding debugging info reader
1854 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
1855 there is probably little to be done.
1857 If you need to add a new object file format, you must first add it to
1858 BFD. This is beyond the scope of this document.
1860 You must then arrange for the BFD code to provide access to the
1861 debugging symbols. Generally @value{GDBN} will have to call swapping routines
1862 from BFD and a few other BFD internal routines to locate the debugging
1863 information. As much as possible, @value{GDBN} should not depend on the BFD
1864 internal data structures.
1866 For some targets (e.g., COFF), there is a special transfer vector used
1867 to call swapping routines, since the external data structures on various
1868 platforms have different sizes and layouts. Specialized routines that
1869 will only ever be implemented by one object file format may be called
1870 directly. This interface should be described in a file
1871 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
1874 @node Language Support
1876 @chapter Language Support
1878 @cindex language support
1879 @value{GDBN}'s language support is mainly driven by the symbol reader,
1880 although it is possible for the user to set the source language
1883 @value{GDBN} chooses the source language by looking at the extension
1884 of the file recorded in the debug info; @file{.c} means C, @file{.f}
1885 means Fortran, etc. It may also use a special-purpose language
1886 identifier if the debug format supports it, like with DWARF.
1888 @section Adding a Source Language to @value{GDBN}
1890 @cindex adding source language
1891 To add other languages to @value{GDBN}'s expression parser, follow the
1895 @item Create the expression parser.
1897 @cindex expression parser
1898 This should reside in a file @file{@var{lang}-exp.y}. Routines for
1899 building parsed expressions into a @code{union exp_element} list are in
1902 @cindex language parser
1903 Since we can't depend upon everyone having Bison, and YACC produces
1904 parsers that define a bunch of global names, the following lines
1905 @strong{must} be included at the top of the YACC parser, to prevent the
1906 various parsers from defining the same global names:
1909 #define yyparse @var{lang}_parse
1910 #define yylex @var{lang}_lex
1911 #define yyerror @var{lang}_error
1912 #define yylval @var{lang}_lval
1913 #define yychar @var{lang}_char
1914 #define yydebug @var{lang}_debug
1915 #define yypact @var{lang}_pact
1916 #define yyr1 @var{lang}_r1
1917 #define yyr2 @var{lang}_r2
1918 #define yydef @var{lang}_def
1919 #define yychk @var{lang}_chk
1920 #define yypgo @var{lang}_pgo
1921 #define yyact @var{lang}_act
1922 #define yyexca @var{lang}_exca
1923 #define yyerrflag @var{lang}_errflag
1924 #define yynerrs @var{lang}_nerrs
1927 At the bottom of your parser, define a @code{struct language_defn} and
1928 initialize it with the right values for your language. Define an
1929 @code{initialize_@var{lang}} routine and have it call
1930 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
1931 that your language exists. You'll need some other supporting variables
1932 and functions, which will be used via pointers from your
1933 @code{@var{lang}_language_defn}. See the declaration of @code{struct
1934 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
1935 for more information.
1937 @item Add any evaluation routines, if necessary
1939 @cindex expression evaluation routines
1940 @findex evaluate_subexp
1941 @findex prefixify_subexp
1942 @findex length_of_subexp
1943 If you need new opcodes (that represent the operations of the language),
1944 add them to the enumerated type in @file{expression.h}. Add support
1945 code for these operations in the @code{evaluate_subexp} function
1946 defined in the file @file{eval.c}. Add cases
1947 for new opcodes in two functions from @file{parse.c}:
1948 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
1949 the number of @code{exp_element}s that a given operation takes up.
1951 @item Update some existing code
1953 Add an enumerated identifier for your language to the enumerated type
1954 @code{enum language} in @file{defs.h}.
1956 Update the routines in @file{language.c} so your language is included.
1957 These routines include type predicates and such, which (in some cases)
1958 are language dependent. If your language does not appear in the switch
1959 statement, an error is reported.
1961 @vindex current_language
1962 Also included in @file{language.c} is the code that updates the variable
1963 @code{current_language}, and the routines that translate the
1964 @code{language_@var{lang}} enumerated identifier into a printable
1967 @findex _initialize_language
1968 Update the function @code{_initialize_language} to include your
1969 language. This function picks the default language upon startup, so is
1970 dependent upon which languages that @value{GDBN} is built for.
1972 @findex allocate_symtab
1973 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
1974 code so that the language of each symtab (source file) is set properly.
1975 This is used to determine the language to use at each stack frame level.
1976 Currently, the language is set based upon the extension of the source
1977 file. If the language can be better inferred from the symbol
1978 information, please set the language of the symtab in the symbol-reading
1981 @findex print_subexp
1982 @findex op_print_tab
1983 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
1984 expression opcodes you have added to @file{expression.h}. Also, add the
1985 printed representations of your operators to @code{op_print_tab}.
1987 @item Add a place of call
1990 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
1991 @code{parse_exp_1} (defined in @file{parse.c}).
1993 @item Use macros to trim code
1995 @cindex trimming language-dependent code
1996 The user has the option of building @value{GDBN} for some or all of the
1997 languages. If the user decides to build @value{GDBN} for the language
1998 @var{lang}, then every file dependent on @file{language.h} will have the
1999 macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to
2000 leave out large routines that the user won't need if he or she is not
2001 using your language.
2003 Note that you do not need to do this in your YACC parser, since if @value{GDBN}
2004 is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the
2005 compiled form of your parser) is not linked into @value{GDBN} at all.
2007 See the file @file{configure.in} for how @value{GDBN} is configured
2008 for different languages.
2010 @item Edit @file{Makefile.in}
2012 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2013 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2014 not get linked in, or, worse yet, it may not get @code{tar}red into the
2019 @node Host Definition
2021 @chapter Host Definition
2023 @emph{Maintainer's note: In theory, new targets no longer need to use
2024 the host framework described below. Instead it should be possible to
2025 handle everything using autoconf. Patches eliminating this framework
2028 With the advent of Autoconf, it's rarely necessary to have host
2029 definition machinery anymore.
2031 @section Adding a New Host
2033 @cindex adding a new host
2034 @cindex host, adding
2035 Most of @value{GDBN}'s host configuration support happens via
2036 Autoconf. New host-specific definitions should be rarely needed.
2037 @value{GDBN} still uses the host-specific definitions and files listed
2038 below, but these mostly exist for historical reasons, and should
2039 eventually disappear.
2041 Several files control @value{GDBN}'s configuration for host systems:
2045 @item gdb/config/@var{arch}/@var{xyz}.mh
2046 Specifies Makefile fragments needed when hosting on machine @var{xyz}.
2047 In particular, this lists the required machine-dependent object files,
2048 by defining @samp{XDEPFILES=@dots{}}. Also specifies the header file
2049 which describes host @var{xyz}, by defining @code{XM_FILE=
2050 xm-@var{xyz}.h}. You can also define @code{CC}, @code{SYSV_DEFINE},
2051 @code{XM_CFLAGS}, @code{XM_ADD_FILES}, @code{XM_CLIBS}, @code{XM_CDEPS},
2052 etc.; see @file{Makefile.in}.
2054 @item gdb/config/@var{arch}/xm-@var{xyz}.h
2055 (@file{xm.h} is a link to this file, created by @code{configure}). Contains C
2056 macro definitions describing the host system environment, such as byte
2057 order, host C compiler and library.
2059 @item gdb/@var{xyz}-xdep.c
2060 Contains any miscellaneous C code required for this machine as a host.
2061 On most machines it doesn't exist at all. If it does exist, put
2062 @file{@var{xyz}-xdep.o} into the @code{XDEPFILES} line in
2063 @file{gdb/config/@var{arch}/@var{xyz}.mh}.
2066 @subheading Generic Host Support Files
2068 @cindex generic host support
2069 There are some ``generic'' versions of routines that can be used by
2070 various systems. These can be customized in various ways by macros
2071 defined in your @file{xm-@var{xyz}.h} file. If these routines work for
2072 the @var{xyz} host, you can just include the generic file's name (with
2073 @samp{.o}, not @samp{.c}) in @code{XDEPFILES}.
2075 Otherwise, if your machine needs custom support routines, you will need
2076 to write routines that perform the same functions as the generic file.
2077 Put them into @code{@var{xyz}-xdep.c}, and put @code{@var{xyz}-xdep.o}
2078 into @code{XDEPFILES}.
2081 @cindex remote debugging support
2082 @cindex serial line support
2084 This contains serial line support for Unix systems. This is always
2085 included, via the makefile variable @code{SER_HARDWIRE}; override this
2086 variable in the @file{.mh} file to avoid it.
2089 This contains serial line support for 32-bit programs running under DOS,
2090 using the DJGPP (a.k.a.@: GO32) execution environment.
2092 @cindex TCP remote support
2094 This contains generic TCP support using sockets.
2097 @section Host Conditionals
2099 When @value{GDBN} is configured and compiled, various macros are
2100 defined or left undefined, to control compilation based on the
2101 attributes of the host system. These macros and their meanings (or if
2102 the meaning is not documented here, then one of the source files where
2103 they are used is indicated) are:
2106 @item @value{GDBN}INIT_FILENAME
2107 The default name of @value{GDBN}'s initialization file (normally
2111 This macro is deprecated.
2114 Define this if your system does not have a @code{<sys/file.h>}.
2116 @item SIGWINCH_HANDLER
2117 If your host defines @code{SIGWINCH}, you can define this to be the name
2118 of a function to be called if @code{SIGWINCH} is received.
2120 @item SIGWINCH_HANDLER_BODY
2121 Define this to expand into code that will define the function named by
2122 the expansion of @code{SIGWINCH_HANDLER}.
2124 @item ALIGN_STACK_ON_STARTUP
2125 @cindex stack alignment
2126 Define this if your system is of a sort that will crash in
2127 @code{tgetent} if the stack happens not to be longword-aligned when
2128 @code{main} is called. This is a rare situation, but is known to occur
2129 on several different types of systems.
2131 @item CRLF_SOURCE_FILES
2132 @cindex DOS text files
2133 Define this if host files use @code{\r\n} rather than @code{\n} as a
2134 line terminator. This will cause source file listings to omit @code{\r}
2135 characters when printing and it will allow @code{\r\n} line endings of files
2136 which are ``sourced'' by gdb. It must be possible to open files in binary
2137 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2139 @item DEFAULT_PROMPT
2141 The default value of the prompt string (normally @code{"(gdb) "}).
2144 @cindex terminal device
2145 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2147 @item FCLOSE_PROVIDED
2148 Define this if the system declares @code{fclose} in the headers included
2149 in @code{defs.h}. This isn't needed unless your compiler is unusually
2153 Define this if binary files are opened the same way as text files.
2155 @item GETENV_PROVIDED
2156 Define this if the system declares @code{getenv} in its headers included
2157 in @code{defs.h}. This isn't needed unless your compiler is unusually
2162 In some cases, use the system call @code{mmap} for reading symbol
2163 tables. For some machines this allows for sharing and quick updates.
2166 Define this if the host system has @code{termio.h}.
2173 Values for host-side constants.
2176 Substitute for isatty, if not available.
2179 This is the longest integer type available on the host. If not defined,
2180 it will default to @code{long long} or @code{long}, depending on
2181 @code{CC_HAS_LONG_LONG}.
2183 @item CC_HAS_LONG_LONG
2184 @cindex @code{long long} data type
2185 Define this if the host C compiler supports @code{long long}. This is set
2186 by the @code{configure} script.
2188 @item PRINTF_HAS_LONG_LONG
2189 Define this if the host can handle printing of long long integers via
2190 the printf format conversion specifier @code{ll}. This is set by the
2191 @code{configure} script.
2193 @item HAVE_LONG_DOUBLE
2194 Define this if the host C compiler supports @code{long double}. This is
2195 set by the @code{configure} script.
2197 @item PRINTF_HAS_LONG_DOUBLE
2198 Define this if the host can handle printing of long double float-point
2199 numbers via the printf format conversion specifier @code{Lg}. This is
2200 set by the @code{configure} script.
2202 @item SCANF_HAS_LONG_DOUBLE
2203 Define this if the host can handle the parsing of long double
2204 float-point numbers via the scanf format conversion specifier
2205 @code{Lg}. This is set by the @code{configure} script.
2207 @item LSEEK_NOT_LINEAR
2208 Define this if @code{lseek (n)} does not necessarily move to byte number
2209 @code{n} in the file. This is only used when reading source files. It
2210 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2213 This macro is used as the argument to @code{lseek} (or, most commonly,
2214 @code{bfd_seek}). FIXME, should be replaced by SEEK_SET instead,
2215 which is the POSIX equivalent.
2217 @item MALLOC_INCOMPATIBLE
2218 Define this if the system's prototype for @code{malloc} differs from the
2219 @sc{ansi} definition.
2221 @item MMAP_BASE_ADDRESS
2222 When using HAVE_MMAP, the first mapping should go at this address.
2224 @item MMAP_INCREMENT
2225 when using HAVE_MMAP, this is the increment between mappings.
2228 If defined, this should be one or more tokens, such as @code{volatile},
2229 that can be used in both the declaration and definition of functions to
2230 indicate that they never return. The default is already set correctly
2231 if compiling with GCC. This will almost never need to be defined.
2234 If defined, this should be one or more tokens, such as
2235 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2236 of functions to indicate that they never return. The default is already
2237 set correctly if compiling with GCC. This will almost never need to be
2240 @item USE_GENERIC_DUMMY_FRAMES
2241 @cindex generic dummy frames
2242 Define this to 1 if the target is using the generic inferior function
2243 call code. See @code{blockframe.c} for more information.
2247 @value{GDBN} will use the @code{mmalloc} library for memory allocation
2248 for symbol reading if this symbol is defined. Be careful defining it
2249 since there are systems on which @code{mmalloc} does not work for some
2250 reason. One example is the DECstation, where its RPC library can't
2251 cope with our redefinition of @code{malloc} to call @code{mmalloc}.
2252 When defining @code{USE_MMALLOC}, you will also have to set
2253 @code{MMALLOC} in the Makefile, to point to the @code{mmalloc} library. This
2254 define is set when you configure with @samp{--with-mmalloc}.
2258 Define this if you are using @code{mmalloc}, but don't want the overhead
2259 of checking the heap with @code{mmcheck}. Note that on some systems,
2260 the C runtime makes calls to @code{malloc} prior to calling @code{main}, and if
2261 @code{free} is ever called with these pointers after calling
2262 @code{mmcheck} to enable checking, a memory corruption abort is certain
2263 to occur. These systems can still use @code{mmalloc}, but must define
2267 Define this to 1 if the C runtime allocates memory prior to
2268 @code{mmcheck} being called, but that memory is never freed so we don't
2269 have to worry about it triggering a memory corruption abort. The
2270 default is 0, which means that @code{mmcheck} will only install the heap
2271 checking functions if there has not yet been any memory allocation
2272 calls, and if it fails to install the functions, @value{GDBN} will issue a
2273 warning. This is currently defined if you configure using
2274 @samp{--with-mmalloc}.
2276 @item NO_SIGINTERRUPT
2277 @findex siginterrupt
2278 Define this to indicate that @code{siginterrupt} is not available.
2282 Define these to appropriate value for the system @code{lseek}, if not already
2286 This is the signal for stopping @value{GDBN}. Defaults to
2287 @code{SIGTSTP}. (Only redefined for the Convex.)
2290 Define this if the interior's tty should be opened with the @code{O_NOCTTY}
2291 flag. (FIXME: This should be a native-only flag, but @file{inflow.c} is
2295 Means that System V (prior to SVR4) include files are in use. (FIXME:
2296 This symbol is abused in @file{infrun.c}, @file{regex.c},
2297 @file{remote-nindy.c}, and @file{utils.c} for other things, at the
2301 Define this to help placate @code{lint} in some situations.
2304 Define this to override the defaults of @code{__volatile__} or
2309 @node Target Architecture Definition
2311 @chapter Target Architecture Definition
2313 @cindex target architecture definition
2314 @value{GDBN}'s target architecture defines what sort of
2315 machine-language programs @value{GDBN} can work with, and how it works
2318 The target architecture object is implemented as the C structure
2319 @code{struct gdbarch *}. The structure, and its methods, are generated
2320 using the Bourne shell script @file{gdbarch.sh}.
2322 @section Registers and Memory
2324 @value{GDBN}'s model of the target machine is rather simple.
2325 @value{GDBN} assumes the machine includes a bank of registers and a
2326 block of memory. Each register may have a different size.
2328 @value{GDBN} does not have a magical way to match up with the
2329 compiler's idea of which registers are which; however, it is critical
2330 that they do match up accurately. The only way to make this work is
2331 to get accurate information about the order that the compiler uses,
2332 and to reflect that in the @code{REGISTER_NAME} and related macros.
2334 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
2336 @section Pointers Are Not Always Addresses
2337 @cindex pointer representation
2338 @cindex address representation
2339 @cindex word-addressed machines
2340 @cindex separate data and code address spaces
2341 @cindex spaces, separate data and code address
2342 @cindex address spaces, separate data and code
2343 @cindex code pointers, word-addressed
2344 @cindex converting between pointers and addresses
2345 @cindex D10V addresses
2347 On almost all 32-bit architectures, the representation of a pointer is
2348 indistinguishable from the representation of some fixed-length number
2349 whose value is the byte address of the object pointed to. On such
2350 machines, the words ``pointer'' and ``address'' can be used interchangeably.
2351 However, architectures with smaller word sizes are often cramped for
2352 address space, so they may choose a pointer representation that breaks this
2353 identity, and allows a larger code address space.
2355 For example, the Mitsubishi D10V is a 16-bit VLIW processor whose
2356 instructions are 32 bits long@footnote{Some D10V instructions are
2357 actually pairs of 16-bit sub-instructions. However, since you can't
2358 jump into the middle of such a pair, code addresses can only refer to
2359 full 32 bit instructions, which is what matters in this explanation.}.
2360 If the D10V used ordinary byte addresses to refer to code locations,
2361 then the processor would only be able to address 64kb of instructions.
2362 However, since instructions must be aligned on four-byte boundaries, the
2363 low two bits of any valid instruction's byte address are always
2364 zero---byte addresses waste two bits. So instead of byte addresses,
2365 the D10V uses word addresses---byte addresses shifted right two bits---to
2366 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
2369 However, this means that code pointers and data pointers have different
2370 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
2371 @code{0xC020} when used as a data address, but refers to byte address
2372 @code{0x30080} when used as a code address.
2374 (The D10V also uses separate code and data address spaces, which also
2375 affects the correspondence between pointers and addresses, but we're
2376 going to ignore that here; this example is already too long.)
2378 To cope with architectures like this---the D10V is not the only
2379 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
2380 byte numbers, and @dfn{pointers}, which are the target's representation
2381 of an address of a particular type of data. In the example above,
2382 @code{0xC020} is the pointer, which refers to one of the addresses
2383 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
2384 @value{GDBN} provides functions for turning a pointer into an address
2385 and vice versa, in the appropriate way for the current architecture.
2387 Unfortunately, since addresses and pointers are identical on almost all
2388 processors, this distinction tends to bit-rot pretty quickly. Thus,
2389 each time you port @value{GDBN} to an architecture which does
2390 distinguish between pointers and addresses, you'll probably need to
2391 clean up some architecture-independent code.
2393 Here are functions which convert between pointers and addresses:
2395 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
2396 Treat the bytes at @var{buf} as a pointer or reference of type
2397 @var{type}, and return the address it represents, in a manner
2398 appropriate for the current architecture. This yields an address
2399 @value{GDBN} can use to read target memory, disassemble, etc. Note that
2400 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2403 For example, if the current architecture is the Intel x86, this function
2404 extracts a little-endian integer of the appropriate length from
2405 @var{buf} and returns it. However, if the current architecture is the
2406 D10V, this function will return a 16-bit integer extracted from
2407 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
2409 If @var{type} is not a pointer or reference type, then this function
2410 will signal an internal error.
2413 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
2414 Store the address @var{addr} in @var{buf}, in the proper format for a
2415 pointer of type @var{type} in the current architecture. Note that
2416 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2419 For example, if the current architecture is the Intel x86, this function
2420 stores @var{addr} unmodified as a little-endian integer of the
2421 appropriate length in @var{buf}. However, if the current architecture
2422 is the D10V, this function divides @var{addr} by four if @var{type} is
2423 a pointer to a function, and then stores it in @var{buf}.
2425 If @var{type} is not a pointer or reference type, then this function
2426 will signal an internal error.
2429 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
2430 Assuming that @var{val} is a pointer, return the address it represents,
2431 as appropriate for the current architecture.
2433 This function actually works on integral values, as well as pointers.
2434 For pointers, it performs architecture-specific conversions as
2435 described above for @code{extract_typed_address}.
2438 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
2439 Create and return a value representing a pointer of type @var{type} to
2440 the address @var{addr}, as appropriate for the current architecture.
2441 This function performs architecture-specific conversions as described
2442 above for @code{store_typed_address}.
2446 @value{GDBN} also provides functions that do the same tasks, but assume
2447 that pointers are simply byte addresses; they aren't sensitive to the
2448 current architecture, beyond knowing the appropriate endianness.
2450 @deftypefun CORE_ADDR extract_address (void *@var{addr}, int len)
2451 Extract a @var{len}-byte number from @var{addr} in the appropriate
2452 endianness for the current architecture, and return it. Note that
2453 @var{addr} refers to @value{GDBN}'s memory, not the inferior's.
2455 This function should only be used in architecture-specific code; it
2456 doesn't have enough information to turn bits into a true address in the
2457 appropriate way for the current architecture. If you can, use
2458 @code{extract_typed_address} instead.
2461 @deftypefun void store_address (void *@var{addr}, int @var{len}, LONGEST @var{val})
2462 Store @var{val} at @var{addr} as a @var{len}-byte integer, in the
2463 appropriate endianness for the current architecture. Note that
2464 @var{addr} refers to a buffer in @value{GDBN}'s memory, not the
2467 This function should only be used in architecture-specific code; it
2468 doesn't have enough information to turn a true address into bits in the
2469 appropriate way for the current architecture. If you can, use
2470 @code{store_typed_address} instead.
2474 Here are some macros which architectures can define to indicate the
2475 relationship between pointers and addresses. These have default
2476 definitions, appropriate for architectures on which all pointers are
2477 simple unsigned byte addresses.
2479 @deftypefn {Target Macro} CORE_ADDR POINTER_TO_ADDRESS (struct type *@var{type}, char *@var{buf})
2480 Assume that @var{buf} holds a pointer of type @var{type}, in the
2481 appropriate format for the current architecture. Return the byte
2482 address the pointer refers to.
2484 This function may safely assume that @var{type} is either a pointer or a
2485 C@t{++} reference type.
2488 @deftypefn {Target Macro} void ADDRESS_TO_POINTER (struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
2489 Store in @var{buf} a pointer of type @var{type} representing the address
2490 @var{addr}, in the appropriate format for the current architecture.
2492 This function may safely assume that @var{type} is either a pointer or a
2493 C@t{++} reference type.
2497 @section Using Different Register and Memory Data Representations
2498 @cindex raw representation
2499 @cindex virtual representation
2500 @cindex representations, raw and virtual
2501 @cindex register data formats, converting
2502 @cindex @code{struct value}, converting register contents to
2504 @emph{Maintainer's note: The way GDB manipulates registers is undergoing
2505 significant change. Many of the macros and functions refered to in the
2506 sections below are likely to be made obsolete. See the file @file{TODO}
2507 for more up-to-date information.}
2509 Some architectures use one representation for a value when it lives in a
2510 register, but use a different representation when it lives in memory.
2511 In @value{GDBN}'s terminology, the @dfn{raw} representation is the one used in
2512 the target registers, and the @dfn{virtual} representation is the one
2513 used in memory, and within @value{GDBN} @code{struct value} objects.
2515 For almost all data types on almost all architectures, the virtual and
2516 raw representations are identical, and no special handling is needed.
2517 However, they do occasionally differ. For example:
2521 The x86 architecture supports an 80-bit @code{long double} type. However, when
2522 we store those values in memory, they occupy twelve bytes: the
2523 floating-point number occupies the first ten, and the final two bytes
2524 are unused. This keeps the values aligned on four-byte boundaries,
2525 allowing more efficient access. Thus, the x86 80-bit floating-point
2526 type is the raw representation, and the twelve-byte loosely-packed
2527 arrangement is the virtual representation.
2530 Some 64-bit MIPS targets present 32-bit registers to @value{GDBN} as 64-bit
2531 registers, with garbage in their upper bits. @value{GDBN} ignores the top 32
2532 bits. Thus, the 64-bit form, with garbage in the upper 32 bits, is the
2533 raw representation, and the trimmed 32-bit representation is the
2534 virtual representation.
2537 In general, the raw representation is determined by the architecture, or
2538 @value{GDBN}'s interface to the architecture, while the virtual representation
2539 can be chosen for @value{GDBN}'s convenience. @value{GDBN}'s register file,
2540 @code{registers}, holds the register contents in raw format, and the
2541 @value{GDBN} remote protocol transmits register values in raw format.
2543 Your architecture may define the following macros to request
2544 conversions between the raw and virtual format:
2546 @deftypefn {Target Macro} int REGISTER_CONVERTIBLE (int @var{reg})
2547 Return non-zero if register number @var{reg}'s value needs different raw
2548 and virtual formats.
2550 You should not use @code{REGISTER_CONVERT_TO_VIRTUAL} for a register
2551 unless this macro returns a non-zero value for that register.
2554 @deftypefn {Target Macro} int REGISTER_RAW_SIZE (int @var{reg})
2555 The size of register number @var{reg}'s raw value. This is the number
2556 of bytes the register will occupy in @code{registers}, or in a @value{GDBN}
2557 remote protocol packet.
2560 @deftypefn {Target Macro} int REGISTER_VIRTUAL_SIZE (int @var{reg})
2561 The size of register number @var{reg}'s value, in its virtual format.
2562 This is the size a @code{struct value}'s buffer will have, holding that
2566 @deftypefn {Target Macro} struct type *REGISTER_VIRTUAL_TYPE (int @var{reg})
2567 This is the type of the virtual representation of register number
2568 @var{reg}. Note that there is no need for a macro giving a type for the
2569 register's raw form; once the register's value has been obtained, @value{GDBN}
2570 always uses the virtual form.
2573 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_VIRTUAL (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
2574 Convert the value of register number @var{reg} to @var{type}, which
2575 should always be @code{REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2576 at @var{from} holds the register's value in raw format; the macro should
2577 convert the value to virtual format, and place it at @var{to}.
2579 Note that @code{REGISTER_CONVERT_TO_VIRTUAL} and
2580 @code{REGISTER_CONVERT_TO_RAW} take their @var{reg} and @var{type}
2581 arguments in different orders.
2583 You should only use @code{REGISTER_CONVERT_TO_VIRTUAL} with registers
2584 for which the @code{REGISTER_CONVERTIBLE} macro returns a non-zero
2588 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_RAW (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
2589 Convert the value of register number @var{reg} to @var{type}, which
2590 should always be @code{REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2591 at @var{from} holds the register's value in raw format; the macro should
2592 convert the value to virtual format, and place it at @var{to}.
2594 Note that REGISTER_CONVERT_TO_VIRTUAL and REGISTER_CONVERT_TO_RAW take
2595 their @var{reg} and @var{type} arguments in different orders.
2599 @section Frame Interpretation
2601 @section Inferior Call Setup
2603 @section Compiler Characteristics
2605 @section Target Conditionals
2607 This section describes the macros that you can use to define the target
2612 @item ADDITIONAL_OPTIONS
2613 @itemx ADDITIONAL_OPTION_CASES
2614 @itemx ADDITIONAL_OPTION_HANDLER
2615 @itemx ADDITIONAL_OPTION_HELP
2616 @findex ADDITIONAL_OPTION_HELP
2617 @findex ADDITIONAL_OPTION_HANDLER
2618 @findex ADDITIONAL_OPTION_CASES
2619 @findex ADDITIONAL_OPTIONS
2620 These are a set of macros that allow the addition of additional command
2621 line options to @value{GDBN}. They are currently used only for the unsupported
2622 i960 Nindy target, and should not be used in any other configuration.
2624 @item ADDR_BITS_REMOVE (addr)
2625 @findex ADDR_BITS_REMOVE
2626 If a raw machine instruction address includes any bits that are not
2627 really part of the address, then define this macro to expand into an
2628 expression that zeroes those bits in @var{addr}. This is only used for
2629 addresses of instructions, and even then not in all contexts.
2631 For example, the two low-order bits of the PC on the Hewlett-Packard PA
2632 2.0 architecture contain the privilege level of the corresponding
2633 instruction. Since instructions must always be aligned on four-byte
2634 boundaries, the processor masks out these bits to generate the actual
2635 address of the instruction. ADDR_BITS_REMOVE should filter out these
2636 bits with an expression such as @code{((addr) & ~3)}.
2638 @item ADDRESS_TO_POINTER (@var{type}, @var{buf}, @var{addr})
2639 @findex ADDRESS_TO_POINTER
2640 Store in @var{buf} a pointer of type @var{type} representing the address
2641 @var{addr}, in the appropriate format for the current architecture.
2642 This macro may safely assume that @var{type} is either a pointer or a
2643 C@t{++} reference type.
2644 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
2646 @item BEFORE_MAIN_LOOP_HOOK
2647 @findex BEFORE_MAIN_LOOP_HOOK
2648 Define this to expand into any code that you want to execute before the
2649 main loop starts. Although this is not, strictly speaking, a target
2650 conditional, that is how it is currently being used. Note that if a
2651 configuration were to define it one way for a host and a different way
2652 for the target, @value{GDBN} will probably not compile, let alone run
2653 correctly. This macro is currently used only for the unsupported i960 Nindy
2654 target, and should not be used in any other configuration.
2656 @item BELIEVE_PCC_PROMOTION
2657 @findex BELIEVE_PCC_PROMOTION
2658 Define if the compiler promotes a @code{short} or @code{char}
2659 parameter to an @code{int}, but still reports the parameter as its
2660 original type, rather than the promoted type.
2662 @item BELIEVE_PCC_PROMOTION_TYPE
2663 @findex BELIEVE_PCC_PROMOTION_TYPE
2664 Define this if @value{GDBN} should believe the type of a @code{short}
2665 argument when compiled by @code{pcc}, but look within a full int space to get
2666 its value. Only defined for Sun-3 at present.
2668 @item BITS_BIG_ENDIAN
2669 @findex BITS_BIG_ENDIAN
2670 Define this if the numbering of bits in the targets does @strong{not} match the
2671 endianness of the target byte order. A value of 1 means that the bits
2672 are numbered in a big-endian bit order, 0 means little-endian.
2676 This is the character array initializer for the bit pattern to put into
2677 memory where a breakpoint is set. Although it's common to use a trap
2678 instruction for a breakpoint, it's not required; for instance, the bit
2679 pattern could be an invalid instruction. The breakpoint must be no
2680 longer than the shortest instruction of the architecture.
2682 @code{BREAKPOINT} has been deprecated in favor of
2683 @code{BREAKPOINT_FROM_PC}.
2685 @item BIG_BREAKPOINT
2686 @itemx LITTLE_BREAKPOINT
2687 @findex LITTLE_BREAKPOINT
2688 @findex BIG_BREAKPOINT
2689 Similar to BREAKPOINT, but used for bi-endian targets.
2691 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
2692 favor of @code{BREAKPOINT_FROM_PC}.
2694 @item REMOTE_BREAKPOINT
2695 @itemx LITTLE_REMOTE_BREAKPOINT
2696 @itemx BIG_REMOTE_BREAKPOINT
2697 @findex BIG_REMOTE_BREAKPOINT
2698 @findex LITTLE_REMOTE_BREAKPOINT
2699 @findex REMOTE_BREAKPOINT
2700 Similar to BREAKPOINT, but used for remote targets.
2702 @code{BIG_REMOTE_BREAKPOINT} and @code{LITTLE_REMOTE_BREAKPOINT} have been
2703 deprecated in favor of @code{BREAKPOINT_FROM_PC}.
2705 @item BREAKPOINT_FROM_PC (@var{pcptr}, @var{lenptr})
2706 @findex BREAKPOINT_FROM_PC
2707 Use the program counter to determine the contents and size of a
2708 breakpoint instruction. It returns a pointer to a string of bytes
2709 that encode a breakpoint instruction, stores the length of the string
2710 to *@var{lenptr}, and adjusts pc (if necessary) to point to the actual
2711 memory location where the breakpoint should be inserted.
2713 Although it is common to use a trap instruction for a breakpoint, it's
2714 not required; for instance, the bit pattern could be an invalid
2715 instruction. The breakpoint must be no longer than the shortest
2716 instruction of the architecture.
2718 Replaces all the other @var{BREAKPOINT} macros.
2720 @item MEMORY_INSERT_BREAKPOINT (@var{addr}, @var{contents_cache})
2721 @itemx MEMORY_REMOVE_BREAKPOINT (@var{addr}, @var{contents_cache})
2722 @findex MEMORY_REMOVE_BREAKPOINT
2723 @findex MEMORY_INSERT_BREAKPOINT
2724 Insert or remove memory based breakpoints. Reasonable defaults
2725 (@code{default_memory_insert_breakpoint} and
2726 @code{default_memory_remove_breakpoint} respectively) have been
2727 provided so that it is not necessary to define these for most
2728 architectures. Architectures which may want to define
2729 @code{MEMORY_INSERT_BREAKPOINT} and @code{MEMORY_REMOVE_BREAKPOINT} will
2730 likely have instructions that are oddly sized or are not stored in a
2731 conventional manner.
2733 It may also be desirable (from an efficiency standpoint) to define
2734 custom breakpoint insertion and removal routines if
2735 @code{BREAKPOINT_FROM_PC} needs to read the target's memory for some
2739 @findex CALL_DUMMY_P
2740 A C expresson that is non-zero when the target suports inferior function
2743 @item CALL_DUMMY_WORDS
2744 @findex CALL_DUMMY_WORDS
2745 Pointer to an array of @code{LONGEST} words of data containing
2746 host-byte-ordered @code{REGISTER_BYTES} sized values that partially
2747 specify the sequence of instructions needed for an inferior function
2750 Should be deprecated in favor of a macro that uses target-byte-ordered
2753 @item SIZEOF_CALL_DUMMY_WORDS
2754 @findex SIZEOF_CALL_DUMMY_WORDS
2755 The size of @code{CALL_DUMMY_WORDS}. When @code{CALL_DUMMY_P} this must
2756 return a positive value. See also @code{CALL_DUMMY_LENGTH}.
2760 A static initializer for @code{CALL_DUMMY_WORDS}. Deprecated.
2762 @item CALL_DUMMY_LOCATION
2763 @findex CALL_DUMMY_LOCATION
2764 See the file @file{inferior.h}.
2766 @item CALL_DUMMY_STACK_ADJUST
2767 @findex CALL_DUMMY_STACK_ADJUST
2768 Stack adjustment needed when performing an inferior function call.
2770 Should be deprecated in favor of something like @code{STACK_ALIGN}.
2772 @item CALL_DUMMY_STACK_ADJUST_P
2773 @findex CALL_DUMMY_STACK_ADJUST_P
2774 Predicate for use of @code{CALL_DUMMY_STACK_ADJUST}.
2776 Should be deprecated in favor of something like @code{STACK_ALIGN}.
2778 @item CANNOT_FETCH_REGISTER (@var{regno})
2779 @findex CANNOT_FETCH_REGISTER
2780 A C expression that should be nonzero if @var{regno} cannot be fetched
2781 from an inferior process. This is only relevant if
2782 @code{FETCH_INFERIOR_REGISTERS} is not defined.
2784 @item CANNOT_STORE_REGISTER (@var{regno})
2785 @findex CANNOT_STORE_REGISTER
2786 A C expression that should be nonzero if @var{regno} should not be
2787 written to the target. This is often the case for program counters,
2788 status words, and other special registers. If this is not defined,
2789 @value{GDBN} will assume that all registers may be written.
2791 @item DO_DEFERRED_STORES
2792 @itemx CLEAR_DEFERRED_STORES
2793 @findex CLEAR_DEFERRED_STORES
2794 @findex DO_DEFERRED_STORES
2795 Define this to execute any deferred stores of registers into the inferior,
2796 and to cancel any deferred stores.
2798 Currently only implemented correctly for native Sparc configurations?
2800 @item COERCE_FLOAT_TO_DOUBLE (@var{formal}, @var{actual})
2801 @findex COERCE_FLOAT_TO_DOUBLE
2802 @cindex promotion to @code{double}
2803 @cindex @code{float} arguments
2804 @cindex prototyped functions, passing arguments to
2805 @cindex passing arguments to prototyped functions
2806 Return non-zero if GDB should promote @code{float} values to
2807 @code{double} when calling a non-prototyped function. The argument
2808 @var{actual} is the type of the value we want to pass to the function.
2809 The argument @var{formal} is the type of this argument, as it appears in
2810 the function's definition. Note that @var{formal} may be zero if we
2811 have no debugging information for the function, or if we're passing more
2812 arguments than are officially declared (for example, varargs). This
2813 macro is never invoked if the function definitely has a prototype.
2815 How you should pass arguments to a function depends on whether it was
2816 defined in K&R style or prototype style. If you define a function using
2817 the K&R syntax that takes a @code{float} argument, then callers must
2818 pass that argument as a @code{double}. If you define the function using
2819 the prototype syntax, then you must pass the argument as a @code{float},
2822 Unfortunately, on certain older platforms, the debug info doesn't
2823 indicate reliably how each function was defined. A function type's
2824 @code{TYPE_FLAG_PROTOTYPED} flag may be unset, even if the function was
2825 defined in prototype style. When calling a function whose
2826 @code{TYPE_FLAG_PROTOTYPED} flag is unset, GDB consults the
2827 @code{COERCE_FLOAT_TO_DOUBLE} macro to decide what to do.
2829 @findex standard_coerce_float_to_double
2830 For modern targets, it is proper to assume that, if the prototype flag
2831 is unset, that can be trusted: @code{float} arguments should be promoted
2832 to @code{double}. You should use the function
2833 @code{standard_coerce_float_to_double} to get this behavior.
2835 @findex default_coerce_float_to_double
2836 For some older targets, if the prototype flag is unset, that doesn't
2837 tell us anything. So we guess that, if we don't have a type for the
2838 formal parameter (@i{i.e.}, the first argument to
2839 @code{COERCE_FLOAT_TO_DOUBLE} is null), then we should promote it;
2840 otherwise, we should leave it alone. The function
2841 @code{default_coerce_float_to_double} provides this behavior; it is the
2842 default value, for compatibility with older configurations.
2845 @findex CPLUS_MARKERz
2846 Define this to expand into the character that G@t{++} uses to distinguish
2847 compiler-generated identifiers from programmer-specified identifiers.
2848 By default, this expands into @code{'$'}. Most System V targets should
2849 define this to @code{'.'}.
2851 @item DBX_PARM_SYMBOL_CLASS
2852 @findex DBX_PARM_SYMBOL_CLASS
2853 Hook for the @code{SYMBOL_CLASS} of a parameter when decoding DBX symbol
2854 information. In the i960, parameters can be stored as locals or as
2855 args, depending on the type of the debug record.
2857 @item DECR_PC_AFTER_BREAK
2858 @findex DECR_PC_AFTER_BREAK
2859 Define this to be the amount by which to decrement the PC after the
2860 program encounters a breakpoint. This is often the number of bytes in
2861 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
2863 @item DECR_PC_AFTER_HW_BREAK
2864 @findex DECR_PC_AFTER_HW_BREAK
2865 Similarly, for hardware breakpoints.
2867 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
2868 @findex DISABLE_UNSETTABLE_BREAK
2869 If defined, this should evaluate to 1 if @var{addr} is in a shared
2870 library in which breakpoints cannot be set and so should be disabled.
2872 @item DO_REGISTERS_INFO
2873 @findex DO_REGISTERS_INFO
2874 If defined, use this to print the value of a register or all registers.
2876 @item DWARF_REG_TO_REGNUM
2877 @findex DWARF_REG_TO_REGNUM
2878 Convert DWARF register number into @value{GDBN} regnum. If not defined,
2879 no conversion will be performed.
2881 @item DWARF2_REG_TO_REGNUM
2882 @findex DWARF2_REG_TO_REGNUM
2883 Convert DWARF2 register number into @value{GDBN} regnum. If not
2884 defined, no conversion will be performed.
2886 @item ECOFF_REG_TO_REGNUM
2887 @findex ECOFF_REG_TO_REGNUM
2888 Convert ECOFF register number into @value{GDBN} regnum. If not defined,
2889 no conversion will be performed.
2891 @item END_OF_TEXT_DEFAULT
2892 @findex END_OF_TEXT_DEFAULT
2893 This is an expression that should designate the end of the text section.
2896 @item EXTRACT_RETURN_VALUE(@var{type}, @var{regbuf}, @var{valbuf})
2897 @findex EXTRACT_RETURN_VALUE
2898 Define this to extract a function's return value of type @var{type} from
2899 the raw register state @var{regbuf} and copy that, in virtual format,
2902 @item EXTRACT_STRUCT_VALUE_ADDRESS(@var{regbuf})
2903 @findex EXTRACT_STRUCT_VALUE_ADDRESS
2904 When defined, extract from the array @var{regbuf} (containing the raw
2905 register state) the @code{CORE_ADDR} at which a function should return
2906 its structure value.
2908 If not defined, @code{EXTRACT_RETURN_VALUE} is used.
2910 @item EXTRACT_STRUCT_VALUE_ADDRESS_P()
2911 @findex EXTRACT_STRUCT_VALUE_ADDRESS_P
2912 Predicate for @code{EXTRACT_STRUCT_VALUE_ADDRESS}.
2916 If defined, then the @samp{info float} command will print information about
2917 the processor's floating point unit.
2921 If the virtual frame pointer is kept in a register, then define this
2922 macro to be the number (greater than or equal to zero) of that register.
2924 This should only need to be defined if @code{TARGET_READ_FP} and
2925 @code{TARGET_WRITE_FP} are not defined.
2927 @item FRAMELESS_FUNCTION_INVOCATION(@var{fi})
2928 @findex FRAMELESS_FUNCTION_INVOCATION
2929 Define this to an expression that returns 1 if the function invocation
2930 represented by @var{fi} does not have a stack frame associated with it.
2933 @item FRAME_ARGS_ADDRESS_CORRECT
2934 @findex FRAME_ARGS_ADDRESS_CORRECT
2937 @item FRAME_CHAIN(@var{frame})
2939 Given @var{frame}, return a pointer to the calling frame.
2941 @item FRAME_CHAIN_COMBINE(@var{chain}, @var{frame})
2942 @findex FRAME_CHAIN_COMBINE
2943 Define this to take the frame chain pointer and the frame's nominal
2944 address and produce the nominal address of the caller's frame.
2945 Presently only defined for HP PA.
2947 @item FRAME_CHAIN_VALID(@var{chain}, @var{thisframe})
2948 @findex FRAME_CHAIN_VALID
2949 Define this to be an expression that returns zero if the given frame is
2950 an outermost frame, with no caller, and nonzero otherwise. Several
2951 common definitions are available:
2955 @code{file_frame_chain_valid} is nonzero if the chain pointer is nonzero
2956 and given frame's PC is not inside the startup file (such as
2960 @code{func_frame_chain_valid} is nonzero if the chain
2961 pointer is nonzero and the given frame's PC is not in @code{main} or a
2962 known entry point function (such as @code{_start}).
2965 @code{generic_file_frame_chain_valid} and
2966 @code{generic_func_frame_chain_valid} are equivalent implementations for
2967 targets using generic dummy frames.
2970 @item FRAME_INIT_SAVED_REGS(@var{frame})
2971 @findex FRAME_INIT_SAVED_REGS
2972 See @file{frame.h}. Determines the address of all registers in the
2973 current stack frame storing each in @code{frame->saved_regs}. Space for
2974 @code{frame->saved_regs} shall be allocated by
2975 @code{FRAME_INIT_SAVED_REGS} using either
2976 @code{frame_saved_regs_zalloc} or @code{frame_obstack_alloc}.
2978 @code{FRAME_FIND_SAVED_REGS} and @code{EXTRA_FRAME_INFO} are deprecated.
2980 @item FRAME_NUM_ARGS (@var{fi})
2981 @findex FRAME_NUM_ARGS
2982 For the frame described by @var{fi} return the number of arguments that
2983 are being passed. If the number of arguments is not known, return
2986 @item FRAME_SAVED_PC(@var{frame})
2987 @findex FRAME_SAVED_PC
2988 Given @var{frame}, return the pc saved there. This is the return
2991 @item FUNCTION_EPILOGUE_SIZE
2992 @findex FUNCTION_EPILOGUE_SIZE
2993 For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the
2994 function end symbol is 0. For such targets, you must define
2995 @code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size of a
2996 function's epilogue.
2998 @item FUNCTION_START_OFFSET
2999 @findex FUNCTION_START_OFFSET
3000 An integer, giving the offset in bytes from a function's address (as
3001 used in the values of symbols, function pointers, etc.), and the
3002 function's first genuine instruction.
3004 This is zero on almost all machines: the function's address is usually
3005 the address of its first instruction. However, on the VAX, for example,
3006 each function starts with two bytes containing a bitmask indicating
3007 which registers to save upon entry to the function. The VAX @code{call}
3008 instructions check this value, and save the appropriate registers
3009 automatically. Thus, since the offset from the function's address to
3010 its first instruction is two bytes, @code{FUNCTION_START_OFFSET} would
3013 @item GCC_COMPILED_FLAG_SYMBOL
3014 @itemx GCC2_COMPILED_FLAG_SYMBOL
3015 @findex GCC2_COMPILED_FLAG_SYMBOL
3016 @findex GCC_COMPILED_FLAG_SYMBOL
3017 If defined, these are the names of the symbols that @value{GDBN} will
3018 look for to detect that GCC compiled the file. The default symbols
3019 are @code{gcc_compiled.} and @code{gcc2_compiled.},
3020 respectively. (Currently only defined for the Delta 68.)
3022 @item @value{GDBN}_MULTI_ARCH
3023 @findex @value{GDBN}_MULTI_ARCH
3024 If defined and non-zero, enables suport for multiple architectures
3025 within @value{GDBN}.
3027 This support can be enabled at two levels. At level one, only
3028 definitions for previously undefined macros are provided; at level two,
3029 a multi-arch definition of all architecture dependant macros will be
3032 @item @value{GDBN}_TARGET_IS_HPPA
3033 @findex @value{GDBN}_TARGET_IS_HPPA
3034 This determines whether horrible kludge code in @file{dbxread.c} and
3035 @file{partial-stab.h} is used to mangle multiple-symbol-table files from
3036 HPPA's. This should all be ripped out, and a scheme like @file{elfread.c}
3039 @item GET_LONGJMP_TARGET
3040 @findex GET_LONGJMP_TARGET
3041 For most machines, this is a target-dependent parameter. On the
3042 DECstation and the Iris, this is a native-dependent parameter, since
3043 trhe header file @file{setjmp.h} is needed to define it.
3045 This macro determines the target PC address that @code{longjmp} will jump to,
3046 assuming that we have just stopped at a @code{longjmp} breakpoint. It takes a
3047 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3048 pointer. It examines the current state of the machine as needed.
3050 @item GET_SAVED_REGISTER
3051 @findex GET_SAVED_REGISTER
3052 @findex get_saved_register
3053 Define this if you need to supply your own definition for the function
3054 @code{get_saved_register}.
3056 @item HAVE_REGISTER_WINDOWS
3057 @findex HAVE_REGISTER_WINDOWS
3058 Define this if the target has register windows.
3060 @item REGISTER_IN_WINDOW_P (@var{regnum})
3061 @findex REGISTER_IN_WINDOW_P
3062 Define this to be an expression that is 1 if the given register is in
3065 @item IBM6000_TARGET
3066 @findex IBM6000_TARGET
3067 Shows that we are configured for an IBM RS/6000 target. This
3068 conditional should be eliminated (FIXME) and replaced by
3069 feature-specific macros. It was introduced in a haste and we are
3070 repenting at leisure.
3072 @item I386_USE_GENERIC_WATCHPOINTS
3073 An x86-based target can define this to use the generic x86 watchpoint
3074 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3076 @item SYMBOLS_CAN_START_WITH_DOLLAR
3077 @findex SYMBOLS_CAN_START_WITH_DOLLAR
3078 Some systems have routines whose names start with @samp{$}. Giving this
3079 macro a non-zero value tells @value{GDBN}'s expression parser to check for such
3080 routines when parsing tokens that begin with @samp{$}.
3082 On HP-UX, certain system routines (millicode) have names beginning with
3083 @samp{$} or @samp{$$}. For example, @code{$$dyncall} is a millicode
3084 routine that handles inter-space procedure calls on PA-RISC.
3088 Define this if the target system uses IEEE-format floating point numbers.
3090 @item INIT_EXTRA_FRAME_INFO (@var{fromleaf}, @var{frame})
3091 @findex INIT_EXTRA_FRAME_INFO
3092 If additional information about the frame is required this should be
3093 stored in @code{frame->extra_info}. Space for @code{frame->extra_info}
3094 is allocated using @code{frame_obstack_alloc}.
3096 @item INIT_FRAME_PC (@var{fromleaf}, @var{prev})
3097 @findex INIT_FRAME_PC
3098 This is a C statement that sets the pc of the frame pointed to by
3099 @var{prev}. [By default...]
3101 @item INNER_THAN (@var{lhs}, @var{rhs})
3103 Returns non-zero if stack address @var{lhs} is inner than (nearer to the
3104 stack top) stack address @var{rhs}. Define this as @code{lhs < rhs} if
3105 the target's stack grows downward in memory, or @code{lhs > rsh} if the
3108 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{pc})
3109 @findex gdbarch_in_function_epilogue_p
3110 Returns non-zero if the given @var{pc} is in the epilogue of a function.
3111 The epilogue of a function is defined as the part of a function where
3112 the stack frame of the function already has been destroyed up to the
3113 final `return from function call' instruction.
3115 @item IN_SIGTRAMP (@var{pc}, @var{name})
3117 Define this to return non-zero if the given @var{pc} and/or @var{name}
3118 indicates that the current function is a @code{sigtramp}.
3120 @item SIGTRAMP_START (@var{pc})
3121 @findex SIGTRAMP_START
3122 @itemx SIGTRAMP_END (@var{pc})
3123 @findex SIGTRAMP_END
3124 Define these to be the start and end address of the @code{sigtramp} for the
3125 given @var{pc}. On machines where the address is just a compile time
3126 constant, the macro expansion will typically just ignore the supplied
3129 @item IN_SOLIB_CALL_TRAMPOLINE (@var{pc}, @var{name})
3130 @findex IN_SOLIB_CALL_TRAMPOLINE
3131 Define this to evaluate to nonzero if the program is stopped in the
3132 trampoline that connects to a shared library.
3134 @item IN_SOLIB_RETURN_TRAMPOLINE (@var{pc}, @var{name})
3135 @findex IN_SOLIB_RETURN_TRAMPOLINE
3136 Define this to evaluate to nonzero if the program is stopped in the
3137 trampoline that returns from a shared library.
3139 @item IN_SOLIB_DYNSYM_RESOLVE_CODE (@var{pc})
3140 @findex IN_SOLIB_DYNSYM_RESOLVE_CODE
3141 Define this to evaluate to nonzero if the program is stopped in the
3144 @item SKIP_SOLIB_RESOLVER (@var{pc})
3145 @findex SKIP_SOLIB_RESOLVER
3146 Define this to evaluate to the (nonzero) address at which execution
3147 should continue to get past the dynamic linker's symbol resolution
3148 function. A zero value indicates that it is not important or necessary
3149 to set a breakpoint to get through the dynamic linker and that single
3150 stepping will suffice.
3152 @item INTEGER_TO_ADDRESS (@var{type}, @var{buf})
3153 @findex INTEGER_TO_ADDRESS
3154 @cindex converting integers to addresses
3155 Define this when the architecture needs to handle non-pointer to address
3156 conversions specially. Converts that value to an address according to
3157 the current architectures conventions.
3159 @emph{Pragmatics: When the user copies a well defined expression from
3160 their source code and passes it, as a parameter, to @value{GDBN}'s
3161 @code{print} command, they should get the same value as would have been
3162 computed by the target program. Any deviation from this rule can cause
3163 major confusion and annoyance, and needs to be justified carefully. In
3164 other words, @value{GDBN} doesn't really have the freedom to do these
3165 conversions in clever and useful ways. It has, however, been pointed
3166 out that users aren't complaining about how @value{GDBN} casts integers
3167 to pointers; they are complaining that they can't take an address from a
3168 disassembly listing and give it to @code{x/i}. Adding an architecture
3169 method like @code{INTEGER_TO_ADDRESS} certainly makes it possible for
3170 @value{GDBN} to ``get it right'' in all circumstances.}
3172 @xref{Target Architecture Definition, , Pointers Are Not Always
3175 @item IS_TRAPPED_INTERNALVAR (@var{name})
3176 @findex IS_TRAPPED_INTERNALVAR
3177 This is an ugly hook to allow the specification of special actions that
3178 should occur as a side-effect of setting the value of a variable
3179 internal to @value{GDBN}. Currently only used by the h8500. Note that this
3180 could be either a host or target conditional.
3182 @item NEED_TEXT_START_END
3183 @findex NEED_TEXT_START_END
3184 Define this if @value{GDBN} should determine the start and end addresses of the
3185 text section. (Seems dubious.)
3187 @item NO_HIF_SUPPORT
3188 @findex NO_HIF_SUPPORT
3189 (Specific to the a29k.)
3191 @item POINTER_TO_ADDRESS (@var{type}, @var{buf})
3192 @findex POINTER_TO_ADDRESS
3193 Assume that @var{buf} holds a pointer of type @var{type}, in the
3194 appropriate format for the current architecture. Return the byte
3195 address the pointer refers to.
3196 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3198 @item REGISTER_CONVERTIBLE (@var{reg})
3199 @findex REGISTER_CONVERTIBLE
3200 Return non-zero if @var{reg} uses different raw and virtual formats.
3201 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3203 @item REGISTER_RAW_SIZE (@var{reg})
3204 @findex REGISTER_RAW_SIZE
3205 Return the raw size of @var{reg}.
3206 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3208 @item REGISTER_VIRTUAL_SIZE (@var{reg})
3209 @findex REGISTER_VIRTUAL_SIZE
3210 Return the virtual size of @var{reg}.
3211 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3213 @item REGISTER_VIRTUAL_TYPE (@var{reg})
3214 @findex REGISTER_VIRTUAL_TYPE
3215 Return the virtual type of @var{reg}.
3216 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3218 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
3219 @findex REGISTER_CONVERT_TO_VIRTUAL
3220 Convert the value of register @var{reg} from its raw form to its virtual
3222 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3224 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
3225 @findex REGISTER_CONVERT_TO_RAW
3226 Convert the value of register @var{reg} from its virtual form to its raw
3228 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3230 @item RETURN_VALUE_ON_STACK(@var{type})
3231 @findex RETURN_VALUE_ON_STACK
3232 @cindex returning structures by value
3233 @cindex structures, returning by value
3235 Return non-zero if values of type TYPE are returned on the stack, using
3236 the ``struct convention'' (i.e., the caller provides a pointer to a
3237 buffer in which the callee should store the return value). This
3238 controls how the @samp{finish} command finds a function's return value,
3239 and whether an inferior function call reserves space on the stack for
3242 The full logic @value{GDBN} uses here is kind of odd.
3246 If the type being returned by value is not a structure, union, or array,
3247 and @code{RETURN_VALUE_ON_STACK} returns zero, then @value{GDBN}
3248 concludes the value is not returned using the struct convention.
3251 Otherwise, @value{GDBN} calls @code{USE_STRUCT_CONVENTION} (see below).
3252 If that returns non-zero, @value{GDBN} assumes the struct convention is
3256 In other words, to indicate that a given type is returned by value using
3257 the struct convention, that type must be either a struct, union, array,
3258 or something @code{RETURN_VALUE_ON_STACK} likes, @emph{and} something
3259 that @code{USE_STRUCT_CONVENTION} likes.
3261 Note that, in C and C@t{++}, arrays are never returned by value. In those
3262 languages, these predicates will always see a pointer type, never an
3263 array type. All the references above to arrays being returned by value
3264 apply only to other languages.
3266 @item SOFTWARE_SINGLE_STEP_P()
3267 @findex SOFTWARE_SINGLE_STEP_P
3268 Define this as 1 if the target does not have a hardware single-step
3269 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
3271 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breapoints_p})
3272 @findex SOFTWARE_SINGLE_STEP
3273 A function that inserts or removes (depending on
3274 @var{insert_breapoints_p}) breakpoints at each possible destinations of
3275 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
3278 @item SOFUN_ADDRESS_MAYBE_MISSING
3279 @findex SOFUN_ADDRESS_MAYBE_MISSING
3280 Somebody clever observed that, the more actual addresses you have in the
3281 debug information, the more time the linker has to spend relocating
3282 them. So whenever there's some other way the debugger could find the
3283 address it needs, you should omit it from the debug info, to make
3286 @code{SOFUN_ADDRESS_MAYBE_MISSING} indicates that a particular set of
3287 hacks of this sort are in use, affecting @code{N_SO} and @code{N_FUN}
3288 entries in stabs-format debugging information. @code{N_SO} stabs mark
3289 the beginning and ending addresses of compilation units in the text
3290 segment. @code{N_FUN} stabs mark the starts and ends of functions.
3292 @code{SOFUN_ADDRESS_MAYBE_MISSING} means two things:
3296 @code{N_FUN} stabs have an address of zero. Instead, you should find the
3297 addresses where the function starts by taking the function name from
3298 the stab, and then looking that up in the minsyms (the
3299 linker/assembler symbol table). In other words, the stab has the
3300 name, and the linker/assembler symbol table is the only place that carries
3304 @code{N_SO} stabs have an address of zero, too. You just look at the
3305 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab,
3306 and guess the starting and ending addresses of the compilation unit from
3310 @item PCC_SOL_BROKEN
3311 @findex PCC_SOL_BROKEN
3312 (Used only in the Convex target.)
3314 @item PC_IN_CALL_DUMMY
3315 @findex PC_IN_CALL_DUMMY
3316 See @file{inferior.h}.
3318 @item PC_LOAD_SEGMENT
3319 @findex PC_LOAD_SEGMENT
3320 If defined, print information about the load segment for the program
3321 counter. (Defined only for the RS/6000.)
3325 If the program counter is kept in a register, then define this macro to
3326 be the number (greater than or equal to zero) of that register.
3328 This should only need to be defined if @code{TARGET_READ_PC} and
3329 @code{TARGET_WRITE_PC} are not defined.
3333 The number of the ``next program counter'' register, if defined.
3337 The number of the ``next next program counter'' register, if defined.
3338 Currently, this is only defined for the Motorola 88K.
3341 @findex PARM_BOUNDARY
3342 If non-zero, round arguments to a boundary of this many bits before
3343 pushing them on the stack.
3345 @item PRINT_REGISTER_HOOK (@var{regno})
3346 @findex PRINT_REGISTER_HOOK
3347 If defined, this must be a function that prints the contents of the
3348 given register to standard output.
3350 @item PRINT_TYPELESS_INTEGER
3351 @findex PRINT_TYPELESS_INTEGER
3352 This is an obscure substitute for @code{print_longest} that seems to
3353 have been defined for the Convex target.
3355 @item PROCESS_LINENUMBER_HOOK
3356 @findex PROCESS_LINENUMBER_HOOK
3357 A hook defined for XCOFF reading.
3359 @item PROLOGUE_FIRSTLINE_OVERLAP
3360 @findex PROLOGUE_FIRSTLINE_OVERLAP
3361 (Only used in unsupported Convex configuration.)
3365 If defined, this is the number of the processor status register. (This
3366 definition is only used in generic code when parsing "$ps".)
3370 @findex call_function_by_hand
3371 @findex return_command
3372 Used in @samp{call_function_by_hand} to remove an artificial stack
3373 frame and in @samp{return_command} to remove a real stack frame.
3375 @item PUSH_ARGUMENTS (@var{nargs}, @var{args}, @var{sp}, @var{struct_return}, @var{struct_addr})
3376 @findex PUSH_ARGUMENTS
3377 Define this to push arguments onto the stack for inferior function
3378 call. Returns the updated stack pointer value.
3380 @item PUSH_DUMMY_FRAME
3381 @findex PUSH_DUMMY_FRAME
3382 Used in @samp{call_function_by_hand} to create an artificial stack frame.
3384 @item REGISTER_BYTES
3385 @findex REGISTER_BYTES
3386 The total amount of space needed to store @value{GDBN}'s copy of the machine's
3389 @item REGISTER_NAME(@var{i})
3390 @findex REGISTER_NAME
3391 Return the name of register @var{i} as a string. May return @code{NULL}
3392 or @code{NUL} to indicate that register @var{i} is not valid.
3394 @item REGISTER_NAMES
3395 @findex REGISTER_NAMES
3396 Deprecated in favor of @code{REGISTER_NAME}.
3398 @item REG_STRUCT_HAS_ADDR (@var{gcc_p}, @var{type})
3399 @findex REG_STRUCT_HAS_ADDR
3400 Define this to return 1 if the given type will be passed by pointer
3401 rather than directly.
3403 @item SAVE_DUMMY_FRAME_TOS (@var{sp})
3404 @findex SAVE_DUMMY_FRAME_TOS
3405 Used in @samp{call_function_by_hand} to notify the target dependent code
3406 of the top-of-stack value that will be passed to the the inferior code.
3407 This is the value of the @code{SP} after both the dummy frame and space
3408 for parameters/results have been allocated on the stack.
3410 @item SDB_REG_TO_REGNUM
3411 @findex SDB_REG_TO_REGNUM
3412 Define this to convert sdb register numbers into @value{GDBN} regnums. If not
3413 defined, no conversion will be done.
3415 @item SHIFT_INST_REGS
3416 @findex SHIFT_INST_REGS
3417 (Only used for m88k targets.)
3419 @item SKIP_PERMANENT_BREAKPOINT
3420 @findex SKIP_PERMANENT_BREAKPOINT
3421 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
3422 steps over a breakpoint by removing it, stepping one instruction, and
3423 re-inserting the breakpoint. However, permanent breakpoints are
3424 hardwired into the inferior, and can't be removed, so this strategy
3425 doesn't work. Calling @code{SKIP_PERMANENT_BREAKPOINT} adjusts the processor's
3426 state so that execution will resume just after the breakpoint. This
3427 macro does the right thing even when the breakpoint is in the delay slot
3428 of a branch or jump.
3430 @item SKIP_PROLOGUE (@var{pc})
3431 @findex SKIP_PROLOGUE
3432 A C expression that returns the address of the ``real'' code beyond the
3433 function entry prologue found at @var{pc}.
3435 @item SKIP_PROLOGUE_FRAMELESS_P
3436 @findex SKIP_PROLOGUE_FRAMELESS_P
3437 A C expression that should behave similarly, but that can stop as soon
3438 as the function is known to have a frame. If not defined,
3439 @code{SKIP_PROLOGUE} will be used instead.
3441 @item SKIP_TRAMPOLINE_CODE (@var{pc})
3442 @findex SKIP_TRAMPOLINE_CODE
3443 If the target machine has trampoline code that sits between callers and
3444 the functions being called, then define this macro to return a new PC
3445 that is at the start of the real function.
3449 If the stack-pointer is kept in a register, then define this macro to be
3450 the number (greater than or equal to zero) of that register.
3452 This should only need to be defined if @code{TARGET_WRITE_SP} and
3453 @code{TARGET_WRITE_SP} are not defined.
3455 @item STAB_REG_TO_REGNUM
3456 @findex STAB_REG_TO_REGNUM
3457 Define this to convert stab register numbers (as gotten from `r'
3458 declarations) into @value{GDBN} regnums. If not defined, no conversion will be
3461 @item STACK_ALIGN (@var{addr})
3463 Define this to adjust the address to the alignment required for the
3466 @item STEP_SKIPS_DELAY (@var{addr})
3467 @findex STEP_SKIPS_DELAY
3468 Define this to return true if the address is of an instruction with a
3469 delay slot. If a breakpoint has been placed in the instruction's delay
3470 slot, @value{GDBN} will single-step over that instruction before resuming
3471 normally. Currently only defined for the Mips.
3473 @item STORE_RETURN_VALUE (@var{type}, @var{valbuf})
3474 @findex STORE_RETURN_VALUE
3475 A C expression that stores a function return value of type @var{type},
3476 where @var{valbuf} is the address of the value to be stored.
3478 @item SUN_FIXED_LBRAC_BUG
3479 @findex SUN_FIXED_LBRAC_BUG
3480 (Used only for Sun-3 and Sun-4 targets.)
3482 @item SYMBOL_RELOADING_DEFAULT
3483 @findex SYMBOL_RELOADING_DEFAULT
3484 The default value of the ``symbol-reloading'' variable. (Never defined in
3487 @item TARGET_BYTE_ORDER_DEFAULT
3488 @findex TARGET_BYTE_ORDER_DEFAULT
3489 The ordering of bytes in the target. This must be either
3490 @code{BFD_ENDIAN_BIG} or @code{BFD_ENDIAN_LITTLE}. This macro replaces
3491 @code{TARGET_BYTE_ORDER} which is deprecated.
3493 @item TARGET_BYTE_ORDER_SELECTABLE_P
3494 @findex TARGET_BYTE_ORDER_SELECTABLE_P
3495 Non-zero if the target has both @code{BIG_ENDIAN} and
3496 @code{BFD_ENDIAN_LITTLE} variants. This macro replaces
3497 @code{TARGET_BYTE_ORDER_SELECTABLE} which is deprecated.
3499 @item TARGET_CHAR_BIT
3500 @findex TARGET_CHAR_BIT
3501 Number of bits in a char; defaults to 8.
3503 @item TARGET_CHAR_SIGNED
3504 @findex TARGET_CHAR_SIGNED
3505 Non-zero if @code{char} is normally signed on this architecture; zero if
3506 it should be unsigned.
3508 The ISO C standard requires the compiler to treat @code{char} as
3509 equivalent to either @code{signed char} or @code{unsigned char}; any
3510 character in the standard execution set is supposed to be positive.
3511 Most compilers treat @code{char} as signed, but @code{char} is unsigned
3512 on the IBM S/390, RS6000, and PowerPC targets.
3514 @item TARGET_COMPLEX_BIT
3515 @findex TARGET_COMPLEX_BIT
3516 Number of bits in a complex number; defaults to @code{2 * TARGET_FLOAT_BIT}.
3518 At present this macro is not used.
3520 @item TARGET_DOUBLE_BIT
3521 @findex TARGET_DOUBLE_BIT
3522 Number of bits in a double float; defaults to @code{8 * TARGET_CHAR_BIT}.
3524 @item TARGET_DOUBLE_COMPLEX_BIT
3525 @findex TARGET_DOUBLE_COMPLEX_BIT
3526 Number of bits in a double complex; defaults to @code{2 * TARGET_DOUBLE_BIT}.
3528 At present this macro is not used.
3530 @item TARGET_FLOAT_BIT
3531 @findex TARGET_FLOAT_BIT
3532 Number of bits in a float; defaults to @code{4 * TARGET_CHAR_BIT}.
3534 @item TARGET_INT_BIT
3535 @findex TARGET_INT_BIT
3536 Number of bits in an integer; defaults to @code{4 * TARGET_CHAR_BIT}.
3538 @item TARGET_LONG_BIT
3539 @findex TARGET_LONG_BIT
3540 Number of bits in a long integer; defaults to @code{4 * TARGET_CHAR_BIT}.
3542 @item TARGET_LONG_DOUBLE_BIT
3543 @findex TARGET_LONG_DOUBLE_BIT
3544 Number of bits in a long double float;
3545 defaults to @code{2 * TARGET_DOUBLE_BIT}.
3547 @item TARGET_LONG_LONG_BIT
3548 @findex TARGET_LONG_LONG_BIT
3549 Number of bits in a long long integer; defaults to @code{2 * TARGET_LONG_BIT}.
3551 @item TARGET_PTR_BIT
3552 @findex TARGET_PTR_BIT
3553 Number of bits in a pointer; defaults to @code{TARGET_INT_BIT}.
3555 @item TARGET_SHORT_BIT
3556 @findex TARGET_SHORT_BIT
3557 Number of bits in a short integer; defaults to @code{2 * TARGET_CHAR_BIT}.
3559 @item TARGET_READ_PC
3560 @findex TARGET_READ_PC
3561 @itemx TARGET_WRITE_PC (@var{val}, @var{pid})
3562 @findex TARGET_WRITE_PC
3563 @itemx TARGET_READ_SP
3564 @findex TARGET_READ_SP
3565 @itemx TARGET_WRITE_SP
3566 @findex TARGET_WRITE_SP
3567 @itemx TARGET_READ_FP
3568 @findex TARGET_READ_FP
3569 @itemx TARGET_WRITE_FP
3570 @findex TARGET_WRITE_FP
3577 These change the behavior of @code{read_pc}, @code{write_pc},
3578 @code{read_sp}, @code{write_sp}, @code{read_fp} and @code{write_fp}.
3579 For most targets, these may be left undefined. @value{GDBN} will call the read
3580 and write register functions with the relevant @code{_REGNUM} argument.
3582 These macros are useful when a target keeps one of these registers in a
3583 hard to get at place; for example, part in a segment register and part
3584 in an ordinary register.
3586 @item TARGET_VIRTUAL_FRAME_POINTER(@var{pc}, @var{regp}, @var{offsetp})
3587 @findex TARGET_VIRTUAL_FRAME_POINTER
3588 Returns a @code{(register, offset)} pair representing the virtual
3589 frame pointer in use at the code address @var{pc}. If virtual
3590 frame pointers are not used, a default definition simply returns
3591 @code{FP_REGNUM}, with an offset of zero.
3593 @item TARGET_HAS_HARDWARE_WATCHPOINTS
3594 If non-zero, the target has support for hardware-assisted
3595 watchpoints. @xref{Algorithms, watchpoints}, for more details and
3596 other related macros.
3598 @item TARGET_PRINT_INSN (@var{addr}, @var{info})
3599 @findex TARGET_PRINT_INSN
3600 This is the function used by @value{GDBN} to print an assembly
3601 instruction. It prints the instruction at address @var{addr} in
3602 debugged memory and returns the length of the instruction, in bytes. If
3603 a target doesn't define its own printing routine, it defaults to an
3604 accessor function for the global pointer @code{tm_print_insn}. This
3605 usually points to a function in the @code{opcodes} library (@pxref{Support
3606 Libraries, ,Opcodes}). @var{info} is a structure (of type
3607 @code{disassemble_info}) defined in @file{include/dis-asm.h} used to
3608 pass information to the instruction decoding routine.
3610 @item USE_STRUCT_CONVENTION (@var{gcc_p}, @var{type})
3611 @findex USE_STRUCT_CONVENTION
3612 If defined, this must be an expression that is nonzero if a value of the
3613 given @var{type} being returned from a function must have space
3614 allocated for it on the stack. @var{gcc_p} is true if the function
3615 being considered is known to have been compiled by GCC; this is helpful
3616 for systems where GCC is known to use different calling convention than
3619 @item VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
3620 @findex VARIABLES_INSIDE_BLOCK
3621 For dbx-style debugging information, if the compiler puts variable
3622 declarations inside LBRAC/RBRAC blocks, this should be defined to be
3623 nonzero. @var{desc} is the value of @code{n_desc} from the
3624 @code{N_RBRAC} symbol, and @var{gcc_p} is true if @value{GDBN} has noticed the
3625 presence of either the @code{GCC_COMPILED_SYMBOL} or the
3626 @code{GCC2_COMPILED_SYMBOL}. By default, this is 0.
3628 @item OS9K_VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
3629 @findex OS9K_VARIABLES_INSIDE_BLOCK
3630 Similarly, for OS/9000. Defaults to 1.
3633 Motorola M68K target conditionals.
3637 Define this to be the 4-bit location of the breakpoint trap vector. If
3638 not defined, it will default to @code{0xf}.
3640 @item REMOTE_BPT_VECTOR
3641 Defaults to @code{1}.
3644 @section Adding a New Target
3646 @cindex adding a target
3647 The following files add a target to @value{GDBN}:
3651 @item gdb/config/@var{arch}/@var{ttt}.mt
3652 Contains a Makefile fragment specific to this target. Specifies what
3653 object files are needed for target @var{ttt}, by defining
3654 @samp{TDEPFILES=@dots{}} and @samp{TDEPLIBS=@dots{}}. Also specifies
3655 the header file which describes @var{ttt}, by defining @samp{TM_FILE=
3658 You can also define @samp{TM_CFLAGS}, @samp{TM_CLIBS}, @samp{TM_CDEPS},
3659 but these are now deprecated, replaced by autoconf, and may go away in
3660 future versions of @value{GDBN}.
3662 @item gdb/@var{ttt}-tdep.c
3663 Contains any miscellaneous code required for this target machine. On
3664 some machines it doesn't exist at all. Sometimes the macros in
3665 @file{tm-@var{ttt}.h} become very complicated, so they are implemented
3666 as functions here instead, and the macro is simply defined to call the
3667 function. This is vastly preferable, since it is easier to understand
3670 @item gdb/@var{arch}-tdep.c
3671 @itemx gdb/@var{arch}-tdep.h
3672 This often exists to describe the basic layout of the target machine's
3673 processor chip (registers, stack, etc.). If used, it is included by
3674 @file{@var{ttt}-tdep.h}. It can be shared among many targets that use
3677 @item gdb/config/@var{arch}/tm-@var{ttt}.h
3678 (@file{tm.h} is a link to this file, created by @code{configure}). Contains
3679 macro definitions about the target machine's registers, stack frame
3680 format and instructions.
3682 New targets do not need this file and should not create it.
3684 @item gdb/config/@var{arch}/tm-@var{arch}.h
3685 This often exists to describe the basic layout of the target machine's
3686 processor chip (registers, stack, etc.). If used, it is included by
3687 @file{tm-@var{ttt}.h}. It can be shared among many targets that use the
3690 New targets do not need this file and should not create it.
3694 If you are adding a new operating system for an existing CPU chip, add a
3695 @file{config/tm-@var{os}.h} file that describes the operating system
3696 facilities that are unusual (extra symbol table info; the breakpoint
3697 instruction needed; etc.). Then write a @file{@var{arch}/tm-@var{os}.h}
3698 that just @code{#include}s @file{tm-@var{arch}.h} and
3699 @file{config/tm-@var{os}.h}.
3702 @node Target Vector Definition
3704 @chapter Target Vector Definition
3705 @cindex target vector
3707 The target vector defines the interface between @value{GDBN}'s
3708 abstract handling of target systems, and the nitty-gritty code that
3709 actually exercises control over a process or a serial port.
3710 @value{GDBN} includes some 30-40 different target vectors; however,
3711 each configuration of @value{GDBN} includes only a few of them.
3713 @section File Targets
3715 Both executables and core files have target vectors.
3717 @section Standard Protocol and Remote Stubs
3719 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code
3720 that runs in the target system. @value{GDBN} provides several sample
3721 @dfn{stubs} that can be integrated into target programs or operating
3722 systems for this purpose; they are named @file{*-stub.c}.
3724 The @value{GDBN} user's manual describes how to put such a stub into
3725 your target code. What follows is a discussion of integrating the
3726 SPARC stub into a complicated operating system (rather than a simple
3727 program), by Stu Grossman, the author of this stub.
3729 The trap handling code in the stub assumes the following upon entry to
3734 %l1 and %l2 contain pc and npc respectively at the time of the trap;
3740 you are in the correct trap window.
3743 As long as your trap handler can guarantee those conditions, then there
3744 is no reason why you shouldn't be able to ``share'' traps with the stub.
3745 The stub has no requirement that it be jumped to directly from the
3746 hardware trap vector. That is why it calls @code{exceptionHandler()},
3747 which is provided by the external environment. For instance, this could
3748 set up the hardware traps to actually execute code which calls the stub
3749 first, and then transfers to its own trap handler.
3751 For the most point, there probably won't be much of an issue with
3752 ``sharing'' traps, as the traps we use are usually not used by the kernel,
3753 and often indicate unrecoverable error conditions. Anyway, this is all
3754 controlled by a table, and is trivial to modify. The most important
3755 trap for us is for @code{ta 1}. Without that, we can't single step or
3756 do breakpoints. Everything else is unnecessary for the proper operation
3757 of the debugger/stub.
3759 From reading the stub, it's probably not obvious how breakpoints work.
3760 They are simply done by deposit/examine operations from @value{GDBN}.
3762 @section ROM Monitor Interface
3764 @section Custom Protocols
3766 @section Transport Layer
3768 @section Builtin Simulator
3771 @node Native Debugging
3773 @chapter Native Debugging
3774 @cindex native debugging
3776 Several files control @value{GDBN}'s configuration for native support:
3780 @item gdb/config/@var{arch}/@var{xyz}.mh
3781 Specifies Makefile fragments needed when hosting @emph{or native} on
3782 machine @var{xyz}. In particular, this lists the required
3783 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
3784 Also specifies the header file which describes native support on
3785 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
3786 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
3787 @samp{NAT_CDEPS}, etc.; see @file{Makefile.in}.
3789 @item gdb/config/@var{arch}/nm-@var{xyz}.h
3790 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
3791 macro definitions describing the native system environment, such as
3792 child process control and core file support.
3794 @item gdb/@var{xyz}-nat.c
3795 Contains any miscellaneous C code required for this native support of
3796 this machine. On some machines it doesn't exist at all.
3799 There are some ``generic'' versions of routines that can be used by
3800 various systems. These can be customized in various ways by macros
3801 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
3802 the @var{xyz} host, you can just include the generic file's name (with
3803 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
3805 Otherwise, if your machine needs custom support routines, you will need
3806 to write routines that perform the same functions as the generic file.
3807 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
3808 into @code{NATDEPFILES}.
3812 This contains the @emph{target_ops vector} that supports Unix child
3813 processes on systems which use ptrace and wait to control the child.
3816 This contains the @emph{target_ops vector} that supports Unix child
3817 processes on systems which use /proc to control the child.
3820 This does the low-level grunge that uses Unix system calls to do a ``fork
3821 and exec'' to start up a child process.
3824 This is the low level interface to inferior processes for systems using
3825 the Unix @code{ptrace} call in a vanilla way.
3828 @section Native core file Support
3829 @cindex native core files
3832 @findex fetch_core_registers
3833 @item core-aout.c::fetch_core_registers()
3834 Support for reading registers out of a core file. This routine calls
3835 @code{register_addr()}, see below. Now that BFD is used to read core
3836 files, virtually all machines should use @code{core-aout.c}, and should
3837 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
3838 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
3840 @item core-aout.c::register_addr()
3841 If your @code{nm-@var{xyz}.h} file defines the macro
3842 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
3843 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
3844 register number @code{regno}. @code{blockend} is the offset within the
3845 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
3846 @file{core-aout.c} will define the @code{register_addr()} function and
3847 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
3848 you are using the standard @code{fetch_core_registers()}, you will need
3849 to define your own version of @code{register_addr()}, put it into your
3850 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
3851 the @code{NATDEPFILES} list. If you have your own
3852 @code{fetch_core_registers()}, you may not need a separate
3853 @code{register_addr()}. Many custom @code{fetch_core_registers()}
3854 implementations simply locate the registers themselves.@refill
3857 When making @value{GDBN} run native on a new operating system, to make it
3858 possible to debug core files, you will need to either write specific
3859 code for parsing your OS's core files, or customize
3860 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
3861 machine uses to define the struct of registers that is accessible
3862 (possibly in the u-area) in a core file (rather than
3863 @file{machine/reg.h}), and an include file that defines whatever header
3864 exists on a core file (e.g. the u-area or a @code{struct core}). Then
3865 modify @code{trad_unix_core_file_p} to use these values to set up the
3866 section information for the data segment, stack segment, any other
3867 segments in the core file (perhaps shared library contents or control
3868 information), ``registers'' segment, and if there are two discontiguous
3869 sets of registers (e.g. integer and float), the ``reg2'' segment. This
3870 section information basically delimits areas in the core file in a
3871 standard way, which the section-reading routines in BFD know how to seek
3874 Then back in @value{GDBN}, you need a matching routine called
3875 @code{fetch_core_registers}. If you can use the generic one, it's in
3876 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
3877 It will be passed a char pointer to the entire ``registers'' segment,
3878 its length, and a zero; or a char pointer to the entire ``regs2''
3879 segment, its length, and a 2. The routine should suck out the supplied
3880 register values and install them into @value{GDBN}'s ``registers'' array.
3882 If your system uses @file{/proc} to control processes, and uses ELF
3883 format core files, then you may be able to use the same routines for
3884 reading the registers out of processes and out of core files.
3892 @section shared libraries
3894 @section Native Conditionals
3895 @cindex native conditionals
3897 When @value{GDBN} is configured and compiled, various macros are
3898 defined or left undefined, to control compilation when the host and
3899 target systems are the same. These macros should be defined (or left
3900 undefined) in @file{nm-@var{system}.h}.
3904 @findex ATTACH_DETACH
3905 If defined, then @value{GDBN} will include support for the @code{attach} and
3906 @code{detach} commands.
3908 @item CHILD_PREPARE_TO_STORE
3909 @findex CHILD_PREPARE_TO_STORE
3910 If the machine stores all registers at once in the child process, then
3911 define this to ensure that all values are correct. This usually entails
3912 a read from the child.
3914 [Note that this is incorrectly defined in @file{xm-@var{system}.h} files
3917 @item FETCH_INFERIOR_REGISTERS
3918 @findex FETCH_INFERIOR_REGISTERS
3919 Define this if the native-dependent code will provide its own routines
3920 @code{fetch_inferior_registers} and @code{store_inferior_registers} in
3921 @file{@var{host}-nat.c}. If this symbol is @emph{not} defined, and
3922 @file{infptrace.c} is included in this configuration, the default
3923 routines in @file{infptrace.c} are used for these functions.
3925 @item FILES_INFO_HOOK
3926 @findex FILES_INFO_HOOK
3927 (Only defined for Convex.)
3931 This macro is normally defined to be the number of the first floating
3932 point register, if the machine has such registers. As such, it would
3933 appear only in target-specific code. However, @file{/proc} support uses this
3934 to decide whether floats are in use on this target.
3936 @item GET_LONGJMP_TARGET
3937 @findex GET_LONGJMP_TARGET
3938 For most machines, this is a target-dependent parameter. On the
3939 DECstation and the Iris, this is a native-dependent parameter, since
3940 @file{setjmp.h} is needed to define it.
3942 This macro determines the target PC address that @code{longjmp} will jump to,
3943 assuming that we have just stopped at a longjmp breakpoint. It takes a
3944 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3945 pointer. It examines the current state of the machine as needed.
3947 @item I386_USE_GENERIC_WATCHPOINTS
3948 An x86-based machine can define this to use the generic x86 watchpoint
3949 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3952 @findex KERNEL_U_ADDR
3953 Define this to the address of the @code{u} structure (the ``user
3954 struct'', also known as the ``u-page'') in kernel virtual memory. @value{GDBN}
3955 needs to know this so that it can subtract this address from absolute
3956 addresses in the upage, that are obtained via ptrace or from core files.
3957 On systems that don't need this value, set it to zero.
3959 @item KERNEL_U_ADDR_BSD
3960 @findex KERNEL_U_ADDR_BSD
3961 Define this to cause @value{GDBN} to determine the address of @code{u} at
3962 runtime, by using Berkeley-style @code{nlist} on the kernel's image in
3965 @item KERNEL_U_ADDR_HPUX
3966 @findex KERNEL_U_ADDR_HPUX
3967 Define this to cause @value{GDBN} to determine the address of @code{u} at
3968 runtime, by using HP-style @code{nlist} on the kernel's image in the
3971 @item ONE_PROCESS_WRITETEXT
3972 @findex ONE_PROCESS_WRITETEXT
3973 Define this to be able to, when a breakpoint insertion fails, warn the
3974 user that another process may be running with the same executable.
3976 @item PREPARE_TO_PROCEED (@var{select_it})
3977 @findex PREPARE_TO_PROCEED
3978 This (ugly) macro allows a native configuration to customize the way the
3979 @code{proceed} function in @file{infrun.c} deals with switching between
3982 In a multi-threaded task we may select another thread and then continue
3983 or step. But if the old thread was stopped at a breakpoint, it will
3984 immediately cause another breakpoint stop without any execution (i.e. it
3985 will report a breakpoint hit incorrectly). So @value{GDBN} must step over it
3988 If defined, @code{PREPARE_TO_PROCEED} should check the current thread
3989 against the thread that reported the most recent event. If a step-over
3990 is required, it returns TRUE. If @var{select_it} is non-zero, it should
3991 reselect the old thread.
3994 @findex PROC_NAME_FMT
3995 Defines the format for the name of a @file{/proc} device. Should be
3996 defined in @file{nm.h} @emph{only} in order to override the default
3997 definition in @file{procfs.c}.
4000 @findex PTRACE_FP_BUG
4001 See @file{mach386-xdep.c}.
4003 @item PTRACE_ARG3_TYPE
4004 @findex PTRACE_ARG3_TYPE
4005 The type of the third argument to the @code{ptrace} system call, if it
4006 exists and is different from @code{int}.
4008 @item REGISTER_U_ADDR
4009 @findex REGISTER_U_ADDR
4010 Defines the offset of the registers in the ``u area''.
4012 @item SHELL_COMMAND_CONCAT
4013 @findex SHELL_COMMAND_CONCAT
4014 If defined, is a string to prefix on the shell command used to start the
4019 If defined, this is the name of the shell to use to run the inferior.
4020 Defaults to @code{"/bin/sh"}.
4022 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
4024 Define this to expand into an expression that will cause the symbols in
4025 @var{filename} to be added to @value{GDBN}'s symbol table. If
4026 @var{readsyms} is zero symbols are not read but any necessary low level
4027 processing for @var{filename} is still done.
4029 @item SOLIB_CREATE_INFERIOR_HOOK
4030 @findex SOLIB_CREATE_INFERIOR_HOOK
4031 Define this to expand into any shared-library-relocation code that you
4032 want to be run just after the child process has been forked.
4034 @item START_INFERIOR_TRAPS_EXPECTED
4035 @findex START_INFERIOR_TRAPS_EXPECTED
4036 When starting an inferior, @value{GDBN} normally expects to trap
4038 the shell execs, and once when the program itself execs. If the actual
4039 number of traps is something other than 2, then define this macro to
4040 expand into the number expected.
4042 @item SVR4_SHARED_LIBS
4043 @findex SVR4_SHARED_LIBS
4044 Define this to indicate that SVR4-style shared libraries are in use.
4048 This determines whether small routines in @file{*-tdep.c}, which
4049 translate register values between @value{GDBN}'s internal
4050 representation and the @file{/proc} representation, are compiled.
4053 @findex U_REGS_OFFSET
4054 This is the offset of the registers in the upage. It need only be
4055 defined if the generic ptrace register access routines in
4056 @file{infptrace.c} are being used (that is, @file{infptrace.c} is
4057 configured in, and @code{FETCH_INFERIOR_REGISTERS} is not defined). If
4058 the default value from @file{infptrace.c} is good enough, leave it
4061 The default value means that u.u_ar0 @emph{points to} the location of
4062 the registers. I'm guessing that @code{#define U_REGS_OFFSET 0} means
4063 that @code{u.u_ar0} @emph{is} the location of the registers.
4067 See @file{objfiles.c}.
4070 @findex DEBUG_PTRACE
4071 Define this to debug @code{ptrace} calls.
4075 @node Support Libraries
4077 @chapter Support Libraries
4082 BFD provides support for @value{GDBN} in several ways:
4085 @item identifying executable and core files
4086 BFD will identify a variety of file types, including a.out, coff, and
4087 several variants thereof, as well as several kinds of core files.
4089 @item access to sections of files
4090 BFD parses the file headers to determine the names, virtual addresses,
4091 sizes, and file locations of all the various named sections in files
4092 (such as the text section or the data section). @value{GDBN} simply
4093 calls BFD to read or write section @var{x} at byte offset @var{y} for
4096 @item specialized core file support
4097 BFD provides routines to determine the failing command name stored in a
4098 core file, the signal with which the program failed, and whether a core
4099 file matches (i.e.@: could be a core dump of) a particular executable
4102 @item locating the symbol information
4103 @value{GDBN} uses an internal interface of BFD to determine where to find the
4104 symbol information in an executable file or symbol-file. @value{GDBN} itself
4105 handles the reading of symbols, since BFD does not ``understand'' debug
4106 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
4111 @cindex opcodes library
4113 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
4114 library because it's also used in binutils, for @file{objdump}).
4123 @cindex regular expressions library
4134 @item SIGN_EXTEND_CHAR
4136 @item SWITCH_ENUM_BUG
4151 This chapter covers topics that are lower-level than the major
4152 algorithms of @value{GDBN}.
4157 Cleanups are a structured way to deal with things that need to be done
4160 When your code does something (e.g., @code{xmalloc} some memory, or
4161 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
4162 the memory or @code{close} the file), it can make a cleanup. The
4163 cleanup will be done at some future point: when the command is finished
4164 and control returns to the top level; when an error occurs and the stack
4165 is unwound; or when your code decides it's time to explicitly perform
4166 cleanups. Alternatively you can elect to discard the cleanups you
4172 @item struct cleanup *@var{old_chain};
4173 Declare a variable which will hold a cleanup chain handle.
4175 @findex make_cleanup
4176 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
4177 Make a cleanup which will cause @var{function} to be called with
4178 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
4179 handle that can later be passed to @code{do_cleanups} or
4180 @code{discard_cleanups}. Unless you are going to call
4181 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
4182 from @code{make_cleanup}.
4185 @item do_cleanups (@var{old_chain});
4186 Do all cleanups added to the chain since the corresponding
4187 @code{make_cleanup} call was made.
4189 @findex discard_cleanups
4190 @item discard_cleanups (@var{old_chain});
4191 Same as @code{do_cleanups} except that it just removes the cleanups from
4192 the chain and does not call the specified functions.
4195 Cleanups are implemented as a chain. The handle returned by
4196 @code{make_cleanups} includes the cleanup passed to the call and any
4197 later cleanups appended to the chain (but not yet discarded or
4201 make_cleanup (a, 0);
4203 struct cleanup *old = make_cleanup (b, 0);
4211 will call @code{c()} and @code{b()} but will not call @code{a()}. The
4212 cleanup that calls @code{a()} will remain in the cleanup chain, and will
4213 be done later unless otherwise discarded.@refill
4215 Your function should explicitly do or discard the cleanups it creates.
4216 Failing to do this leads to non-deterministic behavior since the caller
4217 will arbitrarily do or discard your functions cleanups. This need leads
4218 to two common cleanup styles.
4220 The first style is try/finally. Before it exits, your code-block calls
4221 @code{do_cleanups} with the old cleanup chain and thus ensures that your
4222 code-block's cleanups are always performed. For instance, the following
4223 code-segment avoids a memory leak problem (even when @code{error} is
4224 called and a forced stack unwind occurs) by ensuring that the
4225 @code{xfree} will always be called:
4228 struct cleanup *old = make_cleanup (null_cleanup, 0);
4229 data = xmalloc (sizeof blah);
4230 make_cleanup (xfree, data);
4235 The second style is try/except. Before it exits, your code-block calls
4236 @code{discard_cleanups} with the old cleanup chain and thus ensures that
4237 any created cleanups are not performed. For instance, the following
4238 code segment, ensures that the file will be closed but only if there is
4242 FILE *file = fopen ("afile", "r");
4243 struct cleanup *old = make_cleanup (close_file, file);
4245 discard_cleanups (old);
4249 Some functions, e.g. @code{fputs_filtered()} or @code{error()}, specify
4250 that they ``should not be called when cleanups are not in place''. This
4251 means that any actions you need to reverse in the case of an error or
4252 interruption must be on the cleanup chain before you call these
4253 functions, since they might never return to your code (they
4254 @samp{longjmp} instead).
4256 @section Wrapping Output Lines
4257 @cindex line wrap in output
4260 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
4261 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
4262 added in places that would be good breaking points. The utility
4263 routines will take care of actually wrapping if the line width is
4266 The argument to @code{wrap_here} is an indentation string which is
4267 printed @emph{only} if the line breaks there. This argument is saved
4268 away and used later. It must remain valid until the next call to
4269 @code{wrap_here} or until a newline has been printed through the
4270 @code{*_filtered} functions. Don't pass in a local variable and then
4273 It is usually best to call @code{wrap_here} after printing a comma or
4274 space. If you call it before printing a space, make sure that your
4275 indentation properly accounts for the leading space that will print if
4276 the line wraps there.
4278 Any function or set of functions that produce filtered output must
4279 finish by printing a newline, to flush the wrap buffer, before switching
4280 to unfiltered (@code{printf}) output. Symbol reading routines that
4281 print warnings are a good example.
4283 @section @value{GDBN} Coding Standards
4284 @cindex coding standards
4286 @value{GDBN} follows the GNU coding standards, as described in
4287 @file{etc/standards.texi}. This file is also available for anonymous
4288 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
4289 of the standard; in general, when the GNU standard recommends a practice
4290 but does not require it, @value{GDBN} requires it.
4292 @value{GDBN} follows an additional set of coding standards specific to
4293 @value{GDBN}, as described in the following sections.
4298 @value{GDBN} assumes an ISO-C compliant compiler.
4300 @value{GDBN} does not assume an ISO-C or POSIX compliant C library.
4303 @subsection Memory Management
4305 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
4306 @code{calloc}, @code{free} and @code{asprintf}.
4308 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
4309 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
4310 these functions do not return when the memory pool is empty. Instead,
4311 they unwind the stack using cleanups. These functions return
4312 @code{NULL} when requested to allocate a chunk of memory of size zero.
4314 @emph{Pragmatics: By using these functions, the need to check every
4315 memory allocation is removed. These functions provide portable
4318 @value{GDBN} does not use the function @code{free}.
4320 @value{GDBN} uses the function @code{xfree} to return memory to the
4321 memory pool. Consistent with ISO-C, this function ignores a request to
4322 free a @code{NULL} pointer.
4324 @emph{Pragmatics: On some systems @code{free} fails when passed a
4325 @code{NULL} pointer.}
4327 @value{GDBN} can use the non-portable function @code{alloca} for the
4328 allocation of small temporary values (such as strings).
4330 @emph{Pragmatics: This function is very non-portable. Some systems
4331 restrict the memory being allocated to no more than a few kilobytes.}
4333 @value{GDBN} uses the string function @code{xstrdup} and the print
4334 function @code{xasprintf}.
4336 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
4337 functions such as @code{sprintf} are very prone to buffer overflow
4341 @subsection Compiler Warnings
4342 @cindex compiler warnings
4344 With few exceptions, developers should include the configuration option
4345 @samp{--enable-gdb-build-warnings=,-Werror} when building @value{GDBN}.
4346 The exceptions are listed in the file @file{gdb/MAINTAINERS}.
4348 This option causes @value{GDBN} (when built using GCC) to be compiled
4349 with a carefully selected list of compiler warning flags. Any warnings
4350 from those flags being treated as errors.
4352 The current list of warning flags includes:
4356 Since @value{GDBN} coding standard requires all functions to be declared
4357 using a prototype, the flag has the side effect of ensuring that
4358 prototyped functions are always visible with out resorting to
4359 @samp{-Wstrict-prototypes}.
4362 Such code often appears to work except on instruction set architectures
4363 that use register windows.
4370 Since @value{GDBN} uses the @code{format printf} attribute on all
4371 @code{printf} like functions this checks not just @code{printf} calls
4372 but also calls to functions such as @code{fprintf_unfiltered}.
4375 This warning includes uses of the assignment operator within an
4376 @code{if} statement.
4378 @item -Wpointer-arith
4380 @item -Wuninitialized
4383 @emph{Pragmatics: Due to the way that @value{GDBN} is implemented most
4384 functions have unused parameters. Consequently the warning
4385 @samp{-Wunused-parameter} is precluded from the list. The macro
4386 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
4387 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
4388 is being used. The options @samp{-Wall} and @samp{-Wunused} are also
4389 precluded because they both include @samp{-Wunused-parameter}.}
4391 @emph{Pragmatics: @value{GDBN} has not simply accepted the warnings
4392 enabled by @samp{-Wall -Werror -W...}. Instead it is selecting warnings
4393 when and where their benefits can be demonstrated.}
4395 @subsection Formatting
4397 @cindex source code formatting
4398 The standard GNU recommendations for formatting must be followed
4401 A function declaration should not have its name in column zero. A
4402 function definition should have its name in column zero.
4406 static void foo (void);
4414 @emph{Pragmatics: This simplifies scripting. Function definitions can
4415 be found using @samp{^function-name}.}
4417 There must be a space between a function or macro name and the opening
4418 parenthesis of its argument list (except for macro definitions, as
4419 required by C). There must not be a space after an open paren/bracket
4420 or before a close paren/bracket.
4422 While additional whitespace is generally helpful for reading, do not use
4423 more than one blank line to separate blocks, and avoid adding whitespace
4424 after the end of a program line (as of 1/99, some 600 lines had
4425 whitespace after the semicolon). Excess whitespace causes difficulties
4426 for @code{diff} and @code{patch} utilities.
4428 Pointers are declared using the traditional K&R C style:
4442 @subsection Comments
4444 @cindex comment formatting
4445 The standard GNU requirements on comments must be followed strictly.
4447 Block comments must appear in the following form, with no @code{/*}- or
4448 @code{*/}-only lines, and no leading @code{*}:
4451 /* Wait for control to return from inferior to debugger. If inferior
4452 gets a signal, we may decide to start it up again instead of
4453 returning. That is why there is a loop in this function. When
4454 this function actually returns it means the inferior should be left
4455 stopped and @value{GDBN} should read more commands. */
4458 (Note that this format is encouraged by Emacs; tabbing for a multi-line
4459 comment works correctly, and @kbd{M-q} fills the block consistently.)
4461 Put a blank line between the block comments preceding function or
4462 variable definitions, and the definition itself.
4464 In general, put function-body comments on lines by themselves, rather
4465 than trying to fit them into the 20 characters left at the end of a
4466 line, since either the comment or the code will inevitably get longer
4467 than will fit, and then somebody will have to move it anyhow.
4471 @cindex C data types
4472 Code must not depend on the sizes of C data types, the format of the
4473 host's floating point numbers, the alignment of anything, or the order
4474 of evaluation of expressions.
4476 @cindex function usage
4477 Use functions freely. There are only a handful of compute-bound areas
4478 in @value{GDBN} that might be affected by the overhead of a function
4479 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
4480 limited by the target interface (whether serial line or system call).
4482 However, use functions with moderation. A thousand one-line functions
4483 are just as hard to understand as a single thousand-line function.
4485 @emph{Macros are bad, M'kay.}
4486 (But if you have to use a macro, make sure that the macro arguments are
4487 protected with parentheses.)
4491 Declarations like @samp{struct foo *} should be used in preference to
4492 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
4495 @subsection Function Prototypes
4496 @cindex function prototypes
4498 Prototypes must be used when both @emph{declaring} and @emph{defining}
4499 a function. Prototypes for @value{GDBN} functions must include both the
4500 argument type and name, with the name matching that used in the actual
4501 function definition.
4503 All external functions should have a declaration in a header file that
4504 callers include, except for @code{_initialize_*} functions, which must
4505 be external so that @file{init.c} construction works, but shouldn't be
4506 visible to random source files.
4508 Where a source file needs a forward declaration of a static function,
4509 that declaration must appear in a block near the top of the source file.
4512 @subsection Internal Error Recovery
4514 During its execution, @value{GDBN} can encounter two types of errors.
4515 User errors and internal errors. User errors include not only a user
4516 entering an incorrect command but also problems arising from corrupt
4517 object files and system errors when interacting with the target.
4518 Internal errors include situtations where @value{GDBN} has detected, at
4519 run time, a corrupt or erroneous situtation.
4521 When reporting an internal error, @value{GDBN} uses
4522 @code{internal_error} and @code{gdb_assert}.
4524 @value{GDBN} must not call @code{abort} or @code{assert}.
4526 @emph{Pragmatics: There is no @code{internal_warning} function. Either
4527 the code detected a user error, recovered from it and issued a
4528 @code{warning} or the code failed to correctly recover from the user
4529 error and issued an @code{internal_error}.}
4531 @subsection File Names
4533 Any file used when building the core of @value{GDBN} must be in lower
4534 case. Any file used when building the core of @value{GDBN} must be 8.3
4535 unique. These requirements apply to both source and generated files.
4537 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
4538 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
4539 is introduced to the build process both @file{Makefile.in} and
4540 @file{configure.in} need to be modified accordingly. Compare the
4541 convoluted conversion process needed to transform @file{COPYING} into
4542 @file{copying.c} with the conversion needed to transform
4543 @file{version.in} into @file{version.c}.}
4545 Any file non 8.3 compliant file (that is not used when building the core
4546 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
4548 @emph{Pragmatics: This is clearly a compromise.}
4550 When @value{GDBN} has a local version of a system header file (ex
4551 @file{string.h}) the file name based on the POSIX header prefixed with
4552 @file{gdb_} (@file{gdb_string.h}).
4554 For other files @samp{-} is used as the separator.
4557 @subsection Include Files
4559 All @file{.c} files should include @file{defs.h} first.
4561 All @file{.c} files should explicitly include the headers for any
4562 declarations they refer to. They should not rely on files being
4563 included indirectly.
4565 With the exception of the global definitions supplied by @file{defs.h},
4566 a header file should explictily include the header declaring any
4567 @code{typedefs} et.al.@: it refers to.
4569 @code{extern} declarations should never appear in @code{.c} files.
4571 All include files should be wrapped in:
4574 #ifndef INCLUDE_FILE_NAME_H
4575 #define INCLUDE_FILE_NAME_H
4581 @subsection Clean Design and Portable Implementation
4584 In addition to getting the syntax right, there's the little question of
4585 semantics. Some things are done in certain ways in @value{GDBN} because long
4586 experience has shown that the more obvious ways caused various kinds of
4589 @cindex assumptions about targets
4590 You can't assume the byte order of anything that comes from a target
4591 (including @var{value}s, object files, and instructions). Such things
4592 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
4593 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
4594 such as @code{bfd_get_32}.
4596 You can't assume that you know what interface is being used to talk to
4597 the target system. All references to the target must go through the
4598 current @code{target_ops} vector.
4600 You can't assume that the host and target machines are the same machine
4601 (except in the ``native'' support modules). In particular, you can't
4602 assume that the target machine's header files will be available on the
4603 host machine. Target code must bring along its own header files --
4604 written from scratch or explicitly donated by their owner, to avoid
4608 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
4609 to write the code portably than to conditionalize it for various
4612 @cindex system dependencies
4613 New @code{#ifdef}'s which test for specific compilers or manufacturers
4614 or operating systems are unacceptable. All @code{#ifdef}'s should test
4615 for features. The information about which configurations contain which
4616 features should be segregated into the configuration files. Experience
4617 has proven far too often that a feature unique to one particular system
4618 often creeps into other systems; and that a conditional based on some
4619 predefined macro for your current system will become worthless over
4620 time, as new versions of your system come out that behave differently
4621 with regard to this feature.
4623 Adding code that handles specific architectures, operating systems,
4624 target interfaces, or hosts, is not acceptable in generic code.
4626 @cindex portable file name handling
4627 @cindex file names, portability
4628 One particularly notorious area where system dependencies tend to
4629 creep in is handling of file names. The mainline @value{GDBN} code
4630 assumes Posix semantics of file names: absolute file names begin with
4631 a forward slash @file{/}, slashes are used to separate leading
4632 directories, case-sensitive file names. These assumptions are not
4633 necessarily true on non-Posix systems such as MS-Windows. To avoid
4634 system-dependent code where you need to take apart or construct a file
4635 name, use the following portable macros:
4638 @findex HAVE_DOS_BASED_FILE_SYSTEM
4639 @item HAVE_DOS_BASED_FILE_SYSTEM
4640 This preprocessing symbol is defined to a non-zero value on hosts
4641 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
4642 symbol to write conditional code which should only be compiled for
4645 @findex IS_DIR_SEPARATOR
4646 @item IS_DIR_SEPARATOR (@var{c}
4647 Evaluates to a non-zero value if @var{c} is a directory separator
4648 character. On Unix and GNU/Linux systems, only a slash @file{/} is
4649 such a character, but on Windows, both @file{/} and @file{\} will
4652 @findex IS_ABSOLUTE_PATH
4653 @item IS_ABSOLUTE_PATH (@var{file})
4654 Evaluates to a non-zero value if @var{file} is an absolute file name.
4655 For Unix and GNU/Linux hosts, a name which begins with a slash
4656 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
4657 @file{x:\bar} are also absolute file names.
4659 @findex FILENAME_CMP
4660 @item FILENAME_CMP (@var{f1}, @var{f2})
4661 Calls a function which compares file names @var{f1} and @var{f2} as
4662 appropriate for the underlying host filesystem. For Posix systems,
4663 this simply calls @code{strcmp}; on case-insensitive filesystems it
4664 will call @code{strcasecmp} instead.
4666 @findex DIRNAME_SEPARATOR
4667 @item DIRNAME_SEPARATOR
4668 Evaluates to a character which separates directories in
4669 @code{PATH}-style lists, typically held in environment variables.
4670 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
4672 @findex SLASH_STRING
4674 This evaluates to a constant string you should use to produce an
4675 absolute filename from leading directories and the file's basename.
4676 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
4677 @code{"\\"} for some Windows-based ports.
4680 In addition to using these macros, be sure to use portable library
4681 functions whenever possible. For example, to extract a directory or a
4682 basename part from a file name, use the @code{dirname} and
4683 @code{basename} library functions (available in @code{libiberty} for
4684 platforms which don't provide them), instead of searching for a slash
4685 with @code{strrchr}.
4687 Another way to generalize @value{GDBN} along a particular interface is with an
4688 attribute struct. For example, @value{GDBN} has been generalized to handle
4689 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
4690 by defining the @code{target_ops} structure and having a current target (as
4691 well as a stack of targets below it, for memory references). Whenever
4692 something needs to be done that depends on which remote interface we are
4693 using, a flag in the current target_ops structure is tested (e.g.,
4694 @code{target_has_stack}), or a function is called through a pointer in the
4695 current target_ops structure. In this way, when a new remote interface
4696 is added, only one module needs to be touched---the one that actually
4697 implements the new remote interface. Other examples of
4698 attribute-structs are BFD access to multiple kinds of object file
4699 formats, or @value{GDBN}'s access to multiple source languages.
4701 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
4702 the code interfacing between @code{ptrace} and the rest of
4703 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
4704 something was very painful. In @value{GDBN} 4.x, these have all been
4705 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
4706 with variations between systems the same way any system-independent
4707 file would (hooks, @code{#if defined}, etc.), and machines which are
4708 radically different don't need to use @file{infptrace.c} at all.
4710 All debugging code must be controllable using the @samp{set debug
4711 @var{module}} command. Do not use @code{printf} to print trace
4712 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
4713 @code{#ifdef DEBUG}.
4718 @chapter Porting @value{GDBN}
4719 @cindex porting to new machines
4721 Most of the work in making @value{GDBN} compile on a new machine is in
4722 specifying the configuration of the machine. This is done in a
4723 dizzying variety of header files and configuration scripts, which we
4724 hope to make more sensible soon. Let's say your new host is called an
4725 @var{xyz} (e.g., @samp{sun4}), and its full three-part configuration
4726 name is @code{@var{arch}-@var{xvend}-@var{xos}} (e.g.,
4727 @samp{sparc-sun-sunos4}). In particular:
4731 In the top level directory, edit @file{config.sub} and add @var{arch},
4732 @var{xvend}, and @var{xos} to the lists of supported architectures,
4733 vendors, and operating systems near the bottom of the file. Also, add
4734 @var{xyz} as an alias that maps to
4735 @code{@var{arch}-@var{xvend}-@var{xos}}. You can test your changes by
4739 ./config.sub @var{xyz}
4746 ./config.sub @code{@var{arch}-@var{xvend}-@var{xos}}
4750 which should both respond with @code{@var{arch}-@var{xvend}-@var{xos}}
4751 and no error messages.
4754 You need to port BFD, if that hasn't been done already. Porting BFD is
4755 beyond the scope of this manual.
4758 To configure @value{GDBN} itself, edit @file{gdb/configure.host} to recognize
4759 your system and set @code{gdb_host} to @var{xyz}, and (unless your
4760 desired target is already available) also edit @file{gdb/configure.tgt},
4761 setting @code{gdb_target} to something appropriate (for instance,
4765 Finally, you'll need to specify and define @value{GDBN}'s host-, native-, and
4766 target-dependent @file{.h} and @file{.c} files used for your
4770 @section Configuring @value{GDBN} for Release
4772 @cindex preparing a release
4773 @cindex making a distribution tarball
4774 From the top level directory (containing @file{gdb}, @file{bfd},
4775 @file{libiberty}, and so on):
4778 make -f Makefile.in gdb.tar.gz
4782 This will properly configure, clean, rebuild any files that are
4783 distributed pre-built (e.g. @file{c-exp.tab.c} or @file{refcard.ps}),
4784 and will then make a tarfile. (If the top level directory has already
4785 been configured, you can just do @code{make gdb.tar.gz} instead.)
4787 This procedure requires:
4795 @code{makeinfo} (texinfo2 level);
4804 @code{yacc} or @code{bison}.
4808 @dots{} and the usual slew of utilities (@code{sed}, @code{tar}, etc.).
4810 @subheading TEMPORARY RELEASE PROCEDURE FOR DOCUMENTATION
4812 @file{gdb.texinfo} is currently marked up using the texinfo-2 macros,
4813 which are not yet a default for anything (but we have to start using
4816 For making paper, the only thing this implies is the right generation of
4817 @file{texinfo.tex} needs to be included in the distribution.
4819 For making info files, however, rather than duplicating the texinfo2
4820 distribution, generate @file{gdb-all.texinfo} locally, and include the
4821 files @file{gdb.info*} in the distribution. Note the plural;
4822 @code{makeinfo} will split the document into one overall file and five
4823 or so included files.
4827 @chapter Releasing @value{GDBN}
4828 @cindex making a new release of gdb
4830 @section Before the branch
4832 The most important objective at this stage is to find and fix simple
4833 changes that become a pain to track once the branch is created. For
4834 instance, configuration problems that stop @value{GDBN} from even
4835 building. If you can't get the problem fixed, document it in the
4836 @file{PROBLEMS} file.
4838 @subheading Obsolete any code
4840 Mark as @kbd{OBSOLETE} any uninteresting targets or code files. This
4841 has a number of steps and is slow --- mainly to ensure that people have
4842 had a reasonable chance to respond. Remember, everything on the
4843 internet takes a week.
4847 announce the change on @email{gdb@@sources.redhat.com, GDB mailing list}
4851 announce the change on @email{gdb-announce@@sources.redhat.com, GDB
4852 Announcement mailing list}
4856 post / commit the change
4859 @subheading Refresh any imported files.
4861 A number of files are taken from external repositories. They include:
4865 @file{texinfo/texinfo.tex}
4867 @file{config.guess} et.@: al.@:
4870 and should be refreshed.
4872 @subheading Organize and announce the schedule.
4874 The following is a possible schedule. It is based on the rule-of-thumb
4875 that everything on the Internet takes a week. You may want to even
4876 increase those times further since an analysis of the actual data
4877 strongly suggests that the below is far to aggressive.
4885 announce branch date
4893 start enjoying all the fun
4896 As an aside, the branch tag name is probably regrettable vis:
4897 @file{gdb_N_M-YYYY-MM-DD-@{branch,branchpoint@}}.
4900 @section Building a Release
4902 @subheading Establish a few defaults.
4905 $ b=gdb_5_1_0_1-2002-01-03-branch
4907 $ cd /sourceware/snapshot-tmp/gdbadmin-tmp/$b
4909 /home/gdbadmin/bin/autoconf
4912 NB: Check the autoconf version carefully. You want to be using
4913 @file{gdbadmin}'s version (which is really the version taken from the
4914 binutils snapshot). SWARE may have a different version installed.
4916 @subheading Check out the relevant modules:
4919 $ for m in gdb insight dejagnu; do
4920 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
4924 NB: The reading of @file{.cvsrc} is disabled (@file{-f}) so that there
4925 isn't any confusion between what is written here and what CVS really
4928 @subheading Update the file @file{gdb/version.in} where applicable.
4931 $ for m in gdb insight; do echo $v > $m/src/gdb/version.in ; done
4935 @subheading Mutter something about creating a @file{ChangeLog} entry. (both trunk and branch).
4938 $ emacs gdb/src/gdb/version.in
4940 Bump version to 5.1.0.1.
4944 ditto for @file{insight/src/gdb/version.in}
4946 @subheading Mutter something about updating @file{README}
4948 For dejagnu, edit @file{dejagnu/src/dejagnu/configure.in} and set it to
4949 gdb-$v and then regenerate configure. Mention this in the dejagnu
4953 $ emacs dejagnu/src/dejagnu/configure.in
4956 Bump version to 5.1.0.1.
4957 * configure: Re-generate.
4959 $ ( cd dejagnu/src/dejagnu && autoconf )
4962 @subheading Build the snapshot:
4965 $ for m in gdb insight dejagnu; do
4966 ( cd $m/src && gmake -f Makefile.in $m.tar.bz2 )
4970 @subheading Do another @kbd{CVS update} to see what the damage is.
4973 $ ( cd gdb/src && cvs -q update )
4976 You're looking for files that have mysteriously disappeared as the
4977 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
4978 for the @file{version.in} update cronjob.
4980 @subheading Copy all the @file{.bz2} files to the ftp directory:
4983 cp */src/*.bz2 ~ftp/.....
4986 @subheading Something about @kbd{gzip}'ing them.
4988 @subheading Something about web pages?
4990 @subheading Something about documentation?
4992 @subheading Cleanup the release tree
4994 In particular you'll need to:
4998 Commit the changes to @file{ChangeLog} and @file{version.in}
5004 @section After the release
5006 Remove any @kbd{OBSOLETE} code.
5014 The testsuite is an important component of the @value{GDBN} package.
5015 While it is always worthwhile to encourage user testing, in practice
5016 this is rarely sufficient; users typically use only a small subset of
5017 the available commands, and it has proven all too common for a change
5018 to cause a significant regression that went unnoticed for some time.
5020 The @value{GDBN} testsuite uses the DejaGNU testing framework.
5021 DejaGNU is built using @code{Tcl} and @code{expect}. The tests
5022 themselves are calls to various @code{Tcl} procs; the framework runs all the
5023 procs and summarizes the passes and fails.
5025 @section Using the Testsuite
5027 @cindex running the test suite
5028 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
5029 testsuite's objdir) and type @code{make check}. This just sets up some
5030 environment variables and invokes DejaGNU's @code{runtest} script. While
5031 the testsuite is running, you'll get mentions of which test file is in use,
5032 and a mention of any unexpected passes or fails. When the testsuite is
5033 finished, you'll get a summary that looks like this:
5038 # of expected passes 6016
5039 # of unexpected failures 58
5040 # of unexpected successes 5
5041 # of expected failures 183
5042 # of unresolved testcases 3
5043 # of untested testcases 5
5046 The ideal test run consists of expected passes only; however, reality
5047 conspires to keep us from this ideal. Unexpected failures indicate
5048 real problems, whether in @value{GDBN} or in the testsuite. Expected
5049 failures are still failures, but ones which have been decided are too
5050 hard to deal with at the time; for instance, a test case might work
5051 everywhere except on AIX, and there is no prospect of the AIX case
5052 being fixed in the near future. Expected failures should not be added
5053 lightly, since you may be masking serious bugs in @value{GDBN}.
5054 Unexpected successes are expected fails that are passing for some
5055 reason, while unresolved and untested cases often indicate some minor
5056 catastrophe, such as the compiler being unable to deal with a test
5059 When making any significant change to @value{GDBN}, you should run the
5060 testsuite before and after the change, to confirm that there are no
5061 regressions. Note that truly complete testing would require that you
5062 run the testsuite with all supported configurations and a variety of
5063 compilers; however this is more than really necessary. In many cases
5064 testing with a single configuration is sufficient. Other useful
5065 options are to test one big-endian (Sparc) and one little-endian (x86)
5066 host, a cross config with a builtin simulator (powerpc-eabi,
5067 mips-elf), or a 64-bit host (Alpha).
5069 If you add new functionality to @value{GDBN}, please consider adding
5070 tests for it as well; this way future @value{GDBN} hackers can detect
5071 and fix their changes that break the functionality you added.
5072 Similarly, if you fix a bug that was not previously reported as a test
5073 failure, please add a test case for it. Some cases are extremely
5074 difficult to test, such as code that handles host OS failures or bugs
5075 in particular versions of compilers, and it's OK not to try to write
5076 tests for all of those.
5078 @section Testsuite Organization
5080 @cindex test suite organization
5081 The testsuite is entirely contained in @file{gdb/testsuite}. While the
5082 testsuite includes some makefiles and configury, these are very minimal,
5083 and used for little besides cleaning up, since the tests themselves
5084 handle the compilation of the programs that @value{GDBN} will run. The file
5085 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
5086 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
5087 configuration-specific files, typically used for special-purpose
5088 definitions of procs like @code{gdb_load} and @code{gdb_start}.
5090 The tests themselves are to be found in @file{testsuite/gdb.*} and
5091 subdirectories of those. The names of the test files must always end
5092 with @file{.exp}. DejaGNU collects the test files by wildcarding
5093 in the test directories, so both subdirectories and individual files
5094 get chosen and run in alphabetical order.
5096 The following table lists the main types of subdirectories and what they
5097 are for. Since DejaGNU finds test files no matter where they are
5098 located, and since each test file sets up its own compilation and
5099 execution environment, this organization is simply for convenience and
5104 This is the base testsuite. The tests in it should apply to all
5105 configurations of @value{GDBN} (but generic native-only tests may live here).
5106 The test programs should be in the subset of C that is valid K&R,
5107 ANSI/ISO, and C++ (@code{#ifdef}s are allowed if necessary, for instance
5110 @item gdb.@var{lang}
5111 Language-specific tests for any language @var{lang} besides C. Examples are
5112 @file{gdb.c++} and @file{gdb.java}.
5114 @item gdb.@var{platform}
5115 Non-portable tests. The tests are specific to a specific configuration
5116 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
5119 @item gdb.@var{compiler}
5120 Tests specific to a particular compiler. As of this writing (June
5121 1999), there aren't currently any groups of tests in this category that
5122 couldn't just as sensibly be made platform-specific, but one could
5123 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
5126 @item gdb.@var{subsystem}
5127 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
5128 instance, @file{gdb.disasm} exercises various disassemblers, while
5129 @file{gdb.stabs} tests pathways through the stabs symbol reader.
5132 @section Writing Tests
5133 @cindex writing tests
5135 In many areas, the @value{GDBN} tests are already quite comprehensive; you
5136 should be able to copy existing tests to handle new cases.
5138 You should try to use @code{gdb_test} whenever possible, since it
5139 includes cases to handle all the unexpected errors that might happen.
5140 However, it doesn't cost anything to add new test procedures; for
5141 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
5142 calls @code{gdb_test} multiple times.
5144 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
5145 necessary, such as when @value{GDBN} has several valid responses to a command.
5147 The source language programs do @emph{not} need to be in a consistent
5148 style. Since @value{GDBN} is used to debug programs written in many different
5149 styles, it's worth having a mix of styles in the testsuite; for
5150 instance, some @value{GDBN} bugs involving the display of source lines would
5151 never manifest themselves if the programs used GNU coding style
5158 Check the @file{README} file, it often has useful information that does not
5159 appear anywhere else in the directory.
5162 * Getting Started:: Getting started working on @value{GDBN}
5163 * Debugging GDB:: Debugging @value{GDBN} with itself
5166 @node Getting Started,,, Hints
5168 @section Getting Started
5170 @value{GDBN} is a large and complicated program, and if you first starting to
5171 work on it, it can be hard to know where to start. Fortunately, if you
5172 know how to go about it, there are ways to figure out what is going on.
5174 This manual, the @value{GDBN} Internals manual, has information which applies
5175 generally to many parts of @value{GDBN}.
5177 Information about particular functions or data structures are located in
5178 comments with those functions or data structures. If you run across a
5179 function or a global variable which does not have a comment correctly
5180 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
5181 free to submit a bug report, with a suggested comment if you can figure
5182 out what the comment should say. If you find a comment which is
5183 actually wrong, be especially sure to report that.
5185 Comments explaining the function of macros defined in host, target, or
5186 native dependent files can be in several places. Sometimes they are
5187 repeated every place the macro is defined. Sometimes they are where the
5188 macro is used. Sometimes there is a header file which supplies a
5189 default definition of the macro, and the comment is there. This manual
5190 also documents all the available macros.
5191 @c (@pxref{Host Conditionals}, @pxref{Target
5192 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
5195 Start with the header files. Once you have some idea of how
5196 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
5197 @file{gdbtypes.h}), you will find it much easier to understand the
5198 code which uses and creates those symbol tables.
5200 You may wish to process the information you are getting somehow, to
5201 enhance your understanding of it. Summarize it, translate it to another
5202 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
5203 the code to predict what a test case would do and write the test case
5204 and verify your prediction, etc. If you are reading code and your eyes
5205 are starting to glaze over, this is a sign you need to use a more active
5208 Once you have a part of @value{GDBN} to start with, you can find more
5209 specifically the part you are looking for by stepping through each
5210 function with the @code{next} command. Do not use @code{step} or you
5211 will quickly get distracted; when the function you are stepping through
5212 calls another function try only to get a big-picture understanding
5213 (perhaps using the comment at the beginning of the function being
5214 called) of what it does. This way you can identify which of the
5215 functions being called by the function you are stepping through is the
5216 one which you are interested in. You may need to examine the data
5217 structures generated at each stage, with reference to the comments in
5218 the header files explaining what the data structures are supposed to
5221 Of course, this same technique can be used if you are just reading the
5222 code, rather than actually stepping through it. The same general
5223 principle applies---when the code you are looking at calls something
5224 else, just try to understand generally what the code being called does,
5225 rather than worrying about all its details.
5227 @cindex command implementation
5228 A good place to start when tracking down some particular area is with
5229 a command which invokes that feature. Suppose you want to know how
5230 single-stepping works. As a @value{GDBN} user, you know that the
5231 @code{step} command invokes single-stepping. The command is invoked
5232 via command tables (see @file{command.h}); by convention the function
5233 which actually performs the command is formed by taking the name of
5234 the command and adding @samp{_command}, or in the case of an
5235 @code{info} subcommand, @samp{_info}. For example, the @code{step}
5236 command invokes the @code{step_command} function and the @code{info
5237 display} command invokes @code{display_info}. When this convention is
5238 not followed, you might have to use @code{grep} or @kbd{M-x
5239 tags-search} in emacs, or run @value{GDBN} on itself and set a
5240 breakpoint in @code{execute_command}.
5242 @cindex @code{bug-gdb} mailing list
5243 If all of the above fail, it may be appropriate to ask for information
5244 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
5245 wondering if anyone could give me some tips about understanding
5246 @value{GDBN}''---if we had some magic secret we would put it in this manual.
5247 Suggestions for improving the manual are always welcome, of course.
5249 @node Debugging GDB,,,Hints
5251 @section Debugging @value{GDBN} with itself
5252 @cindex debugging @value{GDBN}
5254 If @value{GDBN} is limping on your machine, this is the preferred way to get it
5255 fully functional. Be warned that in some ancient Unix systems, like
5256 Ultrix 4.2, a program can't be running in one process while it is being
5257 debugged in another. Rather than typing the command @kbd{@w{./gdb
5258 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
5259 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
5261 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
5262 @file{.gdbinit} file that sets up some simple things to make debugging
5263 gdb easier. The @code{info} command, when executed without a subcommand
5264 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
5265 gdb. See @file{.gdbinit} for details.
5267 If you use emacs, you will probably want to do a @code{make TAGS} after
5268 you configure your distribution; this will put the machine dependent
5269 routines for your local machine where they will be accessed first by
5272 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
5273 have run @code{fixincludes} if you are compiling with gcc.
5275 @section Submitting Patches
5277 @cindex submitting patches
5278 Thanks for thinking of offering your changes back to the community of
5279 @value{GDBN} users. In general we like to get well designed enhancements.
5280 Thanks also for checking in advance about the best way to transfer the
5283 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
5284 This manual summarizes what we believe to be clean design for @value{GDBN}.
5286 If the maintainers don't have time to put the patch in when it arrives,
5287 or if there is any question about a patch, it goes into a large queue
5288 with everyone else's patches and bug reports.
5290 @cindex legal papers for code contributions
5291 The legal issue is that to incorporate substantial changes requires a
5292 copyright assignment from you and/or your employer, granting ownership
5293 of the changes to the Free Software Foundation. You can get the
5294 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
5295 and asking for it. We recommend that people write in "All programs
5296 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
5297 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
5299 contributed with only one piece of legalese pushed through the
5300 bureaucracy and filed with the FSF. We can't start merging changes until
5301 this paperwork is received by the FSF (their rules, which we follow
5302 since we maintain it for them).
5304 Technically, the easiest way to receive changes is to receive each
5305 feature as a small context diff or unidiff, suitable for @code{patch}.
5306 Each message sent to me should include the changes to C code and
5307 header files for a single feature, plus @file{ChangeLog} entries for
5308 each directory where files were modified, and diffs for any changes
5309 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
5310 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
5311 single feature, they can be split down into multiple messages.
5313 In this way, if we read and like the feature, we can add it to the
5314 sources with a single patch command, do some testing, and check it in.
5315 If you leave out the @file{ChangeLog}, we have to write one. If you leave
5316 out the doc, we have to puzzle out what needs documenting. Etc., etc.
5318 The reason to send each change in a separate message is that we will not
5319 install some of the changes. They'll be returned to you with questions
5320 or comments. If we're doing our job correctly, the message back to you
5321 will say what you have to fix in order to make the change acceptable.
5322 The reason to have separate messages for separate features is so that
5323 the acceptable changes can be installed while one or more changes are
5324 being reworked. If multiple features are sent in a single message, we
5325 tend to not put in the effort to sort out the acceptable changes from
5326 the unacceptable, so none of the features get installed until all are
5329 If this sounds painful or authoritarian, well, it is. But we get a lot
5330 of bug reports and a lot of patches, and many of them don't get
5331 installed because we don't have the time to finish the job that the bug
5332 reporter or the contributor could have done. Patches that arrive
5333 complete, working, and well designed, tend to get installed on the day
5334 they arrive. The others go into a queue and get installed as time
5335 permits, which, since the maintainers have many demands to meet, may not
5336 be for quite some time.
5338 Please send patches directly to
5339 @email{gdb-patches@@sourceware.cygnus.com, the @value{GDBN} maintainers}.
5341 @section Obsolete Conditionals
5342 @cindex obsolete code
5344 Fragments of old code in @value{GDBN} sometimes reference or set the following
5345 configuration macros. They should not be used by new code, and old uses
5346 should be removed as those parts of the debugger are otherwise touched.
5349 @item STACK_END_ADDR
5350 This macro used to define where the end of the stack appeared, for use
5351 in interpreting core file formats that don't record this address in the
5352 core file itself. This information is now configured in BFD, and @value{GDBN}
5353 gets the info portably from there. The values in @value{GDBN}'s configuration
5354 files should be moved into BFD configuration files (if needed there),
5355 and deleted from all of @value{GDBN}'s config files.
5357 Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR
5358 is so old that it has never been converted to use BFD. Now that's old!
5360 @item PYRAMID_CONTROL_FRAME_DEBUGGING
5364 @item PYRAMID_PTRACE
5367 @item REG_STACK_SEGMENT
5377 @c TeX can handle the contents at the start but makeinfo 3.12 can not