1 \input texinfo @c -*- texinfo -*-
2 @setfilename gdbint.info
4 @dircategory Software development
6 * Gdb-Internals: (gdbint). The GNU debugger's internals.
10 Copyright @copyright{} 1990, 1991, 1992, 1993, 1994, 1996, 1998, 1999,
11 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2008, 2009
12 Free Software Foundation, Inc.
13 Contributed by Cygnus Solutions. Written by John Gilmore.
14 Second Edition by Stan Shebs.
16 Permission is granted to copy, distribute and/or modify this document
17 under the terms of the GNU Free Documentation License, Version 1.1 or
18 any later version published by the Free Software Foundation; with no
19 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
20 Texts. A copy of the license is included in the section entitled ``GNU
21 Free Documentation License''.
25 This file documents the internals of the GNU debugger @value{GDBN}.
30 @setchapternewpage off
31 @settitle @value{GDBN} Internals
37 @title @value{GDBN} Internals
38 @subtitle{A guide to the internals of the GNU debugger}
40 @author Cygnus Solutions
41 @author Second Edition:
43 @author Cygnus Solutions
46 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
47 \xdef\manvers{\$Revision$} % For use in headers, footers too
49 \hfill Cygnus Solutions\par
51 \hfill \TeX{}info \texinfoversion\par
55 @vskip 0pt plus 1filll
62 @c Perhaps this should be the title of the document (but only for info,
63 @c not for TeX). Existing GNU manuals seem inconsistent on this point.
64 @top Scope of this Document
66 This document documents the internals of the GNU debugger, @value{GDBN}. It
67 includes description of @value{GDBN}'s key algorithms and operations, as well
68 as the mechanisms that adapt @value{GDBN} to specific hosts and targets.
81 * Target Architecture Definition::
82 * Target Descriptions::
83 * Target Vector Definition::
88 * Versions and Branches::
89 * Start of New Year Procedure::
94 * GDB Observers:: @value{GDBN} Currently available observers
95 * GNU Free Documentation License:: The license for this documentation
108 @section Requirements
109 @cindex requirements for @value{GDBN}
111 Before diving into the internals, you should understand the formal
112 requirements and other expectations for @value{GDBN}. Although some
113 of these may seem obvious, there have been proposals for @value{GDBN}
114 that have run counter to these requirements.
116 First of all, @value{GDBN} is a debugger. It's not designed to be a
117 front panel for embedded systems. It's not a text editor. It's not a
118 shell. It's not a programming environment.
120 @value{GDBN} is an interactive tool. Although a batch mode is
121 available, @value{GDBN}'s primary role is to interact with a human
124 @value{GDBN} should be responsive to the user. A programmer hot on
125 the trail of a nasty bug, and operating under a looming deadline, is
126 going to be very impatient of everything, including the response time
127 to debugger commands.
129 @value{GDBN} should be relatively permissive, such as for expressions.
130 While the compiler should be picky (or have the option to be made
131 picky), since source code lives for a long time usually, the
132 programmer doing debugging shouldn't be spending time figuring out to
133 mollify the debugger.
135 @value{GDBN} will be called upon to deal with really large programs.
136 Executable sizes of 50 to 100 megabytes occur regularly, and we've
137 heard reports of programs approaching 1 gigabyte in size.
139 @value{GDBN} should be able to run everywhere. No other debugger is
140 available for even half as many configurations as @value{GDBN}
144 @section Contributors
146 The first edition of this document was written by John Gilmore of
147 Cygnus Solutions. The current second edition was written by Stan Shebs
148 of Cygnus Solutions, who continues to update the manual.
150 Over the years, many others have made additions and changes to this
151 document. This section attempts to record the significant contributors
152 to that effort. One of the virtues of free software is that everyone
153 is free to contribute to it; with regret, we cannot actually
154 acknowledge everyone here.
157 @emph{Plea:} This section has only been added relatively recently (four
158 years after publication of the second edition). Additions to this
159 section are particularly welcome. If you or your friends (or enemies,
160 to be evenhanded) have been unfairly omitted from this list, we would
161 like to add your names!
164 A document such as this relies on being kept up to date by numerous
165 small updates by contributing engineers as they make changes to the
166 code base. The file @file{ChangeLog} in the @value{GDBN} distribution
167 approximates a blow-by-blow account. The most prolific contributors to
168 this important, but low profile task are Andrew Cagney (responsible
169 for over half the entries), Daniel Jacobowitz, Mark Kettenis, Jim
170 Blandy and Eli Zaretskii.
172 Eli Zaretskii and Daniel Jacobowitz wrote the sections documenting
175 Jeremy Bennett updated the sections on initializing a new architecture
176 and register representation, and added the section on Frame Interpretation.
179 @node Overall Structure
181 @chapter Overall Structure
183 @value{GDBN} consists of three major subsystems: user interface,
184 symbol handling (the @dfn{symbol side}), and target system handling (the
187 The user interface consists of several actual interfaces, plus
190 The symbol side consists of object file readers, debugging info
191 interpreters, symbol table management, source language expression
192 parsing, type and value printing.
194 The target side consists of execution control, stack frame analysis, and
195 physical target manipulation.
197 The target side/symbol side division is not formal, and there are a
198 number of exceptions. For instance, core file support involves symbolic
199 elements (the basic core file reader is in BFD) and target elements (it
200 supplies the contents of memory and the values of registers). Instead,
201 this division is useful for understanding how the minor subsystems
204 @section The Symbol Side
206 The symbolic side of @value{GDBN} can be thought of as ``everything
207 you can do in @value{GDBN} without having a live program running''.
208 For instance, you can look at the types of variables, and evaluate
209 many kinds of expressions.
211 @section The Target Side
213 The target side of @value{GDBN} is the ``bits and bytes manipulator''.
214 Although it may make reference to symbolic info here and there, most
215 of the target side will run with only a stripped executable
216 available---or even no executable at all, in remote debugging cases.
218 Operations such as disassembly, stack frame crawls, and register
219 display, are able to work with no symbolic info at all. In some cases,
220 such as disassembly, @value{GDBN} will use symbolic info to present addresses
221 relative to symbols rather than as raw numbers, but it will work either
224 @section Configurations
228 @dfn{Host} refers to attributes of the system where @value{GDBN} runs.
229 @dfn{Target} refers to the system where the program being debugged
230 executes. In most cases they are the same machine, in which case a
231 third type of @dfn{Native} attributes come into play.
233 Defines and include files needed to build on the host are host
234 support. Examples are tty support, system defined types, host byte
235 order, host float format. These are all calculated by @code{autoconf}
236 when the debugger is built.
238 Defines and information needed to handle the target format are target
239 dependent. Examples are the stack frame format, instruction set,
240 breakpoint instruction, registers, and how to set up and tear down the stack
243 Information that is only needed when the host and target are the same,
244 is native dependent. One example is Unix child process support; if the
245 host and target are not the same, calling @code{fork} to start the target
246 process is a bad idea. The various macros needed for finding the
247 registers in the @code{upage}, running @code{ptrace}, and such are all
248 in the native-dependent files.
250 Another example of native-dependent code is support for features that
251 are really part of the target environment, but which require
252 @code{#include} files that are only available on the host system. Core
253 file handling and @code{setjmp} handling are two common cases.
255 When you want to make @value{GDBN} work as the traditional native debugger
256 on a system, you will need to supply both target and native information.
258 @section Source Tree Structure
259 @cindex @value{GDBN} source tree structure
261 The @value{GDBN} source directory has a mostly flat structure---there
262 are only a few subdirectories. A file's name usually gives a hint as
263 to what it does; for example, @file{stabsread.c} reads stabs,
264 @file{dwarf2read.c} reads @sc{DWARF 2}, etc.
266 Files that are related to some common task have names that share
267 common substrings. For example, @file{*-thread.c} files deal with
268 debugging threads on various platforms; @file{*read.c} files deal with
269 reading various kinds of symbol and object files; @file{inf*.c} files
270 deal with direct control of the @dfn{inferior program} (@value{GDBN}
271 parlance for the program being debugged).
273 There are several dozens of files in the @file{*-tdep.c} family.
274 @samp{tdep} stands for @dfn{target-dependent code}---each of these
275 files implements debug support for a specific target architecture
276 (sparc, mips, etc). Usually, only one of these will be used in a
277 specific @value{GDBN} configuration (sometimes two, closely related).
279 Similarly, there are many @file{*-nat.c} files, each one for native
280 debugging on a specific system (e.g., @file{sparc-linux-nat.c} is for
281 native debugging of Sparc machines running the Linux kernel).
283 The few subdirectories of the source tree are:
287 Code that implements @dfn{CLI}, the @value{GDBN} Command-Line
288 Interpreter. @xref{User Interface, Command Interpreter}.
291 Code for the @value{GDBN} remote server.
294 Code for Insight, the @value{GDBN} TK-based GUI front-end.
297 The @dfn{GDB/MI}, the @value{GDBN} Machine Interface interpreter.
300 Target signal translation code.
303 Code for @dfn{TUI}, the @value{GDBN} Text-mode full-screen User
304 Interface. @xref{User Interface, TUI}.
312 @value{GDBN} uses a number of debugging-specific algorithms. They are
313 often not very complicated, but get lost in the thicket of special
314 cases and real-world issues. This chapter describes the basic
315 algorithms and mentions some of the specific target definitions that
318 @section Prologue Analysis
320 @cindex prologue analysis
321 @cindex call frame information
322 @cindex CFI (call frame information)
323 To produce a backtrace and allow the user to manipulate older frames'
324 variables and arguments, @value{GDBN} needs to find the base addresses
325 of older frames, and discover where those frames' registers have been
326 saved. Since a frame's ``callee-saves'' registers get saved by
327 younger frames if and when they're reused, a frame's registers may be
328 scattered unpredictably across younger frames. This means that
329 changing the value of a register-allocated variable in an older frame
330 may actually entail writing to a save slot in some younger frame.
332 Modern versions of GCC emit Dwarf call frame information (``CFI''),
333 which describes how to find frame base addresses and saved registers.
334 But CFI is not always available, so as a fallback @value{GDBN} uses a
335 technique called @dfn{prologue analysis} to find frame sizes and saved
336 registers. A prologue analyzer disassembles the function's machine
337 code starting from its entry point, and looks for instructions that
338 allocate frame space, save the stack pointer in a frame pointer
339 register, save registers, and so on. Obviously, this can't be done
340 accurately in general, but it's tractable to do well enough to be very
341 helpful. Prologue analysis predates the GNU toolchain's support for
342 CFI; at one time, prologue analysis was the only mechanism
343 @value{GDBN} used for stack unwinding at all, when the function
344 calling conventions didn't specify a fixed frame layout.
346 In the olden days, function prologues were generated by hand-written,
347 target-specific code in GCC, and treated as opaque and untouchable by
348 optimizers. Looking at this code, it was usually straightforward to
349 write a prologue analyzer for @value{GDBN} that would accurately
350 understand all the prologues GCC would generate. However, over time
351 GCC became more aggressive about instruction scheduling, and began to
352 understand more about the semantics of the prologue instructions
353 themselves; in response, @value{GDBN}'s analyzers became more complex
354 and fragile. Keeping the prologue analyzers working as GCC (and the
355 instruction sets themselves) evolved became a substantial task.
357 @cindex @file{prologue-value.c}
358 @cindex abstract interpretation of function prologues
359 @cindex pseudo-evaluation of function prologues
360 To try to address this problem, the code in @file{prologue-value.h}
361 and @file{prologue-value.c} provides a general framework for writing
362 prologue analyzers that are simpler and more robust than ad-hoc
363 analyzers. When we analyze a prologue using the prologue-value
364 framework, we're really doing ``abstract interpretation'' or
365 ``pseudo-evaluation'': running the function's code in simulation, but
366 using conservative approximations of the values registers and memory
367 would hold when the code actually runs. For example, if our function
368 starts with the instruction:
371 addi r1, 42 # add 42 to r1
374 we don't know exactly what value will be in @code{r1} after executing
375 this instruction, but we do know it'll be 42 greater than its original
378 If we then see an instruction like:
381 addi r1, 22 # add 22 to r1
384 we still don't know what @code{r1's} value is, but again, we can say
385 it is now 64 greater than its original value.
387 If the next instruction were:
390 mov r2, r1 # set r2 to r1's value
393 then we can say that @code{r2's} value is now the original value of
396 It's common for prologues to save registers on the stack, so we'll
397 need to track the values of stack frame slots, as well as the
398 registers. So after an instruction like this:
404 then we'd know that the stack slot four bytes above the frame pointer
405 holds the original value of @code{r1} plus 64.
409 Of course, this can only go so far before it gets unreasonable. If we
410 wanted to be able to say anything about the value of @code{r1} after
414 xor r1, r3 # exclusive-or r1 and r3, place result in r1
417 then things would get pretty complex. But remember, we're just doing
418 a conservative approximation; if exclusive-or instructions aren't
419 relevant to prologues, we can just say @code{r1}'s value is now
420 ``unknown''. We can ignore things that are too complex, if that loss of
421 information is acceptable for our application.
423 So when we say ``conservative approximation'' here, what we mean is an
424 approximation that is either accurate, or marked ``unknown'', but
427 Using this framework, a prologue analyzer is simply an interpreter for
428 machine code, but one that uses conservative approximations for the
429 contents of registers and memory instead of actual values. Starting
430 from the function's entry point, you simulate instructions up to the
431 current PC, or an instruction that you don't know how to simulate.
432 Now you can examine the state of the registers and stack slots you've
438 To see how large your stack frame is, just check the value of the
439 stack pointer register; if it's the original value of the SP
440 minus a constant, then that constant is the stack frame's size.
441 If the SP's value has been marked as ``unknown'', then that means
442 the prologue has done something too complex for us to track, and
443 we don't know the frame size.
446 To see where we've saved the previous frame's registers, we just
447 search the values we've tracked --- stack slots, usually, but
448 registers, too, if you want --- for something equal to the register's
449 original value. If the calling conventions suggest a standard place
450 to save a given register, then we can check there first, but really,
451 anything that will get us back the original value will probably work.
454 This does take some work. But prologue analyzers aren't
455 quick-and-simple pattern patching to recognize a few fixed prologue
456 forms any more; they're big, hairy functions. Along with inferior
457 function calls, prologue analysis accounts for a substantial portion
458 of the time needed to stabilize a @value{GDBN} port. So it's
459 worthwhile to look for an approach that will be easier to understand
460 and maintain. In the approach described above:
465 It's easier to see that the analyzer is correct: you just see
466 whether the analyzer properly (albeit conservatively) simulates
467 the effect of each instruction.
470 It's easier to extend the analyzer: you can add support for new
471 instructions, and know that you haven't broken anything that
472 wasn't already broken before.
475 It's orthogonal: to gather new information, you don't need to
476 complicate the code for each instruction. As long as your domain
477 of conservative values is already detailed enough to tell you
478 what you need, then all the existing instruction simulations are
479 already gathering the right data for you.
483 The file @file{prologue-value.h} contains detailed comments explaining
484 the framework and how to use it.
487 @section Breakpoint Handling
490 In general, a breakpoint is a user-designated location in the program
491 where the user wants to regain control if program execution ever reaches
494 There are two main ways to implement breakpoints; either as ``hardware''
495 breakpoints or as ``software'' breakpoints.
497 @cindex hardware breakpoints
498 @cindex program counter
499 Hardware breakpoints are sometimes available as a builtin debugging
500 features with some chips. Typically these work by having dedicated
501 register into which the breakpoint address may be stored. If the PC
502 (shorthand for @dfn{program counter})
503 ever matches a value in a breakpoint registers, the CPU raises an
504 exception and reports it to @value{GDBN}.
506 Another possibility is when an emulator is in use; many emulators
507 include circuitry that watches the address lines coming out from the
508 processor, and force it to stop if the address matches a breakpoint's
511 A third possibility is that the target already has the ability to do
512 breakpoints somehow; for instance, a ROM monitor may do its own
513 software breakpoints. So although these are not literally ``hardware
514 breakpoints'', from @value{GDBN}'s point of view they work the same;
515 @value{GDBN} need not do anything more than set the breakpoint and wait
516 for something to happen.
518 Since they depend on hardware resources, hardware breakpoints may be
519 limited in number; when the user asks for more, @value{GDBN} will
520 start trying to set software breakpoints. (On some architectures,
521 notably the 32-bit x86 platforms, @value{GDBN} cannot always know
522 whether there's enough hardware resources to insert all the hardware
523 breakpoints and watchpoints. On those platforms, @value{GDBN} prints
524 an error message only when the program being debugged is continued.)
526 @cindex software breakpoints
527 Software breakpoints require @value{GDBN} to do somewhat more work.
528 The basic theory is that @value{GDBN} will replace a program
529 instruction with a trap, illegal divide, or some other instruction
530 that will cause an exception, and then when it's encountered,
531 @value{GDBN} will take the exception and stop the program. When the
532 user says to continue, @value{GDBN} will restore the original
533 instruction, single-step, re-insert the trap, and continue on.
535 Since it literally overwrites the program being tested, the program area
536 must be writable, so this technique won't work on programs in ROM. It
537 can also distort the behavior of programs that examine themselves,
538 although such a situation would be highly unusual.
540 Also, the software breakpoint instruction should be the smallest size of
541 instruction, so it doesn't overwrite an instruction that might be a jump
542 target, and cause disaster when the program jumps into the middle of the
543 breakpoint instruction. (Strictly speaking, the breakpoint must be no
544 larger than the smallest interval between instructions that may be jump
545 targets; perhaps there is an architecture where only even-numbered
546 instructions may jumped to.) Note that it's possible for an instruction
547 set not to have any instructions usable for a software breakpoint,
548 although in practice only the ARC has failed to define such an
551 Basic breakpoint object handling is in @file{breakpoint.c}. However,
552 much of the interesting breakpoint action is in @file{infrun.c}.
555 @cindex insert or remove software breakpoint
556 @findex target_remove_breakpoint
557 @findex target_insert_breakpoint
558 @item target_remove_breakpoint (@var{bp_tgt})
559 @itemx target_insert_breakpoint (@var{bp_tgt})
560 Insert or remove a software breakpoint at address
561 @code{@var{bp_tgt}->placed_address}. Returns zero for success,
562 non-zero for failure. On input, @var{bp_tgt} contains the address of the
563 breakpoint, and is otherwise initialized to zero. The fields of the
564 @code{struct bp_target_info} pointed to by @var{bp_tgt} are updated
565 to contain other information about the breakpoint on output. The field
566 @code{placed_address} may be updated if the breakpoint was placed at a
567 related address; the field @code{shadow_contents} contains the real
568 contents of the bytes where the breakpoint has been inserted,
569 if reading memory would return the breakpoint instead of the
570 underlying memory; the field @code{shadow_len} is the length of
571 memory cached in @code{shadow_contents}, if any; and the field
572 @code{placed_size} is optionally set and used by the target, if
573 it could differ from @code{shadow_len}.
575 For example, the remote target @samp{Z0} packet does not require
576 shadowing memory, so @code{shadow_len} is left at zero. However,
577 the length reported by @code{gdbarch_breakpoint_from_pc} is cached in
578 @code{placed_size}, so that a matching @samp{z0} packet can be
579 used to remove the breakpoint.
581 @cindex insert or remove hardware breakpoint
582 @findex target_remove_hw_breakpoint
583 @findex target_insert_hw_breakpoint
584 @item target_remove_hw_breakpoint (@var{bp_tgt})
585 @itemx target_insert_hw_breakpoint (@var{bp_tgt})
586 Insert or remove a hardware-assisted breakpoint at address
587 @code{@var{bp_tgt}->placed_address}. Returns zero for success,
588 non-zero for failure. See @code{target_insert_breakpoint} for
589 a description of the @code{struct bp_target_info} pointed to by
590 @var{bp_tgt}; the @code{shadow_contents} and
591 @code{shadow_len} members are not used for hardware breakpoints,
592 but @code{placed_size} may be.
595 @section Single Stepping
597 @section Signal Handling
599 @section Thread Handling
601 @section Inferior Function Calls
603 @section Longjmp Support
605 @cindex @code{longjmp} debugging
606 @value{GDBN} has support for figuring out that the target is doing a
607 @code{longjmp} and for stopping at the target of the jump, if we are
608 stepping. This is done with a few specialized internal breakpoints,
609 which are visible in the output of the @samp{maint info breakpoint}
612 @findex gdbarch_get_longjmp_target
613 To make this work, you need to define a function called
614 @code{gdbarch_get_longjmp_target}, which will examine the
615 @code{jmp_buf} structure and extract the @code{longjmp} target address.
616 Since @code{jmp_buf} is target specific and typically defined in a
617 target header not available to @value{GDBN}, you will need to
618 determine the offset of the PC manually and return that; many targets
619 define a @code{jb_pc_offset} field in the tdep structure to save the
620 value once calculated.
625 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
626 breakpoints}) which break when data is accessed rather than when some
627 instruction is executed. When you have data which changes without
628 your knowing what code does that, watchpoints are the silver bullet to
629 hunt down and kill such bugs.
631 @cindex hardware watchpoints
632 @cindex software watchpoints
633 Watchpoints can be either hardware-assisted or not; the latter type is
634 known as ``software watchpoints.'' @value{GDBN} always uses
635 hardware-assisted watchpoints if they are available, and falls back on
636 software watchpoints otherwise. Typical situations where @value{GDBN}
637 will use software watchpoints are:
641 The watched memory region is too large for the underlying hardware
642 watchpoint support. For example, each x86 debug register can watch up
643 to 4 bytes of memory, so trying to watch data structures whose size is
644 more than 16 bytes will cause @value{GDBN} to use software
648 The value of the expression to be watched depends on data held in
649 registers (as opposed to memory).
652 Too many different watchpoints requested. (On some architectures,
653 this situation is impossible to detect until the debugged program is
654 resumed.) Note that x86 debug registers are used both for hardware
655 breakpoints and for watchpoints, so setting too many hardware
656 breakpoints might cause watchpoint insertion to fail.
659 No hardware-assisted watchpoints provided by the target
663 Software watchpoints are very slow, since @value{GDBN} needs to
664 single-step the program being debugged and test the value of the
665 watched expression(s) after each instruction. The rest of this
666 section is mostly irrelevant for software watchpoints.
668 When the inferior stops, @value{GDBN} tries to establish, among other
669 possible reasons, whether it stopped due to a watchpoint being hit.
670 It first uses @code{STOPPED_BY_WATCHPOINT} to see if any watchpoint
671 was hit. If not, all watchpoint checking is skipped.
673 Then @value{GDBN} calls @code{target_stopped_data_address} exactly
674 once. This method returns the address of the watchpoint which
675 triggered, if the target can determine it. If the triggered address
676 is available, @value{GDBN} compares the address returned by this
677 method with each watched memory address in each active watchpoint.
678 For data-read and data-access watchpoints, @value{GDBN} announces
679 every watchpoint that watches the triggered address as being hit.
680 For this reason, data-read and data-access watchpoints
681 @emph{require} that the triggered address be available; if not, read
682 and access watchpoints will never be considered hit. For data-write
683 watchpoints, if the triggered address is available, @value{GDBN}
684 considers only those watchpoints which match that address;
685 otherwise, @value{GDBN} considers all data-write watchpoints. For
686 each data-write watchpoint that @value{GDBN} considers, it evaluates
687 the expression whose value is being watched, and tests whether the
688 watched value has changed. Watchpoints whose watched values have
689 changed are announced as hit.
691 @c FIXME move these to the main lists of target/native defns
693 @value{GDBN} uses several macros and primitives to support hardware
697 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
698 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
699 Return the number of hardware watchpoints of type @var{type} that are
700 possible to be set. The value is positive if @var{count} watchpoints
701 of this type can be set, zero if setting watchpoints of this type is
702 not supported, and negative if @var{count} is more than the maximum
703 number of watchpoints of type @var{type} that can be set. @var{other}
704 is non-zero if other types of watchpoints are currently enabled (there
705 are architectures which cannot set watchpoints of different types at
708 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
709 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
710 Return non-zero if hardware watchpoints can be used to watch a region
711 whose address is @var{addr} and whose length in bytes is @var{len}.
713 @cindex insert or remove hardware watchpoint
714 @findex target_insert_watchpoint
715 @findex target_remove_watchpoint
716 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
717 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
718 Insert or remove a hardware watchpoint starting at @var{addr}, for
719 @var{len} bytes. @var{type} is the watchpoint type, one of the
720 possible values of the enumerated data type @code{target_hw_bp_type},
721 defined by @file{breakpoint.h} as follows:
724 enum target_hw_bp_type
726 hw_write = 0, /* Common (write) HW watchpoint */
727 hw_read = 1, /* Read HW watchpoint */
728 hw_access = 2, /* Access (read or write) HW watchpoint */
729 hw_execute = 3 /* Execute HW breakpoint */
734 These two macros should return 0 for success, non-zero for failure.
736 @findex target_stopped_data_address
737 @item target_stopped_data_address (@var{addr_p})
738 If the inferior has some watchpoint that triggered, place the address
739 associated with the watchpoint at the location pointed to by
740 @var{addr_p} and return non-zero. Otherwise, return zero. This
741 is required for data-read and data-access watchpoints. It is
742 not required for data-write watchpoints, but @value{GDBN} uses
743 it to improve handling of those also.
745 @value{GDBN} will only call this method once per watchpoint stop,
746 immediately after calling @code{STOPPED_BY_WATCHPOINT}. If the
747 target's watchpoint indication is sticky, i.e., stays set after
748 resuming, this method should clear it. For instance, the x86 debug
749 control register has sticky triggered flags.
751 @findex target_watchpoint_addr_within_range
752 @item target_watchpoint_addr_within_range (@var{target}, @var{addr}, @var{start}, @var{length})
753 Check whether @var{addr} (as returned by @code{target_stopped_data_address})
754 lies within the hardware-defined watchpoint region described by
755 @var{start} and @var{length}. This only needs to be provided if the
756 granularity of a watchpoint is greater than one byte, i.e., if the
757 watchpoint can also trigger on nearby addresses outside of the watched
760 @findex HAVE_STEPPABLE_WATCHPOINT
761 @item HAVE_STEPPABLE_WATCHPOINT
762 If defined to a non-zero value, it is not necessary to disable a
763 watchpoint to step over it. Like @code{gdbarch_have_nonsteppable_watchpoint},
764 this is usually set when watchpoints trigger at the instruction
765 which will perform an interesting read or write. It should be
766 set if there is a temporary disable bit which allows the processor
767 to step over the interesting instruction without raising the
768 watchpoint exception again.
770 @findex gdbarch_have_nonsteppable_watchpoint
771 @item int gdbarch_have_nonsteppable_watchpoint (@var{gdbarch})
772 If it returns a non-zero value, @value{GDBN} should disable a
773 watchpoint to step the inferior over it. This is usually set when
774 watchpoints trigger at the instruction which will perform an
775 interesting read or write.
777 @findex HAVE_CONTINUABLE_WATCHPOINT
778 @item HAVE_CONTINUABLE_WATCHPOINT
779 If defined to a non-zero value, it is possible to continue the
780 inferior after a watchpoint has been hit. This is usually set
781 when watchpoints trigger at the instruction following an interesting
784 @findex CANNOT_STEP_HW_WATCHPOINTS
785 @item CANNOT_STEP_HW_WATCHPOINTS
786 If this is defined to a non-zero value, @value{GDBN} will remove all
787 watchpoints before stepping the inferior.
789 @findex STOPPED_BY_WATCHPOINT
790 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
791 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
792 the type @code{struct target_waitstatus}, defined by @file{target.h}.
793 Normally, this macro is defined to invoke the function pointed to by
794 the @code{to_stopped_by_watchpoint} member of the structure (of the
795 type @code{target_ops}, defined on @file{target.h}) that describes the
796 target-specific operations; @code{to_stopped_by_watchpoint} ignores
797 the @var{wait_status} argument.
799 @value{GDBN} does not require the non-zero value returned by
800 @code{STOPPED_BY_WATCHPOINT} to be 100% correct, so if a target cannot
801 determine for sure whether the inferior stopped due to a watchpoint,
802 it could return non-zero ``just in case''.
805 @subsection Watchpoints and Threads
806 @cindex watchpoints, with threads
808 @value{GDBN} only supports process-wide watchpoints, which trigger
809 in all threads. @value{GDBN} uses the thread ID to make watchpoints
810 act as if they were thread-specific, but it cannot set hardware
811 watchpoints that only trigger in a specific thread. Therefore, even
812 if the target supports threads, per-thread debug registers, and
813 watchpoints which only affect a single thread, it should set the
814 per-thread debug registers for all threads to the same value. On
815 @sc{gnu}/Linux native targets, this is accomplished by using
816 @code{ALL_LWPS} in @code{target_insert_watchpoint} and
817 @code{target_remove_watchpoint} and by using
818 @code{linux_set_new_thread} to register a handler for newly created
821 @value{GDBN}'s @sc{gnu}/Linux support only reports a single event
822 at a time, although multiple events can trigger simultaneously for
823 multi-threaded programs. When multiple events occur, @file{linux-nat.c}
824 queues subsequent events and returns them the next time the program
825 is resumed. This means that @code{STOPPED_BY_WATCHPOINT} and
826 @code{target_stopped_data_address} only need to consult the current
827 thread's state---the thread indicated by @code{inferior_ptid}. If
828 two threads have hit watchpoints simultaneously, those routines
829 will be called a second time for the second thread.
831 @subsection x86 Watchpoints
832 @cindex x86 debug registers
833 @cindex watchpoints, on x86
835 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
836 registers designed to facilitate debugging. @value{GDBN} provides a
837 generic library of functions that x86-based ports can use to implement
838 support for watchpoints and hardware-assisted breakpoints. This
839 subsection documents the x86 watchpoint facilities in @value{GDBN}.
841 (At present, the library functions read and write debug registers directly, and are
842 thus only available for native configurations.)
844 To use the generic x86 watchpoint support, a port should do the
848 @findex I386_USE_GENERIC_WATCHPOINTS
850 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
851 target-dependent headers.
854 Include the @file{config/i386/nm-i386.h} header file @emph{after}
855 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
858 Add @file{i386-nat.o} to the value of the Make variable
859 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}).
862 Provide implementations for the @code{I386_DR_LOW_*} macros described
863 below. Typically, each macro should call a target-specific function
864 which does the real work.
867 The x86 watchpoint support works by maintaining mirror images of the
868 debug registers. Values are copied between the mirror images and the
869 real debug registers via a set of macros which each target needs to
873 @findex I386_DR_LOW_SET_CONTROL
874 @item I386_DR_LOW_SET_CONTROL (@var{val})
875 Set the Debug Control (DR7) register to the value @var{val}.
877 @findex I386_DR_LOW_SET_ADDR
878 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
879 Put the address @var{addr} into the debug register number @var{idx}.
881 @findex I386_DR_LOW_RESET_ADDR
882 @item I386_DR_LOW_RESET_ADDR (@var{idx})
883 Reset (i.e.@: zero out) the address stored in the debug register
886 @findex I386_DR_LOW_GET_STATUS
887 @item I386_DR_LOW_GET_STATUS
888 Return the value of the Debug Status (DR6) register. This value is
889 used immediately after it is returned by
890 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
894 For each one of the 4 debug registers (whose indices are from 0 to 3)
895 that store addresses, a reference count is maintained by @value{GDBN},
896 to allow sharing of debug registers by several watchpoints. This
897 allows users to define several watchpoints that watch the same
898 expression, but with different conditions and/or commands, without
899 wasting debug registers which are in short supply. @value{GDBN}
900 maintains the reference counts internally, targets don't have to do
901 anything to use this feature.
903 The x86 debug registers can each watch a region that is 1, 2, or 4
904 bytes long. The ia32 architecture requires that each watched region
905 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
906 region on 4-byte boundary. However, the x86 watchpoint support in
907 @value{GDBN} can watch unaligned regions and regions larger than 4
908 bytes (up to 16 bytes) by allocating several debug registers to watch
909 a single region. This allocation of several registers per a watched
910 region is also done automatically without target code intervention.
912 The generic x86 watchpoint support provides the following API for the
913 @value{GDBN}'s application code:
916 @findex i386_region_ok_for_watchpoint
917 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
918 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
919 this function. It counts the number of debug registers required to
920 watch a given region, and returns a non-zero value if that number is
921 less than 4, the number of debug registers available to x86
924 @findex i386_stopped_data_address
925 @item i386_stopped_data_address (@var{addr_p})
927 @code{target_stopped_data_address} is set to call this function.
929 function examines the breakpoint condition bits in the DR6 Debug
930 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
931 macro, and returns the address associated with the first bit that is
934 @findex i386_stopped_by_watchpoint
935 @item i386_stopped_by_watchpoint (void)
936 The macro @code{STOPPED_BY_WATCHPOINT}
937 is set to call this function. The
938 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
939 function examines the breakpoint condition bits in the DR6 Debug
940 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
941 macro, and returns true if any bit is set. Otherwise, false is
944 @findex i386_insert_watchpoint
945 @findex i386_remove_watchpoint
946 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
947 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
948 Insert or remove a watchpoint. The macros
949 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
950 are set to call these functions. @code{i386_insert_watchpoint} first
951 looks for a debug register which is already set to watch the same
952 region for the same access types; if found, it just increments the
953 reference count of that debug register, thus implementing debug
954 register sharing between watchpoints. If no such register is found,
955 the function looks for a vacant debug register, sets its mirrored
956 value to @var{addr}, sets the mirrored value of DR7 Debug Control
957 register as appropriate for the @var{len} and @var{type} parameters,
958 and then passes the new values of the debug register and DR7 to the
959 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
960 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
961 required to cover the given region, the above process is repeated for
964 @code{i386_remove_watchpoint} does the opposite: it resets the address
965 in the mirrored value of the debug register and its read/write and
966 length bits in the mirrored value of DR7, then passes these new
967 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
968 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
969 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
970 decrements the reference count, and only calls
971 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
972 the count goes to zero.
974 @findex i386_insert_hw_breakpoint
975 @findex i386_remove_hw_breakpoint
976 @item i386_insert_hw_breakpoint (@var{bp_tgt})
977 @itemx i386_remove_hw_breakpoint (@var{bp_tgt})
978 These functions insert and remove hardware-assisted breakpoints. The
979 macros @code{target_insert_hw_breakpoint} and
980 @code{target_remove_hw_breakpoint} are set to call these functions.
981 The argument is a @code{struct bp_target_info *}, as described in
982 the documentation for @code{target_insert_breakpoint}.
983 These functions work like @code{i386_insert_watchpoint} and
984 @code{i386_remove_watchpoint}, respectively, except that they set up
985 the debug registers to watch instruction execution, and each
986 hardware-assisted breakpoint always requires exactly one debug
989 @findex i386_stopped_by_hwbp
990 @item i386_stopped_by_hwbp (void)
991 This function returns non-zero if the inferior has some watchpoint or
992 hardware breakpoint that triggered. It works like
993 @code{i386_stopped_data_address}, except that it doesn't record the
994 address whose watchpoint triggered.
996 @findex i386_cleanup_dregs
997 @item i386_cleanup_dregs (void)
998 This function clears all the reference counts, addresses, and control
999 bits in the mirror images of the debug registers. It doesn't affect
1000 the actual debug registers in the inferior process.
1007 x86 processors support setting watchpoints on I/O reads or writes.
1008 However, since no target supports this (as of March 2001), and since
1009 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
1010 watchpoints, this feature is not yet available to @value{GDBN} running
1014 x86 processors can enable watchpoints locally, for the current task
1015 only, or globally, for all the tasks. For each debug register,
1016 there's a bit in the DR7 Debug Control register that determines
1017 whether the associated address is watched locally or globally. The
1018 current implementation of x86 watchpoint support in @value{GDBN}
1019 always sets watchpoints to be locally enabled, since global
1020 watchpoints might interfere with the underlying OS and are probably
1021 unavailable in many platforms.
1024 @section Checkpoints
1027 In the abstract, a checkpoint is a point in the execution history of
1028 the program, which the user may wish to return to at some later time.
1030 Internally, a checkpoint is a saved copy of the program state, including
1031 whatever information is required in order to restore the program to that
1032 state at a later time. This can be expected to include the state of
1033 registers and memory, and may include external state such as the state
1034 of open files and devices.
1036 There are a number of ways in which checkpoints may be implemented
1037 in gdb, e.g.@: as corefiles, as forked processes, and as some opaque
1038 method implemented on the target side.
1040 A corefile can be used to save an image of target memory and register
1041 state, which can in principle be restored later --- but corefiles do
1042 not typically include information about external entities such as
1043 open files. Currently this method is not implemented in gdb.
1045 A forked process can save the state of user memory and registers,
1046 as well as some subset of external (kernel) state. This method
1047 is used to implement checkpoints on Linux, and in principle might
1048 be used on other systems.
1050 Some targets, e.g.@: simulators, might have their own built-in
1051 method for saving checkpoints, and gdb might be able to take
1052 advantage of that capability without necessarily knowing any
1053 details of how it is done.
1056 @section Observing changes in @value{GDBN} internals
1057 @cindex observer pattern interface
1058 @cindex notifications about changes in internals
1060 In order to function properly, several modules need to be notified when
1061 some changes occur in the @value{GDBN} internals. Traditionally, these
1062 modules have relied on several paradigms, the most common ones being
1063 hooks and gdb-events. Unfortunately, none of these paradigms was
1064 versatile enough to become the standard notification mechanism in
1065 @value{GDBN}. The fact that they only supported one ``client'' was also
1066 a strong limitation.
1068 A new paradigm, based on the Observer pattern of the @cite{Design
1069 Patterns} book, has therefore been implemented. The goal was to provide
1070 a new interface overcoming the issues with the notification mechanisms
1071 previously available. This new interface needed to be strongly typed,
1072 easy to extend, and versatile enough to be used as the standard
1073 interface when adding new notifications.
1075 See @ref{GDB Observers} for a brief description of the observers
1076 currently implemented in GDB. The rationale for the current
1077 implementation is also briefly discussed.
1079 @node User Interface
1081 @chapter User Interface
1083 @value{GDBN} has several user interfaces, of which the traditional
1084 command-line interface is perhaps the most familiar.
1086 @section Command Interpreter
1088 @cindex command interpreter
1090 The command interpreter in @value{GDBN} is fairly simple. It is designed to
1091 allow for the set of commands to be augmented dynamically, and also
1092 has a recursive subcommand capability, where the first argument to
1093 a command may itself direct a lookup on a different command list.
1095 For instance, the @samp{set} command just starts a lookup on the
1096 @code{setlist} command list, while @samp{set thread} recurses
1097 to the @code{set_thread_cmd_list}.
1101 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
1102 the main command list, and should be used for those commands. The usual
1103 place to add commands is in the @code{_initialize_@var{xyz}} routines at
1104 the ends of most source files.
1106 @findex add_setshow_cmd
1107 @findex add_setshow_cmd_full
1108 To add paired @samp{set} and @samp{show} commands, use
1109 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
1110 a slightly simpler interface which is useful when you don't need to
1111 further modify the new command structures, while the latter returns
1112 the new command structures for manipulation.
1114 @cindex deprecating commands
1115 @findex deprecate_cmd
1116 Before removing commands from the command set it is a good idea to
1117 deprecate them for some time. Use @code{deprecate_cmd} on commands or
1118 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
1119 @code{struct cmd_list_element} as it's first argument. You can use the
1120 return value from @code{add_com} or @code{add_cmd} to deprecate the
1121 command immediately after it is created.
1123 The first time a command is used the user will be warned and offered a
1124 replacement (if one exists). Note that the replacement string passed to
1125 @code{deprecate_cmd} should be the full name of the command, i.e., the
1126 entire string the user should type at the command line.
1128 @anchor{UI-Independent Output}
1129 @section UI-Independent Output---the @code{ui_out} Functions
1130 @c This section is based on the documentation written by Fernando
1131 @c Nasser <fnasser@redhat.com>.
1133 @cindex @code{ui_out} functions
1134 The @code{ui_out} functions present an abstraction level for the
1135 @value{GDBN} output code. They hide the specifics of different user
1136 interfaces supported by @value{GDBN}, and thus free the programmer
1137 from the need to write several versions of the same code, one each for
1138 every UI, to produce output.
1140 @subsection Overview and Terminology
1142 In general, execution of each @value{GDBN} command produces some sort
1143 of output, and can even generate an input request.
1145 Output can be generated for the following purposes:
1149 to display a @emph{result} of an operation;
1152 to convey @emph{info} or produce side-effects of a requested
1156 to provide a @emph{notification} of an asynchronous event (including
1157 progress indication of a prolonged asynchronous operation);
1160 to display @emph{error messages} (including warnings);
1163 to show @emph{debug data};
1166 to @emph{query} or prompt a user for input (a special case).
1170 This section mainly concentrates on how to build result output,
1171 although some of it also applies to other kinds of output.
1173 Generation of output that displays the results of an operation
1174 involves one or more of the following:
1178 output of the actual data
1181 formatting the output as appropriate for console output, to make it
1182 easily readable by humans
1185 machine oriented formatting--a more terse formatting to allow for easy
1186 parsing by programs which read @value{GDBN}'s output
1189 annotation, whose purpose is to help legacy GUIs to identify interesting
1193 The @code{ui_out} routines take care of the first three aspects.
1194 Annotations are provided by separate annotation routines. Note that use
1195 of annotations for an interface between a GUI and @value{GDBN} is
1198 Output can be in the form of a single item, which we call a @dfn{field};
1199 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
1200 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
1201 header and a body. In a BNF-like form:
1204 @item <table> @expansion{}
1205 @code{<header> <body>}
1206 @item <header> @expansion{}
1207 @code{@{ <column> @}}
1208 @item <column> @expansion{}
1209 @code{<width> <alignment> <title>}
1210 @item <body> @expansion{}
1215 @subsection General Conventions
1217 Most @code{ui_out} routines are of type @code{void}, the exceptions are
1218 @code{ui_out_stream_new} (which returns a pointer to the newly created
1219 object) and the @code{make_cleanup} routines.
1221 The first parameter is always the @code{ui_out} vector object, a pointer
1222 to a @code{struct ui_out}.
1224 The @var{format} parameter is like in @code{printf} family of functions.
1225 When it is present, there must also be a variable list of arguments
1226 sufficient used to satisfy the @code{%} specifiers in the supplied
1229 When a character string argument is not used in a @code{ui_out} function
1230 call, a @code{NULL} pointer has to be supplied instead.
1233 @subsection Table, Tuple and List Functions
1235 @cindex list output functions
1236 @cindex table output functions
1237 @cindex tuple output functions
1238 This section introduces @code{ui_out} routines for building lists,
1239 tuples and tables. The routines to output the actual data items
1240 (fields) are presented in the next section.
1242 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
1243 containing information about an object; a @dfn{list} is a sequence of
1244 fields where each field describes an identical object.
1246 Use the @dfn{table} functions when your output consists of a list of
1247 rows (tuples) and the console output should include a heading. Use this
1248 even when you are listing just one object but you still want the header.
1250 @cindex nesting level in @code{ui_out} functions
1251 Tables can not be nested. Tuples and lists can be nested up to a
1252 maximum of five levels.
1254 The overall structure of the table output code is something like this:
1269 Here is the description of table-, tuple- and list-related @code{ui_out}
1272 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
1273 The function @code{ui_out_table_begin} marks the beginning of the output
1274 of a table. It should always be called before any other @code{ui_out}
1275 function for a given table. @var{nbrofcols} is the number of columns in
1276 the table. @var{nr_rows} is the number of rows in the table.
1277 @var{tblid} is an optional string identifying the table. The string
1278 pointed to by @var{tblid} is copied by the implementation of
1279 @code{ui_out_table_begin}, so the application can free the string if it
1280 was @code{malloc}ed.
1282 The companion function @code{ui_out_table_end}, described below, marks
1283 the end of the table's output.
1286 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
1287 @code{ui_out_table_header} provides the header information for a single
1288 table column. You call this function several times, one each for every
1289 column of the table, after @code{ui_out_table_begin}, but before
1290 @code{ui_out_table_body}.
1292 The value of @var{width} gives the column width in characters. The
1293 value of @var{alignment} is one of @code{left}, @code{center}, and
1294 @code{right}, and it specifies how to align the header: left-justify,
1295 center, or right-justify it. @var{colhdr} points to a string that
1296 specifies the column header; the implementation copies that string, so
1297 column header strings in @code{malloc}ed storage can be freed after the
1301 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
1302 This function delimits the table header from the table body.
1305 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
1306 This function signals the end of a table's output. It should be called
1307 after the table body has been produced by the list and field output
1310 There should be exactly one call to @code{ui_out_table_end} for each
1311 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
1312 will signal an internal error.
1315 The output of the tuples that represent the table rows must follow the
1316 call to @code{ui_out_table_body} and precede the call to
1317 @code{ui_out_table_end}. You build a tuple by calling
1318 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
1319 calls to functions which actually output fields between them.
1321 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
1322 This function marks the beginning of a tuple output. @var{id} points
1323 to an optional string that identifies the tuple; it is copied by the
1324 implementation, and so strings in @code{malloc}ed storage can be freed
1328 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
1329 This function signals an end of a tuple output. There should be exactly
1330 one call to @code{ui_out_tuple_end} for each call to
1331 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
1335 @deftypefun {struct cleanup *} make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1336 This function first opens the tuple and then establishes a cleanup
1337 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
1338 and correct implementation of the non-portable@footnote{The function
1339 cast is not portable ISO C.} code sequence:
1341 struct cleanup *old_cleanup;
1342 ui_out_tuple_begin (uiout, "...");
1343 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
1348 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
1349 This function marks the beginning of a list output. @var{id} points to
1350 an optional string that identifies the list; it is copied by the
1351 implementation, and so strings in @code{malloc}ed storage can be freed
1355 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
1356 This function signals an end of a list output. There should be exactly
1357 one call to @code{ui_out_list_end} for each call to
1358 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1362 @deftypefun {struct cleanup *} make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1363 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1364 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
1365 that will close the list.
1368 @subsection Item Output Functions
1370 @cindex item output functions
1371 @cindex field output functions
1373 The functions described below produce output for the actual data
1374 items, or fields, which contain information about the object.
1376 Choose the appropriate function accordingly to your particular needs.
1378 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1379 This is the most general output function. It produces the
1380 representation of the data in the variable-length argument list
1381 according to formatting specifications in @var{format}, a
1382 @code{printf}-like format string. The optional argument @var{fldname}
1383 supplies the name of the field. The data items themselves are
1384 supplied as additional arguments after @var{format}.
1386 This generic function should be used only when it is not possible to
1387 use one of the specialized versions (see below).
1390 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1391 This function outputs a value of an @code{int} variable. It uses the
1392 @code{"%d"} output conversion specification. @var{fldname} specifies
1393 the name of the field.
1396 @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})
1397 This function outputs a value of an @code{int} variable. It differs from
1398 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1399 @var{fldname} specifies
1400 the name of the field.
1403 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, struct gdbarch *@var{gdbarch}, CORE_ADDR @var{address})
1404 This function outputs an address as appropriate for @var{gdbarch}.
1407 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1408 This function outputs a string using the @code{"%s"} conversion
1412 Sometimes, there's a need to compose your output piece by piece using
1413 functions that operate on a stream, such as @code{value_print} or
1414 @code{fprintf_symbol_filtered}. These functions accept an argument of
1415 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1416 used to store the data stream used for the output. When you use one
1417 of these functions, you need a way to pass their results stored in a
1418 @code{ui_file} object to the @code{ui_out} functions. To this end,
1419 you first create a @code{ui_stream} object by calling
1420 @code{ui_out_stream_new}, pass the @code{stream} member of that
1421 @code{ui_stream} object to @code{value_print} and similar functions,
1422 and finally call @code{ui_out_field_stream} to output the field you
1423 constructed. When the @code{ui_stream} object is no longer needed,
1424 you should destroy it and free its memory by calling
1425 @code{ui_out_stream_delete}.
1427 @deftypefun {struct ui_stream *} ui_out_stream_new (struct ui_out *@var{uiout})
1428 This function creates a new @code{ui_stream} object which uses the
1429 same output methods as the @code{ui_out} object whose pointer is
1430 passed in @var{uiout}. It returns a pointer to the newly created
1431 @code{ui_stream} object.
1434 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1435 This functions destroys a @code{ui_stream} object specified by
1439 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1440 This function consumes all the data accumulated in
1441 @code{streambuf->stream} and outputs it like
1442 @code{ui_out_field_string} does. After a call to
1443 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1444 the stream is still valid and may be used for producing more fields.
1447 @strong{Important:} If there is any chance that your code could bail
1448 out before completing output generation and reaching the point where
1449 @code{ui_out_stream_delete} is called, it is necessary to set up a
1450 cleanup, to avoid leaking memory and other resources. Here's a
1451 skeleton code to do that:
1454 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1455 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1460 If the function already has the old cleanup chain set (for other kinds
1461 of cleanups), you just have to add your cleanup to it:
1464 mybuf = ui_out_stream_new (uiout);
1465 make_cleanup (ui_out_stream_delete, mybuf);
1468 Note that with cleanups in place, you should not call
1469 @code{ui_out_stream_delete} directly, or you would attempt to free the
1472 @subsection Utility Output Functions
1474 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1475 This function skips a field in a table. Use it if you have to leave
1476 an empty field without disrupting the table alignment. The argument
1477 @var{fldname} specifies a name for the (missing) filed.
1480 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1481 This function outputs the text in @var{string} in a way that makes it
1482 easy to be read by humans. For example, the console implementation of
1483 this method filters the text through a built-in pager, to prevent it
1484 from scrolling off the visible portion of the screen.
1486 Use this function for printing relatively long chunks of text around
1487 the actual field data: the text it produces is not aligned according
1488 to the table's format. Use @code{ui_out_field_string} to output a
1489 string field, and use @code{ui_out_message}, described below, to
1490 output short messages.
1493 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1494 This function outputs @var{nspaces} spaces. It is handy to align the
1495 text produced by @code{ui_out_text} with the rest of the table or
1499 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1500 This function produces a formatted message, provided that the current
1501 verbosity level is at least as large as given by @var{verbosity}. The
1502 current verbosity level is specified by the user with the @samp{set
1503 verbositylevel} command.@footnote{As of this writing (April 2001),
1504 setting verbosity level is not yet implemented, and is always returned
1505 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1506 argument more than zero will cause the message to never be printed.}
1509 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1510 This function gives the console output filter (a paging filter) a hint
1511 of where to break lines which are too long. Ignored for all other
1512 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1513 be printed to indent the wrapped text on the next line; it must remain
1514 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1515 explicit newline is produced by one of the other functions. If
1516 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1519 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1520 This function flushes whatever output has been accumulated so far, if
1521 the UI buffers output.
1525 @subsection Examples of Use of @code{ui_out} functions
1527 @cindex using @code{ui_out} functions
1528 @cindex @code{ui_out} functions, usage examples
1529 This section gives some practical examples of using the @code{ui_out}
1530 functions to generalize the old console-oriented code in
1531 @value{GDBN}. The examples all come from functions defined on the
1532 @file{breakpoints.c} file.
1534 This example, from the @code{breakpoint_1} function, shows how to
1537 The original code was:
1540 if (!found_a_breakpoint++)
1542 annotate_breakpoints_headers ();
1545 printf_filtered ("Num ");
1547 printf_filtered ("Type ");
1549 printf_filtered ("Disp ");
1551 printf_filtered ("Enb ");
1555 printf_filtered ("Address ");
1558 printf_filtered ("What\n");
1560 annotate_breakpoints_table ();
1564 Here's the new version:
1567 nr_printable_breakpoints = @dots{};
1570 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1572 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1574 if (nr_printable_breakpoints > 0)
1575 annotate_breakpoints_headers ();
1576 if (nr_printable_breakpoints > 0)
1578 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1579 if (nr_printable_breakpoints > 0)
1581 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1582 if (nr_printable_breakpoints > 0)
1584 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1585 if (nr_printable_breakpoints > 0)
1587 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1590 if (nr_printable_breakpoints > 0)
1592 if (print_address_bits <= 32)
1593 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1595 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1597 if (nr_printable_breakpoints > 0)
1599 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1600 ui_out_table_body (uiout);
1601 if (nr_printable_breakpoints > 0)
1602 annotate_breakpoints_table ();
1605 This example, from the @code{print_one_breakpoint} function, shows how
1606 to produce the actual data for the table whose structure was defined
1607 in the above example. The original code was:
1612 printf_filtered ("%-3d ", b->number);
1614 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1615 || ((int) b->type != bptypes[(int) b->type].type))
1616 internal_error ("bptypes table does not describe type #%d.",
1618 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1620 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1622 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1626 This is the new version:
1630 ui_out_tuple_begin (uiout, "bkpt");
1632 ui_out_field_int (uiout, "number", b->number);
1634 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1635 || ((int) b->type != bptypes[(int) b->type].type))
1636 internal_error ("bptypes table does not describe type #%d.",
1638 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1640 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1642 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1646 This example, also from @code{print_one_breakpoint}, shows how to
1647 produce a complicated output field using the @code{print_expression}
1648 functions which requires a stream to be passed. It also shows how to
1649 automate stream destruction with cleanups. The original code was:
1653 print_expression (b->exp, gdb_stdout);
1659 struct ui_stream *stb = ui_out_stream_new (uiout);
1660 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1663 print_expression (b->exp, stb->stream);
1664 ui_out_field_stream (uiout, "what", local_stream);
1667 This example, also from @code{print_one_breakpoint}, shows how to use
1668 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1673 if (b->dll_pathname == NULL)
1674 printf_filtered ("<any library> ");
1676 printf_filtered ("library \"%s\" ", b->dll_pathname);
1683 if (b->dll_pathname == NULL)
1685 ui_out_field_string (uiout, "what", "<any library>");
1686 ui_out_spaces (uiout, 1);
1690 ui_out_text (uiout, "library \"");
1691 ui_out_field_string (uiout, "what", b->dll_pathname);
1692 ui_out_text (uiout, "\" ");
1696 The following example from @code{print_one_breakpoint} shows how to
1697 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1702 if (b->forked_inferior_pid != 0)
1703 printf_filtered ("process %d ", b->forked_inferior_pid);
1710 if (b->forked_inferior_pid != 0)
1712 ui_out_text (uiout, "process ");
1713 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1714 ui_out_spaces (uiout, 1);
1718 Here's an example of using @code{ui_out_field_string}. The original
1723 if (b->exec_pathname != NULL)
1724 printf_filtered ("program \"%s\" ", b->exec_pathname);
1731 if (b->exec_pathname != NULL)
1733 ui_out_text (uiout, "program \"");
1734 ui_out_field_string (uiout, "what", b->exec_pathname);
1735 ui_out_text (uiout, "\" ");
1739 Finally, here's an example of printing an address. The original code:
1743 printf_filtered ("%s ",
1744 hex_string_custom ((unsigned long) b->address, 8));
1751 ui_out_field_core_addr (uiout, "Address", b->address);
1755 @section Console Printing
1764 @cindex @code{libgdb}
1765 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1766 to provide an API to @value{GDBN}'s functionality.
1769 @cindex @code{libgdb}
1770 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1771 better able to support graphical and other environments.
1773 Since @code{libgdb} development is on-going, its architecture is still
1774 evolving. The following components have so far been identified:
1778 Observer - @file{gdb-events.h}.
1780 Builder - @file{ui-out.h}
1782 Event Loop - @file{event-loop.h}
1784 Library - @file{gdb.h}
1787 The model that ties these components together is described below.
1789 @section The @code{libgdb} Model
1791 A client of @code{libgdb} interacts with the library in two ways.
1795 As an observer (using @file{gdb-events}) receiving notifications from
1796 @code{libgdb} of any internal state changes (break point changes, run
1799 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1800 obtain various status values from @value{GDBN}.
1803 Since @code{libgdb} could have multiple clients (e.g., a GUI supporting
1804 the existing @value{GDBN} CLI), those clients must co-operate when
1805 controlling @code{libgdb}. In particular, a client must ensure that
1806 @code{libgdb} is idle (i.e.@: no other client is using @code{libgdb})
1807 before responding to a @file{gdb-event} by making a query.
1809 @section CLI support
1811 At present @value{GDBN}'s CLI is very much entangled in with the core of
1812 @code{libgdb}. Consequently, a client wishing to include the CLI in
1813 their interface needs to carefully co-ordinate its own and the CLI's
1816 It is suggested that the client set @code{libgdb} up to be bi-modal
1817 (alternate between CLI and client query modes). The notes below sketch
1822 The client registers itself as an observer of @code{libgdb}.
1824 The client create and install @code{cli-out} builder using its own
1825 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1826 @code{gdb_stdout} streams.
1828 The client creates a separate custom @code{ui-out} builder that is only
1829 used while making direct queries to @code{libgdb}.
1832 When the client receives input intended for the CLI, it simply passes it
1833 along. Since the @code{cli-out} builder is installed by default, all
1834 the CLI output in response to that command is routed (pronounced rooted)
1835 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1836 At the same time, the client is kept abreast of internal changes by
1837 virtue of being a @code{libgdb} observer.
1839 The only restriction on the client is that it must wait until
1840 @code{libgdb} becomes idle before initiating any queries (using the
1841 client's custom builder).
1843 @section @code{libgdb} components
1845 @subheading Observer - @file{gdb-events.h}
1846 @file{gdb-events} provides the client with a very raw mechanism that can
1847 be used to implement an observer. At present it only allows for one
1848 observer and that observer must, internally, handle the need to delay
1849 the processing of any event notifications until after @code{libgdb} has
1850 finished the current command.
1852 @subheading Builder - @file{ui-out.h}
1853 @file{ui-out} provides the infrastructure necessary for a client to
1854 create a builder. That builder is then passed down to @code{libgdb}
1855 when doing any queries.
1857 @subheading Event Loop - @file{event-loop.h}
1858 @c There could be an entire section on the event-loop
1859 @file{event-loop}, currently non-re-entrant, provides a simple event
1860 loop. A client would need to either plug its self into this loop or,
1861 implement a new event-loop that @value{GDBN} would use.
1863 The event-loop will eventually be made re-entrant. This is so that
1864 @value{GDBN} can better handle the problem of some commands blocking
1865 instead of returning.
1867 @subheading Library - @file{gdb.h}
1868 @file{libgdb} is the most obvious component of this system. It provides
1869 the query interface. Each function is parameterized by a @code{ui-out}
1870 builder. The result of the query is constructed using that builder
1871 before the query function returns.
1878 @cindex @code{value} structure
1879 @value{GDBN} uses @code{struct value}, or @dfn{values}, as an internal
1880 abstraction for the representation of a variety of inferior objects
1881 and @value{GDBN} convenience objects.
1883 Values have an associated @code{struct type}, that describes a virtual
1884 view of the raw data or object stored in or accessed through the
1887 A value is in addition discriminated by its lvalue-ness, given its
1888 @code{enum lval_type} enumeration type:
1890 @cindex @code{lval_type} enumeration, for values.
1892 @item @code{not_lval}
1893 This value is not an lval. It can't be assigned to.
1895 @item @code{lval_memory}
1896 This value represents an object in memory.
1898 @item @code{lval_register}
1899 This value represents an object that lives in a register.
1901 @item @code{lval_internalvar}
1902 Represents the value of an internal variable.
1904 @item @code{lval_internalvar_component}
1905 Represents part of a @value{GDBN} internal variable. E.g., a
1908 @cindex computed values
1909 @item @code{lval_computed}
1910 These are ``computed'' values. They allow creating specialized value
1911 objects for specific purposes, all abstracted away from the core value
1912 support code. The creator of such a value writes specialized
1913 functions to handle the reading and writing to/from the value's
1914 backend data, and optionally, a ``copy operator'' and a
1917 Pointers to these functions are stored in a @code{struct lval_funcs}
1918 instance (declared in @file{value.h}), and passed to the
1919 @code{allocate_computed_value} function, as in the example below.
1923 nil_value_read (struct value *v)
1925 /* This callback reads data from some backend, and stores it in V.
1926 In this case, we always read null data. You'll want to fill in
1927 something more interesting. */
1929 memset (value_contents_all_raw (v),
1931 TYPE_LENGTH (value_type (v)));
1935 nil_value_write (struct value *v, struct value *fromval)
1937 /* Takes the data from FROMVAL and stores it in the backend of V. */
1939 to_oblivion (value_contents_all_raw (fromval),
1941 TYPE_LENGTH (value_type (fromval)));
1944 static struct lval_funcs nil_value_funcs =
1951 make_nil_value (void)
1956 type = make_nils_type ();
1957 v = allocate_computed_value (type, &nil_value_funcs, NULL);
1963 See the implementation of the @code{$_siginfo} convenience variable in
1964 @file{infrun.c} as a real example use of lval_computed.
1969 @chapter Stack Frames
1972 @cindex call stack frame
1973 A frame is a construct that @value{GDBN} uses to keep track of calling
1974 and called functions.
1976 @cindex unwind frame
1977 @value{GDBN}'s frame model, a fresh design, was implemented with the
1978 need to support @sc{dwarf}'s Call Frame Information in mind. In fact,
1979 the term ``unwind'' is taken directly from that specification.
1980 Developers wishing to learn more about unwinders, are encouraged to
1981 read the @sc{dwarf} specification, available from
1982 @url{http://www.dwarfstd.org}.
1984 @findex frame_register_unwind
1985 @findex get_frame_register
1986 @value{GDBN}'s model is that you find a frame's registers by
1987 ``unwinding'' them from the next younger frame. That is,
1988 @samp{get_frame_register} which returns the value of a register in
1989 frame #1 (the next-to-youngest frame), is implemented by calling frame
1990 #0's @code{frame_register_unwind} (the youngest frame). But then the
1991 obvious question is: how do you access the registers of the youngest
1994 @cindex sentinel frame
1995 @findex get_frame_type
1996 @vindex SENTINEL_FRAME
1997 To answer this question, @value{GDBN} has the @dfn{sentinel} frame, the
1998 ``-1st'' frame. Unwinding registers from the sentinel frame gives you
1999 the current values of the youngest real frame's registers. If @var{f}
2000 is a sentinel frame, then @code{get_frame_type (@var{f}) @equiv{}
2003 @section Selecting an Unwinder
2005 @findex frame_unwind_prepend_unwinder
2006 @findex frame_unwind_append_unwinder
2007 The architecture registers a list of frame unwinders (@code{struct
2008 frame_unwind}), using the functions
2009 @code{frame_unwind_prepend_unwinder} and
2010 @code{frame_unwind_append_unwinder}. Each unwinder includes a
2011 sniffer. Whenever @value{GDBN} needs to unwind a frame (to fetch the
2012 previous frame's registers or the current frame's ID), it calls
2013 registered sniffers in order to find one which recognizes the frame.
2014 The first time a sniffer returns non-zero, the corresponding unwinder
2015 is assigned to the frame.
2017 @section Unwinding the Frame ID
2020 Every frame has an associated ID, of type @code{struct frame_id}.
2021 The ID includes the stack base and function start address for
2022 the frame. The ID persists through the entire life of the frame,
2023 including while other called frames are running; it is used to
2024 locate an appropriate @code{struct frame_info} from the cache.
2026 Every time the inferior stops, and at various other times, the frame
2027 cache is flushed. Because of this, parts of @value{GDBN} which need
2028 to keep track of individual frames cannot use pointers to @code{struct
2029 frame_info}. A frame ID provides a stable reference to a frame, even
2030 when the unwinder must be run again to generate a new @code{struct
2031 frame_info} for the same frame.
2033 The frame's unwinder's @code{this_id} method is called to find the ID.
2034 Note that this is different from register unwinding, where the next
2035 frame's @code{prev_register} is called to unwind this frame's
2038 Both stack base and function address are required to identify the
2039 frame, because a recursive function has the same function address for
2040 two consecutive frames and a leaf function may have the same stack
2041 address as its caller. On some platforms, a third address is part of
2042 the ID to further disambiguate frames---for instance, on IA-64
2043 the separate register stack address is included in the ID.
2045 An invalid frame ID (@code{null_frame_id}) returned from the
2046 @code{this_id} method means to stop unwinding after this frame.
2048 @section Unwinding Registers
2050 Each unwinder includes a @code{prev_register} method. This method
2051 takes a frame, an associated cache pointer, and a register number.
2052 It returns a @code{struct value *} describing the requested register,
2053 as saved by this frame. This is the value of the register that is
2054 current in this frame's caller.
2056 The returned value must have the same type as the register. It may
2057 have any lvalue type. In most circumstances one of these routines
2058 will generate the appropriate value:
2061 @item frame_unwind_got_optimized
2062 @findex frame_unwind_got_optimized
2063 This register was not saved.
2065 @item frame_unwind_got_register
2066 @findex frame_unwind_got_register
2067 This register was copied into another register in this frame. This
2068 is also used for unchanged registers; they are ``copied'' into the
2071 @item frame_unwind_got_memory
2072 @findex frame_unwind_got_memory
2073 This register was saved in memory.
2075 @item frame_unwind_got_constant
2076 @findex frame_unwind_got_constant
2077 This register was not saved, but the unwinder can compute the previous
2078 value some other way.
2080 @item frame_unwind_got_address
2081 @findex frame_unwind_got_address
2082 Same as @code{frame_unwind_got_constant}, except that the value is a target
2083 address. This is frequently used for the stack pointer, which is not
2084 explicitly saved but has a known offset from this frame's stack
2085 pointer. For architectures with a flat unified address space, this is
2086 generally the same as @code{frame_unwind_got_constant}.
2089 @node Symbol Handling
2091 @chapter Symbol Handling
2093 Symbols are a key part of @value{GDBN}'s operation. Symbols include
2094 variables, functions, and types.
2096 Symbol information for a large program can be truly massive, and
2097 reading of symbol information is one of the major performance
2098 bottlenecks in @value{GDBN}; it can take many minutes to process it
2099 all. Studies have shown that nearly all the time spent is
2100 computational, rather than file reading.
2102 One of the ways for @value{GDBN} to provide a good user experience is
2103 to start up quickly, taking no more than a few seconds. It is simply
2104 not possible to process all of a program's debugging info in that
2105 time, and so we attempt to handle symbols incrementally. For instance,
2106 we create @dfn{partial symbol tables} consisting of only selected
2107 symbols, and only expand them to full symbol tables when necessary.
2109 @section Symbol Reading
2111 @cindex symbol reading
2112 @cindex reading of symbols
2113 @cindex symbol files
2114 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
2115 file is the file containing the program which @value{GDBN} is
2116 debugging. @value{GDBN} can be directed to use a different file for
2117 symbols (with the @samp{symbol-file} command), and it can also read
2118 more symbols via the @samp{add-file} and @samp{load} commands. In
2119 addition, it may bring in more symbols while loading shared
2122 @findex find_sym_fns
2123 Symbol files are initially opened by code in @file{symfile.c} using
2124 the BFD library (@pxref{Support Libraries}). BFD identifies the type
2125 of the file by examining its header. @code{find_sym_fns} then uses
2126 this identification to locate a set of symbol-reading functions.
2128 @findex add_symtab_fns
2129 @cindex @code{sym_fns} structure
2130 @cindex adding a symbol-reading module
2131 Symbol-reading modules identify themselves to @value{GDBN} by calling
2132 @code{add_symtab_fns} during their module initialization. The argument
2133 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
2134 name (or name prefix) of the symbol format, the length of the prefix,
2135 and pointers to four functions. These functions are called at various
2136 times to process symbol files whose identification matches the specified
2139 The functions supplied by each module are:
2142 @item @var{xyz}_symfile_init(struct sym_fns *sf)
2144 @cindex secondary symbol file
2145 Called from @code{symbol_file_add} when we are about to read a new
2146 symbol file. This function should clean up any internal state (possibly
2147 resulting from half-read previous files, for example) and prepare to
2148 read a new symbol file. Note that the symbol file which we are reading
2149 might be a new ``main'' symbol file, or might be a secondary symbol file
2150 whose symbols are being added to the existing symbol table.
2152 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
2153 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
2154 new symbol file being read. Its @code{private} field has been zeroed,
2155 and can be modified as desired. Typically, a struct of private
2156 information will be @code{malloc}'d, and a pointer to it will be placed
2157 in the @code{private} field.
2159 There is no result from @code{@var{xyz}_symfile_init}, but it can call
2160 @code{error} if it detects an unavoidable problem.
2162 @item @var{xyz}_new_init()
2164 Called from @code{symbol_file_add} when discarding existing symbols.
2165 This function needs only handle the symbol-reading module's internal
2166 state; the symbol table data structures visible to the rest of
2167 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
2168 arguments and no result. It may be called after
2169 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
2170 may be called alone if all symbols are simply being discarded.
2172 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
2174 Called from @code{symbol_file_add} to actually read the symbols from a
2175 symbol-file into a set of psymtabs or symtabs.
2177 @code{sf} points to the @code{struct sym_fns} originally passed to
2178 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
2179 the offset between the file's specified start address and its true
2180 address in memory. @code{mainline} is 1 if this is the main symbol
2181 table being read, and 0 if a secondary symbol file (e.g., shared library
2182 or dynamically loaded file) is being read.@refill
2185 In addition, if a symbol-reading module creates psymtabs when
2186 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
2187 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
2188 from any point in the @value{GDBN} symbol-handling code.
2191 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
2193 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
2194 the psymtab has not already been read in and had its @code{pst->symtab}
2195 pointer set. The argument is the psymtab to be fleshed-out into a
2196 symtab. Upon return, @code{pst->readin} should have been set to 1, and
2197 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
2198 zero if there were no symbols in that part of the symbol file.
2201 @section Partial Symbol Tables
2203 @value{GDBN} has three types of symbol tables:
2206 @cindex full symbol table
2209 Full symbol tables (@dfn{symtabs}). These contain the main
2210 information about symbols and addresses.
2214 Partial symbol tables (@dfn{psymtabs}). These contain enough
2215 information to know when to read the corresponding part of the full
2218 @cindex minimal symbol table
2221 Minimal symbol tables (@dfn{msymtabs}). These contain information
2222 gleaned from non-debugging symbols.
2225 @cindex partial symbol table
2226 This section describes partial symbol tables.
2228 A psymtab is constructed by doing a very quick pass over an executable
2229 file's debugging information. Small amounts of information are
2230 extracted---enough to identify which parts of the symbol table will
2231 need to be re-read and fully digested later, when the user needs the
2232 information. The speed of this pass causes @value{GDBN} to start up very
2233 quickly. Later, as the detailed rereading occurs, it occurs in small
2234 pieces, at various times, and the delay therefrom is mostly invisible to
2236 @c (@xref{Symbol Reading}.)
2238 The symbols that show up in a file's psymtab should be, roughly, those
2239 visible to the debugger's user when the program is not running code from
2240 that file. These include external symbols and types, static symbols and
2241 types, and @code{enum} values declared at file scope.
2243 The psymtab also contains the range of instruction addresses that the
2244 full symbol table would represent.
2246 @cindex finding a symbol
2247 @cindex symbol lookup
2248 The idea is that there are only two ways for the user (or much of the
2249 code in the debugger) to reference a symbol:
2252 @findex find_pc_function
2253 @findex find_pc_line
2255 By its address (e.g., execution stops at some address which is inside a
2256 function in this file). The address will be noticed to be in the
2257 range of this psymtab, and the full symtab will be read in.
2258 @code{find_pc_function}, @code{find_pc_line}, and other
2259 @code{find_pc_@dots{}} functions handle this.
2261 @cindex lookup_symbol
2264 (e.g., the user asks to print a variable, or set a breakpoint on a
2265 function). Global names and file-scope names will be found in the
2266 psymtab, which will cause the symtab to be pulled in. Local names will
2267 have to be qualified by a global name, or a file-scope name, in which
2268 case we will have already read in the symtab as we evaluated the
2269 qualifier. Or, a local symbol can be referenced when we are ``in'' a
2270 local scope, in which case the first case applies. @code{lookup_symbol}
2271 does most of the work here.
2274 The only reason that psymtabs exist is to cause a symtab to be read in
2275 at the right moment. Any symbol that can be elided from a psymtab,
2276 while still causing that to happen, should not appear in it. Since
2277 psymtabs don't have the idea of scope, you can't put local symbols in
2278 them anyway. Psymtabs don't have the idea of the type of a symbol,
2279 either, so types need not appear, unless they will be referenced by
2282 It is a bug for @value{GDBN} to behave one way when only a psymtab has
2283 been read, and another way if the corresponding symtab has been read
2284 in. Such bugs are typically caused by a psymtab that does not contain
2285 all the visible symbols, or which has the wrong instruction address
2288 The psymtab for a particular section of a symbol file (objfile) could be
2289 thrown away after the symtab has been read in. The symtab should always
2290 be searched before the psymtab, so the psymtab will never be used (in a
2291 bug-free environment). Currently, psymtabs are allocated on an obstack,
2292 and all the psymbols themselves are allocated in a pair of large arrays
2293 on an obstack, so there is little to be gained by trying to free them
2294 unless you want to do a lot more work.
2298 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
2300 @cindex fundamental types
2301 These are the fundamental types that @value{GDBN} uses internally. Fundamental
2302 types from the various debugging formats (stabs, ELF, etc) are mapped
2303 into one of these. They are basically a union of all fundamental types
2304 that @value{GDBN} knows about for all the languages that @value{GDBN}
2307 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
2310 Each time @value{GDBN} builds an internal type, it marks it with one
2311 of these types. The type may be a fundamental type, such as
2312 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
2313 which is a pointer to another type. Typically, several @code{FT_*}
2314 types map to one @code{TYPE_CODE_*} type, and are distinguished by
2315 other members of the type struct, such as whether the type is signed
2316 or unsigned, and how many bits it uses.
2318 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
2320 These are instances of type structs that roughly correspond to
2321 fundamental types and are created as global types for @value{GDBN} to
2322 use for various ugly historical reasons. We eventually want to
2323 eliminate these. Note for example that @code{builtin_type_int}
2324 initialized in @file{gdbtypes.c} is basically the same as a
2325 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
2326 an @code{FT_INTEGER} fundamental type. The difference is that the
2327 @code{builtin_type} is not associated with any particular objfile, and
2328 only one instance exists, while @file{c-lang.c} builds as many
2329 @code{TYPE_CODE_INT} types as needed, with each one associated with
2330 some particular objfile.
2332 @section Object File Formats
2333 @cindex object file formats
2337 @cindex @code{a.out} format
2338 The @code{a.out} format is the original file format for Unix. It
2339 consists of three sections: @code{text}, @code{data}, and @code{bss},
2340 which are for program code, initialized data, and uninitialized data,
2343 The @code{a.out} format is so simple that it doesn't have any reserved
2344 place for debugging information. (Hey, the original Unix hackers used
2345 @samp{adb}, which is a machine-language debugger!) The only debugging
2346 format for @code{a.out} is stabs, which is encoded as a set of normal
2347 symbols with distinctive attributes.
2349 The basic @code{a.out} reader is in @file{dbxread.c}.
2354 The COFF format was introduced with System V Release 3 (SVR3) Unix.
2355 COFF files may have multiple sections, each prefixed by a header. The
2356 number of sections is limited.
2358 The COFF specification includes support for debugging. Although this
2359 was a step forward, the debugging information was woefully limited.
2360 For instance, it was not possible to represent code that came from an
2361 included file. GNU's COFF-using configs often use stabs-type info,
2362 encapsulated in special sections.
2364 The COFF reader is in @file{coffread.c}.
2368 @cindex ECOFF format
2369 ECOFF is an extended COFF originally introduced for Mips and Alpha
2372 The basic ECOFF reader is in @file{mipsread.c}.
2376 @cindex XCOFF format
2377 The IBM RS/6000 running AIX uses an object file format called XCOFF.
2378 The COFF sections, symbols, and line numbers are used, but debugging
2379 symbols are @code{dbx}-style stabs whose strings are located in the
2380 @code{.debug} section (rather than the string table). For more
2381 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
2383 The shared library scheme has a clean interface for figuring out what
2384 shared libraries are in use, but the catch is that everything which
2385 refers to addresses (symbol tables and breakpoints at least) needs to be
2386 relocated for both shared libraries and the main executable. At least
2387 using the standard mechanism this can only be done once the program has
2388 been run (or the core file has been read).
2392 @cindex PE-COFF format
2393 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
2394 executables. PE is basically COFF with additional headers.
2396 While BFD includes special PE support, @value{GDBN} needs only the basic
2402 The ELF format came with System V Release 4 (SVR4) Unix. ELF is
2403 similar to COFF in being organized into a number of sections, but it
2404 removes many of COFF's limitations. Debugging info may be either stabs
2405 encapsulated in ELF sections, or more commonly these days, DWARF.
2407 The basic ELF reader is in @file{elfread.c}.
2412 SOM is HP's object file and debug format (not to be confused with IBM's
2413 SOM, which is a cross-language ABI).
2415 The SOM reader is in @file{somread.c}.
2417 @section Debugging File Formats
2419 This section describes characteristics of debugging information that
2420 are independent of the object file format.
2424 @cindex stabs debugging info
2425 @code{stabs} started out as special symbols within the @code{a.out}
2426 format. Since then, it has been encapsulated into other file
2427 formats, such as COFF and ELF.
2429 While @file{dbxread.c} does some of the basic stab processing,
2430 including for encapsulated versions, @file{stabsread.c} does
2435 @cindex COFF debugging info
2436 The basic COFF definition includes debugging information. The level
2437 of support is minimal and non-extensible, and is not often used.
2439 @subsection Mips debug (Third Eye)
2441 @cindex ECOFF debugging info
2442 ECOFF includes a definition of a special debug format.
2444 The file @file{mdebugread.c} implements reading for this format.
2446 @c mention DWARF 1 as a formerly-supported format
2450 @cindex DWARF 2 debugging info
2451 DWARF 2 is an improved but incompatible version of DWARF 1.
2453 The DWARF 2 reader is in @file{dwarf2read.c}.
2455 @subsection Compressed DWARF 2
2457 @cindex Compressed DWARF 2 debugging info
2458 Compressed DWARF 2 is not technically a separate debugging format, but
2459 merely DWARF 2 debug information that has been compressed. In this
2460 format, every object-file section holding DWARF 2 debugging
2461 information is compressed and prepended with a header. (The section
2462 is also typically renamed, so a section called @code{.debug_info} in a
2463 DWARF 2 binary would be called @code{.zdebug_info} in a compressed
2464 DWARF 2 binary.) The header is 12 bytes long:
2468 4 bytes: the literal string ``ZLIB''
2470 8 bytes: the uncompressed size of the section, in big-endian byte
2474 The same reader is used for both compressed an normal DWARF 2 info.
2475 Section decompression is done in @code{zlib_decompress_section} in
2476 @file{dwarf2read.c}.
2480 @cindex DWARF 3 debugging info
2481 DWARF 3 is an improved version of DWARF 2.
2485 @cindex SOM debugging info
2486 Like COFF, the SOM definition includes debugging information.
2488 @section Adding a New Symbol Reader to @value{GDBN}
2490 @cindex adding debugging info reader
2491 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
2492 there is probably little to be done.
2494 If you need to add a new object file format, you must first add it to
2495 BFD. This is beyond the scope of this document.
2497 You must then arrange for the BFD code to provide access to the
2498 debugging symbols. Generally @value{GDBN} will have to call swapping
2499 routines from BFD and a few other BFD internal routines to locate the
2500 debugging information. As much as possible, @value{GDBN} should not
2501 depend on the BFD internal data structures.
2503 For some targets (e.g., COFF), there is a special transfer vector used
2504 to call swapping routines, since the external data structures on various
2505 platforms have different sizes and layouts. Specialized routines that
2506 will only ever be implemented by one object file format may be called
2507 directly. This interface should be described in a file
2508 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
2510 @section Memory Management for Symbol Files
2512 Most memory associated with a loaded symbol file is stored on
2513 its @code{objfile_obstack}. This includes symbols, types,
2514 namespace data, and other information produced by the symbol readers.
2516 Because this data lives on the objfile's obstack, it is automatically
2517 released when the objfile is unloaded or reloaded. Therefore one
2518 objfile must not reference symbol or type data from another objfile;
2519 they could be unloaded at different times.
2521 User convenience variables, et cetera, have associated types. Normally
2522 these types live in the associated objfile. However, when the objfile
2523 is unloaded, those types are deep copied to global memory, so that
2524 the values of the user variables and history items are not lost.
2527 @node Language Support
2529 @chapter Language Support
2531 @cindex language support
2532 @value{GDBN}'s language support is mainly driven by the symbol reader,
2533 although it is possible for the user to set the source language
2536 @value{GDBN} chooses the source language by looking at the extension
2537 of the file recorded in the debug info; @file{.c} means C, @file{.f}
2538 means Fortran, etc. It may also use a special-purpose language
2539 identifier if the debug format supports it, like with DWARF.
2541 @section Adding a Source Language to @value{GDBN}
2543 @cindex adding source language
2544 To add other languages to @value{GDBN}'s expression parser, follow the
2548 @item Create the expression parser.
2550 @cindex expression parser
2551 This should reside in a file @file{@var{lang}-exp.y}. Routines for
2552 building parsed expressions into a @code{union exp_element} list are in
2555 @cindex language parser
2556 Since we can't depend upon everyone having Bison, and YACC produces
2557 parsers that define a bunch of global names, the following lines
2558 @strong{must} be included at the top of the YACC parser, to prevent the
2559 various parsers from defining the same global names:
2562 #define yyparse @var{lang}_parse
2563 #define yylex @var{lang}_lex
2564 #define yyerror @var{lang}_error
2565 #define yylval @var{lang}_lval
2566 #define yychar @var{lang}_char
2567 #define yydebug @var{lang}_debug
2568 #define yypact @var{lang}_pact
2569 #define yyr1 @var{lang}_r1
2570 #define yyr2 @var{lang}_r2
2571 #define yydef @var{lang}_def
2572 #define yychk @var{lang}_chk
2573 #define yypgo @var{lang}_pgo
2574 #define yyact @var{lang}_act
2575 #define yyexca @var{lang}_exca
2576 #define yyerrflag @var{lang}_errflag
2577 #define yynerrs @var{lang}_nerrs
2580 At the bottom of your parser, define a @code{struct language_defn} and
2581 initialize it with the right values for your language. Define an
2582 @code{initialize_@var{lang}} routine and have it call
2583 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
2584 that your language exists. You'll need some other supporting variables
2585 and functions, which will be used via pointers from your
2586 @code{@var{lang}_language_defn}. See the declaration of @code{struct
2587 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
2588 for more information.
2590 @item Add any evaluation routines, if necessary
2592 @cindex expression evaluation routines
2593 @findex evaluate_subexp
2594 @findex prefixify_subexp
2595 @findex length_of_subexp
2596 If you need new opcodes (that represent the operations of the language),
2597 add them to the enumerated type in @file{expression.h}. Add support
2598 code for these operations in the @code{evaluate_subexp} function
2599 defined in the file @file{eval.c}. Add cases
2600 for new opcodes in two functions from @file{parse.c}:
2601 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
2602 the number of @code{exp_element}s that a given operation takes up.
2604 @item Update some existing code
2606 Add an enumerated identifier for your language to the enumerated type
2607 @code{enum language} in @file{defs.h}.
2609 Update the routines in @file{language.c} so your language is included.
2610 These routines include type predicates and such, which (in some cases)
2611 are language dependent. If your language does not appear in the switch
2612 statement, an error is reported.
2614 @vindex current_language
2615 Also included in @file{language.c} is the code that updates the variable
2616 @code{current_language}, and the routines that translate the
2617 @code{language_@var{lang}} enumerated identifier into a printable
2620 @findex _initialize_language
2621 Update the function @code{_initialize_language} to include your
2622 language. This function picks the default language upon startup, so is
2623 dependent upon which languages that @value{GDBN} is built for.
2625 @findex allocate_symtab
2626 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2627 code so that the language of each symtab (source file) is set properly.
2628 This is used to determine the language to use at each stack frame level.
2629 Currently, the language is set based upon the extension of the source
2630 file. If the language can be better inferred from the symbol
2631 information, please set the language of the symtab in the symbol-reading
2634 @findex print_subexp
2635 @findex op_print_tab
2636 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2637 expression opcodes you have added to @file{expression.h}. Also, add the
2638 printed representations of your operators to @code{op_print_tab}.
2640 @item Add a place of call
2643 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2644 @code{parse_exp_1} (defined in @file{parse.c}).
2646 @item Edit @file{Makefile.in}
2648 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2649 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2650 not get linked in, or, worse yet, it may not get @code{tar}red into the
2655 @node Host Definition
2657 @chapter Host Definition
2659 With the advent of Autoconf, it's rarely necessary to have host
2660 definition machinery anymore. The following information is provided,
2661 mainly, as an historical reference.
2663 @section Adding a New Host
2665 @cindex adding a new host
2666 @cindex host, adding
2667 @value{GDBN}'s host configuration support normally happens via Autoconf.
2668 New host-specific definitions should not be needed. Older hosts
2669 @value{GDBN} still use the host-specific definitions and files listed
2670 below, but these mostly exist for historical reasons, and will
2671 eventually disappear.
2674 @item gdb/config/@var{arch}/@var{xyz}.mh
2675 This file is a Makefile fragment that once contained both host and
2676 native configuration information (@pxref{Native Debugging}) for the
2677 machine @var{xyz}. The host configuration information is now handled
2680 Host configuration information included definitions for @code{CC},
2681 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2682 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2684 New host-only configurations do not need this file.
2688 (Files named @file{gdb/config/@var{arch}/xm-@var{xyz}.h} were once
2689 used to define host-specific macros, but were no longer needed and
2690 have all been removed.)
2692 @subheading Generic Host Support Files
2694 @cindex generic host support
2695 There are some ``generic'' versions of routines that can be used by
2699 @cindex remote debugging support
2700 @cindex serial line support
2702 This contains serial line support for Unix systems. It is included by
2703 default on all Unix-like hosts.
2706 This contains serial pipe support for Unix systems. It is included by
2707 default on all Unix-like hosts.
2710 This contains serial line support for 32-bit programs running under
2711 Windows using MinGW.
2714 This contains serial line support for 32-bit programs running under DOS,
2715 using the DJGPP (a.k.a.@: GO32) execution environment.
2717 @cindex TCP remote support
2719 This contains generic TCP support using sockets. It is included by
2720 default on all Unix-like hosts and with MinGW.
2723 @section Host Conditionals
2725 When @value{GDBN} is configured and compiled, various macros are
2726 defined or left undefined, to control compilation based on the
2727 attributes of the host system. While formerly they could be set in
2728 host-specific header files, at present they can be changed only by
2729 setting @code{CFLAGS} when building, or by editing the source code.
2731 These macros and their meanings (or if the meaning is not documented
2732 here, then one of the source files where they are used is indicated)
2736 @item @value{GDBN}INIT_FILENAME
2737 The default name of @value{GDBN}'s initialization file (normally
2740 @item SIGWINCH_HANDLER
2741 If your host defines @code{SIGWINCH}, you can define this to be the name
2742 of a function to be called if @code{SIGWINCH} is received.
2744 @item SIGWINCH_HANDLER_BODY
2745 Define this to expand into code that will define the function named by
2746 the expansion of @code{SIGWINCH_HANDLER}.
2748 @item CRLF_SOURCE_FILES
2749 @cindex DOS text files
2750 Define this if host files use @code{\r\n} rather than @code{\n} as a
2751 line terminator. This will cause source file listings to omit @code{\r}
2752 characters when printing and it will allow @code{\r\n} line endings of files
2753 which are ``sourced'' by gdb. It must be possible to open files in binary
2754 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2756 @item DEFAULT_PROMPT
2758 The default value of the prompt string (normally @code{"(gdb) "}).
2761 @cindex terminal device
2762 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2765 Substitute for isatty, if not available.
2768 Define this if binary files are opened the same way as text files.
2770 @item CC_HAS_LONG_LONG
2771 @cindex @code{long long} data type
2772 Define this if the host C compiler supports @code{long long}. This is set
2773 by the @code{configure} script.
2775 @item PRINTF_HAS_LONG_LONG
2776 Define this if the host can handle printing of long long integers via
2777 the printf format conversion specifier @code{ll}. This is set by the
2778 @code{configure} script.
2780 @item LSEEK_NOT_LINEAR
2781 Define this if @code{lseek (n)} does not necessarily move to byte number
2782 @code{n} in the file. This is only used when reading source files. It
2783 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2786 If defined, this should be one or more tokens, such as @code{volatile},
2787 that can be used in both the declaration and definition of functions to
2788 indicate that they never return. The default is already set correctly
2789 if compiling with GCC. This will almost never need to be defined.
2792 If defined, this should be one or more tokens, such as
2793 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2794 of functions to indicate that they never return. The default is already
2795 set correctly if compiling with GCC. This will almost never need to be
2799 Define this to help placate @code{lint} in some situations.
2802 Define this to override the defaults of @code{__volatile__} or
2807 @node Target Architecture Definition
2809 @chapter Target Architecture Definition
2811 @cindex target architecture definition
2812 @value{GDBN}'s target architecture defines what sort of
2813 machine-language programs @value{GDBN} can work with, and how it works
2816 The target architecture object is implemented as the C structure
2817 @code{struct gdbarch *}. The structure, and its methods, are generated
2818 using the Bourne shell script @file{gdbarch.sh}.
2821 * OS ABI Variant Handling::
2822 * Initialize New Architecture::
2823 * Registers and Memory::
2824 * Pointers and Addresses::
2826 * Register Representation::
2827 * Frame Interpretation::
2828 * Inferior Call Setup::
2829 * Adding support for debugging core files::
2830 * Defining Other Architecture Features::
2831 * Adding a New Target::
2834 @node OS ABI Variant Handling
2835 @section Operating System ABI Variant Handling
2836 @cindex OS ABI variants
2838 @value{GDBN} provides a mechanism for handling variations in OS
2839 ABIs. An OS ABI variant may have influence over any number of
2840 variables in the target architecture definition. There are two major
2841 components in the OS ABI mechanism: sniffers and handlers.
2843 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2844 (the architecture may be wildcarded) in an attempt to determine the
2845 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2846 to be @dfn{generic}, while sniffers for a specific architecture are
2847 considered to be @dfn{specific}. A match from a specific sniffer
2848 overrides a match from a generic sniffer. Multiple sniffers for an
2849 architecture/flavour may exist, in order to differentiate between two
2850 different operating systems which use the same basic file format. The
2851 OS ABI framework provides a generic sniffer for ELF-format files which
2852 examines the @code{EI_OSABI} field of the ELF header, as well as note
2853 sections known to be used by several operating systems.
2855 @cindex fine-tuning @code{gdbarch} structure
2856 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2857 selected OS ABI. There may be only one handler for a given OS ABI
2858 for each BFD architecture.
2860 The following OS ABI variants are defined in @file{defs.h}:
2864 @findex GDB_OSABI_UNINITIALIZED
2865 @item GDB_OSABI_UNINITIALIZED
2866 Used for struct gdbarch_info if ABI is still uninitialized.
2868 @findex GDB_OSABI_UNKNOWN
2869 @item GDB_OSABI_UNKNOWN
2870 The ABI of the inferior is unknown. The default @code{gdbarch}
2871 settings for the architecture will be used.
2873 @findex GDB_OSABI_SVR4
2874 @item GDB_OSABI_SVR4
2875 UNIX System V Release 4.
2877 @findex GDB_OSABI_HURD
2878 @item GDB_OSABI_HURD
2879 GNU using the Hurd kernel.
2881 @findex GDB_OSABI_SOLARIS
2882 @item GDB_OSABI_SOLARIS
2885 @findex GDB_OSABI_OSF1
2886 @item GDB_OSABI_OSF1
2887 OSF/1, including Digital UNIX and Compaq Tru64 UNIX.
2889 @findex GDB_OSABI_LINUX
2890 @item GDB_OSABI_LINUX
2891 GNU using the Linux kernel.
2893 @findex GDB_OSABI_FREEBSD_AOUT
2894 @item GDB_OSABI_FREEBSD_AOUT
2895 FreeBSD using the @code{a.out} executable format.
2897 @findex GDB_OSABI_FREEBSD_ELF
2898 @item GDB_OSABI_FREEBSD_ELF
2899 FreeBSD using the ELF executable format.
2901 @findex GDB_OSABI_NETBSD_AOUT
2902 @item GDB_OSABI_NETBSD_AOUT
2903 NetBSD using the @code{a.out} executable format.
2905 @findex GDB_OSABI_NETBSD_ELF
2906 @item GDB_OSABI_NETBSD_ELF
2907 NetBSD using the ELF executable format.
2909 @findex GDB_OSABI_OPENBSD_ELF
2910 @item GDB_OSABI_OPENBSD_ELF
2911 OpenBSD using the ELF executable format.
2913 @findex GDB_OSABI_WINCE
2914 @item GDB_OSABI_WINCE
2917 @findex GDB_OSABI_GO32
2918 @item GDB_OSABI_GO32
2921 @findex GDB_OSABI_IRIX
2922 @item GDB_OSABI_IRIX
2925 @findex GDB_OSABI_INTERIX
2926 @item GDB_OSABI_INTERIX
2927 Interix (Posix layer for MS-Windows systems).
2929 @findex GDB_OSABI_HPUX_ELF
2930 @item GDB_OSABI_HPUX_ELF
2931 HP/UX using the ELF executable format.
2933 @findex GDB_OSABI_HPUX_SOM
2934 @item GDB_OSABI_HPUX_SOM
2935 HP/UX using the SOM executable format.
2937 @findex GDB_OSABI_QNXNTO
2938 @item GDB_OSABI_QNXNTO
2941 @findex GDB_OSABI_CYGWIN
2942 @item GDB_OSABI_CYGWIN
2945 @findex GDB_OSABI_AIX
2951 Here are the functions that make up the OS ABI framework:
2953 @deftypefun {const char *} gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2954 Return the name of the OS ABI corresponding to @var{osabi}.
2957 @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}))
2958 Register the OS ABI handler specified by @var{init_osabi} for the
2959 architecture, machine type and OS ABI specified by @var{arch},
2960 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2961 machine type, which implies the architecture's default machine type,
2965 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2966 Register the OS ABI file sniffer specified by @var{sniffer} for the
2967 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2968 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2969 be generic, and is allowed to examine @var{flavour}-flavoured files for
2973 @deftypefun {enum gdb_osabi} gdbarch_lookup_osabi (bfd *@var{abfd})
2974 Examine the file described by @var{abfd} to determine its OS ABI.
2975 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2979 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2980 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2981 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2982 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2983 architecture, a warning will be issued and the debugging session will continue
2984 with the defaults already established for @var{gdbarch}.
2987 @deftypefun void generic_elf_osabi_sniff_abi_tag_sections (bfd *@var{abfd}, asection *@var{sect}, void *@var{obj})
2988 Helper routine for ELF file sniffers. Examine the file described by
2989 @var{abfd} and look at ABI tag note sections to determine the OS ABI
2990 from the note. This function should be called via
2991 @code{bfd_map_over_sections}.
2994 @node Initialize New Architecture
2995 @section Initializing a New Architecture
2998 * How an Architecture is Represented::
2999 * Looking Up an Existing Architecture::
3000 * Creating a New Architecture::
3003 @node How an Architecture is Represented
3004 @subsection How an Architecture is Represented
3005 @cindex architecture representation
3006 @cindex representation of architecture
3008 Each @code{gdbarch} is associated with a single @sc{bfd} architecture,
3009 via a @code{bfd_arch_@var{arch}} in the @code{bfd_architecture}
3010 enumeration. The @code{gdbarch} is registered by a call to
3011 @code{register_gdbarch_init}, usually from the file's
3012 @code{_initialize_@var{filename}} routine, which will be automatically
3013 called during @value{GDBN} startup. The arguments are a @sc{bfd}
3014 architecture constant and an initialization function.
3016 @findex _initialize_@var{arch}_tdep
3017 @cindex @file{@var{arch}-tdep.c}
3018 A @value{GDBN} description for a new architecture, @var{arch} is created by
3019 defining a global function @code{_initialize_@var{arch}_tdep}, by
3020 convention in the source file @file{@var{arch}-tdep.c}. For example,
3021 in the case of the OpenRISC 1000, this function is called
3022 @code{_initialize_or1k_tdep} and is found in the file
3025 @cindex @file{configure.tgt}
3026 @cindex @code{gdbarch}
3027 @findex gdbarch_register
3028 The resulting object files containing the implementation of the
3029 @code{_initialize_@var{arch}_tdep} function are specified in the @value{GDBN}
3030 @file{configure.tgt} file, which includes a large case statement
3031 pattern matching against the @code{--target} option of the
3032 @code{configure} script. The new @code{struct gdbarch} is created
3033 within the @code{_initialize_@var{arch}_tdep} function by calling
3034 @code{gdbarch_register}:
3037 void gdbarch_register (enum bfd_architecture @var{architecture},
3038 gdbarch_init_ftype *@var{init_func},
3039 gdbarch_dump_tdep_ftype *@var{tdep_dump_func});
3042 The @var{architecture} will identify the unique @sc{bfd} to be
3043 associated with this @code{gdbarch}. The @var{init_func} funciton is
3044 called to create and return the new @code{struct gdbarch}. The
3045 @var{tdep_dump_func} function will dump the target specific details
3046 associated with this architecture.
3048 For example the function @code{_initialize_or1k_tdep} creates its
3049 architecture for 32-bit OpenRISC 1000 architectures by calling:
3052 gdbarch_register (bfd_arch_or32, or1k_gdbarch_init, or1k_dump_tdep);
3055 @node Looking Up an Existing Architecture
3056 @subsection Looking Up an Existing Architecture
3057 @cindex @code{gdbarch} lookup
3059 The initialization function has this prototype:
3062 static struct gdbarch *
3063 @var{arch}_gdbarch_init (struct gdbarch_info @var{info},
3064 struct gdbarch_list *@var{arches})
3067 The @var{info} argument contains parameters used to select the correct
3068 architecture, and @var{arches} is a list of architectures which
3069 have already been created with the same @code{bfd_arch_@var{arch}}
3072 The initialization function should first make sure that @var{info}
3073 is acceptable, and return @code{NULL} if it is not. Then, it should
3074 search through @var{arches} for an exact match to @var{info}, and
3075 return one if found. Lastly, if no exact match was found, it should
3076 create a new architecture based on @var{info} and return it.
3078 @findex gdbarch_list_lookup_by_info
3079 @cindex @code{gdbarch_info}
3080 The lookup is done using @code{gdbarch_list_lookup_by_info}. It is
3081 passed the list of existing architectures, @var{arches}, and the
3082 @code{struct gdbarch_info}, @var{info}, and returns the first matching
3083 architecture it finds, or @code{NULL} if none are found. If an
3084 architecture is found it can be returned as the result from the
3085 initialization function, otherwise a new @code{struct gdbach} will need
3088 The struct gdbarch_info has the following components:
3093 const struct bfd_arch_info *bfd_arch_info;
3096 struct gdbarch_tdep_info *tdep_info;
3097 enum gdb_osabi osabi;
3098 const struct target_desc *target_desc;
3102 @vindex bfd_arch_info
3103 The @code{bfd_arch_info} member holds the key details about the
3104 architecture. The @code{byte_order} member is a value in an
3105 enumeration indicating the endianism. The @code{abfd} member is a
3106 pointer to the full @sc{bfd}, the @code{tdep_info} member is
3107 additional custom target specific information, @code{osabi} identifies
3108 which (if any) of a number of operating specific ABIs are used by this
3109 architecture and the @code{target_desc} member is a set of name-value
3110 pairs with information about register usage in this target.
3112 When the @code{struct gdbarch} initialization function is called, not
3113 all the fields are provided---only those which can be deduced from the
3114 @sc{bfd}. The @code{struct gdbarch_info}, @var{info} is used as a
3115 look-up key with the list of existing architectures, @var{arches} to
3116 see if a suitable architecture already exists. The @var{tdep_info},
3117 @var{osabi} and @var{target_desc} fields may be added before this
3118 lookup to refine the search.
3120 Only information in @var{info} should be used to choose the new
3121 architecture. Historically, @var{info} could be sparse, and
3122 defaults would be collected from the first element on @var{arches}.
3123 However, @value{GDBN} now fills in @var{info} more thoroughly,
3124 so new @code{gdbarch} initialization functions should not take
3125 defaults from @var{arches}.
3127 @node Creating a New Architecture
3128 @subsection Creating a New Architecture
3129 @cindex @code{struct gdbarch} creation
3131 @findex gdbarch_alloc
3132 @cindex @code{gdbarch_tdep} when allocating new @code{gdbarch}
3133 If no architecture is found, then a new architecture must be created,
3134 by calling @code{gdbarch_alloc} using the supplied @code{@w{struct
3135 gdbarch_info}} and any additional custom target specific
3136 information in a @code{struct gdbarch_tdep}. The prototype for
3137 @code{gdbarch_alloc} is:
3140 struct gdbarch *gdbarch_alloc (const struct gdbarch_info *@var{info},
3141 struct gdbarch_tdep *@var{tdep});
3144 @cindex @code{set_gdbarch} functions
3145 @cindex @code{gdbarch} accessor functions
3146 The newly created struct gdbarch must then be populated. Although
3147 there are default values, in most cases they are not what is
3150 For each element, @var{X}, there is are a pair of corresponding accessor
3151 functions, one to set the value of that element,
3152 @code{set_gdbarch_@var{X}}, the second to either get the value of an
3153 element (if it is a variable) or to apply the element (if it is a
3154 function), @code{gdbarch_@var{X}}. Note that both accessor functions
3155 take a pointer to the @code{@w{struct gdbarch}} as first
3156 argument. Populating the new @code{gdbarch} should use the
3157 @code{set_gdbarch} functions.
3159 The following sections identify the main elements that should be set
3160 in this way. This is not the complete list, but represents the
3161 functions and elements that must commonly be specified for a new
3162 architecture. Many of the functions and variables are described in the
3163 header file @file{gdbarch.h}.
3165 This is the main work in defining a new architecture. Implementing the
3166 set of functions to populate the @code{struct gdbarch}.
3168 @cindex @code{gdbarch_tdep} definition
3169 @code{struct gdbarch_tdep} is not defined within @value{GDBN}---it is up
3170 to the user to define this struct if it is needed to hold custom target
3171 information that is not covered by the standard @code{@w{struct
3172 gdbarch}}. For example with the OpenRISC 1000 architecture it is used to
3173 hold the number of matchpoints available in the target (along with other
3176 If there is no additional target specific information, it can be set to
3179 @node Registers and Memory
3180 @section Registers and Memory
3182 @value{GDBN}'s model of the target machine is rather simple.
3183 @value{GDBN} assumes the machine includes a bank of registers and a
3184 block of memory. Each register may have a different size.
3186 @value{GDBN} does not have a magical way to match up with the
3187 compiler's idea of which registers are which; however, it is critical
3188 that they do match up accurately. The only way to make this work is
3189 to get accurate information about the order that the compiler uses,
3190 and to reflect that in the @code{gdbarch_register_name} and related functions.
3192 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
3194 @node Pointers and Addresses
3195 @section Pointers Are Not Always Addresses
3196 @cindex pointer representation
3197 @cindex address representation
3198 @cindex word-addressed machines
3199 @cindex separate data and code address spaces
3200 @cindex spaces, separate data and code address
3201 @cindex address spaces, separate data and code
3202 @cindex code pointers, word-addressed
3203 @cindex converting between pointers and addresses
3204 @cindex D10V addresses
3206 On almost all 32-bit architectures, the representation of a pointer is
3207 indistinguishable from the representation of some fixed-length number
3208 whose value is the byte address of the object pointed to. On such
3209 machines, the words ``pointer'' and ``address'' can be used interchangeably.
3210 However, architectures with smaller word sizes are often cramped for
3211 address space, so they may choose a pointer representation that breaks this
3212 identity, and allows a larger code address space.
3214 @c D10V is gone from sources - more current example?
3216 For example, the Renesas D10V is a 16-bit VLIW processor whose
3217 instructions are 32 bits long@footnote{Some D10V instructions are
3218 actually pairs of 16-bit sub-instructions. However, since you can't
3219 jump into the middle of such a pair, code addresses can only refer to
3220 full 32 bit instructions, which is what matters in this explanation.}.
3221 If the D10V used ordinary byte addresses to refer to code locations,
3222 then the processor would only be able to address 64kb of instructions.
3223 However, since instructions must be aligned on four-byte boundaries, the
3224 low two bits of any valid instruction's byte address are always
3225 zero---byte addresses waste two bits. So instead of byte addresses,
3226 the D10V uses word addresses---byte addresses shifted right two bits---to
3227 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
3230 However, this means that code pointers and data pointers have different
3231 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
3232 @code{0xC020} when used as a data address, but refers to byte address
3233 @code{0x30080} when used as a code address.
3235 (The D10V also uses separate code and data address spaces, which also
3236 affects the correspondence between pointers and addresses, but we're
3237 going to ignore that here; this example is already too long.)
3239 To cope with architectures like this---the D10V is not the only
3240 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
3241 byte numbers, and @dfn{pointers}, which are the target's representation
3242 of an address of a particular type of data. In the example above,
3243 @code{0xC020} is the pointer, which refers to one of the addresses
3244 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
3245 @value{GDBN} provides functions for turning a pointer into an address
3246 and vice versa, in the appropriate way for the current architecture.
3248 Unfortunately, since addresses and pointers are identical on almost all
3249 processors, this distinction tends to bit-rot pretty quickly. Thus,
3250 each time you port @value{GDBN} to an architecture which does
3251 distinguish between pointers and addresses, you'll probably need to
3252 clean up some architecture-independent code.
3254 Here are functions which convert between pointers and addresses:
3256 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
3257 Treat the bytes at @var{buf} as a pointer or reference of type
3258 @var{type}, and return the address it represents, in a manner
3259 appropriate for the current architecture. This yields an address
3260 @value{GDBN} can use to read target memory, disassemble, etc. Note that
3261 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
3264 For example, if the current architecture is the Intel x86, this function
3265 extracts a little-endian integer of the appropriate length from
3266 @var{buf} and returns it. However, if the current architecture is the
3267 D10V, this function will return a 16-bit integer extracted from
3268 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
3270 If @var{type} is not a pointer or reference type, then this function
3271 will signal an internal error.
3274 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
3275 Store the address @var{addr} in @var{buf}, in the proper format for a
3276 pointer of type @var{type} in the current architecture. Note that
3277 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
3280 For example, if the current architecture is the Intel x86, this function
3281 stores @var{addr} unmodified as a little-endian integer of the
3282 appropriate length in @var{buf}. However, if the current architecture
3283 is the D10V, this function divides @var{addr} by four if @var{type} is
3284 a pointer to a function, and then stores it in @var{buf}.
3286 If @var{type} is not a pointer or reference type, then this function
3287 will signal an internal error.
3290 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
3291 Assuming that @var{val} is a pointer, return the address it represents,
3292 as appropriate for the current architecture.
3294 This function actually works on integral values, as well as pointers.
3295 For pointers, it performs architecture-specific conversions as
3296 described above for @code{extract_typed_address}.
3299 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
3300 Create and return a value representing a pointer of type @var{type} to
3301 the address @var{addr}, as appropriate for the current architecture.
3302 This function performs architecture-specific conversions as described
3303 above for @code{store_typed_address}.
3306 Here are two functions which architectures can define to indicate the
3307 relationship between pointers and addresses. These have default
3308 definitions, appropriate for architectures on which all pointers are
3309 simple unsigned byte addresses.
3311 @deftypefun CORE_ADDR gdbarch_pointer_to_address (struct gdbarch *@var{gdbarch}, struct type *@var{type}, char *@var{buf})
3312 Assume that @var{buf} holds a pointer of type @var{type}, in the
3313 appropriate format for the current architecture. Return the byte
3314 address the pointer refers to.
3316 This function may safely assume that @var{type} is either a pointer or a
3317 C@t{++} reference type.
3320 @deftypefun void gdbarch_address_to_pointer (struct gdbarch *@var{gdbarch}, struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
3321 Store in @var{buf} a pointer of type @var{type} representing the address
3322 @var{addr}, in the appropriate format for the current architecture.
3324 This function may safely assume that @var{type} is either a pointer or a
3325 C@t{++} reference type.
3328 @node Address Classes
3329 @section Address Classes
3330 @cindex address classes
3331 @cindex DW_AT_byte_size
3332 @cindex DW_AT_address_class
3334 Sometimes information about different kinds of addresses is available
3335 via the debug information. For example, some programming environments
3336 define addresses of several different sizes. If the debug information
3337 distinguishes these kinds of address classes through either the size
3338 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
3339 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
3340 following macros should be defined in order to disambiguate these
3341 types within @value{GDBN} as well as provide the added information to
3342 a @value{GDBN} user when printing type expressions.
3344 @deftypefun int gdbarch_address_class_type_flags (struct gdbarch *@var{gdbarch}, int @var{byte_size}, int @var{dwarf2_addr_class})
3345 Returns the type flags needed to construct a pointer type whose size
3346 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
3347 This function is normally called from within a symbol reader. See
3348 @file{dwarf2read.c}.
3351 @deftypefun {char *} gdbarch_address_class_type_flags_to_name (struct gdbarch *@var{gdbarch}, int @var{type_flags})
3352 Given the type flags representing an address class qualifier, return
3355 @deftypefun int gdbarch_address_class_name_to_type_flags (struct gdbarch *@var{gdbarch}, int @var{name}, int *@var{type_flags_ptr})
3356 Given an address qualifier name, set the @code{int} referenced by @var{type_flags_ptr} to the type flags
3357 for that address class qualifier.
3360 Since the need for address classes is rather rare, none of
3361 the address class functions are defined by default. Predicate
3362 functions are provided to detect when they are defined.
3364 Consider a hypothetical architecture in which addresses are normally
3365 32-bits wide, but 16-bit addresses are also supported. Furthermore,
3366 suppose that the @w{DWARF 2} information for this architecture simply
3367 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
3368 of these "short" pointers. The following functions could be defined
3369 to implement the address class functions:
3372 somearch_address_class_type_flags (int byte_size,
3373 int dwarf2_addr_class)
3376 return TYPE_FLAG_ADDRESS_CLASS_1;
3382 somearch_address_class_type_flags_to_name (int type_flags)
3384 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
3391 somearch_address_class_name_to_type_flags (char *name,
3392 int *type_flags_ptr)
3394 if (strcmp (name, "short") == 0)
3396 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
3404 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
3405 to indicate the presence of one of these ``short'' pointers. For
3406 example if the debug information indicates that @code{short_ptr_var} is
3407 one of these short pointers, @value{GDBN} might show the following
3411 (gdb) ptype short_ptr_var
3412 type = int * @@short
3416 @node Register Representation
3417 @section Register Representation
3420 * Raw and Cooked Registers::
3421 * Register Architecture Functions & Variables::
3422 * Register Information Functions::
3423 * Register and Memory Data::
3424 * Register Caching::
3427 @node Raw and Cooked Registers
3428 @subsection Raw and Cooked Registers
3429 @cindex raw register representation
3430 @cindex cooked register representation
3431 @cindex representations, raw and cooked registers
3433 @value{GDBN} considers registers to be a set with members numbered
3434 linearly from 0 upwards. The first part of that set corresponds to real
3435 physical registers, the second part to any @dfn{pseudo-registers}.
3436 Pseudo-registers have no independent physical existence, but are useful
3437 representations of information within the architecture. For example the
3438 OpenRISC 1000 architecture has up to 32 general purpose registers, which
3439 are typically represented as 32-bit (or 64-bit) integers. However the
3440 GPRs are also used as operands to the floating point operations, and it
3441 could be convenient to define a set of pseudo-registers, to show the
3442 GPRs represented as floating point values.
3444 For any architecture, the implementer will decide on a mapping from
3445 hardware to @value{GDBN} register numbers. The registers corresponding to real
3446 hardware are referred to as @dfn{raw} registers, the remaining registers are
3447 @dfn{pseudo-registers}. The total register set (raw and pseudo) is called
3448 the @dfn{cooked} register set.
3451 @node Register Architecture Functions & Variables
3452 @subsection Functions and Variables Specifying the Register Architecture
3453 @cindex @code{gdbarch} register architecture functions
3455 These @code{struct gdbarch} functions and variables specify the number
3456 and type of registers in the architecture.
3458 @deftypefn {Architecture Function} CORE_ADDR read_pc (struct regcache *@var{regcache})
3460 @deftypefn {Architecture Function} void write_pc (struct regcache *@var{regcache}, CORE_ADDR @var{val})
3462 Read or write the program counter. The default value of both
3463 functions is @code{NULL} (no function available). If the program
3464 counter is just an ordinary register, it can be specified in
3465 @code{struct gdbarch} instead (see @code{pc_regnum} below) and it will
3466 be read or written using the standard routines to access registers. This
3467 function need only be specified if the program counter is not an
3470 Any register information can be obtained using the supplied register
3471 cache, @var{regcache}. @xref{Register Caching, , Register Caching}.
3475 @deftypefn {Architecture Function} void pseudo_register_read (struct gdbarch *@var{gdbarch}, struct regcache *@var{regcache}, int @var{regnum}, const gdb_byte *@var{buf})
3477 @deftypefn {Architecture Function} void pseudo_register_write (struct gdbarch *@var{gdbarch}, struct regcache *@var{regcache}, int @var{regnum}, const gdb_byte *@var{buf})
3479 These functions should be defined if there are any pseudo-registers.
3480 The default value is @code{NULL}. @var{regnum} is the number of the
3481 register to read or write (which will be a @dfn{cooked} register
3482 number) and @var{buf} is the buffer where the value read will be
3483 placed, or from which the value to be written will be taken. The
3484 value in the buffer may be converted to or from a signed or unsigned
3485 integral value using one of the utility functions (@pxref{Register and
3486 Memory Data, , Using Different Register and Memory Data
3489 The access should be for the specified architecture,
3490 @var{gdbarch}. Any register information can be obtained using the
3491 supplied register cache, @var{regcache}. @xref{Register Caching, ,
3496 @deftypevr {Architecture Variable} int sp_regnum
3498 @cindex stack pointer
3501 This specifies the register holding the stack pointer, which may be a
3502 raw or pseudo-register. It defaults to -1 (not defined), but it is an
3503 error for it not to be defined.
3505 The value of the stack pointer register can be accessed withing
3506 @value{GDBN} as the variable @kbd{$sp}.
3510 @deftypevr {Architecture Variable} int pc_regnum
3512 @cindex program counter
3515 This specifies the register holding the program counter, which may be a
3516 raw or pseudo-register. It defaults to -1 (not defined). If
3517 @code{pc_regnum} is not defined, then the functions @code{read_pc} and
3518 @code{write_pc} (see above) must be defined.
3520 The value of the program counter (whether defined as a register, or
3521 through @code{read_pc} and @code{write_pc}) can be accessed withing
3522 @value{GDBN} as the variable @kbd{$pc}.
3526 @deftypevr {Architecture Variable} int ps_regnum
3528 @cindex processor status register
3529 @cindex status register
3532 This specifies the register holding the processor status (often called
3533 the status register), which may be a raw or pseudo-register. It
3534 defaults to -1 (not defined).
3536 If defined, the value of this register can be accessed withing
3537 @value{GDBN} as the variable @kbd{$ps}.
3541 @deftypevr {Architecture Variable} int fp0_regnum
3543 @cindex first floating point register
3545 This specifies the first floating point register. It defaults to
3546 0. @code{fp0_regnum} is not needed unless the target offers support
3551 @node Register Information Functions
3552 @subsection Functions Giving Register Information
3553 @cindex @code{gdbarch} register information functions
3555 These functions return information about registers.
3557 @deftypefn {Architecture Function} {const char *} register_name (struct gdbarch *@var{gdbarch}, int @var{regnum})
3559 This function should convert a register number (raw or pseudo) to a
3560 register name (as a C @code{const char *}). This is used both to
3561 determine the name of a register for output and to work out the meaning
3562 of any register names used as input. The function may also return
3563 @code{NULL}, to indicate that @var{regnum} is not a valid register.
3565 For example with the OpenRISC 1000, @value{GDBN} registers 0-31 are the
3566 General Purpose Registers, register 32 is the program counter and
3567 register 33 is the supervision register (i.e.@: the processor status
3568 register), which map to the strings @code{"gpr00"} through
3569 @code{"gpr31"}, @code{"pc"} and @code{"sr"} respectively. This means
3570 that the @value{GDBN} command @kbd{print $gpr5} should print the value of
3571 the OR1K general purpose register 5@footnote{
3572 @cindex frame pointer
3574 Historically, @value{GDBN} always had a concept of a frame pointer
3575 register, which could be accessed via the @value{GDBN} variable,
3576 @kbd{$fp}. That concept is now deprecated, recognizing that not all
3577 architectures have a frame pointer. However if an architecture does
3578 have a frame pointer register, and defines a register or
3579 pseudo-register with the name @code{"fp"}, then that register will be
3580 used as the value of the @kbd{$fp} variable.}.
3582 The default value for this function is @code{NULL}, meaning
3583 undefined. It should always be defined.
3585 The access should be for the specified architecture, @var{gdbarch}.
3589 @deftypefn {Architecture Function} {struct type *} register_type (struct gdbarch *@var{gdbarch}, int @var{regnum})
3591 Given a register number, this function identifies the type of data it
3592 may be holding, specified as a @code{struct type}. @value{GDBN} allows
3593 creation of arbitrary types, but a number of built in types are
3594 provided (@code{builtin_type_void}, @code{builtin_type_int32} etc),
3595 together with functions to derive types from these.
3597 Typically the program counter will have a type of ``pointer to
3598 function'' (it points to code), the frame pointer and stack pointer
3599 will have types of ``pointer to void'' (they point to data on the stack)
3600 and all other integer registers will have a type of 32-bit integer or
3603 This information guides the formatting when displaying register
3604 information. The default value is @code{NULL} meaning no information is
3605 available to guide formatting when displaying registers.
3609 @deftypefn {Architecture Function} void print_registers_info (struct gdbarch *@var{gdbarch}, struct ui_file *@var{file}, struct frame_info *@var{frame}, int @var{regnum}, int @var{all})
3611 Define this function to print out one or all of the registers for the
3612 @value{GDBN} @kbd{info registers} command. The default value is the
3613 function @code{default_print_registers_info}, which uses the register
3614 type information (see @code{register_type} above) to determine how each
3615 register should be printed. Define a custom version of this function
3616 for fuller control over how the registers are displayed.
3618 The access should be for the specified architecture, @var{gdbarch},
3619 with output to the the file specified by the User Interface
3620 Independent Output file handle, @var{file} (@pxref{UI-Independent
3621 Output, , UI-Independent Output---the @code{ui_out}
3624 The registers should show their values in the frame specified by
3625 @var{frame}. If @var{regnum} is -1 and @var{all} is zero, then all
3626 the ``significant'' registers should be shown (the implementer should
3627 decide which registers are ``significant''). Otherwise only the value of
3628 the register specified by @var{regnum} should be output. If
3629 @var{regnum} is -1 and @var{all} is non-zero (true), then the value of
3630 all registers should be shown.
3632 By default @code{default_print_registers_info} prints one register per
3633 line, and if @var{all} is zero omits floating-point registers.
3637 @deftypefn {Architecture Function} void print_float_info (struct gdbarch *@var{gdbarch}, struct ui_file *@var{file}, struct frame_info *@var{frame}, const char *@var{args})
3639 Define this function to provide output about the floating point unit and
3640 registers for the @value{GDBN} @kbd{info float} command respectively.
3641 The default value is @code{NULL} (not defined), meaning no information
3644 The @var{gdbarch} and @var{file} and @var{frame} arguments have the same
3645 meaning as in the @code{print_registers_info} function above. The string
3646 @var{args} contains any supplementary arguments to the @kbd{info float}
3649 Define this function if the target supports floating point operations.
3653 @deftypefn {Architecture Function} void print_vector_info (struct gdbarch *@var{gdbarch}, struct ui_file *@var{file}, struct frame_info *@var{frame}, const char *@var{args})
3655 Define this function to provide output about the vector unit and
3656 registers for the @value{GDBN} @kbd{info vector} command respectively.
3657 The default value is @code{NULL} (not defined), meaning no information
3660 The @var{gdbarch}, @var{file} and @var{frame} arguments have the
3661 same meaning as in the @code{print_registers_info} function above. The
3662 string @var{args} contains any supplementary arguments to the @kbd{info
3665 Define this function if the target supports vector operations.
3669 @deftypefn {Architecture Function} int register_reggroup_p (struct gdbarch *@var{gdbarch}, int @var{regnum}, struct reggroup *@var{group})
3671 @value{GDBN} groups registers into different categories (general,
3672 vector, floating point etc). This function, given a register,
3673 @var{regnum}, and group, @var{group}, returns 1 (true) if the register
3674 is in the group and 0 (false) otherwise.
3676 The information should be for the specified architecture,
3679 The default value is the function @code{default_register_reggroup_p}
3680 which will do a reasonable job based on the type of the register (see
3681 the function @code{register_type} above), with groups for general
3682 purpose registers, floating point registers, vector registers and raw
3683 (i.e not pseudo) registers.
3687 @node Register and Memory Data
3688 @subsection Using Different Register and Memory Data Representations
3689 @cindex register representation
3690 @cindex memory representation
3691 @cindex representations, register and memory
3692 @cindex register data formats, converting
3693 @cindex @code{struct value}, converting register contents to
3695 Some architectures have different representations of data objects,
3696 depending whether the object is held in a register or memory. For
3702 The Alpha architecture can represent 32 bit integer values in
3703 floating-point registers.
3706 The x86 architecture supports 80-bit floating-point registers. The
3707 @code{long double} data type occupies 96 bits in memory but only 80
3708 bits when stored in a register.
3712 In general, the register representation of a data type is determined by
3713 the architecture, or @value{GDBN}'s interface to the architecture, while
3714 the memory representation is determined by the Application Binary
3717 For almost all data types on almost all architectures, the two
3718 representations are identical, and no special handling is needed.
3719 However, they do occasionally differ. An architecture may define the
3720 following @code{struct gdbarch} functions to request conversions
3721 between the register and memory representations of a data type:
3723 @deftypefn {Architecture Function} int gdbarch_convert_register_p (struct gdbarch *@var{gdbarch}, int @var{reg})
3725 Return non-zero (true) if the representation of a data value stored in
3726 this register may be different to the representation of that same data
3727 value when stored in memory. The default value is @code{NULL}
3730 If this function is defined and returns non-zero, the @code{struct
3731 gdbarch} functions @code{gdbarch_register_to_value} and
3732 @code{gdbarch_value_to_register} (see below) should be used to perform
3733 any necessary conversion.
3735 If defined, this function should return zero for the register's native
3736 type, when no conversion is necessary.
3739 @deftypefn {Architecture Function} void gdbarch_register_to_value (struct gdbarch *@var{gdbarch}, int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
3741 Convert the value of register number @var{reg} to a data object of
3742 type @var{type}. The buffer at @var{from} holds the register's value
3743 in raw format; the converted value should be placed in the buffer at
3747 @emph{Note:} @code{gdbarch_register_to_value} and
3748 @code{gdbarch_value_to_register} take their @var{reg} and @var{type}
3749 arguments in different orders.
3752 @code{gdbarch_register_to_value} should only be used with registers
3753 for which the @code{gdbarch_convert_register_p} function returns a
3758 @deftypefn {Architecture Function} void gdbarch_value_to_register (struct gdbarch *@var{gdbarch}, struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3760 Convert a data value of type @var{type} to register number @var{reg}'
3764 @emph{Note:} @code{gdbarch_register_to_value} and
3765 @code{gdbarch_value_to_register} take their @var{reg} and @var{type}
3766 arguments in different orders.
3769 @code{gdbarch_value_to_register} should only be used with registers
3770 for which the @code{gdbarch_convert_register_p} function returns a
3775 @node Register Caching
3776 @subsection Register Caching
3777 @cindex register caching
3779 Caching of registers is used, so that the target does not need to be
3780 accessed and reanalyzed multiple times for each register in
3781 circumstances where the register value cannot have changed.
3783 @cindex @code{struct regcache}
3784 @value{GDBN} provides @code{struct regcache}, associated with a
3785 particular @code{struct gdbarch} to hold the cached values of the raw
3786 registers. A set of functions is provided to access both the raw
3787 registers (with @code{raw} in their name) and the full set of cooked
3788 registers (with @code{cooked} in their name). Functions are provided
3789 to ensure the register cache is kept synchronized with the values of
3790 the actual registers in the target.
3792 Accessing registers through the @code{struct regcache} routines will
3793 ensure that the appropriate @code{struct gdbarch} functions are called
3794 when necessary to access the underlying target architecture. In general
3795 users should use the @dfn{cooked} functions, since these will map to the
3796 @dfn{raw} functions automatically as appropriate.
3798 @findex regcache_cooked_read
3799 @findex regcache_cooked_write
3800 @cindex @code{gdb_byte}
3801 @findex regcache_cooked_read_signed
3802 @findex regcache_cooked_read_unsigned
3803 @findex regcache_cooked_write_signed
3804 @findex regcache_cooked_write_unsigned
3805 The two key functions are @code{regcache_cooked_read} and
3806 @code{regcache_cooked_write} which read or write a register from or to
3807 a byte buffer (type @code{gdb_byte *}). For convenience the wrapper
3808 functions @code{regcache_cooked_read_signed},
3809 @code{regcache_cooked_read_unsigned},
3810 @code{regcache_cooked_write_signed} and
3811 @code{regcache_cooked_write_unsigned} are provided, which read or
3812 write the value using the buffer and convert to or from an integral
3813 value as appropriate.
3815 @node Frame Interpretation
3816 @section Frame Interpretation
3819 * All About Stack Frames::
3820 * Frame Handling Terminology::
3822 * Functions and Variable to Analyze Frames::
3823 * Functions to Access Frame Data::
3824 * Analyzing Stacks---Frame Sniffers::
3827 @node All About Stack Frames
3828 @subsection All About Stack Frames
3830 @value{GDBN} needs to understand the stack on which local (automatic)
3831 variables are stored. The area of the stack containing all the local
3832 variables for a function invocation is known as the @dfn{stack frame}
3833 for that function (or colloquially just as the @dfn{frame}). In turn the
3834 function that called the function will have its stack frame, and so on
3835 back through the chain of functions that have been called.
3837 Almost all architectures have one register dedicated to point to the
3838 end of the stack (the @dfn{stack pointer}). Many have a second register
3839 which points to the start of the currently active stack frame (the
3840 @dfn{frame pointer}). The specific arrangements for an architecture are
3841 a key part of the ABI.
3843 A diagram helps to explain this. Here is a simple program to compute
3856 return n * fact (n - 1);
3864 for (i = 0; i < 10; i++)
3867 printf ("%d! = %d\n", i, f);
3872 Consider the state of the stack when the code reaches line 6 after the
3873 main program has called @code{fact@w{ }(3)}. The chain of function
3874 calls will be @code{main ()}, @code{fact@w{ }(3)}, @code{fact@w{
3875 }(2)}, @code{@w{fact (1)}} and @code{fact@w{ }(0)}.
3877 In this illustration the stack is falling (as used for example by the
3878 OpenRISC 1000 ABI). The stack pointer (SP) is at the end of the stack
3879 (lowest address) and the frame pointer (FP) is at the highest address
3880 in the current stack frame. The following diagram shows how the stack
3883 @center @image{stack_frame,14cm}
3885 In each stack frame, offset 0 from the stack pointer is the frame
3886 pointer of the previous frame and offset 4 (this is illustrating a
3887 32-bit architecture) from the stack pointer is the return address.
3888 Local variables are indexed from the frame pointer, with negative
3889 indexes. In the function @code{fact}, offset -4 from the frame
3890 pointer is the argument @var{n}. In the @code{main} function, offset
3891 -4 from the frame pointer is the local variable @var{i} and offset -8
3892 from the frame pointer is the local variable @var{f}@footnote{This is
3893 a simplified example for illustrative purposes only. Good optimizing
3894 compilers would not put anything on the stack for such simple
3895 functions. Indeed they might eliminate the recursion and use of the
3898 It is very easy to get confused when examining stacks. @value{GDBN}
3899 has terminology it uses rigorously throughout. The stack frame of the
3900 function currently executing, or where execution stopped is numbered
3901 zero. In this example frame #0 is the stack frame of the call to
3902 @code{fact@w{ }(0)}. The stack frame of its calling function
3903 (@code{fact@w{ }(1)} in this case) is numbered #1 and so on back
3904 through the chain of calls.
3906 The main @value{GDBN} data structure describing frames is
3907 @code{@w{struct frame_info}}. It is not used directly, but only via
3908 its accessor functions. @code{frame_info} includes information about
3909 the registers in the frame and a pointer to the code of the function
3910 with which the frame is associated. The entire stack is represented as
3911 a linked list of @code{frame_info} structs.
3913 @node Frame Handling Terminology
3914 @subsection Frame Handling Terminology
3916 It is easy to get confused when referencing stack frames. @value{GDBN}
3917 uses some precise terminology.
3923 @cindex stack frame, definition of THIS frame
3924 @cindex frame, definition of THIS frame
3925 @dfn{THIS} frame is the frame currently under consideration.
3929 @cindex stack frame, definition of NEXT frame
3930 @cindex frame, definition of NEXT frame
3931 The @dfn{NEXT} frame, also sometimes called the inner or newer frame is the
3932 frame of the function called by the function of THIS frame.
3935 @cindex PREVIOUS frame
3936 @cindex stack frame, definition of PREVIOUS frame
3937 @cindex frame, definition of PREVIOUS frame
3938 The @dfn{PREVIOUS} frame, also sometimes called the outer or older frame is
3939 the frame of the function which called the function of THIS frame.
3943 So in the example in the previous section (@pxref{All About Stack
3944 Frames, , All About Stack Frames}), if THIS frame is #3 (the call to
3945 @code{fact@w{ }(3)}), the NEXT frame is frame #2 (the call to
3946 @code{fact@w{ }(2)}) and the PREVIOUS frame is frame #4 (the call to
3947 @code{main@w{ }()}).
3949 @cindex innermost frame
3950 @cindex stack frame, definition of innermost frame
3951 @cindex frame, definition of innermost frame
3952 The @dfn{innermost} frame is the frame of the current executing
3953 function, or where the program stopped, in this example, in the middle
3954 of the call to @code{@w{fact (0))}}. It is always numbered frame #0.
3956 @cindex base of a frame
3957 @cindex stack frame, definition of base of a frame
3958 @cindex frame, definition of base of a frame
3959 The @dfn{base} of a frame is the address immediately before the start
3960 of the NEXT frame. For a stack which grows down in memory (a
3961 @dfn{falling} stack) this will be the lowest address and for a stack
3962 which grows up in memory (a @dfn{rising} stack) this will be the
3963 highest address in the frame.
3965 @value{GDBN} functions to analyze the stack are typically given a
3966 pointer to the NEXT frame to determine information about THIS
3967 frame. Information about THIS frame includes data on where the
3968 registers of the PREVIOUS frame are stored in this stack frame. In
3969 this example the frame pointer of the PREVIOUS frame is stored at
3970 offset 0 from the stack pointer of THIS frame.
3973 @cindex stack frame, definition of unwinding
3974 @cindex frame, definition of unwinding
3975 The process whereby a function is given a pointer to the NEXT
3976 frame to work out information about THIS frame is referred to as
3977 @dfn{unwinding}. The @value{GDBN} functions involved in this typically
3978 include unwind in their name.
3981 @cindex stack frame, definition of sniffing
3982 @cindex frame, definition of sniffing
3983 The process of analyzing a target to determine the information that
3984 should go in struct frame_info is called @dfn{sniffing}. The functions
3985 that carry this out are called sniffers and typically include sniffer
3986 in their name. More than one sniffer may be required to extract all
3987 the information for a particular frame.
3989 @cindex sentinel frame
3990 @cindex stack frame, definition of sentinel frame
3991 @cindex frame, definition of sentinel frame
3992 Because so many functions work using the NEXT frame, there is an issue
3993 about addressing the innermost frame---it has no NEXT frame. To solve
3994 this @value{GDBN} creates a dummy frame #-1, known as the
3995 @dfn{sentinel} frame.
3997 @node Prologue Caches
3998 @subsection Prologue Caches
4000 @cindex function prologue
4001 @cindex prologue of a function
4002 All the frame sniffing functions typically examine the code at the
4003 start of the corresponding function, to determine the state of
4004 registers. The ABI will save old values and set new values of key
4005 registers at the start of each function in what is known as the
4006 function @dfn{prologue}.
4008 @cindex prologue cache
4009 For any particular stack frame this data does not change, so all the
4010 standard unwinding functions, in addition to receiving a pointer to
4011 the NEXT frame as their first argument, receive a pointer to a
4012 @dfn{prologue cache} as their second argument. This can be used to store
4013 values associated with a particular frame, for reuse on subsequent
4014 calls involving the same frame.
4016 It is up to the user to define the structure used (it is a
4017 @code{void@w{ }*} pointer) and arrange allocation and deallocation of
4018 storage. However for general use, @value{GDBN} provides
4019 @code{@w{struct trad_frame_cache}}, with a set of accessor
4020 routines. This structure holds the stack and code address of
4021 THIS frame, the base address of the frame, a pointer to the
4022 struct @code{frame_info} for the NEXT frame and details of
4023 where the registers of the PREVIOUS frame may be found in THIS
4026 Typically the first time any sniffer function is called with NEXT
4027 frame, the prologue sniffer for THIS frame will be @code{NULL}. The
4028 sniffer will analyze the frame, allocate a prologue cache structure
4029 and populate it. Subsequent calls using the same NEXT frame will
4030 pass in this prologue cache, so the data can be returned with no
4031 additional analysis.
4033 @node Functions and Variable to Analyze Frames
4034 @subsection Functions and Variable to Analyze Frames
4036 These struct @code{gdbarch} functions and variable should be defined
4037 to provide analysis of the stack frame and allow it to be adjusted as
4040 @deftypefn {Architecture Function} CORE_ADDR skip_prologue (struct gdbarch *@var{gdbarch}, CORE_ADDR @var{pc})
4042 The prologue of a function is the code at the beginning of the
4043 function which sets up the stack frame, saves the return address
4044 etc. The code representing the behavior of the function starts after
4047 This function skips past the prologue of a function if the program
4048 counter, @var{pc}, is within the prologue of a function. The result is
4049 the program counter immediately after the prologue. With modern
4050 optimizing compilers, this may be a far from trivial exercise. However
4051 the required information may be within the binary as DWARF2 debugging
4052 information, making the job much easier.
4054 The default value is @code{NULL} (not defined). This function should always
4055 be provided, but can take advantage of DWARF2 debugging information,
4056 if that is available.
4060 @deftypefn {Architecture Function} int inner_than (CORE_ADDR @var{lhs}, CORE_ADDR @var{rhs})
4061 @findex core_addr_lessthan
4062 @findex core_addr_greaterthan
4064 Given two frame or stack pointers, return non-zero (true) if the first
4065 represents the @dfn{inner} stack frame and 0 (false) otherwise. This
4066 is used to determine whether the target has a stack which grows up in
4067 memory (rising stack) or grows down in memory (falling stack).
4068 @xref{All About Stack Frames, , All About Stack Frames}, for an
4069 explanation of @dfn{inner} frames.
4071 The default value of this function is @code{NULL} and it should always
4072 be defined. However for almost all architectures one of the built-in
4073 functions can be used: @code{core_addr_lessthan} (for stacks growing
4074 down in memory) or @code{core_addr_greaterthan} (for stacks growing up
4079 @anchor{frame_align}
4080 @deftypefn {Architecture Function} CORE_ADDR frame_align (struct gdbarch *@var{gdbarch}, CORE_ADDR @var{address})
4084 The architecture may have constraints on how its frames are
4085 aligned. For example the OpenRISC 1000 ABI requires stack frames to be
4086 double-word aligned, but 32-bit versions of the architecture allocate
4087 single-word values to the stack. Thus extra padding may be needed at
4088 the end of a stack frame.
4090 Given a proposed address for the stack pointer, this function
4091 returns a suitably aligned address (by expanding the stack frame).
4093 The default value is @code{NULL} (undefined). This function should be defined
4094 for any architecture where it is possible the stack could become
4095 misaligned. The utility functions @code{align_down} (for falling
4096 stacks) and @code{align_up} (for rising stacks) will facilitate the
4097 implementation of this function.
4101 @deftypevr {Architecture Variable} int frame_red_zone_size
4103 Some ABIs reserve space beyond the end of the stack for use by leaf
4104 functions without prologue or epilogue or by exception handlers (for
4105 example the OpenRISC 1000).
4107 This is known as a @dfn{red zone} (AMD terminology). The @sc{amd64}
4108 (nee x86-64) ABI documentation refers to the @dfn{red zone} when
4109 describing this scratch area.
4111 The default value is 0. Set this field if the architecture has such a
4112 red zone. The value must be aligned as required by the ABI (see
4113 @code{frame_align} above for an explanation of stack frame alignment).
4117 @node Functions to Access Frame Data
4118 @subsection Functions to Access Frame Data
4120 These functions provide access to key registers and arguments in the
4123 @deftypefn {Architecture Function} CORE_ADDR unwind_pc (struct gdbarch *@var{gdbarch}, struct frame_info *@var{next_frame})
4125 This function is given a pointer to the NEXT stack frame (@pxref{All
4126 About Stack Frames, , All About Stack Frames}, for how frames are
4127 represented) and returns the value of the program counter in the
4128 PREVIOUS frame (i.e.@: the frame of the function that called THIS
4129 one). This is commonly referred to as the @dfn{return address}.
4131 The implementation, which must be frame agnostic (work with any frame),
4132 is typically no more than:
4136 pc = frame_unwind_register_unsigned (next_frame, @var{ARCH}_PC_REGNUM);
4137 return gdbarch_addr_bits_remove (gdbarch, pc);
4142 @deftypefn {Architecture Function} CORE_ADDR unwind_sp (struct gdbarch *@var{gdbarch}, struct frame_info *@var{next_frame})
4144 This function is given a pointer to the NEXT stack frame
4145 (@pxref{All About Stack Frames, , All About Stack Frames} for how
4146 frames are represented) and returns the value of the stack pointer in
4147 the PREVIOUS frame (i.e.@: the frame of the function that called
4150 The implementation, which must be frame agnostic (work with any frame),
4151 is typically no more than:
4155 sp = frame_unwind_register_unsigned (next_frame, @var{ARCH}_SP_REGNUM);
4156 return gdbarch_addr_bits_remove (gdbarch, sp);
4161 @deftypefn {Architecture Function} int frame_num_args (struct gdbarch *@var{gdbarch}, struct frame_info *@var{this_frame})
4163 This function is given a pointer to THIS stack frame (@pxref{All
4164 About Stack Frames, , All About Stack Frames} for how frames are
4165 represented), and returns the number of arguments that are being
4166 passed, or -1 if not known.
4168 The default value is @code{NULL} (undefined), in which case the number of
4169 arguments passed on any stack frame is always unknown. For many
4170 architectures this will be a suitable default.
4174 @node Analyzing Stacks---Frame Sniffers
4175 @subsection Analyzing Stacks---Frame Sniffers
4177 When a program stops, @value{GDBN} needs to construct the chain of
4178 struct @code{frame_info} representing the state of the stack using
4179 appropriate @dfn{sniffers}.
4181 Each architecture requires appropriate sniffers, but they do not form
4182 entries in @code{@w{struct gdbarch}}, since more than one sniffer may
4183 be required and a sniffer may be suitable for more than one
4184 @code{@w{struct gdbarch}}. Instead sniffers are associated with
4185 architectures using the following functions.
4190 @findex frame_unwind_append_sniffer
4191 @code{frame_unwind_append_sniffer} is used to add a new sniffer to
4192 analyze THIS frame when given a pointer to the NEXT frame.
4195 @findex frame_base_append_sniffer
4196 @code{frame_base_append_sniffer} is used to add a new sniffer
4197 which can determine information about the base of a stack frame.
4200 @findex frame_base_set_default
4201 @code{frame_base_set_default} is used to specify the default base
4206 These functions all take a reference to @code{@w{struct gdbarch}}, so
4207 they are associated with a specific architecture. They are usually
4208 called in the @code{gdbarch} initialization function, after the
4209 @code{gdbarch} struct has been set up. Unless a default has been set, the
4210 most recently appended sniffer will be tried first.
4212 The main frame unwinding sniffer (as set by
4213 @code{frame_unwind_append_sniffer)} returns a structure specifying
4214 a set of sniffing functions:
4216 @cindex @code{frame_unwind}
4220 enum frame_type type;
4221 frame_this_id_ftype *this_id;
4222 frame_prev_register_ftype *prev_register;
4223 const struct frame_data *unwind_data;
4224 frame_sniffer_ftype *sniffer;
4225 frame_prev_pc_ftype *prev_pc;
4226 frame_dealloc_cache_ftype *dealloc_cache;
4230 The @code{type} field indicates the type of frame this sniffer can
4231 handle: normal, dummy (@pxref{Functions Creating Dummy Frames, ,
4232 Functions Creating Dummy Frames}), signal handler or sentinel. Signal
4233 handlers sometimes have their own simplified stack structure for
4234 efficiency, so may need their own handlers.
4236 The @code{unwind_data} field holds additional information which may be
4237 relevant to particular types of frame. For example it may hold
4238 additional information for signal handler frames.
4240 The remaining fields define functions that yield different types of
4241 information when given a pointer to the NEXT stack frame. Not all
4242 functions need be provided. If an entry is @code{NULL}, the next sniffer will
4248 @code{this_id} determines the stack pointer and function (code
4249 entry point) for THIS stack frame.
4252 @code{prev_register} determines where the values of registers for
4253 the PREVIOUS stack frame are stored in THIS stack frame.
4256 @code{sniffer} takes a look at THIS frame's registers to
4257 determine if this is the appropriate unwinder.
4260 @code{prev_pc} determines the program counter for THIS
4261 frame. Only needed if the program counter is not an ordinary register
4262 (@pxref{Register Architecture Functions & Variables,
4263 , Functions and Variables Specifying the Register Architecture}).
4266 @code{dealloc_cache} frees any additional memory associated with
4267 the prologue cache for this frame (@pxref{Prologue Caches, , Prologue
4272 In general it is only the @code{this_id} and @code{prev_register}
4273 fields that need be defined for custom sniffers.
4275 The frame base sniffer is much simpler. It is a @code{@w{struct
4276 frame_base}}, which refers to the corresponding @code{frame_unwind}
4277 struct and whose fields refer to functions yielding various addresses
4280 @cindex @code{frame_base}
4284 const struct frame_unwind *unwind;
4285 frame_this_base_ftype *this_base;
4286 frame_this_locals_ftype *this_locals;
4287 frame_this_args_ftype *this_args;
4291 All the functions referred to take a pointer to the NEXT frame as
4292 argument. The function referred to by @code{this_base} returns the
4293 base address of THIS frame, the function referred to by
4294 @code{this_locals} returns the base address of local variables in THIS
4295 frame and the function referred to by @code{this_args} returns the
4296 base address of the function arguments in this frame.
4298 As described above, the base address of a frame is the address
4299 immediately before the start of the NEXT frame. For a falling
4300 stack, this is the lowest address in the frame and for a rising stack
4301 it is the highest address in the frame. For most architectures the
4302 same address is also the base address for local variables and
4303 arguments, in which case the same function can be used for all three
4304 entries@footnote{It is worth noting that if it cannot be determined in any
4305 other way (for example by there being a register with the name
4306 @code{"fp"}), then the result of the @code{this_base} function will be
4307 used as the value of the frame pointer variable @kbd{$fp} in
4308 @value{GDBN}. This is very often not correct (for example with the
4309 OpenRISC 1000, this value is the stack pointer, @kbd{$sp}). In this
4310 case a register (raw or pseudo) with the name @code{"fp"} should be
4311 defined. It will be used in preference as the value of @kbd{$fp}.}.
4313 @node Inferior Call Setup
4314 @section Inferior Call Setup
4315 @cindex calls to the inferior
4318 * About Dummy Frames::
4319 * Functions Creating Dummy Frames::
4322 @node About Dummy Frames
4323 @subsection About Dummy Frames
4324 @cindex dummy frames
4326 @value{GDBN} can call functions in the target code (for example by
4327 using the @kbd{call} or @kbd{print} commands). These functions may be
4328 breakpointed, and it is essential that if a function does hit a
4329 breakpoint, commands like @kbd{backtrace} work correctly.
4331 This is achieved by making the stack look as though the function had
4332 been called from the point where @value{GDBN} had previously stopped.
4333 This requires that @value{GDBN} can set up stack frames appropriate for
4334 such function calls.
4336 @node Functions Creating Dummy Frames
4337 @subsection Functions Creating Dummy Frames
4339 The following functions provide the functionality to set up such
4340 @dfn{dummy} stack frames.
4342 @deftypefn {Architecture Function} CORE_ADDR push_dummy_call (struct gdbarch *@var{gdbarch}, struct value *@var{function}, struct regcache *@var{regcache}, CORE_ADDR @var{bp_addr}, int @var{nargs}, struct value **@var{args}, CORE_ADDR @var{sp}, int @var{struct_return}, CORE_ADDR @var{struct_addr})
4344 This function sets up a dummy stack frame for the function about to be
4345 called. @code{push_dummy_call} is given the arguments to be passed
4346 and must copy them into registers or push them on to the stack as
4347 appropriate for the ABI.
4349 @var{function} is a pointer to the function
4350 that will be called and @var{regcache} the register cache from which
4351 values should be obtained. @var{bp_addr} is the address to which the
4352 function should return (which is breakpointed, so @value{GDBN} can
4353 regain control, hence the name). @var{nargs} is the number of
4354 arguments to pass and @var{args} an array containing the argument
4355 values. @var{struct_return} is non-zero (true) if the function returns
4356 a structure, and if so @var{struct_addr} is the address in which the
4357 structure should be returned.
4359 After calling this function, @value{GDBN} will pass control to the
4360 target at the address of the function, which will find the stack and
4361 registers set up just as expected.
4363 The default value of this function is @code{NULL} (undefined). If the
4364 function is not defined, then @value{GDBN} will not allow the user to
4365 call functions within the target being debugged.
4369 @deftypefn {Architecture Function} {struct frame_id} unwind_dummy_id (struct gdbarch *@var{gdbarch}, struct frame_info *@var{next_frame})
4371 This is the inverse of @code{push_dummy_call} which restores the stack
4372 pointer and program counter after a call to evaluate a function using
4373 a dummy stack frame. The result is a @code{@w{struct frame_id}}, which
4374 contains the value of the stack pointer and program counter to be
4377 The NEXT frame pointer is provided as argument,
4378 @var{next_frame}. THIS frame is the frame of the dummy function,
4379 which can be unwound, to yield the required stack pointer and program
4380 counter from the PREVIOUS frame.
4382 The default value is @code{NULL} (undefined). If @code{push_dummy_call} is
4383 defined, then this function should also be defined.
4387 @deftypefn {Architecture Function} CORE_ADDR push_dummy_code (struct gdbarch *@var{gdbarch}, CORE_ADDR @var{sp}, CORE_ADDR @var{funaddr}, struct value **@var{args}, int @var{nargs}, struct type *@var{value_type}, CORE_ADDR *@var{real_pc}, CORE_ADDR *@var{bp_addr}, struct regcache *@var{regcache})
4389 If this function is not defined (its default value is @code{NULL}), a dummy
4390 call will use the entry point of the currently loaded code on the
4391 target as its return address. A temporary breakpoint will be set
4392 there, so the location must be writable and have room for a
4395 It is possible that this default is not suitable. It might not be
4396 writable (in ROM possibly), or the ABI might require code to be
4397 executed on return from a call to unwind the stack before the
4398 breakpoint is encountered.
4400 If either of these is the case, then push_dummy_code should be defined
4401 to push an instruction sequence onto the end of the stack to which the
4402 dummy call should return.
4404 The arguments are essentially the same as those to
4405 @code{push_dummy_call}. However the function is provided with the
4406 type of the function result, @var{value_type}, @var{bp_addr} is used
4407 to return a value (the address at which the breakpoint instruction
4408 should be inserted) and @var{real pc} is used to specify the resume
4409 address when starting the call sequence. The function should return
4410 the updated innermost stack address.
4413 @emph{Note:} This does require that code in the stack can be executed.
4414 Some Harvard architectures may not allow this.
4419 @node Adding support for debugging core files
4420 @section Adding support for debugging core files
4423 The prerequisite for adding core file support in @value{GDBN} is to have
4424 core file support in BFD.
4426 Once BFD support is available, writing the apropriate
4427 @code{regset_from_core_section} architecture function should be all
4428 that is needed in order to add support for core files in @value{GDBN}.
4430 @node Defining Other Architecture Features
4431 @section Defining Other Architecture Features
4433 This section describes other functions and values in @code{gdbarch},
4434 together with some useful macros, that you can use to define the
4435 target architecture.
4439 @item CORE_ADDR gdbarch_addr_bits_remove (@var{gdbarch}, @var{addr})
4440 @findex gdbarch_addr_bits_remove
4441 If a raw machine instruction address includes any bits that are not
4442 really part of the address, then this function is used to zero those bits in
4443 @var{addr}. This is only used for addresses of instructions, and even then not
4446 For example, the two low-order bits of the PC on the Hewlett-Packard PA
4447 2.0 architecture contain the privilege level of the corresponding
4448 instruction. Since instructions must always be aligned on four-byte
4449 boundaries, the processor masks out these bits to generate the actual
4450 address of the instruction. @code{gdbarch_addr_bits_remove} would then for
4451 example look like that:
4453 arch_addr_bits_remove (CORE_ADDR addr)
4455 return (addr &= ~0x3);
4459 @item int address_class_name_to_type_flags (@var{gdbarch}, @var{name}, @var{type_flags_ptr})
4460 @findex address_class_name_to_type_flags
4461 If @var{name} is a valid address class qualifier name, set the @code{int}
4462 referenced by @var{type_flags_ptr} to the mask representing the qualifier
4463 and return 1. If @var{name} is not a valid address class qualifier name,
4466 The value for @var{type_flags_ptr} should be one of
4467 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
4468 possibly some combination of these values or'd together.
4469 @xref{Target Architecture Definition, , Address Classes}.
4471 @item int address_class_name_to_type_flags_p (@var{gdbarch})
4472 @findex address_class_name_to_type_flags_p
4473 Predicate which indicates whether @code{address_class_name_to_type_flags}
4476 @item int gdbarch_address_class_type_flags (@var{gdbarch}, @var{byte_size}, @var{dwarf2_addr_class})
4477 @findex gdbarch_address_class_type_flags
4478 Given a pointers byte size (as described by the debug information) and
4479 the possible @code{DW_AT_address_class} value, return the type flags
4480 used by @value{GDBN} to represent this address class. The value
4481 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
4482 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
4483 values or'd together.
4484 @xref{Target Architecture Definition, , Address Classes}.
4486 @item int gdbarch_address_class_type_flags_p (@var{gdbarch})
4487 @findex gdbarch_address_class_type_flags_p
4488 Predicate which indicates whether @code{gdbarch_address_class_type_flags_p} has
4491 @item const char *gdbarch_address_class_type_flags_to_name (@var{gdbarch}, @var{type_flags})
4492 @findex gdbarch_address_class_type_flags_to_name
4493 Return the name of the address class qualifier associated with the type
4494 flags given by @var{type_flags}.
4496 @item int gdbarch_address_class_type_flags_to_name_p (@var{gdbarch})
4497 @findex gdbarch_address_class_type_flags_to_name_p
4498 Predicate which indicates whether @code{gdbarch_address_class_type_flags_to_name} has been defined.
4499 @xref{Target Architecture Definition, , Address Classes}.
4501 @item void gdbarch_address_to_pointer (@var{gdbarch}, @var{type}, @var{buf}, @var{addr})
4502 @findex gdbarch_address_to_pointer
4503 Store in @var{buf} a pointer of type @var{type} representing the address
4504 @var{addr}, in the appropriate format for the current architecture.
4505 This function may safely assume that @var{type} is either a pointer or a
4506 C@t{++} reference type.
4507 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
4509 @item int gdbarch_believe_pcc_promotion (@var{gdbarch})
4510 @findex gdbarch_believe_pcc_promotion
4511 Used to notify if the compiler promotes a @code{short} or @code{char}
4512 parameter to an @code{int}, but still reports the parameter as its
4513 original type, rather than the promoted type.
4515 @item gdbarch_bits_big_endian (@var{gdbarch})
4516 @findex gdbarch_bits_big_endian
4517 This is used if the numbering of bits in the targets does @strong{not} match
4518 the endianism of the target byte order. A value of 1 means that the bits
4519 are numbered in a big-endian bit order, 0 means little-endian.
4521 @item set_gdbarch_bits_big_endian (@var{gdbarch}, @var{bits_big_endian})
4522 @findex set_gdbarch_bits_big_endian
4523 Calling set_gdbarch_bits_big_endian with a value of 1 indicates that the
4524 bits in the target are numbered in a big-endian bit order, 0 indicates
4529 This is the character array initializer for the bit pattern to put into
4530 memory where a breakpoint is set. Although it's common to use a trap
4531 instruction for a breakpoint, it's not required; for instance, the bit
4532 pattern could be an invalid instruction. The breakpoint must be no
4533 longer than the shortest instruction of the architecture.
4535 @code{BREAKPOINT} has been deprecated in favor of
4536 @code{gdbarch_breakpoint_from_pc}.
4538 @item BIG_BREAKPOINT
4539 @itemx LITTLE_BREAKPOINT
4540 @findex LITTLE_BREAKPOINT
4541 @findex BIG_BREAKPOINT
4542 Similar to BREAKPOINT, but used for bi-endian targets.
4544 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
4545 favor of @code{gdbarch_breakpoint_from_pc}.
4547 @item const gdb_byte *gdbarch_breakpoint_from_pc (@var{gdbarch}, @var{pcptr}, @var{lenptr})
4548 @findex gdbarch_breakpoint_from_pc
4549 @anchor{gdbarch_breakpoint_from_pc} Use the program counter to determine the
4550 contents and size of a breakpoint instruction. It returns a pointer to
4551 a static string of bytes that encode a breakpoint instruction, stores the
4552 length of the string to @code{*@var{lenptr}}, and adjusts the program
4553 counter (if necessary) to point to the actual memory location where the
4554 breakpoint should be inserted. May return @code{NULL} to indicate that
4555 software breakpoints are not supported.
4557 Although it is common to use a trap instruction for a breakpoint, it's
4558 not required; for instance, the bit pattern could be an invalid
4559 instruction. The breakpoint must be no longer than the shortest
4560 instruction of the architecture.
4562 Provided breakpoint bytes can be also used by @code{bp_loc_is_permanent} to
4563 detect permanent breakpoints. @code{gdbarch_breakpoint_from_pc} should return
4564 an unchanged memory copy if it was called for a location with permanent
4565 breakpoint as some architectures use breakpoint instructions containing
4566 arbitrary parameter value.
4568 Replaces all the other @var{BREAKPOINT} macros.
4570 @item int gdbarch_memory_insert_breakpoint (@var{gdbarch}, @var{bp_tgt})
4571 @itemx gdbarch_memory_remove_breakpoint (@var{gdbarch}, @var{bp_tgt})
4572 @findex gdbarch_memory_remove_breakpoint
4573 @findex gdbarch_memory_insert_breakpoint
4574 Insert or remove memory based breakpoints. Reasonable defaults
4575 (@code{default_memory_insert_breakpoint} and
4576 @code{default_memory_remove_breakpoint} respectively) have been
4577 provided so that it is not necessary to set these for most
4578 architectures. Architectures which may want to set
4579 @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
4580 conventional manner.
4582 It may also be desirable (from an efficiency standpoint) to define
4583 custom breakpoint insertion and removal routines if
4584 @code{gdbarch_breakpoint_from_pc} needs to read the target's memory for some
4587 @item CORE_ADDR gdbarch_adjust_breakpoint_address (@var{gdbarch}, @var{bpaddr})
4588 @findex gdbarch_adjust_breakpoint_address
4589 @cindex breakpoint address adjusted
4590 Given an address at which a breakpoint is desired, return a breakpoint
4591 address adjusted to account for architectural constraints on
4592 breakpoint placement. This method is not needed by most targets.
4594 The FR-V target (see @file{frv-tdep.c}) requires this method.
4595 The FR-V is a VLIW architecture in which a number of RISC-like
4596 instructions are grouped (packed) together into an aggregate
4597 instruction or instruction bundle. When the processor executes
4598 one of these bundles, the component instructions are executed
4601 In the course of optimization, the compiler may group instructions
4602 from distinct source statements into the same bundle. The line number
4603 information associated with one of the latter statements will likely
4604 refer to some instruction other than the first one in the bundle. So,
4605 if the user attempts to place a breakpoint on one of these latter
4606 statements, @value{GDBN} must be careful to @emph{not} place the break
4607 instruction on any instruction other than the first one in the bundle.
4608 (Remember though that the instructions within a bundle execute
4609 in parallel, so the @emph{first} instruction is the instruction
4610 at the lowest address and has nothing to do with execution order.)
4612 The FR-V's @code{gdbarch_adjust_breakpoint_address} method will adjust a
4613 breakpoint's address by scanning backwards for the beginning of
4614 the bundle, returning the address of the bundle.
4616 Since the adjustment of a breakpoint may significantly alter a user's
4617 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
4618 is initially set and each time that that breakpoint is hit.
4620 @item int gdbarch_call_dummy_location (@var{gdbarch})
4621 @findex gdbarch_call_dummy_location
4622 See the file @file{inferior.h}.
4624 This method has been replaced by @code{gdbarch_push_dummy_code}
4625 (@pxref{gdbarch_push_dummy_code}).
4627 @item int gdbarch_cannot_fetch_register (@var{gdbarch}, @var{regum})
4628 @findex gdbarch_cannot_fetch_register
4629 This function should return nonzero if @var{regno} cannot be fetched
4630 from an inferior process.
4632 @item int gdbarch_cannot_store_register (@var{gdbarch}, @var{regnum})
4633 @findex gdbarch_cannot_store_register
4634 This function should return nonzero if @var{regno} should not be
4635 written to the target. This is often the case for program counters,
4636 status words, and other special registers. This function returns 0 as
4637 default so that @value{GDBN} will assume that all registers may be written.
4639 @item int gdbarch_convert_register_p (@var{gdbarch}, @var{regnum}, struct type *@var{type})
4640 @findex gdbarch_convert_register_p
4641 Return non-zero if register @var{regnum} represents data values of type
4642 @var{type} in a non-standard form.
4643 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
4645 @item int gdbarch_fp0_regnum (@var{gdbarch})
4646 @findex gdbarch_fp0_regnum
4647 This function returns the number of the first floating point register,
4648 if the machine has such registers. Otherwise, it returns -1.
4650 @item CORE_ADDR gdbarch_decr_pc_after_break (@var{gdbarch})
4651 @findex gdbarch_decr_pc_after_break
4652 This function shall return the amount by which to decrement the PC after the
4653 program encounters a breakpoint. This is often the number of bytes in
4654 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
4656 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
4657 @findex DISABLE_UNSETTABLE_BREAK
4658 If defined, this should evaluate to 1 if @var{addr} is in a shared
4659 library in which breakpoints cannot be set and so should be disabled.
4661 @item int gdbarch_dwarf2_reg_to_regnum (@var{gdbarch}, @var{dwarf2_regnr})
4662 @findex gdbarch_dwarf2_reg_to_regnum
4663 Convert DWARF2 register number @var{dwarf2_regnr} into @value{GDBN} regnum.
4664 If not defined, no conversion will be performed.
4666 @item int gdbarch_ecoff_reg_to_regnum (@var{gdbarch}, @var{ecoff_regnr})
4667 @findex gdbarch_ecoff_reg_to_regnum
4668 Convert ECOFF register number @var{ecoff_regnr} into @value{GDBN} regnum. If
4669 not defined, no conversion will be performed.
4671 @item GCC_COMPILED_FLAG_SYMBOL
4672 @itemx GCC2_COMPILED_FLAG_SYMBOL
4673 @findex GCC2_COMPILED_FLAG_SYMBOL
4674 @findex GCC_COMPILED_FLAG_SYMBOL
4675 If defined, these are the names of the symbols that @value{GDBN} will
4676 look for to detect that GCC compiled the file. The default symbols
4677 are @code{gcc_compiled.} and @code{gcc2_compiled.},
4678 respectively. (Currently only defined for the Delta 68.)
4680 @item gdbarch_get_longjmp_target
4681 @findex gdbarch_get_longjmp_target
4682 This function determines the target PC address that @code{longjmp}
4683 will jump to, assuming that we have just stopped at a @code{longjmp}
4684 breakpoint. It takes a @code{CORE_ADDR *} as argument, and stores the
4685 target PC value through this pointer. It examines the current state
4686 of the machine as needed, typically by using a manually-determined
4687 offset into the @code{jmp_buf}. (While we might like to get the offset
4688 from the target's @file{jmpbuf.h}, that header file cannot be assumed
4689 to be available when building a cross-debugger.)
4691 @item DEPRECATED_IBM6000_TARGET
4692 @findex DEPRECATED_IBM6000_TARGET
4693 Shows that we are configured for an IBM RS/6000 system. This
4694 conditional should be eliminated (FIXME) and replaced by
4695 feature-specific macros. It was introduced in haste and we are
4696 repenting at leisure.
4698 @item I386_USE_GENERIC_WATCHPOINTS
4699 An x86-based target can define this to use the generic x86 watchpoint
4700 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4702 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{addr})
4703 @findex gdbarch_in_function_epilogue_p
4704 Returns non-zero if the given @var{addr} is in the epilogue of a function.
4705 The epilogue of a function is defined as the part of a function where
4706 the stack frame of the function already has been destroyed up to the
4707 final `return from function call' instruction.
4709 @item int gdbarch_in_solib_return_trampoline (@var{gdbarch}, @var{pc}, @var{name})
4710 @findex gdbarch_in_solib_return_trampoline
4711 Define this function to return nonzero if the program is stopped in the
4712 trampoline that returns from a shared library.
4714 @item target_so_ops.in_dynsym_resolve_code (@var{pc})
4715 @findex in_dynsym_resolve_code
4716 Define this to return nonzero if the program is stopped in the
4719 @item SKIP_SOLIB_RESOLVER (@var{pc})
4720 @findex SKIP_SOLIB_RESOLVER
4721 Define this to evaluate to the (nonzero) address at which execution
4722 should continue to get past the dynamic linker's symbol resolution
4723 function. A zero value indicates that it is not important or necessary
4724 to set a breakpoint to get through the dynamic linker and that single
4725 stepping will suffice.
4727 @item CORE_ADDR gdbarch_integer_to_address (@var{gdbarch}, @var{type}, @var{buf})
4728 @findex gdbarch_integer_to_address
4729 @cindex converting integers to addresses
4730 Define this when the architecture needs to handle non-pointer to address
4731 conversions specially. Converts that value to an address according to
4732 the current architectures conventions.
4734 @emph{Pragmatics: When the user copies a well defined expression from
4735 their source code and passes it, as a parameter, to @value{GDBN}'s
4736 @code{print} command, they should get the same value as would have been
4737 computed by the target program. Any deviation from this rule can cause
4738 major confusion and annoyance, and needs to be justified carefully. In
4739 other words, @value{GDBN} doesn't really have the freedom to do these
4740 conversions in clever and useful ways. It has, however, been pointed
4741 out that users aren't complaining about how @value{GDBN} casts integers
4742 to pointers; they are complaining that they can't take an address from a
4743 disassembly listing and give it to @code{x/i}. Adding an architecture
4744 method like @code{gdbarch_integer_to_address} certainly makes it possible for
4745 @value{GDBN} to ``get it right'' in all circumstances.}
4747 @xref{Target Architecture Definition, , Pointers Are Not Always
4750 @item CORE_ADDR gdbarch_pointer_to_address (@var{gdbarch}, @var{type}, @var{buf})
4751 @findex gdbarch_pointer_to_address
4752 Assume that @var{buf} holds a pointer of type @var{type}, in the
4753 appropriate format for the current architecture. Return the byte
4754 address the pointer refers to.
4755 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
4757 @item void gdbarch_register_to_value(@var{gdbarch}, @var{frame}, @var{regnum}, @var{type}, @var{fur})
4758 @findex gdbarch_register_to_value
4759 Convert the raw contents of register @var{regnum} into a value of type
4761 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
4763 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
4764 @findex REGISTER_CONVERT_TO_VIRTUAL
4765 Convert the value of register @var{reg} from its raw form to its virtual
4767 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
4769 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
4770 @findex REGISTER_CONVERT_TO_RAW
4771 Convert the value of register @var{reg} from its virtual form to its raw
4773 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
4775 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
4776 @findex regset_from_core_section
4777 Return the appropriate register set for a core file section with name
4778 @var{sect_name} and size @var{sect_size}.
4780 @item SOFTWARE_SINGLE_STEP_P()
4781 @findex SOFTWARE_SINGLE_STEP_P
4782 Define this as 1 if the target does not have a hardware single-step
4783 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
4785 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breakpoints_p})
4786 @findex SOFTWARE_SINGLE_STEP
4787 A function that inserts or removes (depending on
4788 @var{insert_breakpoints_p}) breakpoints at each possible destinations of
4789 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
4792 @item set_gdbarch_sofun_address_maybe_missing (@var{gdbarch}, @var{set})
4793 @findex set_gdbarch_sofun_address_maybe_missing
4794 Somebody clever observed that, the more actual addresses you have in the
4795 debug information, the more time the linker has to spend relocating
4796 them. So whenever there's some other way the debugger could find the
4797 address it needs, you should omit it from the debug info, to make
4800 Calling @code{set_gdbarch_sofun_address_maybe_missing} with a non-zero
4801 argument @var{set} indicates that a particular set of hacks of this sort
4802 are in use, affecting @code{N_SO} and @code{N_FUN} entries in stabs-format
4803 debugging information. @code{N_SO} stabs mark the beginning and ending
4804 addresses of compilation units in the text segment. @code{N_FUN} stabs
4805 mark the starts and ends of functions.
4807 In this case, @value{GDBN} assumes two things:
4811 @code{N_FUN} stabs have an address of zero. Instead of using those
4812 addresses, you should find the address where the function starts by
4813 taking the function name from the stab, and then looking that up in the
4814 minsyms (the linker/assembler symbol table). In other words, the stab
4815 has the name, and the linker/assembler symbol table is the only place
4816 that carries the address.
4819 @code{N_SO} stabs have an address of zero, too. You just look at the
4820 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab, and
4821 guess the starting and ending addresses of the compilation unit from them.
4824 @item int gdbarch_stabs_argument_has_addr (@var{gdbarch}, @var{type})
4825 @findex gdbarch_stabs_argument_has_addr
4826 @anchor{gdbarch_stabs_argument_has_addr} Define this function to return
4827 nonzero if a function argument of type @var{type} is passed by reference
4830 @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})
4831 @findex gdbarch_push_dummy_call
4832 @anchor{gdbarch_push_dummy_call} Define this to push the dummy frame's call to
4833 the inferior function onto the stack. In addition to pushing @var{nargs}, the
4834 code should push @var{struct_addr} (when @var{struct_return} is non-zero), and
4835 the return address (@var{bp_addr}).
4837 @var{function} is a pointer to a @code{struct value}; on architectures that use
4838 function descriptors, this contains the function descriptor value.
4840 Returns the updated top-of-stack pointer.
4842 @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})
4843 @findex gdbarch_push_dummy_code
4844 @anchor{gdbarch_push_dummy_code} Given a stack based call dummy, push the
4845 instruction sequence (including space for a breakpoint) to which the
4846 called function should return.
4848 Set @var{bp_addr} to the address at which the breakpoint instruction
4849 should be inserted, @var{real_pc} to the resume address when starting
4850 the call sequence, and return the updated inner-most stack address.
4852 By default, the stack is grown sufficient to hold a frame-aligned
4853 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
4854 reserved for that breakpoint, and @var{real_pc} set to @var{funaddr}.
4856 This method replaces @w{@code{gdbarch_call_dummy_location (@var{gdbarch})}}.
4858 @item int gdbarch_sdb_reg_to_regnum (@var{gdbarch}, @var{sdb_regnr})
4859 @findex gdbarch_sdb_reg_to_regnum
4860 Use this function to convert sdb register @var{sdb_regnr} into @value{GDBN}
4861 regnum. If not defined, no conversion will be done.
4863 @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})
4864 @findex gdbarch_return_value
4865 @anchor{gdbarch_return_value} Given a function with a return-value of
4866 type @var{rettype}, return which return-value convention that function
4869 @value{GDBN} currently recognizes two function return-value conventions:
4870 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
4871 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
4872 value is found in memory and the address of that memory location is
4873 passed in as the function's first parameter.
4875 If the register convention is being used, and @var{writebuf} is
4876 non-@code{NULL}, also copy the return-value in @var{writebuf} into
4879 If the register convention is being used, and @var{readbuf} is
4880 non-@code{NULL}, also copy the return value from @var{regcache} into
4881 @var{readbuf} (@var{regcache} contains a copy of the registers from the
4882 just returned function).
4884 @emph{Maintainer note: This method replaces separate predicate, extract,
4885 store methods. By having only one method, the logic needed to determine
4886 the return-value convention need only be implemented in one place. If
4887 @value{GDBN} were written in an @sc{oo} language, this method would
4888 instead return an object that knew how to perform the register
4889 return-value extract and store.}
4891 @emph{Maintainer note: This method does not take a @var{gcc_p}
4892 parameter, and such a parameter should not be added. If an architecture
4893 that requires per-compiler or per-function information be identified,
4894 then the replacement of @var{rettype} with @code{struct value}
4895 @var{function} should be pursued.}
4897 @emph{Maintainer note: The @var{regcache} parameter limits this methods
4898 to the inner most frame. While replacing @var{regcache} with a
4899 @code{struct frame_info} @var{frame} parameter would remove that
4900 limitation there has yet to be a demonstrated need for such a change.}
4902 @item void gdbarch_skip_permanent_breakpoint (@var{gdbarch}, @var{regcache})
4903 @findex gdbarch_skip_permanent_breakpoint
4904 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
4905 steps over a breakpoint by removing it, stepping one instruction, and
4906 re-inserting the breakpoint. However, permanent breakpoints are
4907 hardwired into the inferior, and can't be removed, so this strategy
4908 doesn't work. Calling @code{gdbarch_skip_permanent_breakpoint} adjusts the
4909 processor's state so that execution will resume just after the breakpoint.
4910 This function does the right thing even when the breakpoint is in the delay slot
4911 of a branch or jump.
4913 @item CORE_ADDR gdbarch_skip_trampoline_code (@var{gdbarch}, @var{frame}, @var{pc})
4914 @findex gdbarch_skip_trampoline_code
4915 If the target machine has trampoline code that sits between callers and
4916 the functions being called, then define this function to return a new PC
4917 that is at the start of the real function.
4919 @item int gdbarch_deprecated_fp_regnum (@var{gdbarch})
4920 @findex gdbarch_deprecated_fp_regnum
4921 If the frame pointer is in a register, use this function to return the
4922 number of that register.
4924 @item int gdbarch_stab_reg_to_regnum (@var{gdbarch}, @var{stab_regnr})
4925 @findex gdbarch_stab_reg_to_regnum
4926 Use this function to convert stab register @var{stab_regnr} into @value{GDBN}
4927 regnum. If not defined, no conversion will be done.
4929 @item SYMBOL_RELOADING_DEFAULT
4930 @findex SYMBOL_RELOADING_DEFAULT
4931 The default value of the ``symbol-reloading'' variable. (Never defined in
4934 @item TARGET_CHAR_BIT
4935 @findex TARGET_CHAR_BIT
4936 Number of bits in a char; defaults to 8.
4938 @item int gdbarch_char_signed (@var{gdbarch})
4939 @findex gdbarch_char_signed
4940 Non-zero if @code{char} is normally signed on this architecture; zero if
4941 it should be unsigned.
4943 The ISO C standard requires the compiler to treat @code{char} as
4944 equivalent to either @code{signed char} or @code{unsigned char}; any
4945 character in the standard execution set is supposed to be positive.
4946 Most compilers treat @code{char} as signed, but @code{char} is unsigned
4947 on the IBM S/390, RS6000, and PowerPC targets.
4949 @item int gdbarch_double_bit (@var{gdbarch})
4950 @findex gdbarch_double_bit
4951 Number of bits in a double float; defaults to @w{@code{8 * TARGET_CHAR_BIT}}.
4953 @item int gdbarch_float_bit (@var{gdbarch})
4954 @findex gdbarch_float_bit
4955 Number of bits in a float; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4957 @item int gdbarch_int_bit (@var{gdbarch})
4958 @findex gdbarch_int_bit
4959 Number of bits in an integer; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4961 @item int gdbarch_long_bit (@var{gdbarch})
4962 @findex gdbarch_long_bit
4963 Number of bits in a long integer; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4965 @item int gdbarch_long_double_bit (@var{gdbarch})
4966 @findex gdbarch_long_double_bit
4967 Number of bits in a long double float;
4968 defaults to @w{@code{2 * gdbarch_double_bit (@var{gdbarch})}}.
4970 @item int gdbarch_long_long_bit (@var{gdbarch})
4971 @findex gdbarch_long_long_bit
4972 Number of bits in a long long integer; defaults to
4973 @w{@code{2 * gdbarch_long_bit (@var{gdbarch})}}.
4975 @item int gdbarch_ptr_bit (@var{gdbarch})
4976 @findex gdbarch_ptr_bit
4977 Number of bits in a pointer; defaults to
4978 @w{@code{gdbarch_int_bit (@var{gdbarch})}}.
4980 @item int gdbarch_short_bit (@var{gdbarch})
4981 @findex gdbarch_short_bit
4982 Number of bits in a short integer; defaults to @w{@code{2 * TARGET_CHAR_BIT}}.
4984 @item void gdbarch_virtual_frame_pointer (@var{gdbarch}, @var{pc}, @var{frame_regnum}, @var{frame_offset})
4985 @findex gdbarch_virtual_frame_pointer
4986 Returns a @code{(@var{register}, @var{offset})} pair representing the virtual
4987 frame pointer in use at the code address @var{pc}. If virtual frame
4988 pointers are not used, a default definition simply returns
4989 @code{gdbarch_deprecated_fp_regnum} (or @code{gdbarch_sp_regnum}, if
4990 no frame pointer is defined), with an offset of zero.
4992 @c need to explain virtual frame pointers, they are recorded in agent
4993 @c expressions for tracepoints
4995 @item TARGET_HAS_HARDWARE_WATCHPOINTS
4996 If non-zero, the target has support for hardware-assisted
4997 watchpoints. @xref{Algorithms, watchpoints}, for more details and
4998 other related macros.
5000 @item int gdbarch_print_insn (@var{gdbarch}, @var{vma}, @var{info})
5001 @findex gdbarch_print_insn
5002 This is the function used by @value{GDBN} to print an assembly
5003 instruction. It prints the instruction at address @var{vma} in
5004 debugged memory and returns the length of the instruction, in bytes.
5005 This usually points to a function in the @code{opcodes} library
5006 (@pxref{Support Libraries, ,Opcodes}). @var{info} is a structure (of
5007 type @code{disassemble_info}) defined in the header file
5008 @file{include/dis-asm.h}, and used to pass information to the
5009 instruction decoding routine.
5011 @item frame_id gdbarch_dummy_id (@var{gdbarch}, @var{frame})
5012 @findex gdbarch_dummy_id
5013 @anchor{gdbarch_dummy_id} Given @var{frame} return a @w{@code{struct
5014 frame_id}} that uniquely identifies an inferior function call's dummy
5015 frame. The value returned must match the dummy frame stack value
5016 previously saved by @code{call_function_by_hand}.
5018 @item void gdbarch_value_to_register (@var{gdbarch}, @var{frame}, @var{type}, @var{buf})
5019 @findex gdbarch_value_to_register
5020 Convert a value of type @var{type} into the raw contents of a register.
5021 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
5025 Motorola M68K target conditionals.
5029 Define this to be the 4-bit location of the breakpoint trap vector. If
5030 not defined, it will default to @code{0xf}.
5032 @item REMOTE_BPT_VECTOR
5033 Defaults to @code{1}.
5037 @node Adding a New Target
5038 @section Adding a New Target
5040 @cindex adding a target
5041 The following files add a target to @value{GDBN}:
5044 @cindex target dependent files
5046 @item gdb/@var{ttt}-tdep.c
5047 Contains any miscellaneous code required for this target machine. On
5048 some machines it doesn't exist at all.
5050 @item gdb/@var{arch}-tdep.c
5051 @itemx gdb/@var{arch}-tdep.h
5052 This is required to describe the basic layout of the target machine's
5053 processor chip (registers, stack, etc.). It can be shared among many
5054 targets that use the same processor architecture.
5058 (Target header files such as
5059 @file{gdb/config/@var{arch}/tm-@var{ttt}.h},
5060 @file{gdb/config/@var{arch}/tm-@var{arch}.h}, and
5061 @file{config/tm-@var{os}.h} are no longer used.)
5063 @findex _initialize_@var{arch}_tdep
5064 A @value{GDBN} description for a new architecture, arch is created by
5065 defining a global function @code{_initialize_@var{arch}_tdep}, by
5066 convention in the source file @file{@var{arch}-tdep.c}. For
5067 example, in the case of the OpenRISC 1000, this function is called
5068 @code{_initialize_or1k_tdep} and is found in the file
5071 The object file resulting from compiling this source file, which will
5072 contain the implementation of the
5073 @code{_initialize_@var{arch}_tdep} function is specified in the
5074 @value{GDBN} @file{configure.tgt} file, which includes a large case
5075 statement pattern matching against the @code{--target} option of the
5076 @kbd{configure} script.
5079 @emph{Note:} If the architecture requires multiple source files, the
5080 corresponding binaries should be included in
5081 @file{configure.tgt}. However if there are header files, the
5082 dependencies on these will not be picked up from the entries in
5083 @file{configure.tgt}. The @file{Makefile.in} file will need extending to
5084 show these dependencies.
5087 @findex gdbarch_register
5088 A new struct gdbarch, defining the new architecture, is created within
5089 the @code{_initialize_@var{arch}_tdep} function by calling
5090 @code{gdbarch_register}:
5093 void gdbarch_register (enum bfd_architecture architecture,
5094 gdbarch_init_ftype *init_func,
5095 gdbarch_dump_tdep_ftype *tdep_dump_func);
5098 This function has been described fully in an earlier
5099 section. @xref{How an Architecture is Represented, , How an
5100 Architecture is Represented}.
5102 The new @code{@w{struct gdbarch}} should contain implementations of
5103 the necessary functions (described in the previous sections) to
5104 describe the basic layout of the target machine's processor chip
5105 (registers, stack, etc.). It can be shared among many targets that use
5106 the same processor architecture.
5108 @node Target Descriptions
5109 @chapter Target Descriptions
5110 @cindex target descriptions
5112 The target architecture definition (@pxref{Target Architecture Definition})
5113 contains @value{GDBN}'s hard-coded knowledge about an architecture. For
5114 some platforms, it is handy to have more flexible knowledge about a specific
5115 instance of the architecture---for instance, a processor or development board.
5116 @dfn{Target descriptions} provide a mechanism for the user to tell @value{GDBN}
5117 more about what their target supports, or for the target to tell @value{GDBN}
5120 For details on writing, automatically supplying, and manually selecting
5121 target descriptions, see @ref{Target Descriptions, , , gdb,
5122 Debugging with @value{GDBN}}. This section will cover some related
5123 topics about the @value{GDBN} internals.
5126 * Target Descriptions Implementation::
5127 * Adding Target Described Register Support::
5130 @node Target Descriptions Implementation
5131 @section Target Descriptions Implementation
5132 @cindex target descriptions, implementation
5134 Before @value{GDBN} connects to a new target, or runs a new program on
5135 an existing target, it discards any existing target description and
5136 reverts to a default gdbarch. Then, after connecting, it looks for a
5137 new target description by calling @code{target_find_description}.
5139 A description may come from a user specified file (XML), the remote
5140 @samp{qXfer:features:read} packet (also XML), or from any custom
5141 @code{to_read_description} routine in the target vector. For instance,
5142 the remote target supports guessing whether a MIPS target is 32-bit or
5143 64-bit based on the size of the @samp{g} packet.
5145 If any target description is found, @value{GDBN} creates a new gdbarch
5146 incorporating the description by calling @code{gdbarch_update_p}. Any
5147 @samp{<architecture>} element is handled first, to determine which
5148 architecture's gdbarch initialization routine is called to create the
5149 new architecture. Then the initialization routine is called, and has
5150 a chance to adjust the constructed architecture based on the contents
5151 of the target description. For instance, it can recognize any
5152 properties set by a @code{to_read_description} routine. Also
5153 see @ref{Adding Target Described Register Support}.
5155 @node Adding Target Described Register Support
5156 @section Adding Target Described Register Support
5157 @cindex target descriptions, adding register support
5159 Target descriptions can report additional registers specific to an
5160 instance of the target. But it takes a little work in the architecture
5161 specific routines to support this.
5163 A target description must either have no registers or a complete
5164 set---this avoids complexity in trying to merge standard registers
5165 with the target defined registers. It is the architecture's
5166 responsibility to validate that a description with registers has
5167 everything it needs. To keep architecture code simple, the same
5168 mechanism is used to assign fixed internal register numbers to
5171 If @code{tdesc_has_registers} returns 1, the description contains
5172 registers. The architecture's @code{gdbarch_init} routine should:
5177 Call @code{tdesc_data_alloc} to allocate storage, early, before
5178 searching for a matching gdbarch or allocating a new one.
5181 Use @code{tdesc_find_feature} to locate standard features by name.
5184 Use @code{tdesc_numbered_register} and @code{tdesc_numbered_register_choices}
5185 to locate the expected registers in the standard features.
5188 Return @code{NULL} if a required feature is missing, or if any standard
5189 feature is missing expected registers. This will produce a warning that
5190 the description was incomplete.
5193 Free the allocated data before returning, unless @code{tdesc_use_registers}
5197 Call @code{set_gdbarch_num_regs} as usual, with a number higher than any
5198 fixed number passed to @code{tdesc_numbered_register}.
5201 Call @code{tdesc_use_registers} after creating a new gdbarch, before
5206 After @code{tdesc_use_registers} has been called, the architecture's
5207 @code{register_name}, @code{register_type}, and @code{register_reggroup_p}
5208 routines will not be called; that information will be taken from
5209 the target description. @code{num_regs} may be increased to account
5210 for any additional registers in the description.
5212 Pseudo-registers require some extra care:
5217 Using @code{tdesc_numbered_register} allows the architecture to give
5218 constant register numbers to standard architectural registers, e.g.@:
5219 as an @code{enum} in @file{@var{arch}-tdep.h}. But because
5220 pseudo-registers are always numbered above @code{num_regs},
5221 which may be increased by the description, constant numbers
5222 can not be used for pseudos. They must be numbered relative to
5223 @code{num_regs} instead.
5226 The description will not describe pseudo-registers, so the
5227 architecture must call @code{set_tdesc_pseudo_register_name},
5228 @code{set_tdesc_pseudo_register_type}, and
5229 @code{set_tdesc_pseudo_register_reggroup_p} to supply routines
5230 describing pseudo registers. These routines will be passed
5231 internal register numbers, so the same routines used for the
5232 gdbarch equivalents are usually suitable.
5237 @node Target Vector Definition
5239 @chapter Target Vector Definition
5240 @cindex target vector
5242 The target vector defines the interface between @value{GDBN}'s
5243 abstract handling of target systems, and the nitty-gritty code that
5244 actually exercises control over a process or a serial port.
5245 @value{GDBN} includes some 30-40 different target vectors; however,
5246 each configuration of @value{GDBN} includes only a few of them.
5249 * Managing Execution State::
5250 * Existing Targets::
5253 @node Managing Execution State
5254 @section Managing Execution State
5255 @cindex execution state
5257 A target vector can be completely inactive (not pushed on the target
5258 stack), active but not running (pushed, but not connected to a fully
5259 manifested inferior), or completely active (pushed, with an accessible
5260 inferior). Most targets are only completely inactive or completely
5261 active, but some support persistent connections to a target even
5262 when the target has exited or not yet started.
5264 For example, connecting to the simulator using @code{target sim} does
5265 not create a running program. Neither registers nor memory are
5266 accessible until @code{run}. Similarly, after @code{kill}, the
5267 program can not continue executing. But in both cases @value{GDBN}
5268 remains connected to the simulator, and target-specific commands
5269 are directed to the simulator.
5271 A target which only supports complete activation should push itself
5272 onto the stack in its @code{to_open} routine (by calling
5273 @code{push_target}), and unpush itself from the stack in its
5274 @code{to_mourn_inferior} routine (by calling @code{unpush_target}).
5276 A target which supports both partial and complete activation should
5277 still call @code{push_target} in @code{to_open}, but not call
5278 @code{unpush_target} in @code{to_mourn_inferior}. Instead, it should
5279 call either @code{target_mark_running} or @code{target_mark_exited}
5280 in its @code{to_open}, depending on whether the target is fully active
5281 after connection. It should also call @code{target_mark_running} any
5282 time the inferior becomes fully active (e.g.@: in
5283 @code{to_create_inferior} and @code{to_attach}), and
5284 @code{target_mark_exited} when the inferior becomes inactive (in
5285 @code{to_mourn_inferior}). The target should also make sure to call
5286 @code{target_mourn_inferior} from its @code{to_kill}, to return the
5287 target to inactive state.
5289 @node Existing Targets
5290 @section Existing Targets
5293 @subsection File Targets
5295 Both executables and core files have target vectors.
5297 @subsection Standard Protocol and Remote Stubs
5299 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code that
5300 runs in the target system. @value{GDBN} provides several sample
5301 @dfn{stubs} that can be integrated into target programs or operating
5302 systems for this purpose; they are named @file{@var{cpu}-stub.c}. Many
5303 operating systems, embedded targets, emulators, and simulators already
5304 have a @value{GDBN} stub built into them, and maintenance of the remote
5305 protocol must be careful to preserve compatibility.
5307 The @value{GDBN} user's manual describes how to put such a stub into
5308 your target code. What follows is a discussion of integrating the
5309 SPARC stub into a complicated operating system (rather than a simple
5310 program), by Stu Grossman, the author of this stub.
5312 The trap handling code in the stub assumes the following upon entry to
5317 %l1 and %l2 contain pc and npc respectively at the time of the trap;
5323 you are in the correct trap window.
5326 As long as your trap handler can guarantee those conditions, then there
5327 is no reason why you shouldn't be able to ``share'' traps with the stub.
5328 The stub has no requirement that it be jumped to directly from the
5329 hardware trap vector. That is why it calls @code{exceptionHandler()},
5330 which is provided by the external environment. For instance, this could
5331 set up the hardware traps to actually execute code which calls the stub
5332 first, and then transfers to its own trap handler.
5334 For the most point, there probably won't be much of an issue with
5335 ``sharing'' traps, as the traps we use are usually not used by the kernel,
5336 and often indicate unrecoverable error conditions. Anyway, this is all
5337 controlled by a table, and is trivial to modify. The most important
5338 trap for us is for @code{ta 1}. Without that, we can't single step or
5339 do breakpoints. Everything else is unnecessary for the proper operation
5340 of the debugger/stub.
5342 From reading the stub, it's probably not obvious how breakpoints work.
5343 They are simply done by deposit/examine operations from @value{GDBN}.
5345 @subsection ROM Monitor Interface
5347 @subsection Custom Protocols
5349 @subsection Transport Layer
5351 @subsection Builtin Simulator
5354 @node Native Debugging
5356 @chapter Native Debugging
5357 @cindex native debugging
5359 Several files control @value{GDBN}'s configuration for native support:
5363 @item gdb/config/@var{arch}/@var{xyz}.mh
5364 Specifies Makefile fragments needed by a @emph{native} configuration on
5365 machine @var{xyz}. In particular, this lists the required
5366 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
5367 Also specifies the header file which describes native support on
5368 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
5369 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
5370 @samp{NAT_CDEPS}, @samp{NAT_GENERATED_FILES}, etc.; see @file{Makefile.in}.
5372 @emph{Maintainer's note: The @file{.mh} suffix is because this file
5373 originally contained @file{Makefile} fragments for hosting @value{GDBN}
5374 on machine @var{xyz}. While the file is no longer used for this
5375 purpose, the @file{.mh} suffix remains. Perhaps someone will
5376 eventually rename these fragments so that they have a @file{.mn}
5379 @item gdb/config/@var{arch}/nm-@var{xyz}.h
5380 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
5381 macro definitions describing the native system environment, such as
5382 child process control and core file support.
5384 @item gdb/@var{xyz}-nat.c
5385 Contains any miscellaneous C code required for this native support of
5386 this machine. On some machines it doesn't exist at all.
5389 There are some ``generic'' versions of routines that can be used by
5390 various systems. These can be customized in various ways by macros
5391 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
5392 the @var{xyz} host, you can just include the generic file's name (with
5393 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
5395 Otherwise, if your machine needs custom support routines, you will need
5396 to write routines that perform the same functions as the generic file.
5397 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
5398 into @code{NATDEPFILES}.
5402 This contains the @emph{target_ops vector} that supports Unix child
5403 processes on systems which use ptrace and wait to control the child.
5406 This contains the @emph{target_ops vector} that supports Unix child
5407 processes on systems which use /proc to control the child.
5410 This does the low-level grunge that uses Unix system calls to do a ``fork
5411 and exec'' to start up a child process.
5414 This is the low level interface to inferior processes for systems using
5415 the Unix @code{ptrace} call in a vanilla way.
5424 @section shared libraries
5426 @section Native Conditionals
5427 @cindex native conditionals
5429 When @value{GDBN} is configured and compiled, various macros are
5430 defined or left undefined, to control compilation when the host and
5431 target systems are the same. These macros should be defined (or left
5432 undefined) in @file{nm-@var{system}.h}.
5436 @item I386_USE_GENERIC_WATCHPOINTS
5437 An x86-based machine can define this to use the generic x86 watchpoint
5438 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
5440 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
5442 Define this to expand into an expression that will cause the symbols in
5443 @var{filename} to be added to @value{GDBN}'s symbol table. If
5444 @var{readsyms} is zero symbols are not read but any necessary low level
5445 processing for @var{filename} is still done.
5447 @item SOLIB_CREATE_INFERIOR_HOOK
5448 @findex SOLIB_CREATE_INFERIOR_HOOK
5449 Define this to expand into any shared-library-relocation code that you
5450 want to be run just after the child process has been forked.
5452 @item START_INFERIOR_TRAPS_EXPECTED
5453 @findex START_INFERIOR_TRAPS_EXPECTED
5454 When starting an inferior, @value{GDBN} normally expects to trap
5456 the shell execs, and once when the program itself execs. If the actual
5457 number of traps is something other than 2, then define this macro to
5458 expand into the number expected.
5462 @node Support Libraries
5464 @chapter Support Libraries
5469 BFD provides support for @value{GDBN} in several ways:
5472 @item identifying executable and core files
5473 BFD will identify a variety of file types, including a.out, coff, and
5474 several variants thereof, as well as several kinds of core files.
5476 @item access to sections of files
5477 BFD parses the file headers to determine the names, virtual addresses,
5478 sizes, and file locations of all the various named sections in files
5479 (such as the text section or the data section). @value{GDBN} simply
5480 calls BFD to read or write section @var{x} at byte offset @var{y} for
5483 @item specialized core file support
5484 BFD provides routines to determine the failing command name stored in a
5485 core file, the signal with which the program failed, and whether a core
5486 file matches (i.e.@: could be a core dump of) a particular executable
5489 @item locating the symbol information
5490 @value{GDBN} uses an internal interface of BFD to determine where to find the
5491 symbol information in an executable file or symbol-file. @value{GDBN} itself
5492 handles the reading of symbols, since BFD does not ``understand'' debug
5493 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
5498 @cindex opcodes library
5500 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
5501 library because it's also used in binutils, for @file{objdump}).
5504 @cindex readline library
5505 The @code{readline} library provides a set of functions for use by applications
5506 that allow users to edit command lines as they are typed in.
5509 @cindex @code{libiberty} library
5511 The @code{libiberty} library provides a set of functions and features
5512 that integrate and improve on functionality found in modern operating
5513 systems. Broadly speaking, such features can be divided into three
5514 groups: supplemental functions (functions that may be missing in some
5515 environments and operating systems), replacement functions (providing
5516 a uniform and easier to use interface for commonly used standard
5517 functions), and extensions (which provide additional functionality
5518 beyond standard functions).
5520 @value{GDBN} uses various features provided by the @code{libiberty}
5521 library, for instance the C@t{++} demangler, the @acronym{IEEE}
5522 floating format support functions, the input options parser
5523 @samp{getopt}, the @samp{obstack} extension, and other functions.
5525 @subsection @code{obstacks} in @value{GDBN}
5526 @cindex @code{obstacks}
5528 The obstack mechanism provides a convenient way to allocate and free
5529 chunks of memory. Each obstack is a pool of memory that is managed
5530 like a stack. Objects (of any nature, size and alignment) are
5531 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
5532 @code{libiberty}'s documentation for a more detailed explanation of
5535 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
5536 object files. There is an obstack associated with each internal
5537 representation of an object file. Lots of things get allocated on
5538 these @code{obstacks}: dictionary entries, blocks, blockvectors,
5539 symbols, minimal symbols, types, vectors of fundamental types, class
5540 fields of types, object files section lists, object files section
5541 offset lists, line tables, symbol tables, partial symbol tables,
5542 string tables, symbol table private data, macros tables, debug
5543 information sections and entries, import and export lists (som),
5544 unwind information (hppa), dwarf2 location expressions data. Plus
5545 various strings such as directory names strings, debug format strings,
5548 An essential and convenient property of all data on @code{obstacks} is
5549 that memory for it gets allocated (with @code{obstack_alloc}) at
5550 various times during a debugging session, but it is released all at
5551 once using the @code{obstack_free} function. The @code{obstack_free}
5552 function takes a pointer to where in the stack it must start the
5553 deletion from (much like the cleanup chains have a pointer to where to
5554 start the cleanups). Because of the stack like structure of the
5555 @code{obstacks}, this allows to free only a top portion of the
5556 obstack. There are a few instances in @value{GDBN} where such thing
5557 happens. Calls to @code{obstack_free} are done after some local data
5558 is allocated to the obstack. Only the local data is deleted from the
5559 obstack. Of course this assumes that nothing between the
5560 @code{obstack_alloc} and the @code{obstack_free} allocates anything
5561 else on the same obstack. For this reason it is best and safest to
5562 use temporary @code{obstacks}.
5564 Releasing the whole obstack is also not safe per se. It is safe only
5565 under the condition that we know the @code{obstacks} memory is no
5566 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
5567 when we get rid of the whole objfile(s), for instance upon reading a
5571 @cindex regular expressions library
5582 @item SIGN_EXTEND_CHAR
5584 @item SWITCH_ENUM_BUG
5593 @section Array Containers
5594 @cindex Array Containers
5597 Often it is necessary to manipulate a dynamic array of a set of
5598 objects. C forces some bookkeeping on this, which can get cumbersome
5599 and repetitive. The @file{vec.h} file contains macros for defining
5600 and using a typesafe vector type. The functions defined will be
5601 inlined when compiling, and so the abstraction cost should be zero.
5602 Domain checks are added to detect programming errors.
5604 An example use would be an array of symbols or section information.
5605 The array can be grown as symbols are read in (or preallocated), and
5606 the accessor macros provided keep care of all the necessary
5607 bookkeeping. Because the arrays are type safe, there is no danger of
5608 accidentally mixing up the contents. Think of these as C++ templates,
5609 but implemented in C.
5611 Because of the different behavior of structure objects, scalar objects
5612 and of pointers, there are three flavors of vector, one for each of
5613 these variants. Both the structure object and pointer variants pass
5614 pointers to objects around --- in the former case the pointers are
5615 stored into the vector and in the latter case the pointers are
5616 dereferenced and the objects copied into the vector. The scalar
5617 object variant is suitable for @code{int}-like objects, and the vector
5618 elements are returned by value.
5620 There are both @code{index} and @code{iterate} accessors. The iterator
5621 returns a boolean iteration condition and updates the iteration
5622 variable passed by reference. Because the iterator will be inlined,
5623 the address-of can be optimized away.
5625 The vectors are implemented using the trailing array idiom, thus they
5626 are not resizeable without changing the address of the vector object
5627 itself. This means you cannot have variables or fields of vector type
5628 --- always use a pointer to a vector. The one exception is the final
5629 field of a structure, which could be a vector type. You will have to
5630 use the @code{embedded_size} & @code{embedded_init} calls to create
5631 such objects, and they will probably not be resizeable (so don't use
5632 the @dfn{safe} allocation variants). The trailing array idiom is used
5633 (rather than a pointer to an array of data), because, if we allow
5634 @code{NULL} to also represent an empty vector, empty vectors occupy
5635 minimal space in the structure containing them.
5637 Each operation that increases the number of active elements is
5638 available in @dfn{quick} and @dfn{safe} variants. The former presumes
5639 that there is sufficient allocated space for the operation to succeed
5640 (it dies if there is not). The latter will reallocate the vector, if
5641 needed. Reallocation causes an exponential increase in vector size.
5642 If you know you will be adding N elements, it would be more efficient
5643 to use the reserve operation before adding the elements with the
5644 @dfn{quick} operation. This will ensure there are at least as many
5645 elements as you ask for, it will exponentially increase if there are
5646 too few spare slots. If you want reserve a specific number of slots,
5647 but do not want the exponential increase (for instance, you know this
5648 is the last allocation), use a negative number for reservation. You
5649 can also create a vector of a specific size from the get go.
5651 You should prefer the push and pop operations, as they append and
5652 remove from the end of the vector. If you need to remove several items
5653 in one go, use the truncate operation. The insert and remove
5654 operations allow you to change elements in the middle of the vector.
5655 There are two remove operations, one which preserves the element
5656 ordering @code{ordered_remove}, and one which does not
5657 @code{unordered_remove}. The latter function copies the end element
5658 into the removed slot, rather than invoke a memmove operation. The
5659 @code{lower_bound} function will determine where to place an item in
5660 the array using insert that will maintain sorted order.
5662 If you need to directly manipulate a vector, then the @code{address}
5663 accessor will return the address of the start of the vector. Also the
5664 @code{space} predicate will tell you whether there is spare capacity in the
5665 vector. You will not normally need to use these two functions.
5667 Vector types are defined using a
5668 @code{DEF_VEC_@{O,P,I@}(@var{typename})} macro. Variables of vector
5669 type are declared using a @code{VEC(@var{typename})} macro. The
5670 characters @code{O}, @code{P} and @code{I} indicate whether
5671 @var{typename} is an object (@code{O}), pointer (@code{P}) or integral
5672 (@code{I}) type. Be careful to pick the correct one, as you'll get an
5673 awkward and inefficient API if you use the wrong one. There is a
5674 check, which results in a compile-time warning, for the @code{P} and
5675 @code{I} versions, but there is no check for the @code{O} versions, as
5676 that is not possible in plain C.
5678 An example of their use would be,
5681 DEF_VEC_P(tree); // non-managed tree vector.
5684 VEC(tree) *v; // A (pointer to) a vector of tree pointers.
5687 struct my_struct *s;
5689 if (VEC_length(tree, s->v)) @{ we have some contents @}
5690 VEC_safe_push(tree, s->v, decl); // append some decl onto the end
5691 for (ix = 0; VEC_iterate(tree, s->v, ix, elt); ix++)
5692 @{ do something with elt @}
5696 The @file{vec.h} file provides details on how to invoke the various
5697 accessors provided. They are enumerated here:
5701 Return the number of items in the array,
5704 Return true if the array has no elements.
5708 Return the last or arbitrary item in the array.
5711 Access an array element and indicate whether the array has been
5716 Create and destroy an array.
5718 @item VEC_embedded_size
5719 @itemx VEC_embedded_init
5720 Helpers for embedding an array as the final element of another struct.
5726 Return the amount of free space in an array.
5729 Ensure a certain amount of free space.
5731 @item VEC_quick_push
5732 @itemx VEC_safe_push
5733 Append to an array, either assuming the space is available, or making
5737 Remove the last item from an array.
5740 Remove several items from the end of an array.
5743 Add several items to the end of an array.
5746 Overwrite an item in the array.
5748 @item VEC_quick_insert
5749 @itemx VEC_safe_insert
5750 Insert an item into the middle of the array. Either the space must
5751 already exist, or the space is created.
5753 @item VEC_ordered_remove
5754 @itemx VEC_unordered_remove
5755 Remove an item from the array, preserving order or not.
5757 @item VEC_block_remove
5758 Remove a set of items from the array.
5761 Provide the address of the first element.
5763 @item VEC_lower_bound
5764 Binary search the array.
5774 This chapter covers topics that are lower-level than the major
5775 algorithms of @value{GDBN}.
5780 Cleanups are a structured way to deal with things that need to be done
5783 When your code does something (e.g., @code{xmalloc} some memory, or
5784 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
5785 the memory or @code{close} the file), it can make a cleanup. The
5786 cleanup will be done at some future point: when the command is finished
5787 and control returns to the top level; when an error occurs and the stack
5788 is unwound; or when your code decides it's time to explicitly perform
5789 cleanups. Alternatively you can elect to discard the cleanups you
5795 @item struct cleanup *@var{old_chain};
5796 Declare a variable which will hold a cleanup chain handle.
5798 @findex make_cleanup
5799 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
5800 Make a cleanup which will cause @var{function} to be called with
5801 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
5802 handle that can later be passed to @code{do_cleanups} or
5803 @code{discard_cleanups}. Unless you are going to call
5804 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
5805 from @code{make_cleanup}.
5808 @item do_cleanups (@var{old_chain});
5809 Do all cleanups added to the chain since the corresponding
5810 @code{make_cleanup} call was made.
5812 @findex discard_cleanups
5813 @item discard_cleanups (@var{old_chain});
5814 Same as @code{do_cleanups} except that it just removes the cleanups from
5815 the chain and does not call the specified functions.
5818 Cleanups are implemented as a chain. The handle returned by
5819 @code{make_cleanups} includes the cleanup passed to the call and any
5820 later cleanups appended to the chain (but not yet discarded or
5824 make_cleanup (a, 0);
5826 struct cleanup *old = make_cleanup (b, 0);
5834 will call @code{c()} and @code{b()} but will not call @code{a()}. The
5835 cleanup that calls @code{a()} will remain in the cleanup chain, and will
5836 be done later unless otherwise discarded.@refill
5838 Your function should explicitly do or discard the cleanups it creates.
5839 Failing to do this leads to non-deterministic behavior since the caller
5840 will arbitrarily do or discard your functions cleanups. This need leads
5841 to two common cleanup styles.
5843 The first style is try/finally. Before it exits, your code-block calls
5844 @code{do_cleanups} with the old cleanup chain and thus ensures that your
5845 code-block's cleanups are always performed. For instance, the following
5846 code-segment avoids a memory leak problem (even when @code{error} is
5847 called and a forced stack unwind occurs) by ensuring that the
5848 @code{xfree} will always be called:
5851 struct cleanup *old = make_cleanup (null_cleanup, 0);
5852 data = xmalloc (sizeof blah);
5853 make_cleanup (xfree, data);
5858 The second style is try/except. Before it exits, your code-block calls
5859 @code{discard_cleanups} with the old cleanup chain and thus ensures that
5860 any created cleanups are not performed. For instance, the following
5861 code segment, ensures that the file will be closed but only if there is
5865 FILE *file = fopen ("afile", "r");
5866 struct cleanup *old = make_cleanup (close_file, file);
5868 discard_cleanups (old);
5872 Some functions, e.g., @code{fputs_filtered()} or @code{error()}, specify
5873 that they ``should not be called when cleanups are not in place''. This
5874 means that any actions you need to reverse in the case of an error or
5875 interruption must be on the cleanup chain before you call these
5876 functions, since they might never return to your code (they
5877 @samp{longjmp} instead).
5879 @section Per-architecture module data
5880 @cindex per-architecture module data
5881 @cindex multi-arch data
5882 @cindex data-pointer, per-architecture/per-module
5884 The multi-arch framework includes a mechanism for adding module
5885 specific per-architecture data-pointers to the @code{struct gdbarch}
5886 architecture object.
5888 A module registers one or more per-architecture data-pointers using:
5890 @deftypefn {Architecture Function} {struct gdbarch_data *} gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
5891 @var{pre_init} is used to, on-demand, allocate an initial value for a
5892 per-architecture data-pointer using the architecture's obstack (passed
5893 in as a parameter). Since @var{pre_init} can be called during
5894 architecture creation, it is not parameterized with the architecture.
5895 and must not call modules that use per-architecture data.
5898 @deftypefn {Architecture Function} {struct gdbarch_data *} gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
5899 @var{post_init} is used to obtain an initial value for a
5900 per-architecture data-pointer @emph{after}. Since @var{post_init} is
5901 always called after architecture creation, it both receives the fully
5902 initialized architecture and is free to call modules that use
5903 per-architecture data (care needs to be taken to ensure that those
5904 other modules do not try to call back to this module as that will
5905 create in cycles in the initialization call graph).
5908 These functions return a @code{struct gdbarch_data} that is used to
5909 identify the per-architecture data-pointer added for that module.
5911 The per-architecture data-pointer is accessed using the function:
5913 @deftypefn {Architecture Function} {void *} gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
5914 Given the architecture @var{arch} and module data handle
5915 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
5916 or @code{gdbarch_data_register_post_init}), this function returns the
5917 current value of the per-architecture data-pointer. If the data
5918 pointer is @code{NULL}, it is first initialized by calling the
5919 corresponding @var{pre_init} or @var{post_init} method.
5922 The examples below assume the following definitions:
5925 struct nozel @{ int total; @};
5926 static struct gdbarch_data *nozel_handle;
5929 A module can extend the architecture vector, adding additional
5930 per-architecture data, using the @var{pre_init} method. The module's
5931 per-architecture data is then initialized during architecture
5934 In the below, the module's per-architecture @emph{nozel} is added. An
5935 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
5936 from @code{gdbarch_init}.
5940 nozel_pre_init (struct obstack *obstack)
5942 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
5949 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
5951 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5952 data->total = nozel;
5956 A module can on-demand create architecture dependent data structures
5957 using @code{post_init}.
5959 In the below, the nozel's total is computed on-demand by
5960 @code{nozel_post_init} using information obtained from the
5965 nozel_post_init (struct gdbarch *gdbarch)
5967 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
5968 nozel->total = gdbarch@dots{} (gdbarch);
5975 nozel_total (struct gdbarch *gdbarch)
5977 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5982 @section Wrapping Output Lines
5983 @cindex line wrap in output
5986 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
5987 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
5988 added in places that would be good breaking points. The utility
5989 routines will take care of actually wrapping if the line width is
5992 The argument to @code{wrap_here} is an indentation string which is
5993 printed @emph{only} if the line breaks there. This argument is saved
5994 away and used later. It must remain valid until the next call to
5995 @code{wrap_here} or until a newline has been printed through the
5996 @code{*_filtered} functions. Don't pass in a local variable and then
5999 It is usually best to call @code{wrap_here} after printing a comma or
6000 space. If you call it before printing a space, make sure that your
6001 indentation properly accounts for the leading space that will print if
6002 the line wraps there.
6004 Any function or set of functions that produce filtered output must
6005 finish by printing a newline, to flush the wrap buffer, before switching
6006 to unfiltered (@code{printf}) output. Symbol reading routines that
6007 print warnings are a good example.
6009 @section @value{GDBN} Coding Standards
6010 @cindex coding standards
6012 @value{GDBN} follows the GNU coding standards, as described in
6013 @file{etc/standards.texi}. This file is also available for anonymous
6014 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
6015 of the standard; in general, when the GNU standard recommends a practice
6016 but does not require it, @value{GDBN} requires it.
6018 @value{GDBN} follows an additional set of coding standards specific to
6019 @value{GDBN}, as described in the following sections.
6024 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
6027 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
6030 @subsection Memory Management
6032 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
6033 @code{calloc}, @code{free} and @code{asprintf}.
6035 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
6036 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
6037 these functions do not return when the memory pool is empty. Instead,
6038 they unwind the stack using cleanups. These functions return
6039 @code{NULL} when requested to allocate a chunk of memory of size zero.
6041 @emph{Pragmatics: By using these functions, the need to check every
6042 memory allocation is removed. These functions provide portable
6045 @value{GDBN} does not use the function @code{free}.
6047 @value{GDBN} uses the function @code{xfree} to return memory to the
6048 memory pool. Consistent with ISO-C, this function ignores a request to
6049 free a @code{NULL} pointer.
6051 @emph{Pragmatics: On some systems @code{free} fails when passed a
6052 @code{NULL} pointer.}
6054 @value{GDBN} can use the non-portable function @code{alloca} for the
6055 allocation of small temporary values (such as strings).
6057 @emph{Pragmatics: This function is very non-portable. Some systems
6058 restrict the memory being allocated to no more than a few kilobytes.}
6060 @value{GDBN} uses the string function @code{xstrdup} and the print
6061 function @code{xstrprintf}.
6063 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
6064 functions such as @code{sprintf} are very prone to buffer overflow
6068 @subsection Compiler Warnings
6069 @cindex compiler warnings
6071 With few exceptions, developers should avoid the configuration option
6072 @samp{--disable-werror} when building @value{GDBN}. The exceptions
6073 are listed in the file @file{gdb/MAINTAINERS}. The default, when
6074 building with @sc{gcc}, is @samp{--enable-werror}.
6076 This option causes @value{GDBN} (when built using GCC) to be compiled
6077 with a carefully selected list of compiler warning flags. Any warnings
6078 from those flags are treated as errors.
6080 The current list of warning flags includes:
6084 Recommended @sc{gcc} warnings.
6086 @item -Wdeclaration-after-statement
6088 @sc{gcc} 3.x (and later) and @sc{c99} allow declarations mixed with
6089 code, but @sc{gcc} 2.x and @sc{c89} do not.
6091 @item -Wpointer-arith
6093 @item -Wformat-nonliteral
6094 Non-literal format strings, with a few exceptions, are bugs - they
6095 might contain unintended user-supplied format specifiers.
6096 Since @value{GDBN} uses the @code{format printf} attribute on all
6097 @code{printf} like functions this checks not just @code{printf} calls
6098 but also calls to functions such as @code{fprintf_unfiltered}.
6100 @item -Wno-pointer-sign
6101 In version 4.0, GCC began warning about pointer argument passing or
6102 assignment even when the source and destination differed only in
6103 signedness. However, most @value{GDBN} code doesn't distinguish
6104 carefully between @code{char} and @code{unsigned char}. In early 2006
6105 the @value{GDBN} developers decided correcting these warnings wasn't
6106 worth the time it would take.
6108 @item -Wno-unused-parameter
6109 Due to the way that @value{GDBN} is implemented many functions have
6110 unused parameters. Consequently this warning is avoided. The macro
6111 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
6112 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
6117 @itemx -Wno-char-subscripts
6118 These are warnings which might be useful for @value{GDBN}, but are
6119 currently too noisy to enable with @samp{-Werror}.
6123 @subsection Formatting
6125 @cindex source code formatting
6126 The standard GNU recommendations for formatting must be followed
6129 A function declaration should not have its name in column zero. A
6130 function definition should have its name in column zero.
6134 static void foo (void);
6142 @emph{Pragmatics: This simplifies scripting. Function definitions can
6143 be found using @samp{^function-name}.}
6145 There must be a space between a function or macro name and the opening
6146 parenthesis of its argument list (except for macro definitions, as
6147 required by C). There must not be a space after an open paren/bracket
6148 or before a close paren/bracket.
6150 While additional whitespace is generally helpful for reading, do not use
6151 more than one blank line to separate blocks, and avoid adding whitespace
6152 after the end of a program line (as of 1/99, some 600 lines had
6153 whitespace after the semicolon). Excess whitespace causes difficulties
6154 for @code{diff} and @code{patch} utilities.
6156 Pointers are declared using the traditional K&R C style:
6170 @subsection Comments
6172 @cindex comment formatting
6173 The standard GNU requirements on comments must be followed strictly.
6175 Block comments must appear in the following form, with no @code{/*}- or
6176 @code{*/}-only lines, and no leading @code{*}:
6179 /* Wait for control to return from inferior to debugger. If inferior
6180 gets a signal, we may decide to start it up again instead of
6181 returning. That is why there is a loop in this function. When
6182 this function actually returns it means the inferior should be left
6183 stopped and @value{GDBN} should read more commands. */
6186 (Note that this format is encouraged by Emacs; tabbing for a multi-line
6187 comment works correctly, and @kbd{M-q} fills the block consistently.)
6189 Put a blank line between the block comments preceding function or
6190 variable definitions, and the definition itself.
6192 In general, put function-body comments on lines by themselves, rather
6193 than trying to fit them into the 20 characters left at the end of a
6194 line, since either the comment or the code will inevitably get longer
6195 than will fit, and then somebody will have to move it anyhow.
6199 @cindex C data types
6200 Code must not depend on the sizes of C data types, the format of the
6201 host's floating point numbers, the alignment of anything, or the order
6202 of evaluation of expressions.
6204 @cindex function usage
6205 Use functions freely. There are only a handful of compute-bound areas
6206 in @value{GDBN} that might be affected by the overhead of a function
6207 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
6208 limited by the target interface (whether serial line or system call).
6210 However, use functions with moderation. A thousand one-line functions
6211 are just as hard to understand as a single thousand-line function.
6213 @emph{Macros are bad, M'kay.}
6214 (But if you have to use a macro, make sure that the macro arguments are
6215 protected with parentheses.)
6219 Declarations like @samp{struct foo *} should be used in preference to
6220 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
6223 @subsection Function Prototypes
6224 @cindex function prototypes
6226 Prototypes must be used when both @emph{declaring} and @emph{defining}
6227 a function. Prototypes for @value{GDBN} functions must include both the
6228 argument type and name, with the name matching that used in the actual
6229 function definition.
6231 All external functions should have a declaration in a header file that
6232 callers include, except for @code{_initialize_*} functions, which must
6233 be external so that @file{init.c} construction works, but shouldn't be
6234 visible to random source files.
6236 Where a source file needs a forward declaration of a static function,
6237 that declaration must appear in a block near the top of the source file.
6240 @subsection Internal Error Recovery
6242 During its execution, @value{GDBN} can encounter two types of errors.
6243 User errors and internal errors. User errors include not only a user
6244 entering an incorrect command but also problems arising from corrupt
6245 object files and system errors when interacting with the target.
6246 Internal errors include situations where @value{GDBN} has detected, at
6247 run time, a corrupt or erroneous situation.
6249 When reporting an internal error, @value{GDBN} uses
6250 @code{internal_error} and @code{gdb_assert}.
6252 @value{GDBN} must not call @code{abort} or @code{assert}.
6254 @emph{Pragmatics: There is no @code{internal_warning} function. Either
6255 the code detected a user error, recovered from it and issued a
6256 @code{warning} or the code failed to correctly recover from the user
6257 error and issued an @code{internal_error}.}
6259 @subsection File Names
6261 Any file used when building the core of @value{GDBN} must be in lower
6262 case. Any file used when building the core of @value{GDBN} must be 8.3
6263 unique. These requirements apply to both source and generated files.
6265 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
6266 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
6267 is introduced to the build process both @file{Makefile.in} and
6268 @file{configure.in} need to be modified accordingly. Compare the
6269 convoluted conversion process needed to transform @file{COPYING} into
6270 @file{copying.c} with the conversion needed to transform
6271 @file{version.in} into @file{version.c}.}
6273 Any file non 8.3 compliant file (that is not used when building the core
6274 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
6276 @emph{Pragmatics: This is clearly a compromise.}
6278 When @value{GDBN} has a local version of a system header file (ex
6279 @file{string.h}) the file name based on the POSIX header prefixed with
6280 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
6281 independent: they should use only macros defined by @file{configure},
6282 the compiler, or the host; they should include only system headers; they
6283 should refer only to system types. They may be shared between multiple
6284 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
6286 For other files @samp{-} is used as the separator.
6289 @subsection Include Files
6291 A @file{.c} file should include @file{defs.h} first.
6293 A @file{.c} file should directly include the @code{.h} file of every
6294 declaration and/or definition it directly refers to. It cannot rely on
6297 A @file{.h} file should directly include the @code{.h} file of every
6298 declaration and/or definition it directly refers to. It cannot rely on
6299 indirect inclusion. Exception: The file @file{defs.h} does not need to
6300 be directly included.
6302 An external declaration should only appear in one include file.
6304 An external declaration should never appear in a @code{.c} file.
6305 Exception: a declaration for the @code{_initialize} function that
6306 pacifies @option{-Wmissing-declaration}.
6308 A @code{typedef} definition should only appear in one include file.
6310 An opaque @code{struct} declaration can appear in multiple @file{.h}
6311 files. Where possible, a @file{.h} file should use an opaque
6312 @code{struct} declaration instead of an include.
6314 All @file{.h} files should be wrapped in:
6317 #ifndef INCLUDE_FILE_NAME_H
6318 #define INCLUDE_FILE_NAME_H
6324 @subsection Clean Design and Portable Implementation
6327 In addition to getting the syntax right, there's the little question of
6328 semantics. Some things are done in certain ways in @value{GDBN} because long
6329 experience has shown that the more obvious ways caused various kinds of
6332 @cindex assumptions about targets
6333 You can't assume the byte order of anything that comes from a target
6334 (including @var{value}s, object files, and instructions). Such things
6335 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
6336 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
6337 such as @code{bfd_get_32}.
6339 You can't assume that you know what interface is being used to talk to
6340 the target system. All references to the target must go through the
6341 current @code{target_ops} vector.
6343 You can't assume that the host and target machines are the same machine
6344 (except in the ``native'' support modules). In particular, you can't
6345 assume that the target machine's header files will be available on the
6346 host machine. Target code must bring along its own header files --
6347 written from scratch or explicitly donated by their owner, to avoid
6351 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
6352 to write the code portably than to conditionalize it for various
6355 @cindex system dependencies
6356 New @code{#ifdef}'s which test for specific compilers or manufacturers
6357 or operating systems are unacceptable. All @code{#ifdef}'s should test
6358 for features. The information about which configurations contain which
6359 features should be segregated into the configuration files. Experience
6360 has proven far too often that a feature unique to one particular system
6361 often creeps into other systems; and that a conditional based on some
6362 predefined macro for your current system will become worthless over
6363 time, as new versions of your system come out that behave differently
6364 with regard to this feature.
6366 Adding code that handles specific architectures, operating systems,
6367 target interfaces, or hosts, is not acceptable in generic code.
6369 @cindex portable file name handling
6370 @cindex file names, portability
6371 One particularly notorious area where system dependencies tend to
6372 creep in is handling of file names. The mainline @value{GDBN} code
6373 assumes Posix semantics of file names: absolute file names begin with
6374 a forward slash @file{/}, slashes are used to separate leading
6375 directories, case-sensitive file names. These assumptions are not
6376 necessarily true on non-Posix systems such as MS-Windows. To avoid
6377 system-dependent code where you need to take apart or construct a file
6378 name, use the following portable macros:
6381 @findex HAVE_DOS_BASED_FILE_SYSTEM
6382 @item HAVE_DOS_BASED_FILE_SYSTEM
6383 This preprocessing symbol is defined to a non-zero value on hosts
6384 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
6385 symbol to write conditional code which should only be compiled for
6388 @findex IS_DIR_SEPARATOR
6389 @item IS_DIR_SEPARATOR (@var{c})
6390 Evaluates to a non-zero value if @var{c} is a directory separator
6391 character. On Unix and GNU/Linux systems, only a slash @file{/} is
6392 such a character, but on Windows, both @file{/} and @file{\} will
6395 @findex IS_ABSOLUTE_PATH
6396 @item IS_ABSOLUTE_PATH (@var{file})
6397 Evaluates to a non-zero value if @var{file} is an absolute file name.
6398 For Unix and GNU/Linux hosts, a name which begins with a slash
6399 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
6400 @file{x:\bar} are also absolute file names.
6402 @findex FILENAME_CMP
6403 @item FILENAME_CMP (@var{f1}, @var{f2})
6404 Calls a function which compares file names @var{f1} and @var{f2} as
6405 appropriate for the underlying host filesystem. For Posix systems,
6406 this simply calls @code{strcmp}; on case-insensitive filesystems it
6407 will call @code{strcasecmp} instead.
6409 @findex DIRNAME_SEPARATOR
6410 @item DIRNAME_SEPARATOR
6411 Evaluates to a character which separates directories in
6412 @code{PATH}-style lists, typically held in environment variables.
6413 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
6415 @findex SLASH_STRING
6417 This evaluates to a constant string you should use to produce an
6418 absolute filename from leading directories and the file's basename.
6419 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
6420 @code{"\\"} for some Windows-based ports.
6423 In addition to using these macros, be sure to use portable library
6424 functions whenever possible. For example, to extract a directory or a
6425 basename part from a file name, use the @code{dirname} and
6426 @code{basename} library functions (available in @code{libiberty} for
6427 platforms which don't provide them), instead of searching for a slash
6428 with @code{strrchr}.
6430 Another way to generalize @value{GDBN} along a particular interface is with an
6431 attribute struct. For example, @value{GDBN} has been generalized to handle
6432 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
6433 by defining the @code{target_ops} structure and having a current target (as
6434 well as a stack of targets below it, for memory references). Whenever
6435 something needs to be done that depends on which remote interface we are
6436 using, a flag in the current target_ops structure is tested (e.g.,
6437 @code{target_has_stack}), or a function is called through a pointer in the
6438 current target_ops structure. In this way, when a new remote interface
6439 is added, only one module needs to be touched---the one that actually
6440 implements the new remote interface. Other examples of
6441 attribute-structs are BFD access to multiple kinds of object file
6442 formats, or @value{GDBN}'s access to multiple source languages.
6444 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
6445 the code interfacing between @code{ptrace} and the rest of
6446 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
6447 something was very painful. In @value{GDBN} 4.x, these have all been
6448 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
6449 with variations between systems the same way any system-independent
6450 file would (hooks, @code{#if defined}, etc.), and machines which are
6451 radically different don't need to use @file{infptrace.c} at all.
6453 All debugging code must be controllable using the @samp{set debug
6454 @var{module}} command. Do not use @code{printf} to print trace
6455 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
6456 @code{#ifdef DEBUG}.
6461 @chapter Porting @value{GDBN}
6462 @cindex porting to new machines
6464 Most of the work in making @value{GDBN} compile on a new machine is in
6465 specifying the configuration of the machine. Porting a new
6466 architecture to @value{GDBN} can be broken into a number of steps.
6471 Ensure a @sc{bfd} exists for executables of the target architecture in
6472 the @file{bfd} directory. If one does not exist, create one by
6473 modifying an existing similar one.
6476 Implement a disassembler for the target architecture in the @file{opcodes}
6480 Define the target architecture in the @file{gdb} directory
6481 (@pxref{Adding a New Target, , Adding a New Target}). Add the pattern
6482 for the new target to @file{configure.tgt} with the names of the files
6483 that contain the code. By convention the target architecture
6484 definition for an architecture @var{arch} is placed in
6485 @file{@var{arch}-tdep.c}.
6487 Within @file{@var{arch}-tdep.c} define the function
6488 @code{_initialize_@var{arch}_tdep} which calls
6489 @code{gdbarch_register} to create the new @code{@w{struct
6490 gdbarch}} for the architecture.
6493 If a new remote target is needed, consider adding a new remote target
6494 by defining a function
6495 @code{_initialize_remote_@var{arch}}. However if at all possible
6496 use the @value{GDBN} @emph{Remote Serial Protocol} for this and implement
6497 the server side protocol independently with the target.
6500 If desired implement a simulator in the @file{sim} directory. This
6501 should create the library @file{libsim.a} implementing the interface
6502 in @file{remote-sim.h} (found in the @file{include} directory).
6505 Build and test. If desired, lobby the @sc{gdb} steering group to
6506 have the new port included in the main distribution!
6509 Add a description of the new architecture to the main @value{GDBN} user
6510 guide (@pxref{Configuration Specific Information, , Configuration
6511 Specific Information, gdb, Debugging with @value{GDBN}}).
6515 @node Versions and Branches
6516 @chapter Versions and Branches
6520 @value{GDBN}'s version is determined by the file
6521 @file{gdb/version.in} and takes one of the following forms:
6524 @item @var{major}.@var{minor}
6525 @itemx @var{major}.@var{minor}.@var{patchlevel}
6526 an official release (e.g., 6.2 or 6.2.1)
6527 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}
6528 a snapshot taken at @var{YYYY}-@var{MM}-@var{DD}-gmt (e.g.,
6529 6.1.50.20020302, 6.1.90.20020304, or 6.1.0.20020308)
6530 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}-cvs
6531 a @sc{cvs} check out drawn on @var{YYYY}-@var{MM}-@var{DD} (e.g.,
6532 6.1.50.20020302-cvs, 6.1.90.20020304-cvs, or 6.1.0.20020308-cvs)
6533 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD} (@var{vendor})
6534 a vendor specific release of @value{GDBN}, that while based on@*
6535 @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD},
6536 may include additional changes
6539 @value{GDBN}'s mainline uses the @var{major} and @var{minor} version
6540 numbers from the most recent release branch, with a @var{patchlevel}
6541 of 50. At the time each new release branch is created, the mainline's
6542 @var{major} and @var{minor} version numbers are updated.
6544 @value{GDBN}'s release branch is similar. When the branch is cut, the
6545 @var{patchlevel} is changed from 50 to 90. As draft releases are
6546 drawn from the branch, the @var{patchlevel} is incremented. Once the
6547 first release (@var{major}.@var{minor}) has been made, the
6548 @var{patchlevel} is set to 0 and updates have an incremented
6551 For snapshots, and @sc{cvs} check outs, it is also possible to
6552 identify the @sc{cvs} origin:
6555 @item @var{major}.@var{minor}.50.@var{YYYY}@var{MM}@var{DD}
6556 drawn from the @sc{head} of mainline @sc{cvs} (e.g., 6.1.50.20020302)
6557 @item @var{major}.@var{minor}.90.@var{YYYY}@var{MM}@var{DD}
6558 @itemx @var{major}.@var{minor}.91.@var{YYYY}@var{MM}@var{DD} @dots{}
6559 drawn from a release branch prior to the release (e.g.,
6561 @item @var{major}.@var{minor}.0.@var{YYYY}@var{MM}@var{DD}
6562 @itemx @var{major}.@var{minor}.1.@var{YYYY}@var{MM}@var{DD} @dots{}
6563 drawn from a release branch after the release (e.g., 6.2.0.20020308)
6566 If the previous @value{GDBN} version is 6.1 and the current version is
6567 6.2, then, substituting 6 for @var{major} and 1 or 2 for @var{minor},
6568 here's an illustration of a typical sequence:
6575 +--------------------------.
6578 6.2.50.20020303-cvs 6.1.90 (draft #1)
6580 6.2.50.20020304-cvs 6.1.90.20020304-cvs
6582 6.2.50.20020305-cvs 6.1.91 (draft #2)
6584 6.2.50.20020306-cvs 6.1.91.20020306-cvs
6586 6.2.50.20020307-cvs 6.2 (release)
6588 6.2.50.20020308-cvs 6.2.0.20020308-cvs
6590 6.2.50.20020309-cvs 6.2.1 (update)
6592 6.2.50.20020310-cvs <branch closed>
6596 +--------------------------.
6599 6.3.50.20020312-cvs 6.2.90 (draft #1)
6603 @section Release Branches
6604 @cindex Release Branches
6606 @value{GDBN} draws a release series (6.2, 6.2.1, @dots{}) from a
6607 single release branch, and identifies that branch using the @sc{cvs}
6611 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-branchpoint
6612 gdb_@var{major}_@var{minor}-branch
6613 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-release
6616 @emph{Pragmatics: To help identify the date at which a branch or
6617 release is made, both the branchpoint and release tags include the
6618 date that they are cut (@var{YYYY}@var{MM}@var{DD}) in the tag. The
6619 branch tag, denoting the head of the branch, does not need this.}
6621 @section Vendor Branches
6622 @cindex vendor branches
6624 To avoid version conflicts, vendors are expected to modify the file
6625 @file{gdb/version.in} to include a vendor unique alphabetic identifier
6626 (an official @value{GDBN} release never uses alphabetic characters in
6627 its version identifier). E.g., @samp{6.2widgit2}, or @samp{6.2 (Widgit
6630 @section Experimental Branches
6631 @cindex experimental branches
6633 @subsection Guidelines
6635 @value{GDBN} permits the creation of branches, cut from the @sc{cvs}
6636 repository, for experimental development. Branches make it possible
6637 for developers to share preliminary work, and maintainers to examine
6638 significant new developments.
6640 The following are a set of guidelines for creating such branches:
6644 @item a branch has an owner
6645 The owner can set further policy for a branch, but may not change the
6646 ground rules. In particular, they can set a policy for commits (be it
6647 adding more reviewers or deciding who can commit).
6649 @item all commits are posted
6650 All changes committed to a branch shall also be posted to
6651 @email{gdb-patches@@sourceware.org, the @value{GDBN} patches
6652 mailing list}. While commentary on such changes are encouraged, people
6653 should remember that the changes only apply to a branch.
6655 @item all commits are covered by an assignment
6656 This ensures that all changes belong to the Free Software Foundation,
6657 and avoids the possibility that the branch may become contaminated.
6659 @item a branch is focused
6660 A focused branch has a single objective or goal, and does not contain
6661 unnecessary or irrelevant changes. Cleanups, where identified, being
6662 be pushed into the mainline as soon as possible.
6664 @item a branch tracks mainline
6665 This keeps the level of divergence under control. It also keeps the
6666 pressure on developers to push cleanups and other stuff into the
6669 @item a branch shall contain the entire @value{GDBN} module
6670 The @value{GDBN} module @code{gdb} should be specified when creating a
6671 branch (branches of individual files should be avoided). @xref{Tags}.
6673 @item a branch shall be branded using @file{version.in}
6674 The file @file{gdb/version.in} shall be modified so that it identifies
6675 the branch @var{owner} and branch @var{name}, e.g.,
6676 @samp{6.2.50.20030303_owner_name} or @samp{6.2 (Owner Name)}.
6683 To simplify the identification of @value{GDBN} branches, the following
6684 branch tagging convention is strongly recommended:
6688 @item @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
6689 @itemx @var{owner}_@var{name}-@var{YYYYMMDD}-branch
6690 The branch point and corresponding branch tag. @var{YYYYMMDD} is the
6691 date that the branch was created. A branch is created using the
6692 sequence: @anchor{experimental branch tags}
6694 cvs rtag @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint gdb
6695 cvs rtag -b -r @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint \
6696 @var{owner}_@var{name}-@var{YYYYMMDD}-branch gdb
6699 @item @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
6700 The tagged point, on the mainline, that was used when merging the branch
6701 on @var{yyyymmdd}. To merge in all changes since the branch was cut,
6702 use a command sequence like:
6704 cvs rtag @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint gdb
6706 -j@var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
6707 -j@var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
6710 Similar sequences can be used to just merge in changes since the last
6716 For further information on @sc{cvs}, see
6717 @uref{http://www.gnu.org/software/cvs/, Concurrent Versions System}.
6719 @node Start of New Year Procedure
6720 @chapter Start of New Year Procedure
6721 @cindex new year procedure
6723 At the start of each new year, the following actions should be performed:
6727 Rotate the ChangeLog file
6729 The current @file{ChangeLog} file should be renamed into
6730 @file{ChangeLog-YYYY} where YYYY is the year that has just passed.
6731 A new @file{ChangeLog} file should be created, and its contents should
6732 contain a reference to the previous ChangeLog. The following should
6733 also be preserved at the end of the new ChangeLog, in order to provide
6734 the appropriate settings when editing this file with Emacs:
6740 version-control: never
6746 Add an entry for the newly created ChangeLog file (@file{ChangeLog-YYYY})
6747 in @file{gdb/config/djgpp/fnchange.lst}.
6750 Update the copyright year in the startup message
6752 Update the copyright year in:
6754 @item file @file{top.c}, function @code{print_gdb_version}
6755 @item file @file{gdbserver/server.c}, function @code{gdbserver_version}
6756 @item file @file{gdbserver/gdbreplay.c}, function @code{gdbreplay_version}
6760 Add the new year in the copyright notices of all source and documentation
6761 files. This can be done semi-automatically by running the @code{copyright.sh}
6762 script. This script requires Emacs 22 or later to be installed.
6768 @chapter Releasing @value{GDBN}
6769 @cindex making a new release of gdb
6771 @section Branch Commit Policy
6773 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
6774 5.1 and 5.2 all used the below:
6778 The @file{gdb/MAINTAINERS} file still holds.
6780 Don't fix something on the branch unless/until it is also fixed in the
6781 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
6782 file is better than committing a hack.
6784 When considering a patch for the branch, suggested criteria include:
6785 Does it fix a build? Does it fix the sequence @kbd{break main; run}
6786 when debugging a static binary?
6788 The further a change is from the core of @value{GDBN}, the less likely
6789 the change will worry anyone (e.g., target specific code).
6791 Only post a proposal to change the core of @value{GDBN} after you've
6792 sent individual bribes to all the people listed in the
6793 @file{MAINTAINERS} file @t{;-)}
6796 @emph{Pragmatics: Provided updates are restricted to non-core
6797 functionality there is little chance that a broken change will be fatal.
6798 This means that changes such as adding a new architectures or (within
6799 reason) support for a new host are considered acceptable.}
6802 @section Obsoleting code
6804 Before anything else, poke the other developers (and around the source
6805 code) to see if there is anything that can be removed from @value{GDBN}
6806 (an old target, an unused file).
6808 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
6809 line. Doing this means that it is easy to identify something that has
6810 been obsoleted when greping through the sources.
6812 The process is done in stages --- this is mainly to ensure that the
6813 wider @value{GDBN} community has a reasonable opportunity to respond.
6814 Remember, everything on the Internet takes a week.
6818 Post the proposal on @email{gdb@@sourceware.org, the GDB mailing
6819 list} Creating a bug report to track the task's state, is also highly
6824 Post the proposal on @email{gdb-announce@@sourceware.org, the GDB
6825 Announcement mailing list}.
6829 Go through and edit all relevant files and lines so that they are
6830 prefixed with the word @code{OBSOLETE}.
6832 Wait until the next GDB version, containing this obsolete code, has been
6835 Remove the obsolete code.
6839 @emph{Maintainer note: While removing old code is regrettable it is
6840 hopefully better for @value{GDBN}'s long term development. Firstly it
6841 helps the developers by removing code that is either no longer relevant
6842 or simply wrong. Secondly since it removes any history associated with
6843 the file (effectively clearing the slate) the developer has a much freer
6844 hand when it comes to fixing broken files.}
6848 @section Before the Branch
6850 The most important objective at this stage is to find and fix simple
6851 changes that become a pain to track once the branch is created. For
6852 instance, configuration problems that stop @value{GDBN} from even
6853 building. If you can't get the problem fixed, document it in the
6854 @file{gdb/PROBLEMS} file.
6856 @subheading Prompt for @file{gdb/NEWS}
6858 People always forget. Send a post reminding them but also if you know
6859 something interesting happened add it yourself. The @code{schedule}
6860 script will mention this in its e-mail.
6862 @subheading Review @file{gdb/README}
6864 Grab one of the nightly snapshots and then walk through the
6865 @file{gdb/README} looking for anything that can be improved. The
6866 @code{schedule} script will mention this in its e-mail.
6868 @subheading Refresh any imported files.
6870 A number of files are taken from external repositories. They include:
6874 @file{texinfo/texinfo.tex}
6876 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
6879 @file{etc/standards.texi}, @file{etc/make-stds.texi}
6882 @subheading Check the ARI
6884 @uref{http://sourceware.org/gdb/ari,,A.R.I.} is an @code{awk} script
6885 (Awk Regression Index ;-) that checks for a number of errors and coding
6886 conventions. The checks include things like using @code{malloc} instead
6887 of @code{xmalloc} and file naming problems. There shouldn't be any
6890 @subsection Review the bug data base
6892 Close anything obviously fixed.
6894 @subsection Check all cross targets build
6896 The targets are listed in @file{gdb/MAINTAINERS}.
6899 @section Cut the Branch
6901 @subheading Create the branch
6906 $ V=`echo $v | sed 's/\./_/g'`
6907 $ D=`date -u +%Y-%m-%d`
6910 $ echo cvs -f -d :ext:sourceware.org:/cvs/src rtag \
6911 -D $D-gmt gdb_$V-$D-branchpoint insight
6912 cvs -f -d :ext:sourceware.org:/cvs/src rtag
6913 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight
6916 $ echo cvs -f -d :ext:sourceware.org:/cvs/src rtag \
6917 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight
6918 cvs -f -d :ext:sourceware.org:/cvs/src rtag \
6919 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight
6927 By using @kbd{-D YYYY-MM-DD-gmt}, the branch is forced to an exact
6930 The trunk is first tagged so that the branch point can easily be found.
6932 Insight, which includes @value{GDBN}, is tagged at the same time.
6934 @file{version.in} gets bumped to avoid version number conflicts.
6936 The reading of @file{.cvsrc} is disabled using @file{-f}.
6939 @subheading Update @file{version.in}
6944 $ V=`echo $v | sed 's/\./_/g'`
6948 $ echo cvs -f -d :ext:sourceware.org:/cvs/src co \
6949 -r gdb_$V-branch src/gdb/version.in
6950 cvs -f -d :ext:sourceware.org:/cvs/src co
6951 -r gdb_5_2-branch src/gdb/version.in
6953 U src/gdb/version.in
6955 $ echo $u.90-0000-00-00-cvs > version.in
6957 5.1.90-0000-00-00-cvs
6958 $ cvs -f commit version.in
6963 @file{0000-00-00} is used as a date to pump prime the version.in update
6966 @file{.90} and the previous branch version are used as fairly arbitrary
6967 initial branch version number.
6971 @subheading Update the web and news pages
6975 @subheading Tweak cron to track the new branch
6977 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
6978 This file needs to be updated so that:
6982 A daily timestamp is added to the file @file{version.in}.
6984 The new branch is included in the snapshot process.
6988 See the file @file{gdbadmin/cron/README} for how to install the updated
6991 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
6992 any changes. That file is copied to both the branch/ and current/
6993 snapshot directories.
6996 @subheading Update the NEWS and README files
6998 The @file{NEWS} file needs to be updated so that on the branch it refers
6999 to @emph{changes in the current release} while on the trunk it also
7000 refers to @emph{changes since the current release}.
7002 The @file{README} file needs to be updated so that it refers to the
7005 @subheading Post the branch info
7007 Send an announcement to the mailing lists:
7011 @email{gdb-announce@@sourceware.org, GDB Announcement mailing list}
7013 @email{gdb@@sourceware.org, GDB Discussion mailing list} and
7014 @email{gdb-testers@@sourceware.org, GDB Testers mailing list}
7017 @emph{Pragmatics: The branch creation is sent to the announce list to
7018 ensure that people people not subscribed to the higher volume discussion
7021 The announcement should include:
7027 How to check out the branch using CVS.
7029 The date/number of weeks until the release.
7031 The branch commit policy still holds.
7034 @section Stabilize the branch
7036 Something goes here.
7038 @section Create a Release
7040 The process of creating and then making available a release is broken
7041 down into a number of stages. The first part addresses the technical
7042 process of creating a releasable tar ball. The later stages address the
7043 process of releasing that tar ball.
7045 When making a release candidate just the first section is needed.
7047 @subsection Create a release candidate
7049 The objective at this stage is to create a set of tar balls that can be
7050 made available as a formal release (or as a less formal release
7053 @subsubheading Freeze the branch
7055 Send out an e-mail notifying everyone that the branch is frozen to
7056 @email{gdb-patches@@sourceware.org}.
7058 @subsubheading Establish a few defaults.
7063 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
7065 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
7069 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
7071 /home/gdbadmin/bin/autoconf
7080 Check the @code{autoconf} version carefully. You want to be using the
7081 version taken from the @file{binutils} snapshot directory, which can be
7082 found at @uref{ftp://sourceware.org/pub/binutils/}. It is very
7083 unlikely that a system installed version of @code{autoconf} (e.g.,
7084 @file{/usr/bin/autoconf}) is correct.
7087 @subsubheading Check out the relevant modules:
7090 $ for m in gdb insight
7092 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
7102 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
7103 any confusion between what is written here and what your local
7104 @code{cvs} really does.
7107 @subsubheading Update relevant files.
7113 Major releases get their comments added as part of the mainline. Minor
7114 releases should probably mention any significant bugs that were fixed.
7116 Don't forget to include the @file{ChangeLog} entry.
7119 $ emacs gdb/src/gdb/NEWS
7124 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
7125 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
7130 You'll need to update:
7142 $ emacs gdb/src/gdb/README
7147 $ cp gdb/src/gdb/README insight/src/gdb/README
7148 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
7151 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
7152 before the initial branch was cut so just a simple substitute is needed
7155 @emph{Maintainer note: Other projects generate @file{README} and
7156 @file{INSTALL} from the core documentation. This might be worth
7159 @item gdb/version.in
7162 $ echo $v > gdb/src/gdb/version.in
7163 $ cat gdb/src/gdb/version.in
7165 $ emacs gdb/src/gdb/version.in
7168 ... Bump to version ...
7170 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
7171 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
7176 @subsubheading Do the dirty work
7178 This is identical to the process used to create the daily snapshot.
7181 $ for m in gdb insight
7183 ( cd $m/src && gmake -f src-release $m.tar )
7187 If the top level source directory does not have @file{src-release}
7188 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
7191 $ for m in gdb insight
7193 ( cd $m/src && gmake -f Makefile.in $m.tar )
7197 @subsubheading Check the source files
7199 You're looking for files that have mysteriously disappeared.
7200 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
7201 for the @file{version.in} update @kbd{cronjob}.
7204 $ ( cd gdb/src && cvs -f -q -n update )
7208 @dots{} lots of generated files @dots{}
7213 @dots{} lots of generated files @dots{}
7218 @emph{Don't worry about the @file{gdb.info-??} or
7219 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
7220 was also generated only something strange with CVS means that they
7221 didn't get suppressed). Fixing it would be nice though.}
7223 @subsubheading Create compressed versions of the release
7229 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
7230 $ for m in gdb insight
7232 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
7233 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
7243 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
7244 in that mode, @code{gzip} does not know the name of the file and, hence,
7245 can not include it in the compressed file. This is also why the release
7246 process runs @code{tar} and @code{bzip2} as separate passes.
7249 @subsection Sanity check the tar ball
7251 Pick a popular machine (Solaris/PPC?) and try the build on that.
7254 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
7259 $ ./gdb/gdb ./gdb/gdb
7263 Breakpoint 1 at 0x80732bc: file main.c, line 734.
7265 Starting program: /tmp/gdb-5.2/gdb/gdb
7267 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
7268 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
7270 $1 = @{argc = 136426532, argv = 0x821b7f0@}
7274 @subsection Make a release candidate available
7276 If this is a release candidate then the only remaining steps are:
7280 Commit @file{version.in} and @file{ChangeLog}
7282 Tweak @file{version.in} (and @file{ChangeLog} to read
7283 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
7284 process can restart.
7286 Make the release candidate available in
7287 @uref{ftp://sourceware.org/pub/gdb/snapshots/branch}
7289 Notify the relevant mailing lists ( @email{gdb@@sourceware.org} and
7290 @email{gdb-testers@@sourceware.org} that the candidate is available.
7293 @subsection Make a formal release available
7295 (And you thought all that was required was to post an e-mail.)
7297 @subsubheading Install on sware
7299 Copy the new files to both the release and the old release directory:
7302 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
7303 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
7307 Clean up the releases directory so that only the most recent releases
7308 are available (e.g.@: keep 5.2 and 5.2.1 but remove 5.1):
7311 $ cd ~ftp/pub/gdb/releases
7316 Update the file @file{README} and @file{.message} in the releases
7323 $ ln README .message
7326 @subsubheading Update the web pages.
7330 @item htdocs/download/ANNOUNCEMENT
7331 This file, which is posted as the official announcement, includes:
7334 General announcement.
7336 News. If making an @var{M}.@var{N}.1 release, retain the news from
7337 earlier @var{M}.@var{N} release.
7342 @item htdocs/index.html
7343 @itemx htdocs/news/index.html
7344 @itemx htdocs/download/index.html
7345 These files include:
7348 Announcement of the most recent release.
7350 News entry (remember to update both the top level and the news directory).
7352 These pages also need to be regenerate using @code{index.sh}.
7354 @item download/onlinedocs/
7355 You need to find the magic command that is used to generate the online
7356 docs from the @file{.tar.bz2}. The best way is to look in the output
7357 from one of the nightly @code{cron} jobs and then just edit accordingly.
7361 $ ~/ss/update-web-docs \
7362 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
7364 /www/sourceware/htdocs/gdb/download/onlinedocs \
7369 Just like the online documentation. Something like:
7372 $ /bin/sh ~/ss/update-web-ari \
7373 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
7375 /www/sourceware/htdocs/gdb/download/ari \
7381 @subsubheading Shadow the pages onto gnu
7383 Something goes here.
7386 @subsubheading Install the @value{GDBN} tar ball on GNU
7388 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
7389 @file{~ftp/gnu/gdb}.
7391 @subsubheading Make the @file{ANNOUNCEMENT}
7393 Post the @file{ANNOUNCEMENT} file you created above to:
7397 @email{gdb-announce@@sourceware.org, GDB Announcement mailing list}
7399 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
7400 day or so to let things get out)
7402 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
7407 The release is out but you're still not finished.
7409 @subsubheading Commit outstanding changes
7411 In particular you'll need to commit any changes to:
7415 @file{gdb/ChangeLog}
7417 @file{gdb/version.in}
7424 @subsubheading Tag the release
7429 $ d=`date -u +%Y-%m-%d`
7432 $ ( cd insight/src/gdb && cvs -f -q update )
7433 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
7436 Insight is used since that contains more of the release than
7439 @subsubheading Mention the release on the trunk
7441 Just put something in the @file{ChangeLog} so that the trunk also
7442 indicates when the release was made.
7444 @subsubheading Restart @file{gdb/version.in}
7446 If @file{gdb/version.in} does not contain an ISO date such as
7447 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
7448 committed all the release changes it can be set to
7449 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
7450 is important - it affects the snapshot process).
7452 Don't forget the @file{ChangeLog}.
7454 @subsubheading Merge into trunk
7456 The files committed to the branch may also need changes merged into the
7459 @subsubheading Revise the release schedule
7461 Post a revised release schedule to @email{gdb@@sourceware.org, GDB
7462 Discussion List} with an updated announcement. The schedule can be
7463 generated by running:
7466 $ ~/ss/schedule `date +%s` schedule
7470 The first parameter is approximate date/time in seconds (from the epoch)
7471 of the most recent release.
7473 Also update the schedule @code{cronjob}.
7475 @section Post release
7477 Remove any @code{OBSOLETE} code.
7484 The testsuite is an important component of the @value{GDBN} package.
7485 While it is always worthwhile to encourage user testing, in practice
7486 this is rarely sufficient; users typically use only a small subset of
7487 the available commands, and it has proven all too common for a change
7488 to cause a significant regression that went unnoticed for some time.
7490 The @value{GDBN} testsuite uses the DejaGNU testing framework. The
7491 tests themselves are calls to various @code{Tcl} procs; the framework
7492 runs all the procs and summarizes the passes and fails.
7494 @section Using the Testsuite
7496 @cindex running the test suite
7497 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
7498 testsuite's objdir) and type @code{make check}. This just sets up some
7499 environment variables and invokes DejaGNU's @code{runtest} script. While
7500 the testsuite is running, you'll get mentions of which test file is in use,
7501 and a mention of any unexpected passes or fails. When the testsuite is
7502 finished, you'll get a summary that looks like this:
7507 # of expected passes 6016
7508 # of unexpected failures 58
7509 # of unexpected successes 5
7510 # of expected failures 183
7511 # of unresolved testcases 3
7512 # of untested testcases 5
7515 To run a specific test script, type:
7517 make check RUNTESTFLAGS='@var{tests}'
7519 where @var{tests} is a list of test script file names, separated by
7522 If you use GNU make, you can use its @option{-j} option to run the
7523 testsuite in parallel. This can greatly reduce the amount of time it
7524 takes for the testsuite to run. In this case, if you set
7525 @code{RUNTESTFLAGS} then, by default, the tests will be run serially
7526 even under @option{-j}. You can override this and force a parallel run
7527 by setting the @code{make} variable @code{FORCE_PARALLEL} to any
7528 non-empty value. Note that the parallel @kbd{make check} assumes
7529 that you want to run the entire testsuite, so it is not compatible
7530 with some dejagnu options, like @option{--directory}.
7532 The ideal test run consists of expected passes only; however, reality
7533 conspires to keep us from this ideal. Unexpected failures indicate
7534 real problems, whether in @value{GDBN} or in the testsuite. Expected
7535 failures are still failures, but ones which have been decided are too
7536 hard to deal with at the time; for instance, a test case might work
7537 everywhere except on AIX, and there is no prospect of the AIX case
7538 being fixed in the near future. Expected failures should not be added
7539 lightly, since you may be masking serious bugs in @value{GDBN}.
7540 Unexpected successes are expected fails that are passing for some
7541 reason, while unresolved and untested cases often indicate some minor
7542 catastrophe, such as the compiler being unable to deal with a test
7545 When making any significant change to @value{GDBN}, you should run the
7546 testsuite before and after the change, to confirm that there are no
7547 regressions. Note that truly complete testing would require that you
7548 run the testsuite with all supported configurations and a variety of
7549 compilers; however this is more than really necessary. In many cases
7550 testing with a single configuration is sufficient. Other useful
7551 options are to test one big-endian (Sparc) and one little-endian (x86)
7552 host, a cross config with a builtin simulator (powerpc-eabi,
7553 mips-elf), or a 64-bit host (Alpha).
7555 If you add new functionality to @value{GDBN}, please consider adding
7556 tests for it as well; this way future @value{GDBN} hackers can detect
7557 and fix their changes that break the functionality you added.
7558 Similarly, if you fix a bug that was not previously reported as a test
7559 failure, please add a test case for it. Some cases are extremely
7560 difficult to test, such as code that handles host OS failures or bugs
7561 in particular versions of compilers, and it's OK not to try to write
7562 tests for all of those.
7564 DejaGNU supports separate build, host, and target machines. However,
7565 some @value{GDBN} test scripts do not work if the build machine and
7566 the host machine are not the same. In such an environment, these scripts
7567 will give a result of ``UNRESOLVED'', like this:
7570 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
7573 Sometimes it is convenient to get a transcript of the commands which
7574 the testsuite sends to @value{GDBN}. For example, if @value{GDBN}
7575 crashes during testing, a transcript can be used to more easily
7576 reconstruct the failure when running @value{GDBN} under @value{GDBN}.
7578 You can instruct the @value{GDBN} testsuite to write transcripts by
7579 setting the DejaGNU variable @code{TRANSCRIPT} (to any value)
7580 before invoking @code{runtest} or @kbd{make check}. The transcripts
7581 will be written into DejaGNU's output directory. One transcript will
7582 be made for each invocation of @value{GDBN}; they will be named
7583 @file{transcript.@var{n}}, where @var{n} is an integer. The first
7584 line of the transcript file will show how @value{GDBN} was invoked;
7585 each subsequent line is a command sent as input to @value{GDBN}.
7588 make check RUNTESTFLAGS=TRANSCRIPT=y
7591 Note that the transcript is not always complete. In particular, tests
7592 of completion can yield partial command lines.
7594 @section Testsuite Organization
7596 @cindex test suite organization
7597 The testsuite is entirely contained in @file{gdb/testsuite}. While the
7598 testsuite includes some makefiles and configury, these are very minimal,
7599 and used for little besides cleaning up, since the tests themselves
7600 handle the compilation of the programs that @value{GDBN} will run. The file
7601 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
7602 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
7603 configuration-specific files, typically used for special-purpose
7604 definitions of procs like @code{gdb_load} and @code{gdb_start}.
7606 The tests themselves are to be found in @file{testsuite/gdb.*} and
7607 subdirectories of those. The names of the test files must always end
7608 with @file{.exp}. DejaGNU collects the test files by wildcarding
7609 in the test directories, so both subdirectories and individual files
7610 get chosen and run in alphabetical order.
7612 The following table lists the main types of subdirectories and what they
7613 are for. Since DejaGNU finds test files no matter where they are
7614 located, and since each test file sets up its own compilation and
7615 execution environment, this organization is simply for convenience and
7620 This is the base testsuite. The tests in it should apply to all
7621 configurations of @value{GDBN} (but generic native-only tests may live here).
7622 The test programs should be in the subset of C that is valid K&R,
7623 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
7626 @item gdb.@var{lang}
7627 Language-specific tests for any language @var{lang} besides C. Examples are
7628 @file{gdb.cp} and @file{gdb.java}.
7630 @item gdb.@var{platform}
7631 Non-portable tests. The tests are specific to a specific configuration
7632 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
7635 @item gdb.@var{compiler}
7636 Tests specific to a particular compiler. As of this writing (June
7637 1999), there aren't currently any groups of tests in this category that
7638 couldn't just as sensibly be made platform-specific, but one could
7639 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
7642 @item gdb.@var{subsystem}
7643 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
7644 instance, @file{gdb.disasm} exercises various disassemblers, while
7645 @file{gdb.stabs} tests pathways through the stabs symbol reader.
7648 @section Writing Tests
7649 @cindex writing tests
7651 In many areas, the @value{GDBN} tests are already quite comprehensive; you
7652 should be able to copy existing tests to handle new cases.
7654 You should try to use @code{gdb_test} whenever possible, since it
7655 includes cases to handle all the unexpected errors that might happen.
7656 However, it doesn't cost anything to add new test procedures; for
7657 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
7658 calls @code{gdb_test} multiple times.
7660 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
7661 necessary. Even if @value{GDBN} has several valid responses to
7662 a command, you can use @code{gdb_test_multiple}. Like @code{gdb_test},
7663 @code{gdb_test_multiple} recognizes internal errors and unexpected
7666 Do not write tests which expect a literal tab character from @value{GDBN}.
7667 On some operating systems (e.g.@: OpenBSD) the TTY layer expands tabs to
7668 spaces, so by the time @value{GDBN}'s output reaches expect the tab is gone.
7670 The source language programs do @emph{not} need to be in a consistent
7671 style. Since @value{GDBN} is used to debug programs written in many different
7672 styles, it's worth having a mix of styles in the testsuite; for
7673 instance, some @value{GDBN} bugs involving the display of source lines would
7674 never manifest themselves if the programs used GNU coding style
7681 Check the @file{README} file, it often has useful information that does not
7682 appear anywhere else in the directory.
7685 * Getting Started:: Getting started working on @value{GDBN}
7686 * Debugging GDB:: Debugging @value{GDBN} with itself
7689 @node Getting Started,,, Hints
7691 @section Getting Started
7693 @value{GDBN} is a large and complicated program, and if you first starting to
7694 work on it, it can be hard to know where to start. Fortunately, if you
7695 know how to go about it, there are ways to figure out what is going on.
7697 This manual, the @value{GDBN} Internals manual, has information which applies
7698 generally to many parts of @value{GDBN}.
7700 Information about particular functions or data structures are located in
7701 comments with those functions or data structures. If you run across a
7702 function or a global variable which does not have a comment correctly
7703 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
7704 free to submit a bug report, with a suggested comment if you can figure
7705 out what the comment should say. If you find a comment which is
7706 actually wrong, be especially sure to report that.
7708 Comments explaining the function of macros defined in host, target, or
7709 native dependent files can be in several places. Sometimes they are
7710 repeated every place the macro is defined. Sometimes they are where the
7711 macro is used. Sometimes there is a header file which supplies a
7712 default definition of the macro, and the comment is there. This manual
7713 also documents all the available macros.
7714 @c (@pxref{Host Conditionals}, @pxref{Target
7715 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
7718 Start with the header files. Once you have some idea of how
7719 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
7720 @file{gdbtypes.h}), you will find it much easier to understand the
7721 code which uses and creates those symbol tables.
7723 You may wish to process the information you are getting somehow, to
7724 enhance your understanding of it. Summarize it, translate it to another
7725 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
7726 the code to predict what a test case would do and write the test case
7727 and verify your prediction, etc. If you are reading code and your eyes
7728 are starting to glaze over, this is a sign you need to use a more active
7731 Once you have a part of @value{GDBN} to start with, you can find more
7732 specifically the part you are looking for by stepping through each
7733 function with the @code{next} command. Do not use @code{step} or you
7734 will quickly get distracted; when the function you are stepping through
7735 calls another function try only to get a big-picture understanding
7736 (perhaps using the comment at the beginning of the function being
7737 called) of what it does. This way you can identify which of the
7738 functions being called by the function you are stepping through is the
7739 one which you are interested in. You may need to examine the data
7740 structures generated at each stage, with reference to the comments in
7741 the header files explaining what the data structures are supposed to
7744 Of course, this same technique can be used if you are just reading the
7745 code, rather than actually stepping through it. The same general
7746 principle applies---when the code you are looking at calls something
7747 else, just try to understand generally what the code being called does,
7748 rather than worrying about all its details.
7750 @cindex command implementation
7751 A good place to start when tracking down some particular area is with
7752 a command which invokes that feature. Suppose you want to know how
7753 single-stepping works. As a @value{GDBN} user, you know that the
7754 @code{step} command invokes single-stepping. The command is invoked
7755 via command tables (see @file{command.h}); by convention the function
7756 which actually performs the command is formed by taking the name of
7757 the command and adding @samp{_command}, or in the case of an
7758 @code{info} subcommand, @samp{_info}. For example, the @code{step}
7759 command invokes the @code{step_command} function and the @code{info
7760 display} command invokes @code{display_info}. When this convention is
7761 not followed, you might have to use @code{grep} or @kbd{M-x
7762 tags-search} in emacs, or run @value{GDBN} on itself and set a
7763 breakpoint in @code{execute_command}.
7765 @cindex @code{bug-gdb} mailing list
7766 If all of the above fail, it may be appropriate to ask for information
7767 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
7768 wondering if anyone could give me some tips about understanding
7769 @value{GDBN}''---if we had some magic secret we would put it in this manual.
7770 Suggestions for improving the manual are always welcome, of course.
7772 @node Debugging GDB,,,Hints
7774 @section Debugging @value{GDBN} with itself
7775 @cindex debugging @value{GDBN}
7777 If @value{GDBN} is limping on your machine, this is the preferred way to get it
7778 fully functional. Be warned that in some ancient Unix systems, like
7779 Ultrix 4.2, a program can't be running in one process while it is being
7780 debugged in another. Rather than typing the command @kbd{@w{./gdb
7781 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
7782 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
7784 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
7785 @file{.gdbinit} file that sets up some simple things to make debugging
7786 gdb easier. The @code{info} command, when executed without a subcommand
7787 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
7788 gdb. See @file{.gdbinit} for details.
7790 If you use emacs, you will probably want to do a @code{make TAGS} after
7791 you configure your distribution; this will put the machine dependent
7792 routines for your local machine where they will be accessed first by
7795 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
7796 have run @code{fixincludes} if you are compiling with gcc.
7798 @section Submitting Patches
7800 @cindex submitting patches
7801 Thanks for thinking of offering your changes back to the community of
7802 @value{GDBN} users. In general we like to get well designed enhancements.
7803 Thanks also for checking in advance about the best way to transfer the
7806 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
7807 This manual summarizes what we believe to be clean design for @value{GDBN}.
7809 If the maintainers don't have time to put the patch in when it arrives,
7810 or if there is any question about a patch, it goes into a large queue
7811 with everyone else's patches and bug reports.
7813 @cindex legal papers for code contributions
7814 The legal issue is that to incorporate substantial changes requires a
7815 copyright assignment from you and/or your employer, granting ownership
7816 of the changes to the Free Software Foundation. You can get the
7817 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
7818 and asking for it. We recommend that people write in "All programs
7819 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
7820 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
7822 contributed with only one piece of legalese pushed through the
7823 bureaucracy and filed with the FSF. We can't start merging changes until
7824 this paperwork is received by the FSF (their rules, which we follow
7825 since we maintain it for them).
7827 Technically, the easiest way to receive changes is to receive each
7828 feature as a small context diff or unidiff, suitable for @code{patch}.
7829 Each message sent to me should include the changes to C code and
7830 header files for a single feature, plus @file{ChangeLog} entries for
7831 each directory where files were modified, and diffs for any changes
7832 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
7833 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
7834 single feature, they can be split down into multiple messages.
7836 In this way, if we read and like the feature, we can add it to the
7837 sources with a single patch command, do some testing, and check it in.
7838 If you leave out the @file{ChangeLog}, we have to write one. If you leave
7839 out the doc, we have to puzzle out what needs documenting. Etc., etc.
7841 The reason to send each change in a separate message is that we will not
7842 install some of the changes. They'll be returned to you with questions
7843 or comments. If we're doing our job correctly, the message back to you
7844 will say what you have to fix in order to make the change acceptable.
7845 The reason to have separate messages for separate features is so that
7846 the acceptable changes can be installed while one or more changes are
7847 being reworked. If multiple features are sent in a single message, we
7848 tend to not put in the effort to sort out the acceptable changes from
7849 the unacceptable, so none of the features get installed until all are
7852 If this sounds painful or authoritarian, well, it is. But we get a lot
7853 of bug reports and a lot of patches, and many of them don't get
7854 installed because we don't have the time to finish the job that the bug
7855 reporter or the contributor could have done. Patches that arrive
7856 complete, working, and well designed, tend to get installed on the day
7857 they arrive. The others go into a queue and get installed as time
7858 permits, which, since the maintainers have many demands to meet, may not
7859 be for quite some time.
7861 Please send patches directly to
7862 @email{gdb-patches@@sourceware.org, the @value{GDBN} maintainers}.
7864 @section Build Script
7866 @cindex build script
7868 The script @file{gdb_buildall.sh} builds @value{GDBN} with flag
7869 @option{--enable-targets=all} set. This builds @value{GDBN} with all supported
7870 targets activated. This helps testing @value{GDBN} when doing changes that
7871 affect more than one architecture and is much faster than using
7872 @file{gdb_mbuild.sh}.
7874 After building @value{GDBN} the script checks which architectures are
7875 supported and then switches the current architecture to each of those to get
7876 information about the architecture. The test results are stored in log files
7877 in the directory the script was called from.
7879 @include observer.texi