1 \input texinfo @c -*- texinfo -*-
2 @setfilename gdbint.info
4 @dircategory Software development
6 * Gdb-Internals: (gdbint). The GNU debugger's internals.
10 This file documents the internals of the GNU debugger @value{GDBN}.
11 Copyright (C) 1990, 1991, 1992, 1993, 1994, 1996, 1998, 1999, 2000, 2001,
12 2002, 2003, 2004, 2005, 2006
13 Free Software Foundation, Inc.
14 Contributed by Cygnus Solutions. Written by John Gilmore.
15 Second Edition by Stan Shebs.
17 Permission is granted to copy, distribute and/or modify this document
18 under the terms of the GNU Free Documentation License, Version 1.1 or
19 any later version published by the Free Software Foundation; with no
20 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
21 Texts. A copy of the license is included in the section entitled ``GNU
22 Free Documentation License''.
25 @setchapternewpage off
26 @settitle @value{GDBN} Internals
32 @title @value{GDBN} Internals
33 @subtitle{A guide to the internals of the GNU debugger}
35 @author Cygnus Solutions
36 @author Second Edition:
38 @author Cygnus Solutions
41 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
42 \xdef\manvers{\$Revision$} % For use in headers, footers too
44 \hfill Cygnus Solutions\par
46 \hfill \TeX{}info \texinfoversion\par
50 @vskip 0pt plus 1filll
51 Copyright @copyright{} 1990,1991,1992,1993,1994,1996,1998,1999,2000,2001,
52 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
54 Permission is granted to copy, distribute and/or modify this document
55 under the terms of the GNU Free Documentation License, Version 1.1 or
56 any later version published by the Free Software Foundation; with no
57 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
58 Texts. A copy of the license is included in the section entitled ``GNU
59 Free Documentation License''.
65 @c Perhaps this should be the title of the document (but only for info,
66 @c not for TeX). Existing GNU manuals seem inconsistent on this point.
67 @top Scope of this Document
69 This document documents the internals of the GNU debugger, @value{GDBN}. It
70 includes description of @value{GDBN}'s key algorithms and operations, as well
71 as the mechanisms that adapt @value{GDBN} to specific hosts and targets.
82 * Target Architecture Definition::
83 * Target Descriptions::
84 * Target Vector Definition::
89 * Versions and Branches::
90 * Start of New Year Procedure::
95 * GDB Observers:: @value{GDBN} Currently available observers
96 * GNU Free Documentation License:: The license for this documentation
102 @chapter Requirements
103 @cindex requirements for @value{GDBN}
105 Before diving into the internals, you should understand the formal
106 requirements and other expectations for @value{GDBN}. Although some
107 of these may seem obvious, there have been proposals for @value{GDBN}
108 that have run counter to these requirements.
110 First of all, @value{GDBN} is a debugger. It's not designed to be a
111 front panel for embedded systems. It's not a text editor. It's not a
112 shell. It's not a programming environment.
114 @value{GDBN} is an interactive tool. Although a batch mode is
115 available, @value{GDBN}'s primary role is to interact with a human
118 @value{GDBN} should be responsive to the user. A programmer hot on
119 the trail of a nasty bug, and operating under a looming deadline, is
120 going to be very impatient of everything, including the response time
121 to debugger commands.
123 @value{GDBN} should be relatively permissive, such as for expressions.
124 While the compiler should be picky (or have the option to be made
125 picky), since source code lives for a long time usually, the
126 programmer doing debugging shouldn't be spending time figuring out to
127 mollify the debugger.
129 @value{GDBN} will be called upon to deal with really large programs.
130 Executable sizes of 50 to 100 megabytes occur regularly, and we've
131 heard reports of programs approaching 1 gigabyte in size.
133 @value{GDBN} should be able to run everywhere. No other debugger is
134 available for even half as many configurations as @value{GDBN}
138 @node Overall Structure
140 @chapter Overall Structure
142 @value{GDBN} consists of three major subsystems: user interface,
143 symbol handling (the @dfn{symbol side}), and target system handling (the
146 The user interface consists of several actual interfaces, plus
149 The symbol side consists of object file readers, debugging info
150 interpreters, symbol table management, source language expression
151 parsing, type and value printing.
153 The target side consists of execution control, stack frame analysis, and
154 physical target manipulation.
156 The target side/symbol side division is not formal, and there are a
157 number of exceptions. For instance, core file support involves symbolic
158 elements (the basic core file reader is in BFD) and target elements (it
159 supplies the contents of memory and the values of registers). Instead,
160 this division is useful for understanding how the minor subsystems
163 @section The Symbol Side
165 The symbolic side of @value{GDBN} can be thought of as ``everything
166 you can do in @value{GDBN} without having a live program running''.
167 For instance, you can look at the types of variables, and evaluate
168 many kinds of expressions.
170 @section The Target Side
172 The target side of @value{GDBN} is the ``bits and bytes manipulator''.
173 Although it may make reference to symbolic info here and there, most
174 of the target side will run with only a stripped executable
175 available---or even no executable at all, in remote debugging cases.
177 Operations such as disassembly, stack frame crawls, and register
178 display, are able to work with no symbolic info at all. In some cases,
179 such as disassembly, @value{GDBN} will use symbolic info to present addresses
180 relative to symbols rather than as raw numbers, but it will work either
183 @section Configurations
187 @dfn{Host} refers to attributes of the system where @value{GDBN} runs.
188 @dfn{Target} refers to the system where the program being debugged
189 executes. In most cases they are the same machine, in which case a
190 third type of @dfn{Native} attributes come into play.
192 Defines and include files needed to build on the host are host support.
193 Examples are tty support, system defined types, host byte order, host
196 Defines and information needed to handle the target format are target
197 dependent. Examples are the stack frame format, instruction set,
198 breakpoint instruction, registers, and how to set up and tear down the stack
201 Information that is only needed when the host and target are the same,
202 is native dependent. One example is Unix child process support; if the
203 host and target are not the same, doing a fork to start the target
204 process is a bad idea. The various macros needed for finding the
205 registers in the @code{upage}, running @code{ptrace}, and such are all
206 in the native-dependent files.
208 Another example of native-dependent code is support for features that
209 are really part of the target environment, but which require
210 @code{#include} files that are only available on the host system. Core
211 file handling and @code{setjmp} handling are two common cases.
213 When you want to make @value{GDBN} work ``native'' on a particular machine, you
214 have to include all three kinds of information.
216 @section Source Tree Structure
217 @cindex @value{GDBN} source tree structure
219 The @value{GDBN} source directory has a mostly flat structure---there
220 are only a few subdirectories. A file's name usually gives a hint as
221 to what it does; for example, @file{stabsread.c} reads stabs,
222 @file{dwarf2read.c} reads @sc{DWARF 2}, etc.
224 Files that are related to some common task have names that share
225 common substrings. For example, @file{*-thread.c} files deal with
226 debugging threads on various platforms; @file{*read.c} files deal with
227 reading various kinds of symbol and object files; @file{inf*.c} files
228 deal with direct control of the @dfn{inferior program} (@value{GDBN}
229 parlance for the program being debugged).
231 There are several dozens of files in the @file{*-tdep.c} family.
232 @samp{tdep} stands for @dfn{target-dependent code}---each of these
233 files implements debug support for a specific target architecture
234 (sparc, mips, etc). Usually, only one of these will be used in a
235 specific @value{GDBN} configuration (sometimes two, closely related).
237 Similarly, there are many @file{*-nat.c} files, each one for native
238 debugging on a specific system (e.g., @file{sparc-linux-nat.c} is for
239 native debugging of Sparc machines running the Linux kernel).
241 The few subdirectories of the source tree are:
245 Code that implements @dfn{CLI}, the @value{GDBN} Command-Line
246 Interpreter. @xref{User Interface, Command Interpreter}.
249 Code for the @value{GDBN} remote server.
252 Code for Insight, the @value{GDBN} TK-based GUI front-end.
255 The @dfn{GDB/MI}, the @value{GDBN} Machine Interface interpreter.
258 Target signal translation code.
261 Code for @dfn{TUI}, the @value{GDBN} Text-mode full-screen User
262 Interface. @xref{User Interface, TUI}.
270 @value{GDBN} uses a number of debugging-specific algorithms. They are
271 often not very complicated, but get lost in the thicket of special
272 cases and real-world issues. This chapter describes the basic
273 algorithms and mentions some of the specific target definitions that
279 @cindex call stack frame
280 A frame is a construct that @value{GDBN} uses to keep track of calling
281 and called functions.
283 @cindex frame, unwind
284 @value{GDBN}'s frame model, a fresh design, was implemented with the
285 need to support @sc{dwarf}'s Call Frame Information in mind. In fact,
286 the term ``unwind'' is taken directly from that specification.
287 Developers wishing to learn more about unwinders, are encouraged to
288 read the @sc{dwarf} specification.
290 @findex frame_register_unwind
291 @findex get_frame_register
292 @value{GDBN}'s model is that you find a frame's registers by
293 ``unwinding'' them from the next younger frame. That is,
294 @samp{get_frame_register} which returns the value of a register in
295 frame #1 (the next-to-youngest frame), is implemented by calling frame
296 #0's @code{frame_register_unwind} (the youngest frame). But then the
297 obvious question is: how do you access the registers of the youngest
300 @cindex sentinel frame
301 @findex get_frame_type
302 @vindex SENTINEL_FRAME
303 To answer this question, GDB has the @dfn{sentinel} frame, the
304 ``-1st'' frame. Unwinding registers from the sentinel frame gives you
305 the current values of the youngest real frame's registers. If @var{f}
306 is a sentinel frame, then @code{get_frame_type (@var{f}) ==
309 @section Prologue Analysis
311 @cindex prologue analysis
312 @cindex call frame information
313 @cindex CFI (call frame information)
314 To produce a backtrace and allow the user to manipulate older frames'
315 variables and arguments, @value{GDBN} needs to find the base addresses
316 of older frames, and discover where those frames' registers have been
317 saved. Since a frame's ``callee-saves'' registers get saved by
318 younger frames if and when they're reused, a frame's registers may be
319 scattered unpredictably across younger frames. This means that
320 changing the value of a register-allocated variable in an older frame
321 may actually entail writing to a save slot in some younger frame.
323 Modern versions of GCC emit Dwarf call frame information (``CFI''),
324 which describes how to find frame base addresses and saved registers.
325 But CFI is not always available, so as a fallback @value{GDBN} uses a
326 technique called @dfn{prologue analysis} to find frame sizes and saved
327 registers. A prologue analyzer disassembles the function's machine
328 code starting from its entry point, and looks for instructions that
329 allocate frame space, save the stack pointer in a frame pointer
330 register, save registers, and so on. Obviously, this can't be done
331 accurately in general, but it's tractable to do well enough to be very
332 helpful. Prologue analysis predates the GNU toolchain's support for
333 CFI; at one time, prologue analysis was the only mechanism
334 @value{GDBN} used for stack unwinding at all, when the function
335 calling conventions didn't specify a fixed frame layout.
337 In the olden days, function prologues were generated by hand-written,
338 target-specific code in GCC, and treated as opaque and untouchable by
339 optimizers. Looking at this code, it was usually straightforward to
340 write a prologue analyzer for @value{GDBN} that would accurately
341 understand all the prologues GCC would generate. However, over time
342 GCC became more aggressive about instruction scheduling, and began to
343 understand more about the semantics of the prologue instructions
344 themselves; in response, @value{GDBN}'s analyzers became more complex
345 and fragile. Keeping the prologue analyzers working as GCC (and the
346 instruction sets themselves) evolved became a substantial task.
348 @cindex @file{prologue-value.c}
349 @cindex abstract interpretation of function prologues
350 @cindex pseudo-evaluation of function prologues
351 To try to address this problem, the code in @file{prologue-value.h}
352 and @file{prologue-value.c} provides a general framework for writing
353 prologue analyzers that are simpler and more robust than ad-hoc
354 analyzers. When we analyze a prologue using the prologue-value
355 framework, we're really doing ``abstract interpretation'' or
356 ``pseudo-evaluation'': running the function's code in simulation, but
357 using conservative approximations of the values registers and memory
358 would hold when the code actually runs. For example, if our function
359 starts with the instruction:
362 addi r1, 42 # add 42 to r1
365 we don't know exactly what value will be in @code{r1} after executing
366 this instruction, but we do know it'll be 42 greater than its original
369 If we then see an instruction like:
372 addi r1, 22 # add 22 to r1
375 we still don't know what @code{r1's} value is, but again, we can say
376 it is now 64 greater than its original value.
378 If the next instruction were:
381 mov r2, r1 # set r2 to r1's value
384 then we can say that @code{r2's} value is now the original value of
387 It's common for prologues to save registers on the stack, so we'll
388 need to track the values of stack frame slots, as well as the
389 registers. So after an instruction like this:
395 then we'd know that the stack slot four bytes above the frame pointer
396 holds the original value of @code{r1} plus 64.
400 Of course, this can only go so far before it gets unreasonable. If we
401 wanted to be able to say anything about the value of @code{r1} after
405 xor r1, r3 # exclusive-or r1 and r3, place result in r1
408 then things would get pretty complex. But remember, we're just doing
409 a conservative approximation; if exclusive-or instructions aren't
410 relevant to prologues, we can just say @code{r1}'s value is now
411 ``unknown''. We can ignore things that are too complex, if that loss of
412 information is acceptable for our application.
414 So when we say ``conservative approximation'' here, what we mean is an
415 approximation that is either accurate, or marked ``unknown'', but
418 Using this framework, a prologue analyzer is simply an interpreter for
419 machine code, but one that uses conservative approximations for the
420 contents of registers and memory instead of actual values. Starting
421 from the function's entry point, you simulate instructions up to the
422 current PC, or an instruction that you don't know how to simulate.
423 Now you can examine the state of the registers and stack slots you've
429 To see how large your stack frame is, just check the value of the
430 stack pointer register; if it's the original value of the SP
431 minus a constant, then that constant is the stack frame's size.
432 If the SP's value has been marked as ``unknown'', then that means
433 the prologue has done something too complex for us to track, and
434 we don't know the frame size.
437 To see where we've saved the previous frame's registers, we just
438 search the values we've tracked --- stack slots, usually, but
439 registers, too, if you want --- for something equal to the register's
440 original value. If the calling conventions suggest a standard place
441 to save a given register, then we can check there first, but really,
442 anything that will get us back the original value will probably work.
445 This does take some work. But prologue analyzers aren't
446 quick-and-simple pattern patching to recognize a few fixed prologue
447 forms any more; they're big, hairy functions. Along with inferior
448 function calls, prologue analysis accounts for a substantial portion
449 of the time needed to stabilize a @value{GDBN} port. So it's
450 worthwhile to look for an approach that will be easier to understand
451 and maintain. In the approach described above:
456 It's easier to see that the analyzer is correct: you just see
457 whether the analyzer properly (albeit conservatively) simulates
458 the effect of each instruction.
461 It's easier to extend the analyzer: you can add support for new
462 instructions, and know that you haven't broken anything that
463 wasn't already broken before.
466 It's orthogonal: to gather new information, you don't need to
467 complicate the code for each instruction. As long as your domain
468 of conservative values is already detailed enough to tell you
469 what you need, then all the existing instruction simulations are
470 already gathering the right data for you.
474 The file @file{prologue-value.h} contains detailed comments explaining
475 the framework and how to use it.
478 @section Breakpoint Handling
481 In general, a breakpoint is a user-designated location in the program
482 where the user wants to regain control if program execution ever reaches
485 There are two main ways to implement breakpoints; either as ``hardware''
486 breakpoints or as ``software'' breakpoints.
488 @cindex hardware breakpoints
489 @cindex program counter
490 Hardware breakpoints are sometimes available as a builtin debugging
491 features with some chips. Typically these work by having dedicated
492 register into which the breakpoint address may be stored. If the PC
493 (shorthand for @dfn{program counter})
494 ever matches a value in a breakpoint registers, the CPU raises an
495 exception and reports it to @value{GDBN}.
497 Another possibility is when an emulator is in use; many emulators
498 include circuitry that watches the address lines coming out from the
499 processor, and force it to stop if the address matches a breakpoint's
502 A third possibility is that the target already has the ability to do
503 breakpoints somehow; for instance, a ROM monitor may do its own
504 software breakpoints. So although these are not literally ``hardware
505 breakpoints'', from @value{GDBN}'s point of view they work the same;
506 @value{GDBN} need not do anything more than set the breakpoint and wait
507 for something to happen.
509 Since they depend on hardware resources, hardware breakpoints may be
510 limited in number; when the user asks for more, @value{GDBN} will
511 start trying to set software breakpoints. (On some architectures,
512 notably the 32-bit x86 platforms, @value{GDBN} cannot always know
513 whether there's enough hardware resources to insert all the hardware
514 breakpoints and watchpoints. On those platforms, @value{GDBN} prints
515 an error message only when the program being debugged is continued.)
517 @cindex software breakpoints
518 Software breakpoints require @value{GDBN} to do somewhat more work.
519 The basic theory is that @value{GDBN} will replace a program
520 instruction with a trap, illegal divide, or some other instruction
521 that will cause an exception, and then when it's encountered,
522 @value{GDBN} will take the exception and stop the program. When the
523 user says to continue, @value{GDBN} will restore the original
524 instruction, single-step, re-insert the trap, and continue on.
526 Since it literally overwrites the program being tested, the program area
527 must be writable, so this technique won't work on programs in ROM. It
528 can also distort the behavior of programs that examine themselves,
529 although such a situation would be highly unusual.
531 Also, the software breakpoint instruction should be the smallest size of
532 instruction, so it doesn't overwrite an instruction that might be a jump
533 target, and cause disaster when the program jumps into the middle of the
534 breakpoint instruction. (Strictly speaking, the breakpoint must be no
535 larger than the smallest interval between instructions that may be jump
536 targets; perhaps there is an architecture where only even-numbered
537 instructions may jumped to.) Note that it's possible for an instruction
538 set not to have any instructions usable for a software breakpoint,
539 although in practice only the ARC has failed to define such an
543 The basic definition of the software breakpoint is the macro
546 Basic breakpoint object handling is in @file{breakpoint.c}. However,
547 much of the interesting breakpoint action is in @file{infrun.c}.
550 @cindex insert or remove software breakpoint
551 @findex target_remove_breakpoint
552 @findex target_insert_breakpoint
553 @item target_remove_breakpoint (@var{bp_tgt})
554 @itemx target_insert_breakpoint (@var{bp_tgt})
555 Insert or remove a software breakpoint at address
556 @code{@var{bp_tgt}->placed_address}. Returns zero for success,
557 non-zero for failure. On input, @var{bp_tgt} contains the address of the
558 breakpoint, and is otherwise initialized to zero. The fields of the
559 @code{struct bp_target_info} pointed to by @var{bp_tgt} are updated
560 to contain other information about the breakpoint on output. The field
561 @code{placed_address} may be updated if the breakpoint was placed at a
562 related address; the field @code{shadow_contents} contains the real
563 contents of the bytes where the breakpoint has been inserted,
564 if reading memory would return the breakpoint instead of the
565 underlying memory; the field @code{shadow_len} is the length of
566 memory cached in @code{shadow_contents}, if any; and the field
567 @code{placed_size} is optionally set and used by the target, if
568 it could differ from @code{shadow_len}.
570 For example, the remote target @samp{Z0} packet does not require
571 shadowing memory, so @code{shadow_len} is left at zero. However,
572 the length reported by @code{gdbarch_breakpoint_from_pc} is cached in
573 @code{placed_size}, so that a matching @samp{z0} packet can be
574 used to remove the breakpoint.
576 @cindex insert or remove hardware breakpoint
577 @findex target_remove_hw_breakpoint
578 @findex target_insert_hw_breakpoint
579 @item target_remove_hw_breakpoint (@var{bp_tgt})
580 @itemx target_insert_hw_breakpoint (@var{bp_tgt})
581 Insert or remove a hardware-assisted breakpoint at address
582 @code{@var{bp_tgt}->placed_address}. Returns zero for success,
583 non-zero for failure. See @code{target_insert_breakpoint} for
584 a description of the @code{struct bp_target_info} pointed to by
585 @var{bp_tgt}; the @code{shadow_contents} and
586 @code{shadow_len} members are not used for hardware breakpoints,
587 but @code{placed_size} may be.
590 @section Single Stepping
592 @section Signal Handling
594 @section Thread Handling
596 @section Inferior Function Calls
598 @section Longjmp Support
600 @cindex @code{longjmp} debugging
601 @value{GDBN} has support for figuring out that the target is doing a
602 @code{longjmp} and for stopping at the target of the jump, if we are
603 stepping. This is done with a few specialized internal breakpoints,
604 which are visible in the output of the @samp{maint info breakpoint}
607 @findex gdbarch_get_longjmp_target
608 To make this work, you need to define a function called
609 @code{gdbarch_get_longjmp_target}, which will examine the @code{jmp_buf}
610 structure and extract the longjmp target address. Since @code{jmp_buf}
611 is target specific, you will need to define it in the appropriate
612 @file{tm-@var{target}.h} file. Look in @file{tm-sun4os4.h} and
613 @file{sparc-tdep.c} for examples of how to do this.
618 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
619 breakpoints}) which break when data is accessed rather than when some
620 instruction is executed. When you have data which changes without
621 your knowing what code does that, watchpoints are the silver bullet to
622 hunt down and kill such bugs.
624 @cindex hardware watchpoints
625 @cindex software watchpoints
626 Watchpoints can be either hardware-assisted or not; the latter type is
627 known as ``software watchpoints.'' @value{GDBN} always uses
628 hardware-assisted watchpoints if they are available, and falls back on
629 software watchpoints otherwise. Typical situations where @value{GDBN}
630 will use software watchpoints are:
634 The watched memory region is too large for the underlying hardware
635 watchpoint support. For example, each x86 debug register can watch up
636 to 4 bytes of memory, so trying to watch data structures whose size is
637 more than 16 bytes will cause @value{GDBN} to use software
641 The value of the expression to be watched depends on data held in
642 registers (as opposed to memory).
645 Too many different watchpoints requested. (On some architectures,
646 this situation is impossible to detect until the debugged program is
647 resumed.) Note that x86 debug registers are used both for hardware
648 breakpoints and for watchpoints, so setting too many hardware
649 breakpoints might cause watchpoint insertion to fail.
652 No hardware-assisted watchpoints provided by the target
656 Software watchpoints are very slow, since @value{GDBN} needs to
657 single-step the program being debugged and test the value of the
658 watched expression(s) after each instruction. The rest of this
659 section is mostly irrelevant for software watchpoints.
661 When the inferior stops, @value{GDBN} tries to establish, among other
662 possible reasons, whether it stopped due to a watchpoint being hit.
663 For a data-write watchpoint, it does so by evaluating, for each
664 watchpoint, the expression whose value is being watched, and testing
665 whether the watched value has changed. For data-read and data-access
666 watchpoints, @value{GDBN} needs the target to supply a primitive that
667 returns the address of the data that was accessed or read (see the
668 description of @code{target_stopped_data_address} below): if this
669 primitive returns a valid address, @value{GDBN} infers that a
670 watchpoint triggered if it watches an expression whose evaluation uses
673 @value{GDBN} uses several macros and primitives to support hardware
677 @findex TARGET_HAS_HARDWARE_WATCHPOINTS
678 @item TARGET_HAS_HARDWARE_WATCHPOINTS
679 If defined, the target supports hardware watchpoints.
681 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
682 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
683 Return the number of hardware watchpoints of type @var{type} that are
684 possible to be set. The value is positive if @var{count} watchpoints
685 of this type can be set, zero if setting watchpoints of this type is
686 not supported, and negative if @var{count} is more than the maximum
687 number of watchpoints of type @var{type} that can be set. @var{other}
688 is non-zero if other types of watchpoints are currently enabled (there
689 are architectures which cannot set watchpoints of different types at
692 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
693 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
694 Return non-zero if hardware watchpoints can be used to watch a region
695 whose address is @var{addr} and whose length in bytes is @var{len}.
697 @cindex insert or remove hardware watchpoint
698 @findex target_insert_watchpoint
699 @findex target_remove_watchpoint
700 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
701 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
702 Insert or remove a hardware watchpoint starting at @var{addr}, for
703 @var{len} bytes. @var{type} is the watchpoint type, one of the
704 possible values of the enumerated data type @code{target_hw_bp_type},
705 defined by @file{breakpoint.h} as follows:
708 enum target_hw_bp_type
710 hw_write = 0, /* Common (write) HW watchpoint */
711 hw_read = 1, /* Read HW watchpoint */
712 hw_access = 2, /* Access (read or write) HW watchpoint */
713 hw_execute = 3 /* Execute HW breakpoint */
718 These two macros should return 0 for success, non-zero for failure.
720 @findex target_stopped_data_address
721 @item target_stopped_data_address (@var{addr_p})
722 If the inferior has some watchpoint that triggered, place the address
723 associated with the watchpoint at the location pointed to by
724 @var{addr_p} and return non-zero. Otherwise, return zero. Note that
725 this primitive is used by @value{GDBN} only on targets that support
726 data-read or data-access type watchpoints, so targets that have
727 support only for data-write watchpoints need not implement these
730 @findex HAVE_STEPPABLE_WATCHPOINT
731 @item HAVE_STEPPABLE_WATCHPOINT
732 If defined to a non-zero value, it is not necessary to disable a
733 watchpoint to step over it.
735 @findex gdbarch_have_nonsteppable_watchpoint
736 @item int gdbarch_have_nonsteppable_watchpoint (@var{gdbarch})
737 If it returns a non-zero value, @value{GDBN} should disable a
738 watchpoint to step the inferior over it.
740 @findex HAVE_CONTINUABLE_WATCHPOINT
741 @item HAVE_CONTINUABLE_WATCHPOINT
742 If defined to a non-zero value, it is possible to continue the
743 inferior after a watchpoint has been hit.
745 @findex CANNOT_STEP_HW_WATCHPOINTS
746 @item CANNOT_STEP_HW_WATCHPOINTS
747 If this is defined to a non-zero value, @value{GDBN} will remove all
748 watchpoints before stepping the inferior.
750 @findex STOPPED_BY_WATCHPOINT
751 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
752 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
753 the type @code{struct target_waitstatus}, defined by @file{target.h}.
754 Normally, this macro is defined to invoke the function pointed to by
755 the @code{to_stopped_by_watchpoint} member of the structure (of the
756 type @code{target_ops}, defined on @file{target.h}) that describes the
757 target-specific operations; @code{to_stopped_by_watchpoint} ignores
758 the @var{wait_status} argument.
760 @value{GDBN} does not require the non-zero value returned by
761 @code{STOPPED_BY_WATCHPOINT} to be 100% correct, so if a target cannot
762 determine for sure whether the inferior stopped due to a watchpoint,
763 it could return non-zero ``just in case''.
766 @subsection x86 Watchpoints
767 @cindex x86 debug registers
768 @cindex watchpoints, on x86
770 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
771 registers designed to facilitate debugging. @value{GDBN} provides a
772 generic library of functions that x86-based ports can use to implement
773 support for watchpoints and hardware-assisted breakpoints. This
774 subsection documents the x86 watchpoint facilities in @value{GDBN}.
776 To use the generic x86 watchpoint support, a port should do the
780 @findex I386_USE_GENERIC_WATCHPOINTS
782 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
783 target-dependent headers.
786 Include the @file{config/i386/nm-i386.h} header file @emph{after}
787 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
790 Add @file{i386-nat.o} to the value of the Make variable
791 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}) or
792 @code{TDEPFILES} (@pxref{Target Architecture Definition, TDEPFILES}).
795 Provide implementations for the @code{I386_DR_LOW_*} macros described
796 below. Typically, each macro should call a target-specific function
797 which does the real work.
800 The x86 watchpoint support works by maintaining mirror images of the
801 debug registers. Values are copied between the mirror images and the
802 real debug registers via a set of macros which each target needs to
806 @findex I386_DR_LOW_SET_CONTROL
807 @item I386_DR_LOW_SET_CONTROL (@var{val})
808 Set the Debug Control (DR7) register to the value @var{val}.
810 @findex I386_DR_LOW_SET_ADDR
811 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
812 Put the address @var{addr} into the debug register number @var{idx}.
814 @findex I386_DR_LOW_RESET_ADDR
815 @item I386_DR_LOW_RESET_ADDR (@var{idx})
816 Reset (i.e.@: zero out) the address stored in the debug register
819 @findex I386_DR_LOW_GET_STATUS
820 @item I386_DR_LOW_GET_STATUS
821 Return the value of the Debug Status (DR6) register. This value is
822 used immediately after it is returned by
823 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
827 For each one of the 4 debug registers (whose indices are from 0 to 3)
828 that store addresses, a reference count is maintained by @value{GDBN},
829 to allow sharing of debug registers by several watchpoints. This
830 allows users to define several watchpoints that watch the same
831 expression, but with different conditions and/or commands, without
832 wasting debug registers which are in short supply. @value{GDBN}
833 maintains the reference counts internally, targets don't have to do
834 anything to use this feature.
836 The x86 debug registers can each watch a region that is 1, 2, or 4
837 bytes long. The ia32 architecture requires that each watched region
838 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
839 region on 4-byte boundary. However, the x86 watchpoint support in
840 @value{GDBN} can watch unaligned regions and regions larger than 4
841 bytes (up to 16 bytes) by allocating several debug registers to watch
842 a single region. This allocation of several registers per a watched
843 region is also done automatically without target code intervention.
845 The generic x86 watchpoint support provides the following API for the
846 @value{GDBN}'s application code:
849 @findex i386_region_ok_for_watchpoint
850 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
851 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
852 this function. It counts the number of debug registers required to
853 watch a given region, and returns a non-zero value if that number is
854 less than 4, the number of debug registers available to x86
857 @findex i386_stopped_data_address
858 @item i386_stopped_data_address (@var{addr_p})
860 @code{target_stopped_data_address} is set to call this function.
862 function examines the breakpoint condition bits in the DR6 Debug
863 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
864 macro, and returns the address associated with the first bit that is
867 @findex i386_stopped_by_watchpoint
868 @item i386_stopped_by_watchpoint (void)
869 The macro @code{STOPPED_BY_WATCHPOINT}
870 is set to call this function. The
871 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
872 function examines the breakpoint condition bits in the DR6 Debug
873 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
874 macro, and returns true if any bit is set. Otherwise, false is
877 @findex i386_insert_watchpoint
878 @findex i386_remove_watchpoint
879 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
880 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
881 Insert or remove a watchpoint. The macros
882 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
883 are set to call these functions. @code{i386_insert_watchpoint} first
884 looks for a debug register which is already set to watch the same
885 region for the same access types; if found, it just increments the
886 reference count of that debug register, thus implementing debug
887 register sharing between watchpoints. If no such register is found,
888 the function looks for a vacant debug register, sets its mirrored
889 value to @var{addr}, sets the mirrored value of DR7 Debug Control
890 register as appropriate for the @var{len} and @var{type} parameters,
891 and then passes the new values of the debug register and DR7 to the
892 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
893 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
894 required to cover the given region, the above process is repeated for
897 @code{i386_remove_watchpoint} does the opposite: it resets the address
898 in the mirrored value of the debug register and its read/write and
899 length bits in the mirrored value of DR7, then passes these new
900 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
901 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
902 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
903 decrements the reference count, and only calls
904 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
905 the count goes to zero.
907 @findex i386_insert_hw_breakpoint
908 @findex i386_remove_hw_breakpoint
909 @item i386_insert_hw_breakpoint (@var{bp_tgt})
910 @itemx i386_remove_hw_breakpoint (@var{bp_tgt})
911 These functions insert and remove hardware-assisted breakpoints. The
912 macros @code{target_insert_hw_breakpoint} and
913 @code{target_remove_hw_breakpoint} are set to call these functions.
914 The argument is a @code{struct bp_target_info *}, as described in
915 the documentation for @code{target_insert_breakpoint}.
916 These functions work like @code{i386_insert_watchpoint} and
917 @code{i386_remove_watchpoint}, respectively, except that they set up
918 the debug registers to watch instruction execution, and each
919 hardware-assisted breakpoint always requires exactly one debug
922 @findex i386_stopped_by_hwbp
923 @item i386_stopped_by_hwbp (void)
924 This function returns non-zero if the inferior has some watchpoint or
925 hardware breakpoint that triggered. It works like
926 @code{i386_stopped_data_address}, except that it doesn't record the
927 address whose watchpoint triggered.
929 @findex i386_cleanup_dregs
930 @item i386_cleanup_dregs (void)
931 This function clears all the reference counts, addresses, and control
932 bits in the mirror images of the debug registers. It doesn't affect
933 the actual debug registers in the inferior process.
940 x86 processors support setting watchpoints on I/O reads or writes.
941 However, since no target supports this (as of March 2001), and since
942 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
943 watchpoints, this feature is not yet available to @value{GDBN} running
947 x86 processors can enable watchpoints locally, for the current task
948 only, or globally, for all the tasks. For each debug register,
949 there's a bit in the DR7 Debug Control register that determines
950 whether the associated address is watched locally or globally. The
951 current implementation of x86 watchpoint support in @value{GDBN}
952 always sets watchpoints to be locally enabled, since global
953 watchpoints might interfere with the underlying OS and are probably
954 unavailable in many platforms.
960 In the abstract, a checkpoint is a point in the execution history of
961 the program, which the user may wish to return to at some later time.
963 Internally, a checkpoint is a saved copy of the program state, including
964 whatever information is required in order to restore the program to that
965 state at a later time. This can be expected to include the state of
966 registers and memory, and may include external state such as the state
967 of open files and devices.
969 There are a number of ways in which checkpoints may be implemented
970 in gdb, e.g.@: as corefiles, as forked processes, and as some opaque
971 method implemented on the target side.
973 A corefile can be used to save an image of target memory and register
974 state, which can in principle be restored later --- but corefiles do
975 not typically include information about external entities such as
976 open files. Currently this method is not implemented in gdb.
978 A forked process can save the state of user memory and registers,
979 as well as some subset of external (kernel) state. This method
980 is used to implement checkpoints on Linux, and in principle might
981 be used on other systems.
983 Some targets, e.g.@: simulators, might have their own built-in
984 method for saving checkpoints, and gdb might be able to take
985 advantage of that capability without necessarily knowing any
986 details of how it is done.
989 @section Observing changes in @value{GDBN} internals
990 @cindex observer pattern interface
991 @cindex notifications about changes in internals
993 In order to function properly, several modules need to be notified when
994 some changes occur in the @value{GDBN} internals. Traditionally, these
995 modules have relied on several paradigms, the most common ones being
996 hooks and gdb-events. Unfortunately, none of these paradigms was
997 versatile enough to become the standard notification mechanism in
998 @value{GDBN}. The fact that they only supported one ``client'' was also
1001 A new paradigm, based on the Observer pattern of the @cite{Design
1002 Patterns} book, has therefore been implemented. The goal was to provide
1003 a new interface overcoming the issues with the notification mechanisms
1004 previously available. This new interface needed to be strongly typed,
1005 easy to extend, and versatile enough to be used as the standard
1006 interface when adding new notifications.
1008 See @ref{GDB Observers} for a brief description of the observers
1009 currently implemented in GDB. The rationale for the current
1010 implementation is also briefly discussed.
1012 @node User Interface
1014 @chapter User Interface
1016 @value{GDBN} has several user interfaces. Although the command-line interface
1017 is the most common and most familiar, there are others.
1019 @section Command Interpreter
1021 @cindex command interpreter
1023 The command interpreter in @value{GDBN} is fairly simple. It is designed to
1024 allow for the set of commands to be augmented dynamically, and also
1025 has a recursive subcommand capability, where the first argument to
1026 a command may itself direct a lookup on a different command list.
1028 For instance, the @samp{set} command just starts a lookup on the
1029 @code{setlist} command list, while @samp{set thread} recurses
1030 to the @code{set_thread_cmd_list}.
1034 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
1035 the main command list, and should be used for those commands. The usual
1036 place to add commands is in the @code{_initialize_@var{xyz}} routines at
1037 the ends of most source files.
1039 @findex add_setshow_cmd
1040 @findex add_setshow_cmd_full
1041 To add paired @samp{set} and @samp{show} commands, use
1042 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
1043 a slightly simpler interface which is useful when you don't need to
1044 further modify the new command structures, while the latter returns
1045 the new command structures for manipulation.
1047 @cindex deprecating commands
1048 @findex deprecate_cmd
1049 Before removing commands from the command set it is a good idea to
1050 deprecate them for some time. Use @code{deprecate_cmd} on commands or
1051 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
1052 @code{struct cmd_list_element} as it's first argument. You can use the
1053 return value from @code{add_com} or @code{add_cmd} to deprecate the
1054 command immediately after it is created.
1056 The first time a command is used the user will be warned and offered a
1057 replacement (if one exists). Note that the replacement string passed to
1058 @code{deprecate_cmd} should be the full name of the command, i.e., the
1059 entire string the user should type at the command line.
1061 @section UI-Independent Output---the @code{ui_out} Functions
1062 @c This section is based on the documentation written by Fernando
1063 @c Nasser <fnasser@redhat.com>.
1065 @cindex @code{ui_out} functions
1066 The @code{ui_out} functions present an abstraction level for the
1067 @value{GDBN} output code. They hide the specifics of different user
1068 interfaces supported by @value{GDBN}, and thus free the programmer
1069 from the need to write several versions of the same code, one each for
1070 every UI, to produce output.
1072 @subsection Overview and Terminology
1074 In general, execution of each @value{GDBN} command produces some sort
1075 of output, and can even generate an input request.
1077 Output can be generated for the following purposes:
1081 to display a @emph{result} of an operation;
1084 to convey @emph{info} or produce side-effects of a requested
1088 to provide a @emph{notification} of an asynchronous event (including
1089 progress indication of a prolonged asynchronous operation);
1092 to display @emph{error messages} (including warnings);
1095 to show @emph{debug data};
1098 to @emph{query} or prompt a user for input (a special case).
1102 This section mainly concentrates on how to build result output,
1103 although some of it also applies to other kinds of output.
1105 Generation of output that displays the results of an operation
1106 involves one or more of the following:
1110 output of the actual data
1113 formatting the output as appropriate for console output, to make it
1114 easily readable by humans
1117 machine oriented formatting--a more terse formatting to allow for easy
1118 parsing by programs which read @value{GDBN}'s output
1121 annotation, whose purpose is to help legacy GUIs to identify interesting
1125 The @code{ui_out} routines take care of the first three aspects.
1126 Annotations are provided by separate annotation routines. Note that use
1127 of annotations for an interface between a GUI and @value{GDBN} is
1130 Output can be in the form of a single item, which we call a @dfn{field};
1131 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
1132 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
1133 header and a body. In a BNF-like form:
1136 @item <table> @expansion{}
1137 @code{<header> <body>}
1138 @item <header> @expansion{}
1139 @code{@{ <column> @}}
1140 @item <column> @expansion{}
1141 @code{<width> <alignment> <title>}
1142 @item <body> @expansion{}
1147 @subsection General Conventions
1149 Most @code{ui_out} routines are of type @code{void}, the exceptions are
1150 @code{ui_out_stream_new} (which returns a pointer to the newly created
1151 object) and the @code{make_cleanup} routines.
1153 The first parameter is always the @code{ui_out} vector object, a pointer
1154 to a @code{struct ui_out}.
1156 The @var{format} parameter is like in @code{printf} family of functions.
1157 When it is present, there must also be a variable list of arguments
1158 sufficient used to satisfy the @code{%} specifiers in the supplied
1161 When a character string argument is not used in a @code{ui_out} function
1162 call, a @code{NULL} pointer has to be supplied instead.
1165 @subsection Table, Tuple and List Functions
1167 @cindex list output functions
1168 @cindex table output functions
1169 @cindex tuple output functions
1170 This section introduces @code{ui_out} routines for building lists,
1171 tuples and tables. The routines to output the actual data items
1172 (fields) are presented in the next section.
1174 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
1175 containing information about an object; a @dfn{list} is a sequence of
1176 fields where each field describes an identical object.
1178 Use the @dfn{table} functions when your output consists of a list of
1179 rows (tuples) and the console output should include a heading. Use this
1180 even when you are listing just one object but you still want the header.
1182 @cindex nesting level in @code{ui_out} functions
1183 Tables can not be nested. Tuples and lists can be nested up to a
1184 maximum of five levels.
1186 The overall structure of the table output code is something like this:
1201 Here is the description of table-, tuple- and list-related @code{ui_out}
1204 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
1205 The function @code{ui_out_table_begin} marks the beginning of the output
1206 of a table. It should always be called before any other @code{ui_out}
1207 function for a given table. @var{nbrofcols} is the number of columns in
1208 the table. @var{nr_rows} is the number of rows in the table.
1209 @var{tblid} is an optional string identifying the table. The string
1210 pointed to by @var{tblid} is copied by the implementation of
1211 @code{ui_out_table_begin}, so the application can free the string if it
1212 was @code{malloc}ed.
1214 The companion function @code{ui_out_table_end}, described below, marks
1215 the end of the table's output.
1218 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
1219 @code{ui_out_table_header} provides the header information for a single
1220 table column. You call this function several times, one each for every
1221 column of the table, after @code{ui_out_table_begin}, but before
1222 @code{ui_out_table_body}.
1224 The value of @var{width} gives the column width in characters. The
1225 value of @var{alignment} is one of @code{left}, @code{center}, and
1226 @code{right}, and it specifies how to align the header: left-justify,
1227 center, or right-justify it. @var{colhdr} points to a string that
1228 specifies the column header; the implementation copies that string, so
1229 column header strings in @code{malloc}ed storage can be freed after the
1233 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
1234 This function delimits the table header from the table body.
1237 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
1238 This function signals the end of a table's output. It should be called
1239 after the table body has been produced by the list and field output
1242 There should be exactly one call to @code{ui_out_table_end} for each
1243 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
1244 will signal an internal error.
1247 The output of the tuples that represent the table rows must follow the
1248 call to @code{ui_out_table_body} and precede the call to
1249 @code{ui_out_table_end}. You build a tuple by calling
1250 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
1251 calls to functions which actually output fields between them.
1253 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
1254 This function marks the beginning of a tuple output. @var{id} points
1255 to an optional string that identifies the tuple; it is copied by the
1256 implementation, and so strings in @code{malloc}ed storage can be freed
1260 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
1261 This function signals an end of a tuple output. There should be exactly
1262 one call to @code{ui_out_tuple_end} for each call to
1263 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
1267 @deftypefun struct cleanup *make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1268 This function first opens the tuple and then establishes a cleanup
1269 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
1270 and correct implementation of the non-portable@footnote{The function
1271 cast is not portable ISO C.} code sequence:
1273 struct cleanup *old_cleanup;
1274 ui_out_tuple_begin (uiout, "...");
1275 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
1280 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
1281 This function marks the beginning of a list output. @var{id} points to
1282 an optional string that identifies the list; it is copied by the
1283 implementation, and so strings in @code{malloc}ed storage can be freed
1287 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
1288 This function signals an end of a list output. There should be exactly
1289 one call to @code{ui_out_list_end} for each call to
1290 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1294 @deftypefun struct cleanup *make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1295 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1296 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
1297 that will close the list.
1300 @subsection Item Output Functions
1302 @cindex item output functions
1303 @cindex field output functions
1305 The functions described below produce output for the actual data
1306 items, or fields, which contain information about the object.
1308 Choose the appropriate function accordingly to your particular needs.
1310 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1311 This is the most general output function. It produces the
1312 representation of the data in the variable-length argument list
1313 according to formatting specifications in @var{format}, a
1314 @code{printf}-like format string. The optional argument @var{fldname}
1315 supplies the name of the field. The data items themselves are
1316 supplied as additional arguments after @var{format}.
1318 This generic function should be used only when it is not possible to
1319 use one of the specialized versions (see below).
1322 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1323 This function outputs a value of an @code{int} variable. It uses the
1324 @code{"%d"} output conversion specification. @var{fldname} specifies
1325 the name of the field.
1328 @deftypefun void ui_out_field_fmt_int (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{fldname}, int @var{value})
1329 This function outputs a value of an @code{int} variable. It differs from
1330 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1331 @var{fldname} specifies
1332 the name of the field.
1335 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1336 This function outputs an address.
1339 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1340 This function outputs a string using the @code{"%s"} conversion
1344 Sometimes, there's a need to compose your output piece by piece using
1345 functions that operate on a stream, such as @code{value_print} or
1346 @code{fprintf_symbol_filtered}. These functions accept an argument of
1347 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1348 used to store the data stream used for the output. When you use one
1349 of these functions, you need a way to pass their results stored in a
1350 @code{ui_file} object to the @code{ui_out} functions. To this end,
1351 you first create a @code{ui_stream} object by calling
1352 @code{ui_out_stream_new}, pass the @code{stream} member of that
1353 @code{ui_stream} object to @code{value_print} and similar functions,
1354 and finally call @code{ui_out_field_stream} to output the field you
1355 constructed. When the @code{ui_stream} object is no longer needed,
1356 you should destroy it and free its memory by calling
1357 @code{ui_out_stream_delete}.
1359 @deftypefun struct ui_stream *ui_out_stream_new (struct ui_out *@var{uiout})
1360 This function creates a new @code{ui_stream} object which uses the
1361 same output methods as the @code{ui_out} object whose pointer is
1362 passed in @var{uiout}. It returns a pointer to the newly created
1363 @code{ui_stream} object.
1366 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1367 This functions destroys a @code{ui_stream} object specified by
1371 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1372 This function consumes all the data accumulated in
1373 @code{streambuf->stream} and outputs it like
1374 @code{ui_out_field_string} does. After a call to
1375 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1376 the stream is still valid and may be used for producing more fields.
1379 @strong{Important:} If there is any chance that your code could bail
1380 out before completing output generation and reaching the point where
1381 @code{ui_out_stream_delete} is called, it is necessary to set up a
1382 cleanup, to avoid leaking memory and other resources. Here's a
1383 skeleton code to do that:
1386 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1387 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1392 If the function already has the old cleanup chain set (for other kinds
1393 of cleanups), you just have to add your cleanup to it:
1396 mybuf = ui_out_stream_new (uiout);
1397 make_cleanup (ui_out_stream_delete, mybuf);
1400 Note that with cleanups in place, you should not call
1401 @code{ui_out_stream_delete} directly, or you would attempt to free the
1404 @subsection Utility Output Functions
1406 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1407 This function skips a field in a table. Use it if you have to leave
1408 an empty field without disrupting the table alignment. The argument
1409 @var{fldname} specifies a name for the (missing) filed.
1412 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1413 This function outputs the text in @var{string} in a way that makes it
1414 easy to be read by humans. For example, the console implementation of
1415 this method filters the text through a built-in pager, to prevent it
1416 from scrolling off the visible portion of the screen.
1418 Use this function for printing relatively long chunks of text around
1419 the actual field data: the text it produces is not aligned according
1420 to the table's format. Use @code{ui_out_field_string} to output a
1421 string field, and use @code{ui_out_message}, described below, to
1422 output short messages.
1425 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1426 This function outputs @var{nspaces} spaces. It is handy to align the
1427 text produced by @code{ui_out_text} with the rest of the table or
1431 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1432 This function produces a formatted message, provided that the current
1433 verbosity level is at least as large as given by @var{verbosity}. The
1434 current verbosity level is specified by the user with the @samp{set
1435 verbositylevel} command.@footnote{As of this writing (April 2001),
1436 setting verbosity level is not yet implemented, and is always returned
1437 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1438 argument more than zero will cause the message to never be printed.}
1441 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1442 This function gives the console output filter (a paging filter) a hint
1443 of where to break lines which are too long. Ignored for all other
1444 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1445 be printed to indent the wrapped text on the next line; it must remain
1446 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1447 explicit newline is produced by one of the other functions. If
1448 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1451 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1452 This function flushes whatever output has been accumulated so far, if
1453 the UI buffers output.
1457 @subsection Examples of Use of @code{ui_out} functions
1459 @cindex using @code{ui_out} functions
1460 @cindex @code{ui_out} functions, usage examples
1461 This section gives some practical examples of using the @code{ui_out}
1462 functions to generalize the old console-oriented code in
1463 @value{GDBN}. The examples all come from functions defined on the
1464 @file{breakpoints.c} file.
1466 This example, from the @code{breakpoint_1} function, shows how to
1469 The original code was:
1472 if (!found_a_breakpoint++)
1474 annotate_breakpoints_headers ();
1477 printf_filtered ("Num ");
1479 printf_filtered ("Type ");
1481 printf_filtered ("Disp ");
1483 printf_filtered ("Enb ");
1487 printf_filtered ("Address ");
1490 printf_filtered ("What\n");
1492 annotate_breakpoints_table ();
1496 Here's the new version:
1499 nr_printable_breakpoints = @dots{};
1502 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1504 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1506 if (nr_printable_breakpoints > 0)
1507 annotate_breakpoints_headers ();
1508 if (nr_printable_breakpoints > 0)
1510 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1511 if (nr_printable_breakpoints > 0)
1513 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1514 if (nr_printable_breakpoints > 0)
1516 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1517 if (nr_printable_breakpoints > 0)
1519 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1522 if (nr_printable_breakpoints > 0)
1524 if (gdbarch_addr_bit (current_gdbarch) <= 32)
1525 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1527 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1529 if (nr_printable_breakpoints > 0)
1531 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1532 ui_out_table_body (uiout);
1533 if (nr_printable_breakpoints > 0)
1534 annotate_breakpoints_table ();
1537 This example, from the @code{print_one_breakpoint} function, shows how
1538 to produce the actual data for the table whose structure was defined
1539 in the above example. The original code was:
1544 printf_filtered ("%-3d ", b->number);
1546 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1547 || ((int) b->type != bptypes[(int) b->type].type))
1548 internal_error ("bptypes table does not describe type #%d.",
1550 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1552 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1554 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1558 This is the new version:
1562 ui_out_tuple_begin (uiout, "bkpt");
1564 ui_out_field_int (uiout, "number", b->number);
1566 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1567 || ((int) b->type != bptypes[(int) b->type].type))
1568 internal_error ("bptypes table does not describe type #%d.",
1570 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1572 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1574 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1578 This example, also from @code{print_one_breakpoint}, shows how to
1579 produce a complicated output field using the @code{print_expression}
1580 functions which requires a stream to be passed. It also shows how to
1581 automate stream destruction with cleanups. The original code was:
1585 print_expression (b->exp, gdb_stdout);
1591 struct ui_stream *stb = ui_out_stream_new (uiout);
1592 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1595 print_expression (b->exp, stb->stream);
1596 ui_out_field_stream (uiout, "what", local_stream);
1599 This example, also from @code{print_one_breakpoint}, shows how to use
1600 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1605 if (b->dll_pathname == NULL)
1606 printf_filtered ("<any library> ");
1608 printf_filtered ("library \"%s\" ", b->dll_pathname);
1615 if (b->dll_pathname == NULL)
1617 ui_out_field_string (uiout, "what", "<any library>");
1618 ui_out_spaces (uiout, 1);
1622 ui_out_text (uiout, "library \"");
1623 ui_out_field_string (uiout, "what", b->dll_pathname);
1624 ui_out_text (uiout, "\" ");
1628 The following example from @code{print_one_breakpoint} shows how to
1629 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1634 if (b->forked_inferior_pid != 0)
1635 printf_filtered ("process %d ", b->forked_inferior_pid);
1642 if (b->forked_inferior_pid != 0)
1644 ui_out_text (uiout, "process ");
1645 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1646 ui_out_spaces (uiout, 1);
1650 Here's an example of using @code{ui_out_field_string}. The original
1655 if (b->exec_pathname != NULL)
1656 printf_filtered ("program \"%s\" ", b->exec_pathname);
1663 if (b->exec_pathname != NULL)
1665 ui_out_text (uiout, "program \"");
1666 ui_out_field_string (uiout, "what", b->exec_pathname);
1667 ui_out_text (uiout, "\" ");
1671 Finally, here's an example of printing an address. The original code:
1675 printf_filtered ("%s ",
1676 hex_string_custom ((unsigned long) b->address, 8));
1683 ui_out_field_core_addr (uiout, "Address", b->address);
1687 @section Console Printing
1696 @cindex @code{libgdb}
1697 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1698 to provide an API to @value{GDBN}'s functionality.
1701 @cindex @code{libgdb}
1702 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1703 better able to support graphical and other environments.
1705 Since @code{libgdb} development is on-going, its architecture is still
1706 evolving. The following components have so far been identified:
1710 Observer - @file{gdb-events.h}.
1712 Builder - @file{ui-out.h}
1714 Event Loop - @file{event-loop.h}
1716 Library - @file{gdb.h}
1719 The model that ties these components together is described below.
1721 @section The @code{libgdb} Model
1723 A client of @code{libgdb} interacts with the library in two ways.
1727 As an observer (using @file{gdb-events}) receiving notifications from
1728 @code{libgdb} of any internal state changes (break point changes, run
1731 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1732 obtain various status values from @value{GDBN}.
1735 Since @code{libgdb} could have multiple clients (e.g., a GUI supporting
1736 the existing @value{GDBN} CLI), those clients must co-operate when
1737 controlling @code{libgdb}. In particular, a client must ensure that
1738 @code{libgdb} is idle (i.e. no other client is using @code{libgdb})
1739 before responding to a @file{gdb-event} by making a query.
1741 @section CLI support
1743 At present @value{GDBN}'s CLI is very much entangled in with the core of
1744 @code{libgdb}. Consequently, a client wishing to include the CLI in
1745 their interface needs to carefully co-ordinate its own and the CLI's
1748 It is suggested that the client set @code{libgdb} up to be bi-modal
1749 (alternate between CLI and client query modes). The notes below sketch
1754 The client registers itself as an observer of @code{libgdb}.
1756 The client create and install @code{cli-out} builder using its own
1757 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1758 @code{gdb_stdout} streams.
1760 The client creates a separate custom @code{ui-out} builder that is only
1761 used while making direct queries to @code{libgdb}.
1764 When the client receives input intended for the CLI, it simply passes it
1765 along. Since the @code{cli-out} builder is installed by default, all
1766 the CLI output in response to that command is routed (pronounced rooted)
1767 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1768 At the same time, the client is kept abreast of internal changes by
1769 virtue of being a @code{libgdb} observer.
1771 The only restriction on the client is that it must wait until
1772 @code{libgdb} becomes idle before initiating any queries (using the
1773 client's custom builder).
1775 @section @code{libgdb} components
1777 @subheading Observer - @file{gdb-events.h}
1778 @file{gdb-events} provides the client with a very raw mechanism that can
1779 be used to implement an observer. At present it only allows for one
1780 observer and that observer must, internally, handle the need to delay
1781 the processing of any event notifications until after @code{libgdb} has
1782 finished the current command.
1784 @subheading Builder - @file{ui-out.h}
1785 @file{ui-out} provides the infrastructure necessary for a client to
1786 create a builder. That builder is then passed down to @code{libgdb}
1787 when doing any queries.
1789 @subheading Event Loop - @file{event-loop.h}
1790 @c There could be an entire section on the event-loop
1791 @file{event-loop}, currently non-re-entrant, provides a simple event
1792 loop. A client would need to either plug its self into this loop or,
1793 implement a new event-loop that GDB would use.
1795 The event-loop will eventually be made re-entrant. This is so that
1796 @value{GDBN} can better handle the problem of some commands blocking
1797 instead of returning.
1799 @subheading Library - @file{gdb.h}
1800 @file{libgdb} is the most obvious component of this system. It provides
1801 the query interface. Each function is parameterized by a @code{ui-out}
1802 builder. The result of the query is constructed using that builder
1803 before the query function returns.
1805 @node Symbol Handling
1807 @chapter Symbol Handling
1809 Symbols are a key part of @value{GDBN}'s operation. Symbols include variables,
1810 functions, and types.
1812 @section Symbol Reading
1814 @cindex symbol reading
1815 @cindex reading of symbols
1816 @cindex symbol files
1817 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
1818 file is the file containing the program which @value{GDBN} is
1819 debugging. @value{GDBN} can be directed to use a different file for
1820 symbols (with the @samp{symbol-file} command), and it can also read
1821 more symbols via the @samp{add-file} and @samp{load} commands, or while
1822 reading symbols from shared libraries.
1824 @findex find_sym_fns
1825 Symbol files are initially opened by code in @file{symfile.c} using
1826 the BFD library (@pxref{Support Libraries}). BFD identifies the type
1827 of the file by examining its header. @code{find_sym_fns} then uses
1828 this identification to locate a set of symbol-reading functions.
1830 @findex add_symtab_fns
1831 @cindex @code{sym_fns} structure
1832 @cindex adding a symbol-reading module
1833 Symbol-reading modules identify themselves to @value{GDBN} by calling
1834 @code{add_symtab_fns} during their module initialization. The argument
1835 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
1836 name (or name prefix) of the symbol format, the length of the prefix,
1837 and pointers to four functions. These functions are called at various
1838 times to process symbol files whose identification matches the specified
1841 The functions supplied by each module are:
1844 @item @var{xyz}_symfile_init(struct sym_fns *sf)
1846 @cindex secondary symbol file
1847 Called from @code{symbol_file_add} when we are about to read a new
1848 symbol file. This function should clean up any internal state (possibly
1849 resulting from half-read previous files, for example) and prepare to
1850 read a new symbol file. Note that the symbol file which we are reading
1851 might be a new ``main'' symbol file, or might be a secondary symbol file
1852 whose symbols are being added to the existing symbol table.
1854 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
1855 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
1856 new symbol file being read. Its @code{private} field has been zeroed,
1857 and can be modified as desired. Typically, a struct of private
1858 information will be @code{malloc}'d, and a pointer to it will be placed
1859 in the @code{private} field.
1861 There is no result from @code{@var{xyz}_symfile_init}, but it can call
1862 @code{error} if it detects an unavoidable problem.
1864 @item @var{xyz}_new_init()
1866 Called from @code{symbol_file_add} when discarding existing symbols.
1867 This function needs only handle the symbol-reading module's internal
1868 state; the symbol table data structures visible to the rest of
1869 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
1870 arguments and no result. It may be called after
1871 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
1872 may be called alone if all symbols are simply being discarded.
1874 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
1876 Called from @code{symbol_file_add} to actually read the symbols from a
1877 symbol-file into a set of psymtabs or symtabs.
1879 @code{sf} points to the @code{struct sym_fns} originally passed to
1880 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
1881 the offset between the file's specified start address and its true
1882 address in memory. @code{mainline} is 1 if this is the main symbol
1883 table being read, and 0 if a secondary symbol file (e.g., shared library
1884 or dynamically loaded file) is being read.@refill
1887 In addition, if a symbol-reading module creates psymtabs when
1888 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
1889 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
1890 from any point in the @value{GDBN} symbol-handling code.
1893 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
1895 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
1896 the psymtab has not already been read in and had its @code{pst->symtab}
1897 pointer set. The argument is the psymtab to be fleshed-out into a
1898 symtab. Upon return, @code{pst->readin} should have been set to 1, and
1899 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
1900 zero if there were no symbols in that part of the symbol file.
1903 @section Partial Symbol Tables
1905 @value{GDBN} has three types of symbol tables:
1908 @cindex full symbol table
1911 Full symbol tables (@dfn{symtabs}). These contain the main
1912 information about symbols and addresses.
1916 Partial symbol tables (@dfn{psymtabs}). These contain enough
1917 information to know when to read the corresponding part of the full
1920 @cindex minimal symbol table
1923 Minimal symbol tables (@dfn{msymtabs}). These contain information
1924 gleaned from non-debugging symbols.
1927 @cindex partial symbol table
1928 This section describes partial symbol tables.
1930 A psymtab is constructed by doing a very quick pass over an executable
1931 file's debugging information. Small amounts of information are
1932 extracted---enough to identify which parts of the symbol table will
1933 need to be re-read and fully digested later, when the user needs the
1934 information. The speed of this pass causes @value{GDBN} to start up very
1935 quickly. Later, as the detailed rereading occurs, it occurs in small
1936 pieces, at various times, and the delay therefrom is mostly invisible to
1938 @c (@xref{Symbol Reading}.)
1940 The symbols that show up in a file's psymtab should be, roughly, those
1941 visible to the debugger's user when the program is not running code from
1942 that file. These include external symbols and types, static symbols and
1943 types, and @code{enum} values declared at file scope.
1945 The psymtab also contains the range of instruction addresses that the
1946 full symbol table would represent.
1948 @cindex finding a symbol
1949 @cindex symbol lookup
1950 The idea is that there are only two ways for the user (or much of the
1951 code in the debugger) to reference a symbol:
1954 @findex find_pc_function
1955 @findex find_pc_line
1957 By its address (e.g., execution stops at some address which is inside a
1958 function in this file). The address will be noticed to be in the
1959 range of this psymtab, and the full symtab will be read in.
1960 @code{find_pc_function}, @code{find_pc_line}, and other
1961 @code{find_pc_@dots{}} functions handle this.
1963 @cindex lookup_symbol
1966 (e.g., the user asks to print a variable, or set a breakpoint on a
1967 function). Global names and file-scope names will be found in the
1968 psymtab, which will cause the symtab to be pulled in. Local names will
1969 have to be qualified by a global name, or a file-scope name, in which
1970 case we will have already read in the symtab as we evaluated the
1971 qualifier. Or, a local symbol can be referenced when we are ``in'' a
1972 local scope, in which case the first case applies. @code{lookup_symbol}
1973 does most of the work here.
1976 The only reason that psymtabs exist is to cause a symtab to be read in
1977 at the right moment. Any symbol that can be elided from a psymtab,
1978 while still causing that to happen, should not appear in it. Since
1979 psymtabs don't have the idea of scope, you can't put local symbols in
1980 them anyway. Psymtabs don't have the idea of the type of a symbol,
1981 either, so types need not appear, unless they will be referenced by
1984 It is a bug for @value{GDBN} to behave one way when only a psymtab has
1985 been read, and another way if the corresponding symtab has been read
1986 in. Such bugs are typically caused by a psymtab that does not contain
1987 all the visible symbols, or which has the wrong instruction address
1990 The psymtab for a particular section of a symbol file (objfile) could be
1991 thrown away after the symtab has been read in. The symtab should always
1992 be searched before the psymtab, so the psymtab will never be used (in a
1993 bug-free environment). Currently, psymtabs are allocated on an obstack,
1994 and all the psymbols themselves are allocated in a pair of large arrays
1995 on an obstack, so there is little to be gained by trying to free them
1996 unless you want to do a lot more work.
2000 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
2002 @cindex fundamental types
2003 These are the fundamental types that @value{GDBN} uses internally. Fundamental
2004 types from the various debugging formats (stabs, ELF, etc) are mapped
2005 into one of these. They are basically a union of all fundamental types
2006 that @value{GDBN} knows about for all the languages that @value{GDBN}
2009 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
2012 Each time @value{GDBN} builds an internal type, it marks it with one
2013 of these types. The type may be a fundamental type, such as
2014 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
2015 which is a pointer to another type. Typically, several @code{FT_*}
2016 types map to one @code{TYPE_CODE_*} type, and are distinguished by
2017 other members of the type struct, such as whether the type is signed
2018 or unsigned, and how many bits it uses.
2020 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
2022 These are instances of type structs that roughly correspond to
2023 fundamental types and are created as global types for @value{GDBN} to
2024 use for various ugly historical reasons. We eventually want to
2025 eliminate these. Note for example that @code{builtin_type_int}
2026 initialized in @file{gdbtypes.c} is basically the same as a
2027 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
2028 an @code{FT_INTEGER} fundamental type. The difference is that the
2029 @code{builtin_type} is not associated with any particular objfile, and
2030 only one instance exists, while @file{c-lang.c} builds as many
2031 @code{TYPE_CODE_INT} types as needed, with each one associated with
2032 some particular objfile.
2034 @section Object File Formats
2035 @cindex object file formats
2039 @cindex @code{a.out} format
2040 The @code{a.out} format is the original file format for Unix. It
2041 consists of three sections: @code{text}, @code{data}, and @code{bss},
2042 which are for program code, initialized data, and uninitialized data,
2045 The @code{a.out} format is so simple that it doesn't have any reserved
2046 place for debugging information. (Hey, the original Unix hackers used
2047 @samp{adb}, which is a machine-language debugger!) The only debugging
2048 format for @code{a.out} is stabs, which is encoded as a set of normal
2049 symbols with distinctive attributes.
2051 The basic @code{a.out} reader is in @file{dbxread.c}.
2056 The COFF format was introduced with System V Release 3 (SVR3) Unix.
2057 COFF files may have multiple sections, each prefixed by a header. The
2058 number of sections is limited.
2060 The COFF specification includes support for debugging. Although this
2061 was a step forward, the debugging information was woefully limited. For
2062 instance, it was not possible to represent code that came from an
2065 The COFF reader is in @file{coffread.c}.
2069 @cindex ECOFF format
2070 ECOFF is an extended COFF originally introduced for Mips and Alpha
2073 The basic ECOFF reader is in @file{mipsread.c}.
2077 @cindex XCOFF format
2078 The IBM RS/6000 running AIX uses an object file format called XCOFF.
2079 The COFF sections, symbols, and line numbers are used, but debugging
2080 symbols are @code{dbx}-style stabs whose strings are located in the
2081 @code{.debug} section (rather than the string table). For more
2082 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
2084 The shared library scheme has a clean interface for figuring out what
2085 shared libraries are in use, but the catch is that everything which
2086 refers to addresses (symbol tables and breakpoints at least) needs to be
2087 relocated for both shared libraries and the main executable. At least
2088 using the standard mechanism this can only be done once the program has
2089 been run (or the core file has been read).
2093 @cindex PE-COFF format
2094 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
2095 executables. PE is basically COFF with additional headers.
2097 While BFD includes special PE support, @value{GDBN} needs only the basic
2103 The ELF format came with System V Release 4 (SVR4) Unix. ELF is similar
2104 to COFF in being organized into a number of sections, but it removes
2105 many of COFF's limitations.
2107 The basic ELF reader is in @file{elfread.c}.
2112 SOM is HP's object file and debug format (not to be confused with IBM's
2113 SOM, which is a cross-language ABI).
2115 The SOM reader is in @file{somread.c}.
2117 @section Debugging File Formats
2119 This section describes characteristics of debugging information that
2120 are independent of the object file format.
2124 @cindex stabs debugging info
2125 @code{stabs} started out as special symbols within the @code{a.out}
2126 format. Since then, it has been encapsulated into other file
2127 formats, such as COFF and ELF.
2129 While @file{dbxread.c} does some of the basic stab processing,
2130 including for encapsulated versions, @file{stabsread.c} does
2135 @cindex COFF debugging info
2136 The basic COFF definition includes debugging information. The level
2137 of support is minimal and non-extensible, and is not often used.
2139 @subsection Mips debug (Third Eye)
2141 @cindex ECOFF debugging info
2142 ECOFF includes a definition of a special debug format.
2144 The file @file{mdebugread.c} implements reading for this format.
2148 @cindex DWARF 2 debugging info
2149 DWARF 2 is an improved but incompatible version of DWARF 1.
2151 The DWARF 2 reader is in @file{dwarf2read.c}.
2155 @cindex SOM debugging info
2156 Like COFF, the SOM definition includes debugging information.
2158 @section Adding a New Symbol Reader to @value{GDBN}
2160 @cindex adding debugging info reader
2161 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
2162 there is probably little to be done.
2164 If you need to add a new object file format, you must first add it to
2165 BFD. This is beyond the scope of this document.
2167 You must then arrange for the BFD code to provide access to the
2168 debugging symbols. Generally @value{GDBN} will have to call swapping routines
2169 from BFD and a few other BFD internal routines to locate the debugging
2170 information. As much as possible, @value{GDBN} should not depend on the BFD
2171 internal data structures.
2173 For some targets (e.g., COFF), there is a special transfer vector used
2174 to call swapping routines, since the external data structures on various
2175 platforms have different sizes and layouts. Specialized routines that
2176 will only ever be implemented by one object file format may be called
2177 directly. This interface should be described in a file
2178 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
2180 @section Memory Management for Symbol Files
2182 Most memory associated with a loaded symbol file is stored on
2183 its @code{objfile_obstack}. This includes symbols, types,
2184 namespace data, and other information produced by the symbol readers.
2186 Because this data lives on the objfile's obstack, it is automatically
2187 released when the objfile is unloaded or reloaded. Therefore one
2188 objfile must not reference symbol or type data from another objfile;
2189 they could be unloaded at different times.
2191 User convenience variables, et cetera, have associated types. Normally
2192 these types live in the associated objfile. However, when the objfile
2193 is unloaded, those types are deep copied to global memory, so that
2194 the values of the user variables and history items are not lost.
2197 @node Language Support
2199 @chapter Language Support
2201 @cindex language support
2202 @value{GDBN}'s language support is mainly driven by the symbol reader,
2203 although it is possible for the user to set the source language
2206 @value{GDBN} chooses the source language by looking at the extension
2207 of the file recorded in the debug info; @file{.c} means C, @file{.f}
2208 means Fortran, etc. It may also use a special-purpose language
2209 identifier if the debug format supports it, like with DWARF.
2211 @section Adding a Source Language to @value{GDBN}
2213 @cindex adding source language
2214 To add other languages to @value{GDBN}'s expression parser, follow the
2218 @item Create the expression parser.
2220 @cindex expression parser
2221 This should reside in a file @file{@var{lang}-exp.y}. Routines for
2222 building parsed expressions into a @code{union exp_element} list are in
2225 @cindex language parser
2226 Since we can't depend upon everyone having Bison, and YACC produces
2227 parsers that define a bunch of global names, the following lines
2228 @strong{must} be included at the top of the YACC parser, to prevent the
2229 various parsers from defining the same global names:
2232 #define yyparse @var{lang}_parse
2233 #define yylex @var{lang}_lex
2234 #define yyerror @var{lang}_error
2235 #define yylval @var{lang}_lval
2236 #define yychar @var{lang}_char
2237 #define yydebug @var{lang}_debug
2238 #define yypact @var{lang}_pact
2239 #define yyr1 @var{lang}_r1
2240 #define yyr2 @var{lang}_r2
2241 #define yydef @var{lang}_def
2242 #define yychk @var{lang}_chk
2243 #define yypgo @var{lang}_pgo
2244 #define yyact @var{lang}_act
2245 #define yyexca @var{lang}_exca
2246 #define yyerrflag @var{lang}_errflag
2247 #define yynerrs @var{lang}_nerrs
2250 At the bottom of your parser, define a @code{struct language_defn} and
2251 initialize it with the right values for your language. Define an
2252 @code{initialize_@var{lang}} routine and have it call
2253 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
2254 that your language exists. You'll need some other supporting variables
2255 and functions, which will be used via pointers from your
2256 @code{@var{lang}_language_defn}. See the declaration of @code{struct
2257 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
2258 for more information.
2260 @item Add any evaluation routines, if necessary
2262 @cindex expression evaluation routines
2263 @findex evaluate_subexp
2264 @findex prefixify_subexp
2265 @findex length_of_subexp
2266 If you need new opcodes (that represent the operations of the language),
2267 add them to the enumerated type in @file{expression.h}. Add support
2268 code for these operations in the @code{evaluate_subexp} function
2269 defined in the file @file{eval.c}. Add cases
2270 for new opcodes in two functions from @file{parse.c}:
2271 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
2272 the number of @code{exp_element}s that a given operation takes up.
2274 @item Update some existing code
2276 Add an enumerated identifier for your language to the enumerated type
2277 @code{enum language} in @file{defs.h}.
2279 Update the routines in @file{language.c} so your language is included.
2280 These routines include type predicates and such, which (in some cases)
2281 are language dependent. If your language does not appear in the switch
2282 statement, an error is reported.
2284 @vindex current_language
2285 Also included in @file{language.c} is the code that updates the variable
2286 @code{current_language}, and the routines that translate the
2287 @code{language_@var{lang}} enumerated identifier into a printable
2290 @findex _initialize_language
2291 Update the function @code{_initialize_language} to include your
2292 language. This function picks the default language upon startup, so is
2293 dependent upon which languages that @value{GDBN} is built for.
2295 @findex allocate_symtab
2296 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2297 code so that the language of each symtab (source file) is set properly.
2298 This is used to determine the language to use at each stack frame level.
2299 Currently, the language is set based upon the extension of the source
2300 file. If the language can be better inferred from the symbol
2301 information, please set the language of the symtab in the symbol-reading
2304 @findex print_subexp
2305 @findex op_print_tab
2306 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2307 expression opcodes you have added to @file{expression.h}. Also, add the
2308 printed representations of your operators to @code{op_print_tab}.
2310 @item Add a place of call
2313 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2314 @code{parse_exp_1} (defined in @file{parse.c}).
2316 @item Use macros to trim code
2318 @cindex trimming language-dependent code
2319 The user has the option of building @value{GDBN} for some or all of the
2320 languages. If the user decides to build @value{GDBN} for the language
2321 @var{lang}, then every file dependent on @file{language.h} will have the
2322 macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to
2323 leave out large routines that the user won't need if he or she is not
2324 using your language.
2326 Note that you do not need to do this in your YACC parser, since if @value{GDBN}
2327 is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the
2328 compiled form of your parser) is not linked into @value{GDBN} at all.
2330 See the file @file{configure.in} for how @value{GDBN} is configured
2331 for different languages.
2333 @item Edit @file{Makefile.in}
2335 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2336 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2337 not get linked in, or, worse yet, it may not get @code{tar}red into the
2342 @node Host Definition
2344 @chapter Host Definition
2346 With the advent of Autoconf, it's rarely necessary to have host
2347 definition machinery anymore. The following information is provided,
2348 mainly, as an historical reference.
2350 @section Adding a New Host
2352 @cindex adding a new host
2353 @cindex host, adding
2354 @value{GDBN}'s host configuration support normally happens via Autoconf.
2355 New host-specific definitions should not be needed. Older hosts
2356 @value{GDBN} still use the host-specific definitions and files listed
2357 below, but these mostly exist for historical reasons, and will
2358 eventually disappear.
2361 @item gdb/config/@var{arch}/@var{xyz}.mh
2362 This file once contained both host and native configuration information
2363 (@pxref{Native Debugging}) for the machine @var{xyz}. The host
2364 configuration information is now handed by Autoconf.
2366 Host configuration information included a definition of
2367 @code{XM_FILE=xm-@var{xyz}.h} and possibly definitions for @code{CC},
2368 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2369 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2371 New host only configurations do not need this file.
2373 @item gdb/config/@var{arch}/xm-@var{xyz}.h
2374 This file once contained definitions and includes required when hosting
2375 gdb on machine @var{xyz}. Those definitions and includes are now
2376 handled by Autoconf.
2378 New host and native configurations do not need this file.
2380 @emph{Maintainer's note: Some hosts continue to use the @file{xm-xyz.h}
2381 file to define the macros @var{HOST_FLOAT_FORMAT},
2382 @var{HOST_DOUBLE_FORMAT} and @var{HOST_LONG_DOUBLE_FORMAT}. That code
2383 also needs to be replaced with either an Autoconf or run-time test.}
2387 @subheading Generic Host Support Files
2389 @cindex generic host support
2390 There are some ``generic'' versions of routines that can be used by
2391 various systems. These can be customized in various ways by macros
2392 defined in your @file{xm-@var{xyz}.h} file. If these routines work for
2393 the @var{xyz} host, you can just include the generic file's name (with
2394 @samp{.o}, not @samp{.c}) in @code{XDEPFILES}.
2396 Otherwise, if your machine needs custom support routines, you will need
2397 to write routines that perform the same functions as the generic file.
2398 Put them into @code{@var{xyz}-xdep.c}, and put @code{@var{xyz}-xdep.o}
2399 into @code{XDEPFILES}.
2402 @cindex remote debugging support
2403 @cindex serial line support
2405 This contains serial line support for Unix systems. This is always
2406 included, via the makefile variable @code{SER_HARDWIRE}; override this
2407 variable in the @file{.mh} file to avoid it.
2410 This contains serial line support for 32-bit programs running under DOS,
2411 using the DJGPP (a.k.a.@: GO32) execution environment.
2413 @cindex TCP remote support
2415 This contains generic TCP support using sockets.
2418 @section Host Conditionals
2420 When @value{GDBN} is configured and compiled, various macros are
2421 defined or left undefined, to control compilation based on the
2422 attributes of the host system. These macros and their meanings (or if
2423 the meaning is not documented here, then one of the source files where
2424 they are used is indicated) are:
2427 @item @value{GDBN}INIT_FILENAME
2428 The default name of @value{GDBN}'s initialization file (normally
2432 This macro is deprecated.
2434 @item SIGWINCH_HANDLER
2435 If your host defines @code{SIGWINCH}, you can define this to be the name
2436 of a function to be called if @code{SIGWINCH} is received.
2438 @item SIGWINCH_HANDLER_BODY
2439 Define this to expand into code that will define the function named by
2440 the expansion of @code{SIGWINCH_HANDLER}.
2442 @item ALIGN_STACK_ON_STARTUP
2443 @cindex stack alignment
2444 Define this if your system is of a sort that will crash in
2445 @code{tgetent} if the stack happens not to be longword-aligned when
2446 @code{main} is called. This is a rare situation, but is known to occur
2447 on several different types of systems.
2449 @item CRLF_SOURCE_FILES
2450 @cindex DOS text files
2451 Define this if host files use @code{\r\n} rather than @code{\n} as a
2452 line terminator. This will cause source file listings to omit @code{\r}
2453 characters when printing and it will allow @code{\r\n} line endings of files
2454 which are ``sourced'' by gdb. It must be possible to open files in binary
2455 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2457 @item DEFAULT_PROMPT
2459 The default value of the prompt string (normally @code{"(gdb) "}).
2462 @cindex terminal device
2463 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2466 Define this if binary files are opened the same way as text files.
2470 In some cases, use the system call @code{mmap} for reading symbol
2471 tables. For some machines this allows for sharing and quick updates.
2474 Define this if the host system has @code{termio.h}.
2481 Values for host-side constants.
2484 Substitute for isatty, if not available.
2487 This is the longest integer type available on the host. If not defined,
2488 it will default to @code{long long} or @code{long}, depending on
2489 @code{CC_HAS_LONG_LONG}.
2491 @item CC_HAS_LONG_LONG
2492 @cindex @code{long long} data type
2493 Define this if the host C compiler supports @code{long long}. This is set
2494 by the @code{configure} script.
2496 @item PRINTF_HAS_LONG_LONG
2497 Define this if the host can handle printing of long long integers via
2498 the printf format conversion specifier @code{ll}. This is set by the
2499 @code{configure} script.
2501 @item HAVE_LONG_DOUBLE
2502 Define this if the host C compiler supports @code{long double}. This is
2503 set by the @code{configure} script.
2505 @item PRINTF_HAS_LONG_DOUBLE
2506 Define this if the host can handle printing of long double float-point
2507 numbers via the printf format conversion specifier @code{Lg}. This is
2508 set by the @code{configure} script.
2510 @item SCANF_HAS_LONG_DOUBLE
2511 Define this if the host can handle the parsing of long double
2512 float-point numbers via the scanf format conversion specifier
2513 @code{Lg}. This is set by the @code{configure} script.
2515 @item LSEEK_NOT_LINEAR
2516 Define this if @code{lseek (n)} does not necessarily move to byte number
2517 @code{n} in the file. This is only used when reading source files. It
2518 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2521 This macro is used as the argument to @code{lseek} (or, most commonly,
2522 @code{bfd_seek}). FIXME, should be replaced by SEEK_SET instead,
2523 which is the POSIX equivalent.
2526 If defined, this should be one or more tokens, such as @code{volatile},
2527 that can be used in both the declaration and definition of functions to
2528 indicate that they never return. The default is already set correctly
2529 if compiling with GCC. This will almost never need to be defined.
2532 If defined, this should be one or more tokens, such as
2533 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2534 of functions to indicate that they never return. The default is already
2535 set correctly if compiling with GCC. This will almost never need to be
2540 Define these to appropriate value for the system @code{lseek}, if not already
2544 This is the signal for stopping @value{GDBN}. Defaults to
2545 @code{SIGTSTP}. (Only redefined for the Convex.)
2548 Means that System V (prior to SVR4) include files are in use. (FIXME:
2549 This symbol is abused in @file{infrun.c}, @file{regex.c}, and
2550 @file{utils.c} for other things, at the moment.)
2553 Define this to help placate @code{lint} in some situations.
2556 Define this to override the defaults of @code{__volatile__} or
2561 @node Target Architecture Definition
2563 @chapter Target Architecture Definition
2565 @cindex target architecture definition
2566 @value{GDBN}'s target architecture defines what sort of
2567 machine-language programs @value{GDBN} can work with, and how it works
2570 The target architecture object is implemented as the C structure
2571 @code{struct gdbarch *}. The structure, and its methods, are generated
2572 using the Bourne shell script @file{gdbarch.sh}.
2575 * OS ABI Variant Handling::
2576 * Initialize New Architecture::
2577 * Registers and Memory::
2578 * Pointers and Addresses::
2580 * Raw and Virtual Registers::
2581 * Register and Memory Data::
2582 * Frame Interpretation::
2583 * Inferior Call Setup::
2584 * Compiler Characteristics::
2585 * Target Conditionals::
2586 * Adding a New Target::
2589 @node OS ABI Variant Handling
2590 @section Operating System ABI Variant Handling
2591 @cindex OS ABI variants
2593 @value{GDBN} provides a mechanism for handling variations in OS
2594 ABIs. An OS ABI variant may have influence over any number of
2595 variables in the target architecture definition. There are two major
2596 components in the OS ABI mechanism: sniffers and handlers.
2598 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2599 (the architecture may be wildcarded) in an attempt to determine the
2600 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2601 to be @dfn{generic}, while sniffers for a specific architecture are
2602 considered to be @dfn{specific}. A match from a specific sniffer
2603 overrides a match from a generic sniffer. Multiple sniffers for an
2604 architecture/flavour may exist, in order to differentiate between two
2605 different operating systems which use the same basic file format. The
2606 OS ABI framework provides a generic sniffer for ELF-format files which
2607 examines the @code{EI_OSABI} field of the ELF header, as well as note
2608 sections known to be used by several operating systems.
2610 @cindex fine-tuning @code{gdbarch} structure
2611 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2612 selected OS ABI. There may be only one handler for a given OS ABI
2613 for each BFD architecture.
2615 The following OS ABI variants are defined in @file{defs.h}:
2619 @findex GDB_OSABI_UNINITIALIZED
2620 @item GDB_OSABI_UNINITIALIZED
2621 Used for struct gdbarch_info if ABI is still uninitialized.
2623 @findex GDB_OSABI_UNKNOWN
2624 @item GDB_OSABI_UNKNOWN
2625 The ABI of the inferior is unknown. The default @code{gdbarch}
2626 settings for the architecture will be used.
2628 @findex GDB_OSABI_SVR4
2629 @item GDB_OSABI_SVR4
2630 UNIX System V Release 4.
2632 @findex GDB_OSABI_HURD
2633 @item GDB_OSABI_HURD
2634 GNU using the Hurd kernel.
2636 @findex GDB_OSABI_SOLARIS
2637 @item GDB_OSABI_SOLARIS
2640 @findex GDB_OSABI_OSF1
2641 @item GDB_OSABI_OSF1
2642 OSF/1, including Digital UNIX and Compaq Tru64 UNIX.
2644 @findex GDB_OSABI_LINUX
2645 @item GDB_OSABI_LINUX
2646 GNU using the Linux kernel.
2648 @findex GDB_OSABI_FREEBSD_AOUT
2649 @item GDB_OSABI_FREEBSD_AOUT
2650 FreeBSD using the @code{a.out} executable format.
2652 @findex GDB_OSABI_FREEBSD_ELF
2653 @item GDB_OSABI_FREEBSD_ELF
2654 FreeBSD using the ELF executable format.
2656 @findex GDB_OSABI_NETBSD_AOUT
2657 @item GDB_OSABI_NETBSD_AOUT
2658 NetBSD using the @code{a.out} executable format.
2660 @findex GDB_OSABI_NETBSD_ELF
2661 @item GDB_OSABI_NETBSD_ELF
2662 NetBSD using the ELF executable format.
2664 @findex GDB_OSABI_OPENBSD_ELF
2665 @item GDB_OSABI_OPENBSD_ELF
2666 OpenBSD using the ELF executable format.
2668 @findex GDB_OSABI_WINCE
2669 @item GDB_OSABI_WINCE
2672 @findex GDB_OSABI_GO32
2673 @item GDB_OSABI_GO32
2676 @findex GDB_OSABI_IRIX
2677 @item GDB_OSABI_IRIX
2680 @findex GDB_OSABI_INTERIX
2681 @item GDB_OSABI_INTERIX
2682 Interix (Posix layer for MS-Windows systems).
2684 @findex GDB_OSABI_HPUX_ELF
2685 @item GDB_OSABI_HPUX_ELF
2686 HP/UX using the ELF executable format.
2688 @findex GDB_OSABI_HPUX_SOM
2689 @item GDB_OSABI_HPUX_SOM
2690 HP/UX using the SOM executable format.
2692 @findex GDB_OSABI_QNXNTO
2693 @item GDB_OSABI_QNXNTO
2696 @findex GDB_OSABI_CYGWIN
2697 @item GDB_OSABI_CYGWIN
2700 @findex GDB_OSABI_AIX
2706 Here are the functions that make up the OS ABI framework:
2708 @deftypefun const char *gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2709 Return the name of the OS ABI corresponding to @var{osabi}.
2712 @deftypefun void gdbarch_register_osabi (enum bfd_architecture @var{arch}, unsigned long @var{machine}, enum gdb_osabi @var{osabi}, void (*@var{init_osabi})(struct gdbarch_info @var{info}, struct gdbarch *@var{gdbarch}))
2713 Register the OS ABI handler specified by @var{init_osabi} for the
2714 architecture, machine type and OS ABI specified by @var{arch},
2715 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2716 machine type, which implies the architecture's default machine type,
2720 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2721 Register the OS ABI file sniffer specified by @var{sniffer} for the
2722 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2723 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2724 be generic, and is allowed to examine @var{flavour}-flavoured files for
2728 @deftypefun enum gdb_osabi gdbarch_lookup_osabi (bfd *@var{abfd})
2729 Examine the file described by @var{abfd} to determine its OS ABI.
2730 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2734 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2735 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2736 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2737 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2738 architecture, a warning will be issued and the debugging session will continue
2739 with the defaults already established for @var{gdbarch}.
2742 @deftypefun void generic_elf_osabi_sniff_abi_tag_sections (bfd *@var{abfd}, asection *@var{sect}, void *@var{obj})
2743 Helper routine for ELF file sniffers. Examine the file described by
2744 @var{abfd} and look at ABI tag note sections to determine the OS ABI
2745 from the note. This function should be called via
2746 @code{bfd_map_over_sections}.
2749 @node Initialize New Architecture
2750 @section Initializing a New Architecture
2752 Each @code{gdbarch} is associated with a single @sc{bfd} architecture,
2753 via a @code{bfd_arch_@var{arch}} constant. The @code{gdbarch} is
2754 registered by a call to @code{register_gdbarch_init}, usually from
2755 the file's @code{_initialize_@var{filename}} routine, which will
2756 be automatically called during @value{GDBN} startup. The arguments
2757 are a @sc{bfd} architecture constant and an initialization function.
2759 The initialization function has this type:
2762 static struct gdbarch *
2763 @var{arch}_gdbarch_init (struct gdbarch_info @var{info},
2764 struct gdbarch_list *@var{arches})
2767 The @var{info} argument contains parameters used to select the correct
2768 architecture, and @var{arches} is a list of architectures which
2769 have already been created with the same @code{bfd_arch_@var{arch}}
2772 The initialization function should first make sure that @var{info}
2773 is acceptable, and return @code{NULL} if it is not. Then, it should
2774 search through @var{arches} for an exact match to @var{info}, and
2775 return one if found. Lastly, if no exact match was found, it should
2776 create a new architecture based on @var{info} and return it.
2778 Only information in @var{info} should be used to choose the new
2779 architecture. Historically, @var{info} could be sparse, and
2780 defaults would be collected from the first element on @var{arches}.
2781 However, @value{GDBN} now fills in @var{info} more thoroughly,
2782 so new @code{gdbarch} initialization functions should not take
2783 defaults from @var{arches}.
2785 @node Registers and Memory
2786 @section Registers and Memory
2788 @value{GDBN}'s model of the target machine is rather simple.
2789 @value{GDBN} assumes the machine includes a bank of registers and a
2790 block of memory. Each register may have a different size.
2792 @value{GDBN} does not have a magical way to match up with the
2793 compiler's idea of which registers are which; however, it is critical
2794 that they do match up accurately. The only way to make this work is
2795 to get accurate information about the order that the compiler uses,
2796 and to reflect that in the @code{gdbarch_register_name} and related functions.
2798 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
2800 @node Pointers and Addresses
2801 @section Pointers Are Not Always Addresses
2802 @cindex pointer representation
2803 @cindex address representation
2804 @cindex word-addressed machines
2805 @cindex separate data and code address spaces
2806 @cindex spaces, separate data and code address
2807 @cindex address spaces, separate data and code
2808 @cindex code pointers, word-addressed
2809 @cindex converting between pointers and addresses
2810 @cindex D10V addresses
2812 On almost all 32-bit architectures, the representation of a pointer is
2813 indistinguishable from the representation of some fixed-length number
2814 whose value is the byte address of the object pointed to. On such
2815 machines, the words ``pointer'' and ``address'' can be used interchangeably.
2816 However, architectures with smaller word sizes are often cramped for
2817 address space, so they may choose a pointer representation that breaks this
2818 identity, and allows a larger code address space.
2820 For example, the Renesas D10V is a 16-bit VLIW processor whose
2821 instructions are 32 bits long@footnote{Some D10V instructions are
2822 actually pairs of 16-bit sub-instructions. However, since you can't
2823 jump into the middle of such a pair, code addresses can only refer to
2824 full 32 bit instructions, which is what matters in this explanation.}.
2825 If the D10V used ordinary byte addresses to refer to code locations,
2826 then the processor would only be able to address 64kb of instructions.
2827 However, since instructions must be aligned on four-byte boundaries, the
2828 low two bits of any valid instruction's byte address are always
2829 zero---byte addresses waste two bits. So instead of byte addresses,
2830 the D10V uses word addresses---byte addresses shifted right two bits---to
2831 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
2834 However, this means that code pointers and data pointers have different
2835 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
2836 @code{0xC020} when used as a data address, but refers to byte address
2837 @code{0x30080} when used as a code address.
2839 (The D10V also uses separate code and data address spaces, which also
2840 affects the correspondence between pointers and addresses, but we're
2841 going to ignore that here; this example is already too long.)
2843 To cope with architectures like this---the D10V is not the only
2844 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
2845 byte numbers, and @dfn{pointers}, which are the target's representation
2846 of an address of a particular type of data. In the example above,
2847 @code{0xC020} is the pointer, which refers to one of the addresses
2848 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
2849 @value{GDBN} provides functions for turning a pointer into an address
2850 and vice versa, in the appropriate way for the current architecture.
2852 Unfortunately, since addresses and pointers are identical on almost all
2853 processors, this distinction tends to bit-rot pretty quickly. Thus,
2854 each time you port @value{GDBN} to an architecture which does
2855 distinguish between pointers and addresses, you'll probably need to
2856 clean up some architecture-independent code.
2858 Here are functions which convert between pointers and addresses:
2860 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
2861 Treat the bytes at @var{buf} as a pointer or reference of type
2862 @var{type}, and return the address it represents, in a manner
2863 appropriate for the current architecture. This yields an address
2864 @value{GDBN} can use to read target memory, disassemble, etc. Note that
2865 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2868 For example, if the current architecture is the Intel x86, this function
2869 extracts a little-endian integer of the appropriate length from
2870 @var{buf} and returns it. However, if the current architecture is the
2871 D10V, this function will return a 16-bit integer extracted from
2872 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
2874 If @var{type} is not a pointer or reference type, then this function
2875 will signal an internal error.
2878 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
2879 Store the address @var{addr} in @var{buf}, in the proper format for a
2880 pointer of type @var{type} in the current architecture. Note that
2881 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2884 For example, if the current architecture is the Intel x86, this function
2885 stores @var{addr} unmodified as a little-endian integer of the
2886 appropriate length in @var{buf}. However, if the current architecture
2887 is the D10V, this function divides @var{addr} by four if @var{type} is
2888 a pointer to a function, and then stores it in @var{buf}.
2890 If @var{type} is not a pointer or reference type, then this function
2891 will signal an internal error.
2894 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
2895 Assuming that @var{val} is a pointer, return the address it represents,
2896 as appropriate for the current architecture.
2898 This function actually works on integral values, as well as pointers.
2899 For pointers, it performs architecture-specific conversions as
2900 described above for @code{extract_typed_address}.
2903 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
2904 Create and return a value representing a pointer of type @var{type} to
2905 the address @var{addr}, as appropriate for the current architecture.
2906 This function performs architecture-specific conversions as described
2907 above for @code{store_typed_address}.
2910 Here are two functions which architectures can define to indicate the
2911 relationship between pointers and addresses. These have default
2912 definitions, appropriate for architectures on which all pointers are
2913 simple unsigned byte addresses.
2915 @deftypefun CORE_ADDR gdbarch_pointer_to_address (struct gdbarch *@var{current_gdbarch}, struct type *@var{type}, char *@var{buf})
2916 Assume that @var{buf} holds a pointer of type @var{type}, in the
2917 appropriate format for the current architecture. Return the byte
2918 address the pointer refers to.
2920 This function may safely assume that @var{type} is either a pointer or a
2921 C@t{++} reference type.
2924 @deftypefun void gdbarch_address_to_pointer (struct gdbarch *@var{current_gdbarch}, struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
2925 Store in @var{buf} a pointer of type @var{type} representing the address
2926 @var{addr}, in the appropriate format for the current architecture.
2928 This function may safely assume that @var{type} is either a pointer or a
2929 C@t{++} reference type.
2932 @node Address Classes
2933 @section Address Classes
2934 @cindex address classes
2935 @cindex DW_AT_byte_size
2936 @cindex DW_AT_address_class
2938 Sometimes information about different kinds of addresses is available
2939 via the debug information. For example, some programming environments
2940 define addresses of several different sizes. If the debug information
2941 distinguishes these kinds of address classes through either the size
2942 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
2943 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
2944 following macros should be defined in order to disambiguate these
2945 types within @value{GDBN} as well as provide the added information to
2946 a @value{GDBN} user when printing type expressions.
2948 @deftypefun int gdbarch_address_class_type_flags (struct gdbarch *@var{current_gdbarch}, int @var{byte_size}, int @var{dwarf2_addr_class})
2949 Returns the type flags needed to construct a pointer type whose size
2950 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
2951 This function is normally called from within a symbol reader. See
2952 @file{dwarf2read.c}.
2955 @deftypefun char *gdbarch_address_class_type_flags_to_name (struct gdbarch *@var{current_gdbarch}, int @var{type_flags})
2956 Given the type flags representing an address class qualifier, return
2959 @deftypefun int gdbarch_address_class_name_to_type_flags (struct gdbarch *@var{current_gdbarch}, int @var{name}, int *var{type_flags_ptr})
2960 Given an address qualifier name, set the @code{int} referenced by @var{type_flags_ptr} to the type flags
2961 for that address class qualifier.
2964 Since the need for address classes is rather rare, none of
2965 the address class functions are defined by default. Predicate
2966 functions are provided to detect when they are defined.
2968 Consider a hypothetical architecture in which addresses are normally
2969 32-bits wide, but 16-bit addresses are also supported. Furthermore,
2970 suppose that the @w{DWARF 2} information for this architecture simply
2971 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
2972 of these "short" pointers. The following functions could be defined
2973 to implement the address class functions:
2976 somearch_address_class_type_flags (int byte_size,
2977 int dwarf2_addr_class)
2980 return TYPE_FLAG_ADDRESS_CLASS_1;
2986 somearch_address_class_type_flags_to_name (int type_flags)
2988 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
2995 somearch_address_class_name_to_type_flags (char *name,
2996 int *type_flags_ptr)
2998 if (strcmp (name, "short") == 0)
3000 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
3008 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
3009 to indicate the presence of one of these "short" pointers. E.g, if
3010 the debug information indicates that @code{short_ptr_var} is one of these
3011 short pointers, @value{GDBN} might show the following behavior:
3014 (gdb) ptype short_ptr_var
3015 type = int * @@short
3019 @node Raw and Virtual Registers
3020 @section Raw and Virtual Register Representations
3021 @cindex raw register representation
3022 @cindex virtual register representation
3023 @cindex representations, raw and virtual registers
3025 @emph{Maintainer note: This section is pretty much obsolete. The
3026 functionality described here has largely been replaced by
3027 pseudo-registers and the mechanisms described in @ref{Target
3028 Architecture Definition, , Using Different Register and Memory Data
3029 Representations}. See also @uref{http://www.gnu.org/software/gdb/bugs/,
3030 Bug Tracking Database} and
3031 @uref{http://sources.redhat.com/gdb/current/ari/, ARI Index} for more
3032 up-to-date information.}
3034 Some architectures use one representation for a value when it lives in a
3035 register, but use a different representation when it lives in memory.
3036 In @value{GDBN}'s terminology, the @dfn{raw} representation is the one used in
3037 the target registers, and the @dfn{virtual} representation is the one
3038 used in memory, and within @value{GDBN} @code{struct value} objects.
3040 @emph{Maintainer note: Notice that the same mechanism is being used to
3041 both convert a register to a @code{struct value} and alternative
3044 For almost all data types on almost all architectures, the virtual and
3045 raw representations are identical, and no special handling is needed.
3046 However, they do occasionally differ. For example:
3050 The x86 architecture supports an 80-bit @code{long double} type. However, when
3051 we store those values in memory, they occupy twelve bytes: the
3052 floating-point number occupies the first ten, and the final two bytes
3053 are unused. This keeps the values aligned on four-byte boundaries,
3054 allowing more efficient access. Thus, the x86 80-bit floating-point
3055 type is the raw representation, and the twelve-byte loosely-packed
3056 arrangement is the virtual representation.
3059 Some 64-bit MIPS targets present 32-bit registers to @value{GDBN} as 64-bit
3060 registers, with garbage in their upper bits. @value{GDBN} ignores the top 32
3061 bits. Thus, the 64-bit form, with garbage in the upper 32 bits, is the
3062 raw representation, and the trimmed 32-bit representation is the
3063 virtual representation.
3066 In general, the raw representation is determined by the architecture, or
3067 @value{GDBN}'s interface to the architecture, while the virtual representation
3068 can be chosen for @value{GDBN}'s convenience. @value{GDBN}'s register file,
3069 @code{registers}, holds the register contents in raw format, and the
3070 @value{GDBN} remote protocol transmits register values in raw format.
3072 Your architecture may define the following macros to request
3073 conversions between the raw and virtual format:
3075 @deftypefn {Target Macro} int REGISTER_CONVERTIBLE (int @var{reg})
3076 Return non-zero if register number @var{reg}'s value needs different raw
3077 and virtual formats.
3079 You should not use @code{REGISTER_CONVERT_TO_VIRTUAL} for a register
3080 unless this macro returns a non-zero value for that register.
3083 @deftypefn {Target Macro} int DEPRECATED_REGISTER_RAW_SIZE (int @var{reg})
3084 The size of register number @var{reg}'s raw value. This is the number
3085 of bytes the register will occupy in @code{registers}, or in a @value{GDBN}
3086 remote protocol packet.
3089 @deftypefn {Target Macro} int DEPRECATED_REGISTER_VIRTUAL_SIZE (int @var{reg})
3090 The size of register number @var{reg}'s value, in its virtual format.
3091 This is the size a @code{struct value}'s buffer will have, holding that
3095 @deftypefn {Target Macro} struct type *DEPRECATED_REGISTER_VIRTUAL_TYPE (int @var{reg})
3096 This is the type of the virtual representation of register number
3097 @var{reg}. Note that there is no need for a macro giving a type for the
3098 register's raw form; once the register's value has been obtained, @value{GDBN}
3099 always uses the virtual form.
3102 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_VIRTUAL (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
3103 Convert the value of register number @var{reg} to @var{type}, which
3104 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
3105 at @var{from} holds the register's value in raw format; the macro should
3106 convert the value to virtual format, and place it at @var{to}.
3108 Note that @code{REGISTER_CONVERT_TO_VIRTUAL} and
3109 @code{REGISTER_CONVERT_TO_RAW} take their @var{reg} and @var{type}
3110 arguments in different orders.
3112 You should only use @code{REGISTER_CONVERT_TO_VIRTUAL} with registers
3113 for which the @code{REGISTER_CONVERTIBLE} macro returns a non-zero
3117 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_RAW (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3118 Convert the value of register number @var{reg} to @var{type}, which
3119 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
3120 at @var{from} holds the register's value in raw format; the macro should
3121 convert the value to virtual format, and place it at @var{to}.
3123 Note that REGISTER_CONVERT_TO_VIRTUAL and REGISTER_CONVERT_TO_RAW take
3124 their @var{reg} and @var{type} arguments in different orders.
3128 @node Register and Memory Data
3129 @section Using Different Register and Memory Data Representations
3130 @cindex register representation
3131 @cindex memory representation
3132 @cindex representations, register and memory
3133 @cindex register data formats, converting
3134 @cindex @code{struct value}, converting register contents to
3136 @emph{Maintainer's note: The way GDB manipulates registers is undergoing
3137 significant change. Many of the macros and functions referred to in this
3138 section are likely to be subject to further revision. See
3139 @uref{http://sources.redhat.com/gdb/current/ari/, A.R. Index} and
3140 @uref{http://www.gnu.org/software/gdb/bugs, Bug Tracking Database} for
3141 further information. cagney/2002-05-06.}
3143 Some architectures can represent a data object in a register using a
3144 form that is different to the objects more normal memory representation.
3150 The Alpha architecture can represent 32 bit integer values in
3151 floating-point registers.
3154 The x86 architecture supports 80-bit floating-point registers. The
3155 @code{long double} data type occupies 96 bits in memory but only 80 bits
3156 when stored in a register.
3160 In general, the register representation of a data type is determined by
3161 the architecture, or @value{GDBN}'s interface to the architecture, while
3162 the memory representation is determined by the Application Binary
3165 For almost all data types on almost all architectures, the two
3166 representations are identical, and no special handling is needed.
3167 However, they do occasionally differ. Your architecture may define the
3168 following macros to request conversions between the register and memory
3169 representations of a data type:
3171 @deftypefun int gdbarch_convert_register_p (struct gdbarch *@var{gdbarch}, int @var{reg})
3172 Return non-zero if the representation of a data value stored in this
3173 register may be different to the representation of that same data value
3174 when stored in memory.
3176 When non-zero, the macros @code{gdbarch_register_to_value} and
3177 @code{value_to_register} are used to perform any necessary conversion.
3180 @deftypefun void gdbarch_register_to_value (struct gdbarch *@var{gdbarch}, int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
3181 Convert the value of register number @var{reg} to a data object of type
3182 @var{type}. The buffer at @var{from} holds the register's value in raw
3183 format; the converted value should be placed in the buffer at @var{to}.
3185 Note that @code{gdbarch_register_to_value} and @code{gdbarch_value_to_register}
3186 take their @var{reg} and @var{type} arguments in different orders.
3188 You should only use @code{gdbarch_register_to_value} with registers for which
3189 the @code{gdbarch_convert_register_p} function returns a non-zero value.
3192 @deftypefun void gdbarch_value_to_register (struct gdbarch *@var{gdbarch}, struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3193 Convert a data value of type @var{type} to register number @var{reg}'
3196 Note that @code{gdbarch_register_to_value} and @code{gdbarch_value_to_register}
3197 take their @var{reg} and @var{type} arguments in different orders.
3199 You should only use @code{gdbarch_value_to_register} with registers for which
3200 the @code{gdbarch_convert_register_p} function returns a non-zero value.
3203 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_TYPE (int @var{regnum}, struct type *@var{type}, char *@var{buf})
3204 See @file{mips-tdep.c}. It does not do what you want.
3207 @node Frame Interpretation
3208 @section Frame Interpretation
3210 @node Inferior Call Setup
3211 @section Inferior Call Setup
3213 @node Compiler Characteristics
3214 @section Compiler Characteristics
3216 @node Target Conditionals
3217 @section Target Conditionals
3219 This section describes the macros and functions that you can use to define the
3224 @item CORE_ADDR gdbarch_addr_bits_remove (@var{gdbarch}, @var{addr})
3225 @findex gdbarch_addr_bits_remove
3226 If a raw machine instruction address includes any bits that are not
3227 really part of the address, then this function is used to zero those bits in
3228 @var{addr}. This is only used for addresses of instructions, and even then not
3231 For example, the two low-order bits of the PC on the Hewlett-Packard PA
3232 2.0 architecture contain the privilege level of the corresponding
3233 instruction. Since instructions must always be aligned on four-byte
3234 boundaries, the processor masks out these bits to generate the actual
3235 address of the instruction. @code{gdbarch_addr_bits_remove} would then for
3236 example look like that:
3238 arch_addr_bits_remove (CORE_ADDR addr)
3240 return (addr &= ~0x3);
3244 @item int address_class_name_to_type_flags (@var{gdbarch}, @var{name}, @var{type_flags_ptr})
3245 @findex address_class_name_to_type_flags
3246 If @var{name} is a valid address class qualifier name, set the @code{int}
3247 referenced by @var{type_flags_ptr} to the mask representing the qualifier
3248 and return 1. If @var{name} is not a valid address class qualifier name,
3251 The value for @var{type_flags_ptr} should be one of
3252 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
3253 possibly some combination of these values or'd together.
3254 @xref{Target Architecture Definition, , Address Classes}.
3256 @item int address_class_name_to_type_flags_p (@var{gdbarch})
3257 @findex address_class_name_to_type_flags_p
3258 Predicate which indicates whether @code{address_class_name_to_type_flags}
3261 @item int gdbarch_address_class_type_flags (@var{gdbarch}, @var{byte_size}, @var{dwarf2_addr_class})
3262 @findex gdbarch_address_class_type_flags
3263 Given a pointers byte size (as described by the debug information) and
3264 the possible @code{DW_AT_address_class} value, return the type flags
3265 used by @value{GDBN} to represent this address class. The value
3266 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
3267 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
3268 values or'd together.
3269 @xref{Target Architecture Definition, , Address Classes}.
3271 @item int gdbarch_address_class_type_flags_p (@var{gdbarch})
3272 @findex gdbarch_address_class_type_flags_p
3273 Predicate which indicates whether @code{gdbarch_address_class_type_flags_p} has
3276 @item const char *gdbarch_address_class_type_flags_to_name (@var{gdbarch}, @var{type_flags})
3277 @findex gdbarch_address_class_type_flags_to_name
3278 Return the name of the address class qualifier associated with the type
3279 flags given by @var{type_flags}.
3281 @item int gdbarch_address_class_type_flags_to_name_p (@var{gdbarch})
3282 @findex gdbarch_address_class_type_flags_to_name_p
3283 Predicate which indicates whether @code{gdbarch_address_class_type_flags_to_name} has been defined.
3284 @xref{Target Architecture Definition, , Address Classes}.
3286 @item void gdbarch_address_to_pointer (@var{gdbarch}, @var{type}, @var{buf}, @var{addr})
3287 @findex gdbarch_address_to_pointer
3288 Store in @var{buf} a pointer of type @var{type} representing the address
3289 @var{addr}, in the appropriate format for the current architecture.
3290 This function may safely assume that @var{type} is either a pointer or a
3291 C@t{++} reference type.
3292 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3294 @item int gdbarch_believe_pcc_promotion (@var{gdbarch})
3295 @findex gdbarch_believe_pcc_promotion
3296 Used to notify if the compiler promotes a @code{short} or @code{char}
3297 parameter to an @code{int}, but still reports the parameter as its
3298 original type, rather than the promoted type.
3300 @item BITS_BIG_ENDIAN
3301 @findex BITS_BIG_ENDIAN
3302 Define this if the numbering of bits in the targets does @strong{not} match the
3303 endianness of the target byte order. A value of 1 means that the bits
3304 are numbered in a big-endian bit order, 0 means little-endian.
3308 This is the character array initializer for the bit pattern to put into
3309 memory where a breakpoint is set. Although it's common to use a trap
3310 instruction for a breakpoint, it's not required; for instance, the bit
3311 pattern could be an invalid instruction. The breakpoint must be no
3312 longer than the shortest instruction of the architecture.
3314 @code{BREAKPOINT} has been deprecated in favor of
3315 @code{gdbarch_breakpoint_from_pc}.
3317 @item BIG_BREAKPOINT
3318 @itemx LITTLE_BREAKPOINT
3319 @findex LITTLE_BREAKPOINT
3320 @findex BIG_BREAKPOINT
3321 Similar to BREAKPOINT, but used for bi-endian targets.
3323 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
3324 favor of @code{gdbarch_breakpoint_from_pc}.
3326 @item const gdb_byte *gdbarch_breakpoint_from_pc (@var{gdbarch}, @var{pcptr}, @var{lenptr})
3327 @findex gdbarch_breakpoint_from_pc
3328 @anchor{gdbarch_breakpoint_from_pc} Use the program counter to determine the
3329 contents and size of a breakpoint instruction. It returns a pointer to
3330 a string of bytes that encode a breakpoint instruction, stores the
3331 length of the string to @code{*@var{lenptr}}, and adjusts the program
3332 counter (if necessary) to point to the actual memory location where the
3333 breakpoint should be inserted.
3335 Although it is common to use a trap instruction for a breakpoint, it's
3336 not required; for instance, the bit pattern could be an invalid
3337 instruction. The breakpoint must be no longer than the shortest
3338 instruction of the architecture.
3340 Replaces all the other @var{BREAKPOINT} macros.
3342 @item int gdbarch_memory_insert_breakpoint (@var{gdbarch}, @var{bp_tgt})
3343 @itemx gdbarch_memory_remove_breakpoint (@var{gdbarch}, @var{bp_tgt})
3344 @findex gdbarch_memory_remove_breakpoint
3345 @findex gdbarch_memory_insert_breakpoint
3346 Insert or remove memory based breakpoints. Reasonable defaults
3347 (@code{default_memory_insert_breakpoint} and
3348 @code{default_memory_remove_breakpoint} respectively) have been
3349 provided so that it is not necessary to set these for most
3350 architectures. Architectures which may want to set
3351 @code{gdbarch_memory_insert_breakpoint} and @code{gdbarch_memory_remove_breakpoint} will likely have instructions that are oddly sized or are not stored in a
3352 conventional manner.
3354 It may also be desirable (from an efficiency standpoint) to define
3355 custom breakpoint insertion and removal routines if
3356 @code{gdbarch_breakpoint_from_pc} needs to read the target's memory for some
3359 @item CORE_ADDR gdbarch_adjust_breakpoint_address (@var{gdbarch}, @var{bpaddr})
3360 @findex gdbarch_adjust_breakpoint_address
3361 @cindex breakpoint address adjusted
3362 Given an address at which a breakpoint is desired, return a breakpoint
3363 address adjusted to account for architectural constraints on
3364 breakpoint placement. This method is not needed by most targets.
3366 The FR-V target (see @file{frv-tdep.c}) requires this method.
3367 The FR-V is a VLIW architecture in which a number of RISC-like
3368 instructions are grouped (packed) together into an aggregate
3369 instruction or instruction bundle. When the processor executes
3370 one of these bundles, the component instructions are executed
3373 In the course of optimization, the compiler may group instructions
3374 from distinct source statements into the same bundle. The line number
3375 information associated with one of the latter statements will likely
3376 refer to some instruction other than the first one in the bundle. So,
3377 if the user attempts to place a breakpoint on one of these latter
3378 statements, @value{GDBN} must be careful to @emph{not} place the break
3379 instruction on any instruction other than the first one in the bundle.
3380 (Remember though that the instructions within a bundle execute
3381 in parallel, so the @emph{first} instruction is the instruction
3382 at the lowest address and has nothing to do with execution order.)
3384 The FR-V's @code{gdbarch_adjust_breakpoint_address} method will adjust a
3385 breakpoint's address by scanning backwards for the beginning of
3386 the bundle, returning the address of the bundle.
3388 Since the adjustment of a breakpoint may significantly alter a user's
3389 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
3390 is initially set and each time that that breakpoint is hit.
3392 @item int gdbarch_call_dummy_location (@var{gdbarch})
3393 @findex gdbarch_call_dummy_location
3394 See the file @file{inferior.h}.
3396 This method has been replaced by @code{gdbarch_push_dummy_code}
3397 (@pxref{gdbarch_push_dummy_code}).
3399 @item int gdbarch_cannot_fetch_register (@var{gdbarch}, @var{regum})
3400 @findex gdbarch_cannot_fetch_register
3401 This function should return nonzero if @var{regno} cannot be fetched
3402 from an inferior process. This is only relevant if
3403 @code{FETCH_INFERIOR_REGISTERS} is not defined.
3405 @item int gdbarch_cannot_store_register (@var{gdbarch}, @var{regnum})
3406 @findex gdbarch_cannot_store_register
3407 This function should return nonzero if @var{regno} should not be
3408 written to the target. This is often the case for program counters,
3409 status words, and other special registers. This function returns 0 as
3410 default so that @value{GDBN} will assume that all registers may be written.
3412 @item int gdbarch_convert_register_p (@var{gdbarch}, @var{regnum}, struct type *@var{type})
3413 @findex gdbarch_convert_register_p
3414 Return non-zero if register @var{regnum} can represent data values in a
3416 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3418 @item CORE_ADDR gdbarch_decr_pc_after_break (@var{gdbarch})
3419 @findex gdbarch_decr_pc_after_break
3420 This function shall return the amount by which to decrement the PC after the
3421 program encounters a breakpoint. This is often the number of bytes in
3422 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
3424 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
3425 @findex DISABLE_UNSETTABLE_BREAK
3426 If defined, this should evaluate to 1 if @var{addr} is in a shared
3427 library in which breakpoints cannot be set and so should be disabled.
3429 @item void gdbarch_print_float_info (@var{gdbarch}, @var{file}, @var{frame}, @var{args})
3430 @findex gdbarch_print_float_info
3431 If defined, then the @samp{info float} command will print information about
3432 the processor's floating point unit.
3434 @item void gdbarch_print_registers_info (@var{gdbarch}, @var{frame}, @var{regnum}, @var{all})
3435 @findex gdbarch_print_registers_info
3436 If defined, pretty print the value of the register @var{regnum} for the
3437 specified @var{frame}. If the value of @var{regnum} is -1, pretty print
3438 either all registers (@var{all} is non zero) or a select subset of
3439 registers (@var{all} is zero).
3441 The default method prints one register per line, and if @var{all} is
3442 zero omits floating-point registers.
3444 @item int gdbarch_print_vector_info (@var{gdbarch}, @var{file}, @var{frame}, @var{args})
3445 @findex gdbarch_print_vector_info
3446 If defined, then the @samp{info vector} command will call this function
3447 to print information about the processor's vector unit.
3449 By default, the @samp{info vector} command will print all vector
3450 registers (the register's type having the vector attribute).
3452 @item int gdbarch_dwarf_reg_to_regnum (@var{gdbarch}, @var{dwarf_regnr})
3453 @findex gdbarch_dwarf_reg_to_regnum
3454 Convert DWARF register number @var{dwarf_regnr} into @value{GDBN} regnum. If
3455 not defined, no conversion will be performed.
3457 @item int gdbarch_dwarf2_reg_to_regnum (@var{gdbarch}, @var{dwarf2_regnr})
3458 @findex gdbarch_dwarf2_reg_to_regnum
3459 Convert DWARF2 register number @var{dwarf2_regnr} into @value{GDBN} regnum.
3460 If not defined, no conversion will be performed.
3462 @item int gdbarch_ecoff_reg_to_regnum (@var{gdbarch}, @var{ecoff_regnr})
3463 @findex gdbarch_ecoff_reg_to_regnum
3464 Convert ECOFF register number @var{ecoff_regnr} into @value{GDBN} regnum. If
3465 not defined, no conversion will be performed.
3467 @item void gdbarch_extract_return_value (@var{gdbarch}, @var{type}, @var{regbuf}, @var{valbuf})
3468 @findex gdbarch_extract_return_value
3469 Define this to extract a function's return value of type @var{type} from
3470 the raw register state @var{regbuf} and copy that, in virtual format,
3473 This method has been deprecated in favour of @code{gdbarch_return_value}
3474 (@pxref{gdbarch_return_value}).
3476 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS(@var{regbuf})
3477 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS
3478 @anchor{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}
3479 When defined, extract from the array @var{regbuf} (containing the raw
3480 register state) the @code{CORE_ADDR} at which a function should return
3481 its structure value.
3483 @xref{gdbarch_return_value}.
3485 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P()
3486 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P
3487 Predicate for @code{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}.
3489 @item DEPRECATED_FP_REGNUM
3490 @findex DEPRECATED_FP_REGNUM
3491 If the virtual frame pointer is kept in a register, then define this
3492 macro to be the number (greater than or equal to zero) of that register.
3494 This should only need to be defined if @code{DEPRECATED_TARGET_READ_FP}
3497 @item DEPRECATED_FRAMELESS_FUNCTION_INVOCATION(@var{fi})
3498 @findex DEPRECATED_FRAMELESS_FUNCTION_INVOCATION
3499 Define this to an expression that returns 1 if the function invocation
3500 represented by @var{fi} does not have a stack frame associated with it.
3503 @item CORE_ADDR frame_align (@var{gdbarch}, @var{address})
3504 @anchor{frame_align}
3506 Define this to adjust @var{address} so that it meets the alignment
3507 requirements for the start of a new stack frame. A stack frame's
3508 alignment requirements are typically stronger than a target processors
3509 stack alignment requirements.
3511 This function is used to ensure that, when creating a dummy frame, both
3512 the initial stack pointer and (if needed) the address of the return
3513 value are correctly aligned.
3515 This function always adjusts the address in the direction of stack
3518 By default, no frame based stack alignment is performed.
3520 @item int gdbarch_frame_red_zone_size (@var{gdbarch})
3521 @findex gdbarch_frame_red_zone_size
3522 The number of bytes, beyond the innermost-stack-address, reserved by the
3523 @sc{abi}. A function is permitted to use this scratch area (instead of
3524 allocating extra stack space).
3526 When performing an inferior function call, to ensure that it does not
3527 modify this area, @value{GDBN} adjusts the innermost-stack-address by
3528 @var{gdbarch_frame_red_zone_size} bytes before pushing parameters onto the
3531 By default, zero bytes are allocated. The value must be aligned
3532 (@pxref{frame_align}).
3534 The @sc{amd64} (nee x86-64) @sc{abi} documentation refers to the
3535 @emph{red zone} when describing this scratch area.
3538 @item DEPRECATED_FRAME_CHAIN(@var{frame})
3539 @findex DEPRECATED_FRAME_CHAIN
3540 Given @var{frame}, return a pointer to the calling frame.
3542 @item DEPRECATED_FRAME_CHAIN_VALID(@var{chain}, @var{thisframe})
3543 @findex DEPRECATED_FRAME_CHAIN_VALID
3544 Define this to be an expression that returns zero if the given frame is an
3545 outermost frame, with no caller, and nonzero otherwise. Most normal
3546 situations can be handled without defining this macro, including @code{NULL}
3547 chain pointers, dummy frames, and frames whose PC values are inside the
3548 startup file (e.g.@: @file{crt0.o}), inside @code{main}, or inside
3551 @item DEPRECATED_FRAME_INIT_SAVED_REGS(@var{frame})
3552 @findex DEPRECATED_FRAME_INIT_SAVED_REGS
3553 See @file{frame.h}. Determines the address of all registers in the
3554 current stack frame storing each in @code{frame->saved_regs}. Space for
3555 @code{frame->saved_regs} shall be allocated by
3556 @code{DEPRECATED_FRAME_INIT_SAVED_REGS} using
3557 @code{frame_saved_regs_zalloc}.
3559 @code{FRAME_FIND_SAVED_REGS} is deprecated.
3561 @item int gdbarch_frame_num_args (@var{gdbarch}, @var{frame})
3562 @findex gdbarch_frame_num_args
3563 For the frame described by @var{frame} return the number of arguments that
3564 are being passed. If the number of arguments is not known, return
3567 @item DEPRECATED_FRAME_SAVED_PC(@var{frame})
3568 @findex DEPRECATED_FRAME_SAVED_PC
3569 @anchor{DEPRECATED_FRAME_SAVED_PC} Given @var{frame}, return the pc
3570 saved there. This is the return address.
3572 This method is deprecated. @xref{gdbarch_unwind_pc}.
3574 @item CORE_ADDR gdbarch_unwind_pc (@var{next_frame})
3575 @findex gdbarch_unwind_pc
3576 @anchor{gdbarch_unwind_pc} Return the instruction address, in
3577 @var{next_frame}'s caller, at which execution will resume after
3578 @var{next_frame} returns. This is commonly referred to as the return address.
3580 The implementation, which must be frame agnostic (work with any frame),
3581 is typically no more than:
3585 pc = frame_unwind_unsigned_register (next_frame, S390_PC_REGNUM);
3586 return gdbarch_addr_bits_remove (gdbarch, pc);
3590 @xref{DEPRECATED_FRAME_SAVED_PC}, which this method replaces.
3592 @item CORE_ADDR gdbarch_unwind_sp (@var{gdbarch}, @var{next_frame})
3593 @findex gdbarch_unwind_sp
3594 @anchor{gdbarch_unwind_sp} Return the frame's inner most stack address. This is
3595 commonly referred to as the frame's @dfn{stack pointer}.
3597 The implementation, which must be frame agnostic (work with any frame),
3598 is typically no more than:
3602 sp = frame_unwind_unsigned_register (next_frame, S390_SP_REGNUM);
3603 return gdbarch_addr_bits_remove (gdbarch, sp);
3607 @xref{TARGET_READ_SP}, which this method replaces.
3609 @item FUNCTION_EPILOGUE_SIZE
3610 @findex FUNCTION_EPILOGUE_SIZE
3611 For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the
3612 function end symbol is 0. For such targets, you must define
3613 @code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size of a
3614 function's epilogue.
3616 @item DEPRECATED_FUNCTION_START_OFFSET
3617 @findex DEPRECATED_FUNCTION_START_OFFSET
3618 An integer, giving the offset in bytes from a function's address (as
3619 used in the values of symbols, function pointers, etc.), and the
3620 function's first genuine instruction.
3622 This is zero on almost all machines: the function's address is usually
3623 the address of its first instruction. However, on the VAX, for
3624 example, each function starts with two bytes containing a bitmask
3625 indicating which registers to save upon entry to the function. The
3626 VAX @code{call} instructions check this value, and save the
3627 appropriate registers automatically. Thus, since the offset from the
3628 function's address to its first instruction is two bytes,
3629 @code{DEPRECATED_FUNCTION_START_OFFSET} would be 2 on the VAX.
3631 @item GCC_COMPILED_FLAG_SYMBOL
3632 @itemx GCC2_COMPILED_FLAG_SYMBOL
3633 @findex GCC2_COMPILED_FLAG_SYMBOL
3634 @findex GCC_COMPILED_FLAG_SYMBOL
3635 If defined, these are the names of the symbols that @value{GDBN} will
3636 look for to detect that GCC compiled the file. The default symbols
3637 are @code{gcc_compiled.} and @code{gcc2_compiled.},
3638 respectively. (Currently only defined for the Delta 68.)
3640 @item gdbarch_get_longjmp_target
3641 @findex gdbarch_get_longjmp_target
3642 For most machines, this is a target-dependent parameter. On the
3643 DECstation and the Iris, this is a native-dependent parameter, since
3644 the header file @file{setjmp.h} is needed to define it.
3646 This macro determines the target PC address that @code{longjmp} will jump to,
3647 assuming that we have just stopped at a @code{longjmp} breakpoint. It takes a
3648 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3649 pointer. It examines the current state of the machine as needed.
3651 @item DEPRECATED_IBM6000_TARGET
3652 @findex DEPRECATED_IBM6000_TARGET
3653 Shows that we are configured for an IBM RS/6000 system. This
3654 conditional should be eliminated (FIXME) and replaced by
3655 feature-specific macros. It was introduced in a haste and we are
3656 repenting at leisure.
3658 @item I386_USE_GENERIC_WATCHPOINTS
3659 An x86-based target can define this to use the generic x86 watchpoint
3660 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3662 @item int gdbarch_inner_than (@var{gdbarch}, @var{lhs}, @var{rhs})
3663 @findex gdbarch_inner_than
3664 Returns non-zero if stack address @var{lhs} is inner than (nearer to the
3665 stack top) stack address @var{rhs}. Let the function return
3666 @w{@code{lhs < rhs}} if the target's stack grows downward in memory, or
3667 @w{@code{lhs > rsh}} if the stack grows upward.
3669 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{addr})
3670 @findex gdbarch_in_function_epilogue_p
3671 Returns non-zero if the given @var{addr} is in the epilogue of a function.
3672 The epilogue of a function is defined as the part of a function where
3673 the stack frame of the function already has been destroyed up to the
3674 final `return from function call' instruction.
3676 @item int gdbarch_in_solib_return_trampoline (@var{gdbarch}, @var{pc}, @var{name})
3677 @findex gdbarch_in_solib_return_trampoline
3678 Define this function to return nonzero if the program is stopped in the
3679 trampoline that returns from a shared library.
3681 @item IN_SOLIB_DYNSYM_RESOLVE_CODE (@var{pc})
3682 @findex IN_SOLIB_DYNSYM_RESOLVE_CODE
3683 Define this to return nonzero if the program is stopped in the
3686 @item SKIP_SOLIB_RESOLVER (@var{pc})
3687 @findex SKIP_SOLIB_RESOLVER
3688 Define this to evaluate to the (nonzero) address at which execution
3689 should continue to get past the dynamic linker's symbol resolution
3690 function. A zero value indicates that it is not important or necessary
3691 to set a breakpoint to get through the dynamic linker and that single
3692 stepping will suffice.
3694 @item CORE_ADDR gdbarch_integer_to_address (@var{gdbarch}, @var{type}, @var{buf})
3695 @findex gdbarch_integer_to_address
3696 @cindex converting integers to addresses
3697 Define this when the architecture needs to handle non-pointer to address
3698 conversions specially. Converts that value to an address according to
3699 the current architectures conventions.
3701 @emph{Pragmatics: When the user copies a well defined expression from
3702 their source code and passes it, as a parameter, to @value{GDBN}'s
3703 @code{print} command, they should get the same value as would have been
3704 computed by the target program. Any deviation from this rule can cause
3705 major confusion and annoyance, and needs to be justified carefully. In
3706 other words, @value{GDBN} doesn't really have the freedom to do these
3707 conversions in clever and useful ways. It has, however, been pointed
3708 out that users aren't complaining about how @value{GDBN} casts integers
3709 to pointers; they are complaining that they can't take an address from a
3710 disassembly listing and give it to @code{x/i}. Adding an architecture
3711 method like @code{gdbarch_integer_to_address} certainly makes it possible for
3712 @value{GDBN} to ``get it right'' in all circumstances.}
3714 @xref{Target Architecture Definition, , Pointers Are Not Always
3717 @item CORE_ADDR gdbarch_pointer_to_address (@var{gdbarch}, @var{type}, @var{buf})
3718 @findex gdbarch_pointer_to_address
3719 Assume that @var{buf} holds a pointer of type @var{type}, in the
3720 appropriate format for the current architecture. Return the byte
3721 address the pointer refers to.
3722 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3724 @item void gdbarch_register_to_value(@var{gdbarch}, @var{frame}, @var{regnum}, @var{type}, @var{fur})
3725 @findex gdbarch_register_to_value
3726 Convert the raw contents of register @var{regnum} into a value of type
3728 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3730 @item register_reggroup_p (@var{gdbarch}, @var{regnum}, @var{reggroup})
3731 @findex register_reggroup_p
3732 @cindex register groups
3733 Return non-zero if register @var{regnum} is a member of the register
3734 group @var{reggroup}.
3736 By default, registers are grouped as follows:
3739 @item float_reggroup
3740 Any register with a valid name and a floating-point type.
3741 @item vector_reggroup
3742 Any register with a valid name and a vector type.
3743 @item general_reggroup
3744 Any register with a valid name and a type other than vector or
3745 floating-point. @samp{float_reggroup}.
3747 @itemx restore_reggroup
3749 Any register with a valid name.
3752 @item DEPRECATED_REGISTER_VIRTUAL_SIZE (@var{reg})
3753 @findex DEPRECATED_REGISTER_VIRTUAL_SIZE
3754 Return the virtual size of @var{reg}; defaults to the size of the
3755 register's virtual type.
3756 Return the virtual size of @var{reg}.
3757 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3759 @item DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})
3760 @findex REGISTER_VIRTUAL_TYPE
3761 Return the virtual type of @var{reg}.
3762 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3764 @item struct type *register_type (@var{gdbarch}, @var{reg})
3765 @findex register_type
3766 If defined, return the type of register @var{reg}. This function
3767 supersedes @code{DEPRECATED_REGISTER_VIRTUAL_TYPE}. @xref{Target Architecture
3768 Definition, , Raw and Virtual Register Representations}.
3770 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
3771 @findex REGISTER_CONVERT_TO_VIRTUAL
3772 Convert the value of register @var{reg} from its raw form to its virtual
3774 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3776 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
3777 @findex REGISTER_CONVERT_TO_RAW
3778 Convert the value of register @var{reg} from its virtual form to its raw
3780 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3782 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
3783 @findex regset_from_core_section
3784 Return the appropriate register set for a core file section with name
3785 @var{sect_name} and size @var{sect_size}.
3787 @item SOFTWARE_SINGLE_STEP_P()
3788 @findex SOFTWARE_SINGLE_STEP_P
3789 Define this as 1 if the target does not have a hardware single-step
3790 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
3792 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breakpoints_p})
3793 @findex SOFTWARE_SINGLE_STEP
3794 A function that inserts or removes (depending on
3795 @var{insert_breakpoints_p}) breakpoints at each possible destinations of
3796 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
3799 @item SOFUN_ADDRESS_MAYBE_MISSING
3800 @findex SOFUN_ADDRESS_MAYBE_MISSING
3801 Somebody clever observed that, the more actual addresses you have in the
3802 debug information, the more time the linker has to spend relocating
3803 them. So whenever there's some other way the debugger could find the
3804 address it needs, you should omit it from the debug info, to make
3807 @code{SOFUN_ADDRESS_MAYBE_MISSING} indicates that a particular set of
3808 hacks of this sort are in use, affecting @code{N_SO} and @code{N_FUN}
3809 entries in stabs-format debugging information. @code{N_SO} stabs mark
3810 the beginning and ending addresses of compilation units in the text
3811 segment. @code{N_FUN} stabs mark the starts and ends of functions.
3813 @code{SOFUN_ADDRESS_MAYBE_MISSING} means two things:
3817 @code{N_FUN} stabs have an address of zero. Instead, you should find the
3818 addresses where the function starts by taking the function name from
3819 the stab, and then looking that up in the minsyms (the
3820 linker/assembler symbol table). In other words, the stab has the
3821 name, and the linker/assembler symbol table is the only place that carries
3825 @code{N_SO} stabs have an address of zero, too. You just look at the
3826 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab,
3827 and guess the starting and ending addresses of the compilation unit from
3831 @item int gdbarch_pc_regnum (@var{gdbarch})
3832 @findex gdbarch_pc_regnum
3833 If the program counter is kept in a register, then let this function return
3834 the number (greater than or equal to zero) of that register.
3836 This should only need to be defined if @code{gdbarch_read_pc} and
3837 @code{gdbarch_write_pc} are not defined.
3839 @item int gdbarch_stabs_argument_has_addr (@var{gdbarch}, @var{type})
3840 @findex gdbarch_stabs_argument_has_addr
3841 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3842 @anchor{gdbarch_stabs_argument_has_addr} Define this function to return
3843 nonzero if a function argument of type @var{type} is passed by reference
3846 This method replaces @code{DEPRECATED_REG_STRUCT_HAS_ADDR}
3847 (@pxref{DEPRECATED_REG_STRUCT_HAS_ADDR}).
3849 @item PROCESS_LINENUMBER_HOOK
3850 @findex PROCESS_LINENUMBER_HOOK
3851 A hook defined for XCOFF reading.
3853 @item gdbarch_ps_regnum (@var{gdbarch}
3854 @findex gdbarch_ps_regnum
3855 If defined, this function returns the number of the processor status
3857 (This definition is only used in generic code when parsing "$ps".)
3859 @item CORE_ADDR gdbarch_push_dummy_call (@var{gdbarch}, @var{function}, @var{regcache}, @var{bp_addr}, @var{nargs}, @var{args}, @var{sp}, @var{struct_return}, @var{struct_addr})
3860 @findex gdbarch_push_dummy_call
3861 @findex DEPRECATED_PUSH_ARGUMENTS.
3862 @anchor{gdbarch_push_dummy_call} Define this to push the dummy frame's call to
3863 the inferior function onto the stack. In addition to pushing @var{nargs}, the
3864 code should push @var{struct_addr} (when @var{struct_return} is non-zero), and
3865 the return address (@var{bp_addr}).
3867 @var{function} is a pointer to a @code{struct value}; on architectures that use
3868 function descriptors, this contains the function descriptor value.
3870 Returns the updated top-of-stack pointer.
3872 This method replaces @code{DEPRECATED_PUSH_ARGUMENTS}.
3874 @item CORE_ADDR gdbarch_push_dummy_code (@var{gdbarch}, @var{sp}, @var{funaddr}, @var{using_gcc}, @var{args}, @var{nargs}, @var{value_type}, @var{real_pc}, @var{bp_addr}, @var{regcache})
3875 @findex gdbarch_push_dummy_code
3876 @anchor{gdbarch_push_dummy_code} Given a stack based call dummy, push the
3877 instruction sequence (including space for a breakpoint) to which the
3878 called function should return.
3880 Set @var{bp_addr} to the address at which the breakpoint instruction
3881 should be inserted, @var{real_pc} to the resume address when starting
3882 the call sequence, and return the updated inner-most stack address.
3884 By default, the stack is grown sufficient to hold a frame-aligned
3885 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
3886 reserved for that breakpoint, and @var{real_pc} set to @var{funaddr}.
3888 This method replaces @w{@code{gdbarch_call_dummy_location (@var{gdbarch})}} and
3889 @code{DEPRECATED_REGISTER_SIZE}.
3891 @item const char *gdbarch_register_name (@var{gdbarch}, @var{regnr})
3892 @findex gdbarch_register_name
3893 Return the name of register @var{regnr} as a string. May return @code{NULL}
3894 to indicate that @var{regnr} is not a valid register.
3896 @item DEPRECATED_REG_STRUCT_HAS_ADDR (@var{gcc_p}, @var{type})
3897 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3898 @anchor{DEPRECATED_REG_STRUCT_HAS_ADDR}Define this to return 1 if the
3899 given type will be passed by pointer rather than directly.
3901 This method has been replaced by @code{gdbarch_stabs_argument_has_addr}
3902 (@pxref{gdbarch_stabs_argument_has_addr}).
3904 @item SAVE_DUMMY_FRAME_TOS (@var{sp})
3905 @findex SAVE_DUMMY_FRAME_TOS
3906 @anchor{SAVE_DUMMY_FRAME_TOS} Used in @samp{call_function_by_hand} to
3907 notify the target dependent code of the top-of-stack value that will be
3908 passed to the inferior code. This is the value of the @code{SP}
3909 after both the dummy frame and space for parameters/results have been
3910 allocated on the stack. @xref{gdbarch_unwind_dummy_id}.
3912 @item int gdbarch_sdb_reg_to_regnum (@var{gdbarch}, @var{sdb_regnr})
3913 @findex gdbarch_sdb_reg_to_regnum
3914 Use this function to convert sdb register @var{sdb_regnr} into @value{GDBN}
3915 regnum. If not defined, no conversion will be done.
3917 @item enum return_value_convention gdbarch_return_value (struct gdbarch *@var{gdbarch}, struct type *@var{valtype}, struct regcache *@var{regcache}, void *@var{readbuf}, const void *@var{writebuf})
3918 @findex gdbarch_return_value
3919 @anchor{gdbarch_return_value} Given a function with a return-value of
3920 type @var{rettype}, return which return-value convention that function
3923 @value{GDBN} currently recognizes two function return-value conventions:
3924 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
3925 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
3926 value is found in memory and the address of that memory location is
3927 passed in as the function's first parameter.
3929 If the register convention is being used, and @var{writebuf} is
3930 non-@code{NULL}, also copy the return-value in @var{writebuf} into
3933 If the register convention is being used, and @var{readbuf} is
3934 non-@code{NULL}, also copy the return value from @var{regcache} into
3935 @var{readbuf} (@var{regcache} contains a copy of the registers from the
3936 just returned function).
3938 @xref{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}, for a description of how
3939 return-values that use the struct convention are handled.
3941 @emph{Maintainer note: This method replaces separate predicate, extract,
3942 store methods. By having only one method, the logic needed to determine
3943 the return-value convention need only be implemented in one place. If
3944 @value{GDBN} were written in an @sc{oo} language, this method would
3945 instead return an object that knew how to perform the register
3946 return-value extract and store.}
3948 @emph{Maintainer note: This method does not take a @var{gcc_p}
3949 parameter, and such a parameter should not be added. If an architecture
3950 that requires per-compiler or per-function information be identified,
3951 then the replacement of @var{rettype} with @code{struct value}
3952 @var{function} should be pursued.}
3954 @emph{Maintainer note: The @var{regcache} parameter limits this methods
3955 to the inner most frame. While replacing @var{regcache} with a
3956 @code{struct frame_info} @var{frame} parameter would remove that
3957 limitation there has yet to be a demonstrated need for such a change.}
3959 @item void gdbarch_skip_permanent_breakpoint (@var{gdbarch}, @var{regcache})
3960 @findex gdbarch_skip_permanent_breakpoint
3961 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
3962 steps over a breakpoint by removing it, stepping one instruction, and
3963 re-inserting the breakpoint. However, permanent breakpoints are
3964 hardwired into the inferior, and can't be removed, so this strategy
3965 doesn't work. Calling @code{gdbarch_skip_permanent_breakpoint} adjusts the
3966 processor's state so that execution will resume just after the breakpoint.
3967 This function does the right thing even when the breakpoint is in the delay slot
3968 of a branch or jump.
3970 @item CORE_ADDR gdbarch_skip_prologue (@var{gdbarch}, @var{ip})
3971 @findex gdbarch_skip_prologue
3972 A function that returns the address of the ``real'' code beyond the
3973 function entry prologue found at @var{ip}.
3975 @item CORE_ADDR gdbarch_skip_trampoline_code (@var{gdbarch}, @var{frame}, @var{pc})
3976 @findex gdbarch_skip_trampoline_code
3977 If the target machine has trampoline code that sits between callers and
3978 the functions being called, then define this function to return a new PC
3979 that is at the start of the real function.
3981 @item int gdbarch_sp_regnum (@var{gdbarch})
3982 @findex gdbarch_sp_regnum
3983 If the stack-pointer is kept in a register, then use this function to return
3984 the number (greater than or equal to zero) of that register, or -1 if
3985 there is no such register.
3987 @item int gdbarch_stab_reg_to_regnum (@var{gdbarch}, @var{stab_regnr})
3988 @findex gdbarch_stab_reg_to_regnum
3989 Use this function to convert stab register @var{stab_regnr} into @value{GDBN}
3990 regnum. If not defined, no conversion will be done.
3992 @item void gdbarch_store_return_value (@var{gdbarch}, @var{type}, @var{regcache}, @var{valbuf})
3993 @findex gdbarch_store_return_value
3994 A function that writes the function return value, found in
3995 @var{valbuf}, into the @var{regcache}. @var{type} is the type of the
3996 value that is to be returned.
3998 This method has been deprecated in favour of @code{gdbarch_return_value}
3999 (@pxref{gdbarch_return_value}).
4001 @item SYMBOL_RELOADING_DEFAULT
4002 @findex SYMBOL_RELOADING_DEFAULT
4003 The default value of the ``symbol-reloading'' variable. (Never defined in
4006 @item TARGET_CHAR_BIT
4007 @findex TARGET_CHAR_BIT
4008 Number of bits in a char; defaults to 8.
4010 @item int gdbarch_char_signed (@var{gdbarch})
4011 @findex gdbarch_char_signed
4012 Non-zero if @code{char} is normally signed on this architecture; zero if
4013 it should be unsigned.
4015 The ISO C standard requires the compiler to treat @code{char} as
4016 equivalent to either @code{signed char} or @code{unsigned char}; any
4017 character in the standard execution set is supposed to be positive.
4018 Most compilers treat @code{char} as signed, but @code{char} is unsigned
4019 on the IBM S/390, RS6000, and PowerPC targets.
4021 @item int gdbarch_double_bit (@var{gdbarch})
4022 @findex gdbarch_double_bit
4023 Number of bits in a double float; defaults to @w{@code{8 * TARGET_CHAR_BIT}}.
4025 @item int gdbarch_float_bit (@var{gdbarch})
4026 @findex gdbarch_float_bit
4027 Number of bits in a float; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4029 @item int gdbarch_int_bit (@var{gdbarch})
4030 @findex gdbarch_int_bit
4031 Number of bits in an integer; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4033 @item int gdbarch_long_bit (@var{gdbarch})
4034 @findex gdbarch_long_bit
4035 Number of bits in a long integer; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4037 @item int gdbarch_long_double_bit (@var{gdbarch})
4038 @findex gdbarch_long_double_bit
4039 Number of bits in a long double float;
4040 defaults to @w{@code{2 * gdbarch_double_bit (@var{gdbarch})}}.
4042 @item int gdbarch_long_long_bit (@var{gdbarch})
4043 @findex gdbarch_long_long_bit
4044 Number of bits in a long long integer; defaults to
4045 @w{@code{2 * gdbarch_long_bit (@var{gdbarch})}}.
4047 @item int gdbarch_ptr_bit (@var{gdbarch})
4048 @findex gdbarch_ptr_bit
4049 Number of bits in a pointer; defaults to
4050 @w{@code{gdbarch_int_bit (@var{gdbarch})}}.
4052 @item int gdbarch_short_bit (@var{gdbarch})
4053 @findex gdbarch_short_bit
4054 Number of bits in a short integer; defaults to @w{@code{2 * TARGET_CHAR_BIT}}.
4056 @item CORE_ADDR gdbarch_read_pc (@var{gdbarch}, @var{regcache})
4057 @findex gdbarch_read_pc
4058 @itemx gdbarch_write_pc (@var{gdbarch}, @var{regcache}, @var{val})
4059 @findex gdbarch_write_pc
4060 @anchor{gdbarch_write_pc}
4061 @itemx TARGET_READ_SP
4062 @findex TARGET_READ_SP
4063 @itemx TARGET_READ_FP
4064 @findex TARGET_READ_FP
4065 @findex gdbarch_read_pc
4066 @findex gdbarch_write_pc
4069 @anchor{TARGET_READ_SP} These change the behavior of @code{gdbarch_read_pc},
4070 @code{gdbarch_write_pc}, and @code{read_sp}. For most targets, these may be
4071 left undefined. @value{GDBN} will call the read and write register
4072 functions with the relevant @code{_REGNUM} argument.
4074 These macros and functions are useful when a target keeps one of these
4075 registers in a hard to get at place; for example, part in a segment register
4076 and part in an ordinary register.
4078 @xref{gdbarch_unwind_sp}, which replaces @code{TARGET_READ_SP}.
4080 @item void gdbarch_virtual_frame_pointer (@var{gdbarch}, @var{pc}, @var{frame_regnum}, @var{frame_offset})
4081 @findex gdbarch_virtual_frame_pointer
4082 Returns a @code{(register, offset)} pair representing the virtual frame
4083 pointer in use at the code address @var{pc}. If virtual frame pointers
4084 are not used, a default definition simply returns
4085 @code{DEPRECATED_FP_REGNUM}, with an offset of zero.
4087 @item TARGET_HAS_HARDWARE_WATCHPOINTS
4088 If non-zero, the target has support for hardware-assisted
4089 watchpoints. @xref{Algorithms, watchpoints}, for more details and
4090 other related macros.
4092 @item int gdbarch_print_insn (@var{gdbarch}, @var{vma}, @var{info})
4093 @findex gdbarch_print_insn
4094 This is the function used by @value{GDBN} to print an assembly
4095 instruction. It prints the instruction at address @var{vma} in
4096 debugged memory and returns the length of the instruction, in bytes. If
4097 a target doesn't define its own printing routine, it defaults to an
4098 accessor function for the global pointer
4099 @code{deprecated_tm_print_insn}. This usually points to a function in
4100 the @code{opcodes} library (@pxref{Support Libraries, ,Opcodes}).
4101 @var{info} is a structure (of type @code{disassemble_info}) defined in
4102 @file{include/dis-asm.h} used to pass information to the instruction
4105 @item frame_id gdbarch_unwind_dummy_id (@var{gdbarch}, @var{frame})
4106 @findex gdbarch_unwind_dummy_id
4107 @anchor{gdbarch_unwind_dummy_id} Given @var{frame} return a @w{@code{struct
4108 frame_id}} that uniquely identifies an inferior function call's dummy
4109 frame. The value returned must match the dummy frame stack value
4110 previously saved using @code{SAVE_DUMMY_FRAME_TOS}.
4111 @xref{SAVE_DUMMY_FRAME_TOS}.
4113 @item DEPRECATED_USE_STRUCT_CONVENTION (@var{gcc_p}, @var{type})
4114 @findex DEPRECATED_USE_STRUCT_CONVENTION
4115 If defined, this must be an expression that is nonzero if a value of the
4116 given @var{type} being returned from a function must have space
4117 allocated for it on the stack. @var{gcc_p} is true if the function
4118 being considered is known to have been compiled by GCC; this is helpful
4119 for systems where GCC is known to use different calling convention than
4122 This method has been deprecated in favour of @code{gdbarch_return_value}
4123 (@pxref{gdbarch_return_value}).
4125 @item void gdbarch_value_to_register (@var{gdbarch}, @var{frame}, @var{type}, @var{buf})
4126 @findex gdbarch_value_to_register
4127 Convert a value of type @var{type} into the raw contents of a register.
4128 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
4130 @item VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
4131 @findex VARIABLES_INSIDE_BLOCK
4132 For dbx-style debugging information, if the compiler puts variable
4133 declarations inside LBRAC/RBRAC blocks, this should be defined to be
4134 nonzero. @var{desc} is the value of @code{n_desc} from the
4135 @code{N_RBRAC} symbol, and @var{gcc_p} is true if @value{GDBN} has noticed the
4136 presence of either the @code{GCC_COMPILED_SYMBOL} or the
4137 @code{GCC2_COMPILED_SYMBOL}. By default, this is 0.
4141 Motorola M68K target conditionals.
4145 Define this to be the 4-bit location of the breakpoint trap vector. If
4146 not defined, it will default to @code{0xf}.
4148 @item REMOTE_BPT_VECTOR
4149 Defaults to @code{1}.
4151 @item const char *gdbarch_name_of_malloc (@var{gdbarch})
4152 @findex gdbarch_name_of_malloc
4153 A string containing the name of the function to call in order to
4154 allocate some memory in the inferior. The default value is "malloc".
4158 @node Adding a New Target
4159 @section Adding a New Target
4161 @cindex adding a target
4162 The following files add a target to @value{GDBN}:
4166 @item gdb/config/@var{arch}/@var{ttt}.mt
4167 Contains a Makefile fragment specific to this target. Specifies what
4168 object files are needed for target @var{ttt}, by defining
4169 @samp{TDEPFILES=@dots{}} and @samp{TDEPLIBS=@dots{}}. Also specifies
4170 the header file which describes @var{ttt}, by defining @samp{TM_FILE=
4173 You can also define @samp{TM_CFLAGS}, @samp{TM_CLIBS}, @samp{TM_CDEPS},
4174 but these are now deprecated, replaced by autoconf, and may go away in
4175 future versions of @value{GDBN}.
4177 @item gdb/@var{ttt}-tdep.c
4178 Contains any miscellaneous code required for this target machine. On
4179 some machines it doesn't exist at all. Sometimes the macros in
4180 @file{tm-@var{ttt}.h} become very complicated, so they are implemented
4181 as functions here instead, and the macro is simply defined to call the
4182 function. This is vastly preferable, since it is easier to understand
4185 @item gdb/@var{arch}-tdep.c
4186 @itemx gdb/@var{arch}-tdep.h
4187 This often exists to describe the basic layout of the target machine's
4188 processor chip (registers, stack, etc.). If used, it is included by
4189 @file{@var{ttt}-tdep.h}. It can be shared among many targets that use
4192 @item gdb/config/@var{arch}/tm-@var{ttt}.h
4193 (@file{tm.h} is a link to this file, created by @code{configure}). Contains
4194 macro definitions about the target machine's registers, stack frame
4195 format and instructions.
4197 New targets do not need this file and should not create it.
4199 @item gdb/config/@var{arch}/tm-@var{arch}.h
4200 This often exists to describe the basic layout of the target machine's
4201 processor chip (registers, stack, etc.). If used, it is included by
4202 @file{tm-@var{ttt}.h}. It can be shared among many targets that use the
4205 New targets do not need this file and should not create it.
4209 If you are adding a new operating system for an existing CPU chip, add a
4210 @file{config/tm-@var{os}.h} file that describes the operating system
4211 facilities that are unusual (extra symbol table info; the breakpoint
4212 instruction needed; etc.). Then write a @file{@var{arch}/tm-@var{os}.h}
4213 that just @code{#include}s @file{tm-@var{arch}.h} and
4214 @file{config/tm-@var{os}.h}.
4216 @node Target Descriptions
4217 @chapter Target Descriptions
4218 @cindex target descriptions
4220 The target architecture definition (@pxref{Target Architecture Definition})
4221 contains @value{GDBN}'s hard-coded knowledge about an architecture. For
4222 some platforms, it is handy to have more flexible knowledge about a specific
4223 instance of the architecture---for instance, a processor or development board.
4224 @dfn{Target descriptions} provide a mechanism for the user to tell @value{GDBN}
4225 more about what their target supports, or for the target to tell @value{GDBN}
4228 For details on writing, automatically supplying, and manually selecting
4229 target descriptions, see @ref{Target Descriptions, , , gdb,
4230 Debugging with @value{GDBN}}. This section will cover some related
4231 topics about the @value{GDBN} internals.
4234 * Target Descriptions Implementation::
4235 * Adding Target Described Register Support::
4238 @node Target Descriptions Implementation
4239 @section Target Descriptions Implementation
4240 @cindex target descriptions, implementation
4242 Before @value{GDBN} connects to a new target, or runs a new program on
4243 an existing target, it discards any existing target description and
4244 reverts to a default gdbarch. Then, after connecting, it looks for a
4245 new target description by calling @code{target_find_description}.
4247 A description may come from a user specified file (XML), the remote
4248 @samp{qXfer:features:read} packet (also XML), or from any custom
4249 @code{to_read_description} routine in the target vector. For instance,
4250 the remote target supports guessing whether a MIPS target is 32-bit or
4251 64-bit based on the size of the @samp{g} packet.
4253 If any target description is found, @value{GDBN} creates a new gdbarch
4254 incorporating the description by calling @code{gdbarch_update_p}. Any
4255 @samp{<architecture>} element is handled first, to determine which
4256 architecture's gdbarch initialization routine is called to create the
4257 new architecture. Then the initialization routine is called, and has
4258 a chance to adjust the constructed architecture based on the contents
4259 of the target description. For instance, it can recognize any
4260 properties set by a @code{to_read_description} routine. Also
4261 see @ref{Adding Target Described Register Support}.
4263 @node Adding Target Described Register Support
4264 @section Adding Target Described Register Support
4265 @cindex target descriptions, adding register support
4267 Target descriptions can report additional registers specific to an
4268 instance of the target. But it takes a little work in the architecture
4269 specific routines to support this.
4271 A target description must either have no registers or a complete
4272 set---this avoids complexity in trying to merge standard registers
4273 with the target defined registers. It is the architecture's
4274 responsibility to validate that a description with registers has
4275 everything it needs. To keep architecture code simple, the same
4276 mechanism is used to assign fixed internal register numbers to
4279 If @code{tdesc_has_registers} returns 1, the description contains
4280 registers. The architecture's @code{gdbarch_init} routine should:
4285 Call @code{tdesc_data_alloc} to allocate storage, early, before
4286 searching for a matching gdbarch or allocating a new one.
4289 Use @code{tdesc_find_feature} to locate standard features by name.
4292 Use @code{tdesc_numbered_register} and @code{tdesc_numbered_register_choices}
4293 to locate the expected registers in the standard features.
4296 Return @code{NULL} if a required feature is missing, or if any standard
4297 feature is missing expected registers. This will produce a warning that
4298 the description was incomplete.
4301 Free the allocated data before returning, unless @code{tdesc_use_registers}
4305 Call @code{set_gdbarch_num_regs} as usual, with a number higher than any
4306 fixed number passed to @code{tdesc_numbered_register}.
4309 Call @code{tdesc_use_registers} after creating a new gdbarch, before
4314 After @code{tdesc_use_registers} has been called, the architecture's
4315 @code{register_name}, @code{register_type}, and @code{register_reggroup_p}
4316 routines will not be called; that information will be taken from
4317 the target description. @code{num_regs} may be increased to account
4318 for any additional registers in the description.
4320 Pseudo-registers require some extra care:
4325 Using @code{tdesc_numbered_register} allows the architecture to give
4326 constant register numbers to standard architectural registers, e.g.@:
4327 as an @code{enum} in @file{@var{arch}-tdep.h}. But because
4328 pseudo-registers are always numbered above @code{num_regs},
4329 which may be increased by the description, constant numbers
4330 can not be used for pseudos. They must be numbered relative to
4331 @code{num_regs} instead.
4334 The description will not describe pseudo-registers, so the
4335 architecture must call @code{set_tdesc_pseudo_register_name},
4336 @code{set_tdesc_pseudo_register_type}, and
4337 @code{set_tdesc_pseudo_register_reggroup_p} to supply routines
4338 describing pseudo registers. These routines will be passed
4339 internal register numbers, so the same routines used for the
4340 gdbarch equivalents are usually suitable.
4345 @node Target Vector Definition
4347 @chapter Target Vector Definition
4348 @cindex target vector
4350 The target vector defines the interface between @value{GDBN}'s
4351 abstract handling of target systems, and the nitty-gritty code that
4352 actually exercises control over a process or a serial port.
4353 @value{GDBN} includes some 30-40 different target vectors; however,
4354 each configuration of @value{GDBN} includes only a few of them.
4357 * Managing Execution State::
4358 * Existing Targets::
4361 @node Managing Execution State
4362 @section Managing Execution State
4363 @cindex execution state
4365 A target vector can be completely inactive (not pushed on the target
4366 stack), active but not running (pushed, but not connected to a fully
4367 manifested inferior), or completely active (pushed, with an accessible
4368 inferior). Most targets are only completely inactive or completely
4369 active, but some support persistent connections to a target even
4370 when the target has exited or not yet started.
4372 For example, connecting to the simulator using @code{target sim} does
4373 not create a running program. Neither registers nor memory are
4374 accessible until @code{run}. Similarly, after @code{kill}, the
4375 program can not continue executing. But in both cases @value{GDBN}
4376 remains connected to the simulator, and target-specific commands
4377 are directed to the simulator.
4379 A target which only supports complete activation should push itself
4380 onto the stack in its @code{to_open} routine (by calling
4381 @code{push_target}), and unpush itself from the stack in its
4382 @code{to_mourn_inferior} routine (by calling @code{unpush_target}).
4384 A target which supports both partial and complete activation should
4385 still call @code{push_target} in @code{to_open}, but not call
4386 @code{unpush_target} in @code{to_mourn_inferior}. Instead, it should
4387 call either @code{target_mark_running} or @code{target_mark_exited}
4388 in its @code{to_open}, depending on whether the target is fully active
4389 after connection. It should also call @code{target_mark_running} any
4390 time the inferior becomes fully active (e.g.@: in
4391 @code{to_create_inferior} and @code{to_attach}), and
4392 @code{target_mark_exited} when the inferior becomes inactive (in
4393 @code{to_mourn_inferior}). The target should also make sure to call
4394 @code{target_mourn_inferior} from its @code{to_kill}, to return the
4395 target to inactive state.
4397 @node Existing Targets
4398 @section Existing Targets
4401 @subsection File Targets
4403 Both executables and core files have target vectors.
4405 @subsection Standard Protocol and Remote Stubs
4407 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code
4408 that runs in the target system. @value{GDBN} provides several sample
4409 @dfn{stubs} that can be integrated into target programs or operating
4410 systems for this purpose; they are named @file{*-stub.c}.
4412 The @value{GDBN} user's manual describes how to put such a stub into
4413 your target code. What follows is a discussion of integrating the
4414 SPARC stub into a complicated operating system (rather than a simple
4415 program), by Stu Grossman, the author of this stub.
4417 The trap handling code in the stub assumes the following upon entry to
4422 %l1 and %l2 contain pc and npc respectively at the time of the trap;
4428 you are in the correct trap window.
4431 As long as your trap handler can guarantee those conditions, then there
4432 is no reason why you shouldn't be able to ``share'' traps with the stub.
4433 The stub has no requirement that it be jumped to directly from the
4434 hardware trap vector. That is why it calls @code{exceptionHandler()},
4435 which is provided by the external environment. For instance, this could
4436 set up the hardware traps to actually execute code which calls the stub
4437 first, and then transfers to its own trap handler.
4439 For the most point, there probably won't be much of an issue with
4440 ``sharing'' traps, as the traps we use are usually not used by the kernel,
4441 and often indicate unrecoverable error conditions. Anyway, this is all
4442 controlled by a table, and is trivial to modify. The most important
4443 trap for us is for @code{ta 1}. Without that, we can't single step or
4444 do breakpoints. Everything else is unnecessary for the proper operation
4445 of the debugger/stub.
4447 From reading the stub, it's probably not obvious how breakpoints work.
4448 They are simply done by deposit/examine operations from @value{GDBN}.
4450 @subsection ROM Monitor Interface
4452 @subsection Custom Protocols
4454 @subsection Transport Layer
4456 @subsection Builtin Simulator
4459 @node Native Debugging
4461 @chapter Native Debugging
4462 @cindex native debugging
4464 Several files control @value{GDBN}'s configuration for native support:
4468 @item gdb/config/@var{arch}/@var{xyz}.mh
4469 Specifies Makefile fragments needed by a @emph{native} configuration on
4470 machine @var{xyz}. In particular, this lists the required
4471 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
4472 Also specifies the header file which describes native support on
4473 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
4474 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
4475 @samp{NAT_CDEPS}, etc.; see @file{Makefile.in}.
4477 @emph{Maintainer's note: The @file{.mh} suffix is because this file
4478 originally contained @file{Makefile} fragments for hosting @value{GDBN}
4479 on machine @var{xyz}. While the file is no longer used for this
4480 purpose, the @file{.mh} suffix remains. Perhaps someone will
4481 eventually rename these fragments so that they have a @file{.mn}
4484 @item gdb/config/@var{arch}/nm-@var{xyz}.h
4485 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
4486 macro definitions describing the native system environment, such as
4487 child process control and core file support.
4489 @item gdb/@var{xyz}-nat.c
4490 Contains any miscellaneous C code required for this native support of
4491 this machine. On some machines it doesn't exist at all.
4494 There are some ``generic'' versions of routines that can be used by
4495 various systems. These can be customized in various ways by macros
4496 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
4497 the @var{xyz} host, you can just include the generic file's name (with
4498 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
4500 Otherwise, if your machine needs custom support routines, you will need
4501 to write routines that perform the same functions as the generic file.
4502 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
4503 into @code{NATDEPFILES}.
4507 This contains the @emph{target_ops vector} that supports Unix child
4508 processes on systems which use ptrace and wait to control the child.
4511 This contains the @emph{target_ops vector} that supports Unix child
4512 processes on systems which use /proc to control the child.
4515 This does the low-level grunge that uses Unix system calls to do a ``fork
4516 and exec'' to start up a child process.
4519 This is the low level interface to inferior processes for systems using
4520 the Unix @code{ptrace} call in a vanilla way.
4523 @section Native core file Support
4524 @cindex native core files
4527 @findex fetch_core_registers
4528 @item core-aout.c::fetch_core_registers()
4529 Support for reading registers out of a core file. This routine calls
4530 @code{register_addr()}, see below. Now that BFD is used to read core
4531 files, virtually all machines should use @code{core-aout.c}, and should
4532 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
4533 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
4535 @item core-aout.c::register_addr()
4536 If your @code{nm-@var{xyz}.h} file defines the macro
4537 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
4538 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
4539 register number @code{regno}. @code{blockend} is the offset within the
4540 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
4541 @file{core-aout.c} will define the @code{register_addr()} function and
4542 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
4543 you are using the standard @code{fetch_core_registers()}, you will need
4544 to define your own version of @code{register_addr()}, put it into your
4545 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
4546 the @code{NATDEPFILES} list. If you have your own
4547 @code{fetch_core_registers()}, you may not need a separate
4548 @code{register_addr()}. Many custom @code{fetch_core_registers()}
4549 implementations simply locate the registers themselves.@refill
4552 When making @value{GDBN} run native on a new operating system, to make it
4553 possible to debug core files, you will need to either write specific
4554 code for parsing your OS's core files, or customize
4555 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
4556 machine uses to define the struct of registers that is accessible
4557 (possibly in the u-area) in a core file (rather than
4558 @file{machine/reg.h}), and an include file that defines whatever header
4559 exists on a core file (e.g., the u-area or a @code{struct core}). Then
4560 modify @code{trad_unix_core_file_p} to use these values to set up the
4561 section information for the data segment, stack segment, any other
4562 segments in the core file (perhaps shared library contents or control
4563 information), ``registers'' segment, and if there are two discontiguous
4564 sets of registers (e.g., integer and float), the ``reg2'' segment. This
4565 section information basically delimits areas in the core file in a
4566 standard way, which the section-reading routines in BFD know how to seek
4569 Then back in @value{GDBN}, you need a matching routine called
4570 @code{fetch_core_registers}. If you can use the generic one, it's in
4571 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
4572 It will be passed a char pointer to the entire ``registers'' segment,
4573 its length, and a zero; or a char pointer to the entire ``regs2''
4574 segment, its length, and a 2. The routine should suck out the supplied
4575 register values and install them into @value{GDBN}'s ``registers'' array.
4577 If your system uses @file{/proc} to control processes, and uses ELF
4578 format core files, then you may be able to use the same routines for
4579 reading the registers out of processes and out of core files.
4587 @section shared libraries
4589 @section Native Conditionals
4590 @cindex native conditionals
4592 When @value{GDBN} is configured and compiled, various macros are
4593 defined or left undefined, to control compilation when the host and
4594 target systems are the same. These macros should be defined (or left
4595 undefined) in @file{nm-@var{system}.h}.
4599 @item CHILD_PREPARE_TO_STORE
4600 @findex CHILD_PREPARE_TO_STORE
4601 If the machine stores all registers at once in the child process, then
4602 define this to ensure that all values are correct. This usually entails
4603 a read from the child.
4605 [Note that this is incorrectly defined in @file{xm-@var{system}.h} files
4608 @item FETCH_INFERIOR_REGISTERS
4609 @findex FETCH_INFERIOR_REGISTERS
4610 Define this if the native-dependent code will provide its own routines
4611 @code{fetch_inferior_registers} and @code{store_inferior_registers} in
4612 @file{@var{host}-nat.c}. If this symbol is @emph{not} defined, and
4613 @file{infptrace.c} is included in this configuration, the default
4614 routines in @file{infptrace.c} are used for these functions.
4616 @item int gdbarch_fp0_regnum (@var{gdbarch})
4617 @findex gdbarch_fp0_regnum
4618 This functions normally returns the number of the first floating
4619 point register, if the machine has such registers. As such, it would
4620 appear only in target-specific code. However, @file{/proc} support uses this
4621 to decide whether floats are in use on this target.
4623 @item int gdbarch_get_longjmp_target (@var{gdbarch})
4624 @findex gdbarch_get_longjmp_target
4625 For most machines, this is a target-dependent parameter. On the
4626 DECstation and the Iris, this is a native-dependent parameter, since
4627 @file{setjmp.h} is needed to define it.
4629 This function determines the target PC address that @code{longjmp} will jump to,
4630 assuming that we have just stopped at a longjmp breakpoint. It takes a
4631 @code{CORE_ADDR *} as argument, and stores the target PC value through this
4632 pointer. It examines the current state of the machine as needed.
4634 @item I386_USE_GENERIC_WATCHPOINTS
4635 An x86-based machine can define this to use the generic x86 watchpoint
4636 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4638 @item ONE_PROCESS_WRITETEXT
4639 @findex ONE_PROCESS_WRITETEXT
4640 Define this to be able to, when a breakpoint insertion fails, warn the
4641 user that another process may be running with the same executable.
4644 @findex PROC_NAME_FMT
4645 Defines the format for the name of a @file{/proc} device. Should be
4646 defined in @file{nm.h} @emph{only} in order to override the default
4647 definition in @file{procfs.c}.
4649 @item SHELL_COMMAND_CONCAT
4650 @findex SHELL_COMMAND_CONCAT
4651 If defined, is a string to prefix on the shell command used to start the
4656 If defined, this is the name of the shell to use to run the inferior.
4657 Defaults to @code{"/bin/sh"}.
4659 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
4661 Define this to expand into an expression that will cause the symbols in
4662 @var{filename} to be added to @value{GDBN}'s symbol table. If
4663 @var{readsyms} is zero symbols are not read but any necessary low level
4664 processing for @var{filename} is still done.
4666 @item SOLIB_CREATE_INFERIOR_HOOK
4667 @findex SOLIB_CREATE_INFERIOR_HOOK
4668 Define this to expand into any shared-library-relocation code that you
4669 want to be run just after the child process has been forked.
4671 @item START_INFERIOR_TRAPS_EXPECTED
4672 @findex START_INFERIOR_TRAPS_EXPECTED
4673 When starting an inferior, @value{GDBN} normally expects to trap
4675 the shell execs, and once when the program itself execs. If the actual
4676 number of traps is something other than 2, then define this macro to
4677 expand into the number expected.
4681 See @file{objfiles.c}.
4685 @node Support Libraries
4687 @chapter Support Libraries
4692 BFD provides support for @value{GDBN} in several ways:
4695 @item identifying executable and core files
4696 BFD will identify a variety of file types, including a.out, coff, and
4697 several variants thereof, as well as several kinds of core files.
4699 @item access to sections of files
4700 BFD parses the file headers to determine the names, virtual addresses,
4701 sizes, and file locations of all the various named sections in files
4702 (such as the text section or the data section). @value{GDBN} simply
4703 calls BFD to read or write section @var{x} at byte offset @var{y} for
4706 @item specialized core file support
4707 BFD provides routines to determine the failing command name stored in a
4708 core file, the signal with which the program failed, and whether a core
4709 file matches (i.e.@: could be a core dump of) a particular executable
4712 @item locating the symbol information
4713 @value{GDBN} uses an internal interface of BFD to determine where to find the
4714 symbol information in an executable file or symbol-file. @value{GDBN} itself
4715 handles the reading of symbols, since BFD does not ``understand'' debug
4716 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
4721 @cindex opcodes library
4723 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
4724 library because it's also used in binutils, for @file{objdump}).
4727 @cindex readline library
4728 The @code{readline} library provides a set of functions for use by applications
4729 that allow users to edit command lines as they are typed in.
4732 @cindex @code{libiberty} library
4734 The @code{libiberty} library provides a set of functions and features
4735 that integrate and improve on functionality found in modern operating
4736 systems. Broadly speaking, such features can be divided into three
4737 groups: supplemental functions (functions that may be missing in some
4738 environments and operating systems), replacement functions (providing
4739 a uniform and easier to use interface for commonly used standard
4740 functions), and extensions (which provide additional functionality
4741 beyond standard functions).
4743 @value{GDBN} uses various features provided by the @code{libiberty}
4744 library, for instance the C@t{++} demangler, the @acronym{IEEE}
4745 floating format support functions, the input options parser
4746 @samp{getopt}, the @samp{obstack} extension, and other functions.
4748 @subsection @code{obstacks} in @value{GDBN}
4749 @cindex @code{obstacks}
4751 The obstack mechanism provides a convenient way to allocate and free
4752 chunks of memory. Each obstack is a pool of memory that is managed
4753 like a stack. Objects (of any nature, size and alignment) are
4754 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
4755 @code{libiberty}'s documentation for a more detailed explanation of
4758 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
4759 object files. There is an obstack associated with each internal
4760 representation of an object file. Lots of things get allocated on
4761 these @code{obstacks}: dictionary entries, blocks, blockvectors,
4762 symbols, minimal symbols, types, vectors of fundamental types, class
4763 fields of types, object files section lists, object files section
4764 offset lists, line tables, symbol tables, partial symbol tables,
4765 string tables, symbol table private data, macros tables, debug
4766 information sections and entries, import and export lists (som),
4767 unwind information (hppa), dwarf2 location expressions data. Plus
4768 various strings such as directory names strings, debug format strings,
4771 An essential and convenient property of all data on @code{obstacks} is
4772 that memory for it gets allocated (with @code{obstack_alloc}) at
4773 various times during a debugging session, but it is released all at
4774 once using the @code{obstack_free} function. The @code{obstack_free}
4775 function takes a pointer to where in the stack it must start the
4776 deletion from (much like the cleanup chains have a pointer to where to
4777 start the cleanups). Because of the stack like structure of the
4778 @code{obstacks}, this allows to free only a top portion of the
4779 obstack. There are a few instances in @value{GDBN} where such thing
4780 happens. Calls to @code{obstack_free} are done after some local data
4781 is allocated to the obstack. Only the local data is deleted from the
4782 obstack. Of course this assumes that nothing between the
4783 @code{obstack_alloc} and the @code{obstack_free} allocates anything
4784 else on the same obstack. For this reason it is best and safest to
4785 use temporary @code{obstacks}.
4787 Releasing the whole obstack is also not safe per se. It is safe only
4788 under the condition that we know the @code{obstacks} memory is no
4789 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
4790 when we get rid of the whole objfile(s), for instance upon reading a
4794 @cindex regular expressions library
4805 @item SIGN_EXTEND_CHAR
4807 @item SWITCH_ENUM_BUG
4816 @section Array Containers
4817 @cindex Array Containers
4820 Often it is necessary to manipulate a dynamic array of a set of
4821 objects. C forces some bookkeeping on this, which can get cumbersome
4822 and repetitive. The @file{vec.h} file contains macros for defining
4823 and using a typesafe vector type. The functions defined will be
4824 inlined when compiling, and so the abstraction cost should be zero.
4825 Domain checks are added to detect programming errors.
4827 An example use would be an array of symbols or section information.
4828 The array can be grown as symbols are read in (or preallocated), and
4829 the accessor macros provided keep care of all the necessary
4830 bookkeeping. Because the arrays are type safe, there is no danger of
4831 accidentally mixing up the contents. Think of these as C++ templates,
4832 but implemented in C.
4834 Because of the different behavior of structure objects, scalar objects
4835 and of pointers, there are three flavors of vector, one for each of
4836 these variants. Both the structure object and pointer variants pass
4837 pointers to objects around --- in the former case the pointers are
4838 stored into the vector and in the latter case the pointers are
4839 dereferenced and the objects copied into the vector. The scalar
4840 object variant is suitable for @code{int}-like objects, and the vector
4841 elements are returned by value.
4843 There are both @code{index} and @code{iterate} accessors. The iterator
4844 returns a boolean iteration condition and updates the iteration
4845 variable passed by reference. Because the iterator will be inlined,
4846 the address-of can be optimized away.
4848 The vectors are implemented using the trailing array idiom, thus they
4849 are not resizeable without changing the address of the vector object
4850 itself. This means you cannot have variables or fields of vector type
4851 --- always use a pointer to a vector. The one exception is the final
4852 field of a structure, which could be a vector type. You will have to
4853 use the @code{embedded_size} & @code{embedded_init} calls to create
4854 such objects, and they will probably not be resizeable (so don't use
4855 the @dfn{safe} allocation variants). The trailing array idiom is used
4856 (rather than a pointer to an array of data), because, if we allow
4857 @code{NULL} to also represent an empty vector, empty vectors occupy
4858 minimal space in the structure containing them.
4860 Each operation that increases the number of active elements is
4861 available in @dfn{quick} and @dfn{safe} variants. The former presumes
4862 that there is sufficient allocated space for the operation to succeed
4863 (it dies if there is not). The latter will reallocate the vector, if
4864 needed. Reallocation causes an exponential increase in vector size.
4865 If you know you will be adding N elements, it would be more efficient
4866 to use the reserve operation before adding the elements with the
4867 @dfn{quick} operation. This will ensure there are at least as many
4868 elements as you ask for, it will exponentially increase if there are
4869 too few spare slots. If you want reserve a specific number of slots,
4870 but do not want the exponential increase (for instance, you know this
4871 is the last allocation), use a negative number for reservation. You
4872 can also create a vector of a specific size from the get go.
4874 You should prefer the push and pop operations, as they append and
4875 remove from the end of the vector. If you need to remove several items
4876 in one go, use the truncate operation. The insert and remove
4877 operations allow you to change elements in the middle of the vector.
4878 There are two remove operations, one which preserves the element
4879 ordering @code{ordered_remove}, and one which does not
4880 @code{unordered_remove}. The latter function copies the end element
4881 into the removed slot, rather than invoke a memmove operation. The
4882 @code{lower_bound} function will determine where to place an item in
4883 the array using insert that will maintain sorted order.
4885 If you need to directly manipulate a vector, then the @code{address}
4886 accessor will return the address of the start of the vector. Also the
4887 @code{space} predicate will tell you whether there is spare capacity in the
4888 vector. You will not normally need to use these two functions.
4890 Vector types are defined using a
4891 @code{DEF_VEC_@{O,P,I@}(@var{typename})} macro. Variables of vector
4892 type are declared using a @code{VEC(@var{typename})} macro. The
4893 characters @code{O}, @code{P} and @code{I} indicate whether
4894 @var{typename} is an object (@code{O}), pointer (@code{P}) or integral
4895 (@code{I}) type. Be careful to pick the correct one, as you'll get an
4896 awkward and inefficient API if you use the wrong one. There is a
4897 check, which results in a compile-time warning, for the @code{P} and
4898 @code{I} versions, but there is no check for the @code{O} versions, as
4899 that is not possible in plain C.
4901 An example of their use would be,
4904 DEF_VEC_P(tree); // non-managed tree vector.
4907 VEC(tree) *v; // A (pointer to) a vector of tree pointers.
4910 struct my_struct *s;
4912 if (VEC_length(tree, s->v)) @{ we have some contents @}
4913 VEC_safe_push(tree, s->v, decl); // append some decl onto the end
4914 for (ix = 0; VEC_iterate(tree, s->v, ix, elt); ix++)
4915 @{ do something with elt @}
4919 The @file{vec.h} file provides details on how to invoke the various
4920 accessors provided. They are enumerated here:
4924 Return the number of items in the array,
4927 Return true if the array has no elements.
4931 Return the last or arbitrary item in the array.
4934 Access an array element and indicate whether the array has been
4939 Create and destroy an array.
4941 @item VEC_embedded_size
4942 @itemx VEC_embedded_init
4943 Helpers for embedding an array as the final element of another struct.
4949 Return the amount of free space in an array.
4952 Ensure a certain amount of free space.
4954 @item VEC_quick_push
4955 @itemx VEC_safe_push
4956 Append to an array, either assuming the space is available, or making
4960 Remove the last item from an array.
4963 Remove several items from the end of an array.
4966 Add several items to the end of an array.
4969 Overwrite an item in the array.
4971 @item VEC_quick_insert
4972 @itemx VEC_safe_insert
4973 Insert an item into the middle of the array. Either the space must
4974 already exist, or the space is created.
4976 @item VEC_ordered_remove
4977 @itemx VEC_unordered_remove
4978 Remove an item from the array, preserving order or not.
4980 @item VEC_block_remove
4981 Remove a set of items from the array.
4984 Provide the address of the first element.
4986 @item VEC_lower_bound
4987 Binary search the array.
4997 This chapter covers topics that are lower-level than the major
4998 algorithms of @value{GDBN}.
5003 Cleanups are a structured way to deal with things that need to be done
5006 When your code does something (e.g., @code{xmalloc} some memory, or
5007 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
5008 the memory or @code{close} the file), it can make a cleanup. The
5009 cleanup will be done at some future point: when the command is finished
5010 and control returns to the top level; when an error occurs and the stack
5011 is unwound; or when your code decides it's time to explicitly perform
5012 cleanups. Alternatively you can elect to discard the cleanups you
5018 @item struct cleanup *@var{old_chain};
5019 Declare a variable which will hold a cleanup chain handle.
5021 @findex make_cleanup
5022 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
5023 Make a cleanup which will cause @var{function} to be called with
5024 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
5025 handle that can later be passed to @code{do_cleanups} or
5026 @code{discard_cleanups}. Unless you are going to call
5027 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
5028 from @code{make_cleanup}.
5031 @item do_cleanups (@var{old_chain});
5032 Do all cleanups added to the chain since the corresponding
5033 @code{make_cleanup} call was made.
5035 @findex discard_cleanups
5036 @item discard_cleanups (@var{old_chain});
5037 Same as @code{do_cleanups} except that it just removes the cleanups from
5038 the chain and does not call the specified functions.
5041 Cleanups are implemented as a chain. The handle returned by
5042 @code{make_cleanups} includes the cleanup passed to the call and any
5043 later cleanups appended to the chain (but not yet discarded or
5047 make_cleanup (a, 0);
5049 struct cleanup *old = make_cleanup (b, 0);
5057 will call @code{c()} and @code{b()} but will not call @code{a()}. The
5058 cleanup that calls @code{a()} will remain in the cleanup chain, and will
5059 be done later unless otherwise discarded.@refill
5061 Your function should explicitly do or discard the cleanups it creates.
5062 Failing to do this leads to non-deterministic behavior since the caller
5063 will arbitrarily do or discard your functions cleanups. This need leads
5064 to two common cleanup styles.
5066 The first style is try/finally. Before it exits, your code-block calls
5067 @code{do_cleanups} with the old cleanup chain and thus ensures that your
5068 code-block's cleanups are always performed. For instance, the following
5069 code-segment avoids a memory leak problem (even when @code{error} is
5070 called and a forced stack unwind occurs) by ensuring that the
5071 @code{xfree} will always be called:
5074 struct cleanup *old = make_cleanup (null_cleanup, 0);
5075 data = xmalloc (sizeof blah);
5076 make_cleanup (xfree, data);
5081 The second style is try/except. Before it exits, your code-block calls
5082 @code{discard_cleanups} with the old cleanup chain and thus ensures that
5083 any created cleanups are not performed. For instance, the following
5084 code segment, ensures that the file will be closed but only if there is
5088 FILE *file = fopen ("afile", "r");
5089 struct cleanup *old = make_cleanup (close_file, file);
5091 discard_cleanups (old);
5095 Some functions, e.g., @code{fputs_filtered()} or @code{error()}, specify
5096 that they ``should not be called when cleanups are not in place''. This
5097 means that any actions you need to reverse in the case of an error or
5098 interruption must be on the cleanup chain before you call these
5099 functions, since they might never return to your code (they
5100 @samp{longjmp} instead).
5102 @section Per-architecture module data
5103 @cindex per-architecture module data
5104 @cindex multi-arch data
5105 @cindex data-pointer, per-architecture/per-module
5107 The multi-arch framework includes a mechanism for adding module
5108 specific per-architecture data-pointers to the @code{struct gdbarch}
5109 architecture object.
5111 A module registers one or more per-architecture data-pointers using:
5113 @deftypefun struct gdbarch_data *gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
5114 @var{pre_init} is used to, on-demand, allocate an initial value for a
5115 per-architecture data-pointer using the architecture's obstack (passed
5116 in as a parameter). Since @var{pre_init} can be called during
5117 architecture creation, it is not parameterized with the architecture.
5118 and must not call modules that use per-architecture data.
5121 @deftypefun struct gdbarch_data *gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
5122 @var{post_init} is used to obtain an initial value for a
5123 per-architecture data-pointer @emph{after}. Since @var{post_init} is
5124 always called after architecture creation, it both receives the fully
5125 initialized architecture and is free to call modules that use
5126 per-architecture data (care needs to be taken to ensure that those
5127 other modules do not try to call back to this module as that will
5128 create in cycles in the initialization call graph).
5131 These functions return a @code{struct gdbarch_data} that is used to
5132 identify the per-architecture data-pointer added for that module.
5134 The per-architecture data-pointer is accessed using the function:
5136 @deftypefun void *gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
5137 Given the architecture @var{arch} and module data handle
5138 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
5139 or @code{gdbarch_data_register_post_init}), this function returns the
5140 current value of the per-architecture data-pointer. If the data
5141 pointer is @code{NULL}, it is first initialized by calling the
5142 corresponding @var{pre_init} or @var{post_init} method.
5145 The examples below assume the following definitions:
5148 struct nozel @{ int total; @};
5149 static struct gdbarch_data *nozel_handle;
5152 A module can extend the architecture vector, adding additional
5153 per-architecture data, using the @var{pre_init} method. The module's
5154 per-architecture data is then initialized during architecture
5157 In the below, the module's per-architecture @emph{nozel} is added. An
5158 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
5159 from @code{gdbarch_init}.
5163 nozel_pre_init (struct obstack *obstack)
5165 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
5172 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
5174 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5175 data->total = nozel;
5179 A module can on-demand create architecture dependant data structures
5180 using @code{post_init}.
5182 In the below, the nozel's total is computed on-demand by
5183 @code{nozel_post_init} using information obtained from the
5188 nozel_post_init (struct gdbarch *gdbarch)
5190 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
5191 nozel->total = gdbarch@dots{} (gdbarch);
5198 nozel_total (struct gdbarch *gdbarch)
5200 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5205 @section Wrapping Output Lines
5206 @cindex line wrap in output
5209 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
5210 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
5211 added in places that would be good breaking points. The utility
5212 routines will take care of actually wrapping if the line width is
5215 The argument to @code{wrap_here} is an indentation string which is
5216 printed @emph{only} if the line breaks there. This argument is saved
5217 away and used later. It must remain valid until the next call to
5218 @code{wrap_here} or until a newline has been printed through the
5219 @code{*_filtered} functions. Don't pass in a local variable and then
5222 It is usually best to call @code{wrap_here} after printing a comma or
5223 space. If you call it before printing a space, make sure that your
5224 indentation properly accounts for the leading space that will print if
5225 the line wraps there.
5227 Any function or set of functions that produce filtered output must
5228 finish by printing a newline, to flush the wrap buffer, before switching
5229 to unfiltered (@code{printf}) output. Symbol reading routines that
5230 print warnings are a good example.
5232 @section @value{GDBN} Coding Standards
5233 @cindex coding standards
5235 @value{GDBN} follows the GNU coding standards, as described in
5236 @file{etc/standards.texi}. This file is also available for anonymous
5237 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
5238 of the standard; in general, when the GNU standard recommends a practice
5239 but does not require it, @value{GDBN} requires it.
5241 @value{GDBN} follows an additional set of coding standards specific to
5242 @value{GDBN}, as described in the following sections.
5247 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
5250 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
5253 @subsection Memory Management
5255 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
5256 @code{calloc}, @code{free} and @code{asprintf}.
5258 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
5259 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
5260 these functions do not return when the memory pool is empty. Instead,
5261 they unwind the stack using cleanups. These functions return
5262 @code{NULL} when requested to allocate a chunk of memory of size zero.
5264 @emph{Pragmatics: By using these functions, the need to check every
5265 memory allocation is removed. These functions provide portable
5268 @value{GDBN} does not use the function @code{free}.
5270 @value{GDBN} uses the function @code{xfree} to return memory to the
5271 memory pool. Consistent with ISO-C, this function ignores a request to
5272 free a @code{NULL} pointer.
5274 @emph{Pragmatics: On some systems @code{free} fails when passed a
5275 @code{NULL} pointer.}
5277 @value{GDBN} can use the non-portable function @code{alloca} for the
5278 allocation of small temporary values (such as strings).
5280 @emph{Pragmatics: This function is very non-portable. Some systems
5281 restrict the memory being allocated to no more than a few kilobytes.}
5283 @value{GDBN} uses the string function @code{xstrdup} and the print
5284 function @code{xstrprintf}.
5286 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
5287 functions such as @code{sprintf} are very prone to buffer overflow
5291 @subsection Compiler Warnings
5292 @cindex compiler warnings
5294 With few exceptions, developers should avoid the configuration option
5295 @samp{--disable-werror} when building @value{GDBN}. The exceptions
5296 are listed in the file @file{gdb/MAINTAINERS}. The default, when
5297 building with @sc{gcc}, is @samp{--enable-werror}.
5299 This option causes @value{GDBN} (when built using GCC) to be compiled
5300 with a carefully selected list of compiler warning flags. Any warnings
5301 from those flags are treated as errors.
5303 The current list of warning flags includes:
5307 Recommended @sc{gcc} warnings.
5309 @item -Wdeclaration-after-statement
5311 @sc{gcc} 3.x (and later) and @sc{c99} allow declarations mixed with
5312 code, but @sc{gcc} 2.x and @sc{c89} do not.
5314 @item -Wpointer-arith
5316 @item -Wformat-nonliteral
5317 Non-literal format strings, with a few exceptions, are bugs - they
5318 might contain unintended user-supplied format specifiers.
5319 Since @value{GDBN} uses the @code{format printf} attribute on all
5320 @code{printf} like functions this checks not just @code{printf} calls
5321 but also calls to functions such as @code{fprintf_unfiltered}.
5323 @item -Wno-pointer-sign
5324 In version 4.0, GCC began warning about pointer argument passing or
5325 assignment even when the source and destination differed only in
5326 signedness. However, most @value{GDBN} code doesn't distinguish
5327 carefully between @code{char} and @code{unsigned char}. In early 2006
5328 the @value{GDBN} developers decided correcting these warnings wasn't
5329 worth the time it would take.
5331 @item -Wno-unused-parameter
5332 Due to the way that @value{GDBN} is implemented many functions have
5333 unused parameters. Consequently this warning is avoided. The macro
5334 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
5335 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
5340 @itemx -Wno-char-subscripts
5341 These are warnings which might be useful for @value{GDBN}, but are
5342 currently too noisy to enable with @samp{-Werror}.
5346 @subsection Formatting
5348 @cindex source code formatting
5349 The standard GNU recommendations for formatting must be followed
5352 A function declaration should not have its name in column zero. A
5353 function definition should have its name in column zero.
5357 static void foo (void);
5365 @emph{Pragmatics: This simplifies scripting. Function definitions can
5366 be found using @samp{^function-name}.}
5368 There must be a space between a function or macro name and the opening
5369 parenthesis of its argument list (except for macro definitions, as
5370 required by C). There must not be a space after an open paren/bracket
5371 or before a close paren/bracket.
5373 While additional whitespace is generally helpful for reading, do not use
5374 more than one blank line to separate blocks, and avoid adding whitespace
5375 after the end of a program line (as of 1/99, some 600 lines had
5376 whitespace after the semicolon). Excess whitespace causes difficulties
5377 for @code{diff} and @code{patch} utilities.
5379 Pointers are declared using the traditional K&R C style:
5393 @subsection Comments
5395 @cindex comment formatting
5396 The standard GNU requirements on comments must be followed strictly.
5398 Block comments must appear in the following form, with no @code{/*}- or
5399 @code{*/}-only lines, and no leading @code{*}:
5402 /* Wait for control to return from inferior to debugger. If inferior
5403 gets a signal, we may decide to start it up again instead of
5404 returning. That is why there is a loop in this function. When
5405 this function actually returns it means the inferior should be left
5406 stopped and @value{GDBN} should read more commands. */
5409 (Note that this format is encouraged by Emacs; tabbing for a multi-line
5410 comment works correctly, and @kbd{M-q} fills the block consistently.)
5412 Put a blank line between the block comments preceding function or
5413 variable definitions, and the definition itself.
5415 In general, put function-body comments on lines by themselves, rather
5416 than trying to fit them into the 20 characters left at the end of a
5417 line, since either the comment or the code will inevitably get longer
5418 than will fit, and then somebody will have to move it anyhow.
5422 @cindex C data types
5423 Code must not depend on the sizes of C data types, the format of the
5424 host's floating point numbers, the alignment of anything, or the order
5425 of evaluation of expressions.
5427 @cindex function usage
5428 Use functions freely. There are only a handful of compute-bound areas
5429 in @value{GDBN} that might be affected by the overhead of a function
5430 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
5431 limited by the target interface (whether serial line or system call).
5433 However, use functions with moderation. A thousand one-line functions
5434 are just as hard to understand as a single thousand-line function.
5436 @emph{Macros are bad, M'kay.}
5437 (But if you have to use a macro, make sure that the macro arguments are
5438 protected with parentheses.)
5442 Declarations like @samp{struct foo *} should be used in preference to
5443 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
5446 @subsection Function Prototypes
5447 @cindex function prototypes
5449 Prototypes must be used when both @emph{declaring} and @emph{defining}
5450 a function. Prototypes for @value{GDBN} functions must include both the
5451 argument type and name, with the name matching that used in the actual
5452 function definition.
5454 All external functions should have a declaration in a header file that
5455 callers include, except for @code{_initialize_*} functions, which must
5456 be external so that @file{init.c} construction works, but shouldn't be
5457 visible to random source files.
5459 Where a source file needs a forward declaration of a static function,
5460 that declaration must appear in a block near the top of the source file.
5463 @subsection Internal Error Recovery
5465 During its execution, @value{GDBN} can encounter two types of errors.
5466 User errors and internal errors. User errors include not only a user
5467 entering an incorrect command but also problems arising from corrupt
5468 object files and system errors when interacting with the target.
5469 Internal errors include situations where @value{GDBN} has detected, at
5470 run time, a corrupt or erroneous situation.
5472 When reporting an internal error, @value{GDBN} uses
5473 @code{internal_error} and @code{gdb_assert}.
5475 @value{GDBN} must not call @code{abort} or @code{assert}.
5477 @emph{Pragmatics: There is no @code{internal_warning} function. Either
5478 the code detected a user error, recovered from it and issued a
5479 @code{warning} or the code failed to correctly recover from the user
5480 error and issued an @code{internal_error}.}
5482 @subsection File Names
5484 Any file used when building the core of @value{GDBN} must be in lower
5485 case. Any file used when building the core of @value{GDBN} must be 8.3
5486 unique. These requirements apply to both source and generated files.
5488 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
5489 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
5490 is introduced to the build process both @file{Makefile.in} and
5491 @file{configure.in} need to be modified accordingly. Compare the
5492 convoluted conversion process needed to transform @file{COPYING} into
5493 @file{copying.c} with the conversion needed to transform
5494 @file{version.in} into @file{version.c}.}
5496 Any file non 8.3 compliant file (that is not used when building the core
5497 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
5499 @emph{Pragmatics: This is clearly a compromise.}
5501 When @value{GDBN} has a local version of a system header file (ex
5502 @file{string.h}) the file name based on the POSIX header prefixed with
5503 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
5504 independent: they should use only macros defined by @file{configure},
5505 the compiler, or the host; they should include only system headers; they
5506 should refer only to system types. They may be shared between multiple
5507 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
5509 For other files @samp{-} is used as the separator.
5512 @subsection Include Files
5514 A @file{.c} file should include @file{defs.h} first.
5516 A @file{.c} file should directly include the @code{.h} file of every
5517 declaration and/or definition it directly refers to. It cannot rely on
5520 A @file{.h} file should directly include the @code{.h} file of every
5521 declaration and/or definition it directly refers to. It cannot rely on
5522 indirect inclusion. Exception: The file @file{defs.h} does not need to
5523 be directly included.
5525 An external declaration should only appear in one include file.
5527 An external declaration should never appear in a @code{.c} file.
5528 Exception: a declaration for the @code{_initialize} function that
5529 pacifies @option{-Wmissing-declaration}.
5531 A @code{typedef} definition should only appear in one include file.
5533 An opaque @code{struct} declaration can appear in multiple @file{.h}
5534 files. Where possible, a @file{.h} file should use an opaque
5535 @code{struct} declaration instead of an include.
5537 All @file{.h} files should be wrapped in:
5540 #ifndef INCLUDE_FILE_NAME_H
5541 #define INCLUDE_FILE_NAME_H
5547 @subsection Clean Design and Portable Implementation
5550 In addition to getting the syntax right, there's the little question of
5551 semantics. Some things are done in certain ways in @value{GDBN} because long
5552 experience has shown that the more obvious ways caused various kinds of
5555 @cindex assumptions about targets
5556 You can't assume the byte order of anything that comes from a target
5557 (including @var{value}s, object files, and instructions). Such things
5558 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
5559 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
5560 such as @code{bfd_get_32}.
5562 You can't assume that you know what interface is being used to talk to
5563 the target system. All references to the target must go through the
5564 current @code{target_ops} vector.
5566 You can't assume that the host and target machines are the same machine
5567 (except in the ``native'' support modules). In particular, you can't
5568 assume that the target machine's header files will be available on the
5569 host machine. Target code must bring along its own header files --
5570 written from scratch or explicitly donated by their owner, to avoid
5574 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
5575 to write the code portably than to conditionalize it for various
5578 @cindex system dependencies
5579 New @code{#ifdef}'s which test for specific compilers or manufacturers
5580 or operating systems are unacceptable. All @code{#ifdef}'s should test
5581 for features. The information about which configurations contain which
5582 features should be segregated into the configuration files. Experience
5583 has proven far too often that a feature unique to one particular system
5584 often creeps into other systems; and that a conditional based on some
5585 predefined macro for your current system will become worthless over
5586 time, as new versions of your system come out that behave differently
5587 with regard to this feature.
5589 Adding code that handles specific architectures, operating systems,
5590 target interfaces, or hosts, is not acceptable in generic code.
5592 @cindex portable file name handling
5593 @cindex file names, portability
5594 One particularly notorious area where system dependencies tend to
5595 creep in is handling of file names. The mainline @value{GDBN} code
5596 assumes Posix semantics of file names: absolute file names begin with
5597 a forward slash @file{/}, slashes are used to separate leading
5598 directories, case-sensitive file names. These assumptions are not
5599 necessarily true on non-Posix systems such as MS-Windows. To avoid
5600 system-dependent code where you need to take apart or construct a file
5601 name, use the following portable macros:
5604 @findex HAVE_DOS_BASED_FILE_SYSTEM
5605 @item HAVE_DOS_BASED_FILE_SYSTEM
5606 This preprocessing symbol is defined to a non-zero value on hosts
5607 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
5608 symbol to write conditional code which should only be compiled for
5611 @findex IS_DIR_SEPARATOR
5612 @item IS_DIR_SEPARATOR (@var{c})
5613 Evaluates to a non-zero value if @var{c} is a directory separator
5614 character. On Unix and GNU/Linux systems, only a slash @file{/} is
5615 such a character, but on Windows, both @file{/} and @file{\} will
5618 @findex IS_ABSOLUTE_PATH
5619 @item IS_ABSOLUTE_PATH (@var{file})
5620 Evaluates to a non-zero value if @var{file} is an absolute file name.
5621 For Unix and GNU/Linux hosts, a name which begins with a slash
5622 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
5623 @file{x:\bar} are also absolute file names.
5625 @findex FILENAME_CMP
5626 @item FILENAME_CMP (@var{f1}, @var{f2})
5627 Calls a function which compares file names @var{f1} and @var{f2} as
5628 appropriate for the underlying host filesystem. For Posix systems,
5629 this simply calls @code{strcmp}; on case-insensitive filesystems it
5630 will call @code{strcasecmp} instead.
5632 @findex DIRNAME_SEPARATOR
5633 @item DIRNAME_SEPARATOR
5634 Evaluates to a character which separates directories in
5635 @code{PATH}-style lists, typically held in environment variables.
5636 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
5638 @findex SLASH_STRING
5640 This evaluates to a constant string you should use to produce an
5641 absolute filename from leading directories and the file's basename.
5642 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
5643 @code{"\\"} for some Windows-based ports.
5646 In addition to using these macros, be sure to use portable library
5647 functions whenever possible. For example, to extract a directory or a
5648 basename part from a file name, use the @code{dirname} and
5649 @code{basename} library functions (available in @code{libiberty} for
5650 platforms which don't provide them), instead of searching for a slash
5651 with @code{strrchr}.
5653 Another way to generalize @value{GDBN} along a particular interface is with an
5654 attribute struct. For example, @value{GDBN} has been generalized to handle
5655 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
5656 by defining the @code{target_ops} structure and having a current target (as
5657 well as a stack of targets below it, for memory references). Whenever
5658 something needs to be done that depends on which remote interface we are
5659 using, a flag in the current target_ops structure is tested (e.g.,
5660 @code{target_has_stack}), or a function is called through a pointer in the
5661 current target_ops structure. In this way, when a new remote interface
5662 is added, only one module needs to be touched---the one that actually
5663 implements the new remote interface. Other examples of
5664 attribute-structs are BFD access to multiple kinds of object file
5665 formats, or @value{GDBN}'s access to multiple source languages.
5667 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
5668 the code interfacing between @code{ptrace} and the rest of
5669 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
5670 something was very painful. In @value{GDBN} 4.x, these have all been
5671 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
5672 with variations between systems the same way any system-independent
5673 file would (hooks, @code{#if defined}, etc.), and machines which are
5674 radically different don't need to use @file{infptrace.c} at all.
5676 All debugging code must be controllable using the @samp{set debug
5677 @var{module}} command. Do not use @code{printf} to print trace
5678 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
5679 @code{#ifdef DEBUG}.
5684 @chapter Porting @value{GDBN}
5685 @cindex porting to new machines
5687 Most of the work in making @value{GDBN} compile on a new machine is in
5688 specifying the configuration of the machine. This is done in a
5689 dizzying variety of header files and configuration scripts, which we
5690 hope to make more sensible soon. Let's say your new host is called an
5691 @var{xyz} (e.g., @samp{sun4}), and its full three-part configuration
5692 name is @code{@var{arch}-@var{xvend}-@var{xos}} (e.g.,
5693 @samp{sparc-sun-sunos4}). In particular:
5697 In the top level directory, edit @file{config.sub} and add @var{arch},
5698 @var{xvend}, and @var{xos} to the lists of supported architectures,
5699 vendors, and operating systems near the bottom of the file. Also, add
5700 @var{xyz} as an alias that maps to
5701 @code{@var{arch}-@var{xvend}-@var{xos}}. You can test your changes by
5705 ./config.sub @var{xyz}
5712 ./config.sub @code{@var{arch}-@var{xvend}-@var{xos}}
5716 which should both respond with @code{@var{arch}-@var{xvend}-@var{xos}}
5717 and no error messages.
5720 You need to port BFD, if that hasn't been done already. Porting BFD is
5721 beyond the scope of this manual.
5724 To configure @value{GDBN} itself, edit @file{gdb/configure.host} to recognize
5725 your system and set @code{gdb_host} to @var{xyz}, and (unless your
5726 desired target is already available) also edit @file{gdb/configure.tgt},
5727 setting @code{gdb_target} to something appropriate (for instance,
5730 @emph{Maintainer's note: Work in progress. The file
5731 @file{gdb/configure.host} originally needed to be modified when either a
5732 new native target or a new host machine was being added to @value{GDBN}.
5733 Recent changes have removed this requirement. The file now only needs
5734 to be modified when adding a new native configuration. This will likely
5735 changed again in the future.}
5738 Finally, you'll need to specify and define @value{GDBN}'s host-, native-, and
5739 target-dependent @file{.h} and @file{.c} files used for your
5743 @node Versions and Branches
5744 @chapter Versions and Branches
5748 @value{GDBN}'s version is determined by the file
5749 @file{gdb/version.in} and takes one of the following forms:
5752 @item @var{major}.@var{minor}
5753 @itemx @var{major}.@var{minor}.@var{patchlevel}
5754 an official release (e.g., 6.2 or 6.2.1)
5755 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}
5756 a snapshot taken at @var{YYYY}-@var{MM}-@var{DD}-gmt (e.g.,
5757 6.1.50.20020302, 6.1.90.20020304, or 6.1.0.20020308)
5758 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}-cvs
5759 a @sc{cvs} check out drawn on @var{YYYY}-@var{MM}-@var{DD} (e.g.,
5760 6.1.50.20020302-cvs, 6.1.90.20020304-cvs, or 6.1.0.20020308-cvs)
5761 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD} (@var{vendor})
5762 a vendor specific release of @value{GDBN}, that while based on@*
5763 @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD},
5764 may include additional changes
5767 @value{GDBN}'s mainline uses the @var{major} and @var{minor} version
5768 numbers from the most recent release branch, with a @var{patchlevel}
5769 of 50. At the time each new release branch is created, the mainline's
5770 @var{major} and @var{minor} version numbers are updated.
5772 @value{GDBN}'s release branch is similar. When the branch is cut, the
5773 @var{patchlevel} is changed from 50 to 90. As draft releases are
5774 drawn from the branch, the @var{patchlevel} is incremented. Once the
5775 first release (@var{major}.@var{minor}) has been made, the
5776 @var{patchlevel} is set to 0 and updates have an incremented
5779 For snapshots, and @sc{cvs} check outs, it is also possible to
5780 identify the @sc{cvs} origin:
5783 @item @var{major}.@var{minor}.50.@var{YYYY}@var{MM}@var{DD}
5784 drawn from the @sc{head} of mainline @sc{cvs} (e.g., 6.1.50.20020302)
5785 @item @var{major}.@var{minor}.90.@var{YYYY}@var{MM}@var{DD}
5786 @itemx @var{major}.@var{minor}.91.@var{YYYY}@var{MM}@var{DD} @dots{}
5787 drawn from a release branch prior to the release (e.g.,
5789 @item @var{major}.@var{minor}.0.@var{YYYY}@var{MM}@var{DD}
5790 @itemx @var{major}.@var{minor}.1.@var{YYYY}@var{MM}@var{DD} @dots{}
5791 drawn from a release branch after the release (e.g., 6.2.0.20020308)
5794 If the previous @value{GDBN} version is 6.1 and the current version is
5795 6.2, then, substituting 6 for @var{major} and 1 or 2 for @var{minor},
5796 here's an illustration of a typical sequence:
5803 +--------------------------.
5806 6.2.50.20020303-cvs 6.1.90 (draft #1)
5808 6.2.50.20020304-cvs 6.1.90.20020304-cvs
5810 6.2.50.20020305-cvs 6.1.91 (draft #2)
5812 6.2.50.20020306-cvs 6.1.91.20020306-cvs
5814 6.2.50.20020307-cvs 6.2 (release)
5816 6.2.50.20020308-cvs 6.2.0.20020308-cvs
5818 6.2.50.20020309-cvs 6.2.1 (update)
5820 6.2.50.20020310-cvs <branch closed>
5824 +--------------------------.
5827 6.3.50.20020312-cvs 6.2.90 (draft #1)
5831 @section Release Branches
5832 @cindex Release Branches
5834 @value{GDBN} draws a release series (6.2, 6.2.1, @dots{}) from a
5835 single release branch, and identifies that branch using the @sc{cvs}
5839 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-branchpoint
5840 gdb_@var{major}_@var{minor}-branch
5841 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-release
5844 @emph{Pragmatics: To help identify the date at which a branch or
5845 release is made, both the branchpoint and release tags include the
5846 date that they are cut (@var{YYYY}@var{MM}@var{DD}) in the tag. The
5847 branch tag, denoting the head of the branch, does not need this.}
5849 @section Vendor Branches
5850 @cindex vendor branches
5852 To avoid version conflicts, vendors are expected to modify the file
5853 @file{gdb/version.in} to include a vendor unique alphabetic identifier
5854 (an official @value{GDBN} release never uses alphabetic characters in
5855 its version identifier). E.g., @samp{6.2widgit2}, or @samp{6.2 (Widgit
5858 @section Experimental Branches
5859 @cindex experimental branches
5861 @subsection Guidelines
5863 @value{GDBN} permits the creation of branches, cut from the @sc{cvs}
5864 repository, for experimental development. Branches make it possible
5865 for developers to share preliminary work, and maintainers to examine
5866 significant new developments.
5868 The following are a set of guidelines for creating such branches:
5872 @item a branch has an owner
5873 The owner can set further policy for a branch, but may not change the
5874 ground rules. In particular, they can set a policy for commits (be it
5875 adding more reviewers or deciding who can commit).
5877 @item all commits are posted
5878 All changes committed to a branch shall also be posted to
5879 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} patches
5880 mailing list}. While commentary on such changes are encouraged, people
5881 should remember that the changes only apply to a branch.
5883 @item all commits are covered by an assignment
5884 This ensures that all changes belong to the Free Software Foundation,
5885 and avoids the possibility that the branch may become contaminated.
5887 @item a branch is focused
5888 A focused branch has a single objective or goal, and does not contain
5889 unnecessary or irrelevant changes. Cleanups, where identified, being
5890 be pushed into the mainline as soon as possible.
5892 @item a branch tracks mainline
5893 This keeps the level of divergence under control. It also keeps the
5894 pressure on developers to push cleanups and other stuff into the
5897 @item a branch shall contain the entire @value{GDBN} module
5898 The @value{GDBN} module @code{gdb} should be specified when creating a
5899 branch (branches of individual files should be avoided). @xref{Tags}.
5901 @item a branch shall be branded using @file{version.in}
5902 The file @file{gdb/version.in} shall be modified so that it identifies
5903 the branch @var{owner} and branch @var{name}, e.g.,
5904 @samp{6.2.50.20030303_owner_name} or @samp{6.2 (Owner Name)}.
5911 To simplify the identification of @value{GDBN} branches, the following
5912 branch tagging convention is strongly recommended:
5916 @item @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5917 @itemx @var{owner}_@var{name}-@var{YYYYMMDD}-branch
5918 The branch point and corresponding branch tag. @var{YYYYMMDD} is the
5919 date that the branch was created. A branch is created using the
5920 sequence: @anchor{experimental branch tags}
5922 cvs rtag @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint gdb
5923 cvs rtag -b -r @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint \
5924 @var{owner}_@var{name}-@var{YYYYMMDD}-branch gdb
5927 @item @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5928 The tagged point, on the mainline, that was used when merging the branch
5929 on @var{yyyymmdd}. To merge in all changes since the branch was cut,
5930 use a command sequence like:
5932 cvs rtag @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint gdb
5934 -j@var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5935 -j@var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5938 Similar sequences can be used to just merge in changes since the last
5944 For further information on @sc{cvs}, see
5945 @uref{http://www.gnu.org/software/cvs/, Concurrent Versions System}.
5947 @node Start of New Year Procedure
5948 @chapter Start of New Year Procedure
5949 @cindex new year procedure
5951 At the start of each new year, the following actions should be performed:
5955 Rotate the ChangeLog file
5957 The current @file{ChangeLog} file should be renamed into
5958 @file{ChangeLog-YYYY} where YYYY is the year that has just passed.
5959 A new @file{ChangeLog} file should be created, and its contents should
5960 contain a reference to the previous ChangeLog. The following should
5961 also be preserved at the end of the new ChangeLog, in order to provide
5962 the appropriate settings when editing this file with Emacs:
5968 version-control: never
5973 Add an entry for the newly created ChangeLog file (@file{ChangeLog-YYYY})
5974 in @file{gdb/config/djgpp/fnchange.lst}.
5977 Update the copyright year in the startup message
5979 Update the copyright year in file @file{top.c}, function
5980 @code{print_gdb_version}.
5985 @chapter Releasing @value{GDBN}
5986 @cindex making a new release of gdb
5988 @section Branch Commit Policy
5990 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
5991 5.1 and 5.2 all used the below:
5995 The @file{gdb/MAINTAINERS} file still holds.
5997 Don't fix something on the branch unless/until it is also fixed in the
5998 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
5999 file is better than committing a hack.
6001 When considering a patch for the branch, suggested criteria include:
6002 Does it fix a build? Does it fix the sequence @kbd{break main; run}
6003 when debugging a static binary?
6005 The further a change is from the core of @value{GDBN}, the less likely
6006 the change will worry anyone (e.g., target specific code).
6008 Only post a proposal to change the core of @value{GDBN} after you've
6009 sent individual bribes to all the people listed in the
6010 @file{MAINTAINERS} file @t{;-)}
6013 @emph{Pragmatics: Provided updates are restricted to non-core
6014 functionality there is little chance that a broken change will be fatal.
6015 This means that changes such as adding a new architectures or (within
6016 reason) support for a new host are considered acceptable.}
6019 @section Obsoleting code
6021 Before anything else, poke the other developers (and around the source
6022 code) to see if there is anything that can be removed from @value{GDBN}
6023 (an old target, an unused file).
6025 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
6026 line. Doing this means that it is easy to identify something that has
6027 been obsoleted when greping through the sources.
6029 The process is done in stages --- this is mainly to ensure that the
6030 wider @value{GDBN} community has a reasonable opportunity to respond.
6031 Remember, everything on the Internet takes a week.
6035 Post the proposal on @email{gdb@@sources.redhat.com, the GDB mailing
6036 list} Creating a bug report to track the task's state, is also highly
6041 Post the proposal on @email{gdb-announce@@sources.redhat.com, the GDB
6042 Announcement mailing list}.
6046 Go through and edit all relevant files and lines so that they are
6047 prefixed with the word @code{OBSOLETE}.
6049 Wait until the next GDB version, containing this obsolete code, has been
6052 Remove the obsolete code.
6056 @emph{Maintainer note: While removing old code is regrettable it is
6057 hopefully better for @value{GDBN}'s long term development. Firstly it
6058 helps the developers by removing code that is either no longer relevant
6059 or simply wrong. Secondly since it removes any history associated with
6060 the file (effectively clearing the slate) the developer has a much freer
6061 hand when it comes to fixing broken files.}
6065 @section Before the Branch
6067 The most important objective at this stage is to find and fix simple
6068 changes that become a pain to track once the branch is created. For
6069 instance, configuration problems that stop @value{GDBN} from even
6070 building. If you can't get the problem fixed, document it in the
6071 @file{gdb/PROBLEMS} file.
6073 @subheading Prompt for @file{gdb/NEWS}
6075 People always forget. Send a post reminding them but also if you know
6076 something interesting happened add it yourself. The @code{schedule}
6077 script will mention this in its e-mail.
6079 @subheading Review @file{gdb/README}
6081 Grab one of the nightly snapshots and then walk through the
6082 @file{gdb/README} looking for anything that can be improved. The
6083 @code{schedule} script will mention this in its e-mail.
6085 @subheading Refresh any imported files.
6087 A number of files are taken from external repositories. They include:
6091 @file{texinfo/texinfo.tex}
6093 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
6096 @file{etc/standards.texi}, @file{etc/make-stds.texi}
6099 @subheading Check the ARI
6101 @uref{http://sources.redhat.com/gdb/ari,,A.R.I.} is an @code{awk} script
6102 (Awk Regression Index ;-) that checks for a number of errors and coding
6103 conventions. The checks include things like using @code{malloc} instead
6104 of @code{xmalloc} and file naming problems. There shouldn't be any
6107 @subsection Review the bug data base
6109 Close anything obviously fixed.
6111 @subsection Check all cross targets build
6113 The targets are listed in @file{gdb/MAINTAINERS}.
6116 @section Cut the Branch
6118 @subheading Create the branch
6123 $ V=`echo $v | sed 's/\./_/g'`
6124 $ D=`date -u +%Y-%m-%d`
6127 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6128 -D $D-gmt gdb_$V-$D-branchpoint insight
6129 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag
6130 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight
6133 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6134 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight
6135 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6136 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight
6144 By using @kbd{-D YYYY-MM-DD-gmt}, the branch is forced to an exact
6147 The trunk is first tagged so that the branch point can easily be found.
6149 Insight, which includes @value{GDBN}, is tagged at the same time.
6151 @file{version.in} gets bumped to avoid version number conflicts.
6153 The reading of @file{.cvsrc} is disabled using @file{-f}.
6156 @subheading Update @file{version.in}
6161 $ V=`echo $v | sed 's/\./_/g'`
6165 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src co \
6166 -r gdb_$V-branch src/gdb/version.in
6167 cvs -f -d :ext:sources.redhat.com:/cvs/src co
6168 -r gdb_5_2-branch src/gdb/version.in
6170 U src/gdb/version.in
6172 $ echo $u.90-0000-00-00-cvs > version.in
6174 5.1.90-0000-00-00-cvs
6175 $ cvs -f commit version.in
6180 @file{0000-00-00} is used as a date to pump prime the version.in update
6183 @file{.90} and the previous branch version are used as fairly arbitrary
6184 initial branch version number.
6188 @subheading Update the web and news pages
6192 @subheading Tweak cron to track the new branch
6194 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
6195 This file needs to be updated so that:
6199 A daily timestamp is added to the file @file{version.in}.
6201 The new branch is included in the snapshot process.
6205 See the file @file{gdbadmin/cron/README} for how to install the updated
6208 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
6209 any changes. That file is copied to both the branch/ and current/
6210 snapshot directories.
6213 @subheading Update the NEWS and README files
6215 The @file{NEWS} file needs to be updated so that on the branch it refers
6216 to @emph{changes in the current release} while on the trunk it also
6217 refers to @emph{changes since the current release}.
6219 The @file{README} file needs to be updated so that it refers to the
6222 @subheading Post the branch info
6224 Send an announcement to the mailing lists:
6228 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6230 @email{gdb@@sources.redhat.com, GDB Discussion mailing list} and
6231 @email{gdb-testers@@sources.redhat.com, GDB Testers mailing list}
6234 @emph{Pragmatics: The branch creation is sent to the announce list to
6235 ensure that people people not subscribed to the higher volume discussion
6238 The announcement should include:
6244 How to check out the branch using CVS.
6246 The date/number of weeks until the release.
6248 The branch commit policy still holds.
6251 @section Stabilize the branch
6253 Something goes here.
6255 @section Create a Release
6257 The process of creating and then making available a release is broken
6258 down into a number of stages. The first part addresses the technical
6259 process of creating a releasable tar ball. The later stages address the
6260 process of releasing that tar ball.
6262 When making a release candidate just the first section is needed.
6264 @subsection Create a release candidate
6266 The objective at this stage is to create a set of tar balls that can be
6267 made available as a formal release (or as a less formal release
6270 @subsubheading Freeze the branch
6272 Send out an e-mail notifying everyone that the branch is frozen to
6273 @email{gdb-patches@@sources.redhat.com}.
6275 @subsubheading Establish a few defaults.
6280 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
6282 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
6286 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
6288 /home/gdbadmin/bin/autoconf
6297 Check the @code{autoconf} version carefully. You want to be using the
6298 version taken from the @file{binutils} snapshot directory, which can be
6299 found at @uref{ftp://sources.redhat.com/pub/binutils/}. It is very
6300 unlikely that a system installed version of @code{autoconf} (e.g.,
6301 @file{/usr/bin/autoconf}) is correct.
6304 @subsubheading Check out the relevant modules:
6307 $ for m in gdb insight
6309 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
6319 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
6320 any confusion between what is written here and what your local
6321 @code{cvs} really does.
6324 @subsubheading Update relevant files.
6330 Major releases get their comments added as part of the mainline. Minor
6331 releases should probably mention any significant bugs that were fixed.
6333 Don't forget to include the @file{ChangeLog} entry.
6336 $ emacs gdb/src/gdb/NEWS
6341 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
6342 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6347 You'll need to update:
6359 $ emacs gdb/src/gdb/README
6364 $ cp gdb/src/gdb/README insight/src/gdb/README
6365 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6368 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
6369 before the initial branch was cut so just a simple substitute is needed
6372 @emph{Maintainer note: Other projects generate @file{README} and
6373 @file{INSTALL} from the core documentation. This might be worth
6376 @item gdb/version.in
6379 $ echo $v > gdb/src/gdb/version.in
6380 $ cat gdb/src/gdb/version.in
6382 $ emacs gdb/src/gdb/version.in
6385 ... Bump to version ...
6387 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
6388 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6393 @subsubheading Do the dirty work
6395 This is identical to the process used to create the daily snapshot.
6398 $ for m in gdb insight
6400 ( cd $m/src && gmake -f src-release $m.tar )
6404 If the top level source directory does not have @file{src-release}
6405 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
6408 $ for m in gdb insight
6410 ( cd $m/src && gmake -f Makefile.in $m.tar )
6414 @subsubheading Check the source files
6416 You're looking for files that have mysteriously disappeared.
6417 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
6418 for the @file{version.in} update @kbd{cronjob}.
6421 $ ( cd gdb/src && cvs -f -q -n update )
6425 @dots{} lots of generated files @dots{}
6430 @dots{} lots of generated files @dots{}
6435 @emph{Don't worry about the @file{gdb.info-??} or
6436 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
6437 was also generated only something strange with CVS means that they
6438 didn't get suppressed). Fixing it would be nice though.}
6440 @subsubheading Create compressed versions of the release
6446 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
6447 $ for m in gdb insight
6449 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
6450 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
6460 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
6461 in that mode, @code{gzip} does not know the name of the file and, hence,
6462 can not include it in the compressed file. This is also why the release
6463 process runs @code{tar} and @code{bzip2} as separate passes.
6466 @subsection Sanity check the tar ball
6468 Pick a popular machine (Solaris/PPC?) and try the build on that.
6471 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
6476 $ ./gdb/gdb ./gdb/gdb
6480 Breakpoint 1 at 0x80732bc: file main.c, line 734.
6482 Starting program: /tmp/gdb-5.2/gdb/gdb
6484 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
6485 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
6487 $1 = @{argc = 136426532, argv = 0x821b7f0@}
6491 @subsection Make a release candidate available
6493 If this is a release candidate then the only remaining steps are:
6497 Commit @file{version.in} and @file{ChangeLog}
6499 Tweak @file{version.in} (and @file{ChangeLog} to read
6500 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
6501 process can restart.
6503 Make the release candidate available in
6504 @uref{ftp://sources.redhat.com/pub/gdb/snapshots/branch}
6506 Notify the relevant mailing lists ( @email{gdb@@sources.redhat.com} and
6507 @email{gdb-testers@@sources.redhat.com} that the candidate is available.
6510 @subsection Make a formal release available
6512 (And you thought all that was required was to post an e-mail.)
6514 @subsubheading Install on sware
6516 Copy the new files to both the release and the old release directory:
6519 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
6520 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
6524 Clean up the releases directory so that only the most recent releases
6525 are available (e.g. keep 5.2 and 5.2.1 but remove 5.1):
6528 $ cd ~ftp/pub/gdb/releases
6533 Update the file @file{README} and @file{.message} in the releases
6540 $ ln README .message
6543 @subsubheading Update the web pages.
6547 @item htdocs/download/ANNOUNCEMENT
6548 This file, which is posted as the official announcement, includes:
6551 General announcement.
6553 News. If making an @var{M}.@var{N}.1 release, retain the news from
6554 earlier @var{M}.@var{N} release.
6559 @item htdocs/index.html
6560 @itemx htdocs/news/index.html
6561 @itemx htdocs/download/index.html
6562 These files include:
6565 Announcement of the most recent release.
6567 News entry (remember to update both the top level and the news directory).
6569 These pages also need to be regenerate using @code{index.sh}.
6571 @item download/onlinedocs/
6572 You need to find the magic command that is used to generate the online
6573 docs from the @file{.tar.bz2}. The best way is to look in the output
6574 from one of the nightly @code{cron} jobs and then just edit accordingly.
6578 $ ~/ss/update-web-docs \
6579 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6581 /www/sourceware/htdocs/gdb/download/onlinedocs \
6586 Just like the online documentation. Something like:
6589 $ /bin/sh ~/ss/update-web-ari \
6590 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6592 /www/sourceware/htdocs/gdb/download/ari \
6598 @subsubheading Shadow the pages onto gnu
6600 Something goes here.
6603 @subsubheading Install the @value{GDBN} tar ball on GNU
6605 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
6606 @file{~ftp/gnu/gdb}.
6608 @subsubheading Make the @file{ANNOUNCEMENT}
6610 Post the @file{ANNOUNCEMENT} file you created above to:
6614 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6616 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
6617 day or so to let things get out)
6619 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
6624 The release is out but you're still not finished.
6626 @subsubheading Commit outstanding changes
6628 In particular you'll need to commit any changes to:
6632 @file{gdb/ChangeLog}
6634 @file{gdb/version.in}
6641 @subsubheading Tag the release
6646 $ d=`date -u +%Y-%m-%d`
6649 $ ( cd insight/src/gdb && cvs -f -q update )
6650 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
6653 Insight is used since that contains more of the release than
6656 @subsubheading Mention the release on the trunk
6658 Just put something in the @file{ChangeLog} so that the trunk also
6659 indicates when the release was made.
6661 @subsubheading Restart @file{gdb/version.in}
6663 If @file{gdb/version.in} does not contain an ISO date such as
6664 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
6665 committed all the release changes it can be set to
6666 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
6667 is important - it affects the snapshot process).
6669 Don't forget the @file{ChangeLog}.
6671 @subsubheading Merge into trunk
6673 The files committed to the branch may also need changes merged into the
6676 @subsubheading Revise the release schedule
6678 Post a revised release schedule to @email{gdb@@sources.redhat.com, GDB
6679 Discussion List} with an updated announcement. The schedule can be
6680 generated by running:
6683 $ ~/ss/schedule `date +%s` schedule
6687 The first parameter is approximate date/time in seconds (from the epoch)
6688 of the most recent release.
6690 Also update the schedule @code{cronjob}.
6692 @section Post release
6694 Remove any @code{OBSOLETE} code.
6701 The testsuite is an important component of the @value{GDBN} package.
6702 While it is always worthwhile to encourage user testing, in practice
6703 this is rarely sufficient; users typically use only a small subset of
6704 the available commands, and it has proven all too common for a change
6705 to cause a significant regression that went unnoticed for some time.
6707 The @value{GDBN} testsuite uses the DejaGNU testing framework. The
6708 tests themselves are calls to various @code{Tcl} procs; the framework
6709 runs all the procs and summarizes the passes and fails.
6711 @section Using the Testsuite
6713 @cindex running the test suite
6714 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
6715 testsuite's objdir) and type @code{make check}. This just sets up some
6716 environment variables and invokes DejaGNU's @code{runtest} script. While
6717 the testsuite is running, you'll get mentions of which test file is in use,
6718 and a mention of any unexpected passes or fails. When the testsuite is
6719 finished, you'll get a summary that looks like this:
6724 # of expected passes 6016
6725 # of unexpected failures 58
6726 # of unexpected successes 5
6727 # of expected failures 183
6728 # of unresolved testcases 3
6729 # of untested testcases 5
6732 To run a specific test script, type:
6734 make check RUNTESTFLAGS='@var{tests}'
6736 where @var{tests} is a list of test script file names, separated by
6739 The ideal test run consists of expected passes only; however, reality
6740 conspires to keep us from this ideal. Unexpected failures indicate
6741 real problems, whether in @value{GDBN} or in the testsuite. Expected
6742 failures are still failures, but ones which have been decided are too
6743 hard to deal with at the time; for instance, a test case might work
6744 everywhere except on AIX, and there is no prospect of the AIX case
6745 being fixed in the near future. Expected failures should not be added
6746 lightly, since you may be masking serious bugs in @value{GDBN}.
6747 Unexpected successes are expected fails that are passing for some
6748 reason, while unresolved and untested cases often indicate some minor
6749 catastrophe, such as the compiler being unable to deal with a test
6752 When making any significant change to @value{GDBN}, you should run the
6753 testsuite before and after the change, to confirm that there are no
6754 regressions. Note that truly complete testing would require that you
6755 run the testsuite with all supported configurations and a variety of
6756 compilers; however this is more than really necessary. In many cases
6757 testing with a single configuration is sufficient. Other useful
6758 options are to test one big-endian (Sparc) and one little-endian (x86)
6759 host, a cross config with a builtin simulator (powerpc-eabi,
6760 mips-elf), or a 64-bit host (Alpha).
6762 If you add new functionality to @value{GDBN}, please consider adding
6763 tests for it as well; this way future @value{GDBN} hackers can detect
6764 and fix their changes that break the functionality you added.
6765 Similarly, if you fix a bug that was not previously reported as a test
6766 failure, please add a test case for it. Some cases are extremely
6767 difficult to test, such as code that handles host OS failures or bugs
6768 in particular versions of compilers, and it's OK not to try to write
6769 tests for all of those.
6771 DejaGNU supports separate build, host, and target machines. However,
6772 some @value{GDBN} test scripts do not work if the build machine and
6773 the host machine are not the same. In such an environment, these scripts
6774 will give a result of ``UNRESOLVED'', like this:
6777 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
6780 @section Testsuite Organization
6782 @cindex test suite organization
6783 The testsuite is entirely contained in @file{gdb/testsuite}. While the
6784 testsuite includes some makefiles and configury, these are very minimal,
6785 and used for little besides cleaning up, since the tests themselves
6786 handle the compilation of the programs that @value{GDBN} will run. The file
6787 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
6788 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
6789 configuration-specific files, typically used for special-purpose
6790 definitions of procs like @code{gdb_load} and @code{gdb_start}.
6792 The tests themselves are to be found in @file{testsuite/gdb.*} and
6793 subdirectories of those. The names of the test files must always end
6794 with @file{.exp}. DejaGNU collects the test files by wildcarding
6795 in the test directories, so both subdirectories and individual files
6796 get chosen and run in alphabetical order.
6798 The following table lists the main types of subdirectories and what they
6799 are for. Since DejaGNU finds test files no matter where they are
6800 located, and since each test file sets up its own compilation and
6801 execution environment, this organization is simply for convenience and
6806 This is the base testsuite. The tests in it should apply to all
6807 configurations of @value{GDBN} (but generic native-only tests may live here).
6808 The test programs should be in the subset of C that is valid K&R,
6809 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
6812 @item gdb.@var{lang}
6813 Language-specific tests for any language @var{lang} besides C. Examples are
6814 @file{gdb.cp} and @file{gdb.java}.
6816 @item gdb.@var{platform}
6817 Non-portable tests. The tests are specific to a specific configuration
6818 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
6821 @item gdb.@var{compiler}
6822 Tests specific to a particular compiler. As of this writing (June
6823 1999), there aren't currently any groups of tests in this category that
6824 couldn't just as sensibly be made platform-specific, but one could
6825 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
6828 @item gdb.@var{subsystem}
6829 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
6830 instance, @file{gdb.disasm} exercises various disassemblers, while
6831 @file{gdb.stabs} tests pathways through the stabs symbol reader.
6834 @section Writing Tests
6835 @cindex writing tests
6837 In many areas, the @value{GDBN} tests are already quite comprehensive; you
6838 should be able to copy existing tests to handle new cases.
6840 You should try to use @code{gdb_test} whenever possible, since it
6841 includes cases to handle all the unexpected errors that might happen.
6842 However, it doesn't cost anything to add new test procedures; for
6843 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
6844 calls @code{gdb_test} multiple times.
6846 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
6847 necessary. Even if @value{GDBN} has several valid responses to
6848 a command, you can use @code{gdb_test_multiple}. Like @code{gdb_test},
6849 @code{gdb_test_multiple} recognizes internal errors and unexpected
6852 Do not write tests which expect a literal tab character from @value{GDBN}.
6853 On some operating systems (e.g.@: OpenBSD) the TTY layer expands tabs to
6854 spaces, so by the time @value{GDBN}'s output reaches expect the tab is gone.
6856 The source language programs do @emph{not} need to be in a consistent
6857 style. Since @value{GDBN} is used to debug programs written in many different
6858 styles, it's worth having a mix of styles in the testsuite; for
6859 instance, some @value{GDBN} bugs involving the display of source lines would
6860 never manifest themselves if the programs used GNU coding style
6867 Check the @file{README} file, it often has useful information that does not
6868 appear anywhere else in the directory.
6871 * Getting Started:: Getting started working on @value{GDBN}
6872 * Debugging GDB:: Debugging @value{GDBN} with itself
6875 @node Getting Started,,, Hints
6877 @section Getting Started
6879 @value{GDBN} is a large and complicated program, and if you first starting to
6880 work on it, it can be hard to know where to start. Fortunately, if you
6881 know how to go about it, there are ways to figure out what is going on.
6883 This manual, the @value{GDBN} Internals manual, has information which applies
6884 generally to many parts of @value{GDBN}.
6886 Information about particular functions or data structures are located in
6887 comments with those functions or data structures. If you run across a
6888 function or a global variable which does not have a comment correctly
6889 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
6890 free to submit a bug report, with a suggested comment if you can figure
6891 out what the comment should say. If you find a comment which is
6892 actually wrong, be especially sure to report that.
6894 Comments explaining the function of macros defined in host, target, or
6895 native dependent files can be in several places. Sometimes they are
6896 repeated every place the macro is defined. Sometimes they are where the
6897 macro is used. Sometimes there is a header file which supplies a
6898 default definition of the macro, and the comment is there. This manual
6899 also documents all the available macros.
6900 @c (@pxref{Host Conditionals}, @pxref{Target
6901 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
6904 Start with the header files. Once you have some idea of how
6905 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
6906 @file{gdbtypes.h}), you will find it much easier to understand the
6907 code which uses and creates those symbol tables.
6909 You may wish to process the information you are getting somehow, to
6910 enhance your understanding of it. Summarize it, translate it to another
6911 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
6912 the code to predict what a test case would do and write the test case
6913 and verify your prediction, etc. If you are reading code and your eyes
6914 are starting to glaze over, this is a sign you need to use a more active
6917 Once you have a part of @value{GDBN} to start with, you can find more
6918 specifically the part you are looking for by stepping through each
6919 function with the @code{next} command. Do not use @code{step} or you
6920 will quickly get distracted; when the function you are stepping through
6921 calls another function try only to get a big-picture understanding
6922 (perhaps using the comment at the beginning of the function being
6923 called) of what it does. This way you can identify which of the
6924 functions being called by the function you are stepping through is the
6925 one which you are interested in. You may need to examine the data
6926 structures generated at each stage, with reference to the comments in
6927 the header files explaining what the data structures are supposed to
6930 Of course, this same technique can be used if you are just reading the
6931 code, rather than actually stepping through it. The same general
6932 principle applies---when the code you are looking at calls something
6933 else, just try to understand generally what the code being called does,
6934 rather than worrying about all its details.
6936 @cindex command implementation
6937 A good place to start when tracking down some particular area is with
6938 a command which invokes that feature. Suppose you want to know how
6939 single-stepping works. As a @value{GDBN} user, you know that the
6940 @code{step} command invokes single-stepping. The command is invoked
6941 via command tables (see @file{command.h}); by convention the function
6942 which actually performs the command is formed by taking the name of
6943 the command and adding @samp{_command}, or in the case of an
6944 @code{info} subcommand, @samp{_info}. For example, the @code{step}
6945 command invokes the @code{step_command} function and the @code{info
6946 display} command invokes @code{display_info}. When this convention is
6947 not followed, you might have to use @code{grep} or @kbd{M-x
6948 tags-search} in emacs, or run @value{GDBN} on itself and set a
6949 breakpoint in @code{execute_command}.
6951 @cindex @code{bug-gdb} mailing list
6952 If all of the above fail, it may be appropriate to ask for information
6953 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
6954 wondering if anyone could give me some tips about understanding
6955 @value{GDBN}''---if we had some magic secret we would put it in this manual.
6956 Suggestions for improving the manual are always welcome, of course.
6958 @node Debugging GDB,,,Hints
6960 @section Debugging @value{GDBN} with itself
6961 @cindex debugging @value{GDBN}
6963 If @value{GDBN} is limping on your machine, this is the preferred way to get it
6964 fully functional. Be warned that in some ancient Unix systems, like
6965 Ultrix 4.2, a program can't be running in one process while it is being
6966 debugged in another. Rather than typing the command @kbd{@w{./gdb
6967 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
6968 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
6970 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
6971 @file{.gdbinit} file that sets up some simple things to make debugging
6972 gdb easier. The @code{info} command, when executed without a subcommand
6973 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
6974 gdb. See @file{.gdbinit} for details.
6976 If you use emacs, you will probably want to do a @code{make TAGS} after
6977 you configure your distribution; this will put the machine dependent
6978 routines for your local machine where they will be accessed first by
6981 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
6982 have run @code{fixincludes} if you are compiling with gcc.
6984 @section Submitting Patches
6986 @cindex submitting patches
6987 Thanks for thinking of offering your changes back to the community of
6988 @value{GDBN} users. In general we like to get well designed enhancements.
6989 Thanks also for checking in advance about the best way to transfer the
6992 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
6993 This manual summarizes what we believe to be clean design for @value{GDBN}.
6995 If the maintainers don't have time to put the patch in when it arrives,
6996 or if there is any question about a patch, it goes into a large queue
6997 with everyone else's patches and bug reports.
6999 @cindex legal papers for code contributions
7000 The legal issue is that to incorporate substantial changes requires a
7001 copyright assignment from you and/or your employer, granting ownership
7002 of the changes to the Free Software Foundation. You can get the
7003 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
7004 and asking for it. We recommend that people write in "All programs
7005 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
7006 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
7008 contributed with only one piece of legalese pushed through the
7009 bureaucracy and filed with the FSF. We can't start merging changes until
7010 this paperwork is received by the FSF (their rules, which we follow
7011 since we maintain it for them).
7013 Technically, the easiest way to receive changes is to receive each
7014 feature as a small context diff or unidiff, suitable for @code{patch}.
7015 Each message sent to me should include the changes to C code and
7016 header files for a single feature, plus @file{ChangeLog} entries for
7017 each directory where files were modified, and diffs for any changes
7018 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
7019 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
7020 single feature, they can be split down into multiple messages.
7022 In this way, if we read and like the feature, we can add it to the
7023 sources with a single patch command, do some testing, and check it in.
7024 If you leave out the @file{ChangeLog}, we have to write one. If you leave
7025 out the doc, we have to puzzle out what needs documenting. Etc., etc.
7027 The reason to send each change in a separate message is that we will not
7028 install some of the changes. They'll be returned to you with questions
7029 or comments. If we're doing our job correctly, the message back to you
7030 will say what you have to fix in order to make the change acceptable.
7031 The reason to have separate messages for separate features is so that
7032 the acceptable changes can be installed while one or more changes are
7033 being reworked. If multiple features are sent in a single message, we
7034 tend to not put in the effort to sort out the acceptable changes from
7035 the unacceptable, so none of the features get installed until all are
7038 If this sounds painful or authoritarian, well, it is. But we get a lot
7039 of bug reports and a lot of patches, and many of them don't get
7040 installed because we don't have the time to finish the job that the bug
7041 reporter or the contributor could have done. Patches that arrive
7042 complete, working, and well designed, tend to get installed on the day
7043 they arrive. The others go into a queue and get installed as time
7044 permits, which, since the maintainers have many demands to meet, may not
7045 be for quite some time.
7047 Please send patches directly to
7048 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} maintainers}.
7050 @section Obsolete Conditionals
7051 @cindex obsolete code
7053 Fragments of old code in @value{GDBN} sometimes reference or set the following
7054 configuration macros. They should not be used by new code, and old uses
7055 should be removed as those parts of the debugger are otherwise touched.
7058 @item STACK_END_ADDR
7059 This macro used to define where the end of the stack appeared, for use
7060 in interpreting core file formats that don't record this address in the
7061 core file itself. This information is now configured in BFD, and @value{GDBN}
7062 gets the info portably from there. The values in @value{GDBN}'s configuration
7063 files should be moved into BFD configuration files (if needed there),
7064 and deleted from all of @value{GDBN}'s config files.
7066 Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR
7067 is so old that it has never been converted to use BFD. Now that's old!
7071 @include observer.texi