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: 1.279 $} % 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 It first uses @code{STOPPED_BY_WATCHPOINT} to see if any watchpoint
664 was hit. If not, all watchpoint checking is skipped.
666 Then @value{GDBN} calls @code{target_stopped_data_address} exactly
667 once. This method returns the address of the watchpoint which
668 triggered, if the target can determine it. If the triggered address
669 is available, @value{GDBN} compares the address returned by this
670 method with each watched memory address in each active watchpoint.
671 For data-read and data-access watchpoints, @value{GDBN} announces
672 every watchpoint that watches the triggered address as being hit.
673 For this reason, data-read and data-access watchpoints
674 @emph{require} that the triggered address be available; if not, read
675 and access watchpoints will never be considered hit. For data-write
676 watchpoints, if the triggered address is available, @value{GDBN}
677 considers only those watchpoints which match that address;
678 otherwise, @value{GDBN} considers all data-write watchpoints. For
679 each data-write watchpoint that @value{GDBN} considers, it evaluates
680 the expression whose value is being watched, and tests whether the
681 watched value has changed. Watchpoints whose watched values have
682 changed are announced as hit.
684 @value{GDBN} uses several macros and primitives to support hardware
688 @findex TARGET_HAS_HARDWARE_WATCHPOINTS
689 @item TARGET_HAS_HARDWARE_WATCHPOINTS
690 If defined, the target supports hardware watchpoints.
692 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
693 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
694 Return the number of hardware watchpoints of type @var{type} that are
695 possible to be set. The value is positive if @var{count} watchpoints
696 of this type can be set, zero if setting watchpoints of this type is
697 not supported, and negative if @var{count} is more than the maximum
698 number of watchpoints of type @var{type} that can be set. @var{other}
699 is non-zero if other types of watchpoints are currently enabled (there
700 are architectures which cannot set watchpoints of different types at
703 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
704 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
705 Return non-zero if hardware watchpoints can be used to watch a region
706 whose address is @var{addr} and whose length in bytes is @var{len}.
708 @cindex insert or remove hardware watchpoint
709 @findex target_insert_watchpoint
710 @findex target_remove_watchpoint
711 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
712 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
713 Insert or remove a hardware watchpoint starting at @var{addr}, for
714 @var{len} bytes. @var{type} is the watchpoint type, one of the
715 possible values of the enumerated data type @code{target_hw_bp_type},
716 defined by @file{breakpoint.h} as follows:
719 enum target_hw_bp_type
721 hw_write = 0, /* Common (write) HW watchpoint */
722 hw_read = 1, /* Read HW watchpoint */
723 hw_access = 2, /* Access (read or write) HW watchpoint */
724 hw_execute = 3 /* Execute HW breakpoint */
729 These two macros should return 0 for success, non-zero for failure.
731 @findex target_stopped_data_address
732 @item target_stopped_data_address (@var{addr_p})
733 If the inferior has some watchpoint that triggered, place the address
734 associated with the watchpoint at the location pointed to by
735 @var{addr_p} and return non-zero. Otherwise, return zero. This
736 is required for data-read and data-access watchpoints. It is
737 not required for data-write watchpoints, but @value{GDBN} uses
738 it to improve handling of those also.
740 @value{GDBN} will only call this method once per watchpoint stop,
741 immediately after calling @code{STOPPED_BY_WATCHPOINT}. If the
742 target's watchpoint indication is sticky, i.e., stays set after
743 resuming, this method should clear it. For instance, the x86 debug
744 control register has sticky triggered flags.
746 @findex HAVE_STEPPABLE_WATCHPOINT
747 @item HAVE_STEPPABLE_WATCHPOINT
748 If defined to a non-zero value, it is not necessary to disable a
749 watchpoint to step over it. Like @code{gdbarch_have_nonsteppable_watchpoint},
750 this is usually set when watchpoints trigger at the instruction
751 which will perform an interesting read or write. It should be
752 set if there is a temporary disable bit which allows the processor
753 to step over the interesting instruction without raising the
754 watchpoint exception again.
756 @findex gdbarch_have_nonsteppable_watchpoint
757 @item int gdbarch_have_nonsteppable_watchpoint (@var{gdbarch})
758 If it returns a non-zero value, @value{GDBN} should disable a
759 watchpoint to step the inferior over it. This is usually set when
760 watchpoints trigger at the instruction which will perform an
761 interesting read or write.
763 @findex HAVE_CONTINUABLE_WATCHPOINT
764 @item HAVE_CONTINUABLE_WATCHPOINT
765 If defined to a non-zero value, it is possible to continue the
766 inferior after a watchpoint has been hit. This is usually set
767 when watchpoints trigger at the instruction following an interesting
770 @findex CANNOT_STEP_HW_WATCHPOINTS
771 @item CANNOT_STEP_HW_WATCHPOINTS
772 If this is defined to a non-zero value, @value{GDBN} will remove all
773 watchpoints before stepping the inferior.
775 @findex STOPPED_BY_WATCHPOINT
776 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
777 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
778 the type @code{struct target_waitstatus}, defined by @file{target.h}.
779 Normally, this macro is defined to invoke the function pointed to by
780 the @code{to_stopped_by_watchpoint} member of the structure (of the
781 type @code{target_ops}, defined on @file{target.h}) that describes the
782 target-specific operations; @code{to_stopped_by_watchpoint} ignores
783 the @var{wait_status} argument.
785 @value{GDBN} does not require the non-zero value returned by
786 @code{STOPPED_BY_WATCHPOINT} to be 100% correct, so if a target cannot
787 determine for sure whether the inferior stopped due to a watchpoint,
788 it could return non-zero ``just in case''.
791 @subsection Watchpoints and Threads
792 @cindex watchpoints, with threads
794 @value{GDBN} only supports process-wide watchpoints, which trigger
795 in all threads. @value{GDBN} uses the thread ID to make watchpoints
796 act as if they were thread-specific, but it cannot set hardware
797 watchpoints that only trigger in a specific thread. Therefore, even
798 if the target supports threads, per-thread debug registers, and
799 watchpoints which only affect a single thread, it should set the
800 per-thread debug registers for all threads to the same value. On
801 @sc{gnu}/Linux native targets, this is accomplished by using
802 @code{ALL_LWPS} in @code{target_insert_watchpoint} and
803 @code{target_remove_watchpoint} and by using
804 @code{linux_set_new_thread} to register a handler for newly created
807 @value{GDBN}'s @sc{gnu}/Linux support only reports a single event
808 at a time, although multiple events can trigger simultaneously for
809 multi-threaded programs. When multiple events occur, @file{linux-nat.c}
810 queues subsequent events and returns them the next time the program
811 is resumed. This means that @code{STOPPED_BY_WATCHPOINT} and
812 @code{target_stopped_data_address} only need to consult the current
813 thread's state---the thread indicated by @code{inferior_ptid}. If
814 two threads have hit watchpoints simultaneously, those routines
815 will be called a second time for the second thread.
817 @subsection x86 Watchpoints
818 @cindex x86 debug registers
819 @cindex watchpoints, on x86
821 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
822 registers designed to facilitate debugging. @value{GDBN} provides a
823 generic library of functions that x86-based ports can use to implement
824 support for watchpoints and hardware-assisted breakpoints. This
825 subsection documents the x86 watchpoint facilities in @value{GDBN}.
827 To use the generic x86 watchpoint support, a port should do the
831 @findex I386_USE_GENERIC_WATCHPOINTS
833 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
834 target-dependent headers.
837 Include the @file{config/i386/nm-i386.h} header file @emph{after}
838 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
841 Add @file{i386-nat.o} to the value of the Make variable
842 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}) or
843 @code{TDEPFILES} (@pxref{Target Architecture Definition, TDEPFILES}).
846 Provide implementations for the @code{I386_DR_LOW_*} macros described
847 below. Typically, each macro should call a target-specific function
848 which does the real work.
851 The x86 watchpoint support works by maintaining mirror images of the
852 debug registers. Values are copied between the mirror images and the
853 real debug registers via a set of macros which each target needs to
857 @findex I386_DR_LOW_SET_CONTROL
858 @item I386_DR_LOW_SET_CONTROL (@var{val})
859 Set the Debug Control (DR7) register to the value @var{val}.
861 @findex I386_DR_LOW_SET_ADDR
862 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
863 Put the address @var{addr} into the debug register number @var{idx}.
865 @findex I386_DR_LOW_RESET_ADDR
866 @item I386_DR_LOW_RESET_ADDR (@var{idx})
867 Reset (i.e.@: zero out) the address stored in the debug register
870 @findex I386_DR_LOW_GET_STATUS
871 @item I386_DR_LOW_GET_STATUS
872 Return the value of the Debug Status (DR6) register. This value is
873 used immediately after it is returned by
874 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
878 For each one of the 4 debug registers (whose indices are from 0 to 3)
879 that store addresses, a reference count is maintained by @value{GDBN},
880 to allow sharing of debug registers by several watchpoints. This
881 allows users to define several watchpoints that watch the same
882 expression, but with different conditions and/or commands, without
883 wasting debug registers which are in short supply. @value{GDBN}
884 maintains the reference counts internally, targets don't have to do
885 anything to use this feature.
887 The x86 debug registers can each watch a region that is 1, 2, or 4
888 bytes long. The ia32 architecture requires that each watched region
889 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
890 region on 4-byte boundary. However, the x86 watchpoint support in
891 @value{GDBN} can watch unaligned regions and regions larger than 4
892 bytes (up to 16 bytes) by allocating several debug registers to watch
893 a single region. This allocation of several registers per a watched
894 region is also done automatically without target code intervention.
896 The generic x86 watchpoint support provides the following API for the
897 @value{GDBN}'s application code:
900 @findex i386_region_ok_for_watchpoint
901 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
902 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
903 this function. It counts the number of debug registers required to
904 watch a given region, and returns a non-zero value if that number is
905 less than 4, the number of debug registers available to x86
908 @findex i386_stopped_data_address
909 @item i386_stopped_data_address (@var{addr_p})
911 @code{target_stopped_data_address} is set to call this function.
913 function examines the breakpoint condition bits in the DR6 Debug
914 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
915 macro, and returns the address associated with the first bit that is
918 @findex i386_stopped_by_watchpoint
919 @item i386_stopped_by_watchpoint (void)
920 The macro @code{STOPPED_BY_WATCHPOINT}
921 is set to call this function. The
922 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
923 function examines the breakpoint condition bits in the DR6 Debug
924 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
925 macro, and returns true if any bit is set. Otherwise, false is
928 @findex i386_insert_watchpoint
929 @findex i386_remove_watchpoint
930 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
931 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
932 Insert or remove a watchpoint. The macros
933 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
934 are set to call these functions. @code{i386_insert_watchpoint} first
935 looks for a debug register which is already set to watch the same
936 region for the same access types; if found, it just increments the
937 reference count of that debug register, thus implementing debug
938 register sharing between watchpoints. If no such register is found,
939 the function looks for a vacant debug register, sets its mirrored
940 value to @var{addr}, sets the mirrored value of DR7 Debug Control
941 register as appropriate for the @var{len} and @var{type} parameters,
942 and then passes the new values of the debug register and DR7 to the
943 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
944 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
945 required to cover the given region, the above process is repeated for
948 @code{i386_remove_watchpoint} does the opposite: it resets the address
949 in the mirrored value of the debug register and its read/write and
950 length bits in the mirrored value of DR7, then passes these new
951 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
952 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
953 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
954 decrements the reference count, and only calls
955 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
956 the count goes to zero.
958 @findex i386_insert_hw_breakpoint
959 @findex i386_remove_hw_breakpoint
960 @item i386_insert_hw_breakpoint (@var{bp_tgt})
961 @itemx i386_remove_hw_breakpoint (@var{bp_tgt})
962 These functions insert and remove hardware-assisted breakpoints. The
963 macros @code{target_insert_hw_breakpoint} and
964 @code{target_remove_hw_breakpoint} are set to call these functions.
965 The argument is a @code{struct bp_target_info *}, as described in
966 the documentation for @code{target_insert_breakpoint}.
967 These functions work like @code{i386_insert_watchpoint} and
968 @code{i386_remove_watchpoint}, respectively, except that they set up
969 the debug registers to watch instruction execution, and each
970 hardware-assisted breakpoint always requires exactly one debug
973 @findex i386_stopped_by_hwbp
974 @item i386_stopped_by_hwbp (void)
975 This function returns non-zero if the inferior has some watchpoint or
976 hardware breakpoint that triggered. It works like
977 @code{i386_stopped_data_address}, except that it doesn't record the
978 address whose watchpoint triggered.
980 @findex i386_cleanup_dregs
981 @item i386_cleanup_dregs (void)
982 This function clears all the reference counts, addresses, and control
983 bits in the mirror images of the debug registers. It doesn't affect
984 the actual debug registers in the inferior process.
991 x86 processors support setting watchpoints on I/O reads or writes.
992 However, since no target supports this (as of March 2001), and since
993 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
994 watchpoints, this feature is not yet available to @value{GDBN} running
998 x86 processors can enable watchpoints locally, for the current task
999 only, or globally, for all the tasks. For each debug register,
1000 there's a bit in the DR7 Debug Control register that determines
1001 whether the associated address is watched locally or globally. The
1002 current implementation of x86 watchpoint support in @value{GDBN}
1003 always sets watchpoints to be locally enabled, since global
1004 watchpoints might interfere with the underlying OS and are probably
1005 unavailable in many platforms.
1008 @section Checkpoints
1011 In the abstract, a checkpoint is a point in the execution history of
1012 the program, which the user may wish to return to at some later time.
1014 Internally, a checkpoint is a saved copy of the program state, including
1015 whatever information is required in order to restore the program to that
1016 state at a later time. This can be expected to include the state of
1017 registers and memory, and may include external state such as the state
1018 of open files and devices.
1020 There are a number of ways in which checkpoints may be implemented
1021 in gdb, e.g.@: as corefiles, as forked processes, and as some opaque
1022 method implemented on the target side.
1024 A corefile can be used to save an image of target memory and register
1025 state, which can in principle be restored later --- but corefiles do
1026 not typically include information about external entities such as
1027 open files. Currently this method is not implemented in gdb.
1029 A forked process can save the state of user memory and registers,
1030 as well as some subset of external (kernel) state. This method
1031 is used to implement checkpoints on Linux, and in principle might
1032 be used on other systems.
1034 Some targets, e.g.@: simulators, might have their own built-in
1035 method for saving checkpoints, and gdb might be able to take
1036 advantage of that capability without necessarily knowing any
1037 details of how it is done.
1040 @section Observing changes in @value{GDBN} internals
1041 @cindex observer pattern interface
1042 @cindex notifications about changes in internals
1044 In order to function properly, several modules need to be notified when
1045 some changes occur in the @value{GDBN} internals. Traditionally, these
1046 modules have relied on several paradigms, the most common ones being
1047 hooks and gdb-events. Unfortunately, none of these paradigms was
1048 versatile enough to become the standard notification mechanism in
1049 @value{GDBN}. The fact that they only supported one ``client'' was also
1050 a strong limitation.
1052 A new paradigm, based on the Observer pattern of the @cite{Design
1053 Patterns} book, has therefore been implemented. The goal was to provide
1054 a new interface overcoming the issues with the notification mechanisms
1055 previously available. This new interface needed to be strongly typed,
1056 easy to extend, and versatile enough to be used as the standard
1057 interface when adding new notifications.
1059 See @ref{GDB Observers} for a brief description of the observers
1060 currently implemented in GDB. The rationale for the current
1061 implementation is also briefly discussed.
1063 @node User Interface
1065 @chapter User Interface
1067 @value{GDBN} has several user interfaces. Although the command-line interface
1068 is the most common and most familiar, there are others.
1070 @section Command Interpreter
1072 @cindex command interpreter
1074 The command interpreter in @value{GDBN} is fairly simple. It is designed to
1075 allow for the set of commands to be augmented dynamically, and also
1076 has a recursive subcommand capability, where the first argument to
1077 a command may itself direct a lookup on a different command list.
1079 For instance, the @samp{set} command just starts a lookup on the
1080 @code{setlist} command list, while @samp{set thread} recurses
1081 to the @code{set_thread_cmd_list}.
1085 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
1086 the main command list, and should be used for those commands. The usual
1087 place to add commands is in the @code{_initialize_@var{xyz}} routines at
1088 the ends of most source files.
1090 @findex add_setshow_cmd
1091 @findex add_setshow_cmd_full
1092 To add paired @samp{set} and @samp{show} commands, use
1093 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
1094 a slightly simpler interface which is useful when you don't need to
1095 further modify the new command structures, while the latter returns
1096 the new command structures for manipulation.
1098 @cindex deprecating commands
1099 @findex deprecate_cmd
1100 Before removing commands from the command set it is a good idea to
1101 deprecate them for some time. Use @code{deprecate_cmd} on commands or
1102 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
1103 @code{struct cmd_list_element} as it's first argument. You can use the
1104 return value from @code{add_com} or @code{add_cmd} to deprecate the
1105 command immediately after it is created.
1107 The first time a command is used the user will be warned and offered a
1108 replacement (if one exists). Note that the replacement string passed to
1109 @code{deprecate_cmd} should be the full name of the command, i.e., the
1110 entire string the user should type at the command line.
1112 @section UI-Independent Output---the @code{ui_out} Functions
1113 @c This section is based on the documentation written by Fernando
1114 @c Nasser <fnasser@redhat.com>.
1116 @cindex @code{ui_out} functions
1117 The @code{ui_out} functions present an abstraction level for the
1118 @value{GDBN} output code. They hide the specifics of different user
1119 interfaces supported by @value{GDBN}, and thus free the programmer
1120 from the need to write several versions of the same code, one each for
1121 every UI, to produce output.
1123 @subsection Overview and Terminology
1125 In general, execution of each @value{GDBN} command produces some sort
1126 of output, and can even generate an input request.
1128 Output can be generated for the following purposes:
1132 to display a @emph{result} of an operation;
1135 to convey @emph{info} or produce side-effects of a requested
1139 to provide a @emph{notification} of an asynchronous event (including
1140 progress indication of a prolonged asynchronous operation);
1143 to display @emph{error messages} (including warnings);
1146 to show @emph{debug data};
1149 to @emph{query} or prompt a user for input (a special case).
1153 This section mainly concentrates on how to build result output,
1154 although some of it also applies to other kinds of output.
1156 Generation of output that displays the results of an operation
1157 involves one or more of the following:
1161 output of the actual data
1164 formatting the output as appropriate for console output, to make it
1165 easily readable by humans
1168 machine oriented formatting--a more terse formatting to allow for easy
1169 parsing by programs which read @value{GDBN}'s output
1172 annotation, whose purpose is to help legacy GUIs to identify interesting
1176 The @code{ui_out} routines take care of the first three aspects.
1177 Annotations are provided by separate annotation routines. Note that use
1178 of annotations for an interface between a GUI and @value{GDBN} is
1181 Output can be in the form of a single item, which we call a @dfn{field};
1182 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
1183 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
1184 header and a body. In a BNF-like form:
1187 @item <table> @expansion{}
1188 @code{<header> <body>}
1189 @item <header> @expansion{}
1190 @code{@{ <column> @}}
1191 @item <column> @expansion{}
1192 @code{<width> <alignment> <title>}
1193 @item <body> @expansion{}
1198 @subsection General Conventions
1200 Most @code{ui_out} routines are of type @code{void}, the exceptions are
1201 @code{ui_out_stream_new} (which returns a pointer to the newly created
1202 object) and the @code{make_cleanup} routines.
1204 The first parameter is always the @code{ui_out} vector object, a pointer
1205 to a @code{struct ui_out}.
1207 The @var{format} parameter is like in @code{printf} family of functions.
1208 When it is present, there must also be a variable list of arguments
1209 sufficient used to satisfy the @code{%} specifiers in the supplied
1212 When a character string argument is not used in a @code{ui_out} function
1213 call, a @code{NULL} pointer has to be supplied instead.
1216 @subsection Table, Tuple and List Functions
1218 @cindex list output functions
1219 @cindex table output functions
1220 @cindex tuple output functions
1221 This section introduces @code{ui_out} routines for building lists,
1222 tuples and tables. The routines to output the actual data items
1223 (fields) are presented in the next section.
1225 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
1226 containing information about an object; a @dfn{list} is a sequence of
1227 fields where each field describes an identical object.
1229 Use the @dfn{table} functions when your output consists of a list of
1230 rows (tuples) and the console output should include a heading. Use this
1231 even when you are listing just one object but you still want the header.
1233 @cindex nesting level in @code{ui_out} functions
1234 Tables can not be nested. Tuples and lists can be nested up to a
1235 maximum of five levels.
1237 The overall structure of the table output code is something like this:
1252 Here is the description of table-, tuple- and list-related @code{ui_out}
1255 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
1256 The function @code{ui_out_table_begin} marks the beginning of the output
1257 of a table. It should always be called before any other @code{ui_out}
1258 function for a given table. @var{nbrofcols} is the number of columns in
1259 the table. @var{nr_rows} is the number of rows in the table.
1260 @var{tblid} is an optional string identifying the table. The string
1261 pointed to by @var{tblid} is copied by the implementation of
1262 @code{ui_out_table_begin}, so the application can free the string if it
1263 was @code{malloc}ed.
1265 The companion function @code{ui_out_table_end}, described below, marks
1266 the end of the table's output.
1269 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
1270 @code{ui_out_table_header} provides the header information for a single
1271 table column. You call this function several times, one each for every
1272 column of the table, after @code{ui_out_table_begin}, but before
1273 @code{ui_out_table_body}.
1275 The value of @var{width} gives the column width in characters. The
1276 value of @var{alignment} is one of @code{left}, @code{center}, and
1277 @code{right}, and it specifies how to align the header: left-justify,
1278 center, or right-justify it. @var{colhdr} points to a string that
1279 specifies the column header; the implementation copies that string, so
1280 column header strings in @code{malloc}ed storage can be freed after the
1284 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
1285 This function delimits the table header from the table body.
1288 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
1289 This function signals the end of a table's output. It should be called
1290 after the table body has been produced by the list and field output
1293 There should be exactly one call to @code{ui_out_table_end} for each
1294 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
1295 will signal an internal error.
1298 The output of the tuples that represent the table rows must follow the
1299 call to @code{ui_out_table_body} and precede the call to
1300 @code{ui_out_table_end}. You build a tuple by calling
1301 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
1302 calls to functions which actually output fields between them.
1304 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
1305 This function marks the beginning of a tuple output. @var{id} points
1306 to an optional string that identifies the tuple; it is copied by the
1307 implementation, and so strings in @code{malloc}ed storage can be freed
1311 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
1312 This function signals an end of a tuple output. There should be exactly
1313 one call to @code{ui_out_tuple_end} for each call to
1314 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
1318 @deftypefun struct cleanup *make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1319 This function first opens the tuple and then establishes a cleanup
1320 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
1321 and correct implementation of the non-portable@footnote{The function
1322 cast is not portable ISO C.} code sequence:
1324 struct cleanup *old_cleanup;
1325 ui_out_tuple_begin (uiout, "...");
1326 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
1331 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
1332 This function marks the beginning of a list output. @var{id} points to
1333 an optional string that identifies the list; it is copied by the
1334 implementation, and so strings in @code{malloc}ed storage can be freed
1338 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
1339 This function signals an end of a list output. There should be exactly
1340 one call to @code{ui_out_list_end} for each call to
1341 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1345 @deftypefun struct cleanup *make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1346 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1347 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
1348 that will close the list.
1351 @subsection Item Output Functions
1353 @cindex item output functions
1354 @cindex field output functions
1356 The functions described below produce output for the actual data
1357 items, or fields, which contain information about the object.
1359 Choose the appropriate function accordingly to your particular needs.
1361 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1362 This is the most general output function. It produces the
1363 representation of the data in the variable-length argument list
1364 according to formatting specifications in @var{format}, a
1365 @code{printf}-like format string. The optional argument @var{fldname}
1366 supplies the name of the field. The data items themselves are
1367 supplied as additional arguments after @var{format}.
1369 This generic function should be used only when it is not possible to
1370 use one of the specialized versions (see below).
1373 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1374 This function outputs a value of an @code{int} variable. It uses the
1375 @code{"%d"} output conversion specification. @var{fldname} specifies
1376 the name of the field.
1379 @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})
1380 This function outputs a value of an @code{int} variable. It differs from
1381 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1382 @var{fldname} specifies
1383 the name of the field.
1386 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1387 This function outputs an address.
1390 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1391 This function outputs a string using the @code{"%s"} conversion
1395 Sometimes, there's a need to compose your output piece by piece using
1396 functions that operate on a stream, such as @code{value_print} or
1397 @code{fprintf_symbol_filtered}. These functions accept an argument of
1398 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1399 used to store the data stream used for the output. When you use one
1400 of these functions, you need a way to pass their results stored in a
1401 @code{ui_file} object to the @code{ui_out} functions. To this end,
1402 you first create a @code{ui_stream} object by calling
1403 @code{ui_out_stream_new}, pass the @code{stream} member of that
1404 @code{ui_stream} object to @code{value_print} and similar functions,
1405 and finally call @code{ui_out_field_stream} to output the field you
1406 constructed. When the @code{ui_stream} object is no longer needed,
1407 you should destroy it and free its memory by calling
1408 @code{ui_out_stream_delete}.
1410 @deftypefun struct ui_stream *ui_out_stream_new (struct ui_out *@var{uiout})
1411 This function creates a new @code{ui_stream} object which uses the
1412 same output methods as the @code{ui_out} object whose pointer is
1413 passed in @var{uiout}. It returns a pointer to the newly created
1414 @code{ui_stream} object.
1417 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1418 This functions destroys a @code{ui_stream} object specified by
1422 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1423 This function consumes all the data accumulated in
1424 @code{streambuf->stream} and outputs it like
1425 @code{ui_out_field_string} does. After a call to
1426 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1427 the stream is still valid and may be used for producing more fields.
1430 @strong{Important:} If there is any chance that your code could bail
1431 out before completing output generation and reaching the point where
1432 @code{ui_out_stream_delete} is called, it is necessary to set up a
1433 cleanup, to avoid leaking memory and other resources. Here's a
1434 skeleton code to do that:
1437 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1438 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1443 If the function already has the old cleanup chain set (for other kinds
1444 of cleanups), you just have to add your cleanup to it:
1447 mybuf = ui_out_stream_new (uiout);
1448 make_cleanup (ui_out_stream_delete, mybuf);
1451 Note that with cleanups in place, you should not call
1452 @code{ui_out_stream_delete} directly, or you would attempt to free the
1455 @subsection Utility Output Functions
1457 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1458 This function skips a field in a table. Use it if you have to leave
1459 an empty field without disrupting the table alignment. The argument
1460 @var{fldname} specifies a name for the (missing) filed.
1463 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1464 This function outputs the text in @var{string} in a way that makes it
1465 easy to be read by humans. For example, the console implementation of
1466 this method filters the text through a built-in pager, to prevent it
1467 from scrolling off the visible portion of the screen.
1469 Use this function for printing relatively long chunks of text around
1470 the actual field data: the text it produces is not aligned according
1471 to the table's format. Use @code{ui_out_field_string} to output a
1472 string field, and use @code{ui_out_message}, described below, to
1473 output short messages.
1476 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1477 This function outputs @var{nspaces} spaces. It is handy to align the
1478 text produced by @code{ui_out_text} with the rest of the table or
1482 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1483 This function produces a formatted message, provided that the current
1484 verbosity level is at least as large as given by @var{verbosity}. The
1485 current verbosity level is specified by the user with the @samp{set
1486 verbositylevel} command.@footnote{As of this writing (April 2001),
1487 setting verbosity level is not yet implemented, and is always returned
1488 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1489 argument more than zero will cause the message to never be printed.}
1492 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1493 This function gives the console output filter (a paging filter) a hint
1494 of where to break lines which are too long. Ignored for all other
1495 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1496 be printed to indent the wrapped text on the next line; it must remain
1497 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1498 explicit newline is produced by one of the other functions. If
1499 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1502 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1503 This function flushes whatever output has been accumulated so far, if
1504 the UI buffers output.
1508 @subsection Examples of Use of @code{ui_out} functions
1510 @cindex using @code{ui_out} functions
1511 @cindex @code{ui_out} functions, usage examples
1512 This section gives some practical examples of using the @code{ui_out}
1513 functions to generalize the old console-oriented code in
1514 @value{GDBN}. The examples all come from functions defined on the
1515 @file{breakpoints.c} file.
1517 This example, from the @code{breakpoint_1} function, shows how to
1520 The original code was:
1523 if (!found_a_breakpoint++)
1525 annotate_breakpoints_headers ();
1528 printf_filtered ("Num ");
1530 printf_filtered ("Type ");
1532 printf_filtered ("Disp ");
1534 printf_filtered ("Enb ");
1538 printf_filtered ("Address ");
1541 printf_filtered ("What\n");
1543 annotate_breakpoints_table ();
1547 Here's the new version:
1550 nr_printable_breakpoints = @dots{};
1553 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1555 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1557 if (nr_printable_breakpoints > 0)
1558 annotate_breakpoints_headers ();
1559 if (nr_printable_breakpoints > 0)
1561 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1562 if (nr_printable_breakpoints > 0)
1564 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1565 if (nr_printable_breakpoints > 0)
1567 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1568 if (nr_printable_breakpoints > 0)
1570 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1573 if (nr_printable_breakpoints > 0)
1575 if (gdbarch_addr_bit (current_gdbarch) <= 32)
1576 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1578 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1580 if (nr_printable_breakpoints > 0)
1582 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1583 ui_out_table_body (uiout);
1584 if (nr_printable_breakpoints > 0)
1585 annotate_breakpoints_table ();
1588 This example, from the @code{print_one_breakpoint} function, shows how
1589 to produce the actual data for the table whose structure was defined
1590 in the above example. The original code was:
1595 printf_filtered ("%-3d ", b->number);
1597 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1598 || ((int) b->type != bptypes[(int) b->type].type))
1599 internal_error ("bptypes table does not describe type #%d.",
1601 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1603 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1605 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1609 This is the new version:
1613 ui_out_tuple_begin (uiout, "bkpt");
1615 ui_out_field_int (uiout, "number", b->number);
1617 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1618 || ((int) b->type != bptypes[(int) b->type].type))
1619 internal_error ("bptypes table does not describe type #%d.",
1621 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1623 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1625 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1629 This example, also from @code{print_one_breakpoint}, shows how to
1630 produce a complicated output field using the @code{print_expression}
1631 functions which requires a stream to be passed. It also shows how to
1632 automate stream destruction with cleanups. The original code was:
1636 print_expression (b->exp, gdb_stdout);
1642 struct ui_stream *stb = ui_out_stream_new (uiout);
1643 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1646 print_expression (b->exp, stb->stream);
1647 ui_out_field_stream (uiout, "what", local_stream);
1650 This example, also from @code{print_one_breakpoint}, shows how to use
1651 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1656 if (b->dll_pathname == NULL)
1657 printf_filtered ("<any library> ");
1659 printf_filtered ("library \"%s\" ", b->dll_pathname);
1666 if (b->dll_pathname == NULL)
1668 ui_out_field_string (uiout, "what", "<any library>");
1669 ui_out_spaces (uiout, 1);
1673 ui_out_text (uiout, "library \"");
1674 ui_out_field_string (uiout, "what", b->dll_pathname);
1675 ui_out_text (uiout, "\" ");
1679 The following example from @code{print_one_breakpoint} shows how to
1680 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1685 if (b->forked_inferior_pid != 0)
1686 printf_filtered ("process %d ", b->forked_inferior_pid);
1693 if (b->forked_inferior_pid != 0)
1695 ui_out_text (uiout, "process ");
1696 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1697 ui_out_spaces (uiout, 1);
1701 Here's an example of using @code{ui_out_field_string}. The original
1706 if (b->exec_pathname != NULL)
1707 printf_filtered ("program \"%s\" ", b->exec_pathname);
1714 if (b->exec_pathname != NULL)
1716 ui_out_text (uiout, "program \"");
1717 ui_out_field_string (uiout, "what", b->exec_pathname);
1718 ui_out_text (uiout, "\" ");
1722 Finally, here's an example of printing an address. The original code:
1726 printf_filtered ("%s ",
1727 hex_string_custom ((unsigned long) b->address, 8));
1734 ui_out_field_core_addr (uiout, "Address", b->address);
1738 @section Console Printing
1747 @cindex @code{libgdb}
1748 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1749 to provide an API to @value{GDBN}'s functionality.
1752 @cindex @code{libgdb}
1753 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1754 better able to support graphical and other environments.
1756 Since @code{libgdb} development is on-going, its architecture is still
1757 evolving. The following components have so far been identified:
1761 Observer - @file{gdb-events.h}.
1763 Builder - @file{ui-out.h}
1765 Event Loop - @file{event-loop.h}
1767 Library - @file{gdb.h}
1770 The model that ties these components together is described below.
1772 @section The @code{libgdb} Model
1774 A client of @code{libgdb} interacts with the library in two ways.
1778 As an observer (using @file{gdb-events}) receiving notifications from
1779 @code{libgdb} of any internal state changes (break point changes, run
1782 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1783 obtain various status values from @value{GDBN}.
1786 Since @code{libgdb} could have multiple clients (e.g., a GUI supporting
1787 the existing @value{GDBN} CLI), those clients must co-operate when
1788 controlling @code{libgdb}. In particular, a client must ensure that
1789 @code{libgdb} is idle (i.e. no other client is using @code{libgdb})
1790 before responding to a @file{gdb-event} by making a query.
1792 @section CLI support
1794 At present @value{GDBN}'s CLI is very much entangled in with the core of
1795 @code{libgdb}. Consequently, a client wishing to include the CLI in
1796 their interface needs to carefully co-ordinate its own and the CLI's
1799 It is suggested that the client set @code{libgdb} up to be bi-modal
1800 (alternate between CLI and client query modes). The notes below sketch
1805 The client registers itself as an observer of @code{libgdb}.
1807 The client create and install @code{cli-out} builder using its own
1808 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1809 @code{gdb_stdout} streams.
1811 The client creates a separate custom @code{ui-out} builder that is only
1812 used while making direct queries to @code{libgdb}.
1815 When the client receives input intended for the CLI, it simply passes it
1816 along. Since the @code{cli-out} builder is installed by default, all
1817 the CLI output in response to that command is routed (pronounced rooted)
1818 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1819 At the same time, the client is kept abreast of internal changes by
1820 virtue of being a @code{libgdb} observer.
1822 The only restriction on the client is that it must wait until
1823 @code{libgdb} becomes idle before initiating any queries (using the
1824 client's custom builder).
1826 @section @code{libgdb} components
1828 @subheading Observer - @file{gdb-events.h}
1829 @file{gdb-events} provides the client with a very raw mechanism that can
1830 be used to implement an observer. At present it only allows for one
1831 observer and that observer must, internally, handle the need to delay
1832 the processing of any event notifications until after @code{libgdb} has
1833 finished the current command.
1835 @subheading Builder - @file{ui-out.h}
1836 @file{ui-out} provides the infrastructure necessary for a client to
1837 create a builder. That builder is then passed down to @code{libgdb}
1838 when doing any queries.
1840 @subheading Event Loop - @file{event-loop.h}
1841 @c There could be an entire section on the event-loop
1842 @file{event-loop}, currently non-re-entrant, provides a simple event
1843 loop. A client would need to either plug its self into this loop or,
1844 implement a new event-loop that GDB would use.
1846 The event-loop will eventually be made re-entrant. This is so that
1847 @value{GDBN} can better handle the problem of some commands blocking
1848 instead of returning.
1850 @subheading Library - @file{gdb.h}
1851 @file{libgdb} is the most obvious component of this system. It provides
1852 the query interface. Each function is parameterized by a @code{ui-out}
1853 builder. The result of the query is constructed using that builder
1854 before the query function returns.
1856 @node Symbol Handling
1858 @chapter Symbol Handling
1860 Symbols are a key part of @value{GDBN}'s operation. Symbols include variables,
1861 functions, and types.
1863 @section Symbol Reading
1865 @cindex symbol reading
1866 @cindex reading of symbols
1867 @cindex symbol files
1868 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
1869 file is the file containing the program which @value{GDBN} is
1870 debugging. @value{GDBN} can be directed to use a different file for
1871 symbols (with the @samp{symbol-file} command), and it can also read
1872 more symbols via the @samp{add-file} and @samp{load} commands, or while
1873 reading symbols from shared libraries.
1875 @findex find_sym_fns
1876 Symbol files are initially opened by code in @file{symfile.c} using
1877 the BFD library (@pxref{Support Libraries}). BFD identifies the type
1878 of the file by examining its header. @code{find_sym_fns} then uses
1879 this identification to locate a set of symbol-reading functions.
1881 @findex add_symtab_fns
1882 @cindex @code{sym_fns} structure
1883 @cindex adding a symbol-reading module
1884 Symbol-reading modules identify themselves to @value{GDBN} by calling
1885 @code{add_symtab_fns} during their module initialization. The argument
1886 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
1887 name (or name prefix) of the symbol format, the length of the prefix,
1888 and pointers to four functions. These functions are called at various
1889 times to process symbol files whose identification matches the specified
1892 The functions supplied by each module are:
1895 @item @var{xyz}_symfile_init(struct sym_fns *sf)
1897 @cindex secondary symbol file
1898 Called from @code{symbol_file_add} when we are about to read a new
1899 symbol file. This function should clean up any internal state (possibly
1900 resulting from half-read previous files, for example) and prepare to
1901 read a new symbol file. Note that the symbol file which we are reading
1902 might be a new ``main'' symbol file, or might be a secondary symbol file
1903 whose symbols are being added to the existing symbol table.
1905 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
1906 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
1907 new symbol file being read. Its @code{private} field has been zeroed,
1908 and can be modified as desired. Typically, a struct of private
1909 information will be @code{malloc}'d, and a pointer to it will be placed
1910 in the @code{private} field.
1912 There is no result from @code{@var{xyz}_symfile_init}, but it can call
1913 @code{error} if it detects an unavoidable problem.
1915 @item @var{xyz}_new_init()
1917 Called from @code{symbol_file_add} when discarding existing symbols.
1918 This function needs only handle the symbol-reading module's internal
1919 state; the symbol table data structures visible to the rest of
1920 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
1921 arguments and no result. It may be called after
1922 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
1923 may be called alone if all symbols are simply being discarded.
1925 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
1927 Called from @code{symbol_file_add} to actually read the symbols from a
1928 symbol-file into a set of psymtabs or symtabs.
1930 @code{sf} points to the @code{struct sym_fns} originally passed to
1931 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
1932 the offset between the file's specified start address and its true
1933 address in memory. @code{mainline} is 1 if this is the main symbol
1934 table being read, and 0 if a secondary symbol file (e.g., shared library
1935 or dynamically loaded file) is being read.@refill
1938 In addition, if a symbol-reading module creates psymtabs when
1939 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
1940 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
1941 from any point in the @value{GDBN} symbol-handling code.
1944 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
1946 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
1947 the psymtab has not already been read in and had its @code{pst->symtab}
1948 pointer set. The argument is the psymtab to be fleshed-out into a
1949 symtab. Upon return, @code{pst->readin} should have been set to 1, and
1950 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
1951 zero if there were no symbols in that part of the symbol file.
1954 @section Partial Symbol Tables
1956 @value{GDBN} has three types of symbol tables:
1959 @cindex full symbol table
1962 Full symbol tables (@dfn{symtabs}). These contain the main
1963 information about symbols and addresses.
1967 Partial symbol tables (@dfn{psymtabs}). These contain enough
1968 information to know when to read the corresponding part of the full
1971 @cindex minimal symbol table
1974 Minimal symbol tables (@dfn{msymtabs}). These contain information
1975 gleaned from non-debugging symbols.
1978 @cindex partial symbol table
1979 This section describes partial symbol tables.
1981 A psymtab is constructed by doing a very quick pass over an executable
1982 file's debugging information. Small amounts of information are
1983 extracted---enough to identify which parts of the symbol table will
1984 need to be re-read and fully digested later, when the user needs the
1985 information. The speed of this pass causes @value{GDBN} to start up very
1986 quickly. Later, as the detailed rereading occurs, it occurs in small
1987 pieces, at various times, and the delay therefrom is mostly invisible to
1989 @c (@xref{Symbol Reading}.)
1991 The symbols that show up in a file's psymtab should be, roughly, those
1992 visible to the debugger's user when the program is not running code from
1993 that file. These include external symbols and types, static symbols and
1994 types, and @code{enum} values declared at file scope.
1996 The psymtab also contains the range of instruction addresses that the
1997 full symbol table would represent.
1999 @cindex finding a symbol
2000 @cindex symbol lookup
2001 The idea is that there are only two ways for the user (or much of the
2002 code in the debugger) to reference a symbol:
2005 @findex find_pc_function
2006 @findex find_pc_line
2008 By its address (e.g., execution stops at some address which is inside a
2009 function in this file). The address will be noticed to be in the
2010 range of this psymtab, and the full symtab will be read in.
2011 @code{find_pc_function}, @code{find_pc_line}, and other
2012 @code{find_pc_@dots{}} functions handle this.
2014 @cindex lookup_symbol
2017 (e.g., the user asks to print a variable, or set a breakpoint on a
2018 function). Global names and file-scope names will be found in the
2019 psymtab, which will cause the symtab to be pulled in. Local names will
2020 have to be qualified by a global name, or a file-scope name, in which
2021 case we will have already read in the symtab as we evaluated the
2022 qualifier. Or, a local symbol can be referenced when we are ``in'' a
2023 local scope, in which case the first case applies. @code{lookup_symbol}
2024 does most of the work here.
2027 The only reason that psymtabs exist is to cause a symtab to be read in
2028 at the right moment. Any symbol that can be elided from a psymtab,
2029 while still causing that to happen, should not appear in it. Since
2030 psymtabs don't have the idea of scope, you can't put local symbols in
2031 them anyway. Psymtabs don't have the idea of the type of a symbol,
2032 either, so types need not appear, unless they will be referenced by
2035 It is a bug for @value{GDBN} to behave one way when only a psymtab has
2036 been read, and another way if the corresponding symtab has been read
2037 in. Such bugs are typically caused by a psymtab that does not contain
2038 all the visible symbols, or which has the wrong instruction address
2041 The psymtab for a particular section of a symbol file (objfile) could be
2042 thrown away after the symtab has been read in. The symtab should always
2043 be searched before the psymtab, so the psymtab will never be used (in a
2044 bug-free environment). Currently, psymtabs are allocated on an obstack,
2045 and all the psymbols themselves are allocated in a pair of large arrays
2046 on an obstack, so there is little to be gained by trying to free them
2047 unless you want to do a lot more work.
2051 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
2053 @cindex fundamental types
2054 These are the fundamental types that @value{GDBN} uses internally. Fundamental
2055 types from the various debugging formats (stabs, ELF, etc) are mapped
2056 into one of these. They are basically a union of all fundamental types
2057 that @value{GDBN} knows about for all the languages that @value{GDBN}
2060 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
2063 Each time @value{GDBN} builds an internal type, it marks it with one
2064 of these types. The type may be a fundamental type, such as
2065 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
2066 which is a pointer to another type. Typically, several @code{FT_*}
2067 types map to one @code{TYPE_CODE_*} type, and are distinguished by
2068 other members of the type struct, such as whether the type is signed
2069 or unsigned, and how many bits it uses.
2071 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
2073 These are instances of type structs that roughly correspond to
2074 fundamental types and are created as global types for @value{GDBN} to
2075 use for various ugly historical reasons. We eventually want to
2076 eliminate these. Note for example that @code{builtin_type_int}
2077 initialized in @file{gdbtypes.c} is basically the same as a
2078 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
2079 an @code{FT_INTEGER} fundamental type. The difference is that the
2080 @code{builtin_type} is not associated with any particular objfile, and
2081 only one instance exists, while @file{c-lang.c} builds as many
2082 @code{TYPE_CODE_INT} types as needed, with each one associated with
2083 some particular objfile.
2085 @section Object File Formats
2086 @cindex object file formats
2090 @cindex @code{a.out} format
2091 The @code{a.out} format is the original file format for Unix. It
2092 consists of three sections: @code{text}, @code{data}, and @code{bss},
2093 which are for program code, initialized data, and uninitialized data,
2096 The @code{a.out} format is so simple that it doesn't have any reserved
2097 place for debugging information. (Hey, the original Unix hackers used
2098 @samp{adb}, which is a machine-language debugger!) The only debugging
2099 format for @code{a.out} is stabs, which is encoded as a set of normal
2100 symbols with distinctive attributes.
2102 The basic @code{a.out} reader is in @file{dbxread.c}.
2107 The COFF format was introduced with System V Release 3 (SVR3) Unix.
2108 COFF files may have multiple sections, each prefixed by a header. The
2109 number of sections is limited.
2111 The COFF specification includes support for debugging. Although this
2112 was a step forward, the debugging information was woefully limited. For
2113 instance, it was not possible to represent code that came from an
2116 The COFF reader is in @file{coffread.c}.
2120 @cindex ECOFF format
2121 ECOFF is an extended COFF originally introduced for Mips and Alpha
2124 The basic ECOFF reader is in @file{mipsread.c}.
2128 @cindex XCOFF format
2129 The IBM RS/6000 running AIX uses an object file format called XCOFF.
2130 The COFF sections, symbols, and line numbers are used, but debugging
2131 symbols are @code{dbx}-style stabs whose strings are located in the
2132 @code{.debug} section (rather than the string table). For more
2133 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
2135 The shared library scheme has a clean interface for figuring out what
2136 shared libraries are in use, but the catch is that everything which
2137 refers to addresses (symbol tables and breakpoints at least) needs to be
2138 relocated for both shared libraries and the main executable. At least
2139 using the standard mechanism this can only be done once the program has
2140 been run (or the core file has been read).
2144 @cindex PE-COFF format
2145 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
2146 executables. PE is basically COFF with additional headers.
2148 While BFD includes special PE support, @value{GDBN} needs only the basic
2154 The ELF format came with System V Release 4 (SVR4) Unix. ELF is similar
2155 to COFF in being organized into a number of sections, but it removes
2156 many of COFF's limitations.
2158 The basic ELF reader is in @file{elfread.c}.
2163 SOM is HP's object file and debug format (not to be confused with IBM's
2164 SOM, which is a cross-language ABI).
2166 The SOM reader is in @file{somread.c}.
2168 @section Debugging File Formats
2170 This section describes characteristics of debugging information that
2171 are independent of the object file format.
2175 @cindex stabs debugging info
2176 @code{stabs} started out as special symbols within the @code{a.out}
2177 format. Since then, it has been encapsulated into other file
2178 formats, such as COFF and ELF.
2180 While @file{dbxread.c} does some of the basic stab processing,
2181 including for encapsulated versions, @file{stabsread.c} does
2186 @cindex COFF debugging info
2187 The basic COFF definition includes debugging information. The level
2188 of support is minimal and non-extensible, and is not often used.
2190 @subsection Mips debug (Third Eye)
2192 @cindex ECOFF debugging info
2193 ECOFF includes a definition of a special debug format.
2195 The file @file{mdebugread.c} implements reading for this format.
2199 @cindex DWARF 2 debugging info
2200 DWARF 2 is an improved but incompatible version of DWARF 1.
2202 The DWARF 2 reader is in @file{dwarf2read.c}.
2206 @cindex SOM debugging info
2207 Like COFF, the SOM definition includes debugging information.
2209 @section Adding a New Symbol Reader to @value{GDBN}
2211 @cindex adding debugging info reader
2212 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
2213 there is probably little to be done.
2215 If you need to add a new object file format, you must first add it to
2216 BFD. This is beyond the scope of this document.
2218 You must then arrange for the BFD code to provide access to the
2219 debugging symbols. Generally @value{GDBN} will have to call swapping routines
2220 from BFD and a few other BFD internal routines to locate the debugging
2221 information. As much as possible, @value{GDBN} should not depend on the BFD
2222 internal data structures.
2224 For some targets (e.g., COFF), there is a special transfer vector used
2225 to call swapping routines, since the external data structures on various
2226 platforms have different sizes and layouts. Specialized routines that
2227 will only ever be implemented by one object file format may be called
2228 directly. This interface should be described in a file
2229 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
2231 @section Memory Management for Symbol Files
2233 Most memory associated with a loaded symbol file is stored on
2234 its @code{objfile_obstack}. This includes symbols, types,
2235 namespace data, and other information produced by the symbol readers.
2237 Because this data lives on the objfile's obstack, it is automatically
2238 released when the objfile is unloaded or reloaded. Therefore one
2239 objfile must not reference symbol or type data from another objfile;
2240 they could be unloaded at different times.
2242 User convenience variables, et cetera, have associated types. Normally
2243 these types live in the associated objfile. However, when the objfile
2244 is unloaded, those types are deep copied to global memory, so that
2245 the values of the user variables and history items are not lost.
2248 @node Language Support
2250 @chapter Language Support
2252 @cindex language support
2253 @value{GDBN}'s language support is mainly driven by the symbol reader,
2254 although it is possible for the user to set the source language
2257 @value{GDBN} chooses the source language by looking at the extension
2258 of the file recorded in the debug info; @file{.c} means C, @file{.f}
2259 means Fortran, etc. It may also use a special-purpose language
2260 identifier if the debug format supports it, like with DWARF.
2262 @section Adding a Source Language to @value{GDBN}
2264 @cindex adding source language
2265 To add other languages to @value{GDBN}'s expression parser, follow the
2269 @item Create the expression parser.
2271 @cindex expression parser
2272 This should reside in a file @file{@var{lang}-exp.y}. Routines for
2273 building parsed expressions into a @code{union exp_element} list are in
2276 @cindex language parser
2277 Since we can't depend upon everyone having Bison, and YACC produces
2278 parsers that define a bunch of global names, the following lines
2279 @strong{must} be included at the top of the YACC parser, to prevent the
2280 various parsers from defining the same global names:
2283 #define yyparse @var{lang}_parse
2284 #define yylex @var{lang}_lex
2285 #define yyerror @var{lang}_error
2286 #define yylval @var{lang}_lval
2287 #define yychar @var{lang}_char
2288 #define yydebug @var{lang}_debug
2289 #define yypact @var{lang}_pact
2290 #define yyr1 @var{lang}_r1
2291 #define yyr2 @var{lang}_r2
2292 #define yydef @var{lang}_def
2293 #define yychk @var{lang}_chk
2294 #define yypgo @var{lang}_pgo
2295 #define yyact @var{lang}_act
2296 #define yyexca @var{lang}_exca
2297 #define yyerrflag @var{lang}_errflag
2298 #define yynerrs @var{lang}_nerrs
2301 At the bottom of your parser, define a @code{struct language_defn} and
2302 initialize it with the right values for your language. Define an
2303 @code{initialize_@var{lang}} routine and have it call
2304 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
2305 that your language exists. You'll need some other supporting variables
2306 and functions, which will be used via pointers from your
2307 @code{@var{lang}_language_defn}. See the declaration of @code{struct
2308 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
2309 for more information.
2311 @item Add any evaluation routines, if necessary
2313 @cindex expression evaluation routines
2314 @findex evaluate_subexp
2315 @findex prefixify_subexp
2316 @findex length_of_subexp
2317 If you need new opcodes (that represent the operations of the language),
2318 add them to the enumerated type in @file{expression.h}. Add support
2319 code for these operations in the @code{evaluate_subexp} function
2320 defined in the file @file{eval.c}. Add cases
2321 for new opcodes in two functions from @file{parse.c}:
2322 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
2323 the number of @code{exp_element}s that a given operation takes up.
2325 @item Update some existing code
2327 Add an enumerated identifier for your language to the enumerated type
2328 @code{enum language} in @file{defs.h}.
2330 Update the routines in @file{language.c} so your language is included.
2331 These routines include type predicates and such, which (in some cases)
2332 are language dependent. If your language does not appear in the switch
2333 statement, an error is reported.
2335 @vindex current_language
2336 Also included in @file{language.c} is the code that updates the variable
2337 @code{current_language}, and the routines that translate the
2338 @code{language_@var{lang}} enumerated identifier into a printable
2341 @findex _initialize_language
2342 Update the function @code{_initialize_language} to include your
2343 language. This function picks the default language upon startup, so is
2344 dependent upon which languages that @value{GDBN} is built for.
2346 @findex allocate_symtab
2347 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2348 code so that the language of each symtab (source file) is set properly.
2349 This is used to determine the language to use at each stack frame level.
2350 Currently, the language is set based upon the extension of the source
2351 file. If the language can be better inferred from the symbol
2352 information, please set the language of the symtab in the symbol-reading
2355 @findex print_subexp
2356 @findex op_print_tab
2357 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2358 expression opcodes you have added to @file{expression.h}. Also, add the
2359 printed representations of your operators to @code{op_print_tab}.
2361 @item Add a place of call
2364 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2365 @code{parse_exp_1} (defined in @file{parse.c}).
2367 @item Use macros to trim code
2369 @cindex trimming language-dependent code
2370 The user has the option of building @value{GDBN} for some or all of the
2371 languages. If the user decides to build @value{GDBN} for the language
2372 @var{lang}, then every file dependent on @file{language.h} will have the
2373 macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to
2374 leave out large routines that the user won't need if he or she is not
2375 using your language.
2377 Note that you do not need to do this in your YACC parser, since if @value{GDBN}
2378 is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the
2379 compiled form of your parser) is not linked into @value{GDBN} at all.
2381 See the file @file{configure.in} for how @value{GDBN} is configured
2382 for different languages.
2384 @item Edit @file{Makefile.in}
2386 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2387 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2388 not get linked in, or, worse yet, it may not get @code{tar}red into the
2393 @node Host Definition
2395 @chapter Host Definition
2397 With the advent of Autoconf, it's rarely necessary to have host
2398 definition machinery anymore. The following information is provided,
2399 mainly, as an historical reference.
2401 @section Adding a New Host
2403 @cindex adding a new host
2404 @cindex host, adding
2405 @value{GDBN}'s host configuration support normally happens via Autoconf.
2406 New host-specific definitions should not be needed. Older hosts
2407 @value{GDBN} still use the host-specific definitions and files listed
2408 below, but these mostly exist for historical reasons, and will
2409 eventually disappear.
2412 @item gdb/config/@var{arch}/@var{xyz}.mh
2413 This file once contained both host and native configuration information
2414 (@pxref{Native Debugging}) for the machine @var{xyz}. The host
2415 configuration information is now handed by Autoconf.
2417 Host configuration information included a definition of
2418 @code{XM_FILE=xm-@var{xyz}.h} and possibly definitions for @code{CC},
2419 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2420 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2422 New host only configurations do not need this file.
2424 @item gdb/config/@var{arch}/xm-@var{xyz}.h
2425 This file once contained definitions and includes required when hosting
2426 gdb on machine @var{xyz}. Those definitions and includes are now
2427 handled by Autoconf.
2429 New host and native configurations do not need this file.
2431 @emph{Maintainer's note: Some hosts continue to use the @file{xm-xyz.h}
2432 file to define the macros @var{HOST_FLOAT_FORMAT},
2433 @var{HOST_DOUBLE_FORMAT} and @var{HOST_LONG_DOUBLE_FORMAT}. That code
2434 also needs to be replaced with either an Autoconf or run-time test.}
2438 @subheading Generic Host Support Files
2440 @cindex generic host support
2441 There are some ``generic'' versions of routines that can be used by
2442 various systems. These can be customized in various ways by macros
2443 defined in your @file{xm-@var{xyz}.h} file. If these routines work for
2444 the @var{xyz} host, you can just include the generic file's name (with
2445 @samp{.o}, not @samp{.c}) in @code{XDEPFILES}.
2447 Otherwise, if your machine needs custom support routines, you will need
2448 to write routines that perform the same functions as the generic file.
2449 Put them into @code{@var{xyz}-xdep.c}, and put @code{@var{xyz}-xdep.o}
2450 into @code{XDEPFILES}.
2453 @cindex remote debugging support
2454 @cindex serial line support
2456 This contains serial line support for Unix systems. This is always
2457 included, via the makefile variable @code{SER_HARDWIRE}; override this
2458 variable in the @file{.mh} file to avoid it.
2461 This contains serial line support for 32-bit programs running under DOS,
2462 using the DJGPP (a.k.a.@: GO32) execution environment.
2464 @cindex TCP remote support
2466 This contains generic TCP support using sockets.
2469 @section Host Conditionals
2471 When @value{GDBN} is configured and compiled, various macros are
2472 defined or left undefined, to control compilation based on the
2473 attributes of the host system. These macros and their meanings (or if
2474 the meaning is not documented here, then one of the source files where
2475 they are used is indicated) are:
2478 @item @value{GDBN}INIT_FILENAME
2479 The default name of @value{GDBN}'s initialization file (normally
2483 This macro is deprecated.
2485 @item SIGWINCH_HANDLER
2486 If your host defines @code{SIGWINCH}, you can define this to be the name
2487 of a function to be called if @code{SIGWINCH} is received.
2489 @item SIGWINCH_HANDLER_BODY
2490 Define this to expand into code that will define the function named by
2491 the expansion of @code{SIGWINCH_HANDLER}.
2493 @item CRLF_SOURCE_FILES
2494 @cindex DOS text files
2495 Define this if host files use @code{\r\n} rather than @code{\n} as a
2496 line terminator. This will cause source file listings to omit @code{\r}
2497 characters when printing and it will allow @code{\r\n} line endings of files
2498 which are ``sourced'' by gdb. It must be possible to open files in binary
2499 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2501 @item DEFAULT_PROMPT
2503 The default value of the prompt string (normally @code{"(gdb) "}).
2506 @cindex terminal device
2507 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2510 Define this if binary files are opened the same way as text files.
2514 In some cases, use the system call @code{mmap} for reading symbol
2515 tables. For some machines this allows for sharing and quick updates.
2518 Define this if the host system has @code{termio.h}.
2525 Values for host-side constants.
2528 Substitute for isatty, if not available.
2531 This is the longest integer type available on the host. If not defined,
2532 it will default to @code{long long} or @code{long}, depending on
2533 @code{CC_HAS_LONG_LONG}.
2535 @item CC_HAS_LONG_LONG
2536 @cindex @code{long long} data type
2537 Define this if the host C compiler supports @code{long long}. This is set
2538 by the @code{configure} script.
2540 @item PRINTF_HAS_LONG_LONG
2541 Define this if the host can handle printing of long long integers via
2542 the printf format conversion specifier @code{ll}. This is set by the
2543 @code{configure} script.
2545 @item HAVE_LONG_DOUBLE
2546 Define this if the host C compiler supports @code{long double}. This is
2547 set by the @code{configure} script.
2549 @item PRINTF_HAS_LONG_DOUBLE
2550 Define this if the host can handle printing of long double float-point
2551 numbers via the printf format conversion specifier @code{Lg}. This is
2552 set by the @code{configure} script.
2554 @item SCANF_HAS_LONG_DOUBLE
2555 Define this if the host can handle the parsing of long double
2556 float-point numbers via the scanf format conversion specifier
2557 @code{Lg}. This is set by the @code{configure} script.
2559 @item LSEEK_NOT_LINEAR
2560 Define this if @code{lseek (n)} does not necessarily move to byte number
2561 @code{n} in the file. This is only used when reading source files. It
2562 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2565 This macro is used as the argument to @code{lseek} (or, most commonly,
2566 @code{bfd_seek}). FIXME, should be replaced by SEEK_SET instead,
2567 which is the POSIX equivalent.
2570 If defined, this should be one or more tokens, such as @code{volatile},
2571 that can be used in both the declaration and definition of functions to
2572 indicate that they never return. The default is already set correctly
2573 if compiling with GCC. This will almost never need to be defined.
2576 If defined, this should be one or more tokens, such as
2577 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2578 of functions to indicate that they never return. The default is already
2579 set correctly if compiling with GCC. This will almost never need to be
2584 Define these to appropriate value for the system @code{lseek}, if not already
2588 This is the signal for stopping @value{GDBN}. Defaults to
2589 @code{SIGTSTP}. (Only redefined for the Convex.)
2592 Means that System V (prior to SVR4) include files are in use. (FIXME:
2593 This symbol is abused in @file{infrun.c}, @file{regex.c}, and
2594 @file{utils.c} for other things, at the moment.)
2597 Define this to help placate @code{lint} in some situations.
2600 Define this to override the defaults of @code{__volatile__} or
2605 @node Target Architecture Definition
2607 @chapter Target Architecture Definition
2609 @cindex target architecture definition
2610 @value{GDBN}'s target architecture defines what sort of
2611 machine-language programs @value{GDBN} can work with, and how it works
2614 The target architecture object is implemented as the C structure
2615 @code{struct gdbarch *}. The structure, and its methods, are generated
2616 using the Bourne shell script @file{gdbarch.sh}.
2619 * OS ABI Variant Handling::
2620 * Initialize New Architecture::
2621 * Registers and Memory::
2622 * Pointers and Addresses::
2624 * Raw and Virtual Registers::
2625 * Register and Memory Data::
2626 * Frame Interpretation::
2627 * Inferior Call Setup::
2628 * Compiler Characteristics::
2629 * Target Conditionals::
2630 * Adding a New Target::
2633 @node OS ABI Variant Handling
2634 @section Operating System ABI Variant Handling
2635 @cindex OS ABI variants
2637 @value{GDBN} provides a mechanism for handling variations in OS
2638 ABIs. An OS ABI variant may have influence over any number of
2639 variables in the target architecture definition. There are two major
2640 components in the OS ABI mechanism: sniffers and handlers.
2642 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2643 (the architecture may be wildcarded) in an attempt to determine the
2644 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2645 to be @dfn{generic}, while sniffers for a specific architecture are
2646 considered to be @dfn{specific}. A match from a specific sniffer
2647 overrides a match from a generic sniffer. Multiple sniffers for an
2648 architecture/flavour may exist, in order to differentiate between two
2649 different operating systems which use the same basic file format. The
2650 OS ABI framework provides a generic sniffer for ELF-format files which
2651 examines the @code{EI_OSABI} field of the ELF header, as well as note
2652 sections known to be used by several operating systems.
2654 @cindex fine-tuning @code{gdbarch} structure
2655 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2656 selected OS ABI. There may be only one handler for a given OS ABI
2657 for each BFD architecture.
2659 The following OS ABI variants are defined in @file{defs.h}:
2663 @findex GDB_OSABI_UNINITIALIZED
2664 @item GDB_OSABI_UNINITIALIZED
2665 Used for struct gdbarch_info if ABI is still uninitialized.
2667 @findex GDB_OSABI_UNKNOWN
2668 @item GDB_OSABI_UNKNOWN
2669 The ABI of the inferior is unknown. The default @code{gdbarch}
2670 settings for the architecture will be used.
2672 @findex GDB_OSABI_SVR4
2673 @item GDB_OSABI_SVR4
2674 UNIX System V Release 4.
2676 @findex GDB_OSABI_HURD
2677 @item GDB_OSABI_HURD
2678 GNU using the Hurd kernel.
2680 @findex GDB_OSABI_SOLARIS
2681 @item GDB_OSABI_SOLARIS
2684 @findex GDB_OSABI_OSF1
2685 @item GDB_OSABI_OSF1
2686 OSF/1, including Digital UNIX and Compaq Tru64 UNIX.
2688 @findex GDB_OSABI_LINUX
2689 @item GDB_OSABI_LINUX
2690 GNU using the Linux kernel.
2692 @findex GDB_OSABI_FREEBSD_AOUT
2693 @item GDB_OSABI_FREEBSD_AOUT
2694 FreeBSD using the @code{a.out} executable format.
2696 @findex GDB_OSABI_FREEBSD_ELF
2697 @item GDB_OSABI_FREEBSD_ELF
2698 FreeBSD using the ELF executable format.
2700 @findex GDB_OSABI_NETBSD_AOUT
2701 @item GDB_OSABI_NETBSD_AOUT
2702 NetBSD using the @code{a.out} executable format.
2704 @findex GDB_OSABI_NETBSD_ELF
2705 @item GDB_OSABI_NETBSD_ELF
2706 NetBSD using the ELF executable format.
2708 @findex GDB_OSABI_OPENBSD_ELF
2709 @item GDB_OSABI_OPENBSD_ELF
2710 OpenBSD using the ELF executable format.
2712 @findex GDB_OSABI_WINCE
2713 @item GDB_OSABI_WINCE
2716 @findex GDB_OSABI_GO32
2717 @item GDB_OSABI_GO32
2720 @findex GDB_OSABI_IRIX
2721 @item GDB_OSABI_IRIX
2724 @findex GDB_OSABI_INTERIX
2725 @item GDB_OSABI_INTERIX
2726 Interix (Posix layer for MS-Windows systems).
2728 @findex GDB_OSABI_HPUX_ELF
2729 @item GDB_OSABI_HPUX_ELF
2730 HP/UX using the ELF executable format.
2732 @findex GDB_OSABI_HPUX_SOM
2733 @item GDB_OSABI_HPUX_SOM
2734 HP/UX using the SOM executable format.
2736 @findex GDB_OSABI_QNXNTO
2737 @item GDB_OSABI_QNXNTO
2740 @findex GDB_OSABI_CYGWIN
2741 @item GDB_OSABI_CYGWIN
2744 @findex GDB_OSABI_AIX
2750 Here are the functions that make up the OS ABI framework:
2752 @deftypefun const char *gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2753 Return the name of the OS ABI corresponding to @var{osabi}.
2756 @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}))
2757 Register the OS ABI handler specified by @var{init_osabi} for the
2758 architecture, machine type and OS ABI specified by @var{arch},
2759 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2760 machine type, which implies the architecture's default machine type,
2764 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2765 Register the OS ABI file sniffer specified by @var{sniffer} for the
2766 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2767 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2768 be generic, and is allowed to examine @var{flavour}-flavoured files for
2772 @deftypefun enum gdb_osabi gdbarch_lookup_osabi (bfd *@var{abfd})
2773 Examine the file described by @var{abfd} to determine its OS ABI.
2774 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2778 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2779 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2780 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2781 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2782 architecture, a warning will be issued and the debugging session will continue
2783 with the defaults already established for @var{gdbarch}.
2786 @deftypefun void generic_elf_osabi_sniff_abi_tag_sections (bfd *@var{abfd}, asection *@var{sect}, void *@var{obj})
2787 Helper routine for ELF file sniffers. Examine the file described by
2788 @var{abfd} and look at ABI tag note sections to determine the OS ABI
2789 from the note. This function should be called via
2790 @code{bfd_map_over_sections}.
2793 @node Initialize New Architecture
2794 @section Initializing a New Architecture
2796 Each @code{gdbarch} is associated with a single @sc{bfd} architecture,
2797 via a @code{bfd_arch_@var{arch}} constant. The @code{gdbarch} is
2798 registered by a call to @code{register_gdbarch_init}, usually from
2799 the file's @code{_initialize_@var{filename}} routine, which will
2800 be automatically called during @value{GDBN} startup. The arguments
2801 are a @sc{bfd} architecture constant and an initialization function.
2803 The initialization function has this type:
2806 static struct gdbarch *
2807 @var{arch}_gdbarch_init (struct gdbarch_info @var{info},
2808 struct gdbarch_list *@var{arches})
2811 The @var{info} argument contains parameters used to select the correct
2812 architecture, and @var{arches} is a list of architectures which
2813 have already been created with the same @code{bfd_arch_@var{arch}}
2816 The initialization function should first make sure that @var{info}
2817 is acceptable, and return @code{NULL} if it is not. Then, it should
2818 search through @var{arches} for an exact match to @var{info}, and
2819 return one if found. Lastly, if no exact match was found, it should
2820 create a new architecture based on @var{info} and return it.
2822 Only information in @var{info} should be used to choose the new
2823 architecture. Historically, @var{info} could be sparse, and
2824 defaults would be collected from the first element on @var{arches}.
2825 However, @value{GDBN} now fills in @var{info} more thoroughly,
2826 so new @code{gdbarch} initialization functions should not take
2827 defaults from @var{arches}.
2829 @node Registers and Memory
2830 @section Registers and Memory
2832 @value{GDBN}'s model of the target machine is rather simple.
2833 @value{GDBN} assumes the machine includes a bank of registers and a
2834 block of memory. Each register may have a different size.
2836 @value{GDBN} does not have a magical way to match up with the
2837 compiler's idea of which registers are which; however, it is critical
2838 that they do match up accurately. The only way to make this work is
2839 to get accurate information about the order that the compiler uses,
2840 and to reflect that in the @code{gdbarch_register_name} and related functions.
2842 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
2844 @node Pointers and Addresses
2845 @section Pointers Are Not Always Addresses
2846 @cindex pointer representation
2847 @cindex address representation
2848 @cindex word-addressed machines
2849 @cindex separate data and code address spaces
2850 @cindex spaces, separate data and code address
2851 @cindex address spaces, separate data and code
2852 @cindex code pointers, word-addressed
2853 @cindex converting between pointers and addresses
2854 @cindex D10V addresses
2856 On almost all 32-bit architectures, the representation of a pointer is
2857 indistinguishable from the representation of some fixed-length number
2858 whose value is the byte address of the object pointed to. On such
2859 machines, the words ``pointer'' and ``address'' can be used interchangeably.
2860 However, architectures with smaller word sizes are often cramped for
2861 address space, so they may choose a pointer representation that breaks this
2862 identity, and allows a larger code address space.
2864 For example, the Renesas D10V is a 16-bit VLIW processor whose
2865 instructions are 32 bits long@footnote{Some D10V instructions are
2866 actually pairs of 16-bit sub-instructions. However, since you can't
2867 jump into the middle of such a pair, code addresses can only refer to
2868 full 32 bit instructions, which is what matters in this explanation.}.
2869 If the D10V used ordinary byte addresses to refer to code locations,
2870 then the processor would only be able to address 64kb of instructions.
2871 However, since instructions must be aligned on four-byte boundaries, the
2872 low two bits of any valid instruction's byte address are always
2873 zero---byte addresses waste two bits. So instead of byte addresses,
2874 the D10V uses word addresses---byte addresses shifted right two bits---to
2875 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
2878 However, this means that code pointers and data pointers have different
2879 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
2880 @code{0xC020} when used as a data address, but refers to byte address
2881 @code{0x30080} when used as a code address.
2883 (The D10V also uses separate code and data address spaces, which also
2884 affects the correspondence between pointers and addresses, but we're
2885 going to ignore that here; this example is already too long.)
2887 To cope with architectures like this---the D10V is not the only
2888 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
2889 byte numbers, and @dfn{pointers}, which are the target's representation
2890 of an address of a particular type of data. In the example above,
2891 @code{0xC020} is the pointer, which refers to one of the addresses
2892 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
2893 @value{GDBN} provides functions for turning a pointer into an address
2894 and vice versa, in the appropriate way for the current architecture.
2896 Unfortunately, since addresses and pointers are identical on almost all
2897 processors, this distinction tends to bit-rot pretty quickly. Thus,
2898 each time you port @value{GDBN} to an architecture which does
2899 distinguish between pointers and addresses, you'll probably need to
2900 clean up some architecture-independent code.
2902 Here are functions which convert between pointers and addresses:
2904 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
2905 Treat the bytes at @var{buf} as a pointer or reference of type
2906 @var{type}, and return the address it represents, in a manner
2907 appropriate for the current architecture. This yields an address
2908 @value{GDBN} can use to read target memory, disassemble, etc. Note that
2909 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2912 For example, if the current architecture is the Intel x86, this function
2913 extracts a little-endian integer of the appropriate length from
2914 @var{buf} and returns it. However, if the current architecture is the
2915 D10V, this function will return a 16-bit integer extracted from
2916 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
2918 If @var{type} is not a pointer or reference type, then this function
2919 will signal an internal error.
2922 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
2923 Store the address @var{addr} in @var{buf}, in the proper format for a
2924 pointer of type @var{type} in the current architecture. Note that
2925 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2928 For example, if the current architecture is the Intel x86, this function
2929 stores @var{addr} unmodified as a little-endian integer of the
2930 appropriate length in @var{buf}. However, if the current architecture
2931 is the D10V, this function divides @var{addr} by four if @var{type} is
2932 a pointer to a function, and then stores it in @var{buf}.
2934 If @var{type} is not a pointer or reference type, then this function
2935 will signal an internal error.
2938 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
2939 Assuming that @var{val} is a pointer, return the address it represents,
2940 as appropriate for the current architecture.
2942 This function actually works on integral values, as well as pointers.
2943 For pointers, it performs architecture-specific conversions as
2944 described above for @code{extract_typed_address}.
2947 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
2948 Create and return a value representing a pointer of type @var{type} to
2949 the address @var{addr}, as appropriate for the current architecture.
2950 This function performs architecture-specific conversions as described
2951 above for @code{store_typed_address}.
2954 Here are two functions which architectures can define to indicate the
2955 relationship between pointers and addresses. These have default
2956 definitions, appropriate for architectures on which all pointers are
2957 simple unsigned byte addresses.
2959 @deftypefun CORE_ADDR gdbarch_pointer_to_address (struct gdbarch *@var{current_gdbarch}, struct type *@var{type}, char *@var{buf})
2960 Assume that @var{buf} holds a pointer of type @var{type}, in the
2961 appropriate format for the current architecture. Return the byte
2962 address the pointer refers to.
2964 This function may safely assume that @var{type} is either a pointer or a
2965 C@t{++} reference type.
2968 @deftypefun void gdbarch_address_to_pointer (struct gdbarch *@var{current_gdbarch}, struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
2969 Store in @var{buf} a pointer of type @var{type} representing the address
2970 @var{addr}, in the appropriate format for the current architecture.
2972 This function may safely assume that @var{type} is either a pointer or a
2973 C@t{++} reference type.
2976 @node Address Classes
2977 @section Address Classes
2978 @cindex address classes
2979 @cindex DW_AT_byte_size
2980 @cindex DW_AT_address_class
2982 Sometimes information about different kinds of addresses is available
2983 via the debug information. For example, some programming environments
2984 define addresses of several different sizes. If the debug information
2985 distinguishes these kinds of address classes through either the size
2986 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
2987 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
2988 following macros should be defined in order to disambiguate these
2989 types within @value{GDBN} as well as provide the added information to
2990 a @value{GDBN} user when printing type expressions.
2992 @deftypefun int gdbarch_address_class_type_flags (struct gdbarch *@var{current_gdbarch}, int @var{byte_size}, int @var{dwarf2_addr_class})
2993 Returns the type flags needed to construct a pointer type whose size
2994 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
2995 This function is normally called from within a symbol reader. See
2996 @file{dwarf2read.c}.
2999 @deftypefun char *gdbarch_address_class_type_flags_to_name (struct gdbarch *@var{current_gdbarch}, int @var{type_flags})
3000 Given the type flags representing an address class qualifier, return
3003 @deftypefun int gdbarch_address_class_name_to_type_flags (struct gdbarch *@var{current_gdbarch}, int @var{name}, int *var{type_flags_ptr})
3004 Given an address qualifier name, set the @code{int} referenced by @var{type_flags_ptr} to the type flags
3005 for that address class qualifier.
3008 Since the need for address classes is rather rare, none of
3009 the address class functions are defined by default. Predicate
3010 functions are provided to detect when they are defined.
3012 Consider a hypothetical architecture in which addresses are normally
3013 32-bits wide, but 16-bit addresses are also supported. Furthermore,
3014 suppose that the @w{DWARF 2} information for this architecture simply
3015 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
3016 of these "short" pointers. The following functions could be defined
3017 to implement the address class functions:
3020 somearch_address_class_type_flags (int byte_size,
3021 int dwarf2_addr_class)
3024 return TYPE_FLAG_ADDRESS_CLASS_1;
3030 somearch_address_class_type_flags_to_name (int type_flags)
3032 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
3039 somearch_address_class_name_to_type_flags (char *name,
3040 int *type_flags_ptr)
3042 if (strcmp (name, "short") == 0)
3044 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
3052 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
3053 to indicate the presence of one of these "short" pointers. E.g, if
3054 the debug information indicates that @code{short_ptr_var} is one of these
3055 short pointers, @value{GDBN} might show the following behavior:
3058 (gdb) ptype short_ptr_var
3059 type = int * @@short
3063 @node Raw and Virtual Registers
3064 @section Raw and Virtual Register Representations
3065 @cindex raw register representation
3066 @cindex virtual register representation
3067 @cindex representations, raw and virtual registers
3069 @emph{Maintainer note: This section is pretty much obsolete. The
3070 functionality described here has largely been replaced by
3071 pseudo-registers and the mechanisms described in @ref{Target
3072 Architecture Definition, , Using Different Register and Memory Data
3073 Representations}. See also @uref{http://www.gnu.org/software/gdb/bugs/,
3074 Bug Tracking Database} and
3075 @uref{http://sources.redhat.com/gdb/current/ari/, ARI Index} for more
3076 up-to-date information.}
3078 Some architectures use one representation for a value when it lives in a
3079 register, but use a different representation when it lives in memory.
3080 In @value{GDBN}'s terminology, the @dfn{raw} representation is the one used in
3081 the target registers, and the @dfn{virtual} representation is the one
3082 used in memory, and within @value{GDBN} @code{struct value} objects.
3084 @emph{Maintainer note: Notice that the same mechanism is being used to
3085 both convert a register to a @code{struct value} and alternative
3088 For almost all data types on almost all architectures, the virtual and
3089 raw representations are identical, and no special handling is needed.
3090 However, they do occasionally differ. For example:
3094 The x86 architecture supports an 80-bit @code{long double} type. However, when
3095 we store those values in memory, they occupy twelve bytes: the
3096 floating-point number occupies the first ten, and the final two bytes
3097 are unused. This keeps the values aligned on four-byte boundaries,
3098 allowing more efficient access. Thus, the x86 80-bit floating-point
3099 type is the raw representation, and the twelve-byte loosely-packed
3100 arrangement is the virtual representation.
3103 Some 64-bit MIPS targets present 32-bit registers to @value{GDBN} as 64-bit
3104 registers, with garbage in their upper bits. @value{GDBN} ignores the top 32
3105 bits. Thus, the 64-bit form, with garbage in the upper 32 bits, is the
3106 raw representation, and the trimmed 32-bit representation is the
3107 virtual representation.
3110 In general, the raw representation is determined by the architecture, or
3111 @value{GDBN}'s interface to the architecture, while the virtual representation
3112 can be chosen for @value{GDBN}'s convenience. @value{GDBN}'s register file,
3113 @code{registers}, holds the register contents in raw format, and the
3114 @value{GDBN} remote protocol transmits register values in raw format.
3116 Your architecture may define the following macros to request
3117 conversions between the raw and virtual format:
3119 @deftypefn {Target Macro} int REGISTER_CONVERTIBLE (int @var{reg})
3120 Return non-zero if register number @var{reg}'s value needs different raw
3121 and virtual formats.
3123 You should not use @code{REGISTER_CONVERT_TO_VIRTUAL} for a register
3124 unless this macro returns a non-zero value for that register.
3127 @deftypefn {Target Macro} int DEPRECATED_REGISTER_RAW_SIZE (int @var{reg})
3128 The size of register number @var{reg}'s raw value. This is the number
3129 of bytes the register will occupy in @code{registers}, or in a @value{GDBN}
3130 remote protocol packet.
3133 @deftypefn {Target Macro} int DEPRECATED_REGISTER_VIRTUAL_SIZE (int @var{reg})
3134 The size of register number @var{reg}'s value, in its virtual format.
3135 This is the size a @code{struct value}'s buffer will have, holding that
3139 @deftypefn {Target Macro} struct type *DEPRECATED_REGISTER_VIRTUAL_TYPE (int @var{reg})
3140 This is the type of the virtual representation of register number
3141 @var{reg}. Note that there is no need for a macro giving a type for the
3142 register's raw form; once the register's value has been obtained, @value{GDBN}
3143 always uses the virtual form.
3146 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_VIRTUAL (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
3147 Convert the value of register number @var{reg} to @var{type}, which
3148 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
3149 at @var{from} holds the register's value in raw format; the macro should
3150 convert the value to virtual format, and place it at @var{to}.
3152 Note that @code{REGISTER_CONVERT_TO_VIRTUAL} and
3153 @code{REGISTER_CONVERT_TO_RAW} take their @var{reg} and @var{type}
3154 arguments in different orders.
3156 You should only use @code{REGISTER_CONVERT_TO_VIRTUAL} with registers
3157 for which the @code{REGISTER_CONVERTIBLE} macro returns a non-zero
3161 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_RAW (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3162 Convert the value of register number @var{reg} to @var{type}, which
3163 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
3164 at @var{from} holds the register's value in raw format; the macro should
3165 convert the value to virtual format, and place it at @var{to}.
3167 Note that REGISTER_CONVERT_TO_VIRTUAL and REGISTER_CONVERT_TO_RAW take
3168 their @var{reg} and @var{type} arguments in different orders.
3172 @node Register and Memory Data
3173 @section Using Different Register and Memory Data Representations
3174 @cindex register representation
3175 @cindex memory representation
3176 @cindex representations, register and memory
3177 @cindex register data formats, converting
3178 @cindex @code{struct value}, converting register contents to
3180 @emph{Maintainer's note: The way GDB manipulates registers is undergoing
3181 significant change. Many of the macros and functions referred to in this
3182 section are likely to be subject to further revision. See
3183 @uref{http://sources.redhat.com/gdb/current/ari/, A.R. Index} and
3184 @uref{http://www.gnu.org/software/gdb/bugs, Bug Tracking Database} for
3185 further information. cagney/2002-05-06.}
3187 Some architectures can represent a data object in a register using a
3188 form that is different to the objects more normal memory representation.
3194 The Alpha architecture can represent 32 bit integer values in
3195 floating-point registers.
3198 The x86 architecture supports 80-bit floating-point registers. The
3199 @code{long double} data type occupies 96 bits in memory but only 80 bits
3200 when stored in a register.
3204 In general, the register representation of a data type is determined by
3205 the architecture, or @value{GDBN}'s interface to the architecture, while
3206 the memory representation is determined by the Application Binary
3209 For almost all data types on almost all architectures, the two
3210 representations are identical, and no special handling is needed.
3211 However, they do occasionally differ. Your architecture may define the
3212 following macros to request conversions between the register and memory
3213 representations of a data type:
3215 @deftypefun int gdbarch_convert_register_p (struct gdbarch *@var{gdbarch}, int @var{reg})
3216 Return non-zero if the representation of a data value stored in this
3217 register may be different to the representation of that same data value
3218 when stored in memory.
3220 When non-zero, the macros @code{gdbarch_register_to_value} and
3221 @code{value_to_register} are used to perform any necessary conversion.
3223 This function should return zero for the register's native type, when
3224 no conversion is necessary.
3227 @deftypefun void gdbarch_register_to_value (struct gdbarch *@var{gdbarch}, int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
3228 Convert the value of register number @var{reg} to a data object of type
3229 @var{type}. The buffer at @var{from} holds the register's value in raw
3230 format; the converted value should be placed in the buffer at @var{to}.
3232 Note that @code{gdbarch_register_to_value} and @code{gdbarch_value_to_register}
3233 take their @var{reg} and @var{type} arguments in different orders.
3235 You should only use @code{gdbarch_register_to_value} with registers for which
3236 the @code{gdbarch_convert_register_p} function returns a non-zero value.
3239 @deftypefun void gdbarch_value_to_register (struct gdbarch *@var{gdbarch}, struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3240 Convert a data value of type @var{type} to register number @var{reg}'
3243 Note that @code{gdbarch_register_to_value} and @code{gdbarch_value_to_register}
3244 take their @var{reg} and @var{type} arguments in different orders.
3246 You should only use @code{gdbarch_value_to_register} with registers for which
3247 the @code{gdbarch_convert_register_p} function returns a non-zero value.
3250 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_TYPE (int @var{regnum}, struct type *@var{type}, char *@var{buf})
3251 See @file{mips-tdep.c}. It does not do what you want.
3254 @node Frame Interpretation
3255 @section Frame Interpretation
3257 @node Inferior Call Setup
3258 @section Inferior Call Setup
3260 @node Compiler Characteristics
3261 @section Compiler Characteristics
3263 @node Target Conditionals
3264 @section Target Conditionals
3266 This section describes the macros and functions that you can use to define the
3271 @item CORE_ADDR gdbarch_addr_bits_remove (@var{gdbarch}, @var{addr})
3272 @findex gdbarch_addr_bits_remove
3273 If a raw machine instruction address includes any bits that are not
3274 really part of the address, then this function is used to zero those bits in
3275 @var{addr}. This is only used for addresses of instructions, and even then not
3278 For example, the two low-order bits of the PC on the Hewlett-Packard PA
3279 2.0 architecture contain the privilege level of the corresponding
3280 instruction. Since instructions must always be aligned on four-byte
3281 boundaries, the processor masks out these bits to generate the actual
3282 address of the instruction. @code{gdbarch_addr_bits_remove} would then for
3283 example look like that:
3285 arch_addr_bits_remove (CORE_ADDR addr)
3287 return (addr &= ~0x3);
3291 @item int address_class_name_to_type_flags (@var{gdbarch}, @var{name}, @var{type_flags_ptr})
3292 @findex address_class_name_to_type_flags
3293 If @var{name} is a valid address class qualifier name, set the @code{int}
3294 referenced by @var{type_flags_ptr} to the mask representing the qualifier
3295 and return 1. If @var{name} is not a valid address class qualifier name,
3298 The value for @var{type_flags_ptr} should be one of
3299 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
3300 possibly some combination of these values or'd together.
3301 @xref{Target Architecture Definition, , Address Classes}.
3303 @item int address_class_name_to_type_flags_p (@var{gdbarch})
3304 @findex address_class_name_to_type_flags_p
3305 Predicate which indicates whether @code{address_class_name_to_type_flags}
3308 @item int gdbarch_address_class_type_flags (@var{gdbarch}, @var{byte_size}, @var{dwarf2_addr_class})
3309 @findex gdbarch_address_class_type_flags
3310 Given a pointers byte size (as described by the debug information) and
3311 the possible @code{DW_AT_address_class} value, return the type flags
3312 used by @value{GDBN} to represent this address class. The value
3313 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
3314 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
3315 values or'd together.
3316 @xref{Target Architecture Definition, , Address Classes}.
3318 @item int gdbarch_address_class_type_flags_p (@var{gdbarch})
3319 @findex gdbarch_address_class_type_flags_p
3320 Predicate which indicates whether @code{gdbarch_address_class_type_flags_p} has
3323 @item const char *gdbarch_address_class_type_flags_to_name (@var{gdbarch}, @var{type_flags})
3324 @findex gdbarch_address_class_type_flags_to_name
3325 Return the name of the address class qualifier associated with the type
3326 flags given by @var{type_flags}.
3328 @item int gdbarch_address_class_type_flags_to_name_p (@var{gdbarch})
3329 @findex gdbarch_address_class_type_flags_to_name_p
3330 Predicate which indicates whether @code{gdbarch_address_class_type_flags_to_name} has been defined.
3331 @xref{Target Architecture Definition, , Address Classes}.
3333 @item void gdbarch_address_to_pointer (@var{gdbarch}, @var{type}, @var{buf}, @var{addr})
3334 @findex gdbarch_address_to_pointer
3335 Store in @var{buf} a pointer of type @var{type} representing the address
3336 @var{addr}, in the appropriate format for the current architecture.
3337 This function may safely assume that @var{type} is either a pointer or a
3338 C@t{++} reference type.
3339 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3341 @item int gdbarch_believe_pcc_promotion (@var{gdbarch})
3342 @findex gdbarch_believe_pcc_promotion
3343 Used to notify if the compiler promotes a @code{short} or @code{char}
3344 parameter to an @code{int}, but still reports the parameter as its
3345 original type, rather than the promoted type.
3347 @item gdbarch_bits_big_endian (@var{gdbarch})
3348 @findex gdbarch_bits_big_endian
3349 This is used if the numbering of bits in the targets does @strong{not} match
3350 the endianness of the target byte order. A value of 1 means that the bits
3351 are numbered in a big-endian bit order, 0 means little-endian.
3353 @item set_gdbarch_bits_big_endian (@var{gdbarch}, @var{bits_big_endian})
3354 @findex set_gdbarch_bits_big_endian
3355 Calling set_gdbarch_bits_big_endian with a value of 1 indicates that the
3356 bits in the target are numbered in a big-endian bit order, 0 indicates
3361 This is the character array initializer for the bit pattern to put into
3362 memory where a breakpoint is set. Although it's common to use a trap
3363 instruction for a breakpoint, it's not required; for instance, the bit
3364 pattern could be an invalid instruction. The breakpoint must be no
3365 longer than the shortest instruction of the architecture.
3367 @code{BREAKPOINT} has been deprecated in favor of
3368 @code{gdbarch_breakpoint_from_pc}.
3370 @item BIG_BREAKPOINT
3371 @itemx LITTLE_BREAKPOINT
3372 @findex LITTLE_BREAKPOINT
3373 @findex BIG_BREAKPOINT
3374 Similar to BREAKPOINT, but used for bi-endian targets.
3376 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
3377 favor of @code{gdbarch_breakpoint_from_pc}.
3379 @item const gdb_byte *gdbarch_breakpoint_from_pc (@var{gdbarch}, @var{pcptr}, @var{lenptr})
3380 @findex gdbarch_breakpoint_from_pc
3381 @anchor{gdbarch_breakpoint_from_pc} Use the program counter to determine the
3382 contents and size of a breakpoint instruction. It returns a pointer to
3383 a string of bytes that encode a breakpoint instruction, stores the
3384 length of the string to @code{*@var{lenptr}}, and adjusts the program
3385 counter (if necessary) to point to the actual memory location where the
3386 breakpoint should be inserted.
3388 Although it is common to use a trap instruction for a breakpoint, it's
3389 not required; for instance, the bit pattern could be an invalid
3390 instruction. The breakpoint must be no longer than the shortest
3391 instruction of the architecture.
3393 Replaces all the other @var{BREAKPOINT} macros.
3395 @item int gdbarch_memory_insert_breakpoint (@var{gdbarch}, @var{bp_tgt})
3396 @itemx gdbarch_memory_remove_breakpoint (@var{gdbarch}, @var{bp_tgt})
3397 @findex gdbarch_memory_remove_breakpoint
3398 @findex gdbarch_memory_insert_breakpoint
3399 Insert or remove memory based breakpoints. Reasonable defaults
3400 (@code{default_memory_insert_breakpoint} and
3401 @code{default_memory_remove_breakpoint} respectively) have been
3402 provided so that it is not necessary to set these for most
3403 architectures. Architectures which may want to set
3404 @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
3405 conventional manner.
3407 It may also be desirable (from an efficiency standpoint) to define
3408 custom breakpoint insertion and removal routines if
3409 @code{gdbarch_breakpoint_from_pc} needs to read the target's memory for some
3412 @item CORE_ADDR gdbarch_adjust_breakpoint_address (@var{gdbarch}, @var{bpaddr})
3413 @findex gdbarch_adjust_breakpoint_address
3414 @cindex breakpoint address adjusted
3415 Given an address at which a breakpoint is desired, return a breakpoint
3416 address adjusted to account for architectural constraints on
3417 breakpoint placement. This method is not needed by most targets.
3419 The FR-V target (see @file{frv-tdep.c}) requires this method.
3420 The FR-V is a VLIW architecture in which a number of RISC-like
3421 instructions are grouped (packed) together into an aggregate
3422 instruction or instruction bundle. When the processor executes
3423 one of these bundles, the component instructions are executed
3426 In the course of optimization, the compiler may group instructions
3427 from distinct source statements into the same bundle. The line number
3428 information associated with one of the latter statements will likely
3429 refer to some instruction other than the first one in the bundle. So,
3430 if the user attempts to place a breakpoint on one of these latter
3431 statements, @value{GDBN} must be careful to @emph{not} place the break
3432 instruction on any instruction other than the first one in the bundle.
3433 (Remember though that the instructions within a bundle execute
3434 in parallel, so the @emph{first} instruction is the instruction
3435 at the lowest address and has nothing to do with execution order.)
3437 The FR-V's @code{gdbarch_adjust_breakpoint_address} method will adjust a
3438 breakpoint's address by scanning backwards for the beginning of
3439 the bundle, returning the address of the bundle.
3441 Since the adjustment of a breakpoint may significantly alter a user's
3442 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
3443 is initially set and each time that that breakpoint is hit.
3445 @item int gdbarch_call_dummy_location (@var{gdbarch})
3446 @findex gdbarch_call_dummy_location
3447 See the file @file{inferior.h}.
3449 This method has been replaced by @code{gdbarch_push_dummy_code}
3450 (@pxref{gdbarch_push_dummy_code}).
3452 @item int gdbarch_cannot_fetch_register (@var{gdbarch}, @var{regum})
3453 @findex gdbarch_cannot_fetch_register
3454 This function should return nonzero if @var{regno} cannot be fetched
3455 from an inferior process. This is only relevant if
3456 @code{FETCH_INFERIOR_REGISTERS} is not defined.
3458 @item int gdbarch_cannot_store_register (@var{gdbarch}, @var{regnum})
3459 @findex gdbarch_cannot_store_register
3460 This function should return nonzero if @var{regno} should not be
3461 written to the target. This is often the case for program counters,
3462 status words, and other special registers. This function returns 0 as
3463 default so that @value{GDBN} will assume that all registers may be written.
3465 @item int gdbarch_convert_register_p (@var{gdbarch}, @var{regnum}, struct type *@var{type})
3466 @findex gdbarch_convert_register_p
3467 Return non-zero if register @var{regnum} represents data values of type
3468 @var{type} in a non-standard form.
3469 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3471 @item CORE_ADDR gdbarch_decr_pc_after_break (@var{gdbarch})
3472 @findex gdbarch_decr_pc_after_break
3473 This function shall return the amount by which to decrement the PC after the
3474 program encounters a breakpoint. This is often the number of bytes in
3475 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
3477 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
3478 @findex DISABLE_UNSETTABLE_BREAK
3479 If defined, this should evaluate to 1 if @var{addr} is in a shared
3480 library in which breakpoints cannot be set and so should be disabled.
3482 @item void gdbarch_print_float_info (@var{gdbarch}, @var{file}, @var{frame}, @var{args})
3483 @findex gdbarch_print_float_info
3484 If defined, then the @samp{info float} command will print information about
3485 the processor's floating point unit.
3487 @item void gdbarch_print_registers_info (@var{gdbarch}, @var{frame}, @var{regnum}, @var{all})
3488 @findex gdbarch_print_registers_info
3489 If defined, pretty print the value of the register @var{regnum} for the
3490 specified @var{frame}. If the value of @var{regnum} is -1, pretty print
3491 either all registers (@var{all} is non zero) or a select subset of
3492 registers (@var{all} is zero).
3494 The default method prints one register per line, and if @var{all} is
3495 zero omits floating-point registers.
3497 @item int gdbarch_print_vector_info (@var{gdbarch}, @var{file}, @var{frame}, @var{args})
3498 @findex gdbarch_print_vector_info
3499 If defined, then the @samp{info vector} command will call this function
3500 to print information about the processor's vector unit.
3502 By default, the @samp{info vector} command will print all vector
3503 registers (the register's type having the vector attribute).
3505 @item int gdbarch_dwarf_reg_to_regnum (@var{gdbarch}, @var{dwarf_regnr})
3506 @findex gdbarch_dwarf_reg_to_regnum
3507 Convert DWARF register number @var{dwarf_regnr} into @value{GDBN} regnum. If
3508 not defined, no conversion will be performed.
3510 @item int gdbarch_dwarf2_reg_to_regnum (@var{gdbarch}, @var{dwarf2_regnr})
3511 @findex gdbarch_dwarf2_reg_to_regnum
3512 Convert DWARF2 register number @var{dwarf2_regnr} into @value{GDBN} regnum.
3513 If not defined, no conversion will be performed.
3515 @item int gdbarch_ecoff_reg_to_regnum (@var{gdbarch}, @var{ecoff_regnr})
3516 @findex gdbarch_ecoff_reg_to_regnum
3517 Convert ECOFF register number @var{ecoff_regnr} into @value{GDBN} regnum. If
3518 not defined, no conversion will be performed.
3520 @item DEPRECATED_FP_REGNUM
3521 @findex DEPRECATED_FP_REGNUM
3522 If the virtual frame pointer is kept in a register, then define this
3523 macro to be the number (greater than or equal to zero) of that register.
3525 This should only need to be defined if @code{DEPRECATED_TARGET_READ_FP}
3528 @item DEPRECATED_FRAMELESS_FUNCTION_INVOCATION(@var{fi})
3529 @findex DEPRECATED_FRAMELESS_FUNCTION_INVOCATION
3530 Define this to an expression that returns 1 if the function invocation
3531 represented by @var{fi} does not have a stack frame associated with it.
3534 @item CORE_ADDR frame_align (@var{gdbarch}, @var{address})
3535 @anchor{frame_align}
3537 Define this to adjust @var{address} so that it meets the alignment
3538 requirements for the start of a new stack frame. A stack frame's
3539 alignment requirements are typically stronger than a target processors
3540 stack alignment requirements.
3542 This function is used to ensure that, when creating a dummy frame, both
3543 the initial stack pointer and (if needed) the address of the return
3544 value are correctly aligned.
3546 This function always adjusts the address in the direction of stack
3549 By default, no frame based stack alignment is performed.
3551 @item int gdbarch_frame_red_zone_size (@var{gdbarch})
3552 @findex gdbarch_frame_red_zone_size
3553 The number of bytes, beyond the innermost-stack-address, reserved by the
3554 @sc{abi}. A function is permitted to use this scratch area (instead of
3555 allocating extra stack space).
3557 When performing an inferior function call, to ensure that it does not
3558 modify this area, @value{GDBN} adjusts the innermost-stack-address by
3559 @var{gdbarch_frame_red_zone_size} bytes before pushing parameters onto the
3562 By default, zero bytes are allocated. The value must be aligned
3563 (@pxref{frame_align}).
3565 The @sc{amd64} (nee x86-64) @sc{abi} documentation refers to the
3566 @emph{red zone} when describing this scratch area.
3569 @item DEPRECATED_FRAME_CHAIN(@var{frame})
3570 @findex DEPRECATED_FRAME_CHAIN
3571 Given @var{frame}, return a pointer to the calling frame.
3573 @item DEPRECATED_FRAME_CHAIN_VALID(@var{chain}, @var{thisframe})
3574 @findex DEPRECATED_FRAME_CHAIN_VALID
3575 Define this to be an expression that returns zero if the given frame is an
3576 outermost frame, with no caller, and nonzero otherwise. Most normal
3577 situations can be handled without defining this macro, including @code{NULL}
3578 chain pointers, dummy frames, and frames whose PC values are inside the
3579 startup file (e.g.@: @file{crt0.o}), inside @code{main}, or inside
3582 @item DEPRECATED_FRAME_INIT_SAVED_REGS(@var{frame})
3583 @findex DEPRECATED_FRAME_INIT_SAVED_REGS
3584 See @file{frame.h}. Determines the address of all registers in the
3585 current stack frame storing each in @code{frame->saved_regs}. Space for
3586 @code{frame->saved_regs} shall be allocated by
3587 @code{DEPRECATED_FRAME_INIT_SAVED_REGS} using
3588 @code{frame_saved_regs_zalloc}.
3590 @code{FRAME_FIND_SAVED_REGS} is deprecated.
3592 @item int gdbarch_frame_num_args (@var{gdbarch}, @var{frame})
3593 @findex gdbarch_frame_num_args
3594 For the frame described by @var{frame} return the number of arguments that
3595 are being passed. If the number of arguments is not known, return
3598 @item DEPRECATED_FRAME_SAVED_PC(@var{frame})
3599 @findex DEPRECATED_FRAME_SAVED_PC
3600 @anchor{DEPRECATED_FRAME_SAVED_PC} Given @var{frame}, return the pc
3601 saved there. This is the return address.
3603 This method is deprecated. @xref{gdbarch_unwind_pc}.
3605 @item CORE_ADDR gdbarch_unwind_pc (@var{next_frame})
3606 @findex gdbarch_unwind_pc
3607 @anchor{gdbarch_unwind_pc} Return the instruction address, in
3608 @var{next_frame}'s caller, at which execution will resume after
3609 @var{next_frame} returns. This is commonly referred to as the return address.
3611 The implementation, which must be frame agnostic (work with any frame),
3612 is typically no more than:
3616 pc = frame_unwind_register_unsigned (next_frame, S390_PC_REGNUM);
3617 return gdbarch_addr_bits_remove (gdbarch, pc);
3621 @xref{DEPRECATED_FRAME_SAVED_PC}, which this method replaces.
3623 @item CORE_ADDR gdbarch_unwind_sp (@var{gdbarch}, @var{next_frame})
3624 @findex gdbarch_unwind_sp
3625 @anchor{gdbarch_unwind_sp} Return the frame's inner most stack address. This is
3626 commonly referred to as the frame's @dfn{stack pointer}.
3628 The implementation, which must be frame agnostic (work with any frame),
3629 is typically no more than:
3633 sp = frame_unwind_register_unsigned (next_frame, S390_SP_REGNUM);
3634 return gdbarch_addr_bits_remove (gdbarch, sp);
3638 @xref{TARGET_READ_SP}, which this method replaces.
3640 @item FUNCTION_EPILOGUE_SIZE
3641 @findex FUNCTION_EPILOGUE_SIZE
3642 For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the
3643 function end symbol is 0. For such targets, you must define
3644 @code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size of a
3645 function's epilogue.
3647 @item DEPRECATED_FUNCTION_START_OFFSET
3648 @findex DEPRECATED_FUNCTION_START_OFFSET
3649 An integer, giving the offset in bytes from a function's address (as
3650 used in the values of symbols, function pointers, etc.), and the
3651 function's first genuine instruction.
3653 This is zero on almost all machines: the function's address is usually
3654 the address of its first instruction. However, on the VAX, for
3655 example, each function starts with two bytes containing a bitmask
3656 indicating which registers to save upon entry to the function. The
3657 VAX @code{call} instructions check this value, and save the
3658 appropriate registers automatically. Thus, since the offset from the
3659 function's address to its first instruction is two bytes,
3660 @code{DEPRECATED_FUNCTION_START_OFFSET} would be 2 on the VAX.
3662 @item GCC_COMPILED_FLAG_SYMBOL
3663 @itemx GCC2_COMPILED_FLAG_SYMBOL
3664 @findex GCC2_COMPILED_FLAG_SYMBOL
3665 @findex GCC_COMPILED_FLAG_SYMBOL
3666 If defined, these are the names of the symbols that @value{GDBN} will
3667 look for to detect that GCC compiled the file. The default symbols
3668 are @code{gcc_compiled.} and @code{gcc2_compiled.},
3669 respectively. (Currently only defined for the Delta 68.)
3671 @item gdbarch_get_longjmp_target
3672 @findex gdbarch_get_longjmp_target
3673 For most machines, this is a target-dependent parameter. On the
3674 DECstation and the Iris, this is a native-dependent parameter, since
3675 the header file @file{setjmp.h} is needed to define it.
3677 This macro determines the target PC address that @code{longjmp} will jump to,
3678 assuming that we have just stopped at a @code{longjmp} breakpoint. It takes a
3679 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3680 pointer. It examines the current state of the machine as needed.
3682 @item DEPRECATED_IBM6000_TARGET
3683 @findex DEPRECATED_IBM6000_TARGET
3684 Shows that we are configured for an IBM RS/6000 system. This
3685 conditional should be eliminated (FIXME) and replaced by
3686 feature-specific macros. It was introduced in a haste and we are
3687 repenting at leisure.
3689 @item I386_USE_GENERIC_WATCHPOINTS
3690 An x86-based target can define this to use the generic x86 watchpoint
3691 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3693 @item int gdbarch_inner_than (@var{gdbarch}, @var{lhs}, @var{rhs})
3694 @findex gdbarch_inner_than
3695 Returns non-zero if stack address @var{lhs} is inner than (nearer to the
3696 stack top) stack address @var{rhs}. Let the function return
3697 @w{@code{lhs < rhs}} if the target's stack grows downward in memory, or
3698 @w{@code{lhs > rsh}} if the stack grows upward.
3700 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{addr})
3701 @findex gdbarch_in_function_epilogue_p
3702 Returns non-zero if the given @var{addr} is in the epilogue of a function.
3703 The epilogue of a function is defined as the part of a function where
3704 the stack frame of the function already has been destroyed up to the
3705 final `return from function call' instruction.
3707 @item int gdbarch_in_solib_return_trampoline (@var{gdbarch}, @var{pc}, @var{name})
3708 @findex gdbarch_in_solib_return_trampoline
3709 Define this function to return nonzero if the program is stopped in the
3710 trampoline that returns from a shared library.
3712 @item IN_SOLIB_DYNSYM_RESOLVE_CODE (@var{pc})
3713 @findex IN_SOLIB_DYNSYM_RESOLVE_CODE
3714 Define this to return nonzero if the program is stopped in the
3717 @item SKIP_SOLIB_RESOLVER (@var{pc})
3718 @findex SKIP_SOLIB_RESOLVER
3719 Define this to evaluate to the (nonzero) address at which execution
3720 should continue to get past the dynamic linker's symbol resolution
3721 function. A zero value indicates that it is not important or necessary
3722 to set a breakpoint to get through the dynamic linker and that single
3723 stepping will suffice.
3725 @item CORE_ADDR gdbarch_integer_to_address (@var{gdbarch}, @var{type}, @var{buf})
3726 @findex gdbarch_integer_to_address
3727 @cindex converting integers to addresses
3728 Define this when the architecture needs to handle non-pointer to address
3729 conversions specially. Converts that value to an address according to
3730 the current architectures conventions.
3732 @emph{Pragmatics: When the user copies a well defined expression from
3733 their source code and passes it, as a parameter, to @value{GDBN}'s
3734 @code{print} command, they should get the same value as would have been
3735 computed by the target program. Any deviation from this rule can cause
3736 major confusion and annoyance, and needs to be justified carefully. In
3737 other words, @value{GDBN} doesn't really have the freedom to do these
3738 conversions in clever and useful ways. It has, however, been pointed
3739 out that users aren't complaining about how @value{GDBN} casts integers
3740 to pointers; they are complaining that they can't take an address from a
3741 disassembly listing and give it to @code{x/i}. Adding an architecture
3742 method like @code{gdbarch_integer_to_address} certainly makes it possible for
3743 @value{GDBN} to ``get it right'' in all circumstances.}
3745 @xref{Target Architecture Definition, , Pointers Are Not Always
3748 @item CORE_ADDR gdbarch_pointer_to_address (@var{gdbarch}, @var{type}, @var{buf})
3749 @findex gdbarch_pointer_to_address
3750 Assume that @var{buf} holds a pointer of type @var{type}, in the
3751 appropriate format for the current architecture. Return the byte
3752 address the pointer refers to.
3753 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3755 @item void gdbarch_register_to_value(@var{gdbarch}, @var{frame}, @var{regnum}, @var{type}, @var{fur})
3756 @findex gdbarch_register_to_value
3757 Convert the raw contents of register @var{regnum} into a value of type
3759 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3761 @item register_reggroup_p (@var{gdbarch}, @var{regnum}, @var{reggroup})
3762 @findex register_reggroup_p
3763 @cindex register groups
3764 Return non-zero if register @var{regnum} is a member of the register
3765 group @var{reggroup}.
3767 By default, registers are grouped as follows:
3770 @item float_reggroup
3771 Any register with a valid name and a floating-point type.
3772 @item vector_reggroup
3773 Any register with a valid name and a vector type.
3774 @item general_reggroup
3775 Any register with a valid name and a type other than vector or
3776 floating-point. @samp{float_reggroup}.
3778 @itemx restore_reggroup
3780 Any register with a valid name.
3783 @item DEPRECATED_REGISTER_VIRTUAL_SIZE (@var{reg})
3784 @findex DEPRECATED_REGISTER_VIRTUAL_SIZE
3785 Return the virtual size of @var{reg}; defaults to the size of the
3786 register's virtual type.
3787 Return the virtual size of @var{reg}.
3788 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3790 @item DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})
3791 @findex REGISTER_VIRTUAL_TYPE
3792 Return the virtual type of @var{reg}.
3793 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3795 @item struct type *register_type (@var{gdbarch}, @var{reg})
3796 @findex register_type
3797 If defined, return the type of register @var{reg}. This function
3798 supersedes @code{DEPRECATED_REGISTER_VIRTUAL_TYPE}. @xref{Target Architecture
3799 Definition, , Raw and Virtual Register Representations}.
3801 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
3802 @findex REGISTER_CONVERT_TO_VIRTUAL
3803 Convert the value of register @var{reg} from its raw form to its virtual
3805 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3807 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
3808 @findex REGISTER_CONVERT_TO_RAW
3809 Convert the value of register @var{reg} from its virtual form to its raw
3811 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3813 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
3814 @findex regset_from_core_section
3815 Return the appropriate register set for a core file section with name
3816 @var{sect_name} and size @var{sect_size}.
3818 @item SOFTWARE_SINGLE_STEP_P()
3819 @findex SOFTWARE_SINGLE_STEP_P
3820 Define this as 1 if the target does not have a hardware single-step
3821 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
3823 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breakpoints_p})
3824 @findex SOFTWARE_SINGLE_STEP
3825 A function that inserts or removes (depending on
3826 @var{insert_breakpoints_p}) breakpoints at each possible destinations of
3827 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
3830 @item set_gdbarch_sofun_address_maybe_missing (@var{gdbarch}, @var{set})
3831 @findex set_gdbarch_sofun_address_maybe_missing
3832 Somebody clever observed that, the more actual addresses you have in the
3833 debug information, the more time the linker has to spend relocating
3834 them. So whenever there's some other way the debugger could find the
3835 address it needs, you should omit it from the debug info, to make
3838 Calling @code{set_gdbarch_sofun_address_maybe_missing} with a non-zero
3839 argument @var{set} indicates that a particular set of hacks of this sort
3840 are in use, affecting @code{N_SO} and @code{N_FUN} entries in stabs-format
3841 debugging information. @code{N_SO} stabs mark the beginning and ending
3842 addresses of compilation units in the text segment. @code{N_FUN} stabs
3843 mark the starts and ends of functions.
3845 In this case, @value{GDBN} assumes two things:
3849 @code{N_FUN} stabs have an address of zero. Instead of using those
3850 addresses, you should find the address where the function starts by
3851 taking the function name from the stab, and then looking that up in the
3852 minsyms (the linker/assembler symbol table). In other words, the stab
3853 has the name, and the linker/assembler symbol table is the only place
3854 that carries the address.
3857 @code{N_SO} stabs have an address of zero, too. You just look at the
3858 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab, and
3859 guess the starting and ending addresses of the compilation unit from them.
3862 @item int gdbarch_pc_regnum (@var{gdbarch})
3863 @findex gdbarch_pc_regnum
3864 If the program counter is kept in a register, then let this function return
3865 the number (greater than or equal to zero) of that register.
3867 This should only need to be defined if @code{gdbarch_read_pc} and
3868 @code{gdbarch_write_pc} are not defined.
3870 @item int gdbarch_stabs_argument_has_addr (@var{gdbarch}, @var{type})
3871 @findex gdbarch_stabs_argument_has_addr
3872 @anchor{gdbarch_stabs_argument_has_addr} Define this function to return
3873 nonzero if a function argument of type @var{type} is passed by reference
3876 @item PROCESS_LINENUMBER_HOOK
3877 @findex PROCESS_LINENUMBER_HOOK
3878 A hook defined for XCOFF reading.
3880 @item gdbarch_ps_regnum (@var{gdbarch}
3881 @findex gdbarch_ps_regnum
3882 If defined, this function returns the number of the processor status
3884 (This definition is only used in generic code when parsing "$ps".)
3886 @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})
3887 @findex gdbarch_push_dummy_call
3888 @findex DEPRECATED_PUSH_ARGUMENTS.
3889 @anchor{gdbarch_push_dummy_call} Define this to push the dummy frame's call to
3890 the inferior function onto the stack. In addition to pushing @var{nargs}, the
3891 code should push @var{struct_addr} (when @var{struct_return} is non-zero), and
3892 the return address (@var{bp_addr}).
3894 @var{function} is a pointer to a @code{struct value}; on architectures that use
3895 function descriptors, this contains the function descriptor value.
3897 Returns the updated top-of-stack pointer.
3899 This method replaces @code{DEPRECATED_PUSH_ARGUMENTS}.
3901 @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})
3902 @findex gdbarch_push_dummy_code
3903 @anchor{gdbarch_push_dummy_code} Given a stack based call dummy, push the
3904 instruction sequence (including space for a breakpoint) to which the
3905 called function should return.
3907 Set @var{bp_addr} to the address at which the breakpoint instruction
3908 should be inserted, @var{real_pc} to the resume address when starting
3909 the call sequence, and return the updated inner-most stack address.
3911 By default, the stack is grown sufficient to hold a frame-aligned
3912 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
3913 reserved for that breakpoint, and @var{real_pc} set to @var{funaddr}.
3915 This method replaces @w{@code{gdbarch_call_dummy_location (@var{gdbarch})}} and
3916 @code{DEPRECATED_REGISTER_SIZE}.
3918 @item const char *gdbarch_register_name (@var{gdbarch}, @var{regnr})
3919 @findex gdbarch_register_name
3920 Return the name of register @var{regnr} as a string. May return @code{NULL}
3921 to indicate that @var{regnr} is not a valid register.
3923 @item SAVE_DUMMY_FRAME_TOS (@var{sp})
3924 @findex SAVE_DUMMY_FRAME_TOS
3925 @anchor{SAVE_DUMMY_FRAME_TOS} Used in @samp{call_function_by_hand} to
3926 notify the target dependent code of the top-of-stack value that will be
3927 passed to the inferior code. This is the value of the @code{SP}
3928 after both the dummy frame and space for parameters/results have been
3929 allocated on the stack. @xref{gdbarch_unwind_dummy_id}.
3931 @item int gdbarch_sdb_reg_to_regnum (@var{gdbarch}, @var{sdb_regnr})
3932 @findex gdbarch_sdb_reg_to_regnum
3933 Use this function to convert sdb register @var{sdb_regnr} into @value{GDBN}
3934 regnum. If not defined, no conversion will be done.
3936 @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})
3937 @findex gdbarch_return_value
3938 @anchor{gdbarch_return_value} Given a function with a return-value of
3939 type @var{rettype}, return which return-value convention that function
3942 @value{GDBN} currently recognizes two function return-value conventions:
3943 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
3944 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
3945 value is found in memory and the address of that memory location is
3946 passed in as the function's first parameter.
3948 If the register convention is being used, and @var{writebuf} is
3949 non-@code{NULL}, also copy the return-value in @var{writebuf} into
3952 If the register convention is being used, and @var{readbuf} is
3953 non-@code{NULL}, also copy the return value from @var{regcache} into
3954 @var{readbuf} (@var{regcache} contains a copy of the registers from the
3955 just returned function).
3957 @emph{Maintainer note: This method replaces separate predicate, extract,
3958 store methods. By having only one method, the logic needed to determine
3959 the return-value convention need only be implemented in one place. If
3960 @value{GDBN} were written in an @sc{oo} language, this method would
3961 instead return an object that knew how to perform the register
3962 return-value extract and store.}
3964 @emph{Maintainer note: This method does not take a @var{gcc_p}
3965 parameter, and such a parameter should not be added. If an architecture
3966 that requires per-compiler or per-function information be identified,
3967 then the replacement of @var{rettype} with @code{struct value}
3968 @var{function} should be pursued.}
3970 @emph{Maintainer note: The @var{regcache} parameter limits this methods
3971 to the inner most frame. While replacing @var{regcache} with a
3972 @code{struct frame_info} @var{frame} parameter would remove that
3973 limitation there has yet to be a demonstrated need for such a change.}
3975 @item void gdbarch_skip_permanent_breakpoint (@var{gdbarch}, @var{regcache})
3976 @findex gdbarch_skip_permanent_breakpoint
3977 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
3978 steps over a breakpoint by removing it, stepping one instruction, and
3979 re-inserting the breakpoint. However, permanent breakpoints are
3980 hardwired into the inferior, and can't be removed, so this strategy
3981 doesn't work. Calling @code{gdbarch_skip_permanent_breakpoint} adjusts the
3982 processor's state so that execution will resume just after the breakpoint.
3983 This function does the right thing even when the breakpoint is in the delay slot
3984 of a branch or jump.
3986 @item CORE_ADDR gdbarch_skip_prologue (@var{gdbarch}, @var{ip})
3987 @findex gdbarch_skip_prologue
3988 A function that returns the address of the ``real'' code beyond the
3989 function entry prologue found at @var{ip}.
3991 @item CORE_ADDR gdbarch_skip_trampoline_code (@var{gdbarch}, @var{frame}, @var{pc})
3992 @findex gdbarch_skip_trampoline_code
3993 If the target machine has trampoline code that sits between callers and
3994 the functions being called, then define this function to return a new PC
3995 that is at the start of the real function.
3997 @item int gdbarch_sp_regnum (@var{gdbarch})
3998 @findex gdbarch_sp_regnum
3999 If the stack-pointer is kept in a register, then use this function to return
4000 the number (greater than or equal to zero) of that register, or -1 if
4001 there is no such register.
4003 @item int gdbarch_stab_reg_to_regnum (@var{gdbarch}, @var{stab_regnr})
4004 @findex gdbarch_stab_reg_to_regnum
4005 Use this function to convert stab register @var{stab_regnr} into @value{GDBN}
4006 regnum. If not defined, no conversion will be done.
4008 @item SYMBOL_RELOADING_DEFAULT
4009 @findex SYMBOL_RELOADING_DEFAULT
4010 The default value of the ``symbol-reloading'' variable. (Never defined in
4013 @item TARGET_CHAR_BIT
4014 @findex TARGET_CHAR_BIT
4015 Number of bits in a char; defaults to 8.
4017 @item int gdbarch_char_signed (@var{gdbarch})
4018 @findex gdbarch_char_signed
4019 Non-zero if @code{char} is normally signed on this architecture; zero if
4020 it should be unsigned.
4022 The ISO C standard requires the compiler to treat @code{char} as
4023 equivalent to either @code{signed char} or @code{unsigned char}; any
4024 character in the standard execution set is supposed to be positive.
4025 Most compilers treat @code{char} as signed, but @code{char} is unsigned
4026 on the IBM S/390, RS6000, and PowerPC targets.
4028 @item int gdbarch_double_bit (@var{gdbarch})
4029 @findex gdbarch_double_bit
4030 Number of bits in a double float; defaults to @w{@code{8 * TARGET_CHAR_BIT}}.
4032 @item int gdbarch_float_bit (@var{gdbarch})
4033 @findex gdbarch_float_bit
4034 Number of bits in a float; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4036 @item int gdbarch_int_bit (@var{gdbarch})
4037 @findex gdbarch_int_bit
4038 Number of bits in an integer; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4040 @item int gdbarch_long_bit (@var{gdbarch})
4041 @findex gdbarch_long_bit
4042 Number of bits in a long integer; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4044 @item int gdbarch_long_double_bit (@var{gdbarch})
4045 @findex gdbarch_long_double_bit
4046 Number of bits in a long double float;
4047 defaults to @w{@code{2 * gdbarch_double_bit (@var{gdbarch})}}.
4049 @item int gdbarch_long_long_bit (@var{gdbarch})
4050 @findex gdbarch_long_long_bit
4051 Number of bits in a long long integer; defaults to
4052 @w{@code{2 * gdbarch_long_bit (@var{gdbarch})}}.
4054 @item int gdbarch_ptr_bit (@var{gdbarch})
4055 @findex gdbarch_ptr_bit
4056 Number of bits in a pointer; defaults to
4057 @w{@code{gdbarch_int_bit (@var{gdbarch})}}.
4059 @item int gdbarch_short_bit (@var{gdbarch})
4060 @findex gdbarch_short_bit
4061 Number of bits in a short integer; defaults to @w{@code{2 * TARGET_CHAR_BIT}}.
4063 @item CORE_ADDR gdbarch_read_pc (@var{gdbarch}, @var{regcache})
4064 @findex gdbarch_read_pc
4065 @itemx gdbarch_write_pc (@var{gdbarch}, @var{regcache}, @var{val})
4066 @findex gdbarch_write_pc
4067 @anchor{gdbarch_write_pc}
4068 @itemx TARGET_READ_SP
4069 @findex TARGET_READ_SP
4070 @itemx TARGET_READ_FP
4071 @findex TARGET_READ_FP
4072 @findex gdbarch_read_pc
4073 @findex gdbarch_write_pc
4076 @anchor{TARGET_READ_SP} These change the behavior of @code{gdbarch_read_pc},
4077 @code{gdbarch_write_pc}, and @code{read_sp}. For most targets, these may be
4078 left undefined. @value{GDBN} will call the read and write register
4079 functions with the relevant @code{_REGNUM} argument.
4081 These macros and functions are useful when a target keeps one of these
4082 registers in a hard to get at place; for example, part in a segment register
4083 and part in an ordinary register.
4085 @xref{gdbarch_unwind_sp}, which replaces @code{TARGET_READ_SP}.
4087 @item void gdbarch_virtual_frame_pointer (@var{gdbarch}, @var{pc}, @var{frame_regnum}, @var{frame_offset})
4088 @findex gdbarch_virtual_frame_pointer
4089 Returns a @code{(register, offset)} pair representing the virtual frame
4090 pointer in use at the code address @var{pc}. If virtual frame pointers
4091 are not used, a default definition simply returns
4092 @code{DEPRECATED_FP_REGNUM}, with an offset of zero.
4094 @item TARGET_HAS_HARDWARE_WATCHPOINTS
4095 If non-zero, the target has support for hardware-assisted
4096 watchpoints. @xref{Algorithms, watchpoints}, for more details and
4097 other related macros.
4099 @item int gdbarch_print_insn (@var{gdbarch}, @var{vma}, @var{info})
4100 @findex gdbarch_print_insn
4101 This is the function used by @value{GDBN} to print an assembly
4102 instruction. It prints the instruction at address @var{vma} in
4103 debugged memory and returns the length of the instruction, in bytes. If
4104 a target doesn't define its own printing routine, it defaults to an
4105 accessor function for the global pointer
4106 @code{deprecated_tm_print_insn}. This usually points to a function in
4107 the @code{opcodes} library (@pxref{Support Libraries, ,Opcodes}).
4108 @var{info} is a structure (of type @code{disassemble_info}) defined in
4109 @file{include/dis-asm.h} used to pass information to the instruction
4112 @item frame_id gdbarch_unwind_dummy_id (@var{gdbarch}, @var{frame})
4113 @findex gdbarch_unwind_dummy_id
4114 @anchor{gdbarch_unwind_dummy_id} Given @var{frame} return a @w{@code{struct
4115 frame_id}} that uniquely identifies an inferior function call's dummy
4116 frame. The value returned must match the dummy frame stack value
4117 previously saved using @code{SAVE_DUMMY_FRAME_TOS}.
4118 @xref{SAVE_DUMMY_FRAME_TOS}.
4120 @item DEPRECATED_USE_STRUCT_CONVENTION (@var{gcc_p}, @var{type})
4121 @findex DEPRECATED_USE_STRUCT_CONVENTION
4122 If defined, this must be an expression that is nonzero if a value of the
4123 given @var{type} being returned from a function must have space
4124 allocated for it on the stack. @var{gcc_p} is true if the function
4125 being considered is known to have been compiled by GCC; this is helpful
4126 for systems where GCC is known to use different calling convention than
4129 This method has been deprecated in favour of @code{gdbarch_return_value}
4130 (@pxref{gdbarch_return_value}).
4132 @item void gdbarch_value_to_register (@var{gdbarch}, @var{frame}, @var{type}, @var{buf})
4133 @findex gdbarch_value_to_register
4134 Convert a value of type @var{type} into the raw contents of a register.
4135 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
4139 Motorola M68K target conditionals.
4143 Define this to be the 4-bit location of the breakpoint trap vector. If
4144 not defined, it will default to @code{0xf}.
4146 @item REMOTE_BPT_VECTOR
4147 Defaults to @code{1}.
4149 @item const char *gdbarch_name_of_malloc (@var{gdbarch})
4150 @findex gdbarch_name_of_malloc
4151 A string containing the name of the function to call in order to
4152 allocate some memory in the inferior. The default value is "malloc".
4156 @node Adding a New Target
4157 @section Adding a New Target
4159 @cindex adding a target
4160 The following files add a target to @value{GDBN}:
4164 @item gdb/config/@var{arch}/@var{ttt}.mt
4165 Contains a Makefile fragment specific to this target. Specifies what
4166 object files are needed for target @var{ttt}, by defining
4167 @samp{TDEPFILES=@dots{}} and @samp{TDEPLIBS=@dots{}}. Also specifies
4168 the header file which describes @var{ttt}, by defining @samp{TM_FILE=
4171 You can also define @samp{TM_CFLAGS}, @samp{TM_CLIBS}, @samp{TM_CDEPS},
4172 but these are now deprecated, replaced by autoconf, and may go away in
4173 future versions of @value{GDBN}.
4175 @item gdb/@var{ttt}-tdep.c
4176 Contains any miscellaneous code required for this target machine. On
4177 some machines it doesn't exist at all. Sometimes the macros in
4178 @file{tm-@var{ttt}.h} become very complicated, so they are implemented
4179 as functions here instead, and the macro is simply defined to call the
4180 function. This is vastly preferable, since it is easier to understand
4183 @item gdb/@var{arch}-tdep.c
4184 @itemx gdb/@var{arch}-tdep.h
4185 This often exists to describe the basic layout of the target machine's
4186 processor chip (registers, stack, etc.). If used, it is included by
4187 @file{@var{ttt}-tdep.h}. It can be shared among many targets that use
4190 @item gdb/config/@var{arch}/tm-@var{ttt}.h
4191 (@file{tm.h} is a link to this file, created by @code{configure}). Contains
4192 macro definitions about the target machine's registers, stack frame
4193 format and instructions.
4195 New targets do not need this file and should not create it.
4197 @item gdb/config/@var{arch}/tm-@var{arch}.h
4198 This often exists to describe the basic layout of the target machine's
4199 processor chip (registers, stack, etc.). If used, it is included by
4200 @file{tm-@var{ttt}.h}. It can be shared among many targets that use the
4203 New targets do not need this file and should not create it.
4207 If you are adding a new operating system for an existing CPU chip, add a
4208 @file{config/tm-@var{os}.h} file that describes the operating system
4209 facilities that are unusual (extra symbol table info; the breakpoint
4210 instruction needed; etc.). Then write a @file{@var{arch}/tm-@var{os}.h}
4211 that just @code{#include}s @file{tm-@var{arch}.h} and
4212 @file{config/tm-@var{os}.h}.
4214 @node Target Descriptions
4215 @chapter Target Descriptions
4216 @cindex target descriptions
4218 The target architecture definition (@pxref{Target Architecture Definition})
4219 contains @value{GDBN}'s hard-coded knowledge about an architecture. For
4220 some platforms, it is handy to have more flexible knowledge about a specific
4221 instance of the architecture---for instance, a processor or development board.
4222 @dfn{Target descriptions} provide a mechanism for the user to tell @value{GDBN}
4223 more about what their target supports, or for the target to tell @value{GDBN}
4226 For details on writing, automatically supplying, and manually selecting
4227 target descriptions, see @ref{Target Descriptions, , , gdb,
4228 Debugging with @value{GDBN}}. This section will cover some related
4229 topics about the @value{GDBN} internals.
4232 * Target Descriptions Implementation::
4233 * Adding Target Described Register Support::
4236 @node Target Descriptions Implementation
4237 @section Target Descriptions Implementation
4238 @cindex target descriptions, implementation
4240 Before @value{GDBN} connects to a new target, or runs a new program on
4241 an existing target, it discards any existing target description and
4242 reverts to a default gdbarch. Then, after connecting, it looks for a
4243 new target description by calling @code{target_find_description}.
4245 A description may come from a user specified file (XML), the remote
4246 @samp{qXfer:features:read} packet (also XML), or from any custom
4247 @code{to_read_description} routine in the target vector. For instance,
4248 the remote target supports guessing whether a MIPS target is 32-bit or
4249 64-bit based on the size of the @samp{g} packet.
4251 If any target description is found, @value{GDBN} creates a new gdbarch
4252 incorporating the description by calling @code{gdbarch_update_p}. Any
4253 @samp{<architecture>} element is handled first, to determine which
4254 architecture's gdbarch initialization routine is called to create the
4255 new architecture. Then the initialization routine is called, and has
4256 a chance to adjust the constructed architecture based on the contents
4257 of the target description. For instance, it can recognize any
4258 properties set by a @code{to_read_description} routine. Also
4259 see @ref{Adding Target Described Register Support}.
4261 @node Adding Target Described Register Support
4262 @section Adding Target Described Register Support
4263 @cindex target descriptions, adding register support
4265 Target descriptions can report additional registers specific to an
4266 instance of the target. But it takes a little work in the architecture
4267 specific routines to support this.
4269 A target description must either have no registers or a complete
4270 set---this avoids complexity in trying to merge standard registers
4271 with the target defined registers. It is the architecture's
4272 responsibility to validate that a description with registers has
4273 everything it needs. To keep architecture code simple, the same
4274 mechanism is used to assign fixed internal register numbers to
4277 If @code{tdesc_has_registers} returns 1, the description contains
4278 registers. The architecture's @code{gdbarch_init} routine should:
4283 Call @code{tdesc_data_alloc} to allocate storage, early, before
4284 searching for a matching gdbarch or allocating a new one.
4287 Use @code{tdesc_find_feature} to locate standard features by name.
4290 Use @code{tdesc_numbered_register} and @code{tdesc_numbered_register_choices}
4291 to locate the expected registers in the standard features.
4294 Return @code{NULL} if a required feature is missing, or if any standard
4295 feature is missing expected registers. This will produce a warning that
4296 the description was incomplete.
4299 Free the allocated data before returning, unless @code{tdesc_use_registers}
4303 Call @code{set_gdbarch_num_regs} as usual, with a number higher than any
4304 fixed number passed to @code{tdesc_numbered_register}.
4307 Call @code{tdesc_use_registers} after creating a new gdbarch, before
4312 After @code{tdesc_use_registers} has been called, the architecture's
4313 @code{register_name}, @code{register_type}, and @code{register_reggroup_p}
4314 routines will not be called; that information will be taken from
4315 the target description. @code{num_regs} may be increased to account
4316 for any additional registers in the description.
4318 Pseudo-registers require some extra care:
4323 Using @code{tdesc_numbered_register} allows the architecture to give
4324 constant register numbers to standard architectural registers, e.g.@:
4325 as an @code{enum} in @file{@var{arch}-tdep.h}. But because
4326 pseudo-registers are always numbered above @code{num_regs},
4327 which may be increased by the description, constant numbers
4328 can not be used for pseudos. They must be numbered relative to
4329 @code{num_regs} instead.
4332 The description will not describe pseudo-registers, so the
4333 architecture must call @code{set_tdesc_pseudo_register_name},
4334 @code{set_tdesc_pseudo_register_type}, and
4335 @code{set_tdesc_pseudo_register_reggroup_p} to supply routines
4336 describing pseudo registers. These routines will be passed
4337 internal register numbers, so the same routines used for the
4338 gdbarch equivalents are usually suitable.
4343 @node Target Vector Definition
4345 @chapter Target Vector Definition
4346 @cindex target vector
4348 The target vector defines the interface between @value{GDBN}'s
4349 abstract handling of target systems, and the nitty-gritty code that
4350 actually exercises control over a process or a serial port.
4351 @value{GDBN} includes some 30-40 different target vectors; however,
4352 each configuration of @value{GDBN} includes only a few of them.
4355 * Managing Execution State::
4356 * Existing Targets::
4359 @node Managing Execution State
4360 @section Managing Execution State
4361 @cindex execution state
4363 A target vector can be completely inactive (not pushed on the target
4364 stack), active but not running (pushed, but not connected to a fully
4365 manifested inferior), or completely active (pushed, with an accessible
4366 inferior). Most targets are only completely inactive or completely
4367 active, but some support persistent connections to a target even
4368 when the target has exited or not yet started.
4370 For example, connecting to the simulator using @code{target sim} does
4371 not create a running program. Neither registers nor memory are
4372 accessible until @code{run}. Similarly, after @code{kill}, the
4373 program can not continue executing. But in both cases @value{GDBN}
4374 remains connected to the simulator, and target-specific commands
4375 are directed to the simulator.
4377 A target which only supports complete activation should push itself
4378 onto the stack in its @code{to_open} routine (by calling
4379 @code{push_target}), and unpush itself from the stack in its
4380 @code{to_mourn_inferior} routine (by calling @code{unpush_target}).
4382 A target which supports both partial and complete activation should
4383 still call @code{push_target} in @code{to_open}, but not call
4384 @code{unpush_target} in @code{to_mourn_inferior}. Instead, it should
4385 call either @code{target_mark_running} or @code{target_mark_exited}
4386 in its @code{to_open}, depending on whether the target is fully active
4387 after connection. It should also call @code{target_mark_running} any
4388 time the inferior becomes fully active (e.g.@: in
4389 @code{to_create_inferior} and @code{to_attach}), and
4390 @code{target_mark_exited} when the inferior becomes inactive (in
4391 @code{to_mourn_inferior}). The target should also make sure to call
4392 @code{target_mourn_inferior} from its @code{to_kill}, to return the
4393 target to inactive state.
4395 @node Existing Targets
4396 @section Existing Targets
4399 @subsection File Targets
4401 Both executables and core files have target vectors.
4403 @subsection Standard Protocol and Remote Stubs
4405 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code
4406 that runs in the target system. @value{GDBN} provides several sample
4407 @dfn{stubs} that can be integrated into target programs or operating
4408 systems for this purpose; they are named @file{*-stub.c}.
4410 The @value{GDBN} user's manual describes how to put such a stub into
4411 your target code. What follows is a discussion of integrating the
4412 SPARC stub into a complicated operating system (rather than a simple
4413 program), by Stu Grossman, the author of this stub.
4415 The trap handling code in the stub assumes the following upon entry to
4420 %l1 and %l2 contain pc and npc respectively at the time of the trap;
4426 you are in the correct trap window.
4429 As long as your trap handler can guarantee those conditions, then there
4430 is no reason why you shouldn't be able to ``share'' traps with the stub.
4431 The stub has no requirement that it be jumped to directly from the
4432 hardware trap vector. That is why it calls @code{exceptionHandler()},
4433 which is provided by the external environment. For instance, this could
4434 set up the hardware traps to actually execute code which calls the stub
4435 first, and then transfers to its own trap handler.
4437 For the most point, there probably won't be much of an issue with
4438 ``sharing'' traps, as the traps we use are usually not used by the kernel,
4439 and often indicate unrecoverable error conditions. Anyway, this is all
4440 controlled by a table, and is trivial to modify. The most important
4441 trap for us is for @code{ta 1}. Without that, we can't single step or
4442 do breakpoints. Everything else is unnecessary for the proper operation
4443 of the debugger/stub.
4445 From reading the stub, it's probably not obvious how breakpoints work.
4446 They are simply done by deposit/examine operations from @value{GDBN}.
4448 @subsection ROM Monitor Interface
4450 @subsection Custom Protocols
4452 @subsection Transport Layer
4454 @subsection Builtin Simulator
4457 @node Native Debugging
4459 @chapter Native Debugging
4460 @cindex native debugging
4462 Several files control @value{GDBN}'s configuration for native support:
4466 @item gdb/config/@var{arch}/@var{xyz}.mh
4467 Specifies Makefile fragments needed by a @emph{native} configuration on
4468 machine @var{xyz}. In particular, this lists the required
4469 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
4470 Also specifies the header file which describes native support on
4471 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
4472 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
4473 @samp{NAT_CDEPS}, etc.; see @file{Makefile.in}.
4475 @emph{Maintainer's note: The @file{.mh} suffix is because this file
4476 originally contained @file{Makefile} fragments for hosting @value{GDBN}
4477 on machine @var{xyz}. While the file is no longer used for this
4478 purpose, the @file{.mh} suffix remains. Perhaps someone will
4479 eventually rename these fragments so that they have a @file{.mn}
4482 @item gdb/config/@var{arch}/nm-@var{xyz}.h
4483 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
4484 macro definitions describing the native system environment, such as
4485 child process control and core file support.
4487 @item gdb/@var{xyz}-nat.c
4488 Contains any miscellaneous C code required for this native support of
4489 this machine. On some machines it doesn't exist at all.
4492 There are some ``generic'' versions of routines that can be used by
4493 various systems. These can be customized in various ways by macros
4494 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
4495 the @var{xyz} host, you can just include the generic file's name (with
4496 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
4498 Otherwise, if your machine needs custom support routines, you will need
4499 to write routines that perform the same functions as the generic file.
4500 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
4501 into @code{NATDEPFILES}.
4505 This contains the @emph{target_ops vector} that supports Unix child
4506 processes on systems which use ptrace and wait to control the child.
4509 This contains the @emph{target_ops vector} that supports Unix child
4510 processes on systems which use /proc to control the child.
4513 This does the low-level grunge that uses Unix system calls to do a ``fork
4514 and exec'' to start up a child process.
4517 This is the low level interface to inferior processes for systems using
4518 the Unix @code{ptrace} call in a vanilla way.
4521 @section Native core file Support
4522 @cindex native core files
4525 @findex fetch_core_registers
4526 @item core-aout.c::fetch_core_registers()
4527 Support for reading registers out of a core file. This routine calls
4528 @code{register_addr()}, see below. Now that BFD is used to read core
4529 files, virtually all machines should use @code{core-aout.c}, and should
4530 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
4531 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
4533 @item core-aout.c::register_addr()
4534 If your @code{nm-@var{xyz}.h} file defines the macro
4535 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
4536 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
4537 register number @code{regno}. @code{blockend} is the offset within the
4538 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
4539 @file{core-aout.c} will define the @code{register_addr()} function and
4540 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
4541 you are using the standard @code{fetch_core_registers()}, you will need
4542 to define your own version of @code{register_addr()}, put it into your
4543 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
4544 the @code{NATDEPFILES} list. If you have your own
4545 @code{fetch_core_registers()}, you may not need a separate
4546 @code{register_addr()}. Many custom @code{fetch_core_registers()}
4547 implementations simply locate the registers themselves.@refill
4550 When making @value{GDBN} run native on a new operating system, to make it
4551 possible to debug core files, you will need to either write specific
4552 code for parsing your OS's core files, or customize
4553 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
4554 machine uses to define the struct of registers that is accessible
4555 (possibly in the u-area) in a core file (rather than
4556 @file{machine/reg.h}), and an include file that defines whatever header
4557 exists on a core file (e.g., the u-area or a @code{struct core}). Then
4558 modify @code{trad_unix_core_file_p} to use these values to set up the
4559 section information for the data segment, stack segment, any other
4560 segments in the core file (perhaps shared library contents or control
4561 information), ``registers'' segment, and if there are two discontiguous
4562 sets of registers (e.g., integer and float), the ``reg2'' segment. This
4563 section information basically delimits areas in the core file in a
4564 standard way, which the section-reading routines in BFD know how to seek
4567 Then back in @value{GDBN}, you need a matching routine called
4568 @code{fetch_core_registers}. If you can use the generic one, it's in
4569 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
4570 It will be passed a char pointer to the entire ``registers'' segment,
4571 its length, and a zero; or a char pointer to the entire ``regs2''
4572 segment, its length, and a 2. The routine should suck out the supplied
4573 register values and install them into @value{GDBN}'s ``registers'' array.
4575 If your system uses @file{/proc} to control processes, and uses ELF
4576 format core files, then you may be able to use the same routines for
4577 reading the registers out of processes and out of core files.
4585 @section shared libraries
4587 @section Native Conditionals
4588 @cindex native conditionals
4590 When @value{GDBN} is configured and compiled, various macros are
4591 defined or left undefined, to control compilation when the host and
4592 target systems are the same. These macros should be defined (or left
4593 undefined) in @file{nm-@var{system}.h}.
4597 @item CHILD_PREPARE_TO_STORE
4598 @findex CHILD_PREPARE_TO_STORE
4599 If the machine stores all registers at once in the child process, then
4600 define this to ensure that all values are correct. This usually entails
4601 a read from the child.
4603 [Note that this is incorrectly defined in @file{xm-@var{system}.h} files
4606 @item FETCH_INFERIOR_REGISTERS
4607 @findex FETCH_INFERIOR_REGISTERS
4608 Define this if the native-dependent code will provide its own routines
4609 @code{fetch_inferior_registers} and @code{store_inferior_registers} in
4610 @file{@var{host}-nat.c}. If this symbol is @emph{not} defined, and
4611 @file{infptrace.c} is included in this configuration, the default
4612 routines in @file{infptrace.c} are used for these functions.
4614 @item int gdbarch_fp0_regnum (@var{gdbarch})
4615 @findex gdbarch_fp0_regnum
4616 This functions normally returns the number of the first floating
4617 point register, if the machine has such registers. As such, it would
4618 appear only in target-specific code. However, @file{/proc} support uses this
4619 to decide whether floats are in use on this target.
4621 @item int gdbarch_get_longjmp_target (@var{gdbarch})
4622 @findex gdbarch_get_longjmp_target
4623 For most machines, this is a target-dependent parameter. On the
4624 DECstation and the Iris, this is a native-dependent parameter, since
4625 @file{setjmp.h} is needed to define it.
4627 This function determines the target PC address that @code{longjmp} will jump to,
4628 assuming that we have just stopped at a longjmp breakpoint. It takes a
4629 @code{CORE_ADDR *} as argument, and stores the target PC value through this
4630 pointer. It examines the current state of the machine as needed.
4632 @item I386_USE_GENERIC_WATCHPOINTS
4633 An x86-based machine can define this to use the generic x86 watchpoint
4634 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4636 @item ONE_PROCESS_WRITETEXT
4637 @findex ONE_PROCESS_WRITETEXT
4638 Define this to be able to, when a breakpoint insertion fails, warn the
4639 user that another process may be running with the same executable.
4642 @findex PROC_NAME_FMT
4643 Defines the format for the name of a @file{/proc} device. Should be
4644 defined in @file{nm.h} @emph{only} in order to override the default
4645 definition in @file{procfs.c}.
4647 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
4649 Define this to expand into an expression that will cause the symbols in
4650 @var{filename} to be added to @value{GDBN}'s symbol table. If
4651 @var{readsyms} is zero symbols are not read but any necessary low level
4652 processing for @var{filename} is still done.
4654 @item SOLIB_CREATE_INFERIOR_HOOK
4655 @findex SOLIB_CREATE_INFERIOR_HOOK
4656 Define this to expand into any shared-library-relocation code that you
4657 want to be run just after the child process has been forked.
4659 @item START_INFERIOR_TRAPS_EXPECTED
4660 @findex START_INFERIOR_TRAPS_EXPECTED
4661 When starting an inferior, @value{GDBN} normally expects to trap
4663 the shell execs, and once when the program itself execs. If the actual
4664 number of traps is something other than 2, then define this macro to
4665 expand into the number expected.
4669 @node Support Libraries
4671 @chapter Support Libraries
4676 BFD provides support for @value{GDBN} in several ways:
4679 @item identifying executable and core files
4680 BFD will identify a variety of file types, including a.out, coff, and
4681 several variants thereof, as well as several kinds of core files.
4683 @item access to sections of files
4684 BFD parses the file headers to determine the names, virtual addresses,
4685 sizes, and file locations of all the various named sections in files
4686 (such as the text section or the data section). @value{GDBN} simply
4687 calls BFD to read or write section @var{x} at byte offset @var{y} for
4690 @item specialized core file support
4691 BFD provides routines to determine the failing command name stored in a
4692 core file, the signal with which the program failed, and whether a core
4693 file matches (i.e.@: could be a core dump of) a particular executable
4696 @item locating the symbol information
4697 @value{GDBN} uses an internal interface of BFD to determine where to find the
4698 symbol information in an executable file or symbol-file. @value{GDBN} itself
4699 handles the reading of symbols, since BFD does not ``understand'' debug
4700 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
4705 @cindex opcodes library
4707 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
4708 library because it's also used in binutils, for @file{objdump}).
4711 @cindex readline library
4712 The @code{readline} library provides a set of functions for use by applications
4713 that allow users to edit command lines as they are typed in.
4716 @cindex @code{libiberty} library
4718 The @code{libiberty} library provides a set of functions and features
4719 that integrate and improve on functionality found in modern operating
4720 systems. Broadly speaking, such features can be divided into three
4721 groups: supplemental functions (functions that may be missing in some
4722 environments and operating systems), replacement functions (providing
4723 a uniform and easier to use interface for commonly used standard
4724 functions), and extensions (which provide additional functionality
4725 beyond standard functions).
4727 @value{GDBN} uses various features provided by the @code{libiberty}
4728 library, for instance the C@t{++} demangler, the @acronym{IEEE}
4729 floating format support functions, the input options parser
4730 @samp{getopt}, the @samp{obstack} extension, and other functions.
4732 @subsection @code{obstacks} in @value{GDBN}
4733 @cindex @code{obstacks}
4735 The obstack mechanism provides a convenient way to allocate and free
4736 chunks of memory. Each obstack is a pool of memory that is managed
4737 like a stack. Objects (of any nature, size and alignment) are
4738 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
4739 @code{libiberty}'s documentation for a more detailed explanation of
4742 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
4743 object files. There is an obstack associated with each internal
4744 representation of an object file. Lots of things get allocated on
4745 these @code{obstacks}: dictionary entries, blocks, blockvectors,
4746 symbols, minimal symbols, types, vectors of fundamental types, class
4747 fields of types, object files section lists, object files section
4748 offset lists, line tables, symbol tables, partial symbol tables,
4749 string tables, symbol table private data, macros tables, debug
4750 information sections and entries, import and export lists (som),
4751 unwind information (hppa), dwarf2 location expressions data. Plus
4752 various strings such as directory names strings, debug format strings,
4755 An essential and convenient property of all data on @code{obstacks} is
4756 that memory for it gets allocated (with @code{obstack_alloc}) at
4757 various times during a debugging session, but it is released all at
4758 once using the @code{obstack_free} function. The @code{obstack_free}
4759 function takes a pointer to where in the stack it must start the
4760 deletion from (much like the cleanup chains have a pointer to where to
4761 start the cleanups). Because of the stack like structure of the
4762 @code{obstacks}, this allows to free only a top portion of the
4763 obstack. There are a few instances in @value{GDBN} where such thing
4764 happens. Calls to @code{obstack_free} are done after some local data
4765 is allocated to the obstack. Only the local data is deleted from the
4766 obstack. Of course this assumes that nothing between the
4767 @code{obstack_alloc} and the @code{obstack_free} allocates anything
4768 else on the same obstack. For this reason it is best and safest to
4769 use temporary @code{obstacks}.
4771 Releasing the whole obstack is also not safe per se. It is safe only
4772 under the condition that we know the @code{obstacks} memory is no
4773 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
4774 when we get rid of the whole objfile(s), for instance upon reading a
4778 @cindex regular expressions library
4789 @item SIGN_EXTEND_CHAR
4791 @item SWITCH_ENUM_BUG
4800 @section Array Containers
4801 @cindex Array Containers
4804 Often it is necessary to manipulate a dynamic array of a set of
4805 objects. C forces some bookkeeping on this, which can get cumbersome
4806 and repetitive. The @file{vec.h} file contains macros for defining
4807 and using a typesafe vector type. The functions defined will be
4808 inlined when compiling, and so the abstraction cost should be zero.
4809 Domain checks are added to detect programming errors.
4811 An example use would be an array of symbols or section information.
4812 The array can be grown as symbols are read in (or preallocated), and
4813 the accessor macros provided keep care of all the necessary
4814 bookkeeping. Because the arrays are type safe, there is no danger of
4815 accidentally mixing up the contents. Think of these as C++ templates,
4816 but implemented in C.
4818 Because of the different behavior of structure objects, scalar objects
4819 and of pointers, there are three flavors of vector, one for each of
4820 these variants. Both the structure object and pointer variants pass
4821 pointers to objects around --- in the former case the pointers are
4822 stored into the vector and in the latter case the pointers are
4823 dereferenced and the objects copied into the vector. The scalar
4824 object variant is suitable for @code{int}-like objects, and the vector
4825 elements are returned by value.
4827 There are both @code{index} and @code{iterate} accessors. The iterator
4828 returns a boolean iteration condition and updates the iteration
4829 variable passed by reference. Because the iterator will be inlined,
4830 the address-of can be optimized away.
4832 The vectors are implemented using the trailing array idiom, thus they
4833 are not resizeable without changing the address of the vector object
4834 itself. This means you cannot have variables or fields of vector type
4835 --- always use a pointer to a vector. The one exception is the final
4836 field of a structure, which could be a vector type. You will have to
4837 use the @code{embedded_size} & @code{embedded_init} calls to create
4838 such objects, and they will probably not be resizeable (so don't use
4839 the @dfn{safe} allocation variants). The trailing array idiom is used
4840 (rather than a pointer to an array of data), because, if we allow
4841 @code{NULL} to also represent an empty vector, empty vectors occupy
4842 minimal space in the structure containing them.
4844 Each operation that increases the number of active elements is
4845 available in @dfn{quick} and @dfn{safe} variants. The former presumes
4846 that there is sufficient allocated space for the operation to succeed
4847 (it dies if there is not). The latter will reallocate the vector, if
4848 needed. Reallocation causes an exponential increase in vector size.
4849 If you know you will be adding N elements, it would be more efficient
4850 to use the reserve operation before adding the elements with the
4851 @dfn{quick} operation. This will ensure there are at least as many
4852 elements as you ask for, it will exponentially increase if there are
4853 too few spare slots. If you want reserve a specific number of slots,
4854 but do not want the exponential increase (for instance, you know this
4855 is the last allocation), use a negative number for reservation. You
4856 can also create a vector of a specific size from the get go.
4858 You should prefer the push and pop operations, as they append and
4859 remove from the end of the vector. If you need to remove several items
4860 in one go, use the truncate operation. The insert and remove
4861 operations allow you to change elements in the middle of the vector.
4862 There are two remove operations, one which preserves the element
4863 ordering @code{ordered_remove}, and one which does not
4864 @code{unordered_remove}. The latter function copies the end element
4865 into the removed slot, rather than invoke a memmove operation. The
4866 @code{lower_bound} function will determine where to place an item in
4867 the array using insert that will maintain sorted order.
4869 If you need to directly manipulate a vector, then the @code{address}
4870 accessor will return the address of the start of the vector. Also the
4871 @code{space} predicate will tell you whether there is spare capacity in the
4872 vector. You will not normally need to use these two functions.
4874 Vector types are defined using a
4875 @code{DEF_VEC_@{O,P,I@}(@var{typename})} macro. Variables of vector
4876 type are declared using a @code{VEC(@var{typename})} macro. The
4877 characters @code{O}, @code{P} and @code{I} indicate whether
4878 @var{typename} is an object (@code{O}), pointer (@code{P}) or integral
4879 (@code{I}) type. Be careful to pick the correct one, as you'll get an
4880 awkward and inefficient API if you use the wrong one. There is a
4881 check, which results in a compile-time warning, for the @code{P} and
4882 @code{I} versions, but there is no check for the @code{O} versions, as
4883 that is not possible in plain C.
4885 An example of their use would be,
4888 DEF_VEC_P(tree); // non-managed tree vector.
4891 VEC(tree) *v; // A (pointer to) a vector of tree pointers.
4894 struct my_struct *s;
4896 if (VEC_length(tree, s->v)) @{ we have some contents @}
4897 VEC_safe_push(tree, s->v, decl); // append some decl onto the end
4898 for (ix = 0; VEC_iterate(tree, s->v, ix, elt); ix++)
4899 @{ do something with elt @}
4903 The @file{vec.h} file provides details on how to invoke the various
4904 accessors provided. They are enumerated here:
4908 Return the number of items in the array,
4911 Return true if the array has no elements.
4915 Return the last or arbitrary item in the array.
4918 Access an array element and indicate whether the array has been
4923 Create and destroy an array.
4925 @item VEC_embedded_size
4926 @itemx VEC_embedded_init
4927 Helpers for embedding an array as the final element of another struct.
4933 Return the amount of free space in an array.
4936 Ensure a certain amount of free space.
4938 @item VEC_quick_push
4939 @itemx VEC_safe_push
4940 Append to an array, either assuming the space is available, or making
4944 Remove the last item from an array.
4947 Remove several items from the end of an array.
4950 Add several items to the end of an array.
4953 Overwrite an item in the array.
4955 @item VEC_quick_insert
4956 @itemx VEC_safe_insert
4957 Insert an item into the middle of the array. Either the space must
4958 already exist, or the space is created.
4960 @item VEC_ordered_remove
4961 @itemx VEC_unordered_remove
4962 Remove an item from the array, preserving order or not.
4964 @item VEC_block_remove
4965 Remove a set of items from the array.
4968 Provide the address of the first element.
4970 @item VEC_lower_bound
4971 Binary search the array.
4981 This chapter covers topics that are lower-level than the major
4982 algorithms of @value{GDBN}.
4987 Cleanups are a structured way to deal with things that need to be done
4990 When your code does something (e.g., @code{xmalloc} some memory, or
4991 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
4992 the memory or @code{close} the file), it can make a cleanup. The
4993 cleanup will be done at some future point: when the command is finished
4994 and control returns to the top level; when an error occurs and the stack
4995 is unwound; or when your code decides it's time to explicitly perform
4996 cleanups. Alternatively you can elect to discard the cleanups you
5002 @item struct cleanup *@var{old_chain};
5003 Declare a variable which will hold a cleanup chain handle.
5005 @findex make_cleanup
5006 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
5007 Make a cleanup which will cause @var{function} to be called with
5008 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
5009 handle that can later be passed to @code{do_cleanups} or
5010 @code{discard_cleanups}. Unless you are going to call
5011 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
5012 from @code{make_cleanup}.
5015 @item do_cleanups (@var{old_chain});
5016 Do all cleanups added to the chain since the corresponding
5017 @code{make_cleanup} call was made.
5019 @findex discard_cleanups
5020 @item discard_cleanups (@var{old_chain});
5021 Same as @code{do_cleanups} except that it just removes the cleanups from
5022 the chain and does not call the specified functions.
5025 Cleanups are implemented as a chain. The handle returned by
5026 @code{make_cleanups} includes the cleanup passed to the call and any
5027 later cleanups appended to the chain (but not yet discarded or
5031 make_cleanup (a, 0);
5033 struct cleanup *old = make_cleanup (b, 0);
5041 will call @code{c()} and @code{b()} but will not call @code{a()}. The
5042 cleanup that calls @code{a()} will remain in the cleanup chain, and will
5043 be done later unless otherwise discarded.@refill
5045 Your function should explicitly do or discard the cleanups it creates.
5046 Failing to do this leads to non-deterministic behavior since the caller
5047 will arbitrarily do or discard your functions cleanups. This need leads
5048 to two common cleanup styles.
5050 The first style is try/finally. Before it exits, your code-block calls
5051 @code{do_cleanups} with the old cleanup chain and thus ensures that your
5052 code-block's cleanups are always performed. For instance, the following
5053 code-segment avoids a memory leak problem (even when @code{error} is
5054 called and a forced stack unwind occurs) by ensuring that the
5055 @code{xfree} will always be called:
5058 struct cleanup *old = make_cleanup (null_cleanup, 0);
5059 data = xmalloc (sizeof blah);
5060 make_cleanup (xfree, data);
5065 The second style is try/except. Before it exits, your code-block calls
5066 @code{discard_cleanups} with the old cleanup chain and thus ensures that
5067 any created cleanups are not performed. For instance, the following
5068 code segment, ensures that the file will be closed but only if there is
5072 FILE *file = fopen ("afile", "r");
5073 struct cleanup *old = make_cleanup (close_file, file);
5075 discard_cleanups (old);
5079 Some functions, e.g., @code{fputs_filtered()} or @code{error()}, specify
5080 that they ``should not be called when cleanups are not in place''. This
5081 means that any actions you need to reverse in the case of an error or
5082 interruption must be on the cleanup chain before you call these
5083 functions, since they might never return to your code (they
5084 @samp{longjmp} instead).
5086 @section Per-architecture module data
5087 @cindex per-architecture module data
5088 @cindex multi-arch data
5089 @cindex data-pointer, per-architecture/per-module
5091 The multi-arch framework includes a mechanism for adding module
5092 specific per-architecture data-pointers to the @code{struct gdbarch}
5093 architecture object.
5095 A module registers one or more per-architecture data-pointers using:
5097 @deftypefun struct gdbarch_data *gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
5098 @var{pre_init} is used to, on-demand, allocate an initial value for a
5099 per-architecture data-pointer using the architecture's obstack (passed
5100 in as a parameter). Since @var{pre_init} can be called during
5101 architecture creation, it is not parameterized with the architecture.
5102 and must not call modules that use per-architecture data.
5105 @deftypefun struct gdbarch_data *gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
5106 @var{post_init} is used to obtain an initial value for a
5107 per-architecture data-pointer @emph{after}. Since @var{post_init} is
5108 always called after architecture creation, it both receives the fully
5109 initialized architecture and is free to call modules that use
5110 per-architecture data (care needs to be taken to ensure that those
5111 other modules do not try to call back to this module as that will
5112 create in cycles in the initialization call graph).
5115 These functions return a @code{struct gdbarch_data} that is used to
5116 identify the per-architecture data-pointer added for that module.
5118 The per-architecture data-pointer is accessed using the function:
5120 @deftypefun void *gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
5121 Given the architecture @var{arch} and module data handle
5122 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
5123 or @code{gdbarch_data_register_post_init}), this function returns the
5124 current value of the per-architecture data-pointer. If the data
5125 pointer is @code{NULL}, it is first initialized by calling the
5126 corresponding @var{pre_init} or @var{post_init} method.
5129 The examples below assume the following definitions:
5132 struct nozel @{ int total; @};
5133 static struct gdbarch_data *nozel_handle;
5136 A module can extend the architecture vector, adding additional
5137 per-architecture data, using the @var{pre_init} method. The module's
5138 per-architecture data is then initialized during architecture
5141 In the below, the module's per-architecture @emph{nozel} is added. An
5142 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
5143 from @code{gdbarch_init}.
5147 nozel_pre_init (struct obstack *obstack)
5149 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
5156 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
5158 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5159 data->total = nozel;
5163 A module can on-demand create architecture dependant data structures
5164 using @code{post_init}.
5166 In the below, the nozel's total is computed on-demand by
5167 @code{nozel_post_init} using information obtained from the
5172 nozel_post_init (struct gdbarch *gdbarch)
5174 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
5175 nozel->total = gdbarch@dots{} (gdbarch);
5182 nozel_total (struct gdbarch *gdbarch)
5184 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5189 @section Wrapping Output Lines
5190 @cindex line wrap in output
5193 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
5194 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
5195 added in places that would be good breaking points. The utility
5196 routines will take care of actually wrapping if the line width is
5199 The argument to @code{wrap_here} is an indentation string which is
5200 printed @emph{only} if the line breaks there. This argument is saved
5201 away and used later. It must remain valid until the next call to
5202 @code{wrap_here} or until a newline has been printed through the
5203 @code{*_filtered} functions. Don't pass in a local variable and then
5206 It is usually best to call @code{wrap_here} after printing a comma or
5207 space. If you call it before printing a space, make sure that your
5208 indentation properly accounts for the leading space that will print if
5209 the line wraps there.
5211 Any function or set of functions that produce filtered output must
5212 finish by printing a newline, to flush the wrap buffer, before switching
5213 to unfiltered (@code{printf}) output. Symbol reading routines that
5214 print warnings are a good example.
5216 @section @value{GDBN} Coding Standards
5217 @cindex coding standards
5219 @value{GDBN} follows the GNU coding standards, as described in
5220 @file{etc/standards.texi}. This file is also available for anonymous
5221 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
5222 of the standard; in general, when the GNU standard recommends a practice
5223 but does not require it, @value{GDBN} requires it.
5225 @value{GDBN} follows an additional set of coding standards specific to
5226 @value{GDBN}, as described in the following sections.
5231 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
5234 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
5237 @subsection Memory Management
5239 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
5240 @code{calloc}, @code{free} and @code{asprintf}.
5242 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
5243 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
5244 these functions do not return when the memory pool is empty. Instead,
5245 they unwind the stack using cleanups. These functions return
5246 @code{NULL} when requested to allocate a chunk of memory of size zero.
5248 @emph{Pragmatics: By using these functions, the need to check every
5249 memory allocation is removed. These functions provide portable
5252 @value{GDBN} does not use the function @code{free}.
5254 @value{GDBN} uses the function @code{xfree} to return memory to the
5255 memory pool. Consistent with ISO-C, this function ignores a request to
5256 free a @code{NULL} pointer.
5258 @emph{Pragmatics: On some systems @code{free} fails when passed a
5259 @code{NULL} pointer.}
5261 @value{GDBN} can use the non-portable function @code{alloca} for the
5262 allocation of small temporary values (such as strings).
5264 @emph{Pragmatics: This function is very non-portable. Some systems
5265 restrict the memory being allocated to no more than a few kilobytes.}
5267 @value{GDBN} uses the string function @code{xstrdup} and the print
5268 function @code{xstrprintf}.
5270 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
5271 functions such as @code{sprintf} are very prone to buffer overflow
5275 @subsection Compiler Warnings
5276 @cindex compiler warnings
5278 With few exceptions, developers should avoid the configuration option
5279 @samp{--disable-werror} when building @value{GDBN}. The exceptions
5280 are listed in the file @file{gdb/MAINTAINERS}. The default, when
5281 building with @sc{gcc}, is @samp{--enable-werror}.
5283 This option causes @value{GDBN} (when built using GCC) to be compiled
5284 with a carefully selected list of compiler warning flags. Any warnings
5285 from those flags are treated as errors.
5287 The current list of warning flags includes:
5291 Recommended @sc{gcc} warnings.
5293 @item -Wdeclaration-after-statement
5295 @sc{gcc} 3.x (and later) and @sc{c99} allow declarations mixed with
5296 code, but @sc{gcc} 2.x and @sc{c89} do not.
5298 @item -Wpointer-arith
5300 @item -Wformat-nonliteral
5301 Non-literal format strings, with a few exceptions, are bugs - they
5302 might contain unintended user-supplied format specifiers.
5303 Since @value{GDBN} uses the @code{format printf} attribute on all
5304 @code{printf} like functions this checks not just @code{printf} calls
5305 but also calls to functions such as @code{fprintf_unfiltered}.
5307 @item -Wno-pointer-sign
5308 In version 4.0, GCC began warning about pointer argument passing or
5309 assignment even when the source and destination differed only in
5310 signedness. However, most @value{GDBN} code doesn't distinguish
5311 carefully between @code{char} and @code{unsigned char}. In early 2006
5312 the @value{GDBN} developers decided correcting these warnings wasn't
5313 worth the time it would take.
5315 @item -Wno-unused-parameter
5316 Due to the way that @value{GDBN} is implemented many functions have
5317 unused parameters. Consequently this warning is avoided. The macro
5318 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
5319 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
5324 @itemx -Wno-char-subscripts
5325 These are warnings which might be useful for @value{GDBN}, but are
5326 currently too noisy to enable with @samp{-Werror}.
5330 @subsection Formatting
5332 @cindex source code formatting
5333 The standard GNU recommendations for formatting must be followed
5336 A function declaration should not have its name in column zero. A
5337 function definition should have its name in column zero.
5341 static void foo (void);
5349 @emph{Pragmatics: This simplifies scripting. Function definitions can
5350 be found using @samp{^function-name}.}
5352 There must be a space between a function or macro name and the opening
5353 parenthesis of its argument list (except for macro definitions, as
5354 required by C). There must not be a space after an open paren/bracket
5355 or before a close paren/bracket.
5357 While additional whitespace is generally helpful for reading, do not use
5358 more than one blank line to separate blocks, and avoid adding whitespace
5359 after the end of a program line (as of 1/99, some 600 lines had
5360 whitespace after the semicolon). Excess whitespace causes difficulties
5361 for @code{diff} and @code{patch} utilities.
5363 Pointers are declared using the traditional K&R C style:
5377 @subsection Comments
5379 @cindex comment formatting
5380 The standard GNU requirements on comments must be followed strictly.
5382 Block comments must appear in the following form, with no @code{/*}- or
5383 @code{*/}-only lines, and no leading @code{*}:
5386 /* Wait for control to return from inferior to debugger. If inferior
5387 gets a signal, we may decide to start it up again instead of
5388 returning. That is why there is a loop in this function. When
5389 this function actually returns it means the inferior should be left
5390 stopped and @value{GDBN} should read more commands. */
5393 (Note that this format is encouraged by Emacs; tabbing for a multi-line
5394 comment works correctly, and @kbd{M-q} fills the block consistently.)
5396 Put a blank line between the block comments preceding function or
5397 variable definitions, and the definition itself.
5399 In general, put function-body comments on lines by themselves, rather
5400 than trying to fit them into the 20 characters left at the end of a
5401 line, since either the comment or the code will inevitably get longer
5402 than will fit, and then somebody will have to move it anyhow.
5406 @cindex C data types
5407 Code must not depend on the sizes of C data types, the format of the
5408 host's floating point numbers, the alignment of anything, or the order
5409 of evaluation of expressions.
5411 @cindex function usage
5412 Use functions freely. There are only a handful of compute-bound areas
5413 in @value{GDBN} that might be affected by the overhead of a function
5414 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
5415 limited by the target interface (whether serial line or system call).
5417 However, use functions with moderation. A thousand one-line functions
5418 are just as hard to understand as a single thousand-line function.
5420 @emph{Macros are bad, M'kay.}
5421 (But if you have to use a macro, make sure that the macro arguments are
5422 protected with parentheses.)
5426 Declarations like @samp{struct foo *} should be used in preference to
5427 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
5430 @subsection Function Prototypes
5431 @cindex function prototypes
5433 Prototypes must be used when both @emph{declaring} and @emph{defining}
5434 a function. Prototypes for @value{GDBN} functions must include both the
5435 argument type and name, with the name matching that used in the actual
5436 function definition.
5438 All external functions should have a declaration in a header file that
5439 callers include, except for @code{_initialize_*} functions, which must
5440 be external so that @file{init.c} construction works, but shouldn't be
5441 visible to random source files.
5443 Where a source file needs a forward declaration of a static function,
5444 that declaration must appear in a block near the top of the source file.
5447 @subsection Internal Error Recovery
5449 During its execution, @value{GDBN} can encounter two types of errors.
5450 User errors and internal errors. User errors include not only a user
5451 entering an incorrect command but also problems arising from corrupt
5452 object files and system errors when interacting with the target.
5453 Internal errors include situations where @value{GDBN} has detected, at
5454 run time, a corrupt or erroneous situation.
5456 When reporting an internal error, @value{GDBN} uses
5457 @code{internal_error} and @code{gdb_assert}.
5459 @value{GDBN} must not call @code{abort} or @code{assert}.
5461 @emph{Pragmatics: There is no @code{internal_warning} function. Either
5462 the code detected a user error, recovered from it and issued a
5463 @code{warning} or the code failed to correctly recover from the user
5464 error and issued an @code{internal_error}.}
5466 @subsection File Names
5468 Any file used when building the core of @value{GDBN} must be in lower
5469 case. Any file used when building the core of @value{GDBN} must be 8.3
5470 unique. These requirements apply to both source and generated files.
5472 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
5473 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
5474 is introduced to the build process both @file{Makefile.in} and
5475 @file{configure.in} need to be modified accordingly. Compare the
5476 convoluted conversion process needed to transform @file{COPYING} into
5477 @file{copying.c} with the conversion needed to transform
5478 @file{version.in} into @file{version.c}.}
5480 Any file non 8.3 compliant file (that is not used when building the core
5481 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
5483 @emph{Pragmatics: This is clearly a compromise.}
5485 When @value{GDBN} has a local version of a system header file (ex
5486 @file{string.h}) the file name based on the POSIX header prefixed with
5487 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
5488 independent: they should use only macros defined by @file{configure},
5489 the compiler, or the host; they should include only system headers; they
5490 should refer only to system types. They may be shared between multiple
5491 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
5493 For other files @samp{-} is used as the separator.
5496 @subsection Include Files
5498 A @file{.c} file should include @file{defs.h} first.
5500 A @file{.c} file should directly include the @code{.h} file of every
5501 declaration and/or definition it directly refers to. It cannot rely on
5504 A @file{.h} file should directly include the @code{.h} file of every
5505 declaration and/or definition it directly refers to. It cannot rely on
5506 indirect inclusion. Exception: The file @file{defs.h} does not need to
5507 be directly included.
5509 An external declaration should only appear in one include file.
5511 An external declaration should never appear in a @code{.c} file.
5512 Exception: a declaration for the @code{_initialize} function that
5513 pacifies @option{-Wmissing-declaration}.
5515 A @code{typedef} definition should only appear in one include file.
5517 An opaque @code{struct} declaration can appear in multiple @file{.h}
5518 files. Where possible, a @file{.h} file should use an opaque
5519 @code{struct} declaration instead of an include.
5521 All @file{.h} files should be wrapped in:
5524 #ifndef INCLUDE_FILE_NAME_H
5525 #define INCLUDE_FILE_NAME_H
5531 @subsection Clean Design and Portable Implementation
5534 In addition to getting the syntax right, there's the little question of
5535 semantics. Some things are done in certain ways in @value{GDBN} because long
5536 experience has shown that the more obvious ways caused various kinds of
5539 @cindex assumptions about targets
5540 You can't assume the byte order of anything that comes from a target
5541 (including @var{value}s, object files, and instructions). Such things
5542 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
5543 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
5544 such as @code{bfd_get_32}.
5546 You can't assume that you know what interface is being used to talk to
5547 the target system. All references to the target must go through the
5548 current @code{target_ops} vector.
5550 You can't assume that the host and target machines are the same machine
5551 (except in the ``native'' support modules). In particular, you can't
5552 assume that the target machine's header files will be available on the
5553 host machine. Target code must bring along its own header files --
5554 written from scratch or explicitly donated by their owner, to avoid
5558 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
5559 to write the code portably than to conditionalize it for various
5562 @cindex system dependencies
5563 New @code{#ifdef}'s which test for specific compilers or manufacturers
5564 or operating systems are unacceptable. All @code{#ifdef}'s should test
5565 for features. The information about which configurations contain which
5566 features should be segregated into the configuration files. Experience
5567 has proven far too often that a feature unique to one particular system
5568 often creeps into other systems; and that a conditional based on some
5569 predefined macro for your current system will become worthless over
5570 time, as new versions of your system come out that behave differently
5571 with regard to this feature.
5573 Adding code that handles specific architectures, operating systems,
5574 target interfaces, or hosts, is not acceptable in generic code.
5576 @cindex portable file name handling
5577 @cindex file names, portability
5578 One particularly notorious area where system dependencies tend to
5579 creep in is handling of file names. The mainline @value{GDBN} code
5580 assumes Posix semantics of file names: absolute file names begin with
5581 a forward slash @file{/}, slashes are used to separate leading
5582 directories, case-sensitive file names. These assumptions are not
5583 necessarily true on non-Posix systems such as MS-Windows. To avoid
5584 system-dependent code where you need to take apart or construct a file
5585 name, use the following portable macros:
5588 @findex HAVE_DOS_BASED_FILE_SYSTEM
5589 @item HAVE_DOS_BASED_FILE_SYSTEM
5590 This preprocessing symbol is defined to a non-zero value on hosts
5591 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
5592 symbol to write conditional code which should only be compiled for
5595 @findex IS_DIR_SEPARATOR
5596 @item IS_DIR_SEPARATOR (@var{c})
5597 Evaluates to a non-zero value if @var{c} is a directory separator
5598 character. On Unix and GNU/Linux systems, only a slash @file{/} is
5599 such a character, but on Windows, both @file{/} and @file{\} will
5602 @findex IS_ABSOLUTE_PATH
5603 @item IS_ABSOLUTE_PATH (@var{file})
5604 Evaluates to a non-zero value if @var{file} is an absolute file name.
5605 For Unix and GNU/Linux hosts, a name which begins with a slash
5606 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
5607 @file{x:\bar} are also absolute file names.
5609 @findex FILENAME_CMP
5610 @item FILENAME_CMP (@var{f1}, @var{f2})
5611 Calls a function which compares file names @var{f1} and @var{f2} as
5612 appropriate for the underlying host filesystem. For Posix systems,
5613 this simply calls @code{strcmp}; on case-insensitive filesystems it
5614 will call @code{strcasecmp} instead.
5616 @findex DIRNAME_SEPARATOR
5617 @item DIRNAME_SEPARATOR
5618 Evaluates to a character which separates directories in
5619 @code{PATH}-style lists, typically held in environment variables.
5620 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
5622 @findex SLASH_STRING
5624 This evaluates to a constant string you should use to produce an
5625 absolute filename from leading directories and the file's basename.
5626 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
5627 @code{"\\"} for some Windows-based ports.
5630 In addition to using these macros, be sure to use portable library
5631 functions whenever possible. For example, to extract a directory or a
5632 basename part from a file name, use the @code{dirname} and
5633 @code{basename} library functions (available in @code{libiberty} for
5634 platforms which don't provide them), instead of searching for a slash
5635 with @code{strrchr}.
5637 Another way to generalize @value{GDBN} along a particular interface is with an
5638 attribute struct. For example, @value{GDBN} has been generalized to handle
5639 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
5640 by defining the @code{target_ops} structure and having a current target (as
5641 well as a stack of targets below it, for memory references). Whenever
5642 something needs to be done that depends on which remote interface we are
5643 using, a flag in the current target_ops structure is tested (e.g.,
5644 @code{target_has_stack}), or a function is called through a pointer in the
5645 current target_ops structure. In this way, when a new remote interface
5646 is added, only one module needs to be touched---the one that actually
5647 implements the new remote interface. Other examples of
5648 attribute-structs are BFD access to multiple kinds of object file
5649 formats, or @value{GDBN}'s access to multiple source languages.
5651 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
5652 the code interfacing between @code{ptrace} and the rest of
5653 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
5654 something was very painful. In @value{GDBN} 4.x, these have all been
5655 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
5656 with variations between systems the same way any system-independent
5657 file would (hooks, @code{#if defined}, etc.), and machines which are
5658 radically different don't need to use @file{infptrace.c} at all.
5660 All debugging code must be controllable using the @samp{set debug
5661 @var{module}} command. Do not use @code{printf} to print trace
5662 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
5663 @code{#ifdef DEBUG}.
5668 @chapter Porting @value{GDBN}
5669 @cindex porting to new machines
5671 Most of the work in making @value{GDBN} compile on a new machine is in
5672 specifying the configuration of the machine. This is done in a
5673 dizzying variety of header files and configuration scripts, which we
5674 hope to make more sensible soon. Let's say your new host is called an
5675 @var{xyz} (e.g., @samp{sun4}), and its full three-part configuration
5676 name is @code{@var{arch}-@var{xvend}-@var{xos}} (e.g.,
5677 @samp{sparc-sun-sunos4}). In particular:
5681 In the top level directory, edit @file{config.sub} and add @var{arch},
5682 @var{xvend}, and @var{xos} to the lists of supported architectures,
5683 vendors, and operating systems near the bottom of the file. Also, add
5684 @var{xyz} as an alias that maps to
5685 @code{@var{arch}-@var{xvend}-@var{xos}}. You can test your changes by
5689 ./config.sub @var{xyz}
5696 ./config.sub @code{@var{arch}-@var{xvend}-@var{xos}}
5700 which should both respond with @code{@var{arch}-@var{xvend}-@var{xos}}
5701 and no error messages.
5704 You need to port BFD, if that hasn't been done already. Porting BFD is
5705 beyond the scope of this manual.
5708 To configure @value{GDBN} itself, edit @file{gdb/configure.host} to recognize
5709 your system and set @code{gdb_host} to @var{xyz}, and (unless your
5710 desired target is already available) also edit @file{gdb/configure.tgt},
5711 setting @code{gdb_target} to something appropriate (for instance,
5714 @emph{Maintainer's note: Work in progress. The file
5715 @file{gdb/configure.host} originally needed to be modified when either a
5716 new native target or a new host machine was being added to @value{GDBN}.
5717 Recent changes have removed this requirement. The file now only needs
5718 to be modified when adding a new native configuration. This will likely
5719 changed again in the future.}
5722 Finally, you'll need to specify and define @value{GDBN}'s host-, native-, and
5723 target-dependent @file{.h} and @file{.c} files used for your
5727 @node Versions and Branches
5728 @chapter Versions and Branches
5732 @value{GDBN}'s version is determined by the file
5733 @file{gdb/version.in} and takes one of the following forms:
5736 @item @var{major}.@var{minor}
5737 @itemx @var{major}.@var{minor}.@var{patchlevel}
5738 an official release (e.g., 6.2 or 6.2.1)
5739 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}
5740 a snapshot taken at @var{YYYY}-@var{MM}-@var{DD}-gmt (e.g.,
5741 6.1.50.20020302, 6.1.90.20020304, or 6.1.0.20020308)
5742 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}-cvs
5743 a @sc{cvs} check out drawn on @var{YYYY}-@var{MM}-@var{DD} (e.g.,
5744 6.1.50.20020302-cvs, 6.1.90.20020304-cvs, or 6.1.0.20020308-cvs)
5745 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD} (@var{vendor})
5746 a vendor specific release of @value{GDBN}, that while based on@*
5747 @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD},
5748 may include additional changes
5751 @value{GDBN}'s mainline uses the @var{major} and @var{minor} version
5752 numbers from the most recent release branch, with a @var{patchlevel}
5753 of 50. At the time each new release branch is created, the mainline's
5754 @var{major} and @var{minor} version numbers are updated.
5756 @value{GDBN}'s release branch is similar. When the branch is cut, the
5757 @var{patchlevel} is changed from 50 to 90. As draft releases are
5758 drawn from the branch, the @var{patchlevel} is incremented. Once the
5759 first release (@var{major}.@var{minor}) has been made, the
5760 @var{patchlevel} is set to 0 and updates have an incremented
5763 For snapshots, and @sc{cvs} check outs, it is also possible to
5764 identify the @sc{cvs} origin:
5767 @item @var{major}.@var{minor}.50.@var{YYYY}@var{MM}@var{DD}
5768 drawn from the @sc{head} of mainline @sc{cvs} (e.g., 6.1.50.20020302)
5769 @item @var{major}.@var{minor}.90.@var{YYYY}@var{MM}@var{DD}
5770 @itemx @var{major}.@var{minor}.91.@var{YYYY}@var{MM}@var{DD} @dots{}
5771 drawn from a release branch prior to the release (e.g.,
5773 @item @var{major}.@var{minor}.0.@var{YYYY}@var{MM}@var{DD}
5774 @itemx @var{major}.@var{minor}.1.@var{YYYY}@var{MM}@var{DD} @dots{}
5775 drawn from a release branch after the release (e.g., 6.2.0.20020308)
5778 If the previous @value{GDBN} version is 6.1 and the current version is
5779 6.2, then, substituting 6 for @var{major} and 1 or 2 for @var{minor},
5780 here's an illustration of a typical sequence:
5787 +--------------------------.
5790 6.2.50.20020303-cvs 6.1.90 (draft #1)
5792 6.2.50.20020304-cvs 6.1.90.20020304-cvs
5794 6.2.50.20020305-cvs 6.1.91 (draft #2)
5796 6.2.50.20020306-cvs 6.1.91.20020306-cvs
5798 6.2.50.20020307-cvs 6.2 (release)
5800 6.2.50.20020308-cvs 6.2.0.20020308-cvs
5802 6.2.50.20020309-cvs 6.2.1 (update)
5804 6.2.50.20020310-cvs <branch closed>
5808 +--------------------------.
5811 6.3.50.20020312-cvs 6.2.90 (draft #1)
5815 @section Release Branches
5816 @cindex Release Branches
5818 @value{GDBN} draws a release series (6.2, 6.2.1, @dots{}) from a
5819 single release branch, and identifies that branch using the @sc{cvs}
5823 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-branchpoint
5824 gdb_@var{major}_@var{minor}-branch
5825 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-release
5828 @emph{Pragmatics: To help identify the date at which a branch or
5829 release is made, both the branchpoint and release tags include the
5830 date that they are cut (@var{YYYY}@var{MM}@var{DD}) in the tag. The
5831 branch tag, denoting the head of the branch, does not need this.}
5833 @section Vendor Branches
5834 @cindex vendor branches
5836 To avoid version conflicts, vendors are expected to modify the file
5837 @file{gdb/version.in} to include a vendor unique alphabetic identifier
5838 (an official @value{GDBN} release never uses alphabetic characters in
5839 its version identifier). E.g., @samp{6.2widgit2}, or @samp{6.2 (Widgit
5842 @section Experimental Branches
5843 @cindex experimental branches
5845 @subsection Guidelines
5847 @value{GDBN} permits the creation of branches, cut from the @sc{cvs}
5848 repository, for experimental development. Branches make it possible
5849 for developers to share preliminary work, and maintainers to examine
5850 significant new developments.
5852 The following are a set of guidelines for creating such branches:
5856 @item a branch has an owner
5857 The owner can set further policy for a branch, but may not change the
5858 ground rules. In particular, they can set a policy for commits (be it
5859 adding more reviewers or deciding who can commit).
5861 @item all commits are posted
5862 All changes committed to a branch shall also be posted to
5863 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} patches
5864 mailing list}. While commentary on such changes are encouraged, people
5865 should remember that the changes only apply to a branch.
5867 @item all commits are covered by an assignment
5868 This ensures that all changes belong to the Free Software Foundation,
5869 and avoids the possibility that the branch may become contaminated.
5871 @item a branch is focused
5872 A focused branch has a single objective or goal, and does not contain
5873 unnecessary or irrelevant changes. Cleanups, where identified, being
5874 be pushed into the mainline as soon as possible.
5876 @item a branch tracks mainline
5877 This keeps the level of divergence under control. It also keeps the
5878 pressure on developers to push cleanups and other stuff into the
5881 @item a branch shall contain the entire @value{GDBN} module
5882 The @value{GDBN} module @code{gdb} should be specified when creating a
5883 branch (branches of individual files should be avoided). @xref{Tags}.
5885 @item a branch shall be branded using @file{version.in}
5886 The file @file{gdb/version.in} shall be modified so that it identifies
5887 the branch @var{owner} and branch @var{name}, e.g.,
5888 @samp{6.2.50.20030303_owner_name} or @samp{6.2 (Owner Name)}.
5895 To simplify the identification of @value{GDBN} branches, the following
5896 branch tagging convention is strongly recommended:
5900 @item @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5901 @itemx @var{owner}_@var{name}-@var{YYYYMMDD}-branch
5902 The branch point and corresponding branch tag. @var{YYYYMMDD} is the
5903 date that the branch was created. A branch is created using the
5904 sequence: @anchor{experimental branch tags}
5906 cvs rtag @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint gdb
5907 cvs rtag -b -r @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint \
5908 @var{owner}_@var{name}-@var{YYYYMMDD}-branch gdb
5911 @item @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5912 The tagged point, on the mainline, that was used when merging the branch
5913 on @var{yyyymmdd}. To merge in all changes since the branch was cut,
5914 use a command sequence like:
5916 cvs rtag @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint gdb
5918 -j@var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5919 -j@var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5922 Similar sequences can be used to just merge in changes since the last
5928 For further information on @sc{cvs}, see
5929 @uref{http://www.gnu.org/software/cvs/, Concurrent Versions System}.
5931 @node Start of New Year Procedure
5932 @chapter Start of New Year Procedure
5933 @cindex new year procedure
5935 At the start of each new year, the following actions should be performed:
5939 Rotate the ChangeLog file
5941 The current @file{ChangeLog} file should be renamed into
5942 @file{ChangeLog-YYYY} where YYYY is the year that has just passed.
5943 A new @file{ChangeLog} file should be created, and its contents should
5944 contain a reference to the previous ChangeLog. The following should
5945 also be preserved at the end of the new ChangeLog, in order to provide
5946 the appropriate settings when editing this file with Emacs:
5952 version-control: never
5957 Add an entry for the newly created ChangeLog file (@file{ChangeLog-YYYY})
5958 in @file{gdb/config/djgpp/fnchange.lst}.
5961 Update the copyright year in the startup message
5963 Update the copyright year in file @file{top.c}, function
5964 @code{print_gdb_version}.
5967 Add the new year in the copyright notices of all source and documentation
5968 files. This can be done semi-automatically by running the @code{copyright.sh}
5969 script. This script requires Emacs 22 or later to be installed.
5975 @chapter Releasing @value{GDBN}
5976 @cindex making a new release of gdb
5978 @section Branch Commit Policy
5980 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
5981 5.1 and 5.2 all used the below:
5985 The @file{gdb/MAINTAINERS} file still holds.
5987 Don't fix something on the branch unless/until it is also fixed in the
5988 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
5989 file is better than committing a hack.
5991 When considering a patch for the branch, suggested criteria include:
5992 Does it fix a build? Does it fix the sequence @kbd{break main; run}
5993 when debugging a static binary?
5995 The further a change is from the core of @value{GDBN}, the less likely
5996 the change will worry anyone (e.g., target specific code).
5998 Only post a proposal to change the core of @value{GDBN} after you've
5999 sent individual bribes to all the people listed in the
6000 @file{MAINTAINERS} file @t{;-)}
6003 @emph{Pragmatics: Provided updates are restricted to non-core
6004 functionality there is little chance that a broken change will be fatal.
6005 This means that changes such as adding a new architectures or (within
6006 reason) support for a new host are considered acceptable.}
6009 @section Obsoleting code
6011 Before anything else, poke the other developers (and around the source
6012 code) to see if there is anything that can be removed from @value{GDBN}
6013 (an old target, an unused file).
6015 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
6016 line. Doing this means that it is easy to identify something that has
6017 been obsoleted when greping through the sources.
6019 The process is done in stages --- this is mainly to ensure that the
6020 wider @value{GDBN} community has a reasonable opportunity to respond.
6021 Remember, everything on the Internet takes a week.
6025 Post the proposal on @email{gdb@@sources.redhat.com, the GDB mailing
6026 list} Creating a bug report to track the task's state, is also highly
6031 Post the proposal on @email{gdb-announce@@sources.redhat.com, the GDB
6032 Announcement mailing list}.
6036 Go through and edit all relevant files and lines so that they are
6037 prefixed with the word @code{OBSOLETE}.
6039 Wait until the next GDB version, containing this obsolete code, has been
6042 Remove the obsolete code.
6046 @emph{Maintainer note: While removing old code is regrettable it is
6047 hopefully better for @value{GDBN}'s long term development. Firstly it
6048 helps the developers by removing code that is either no longer relevant
6049 or simply wrong. Secondly since it removes any history associated with
6050 the file (effectively clearing the slate) the developer has a much freer
6051 hand when it comes to fixing broken files.}
6055 @section Before the Branch
6057 The most important objective at this stage is to find and fix simple
6058 changes that become a pain to track once the branch is created. For
6059 instance, configuration problems that stop @value{GDBN} from even
6060 building. If you can't get the problem fixed, document it in the
6061 @file{gdb/PROBLEMS} file.
6063 @subheading Prompt for @file{gdb/NEWS}
6065 People always forget. Send a post reminding them but also if you know
6066 something interesting happened add it yourself. The @code{schedule}
6067 script will mention this in its e-mail.
6069 @subheading Review @file{gdb/README}
6071 Grab one of the nightly snapshots and then walk through the
6072 @file{gdb/README} looking for anything that can be improved. The
6073 @code{schedule} script will mention this in its e-mail.
6075 @subheading Refresh any imported files.
6077 A number of files are taken from external repositories. They include:
6081 @file{texinfo/texinfo.tex}
6083 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
6086 @file{etc/standards.texi}, @file{etc/make-stds.texi}
6089 @subheading Check the ARI
6091 @uref{http://sources.redhat.com/gdb/ari,,A.R.I.} is an @code{awk} script
6092 (Awk Regression Index ;-) that checks for a number of errors and coding
6093 conventions. The checks include things like using @code{malloc} instead
6094 of @code{xmalloc} and file naming problems. There shouldn't be any
6097 @subsection Review the bug data base
6099 Close anything obviously fixed.
6101 @subsection Check all cross targets build
6103 The targets are listed in @file{gdb/MAINTAINERS}.
6106 @section Cut the Branch
6108 @subheading Create the branch
6113 $ V=`echo $v | sed 's/\./_/g'`
6114 $ D=`date -u +%Y-%m-%d`
6117 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6118 -D $D-gmt gdb_$V-$D-branchpoint insight
6119 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag
6120 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight
6123 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6124 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight
6125 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6126 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight
6134 By using @kbd{-D YYYY-MM-DD-gmt}, the branch is forced to an exact
6137 The trunk is first tagged so that the branch point can easily be found.
6139 Insight, which includes @value{GDBN}, is tagged at the same time.
6141 @file{version.in} gets bumped to avoid version number conflicts.
6143 The reading of @file{.cvsrc} is disabled using @file{-f}.
6146 @subheading Update @file{version.in}
6151 $ V=`echo $v | sed 's/\./_/g'`
6155 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src co \
6156 -r gdb_$V-branch src/gdb/version.in
6157 cvs -f -d :ext:sources.redhat.com:/cvs/src co
6158 -r gdb_5_2-branch src/gdb/version.in
6160 U src/gdb/version.in
6162 $ echo $u.90-0000-00-00-cvs > version.in
6164 5.1.90-0000-00-00-cvs
6165 $ cvs -f commit version.in
6170 @file{0000-00-00} is used as a date to pump prime the version.in update
6173 @file{.90} and the previous branch version are used as fairly arbitrary
6174 initial branch version number.
6178 @subheading Update the web and news pages
6182 @subheading Tweak cron to track the new branch
6184 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
6185 This file needs to be updated so that:
6189 A daily timestamp is added to the file @file{version.in}.
6191 The new branch is included in the snapshot process.
6195 See the file @file{gdbadmin/cron/README} for how to install the updated
6198 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
6199 any changes. That file is copied to both the branch/ and current/
6200 snapshot directories.
6203 @subheading Update the NEWS and README files
6205 The @file{NEWS} file needs to be updated so that on the branch it refers
6206 to @emph{changes in the current release} while on the trunk it also
6207 refers to @emph{changes since the current release}.
6209 The @file{README} file needs to be updated so that it refers to the
6212 @subheading Post the branch info
6214 Send an announcement to the mailing lists:
6218 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6220 @email{gdb@@sources.redhat.com, GDB Discussion mailing list} and
6221 @email{gdb-testers@@sources.redhat.com, GDB Testers mailing list}
6224 @emph{Pragmatics: The branch creation is sent to the announce list to
6225 ensure that people people not subscribed to the higher volume discussion
6228 The announcement should include:
6234 How to check out the branch using CVS.
6236 The date/number of weeks until the release.
6238 The branch commit policy still holds.
6241 @section Stabilize the branch
6243 Something goes here.
6245 @section Create a Release
6247 The process of creating and then making available a release is broken
6248 down into a number of stages. The first part addresses the technical
6249 process of creating a releasable tar ball. The later stages address the
6250 process of releasing that tar ball.
6252 When making a release candidate just the first section is needed.
6254 @subsection Create a release candidate
6256 The objective at this stage is to create a set of tar balls that can be
6257 made available as a formal release (or as a less formal release
6260 @subsubheading Freeze the branch
6262 Send out an e-mail notifying everyone that the branch is frozen to
6263 @email{gdb-patches@@sources.redhat.com}.
6265 @subsubheading Establish a few defaults.
6270 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
6272 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
6276 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
6278 /home/gdbadmin/bin/autoconf
6287 Check the @code{autoconf} version carefully. You want to be using the
6288 version taken from the @file{binutils} snapshot directory, which can be
6289 found at @uref{ftp://sources.redhat.com/pub/binutils/}. It is very
6290 unlikely that a system installed version of @code{autoconf} (e.g.,
6291 @file{/usr/bin/autoconf}) is correct.
6294 @subsubheading Check out the relevant modules:
6297 $ for m in gdb insight
6299 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
6309 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
6310 any confusion between what is written here and what your local
6311 @code{cvs} really does.
6314 @subsubheading Update relevant files.
6320 Major releases get their comments added as part of the mainline. Minor
6321 releases should probably mention any significant bugs that were fixed.
6323 Don't forget to include the @file{ChangeLog} entry.
6326 $ emacs gdb/src/gdb/NEWS
6331 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
6332 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6337 You'll need to update:
6349 $ emacs gdb/src/gdb/README
6354 $ cp gdb/src/gdb/README insight/src/gdb/README
6355 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6358 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
6359 before the initial branch was cut so just a simple substitute is needed
6362 @emph{Maintainer note: Other projects generate @file{README} and
6363 @file{INSTALL} from the core documentation. This might be worth
6366 @item gdb/version.in
6369 $ echo $v > gdb/src/gdb/version.in
6370 $ cat gdb/src/gdb/version.in
6372 $ emacs gdb/src/gdb/version.in
6375 ... Bump to version ...
6377 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
6378 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6383 @subsubheading Do the dirty work
6385 This is identical to the process used to create the daily snapshot.
6388 $ for m in gdb insight
6390 ( cd $m/src && gmake -f src-release $m.tar )
6394 If the top level source directory does not have @file{src-release}
6395 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
6398 $ for m in gdb insight
6400 ( cd $m/src && gmake -f Makefile.in $m.tar )
6404 @subsubheading Check the source files
6406 You're looking for files that have mysteriously disappeared.
6407 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
6408 for the @file{version.in} update @kbd{cronjob}.
6411 $ ( cd gdb/src && cvs -f -q -n update )
6415 @dots{} lots of generated files @dots{}
6420 @dots{} lots of generated files @dots{}
6425 @emph{Don't worry about the @file{gdb.info-??} or
6426 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
6427 was also generated only something strange with CVS means that they
6428 didn't get suppressed). Fixing it would be nice though.}
6430 @subsubheading Create compressed versions of the release
6436 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
6437 $ for m in gdb insight
6439 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
6440 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
6450 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
6451 in that mode, @code{gzip} does not know the name of the file and, hence,
6452 can not include it in the compressed file. This is also why the release
6453 process runs @code{tar} and @code{bzip2} as separate passes.
6456 @subsection Sanity check the tar ball
6458 Pick a popular machine (Solaris/PPC?) and try the build on that.
6461 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
6466 $ ./gdb/gdb ./gdb/gdb
6470 Breakpoint 1 at 0x80732bc: file main.c, line 734.
6472 Starting program: /tmp/gdb-5.2/gdb/gdb
6474 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
6475 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
6477 $1 = @{argc = 136426532, argv = 0x821b7f0@}
6481 @subsection Make a release candidate available
6483 If this is a release candidate then the only remaining steps are:
6487 Commit @file{version.in} and @file{ChangeLog}
6489 Tweak @file{version.in} (and @file{ChangeLog} to read
6490 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
6491 process can restart.
6493 Make the release candidate available in
6494 @uref{ftp://sources.redhat.com/pub/gdb/snapshots/branch}
6496 Notify the relevant mailing lists ( @email{gdb@@sources.redhat.com} and
6497 @email{gdb-testers@@sources.redhat.com} that the candidate is available.
6500 @subsection Make a formal release available
6502 (And you thought all that was required was to post an e-mail.)
6504 @subsubheading Install on sware
6506 Copy the new files to both the release and the old release directory:
6509 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
6510 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
6514 Clean up the releases directory so that only the most recent releases
6515 are available (e.g. keep 5.2 and 5.2.1 but remove 5.1):
6518 $ cd ~ftp/pub/gdb/releases
6523 Update the file @file{README} and @file{.message} in the releases
6530 $ ln README .message
6533 @subsubheading Update the web pages.
6537 @item htdocs/download/ANNOUNCEMENT
6538 This file, which is posted as the official announcement, includes:
6541 General announcement.
6543 News. If making an @var{M}.@var{N}.1 release, retain the news from
6544 earlier @var{M}.@var{N} release.
6549 @item htdocs/index.html
6550 @itemx htdocs/news/index.html
6551 @itemx htdocs/download/index.html
6552 These files include:
6555 Announcement of the most recent release.
6557 News entry (remember to update both the top level and the news directory).
6559 These pages also need to be regenerate using @code{index.sh}.
6561 @item download/onlinedocs/
6562 You need to find the magic command that is used to generate the online
6563 docs from the @file{.tar.bz2}. The best way is to look in the output
6564 from one of the nightly @code{cron} jobs and then just edit accordingly.
6568 $ ~/ss/update-web-docs \
6569 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6571 /www/sourceware/htdocs/gdb/download/onlinedocs \
6576 Just like the online documentation. Something like:
6579 $ /bin/sh ~/ss/update-web-ari \
6580 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6582 /www/sourceware/htdocs/gdb/download/ari \
6588 @subsubheading Shadow the pages onto gnu
6590 Something goes here.
6593 @subsubheading Install the @value{GDBN} tar ball on GNU
6595 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
6596 @file{~ftp/gnu/gdb}.
6598 @subsubheading Make the @file{ANNOUNCEMENT}
6600 Post the @file{ANNOUNCEMENT} file you created above to:
6604 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6606 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
6607 day or so to let things get out)
6609 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
6614 The release is out but you're still not finished.
6616 @subsubheading Commit outstanding changes
6618 In particular you'll need to commit any changes to:
6622 @file{gdb/ChangeLog}
6624 @file{gdb/version.in}
6631 @subsubheading Tag the release
6636 $ d=`date -u +%Y-%m-%d`
6639 $ ( cd insight/src/gdb && cvs -f -q update )
6640 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
6643 Insight is used since that contains more of the release than
6646 @subsubheading Mention the release on the trunk
6648 Just put something in the @file{ChangeLog} so that the trunk also
6649 indicates when the release was made.
6651 @subsubheading Restart @file{gdb/version.in}
6653 If @file{gdb/version.in} does not contain an ISO date such as
6654 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
6655 committed all the release changes it can be set to
6656 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
6657 is important - it affects the snapshot process).
6659 Don't forget the @file{ChangeLog}.
6661 @subsubheading Merge into trunk
6663 The files committed to the branch may also need changes merged into the
6666 @subsubheading Revise the release schedule
6668 Post a revised release schedule to @email{gdb@@sources.redhat.com, GDB
6669 Discussion List} with an updated announcement. The schedule can be
6670 generated by running:
6673 $ ~/ss/schedule `date +%s` schedule
6677 The first parameter is approximate date/time in seconds (from the epoch)
6678 of the most recent release.
6680 Also update the schedule @code{cronjob}.
6682 @section Post release
6684 Remove any @code{OBSOLETE} code.
6691 The testsuite is an important component of the @value{GDBN} package.
6692 While it is always worthwhile to encourage user testing, in practice
6693 this is rarely sufficient; users typically use only a small subset of
6694 the available commands, and it has proven all too common for a change
6695 to cause a significant regression that went unnoticed for some time.
6697 The @value{GDBN} testsuite uses the DejaGNU testing framework. The
6698 tests themselves are calls to various @code{Tcl} procs; the framework
6699 runs all the procs and summarizes the passes and fails.
6701 @section Using the Testsuite
6703 @cindex running the test suite
6704 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
6705 testsuite's objdir) and type @code{make check}. This just sets up some
6706 environment variables and invokes DejaGNU's @code{runtest} script. While
6707 the testsuite is running, you'll get mentions of which test file is in use,
6708 and a mention of any unexpected passes or fails. When the testsuite is
6709 finished, you'll get a summary that looks like this:
6714 # of expected passes 6016
6715 # of unexpected failures 58
6716 # of unexpected successes 5
6717 # of expected failures 183
6718 # of unresolved testcases 3
6719 # of untested testcases 5
6722 To run a specific test script, type:
6724 make check RUNTESTFLAGS='@var{tests}'
6726 where @var{tests} is a list of test script file names, separated by
6729 The ideal test run consists of expected passes only; however, reality
6730 conspires to keep us from this ideal. Unexpected failures indicate
6731 real problems, whether in @value{GDBN} or in the testsuite. Expected
6732 failures are still failures, but ones which have been decided are too
6733 hard to deal with at the time; for instance, a test case might work
6734 everywhere except on AIX, and there is no prospect of the AIX case
6735 being fixed in the near future. Expected failures should not be added
6736 lightly, since you may be masking serious bugs in @value{GDBN}.
6737 Unexpected successes are expected fails that are passing for some
6738 reason, while unresolved and untested cases often indicate some minor
6739 catastrophe, such as the compiler being unable to deal with a test
6742 When making any significant change to @value{GDBN}, you should run the
6743 testsuite before and after the change, to confirm that there are no
6744 regressions. Note that truly complete testing would require that you
6745 run the testsuite with all supported configurations and a variety of
6746 compilers; however this is more than really necessary. In many cases
6747 testing with a single configuration is sufficient. Other useful
6748 options are to test one big-endian (Sparc) and one little-endian (x86)
6749 host, a cross config with a builtin simulator (powerpc-eabi,
6750 mips-elf), or a 64-bit host (Alpha).
6752 If you add new functionality to @value{GDBN}, please consider adding
6753 tests for it as well; this way future @value{GDBN} hackers can detect
6754 and fix their changes that break the functionality you added.
6755 Similarly, if you fix a bug that was not previously reported as a test
6756 failure, please add a test case for it. Some cases are extremely
6757 difficult to test, such as code that handles host OS failures or bugs
6758 in particular versions of compilers, and it's OK not to try to write
6759 tests for all of those.
6761 DejaGNU supports separate build, host, and target machines. However,
6762 some @value{GDBN} test scripts do not work if the build machine and
6763 the host machine are not the same. In such an environment, these scripts
6764 will give a result of ``UNRESOLVED'', like this:
6767 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
6770 @section Testsuite Organization
6772 @cindex test suite organization
6773 The testsuite is entirely contained in @file{gdb/testsuite}. While the
6774 testsuite includes some makefiles and configury, these are very minimal,
6775 and used for little besides cleaning up, since the tests themselves
6776 handle the compilation of the programs that @value{GDBN} will run. The file
6777 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
6778 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
6779 configuration-specific files, typically used for special-purpose
6780 definitions of procs like @code{gdb_load} and @code{gdb_start}.
6782 The tests themselves are to be found in @file{testsuite/gdb.*} and
6783 subdirectories of those. The names of the test files must always end
6784 with @file{.exp}. DejaGNU collects the test files by wildcarding
6785 in the test directories, so both subdirectories and individual files
6786 get chosen and run in alphabetical order.
6788 The following table lists the main types of subdirectories and what they
6789 are for. Since DejaGNU finds test files no matter where they are
6790 located, and since each test file sets up its own compilation and
6791 execution environment, this organization is simply for convenience and
6796 This is the base testsuite. The tests in it should apply to all
6797 configurations of @value{GDBN} (but generic native-only tests may live here).
6798 The test programs should be in the subset of C that is valid K&R,
6799 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
6802 @item gdb.@var{lang}
6803 Language-specific tests for any language @var{lang} besides C. Examples are
6804 @file{gdb.cp} and @file{gdb.java}.
6806 @item gdb.@var{platform}
6807 Non-portable tests. The tests are specific to a specific configuration
6808 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
6811 @item gdb.@var{compiler}
6812 Tests specific to a particular compiler. As of this writing (June
6813 1999), there aren't currently any groups of tests in this category that
6814 couldn't just as sensibly be made platform-specific, but one could
6815 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
6818 @item gdb.@var{subsystem}
6819 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
6820 instance, @file{gdb.disasm} exercises various disassemblers, while
6821 @file{gdb.stabs} tests pathways through the stabs symbol reader.
6824 @section Writing Tests
6825 @cindex writing tests
6827 In many areas, the @value{GDBN} tests are already quite comprehensive; you
6828 should be able to copy existing tests to handle new cases.
6830 You should try to use @code{gdb_test} whenever possible, since it
6831 includes cases to handle all the unexpected errors that might happen.
6832 However, it doesn't cost anything to add new test procedures; for
6833 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
6834 calls @code{gdb_test} multiple times.
6836 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
6837 necessary. Even if @value{GDBN} has several valid responses to
6838 a command, you can use @code{gdb_test_multiple}. Like @code{gdb_test},
6839 @code{gdb_test_multiple} recognizes internal errors and unexpected
6842 Do not write tests which expect a literal tab character from @value{GDBN}.
6843 On some operating systems (e.g.@: OpenBSD) the TTY layer expands tabs to
6844 spaces, so by the time @value{GDBN}'s output reaches expect the tab is gone.
6846 The source language programs do @emph{not} need to be in a consistent
6847 style. Since @value{GDBN} is used to debug programs written in many different
6848 styles, it's worth having a mix of styles in the testsuite; for
6849 instance, some @value{GDBN} bugs involving the display of source lines would
6850 never manifest themselves if the programs used GNU coding style
6857 Check the @file{README} file, it often has useful information that does not
6858 appear anywhere else in the directory.
6861 * Getting Started:: Getting started working on @value{GDBN}
6862 * Debugging GDB:: Debugging @value{GDBN} with itself
6865 @node Getting Started,,, Hints
6867 @section Getting Started
6869 @value{GDBN} is a large and complicated program, and if you first starting to
6870 work on it, it can be hard to know where to start. Fortunately, if you
6871 know how to go about it, there are ways to figure out what is going on.
6873 This manual, the @value{GDBN} Internals manual, has information which applies
6874 generally to many parts of @value{GDBN}.
6876 Information about particular functions or data structures are located in
6877 comments with those functions or data structures. If you run across a
6878 function or a global variable which does not have a comment correctly
6879 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
6880 free to submit a bug report, with a suggested comment if you can figure
6881 out what the comment should say. If you find a comment which is
6882 actually wrong, be especially sure to report that.
6884 Comments explaining the function of macros defined in host, target, or
6885 native dependent files can be in several places. Sometimes they are
6886 repeated every place the macro is defined. Sometimes they are where the
6887 macro is used. Sometimes there is a header file which supplies a
6888 default definition of the macro, and the comment is there. This manual
6889 also documents all the available macros.
6890 @c (@pxref{Host Conditionals}, @pxref{Target
6891 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
6894 Start with the header files. Once you have some idea of how
6895 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
6896 @file{gdbtypes.h}), you will find it much easier to understand the
6897 code which uses and creates those symbol tables.
6899 You may wish to process the information you are getting somehow, to
6900 enhance your understanding of it. Summarize it, translate it to another
6901 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
6902 the code to predict what a test case would do and write the test case
6903 and verify your prediction, etc. If you are reading code and your eyes
6904 are starting to glaze over, this is a sign you need to use a more active
6907 Once you have a part of @value{GDBN} to start with, you can find more
6908 specifically the part you are looking for by stepping through each
6909 function with the @code{next} command. Do not use @code{step} or you
6910 will quickly get distracted; when the function you are stepping through
6911 calls another function try only to get a big-picture understanding
6912 (perhaps using the comment at the beginning of the function being
6913 called) of what it does. This way you can identify which of the
6914 functions being called by the function you are stepping through is the
6915 one which you are interested in. You may need to examine the data
6916 structures generated at each stage, with reference to the comments in
6917 the header files explaining what the data structures are supposed to
6920 Of course, this same technique can be used if you are just reading the
6921 code, rather than actually stepping through it. The same general
6922 principle applies---when the code you are looking at calls something
6923 else, just try to understand generally what the code being called does,
6924 rather than worrying about all its details.
6926 @cindex command implementation
6927 A good place to start when tracking down some particular area is with
6928 a command which invokes that feature. Suppose you want to know how
6929 single-stepping works. As a @value{GDBN} user, you know that the
6930 @code{step} command invokes single-stepping. The command is invoked
6931 via command tables (see @file{command.h}); by convention the function
6932 which actually performs the command is formed by taking the name of
6933 the command and adding @samp{_command}, or in the case of an
6934 @code{info} subcommand, @samp{_info}. For example, the @code{step}
6935 command invokes the @code{step_command} function and the @code{info
6936 display} command invokes @code{display_info}. When this convention is
6937 not followed, you might have to use @code{grep} or @kbd{M-x
6938 tags-search} in emacs, or run @value{GDBN} on itself and set a
6939 breakpoint in @code{execute_command}.
6941 @cindex @code{bug-gdb} mailing list
6942 If all of the above fail, it may be appropriate to ask for information
6943 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
6944 wondering if anyone could give me some tips about understanding
6945 @value{GDBN}''---if we had some magic secret we would put it in this manual.
6946 Suggestions for improving the manual are always welcome, of course.
6948 @node Debugging GDB,,,Hints
6950 @section Debugging @value{GDBN} with itself
6951 @cindex debugging @value{GDBN}
6953 If @value{GDBN} is limping on your machine, this is the preferred way to get it
6954 fully functional. Be warned that in some ancient Unix systems, like
6955 Ultrix 4.2, a program can't be running in one process while it is being
6956 debugged in another. Rather than typing the command @kbd{@w{./gdb
6957 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
6958 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
6960 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
6961 @file{.gdbinit} file that sets up some simple things to make debugging
6962 gdb easier. The @code{info} command, when executed without a subcommand
6963 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
6964 gdb. See @file{.gdbinit} for details.
6966 If you use emacs, you will probably want to do a @code{make TAGS} after
6967 you configure your distribution; this will put the machine dependent
6968 routines for your local machine where they will be accessed first by
6971 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
6972 have run @code{fixincludes} if you are compiling with gcc.
6974 @section Submitting Patches
6976 @cindex submitting patches
6977 Thanks for thinking of offering your changes back to the community of
6978 @value{GDBN} users. In general we like to get well designed enhancements.
6979 Thanks also for checking in advance about the best way to transfer the
6982 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
6983 This manual summarizes what we believe to be clean design for @value{GDBN}.
6985 If the maintainers don't have time to put the patch in when it arrives,
6986 or if there is any question about a patch, it goes into a large queue
6987 with everyone else's patches and bug reports.
6989 @cindex legal papers for code contributions
6990 The legal issue is that to incorporate substantial changes requires a
6991 copyright assignment from you and/or your employer, granting ownership
6992 of the changes to the Free Software Foundation. You can get the
6993 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
6994 and asking for it. We recommend that people write in "All programs
6995 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
6996 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
6998 contributed with only one piece of legalese pushed through the
6999 bureaucracy and filed with the FSF. We can't start merging changes until
7000 this paperwork is received by the FSF (their rules, which we follow
7001 since we maintain it for them).
7003 Technically, the easiest way to receive changes is to receive each
7004 feature as a small context diff or unidiff, suitable for @code{patch}.
7005 Each message sent to me should include the changes to C code and
7006 header files for a single feature, plus @file{ChangeLog} entries for
7007 each directory where files were modified, and diffs for any changes
7008 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
7009 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
7010 single feature, they can be split down into multiple messages.
7012 In this way, if we read and like the feature, we can add it to the
7013 sources with a single patch command, do some testing, and check it in.
7014 If you leave out the @file{ChangeLog}, we have to write one. If you leave
7015 out the doc, we have to puzzle out what needs documenting. Etc., etc.
7017 The reason to send each change in a separate message is that we will not
7018 install some of the changes. They'll be returned to you with questions
7019 or comments. If we're doing our job correctly, the message back to you
7020 will say what you have to fix in order to make the change acceptable.
7021 The reason to have separate messages for separate features is so that
7022 the acceptable changes can be installed while one or more changes are
7023 being reworked. If multiple features are sent in a single message, we
7024 tend to not put in the effort to sort out the acceptable changes from
7025 the unacceptable, so none of the features get installed until all are
7028 If this sounds painful or authoritarian, well, it is. But we get a lot
7029 of bug reports and a lot of patches, and many of them don't get
7030 installed because we don't have the time to finish the job that the bug
7031 reporter or the contributor could have done. Patches that arrive
7032 complete, working, and well designed, tend to get installed on the day
7033 they arrive. The others go into a queue and get installed as time
7034 permits, which, since the maintainers have many demands to meet, may not
7035 be for quite some time.
7037 Please send patches directly to
7038 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} maintainers}.
7040 @section Obsolete Conditionals
7041 @cindex obsolete code
7043 Fragments of old code in @value{GDBN} sometimes reference or set the following
7044 configuration macros. They should not be used by new code, and old uses
7045 should be removed as those parts of the debugger are otherwise touched.
7048 @item STACK_END_ADDR
7049 This macro used to define where the end of the stack appeared, for use
7050 in interpreting core file formats that don't record this address in the
7051 core file itself. This information is now configured in BFD, and @value{GDBN}
7052 gets the info portably from there. The values in @value{GDBN}'s configuration
7053 files should be moved into BFD configuration files (if needed there),
7054 and deleted from all of @value{GDBN}'s config files.
7056 Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR
7057 is so old that it has never been converted to use BFD. Now that's old!
7061 @include observer.texi