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_HAS_HARDWARE_WATCHPOINTS
698 @item TARGET_HAS_HARDWARE_WATCHPOINTS
699 If defined, the target supports hardware watchpoints.
700 (Currently only used for several native configs.)
702 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
703 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
704 Return the number of hardware watchpoints of type @var{type} that are
705 possible to be set. The value is positive if @var{count} watchpoints
706 of this type can be set, zero if setting watchpoints of this type is
707 not supported, and negative if @var{count} is more than the maximum
708 number of watchpoints of type @var{type} that can be set. @var{other}
709 is non-zero if other types of watchpoints are currently enabled (there
710 are architectures which cannot set watchpoints of different types at
713 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
714 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
715 Return non-zero if hardware watchpoints can be used to watch a region
716 whose address is @var{addr} and whose length in bytes is @var{len}.
718 @cindex insert or remove hardware watchpoint
719 @findex target_insert_watchpoint
720 @findex target_remove_watchpoint
721 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
722 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
723 Insert or remove a hardware watchpoint starting at @var{addr}, for
724 @var{len} bytes. @var{type} is the watchpoint type, one of the
725 possible values of the enumerated data type @code{target_hw_bp_type},
726 defined by @file{breakpoint.h} as follows:
729 enum target_hw_bp_type
731 hw_write = 0, /* Common (write) HW watchpoint */
732 hw_read = 1, /* Read HW watchpoint */
733 hw_access = 2, /* Access (read or write) HW watchpoint */
734 hw_execute = 3 /* Execute HW breakpoint */
739 These two macros should return 0 for success, non-zero for failure.
741 @findex target_stopped_data_address
742 @item target_stopped_data_address (@var{addr_p})
743 If the inferior has some watchpoint that triggered, place the address
744 associated with the watchpoint at the location pointed to by
745 @var{addr_p} and return non-zero. Otherwise, return zero. This
746 is required for data-read and data-access watchpoints. It is
747 not required for data-write watchpoints, but @value{GDBN} uses
748 it to improve handling of those also.
750 @value{GDBN} will only call this method once per watchpoint stop,
751 immediately after calling @code{STOPPED_BY_WATCHPOINT}. If the
752 target's watchpoint indication is sticky, i.e., stays set after
753 resuming, this method should clear it. For instance, the x86 debug
754 control register has sticky triggered flags.
756 @findex target_watchpoint_addr_within_range
757 @item target_watchpoint_addr_within_range (@var{target}, @var{addr}, @var{start}, @var{length})
758 Check whether @var{addr} (as returned by @code{target_stopped_data_address})
759 lies within the hardware-defined watchpoint region described by
760 @var{start} and @var{length}. This only needs to be provided if the
761 granularity of a watchpoint is greater than one byte, i.e., if the
762 watchpoint can also trigger on nearby addresses outside of the watched
765 @findex HAVE_STEPPABLE_WATCHPOINT
766 @item HAVE_STEPPABLE_WATCHPOINT
767 If defined to a non-zero value, it is not necessary to disable a
768 watchpoint to step over it. Like @code{gdbarch_have_nonsteppable_watchpoint},
769 this is usually set when watchpoints trigger at the instruction
770 which will perform an interesting read or write. It should be
771 set if there is a temporary disable bit which allows the processor
772 to step over the interesting instruction without raising the
773 watchpoint exception again.
775 @findex gdbarch_have_nonsteppable_watchpoint
776 @item int gdbarch_have_nonsteppable_watchpoint (@var{gdbarch})
777 If it returns a non-zero value, @value{GDBN} should disable a
778 watchpoint to step the inferior over it. This is usually set when
779 watchpoints trigger at the instruction which will perform an
780 interesting read or write.
782 @findex HAVE_CONTINUABLE_WATCHPOINT
783 @item HAVE_CONTINUABLE_WATCHPOINT
784 If defined to a non-zero value, it is possible to continue the
785 inferior after a watchpoint has been hit. This is usually set
786 when watchpoints trigger at the instruction following an interesting
789 @findex CANNOT_STEP_HW_WATCHPOINTS
790 @item CANNOT_STEP_HW_WATCHPOINTS
791 If this is defined to a non-zero value, @value{GDBN} will remove all
792 watchpoints before stepping the inferior.
794 @findex STOPPED_BY_WATCHPOINT
795 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
796 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
797 the type @code{struct target_waitstatus}, defined by @file{target.h}.
798 Normally, this macro is defined to invoke the function pointed to by
799 the @code{to_stopped_by_watchpoint} member of the structure (of the
800 type @code{target_ops}, defined on @file{target.h}) that describes the
801 target-specific operations; @code{to_stopped_by_watchpoint} ignores
802 the @var{wait_status} argument.
804 @value{GDBN} does not require the non-zero value returned by
805 @code{STOPPED_BY_WATCHPOINT} to be 100% correct, so if a target cannot
806 determine for sure whether the inferior stopped due to a watchpoint,
807 it could return non-zero ``just in case''.
810 @subsection Watchpoints and Threads
811 @cindex watchpoints, with threads
813 @value{GDBN} only supports process-wide watchpoints, which trigger
814 in all threads. @value{GDBN} uses the thread ID to make watchpoints
815 act as if they were thread-specific, but it cannot set hardware
816 watchpoints that only trigger in a specific thread. Therefore, even
817 if the target supports threads, per-thread debug registers, and
818 watchpoints which only affect a single thread, it should set the
819 per-thread debug registers for all threads to the same value. On
820 @sc{gnu}/Linux native targets, this is accomplished by using
821 @code{ALL_LWPS} in @code{target_insert_watchpoint} and
822 @code{target_remove_watchpoint} and by using
823 @code{linux_set_new_thread} to register a handler for newly created
826 @value{GDBN}'s @sc{gnu}/Linux support only reports a single event
827 at a time, although multiple events can trigger simultaneously for
828 multi-threaded programs. When multiple events occur, @file{linux-nat.c}
829 queues subsequent events and returns them the next time the program
830 is resumed. This means that @code{STOPPED_BY_WATCHPOINT} and
831 @code{target_stopped_data_address} only need to consult the current
832 thread's state---the thread indicated by @code{inferior_ptid}. If
833 two threads have hit watchpoints simultaneously, those routines
834 will be called a second time for the second thread.
836 @subsection x86 Watchpoints
837 @cindex x86 debug registers
838 @cindex watchpoints, on x86
840 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
841 registers designed to facilitate debugging. @value{GDBN} provides a
842 generic library of functions that x86-based ports can use to implement
843 support for watchpoints and hardware-assisted breakpoints. This
844 subsection documents the x86 watchpoint facilities in @value{GDBN}.
846 (At present, the library functions read and write debug registers directly, and are
847 thus only available for native configurations.)
849 To use the generic x86 watchpoint support, a port should do the
853 @findex I386_USE_GENERIC_WATCHPOINTS
855 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
856 target-dependent headers.
859 Include the @file{config/i386/nm-i386.h} header file @emph{after}
860 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
863 Add @file{i386-nat.o} to the value of the Make variable
864 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}).
867 Provide implementations for the @code{I386_DR_LOW_*} macros described
868 below. Typically, each macro should call a target-specific function
869 which does the real work.
872 The x86 watchpoint support works by maintaining mirror images of the
873 debug registers. Values are copied between the mirror images and the
874 real debug registers via a set of macros which each target needs to
878 @findex I386_DR_LOW_SET_CONTROL
879 @item I386_DR_LOW_SET_CONTROL (@var{val})
880 Set the Debug Control (DR7) register to the value @var{val}.
882 @findex I386_DR_LOW_SET_ADDR
883 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
884 Put the address @var{addr} into the debug register number @var{idx}.
886 @findex I386_DR_LOW_RESET_ADDR
887 @item I386_DR_LOW_RESET_ADDR (@var{idx})
888 Reset (i.e.@: zero out) the address stored in the debug register
891 @findex I386_DR_LOW_GET_STATUS
892 @item I386_DR_LOW_GET_STATUS
893 Return the value of the Debug Status (DR6) register. This value is
894 used immediately after it is returned by
895 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
899 For each one of the 4 debug registers (whose indices are from 0 to 3)
900 that store addresses, a reference count is maintained by @value{GDBN},
901 to allow sharing of debug registers by several watchpoints. This
902 allows users to define several watchpoints that watch the same
903 expression, but with different conditions and/or commands, without
904 wasting debug registers which are in short supply. @value{GDBN}
905 maintains the reference counts internally, targets don't have to do
906 anything to use this feature.
908 The x86 debug registers can each watch a region that is 1, 2, or 4
909 bytes long. The ia32 architecture requires that each watched region
910 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
911 region on 4-byte boundary. However, the x86 watchpoint support in
912 @value{GDBN} can watch unaligned regions and regions larger than 4
913 bytes (up to 16 bytes) by allocating several debug registers to watch
914 a single region. This allocation of several registers per a watched
915 region is also done automatically without target code intervention.
917 The generic x86 watchpoint support provides the following API for the
918 @value{GDBN}'s application code:
921 @findex i386_region_ok_for_watchpoint
922 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
923 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
924 this function. It counts the number of debug registers required to
925 watch a given region, and returns a non-zero value if that number is
926 less than 4, the number of debug registers available to x86
929 @findex i386_stopped_data_address
930 @item i386_stopped_data_address (@var{addr_p})
932 @code{target_stopped_data_address} is set to call this function.
934 function examines the breakpoint condition bits in the DR6 Debug
935 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
936 macro, and returns the address associated with the first bit that is
939 @findex i386_stopped_by_watchpoint
940 @item i386_stopped_by_watchpoint (void)
941 The macro @code{STOPPED_BY_WATCHPOINT}
942 is set to call this function. The
943 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
944 function examines the breakpoint condition bits in the DR6 Debug
945 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
946 macro, and returns true if any bit is set. Otherwise, false is
949 @findex i386_insert_watchpoint
950 @findex i386_remove_watchpoint
951 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
952 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
953 Insert or remove a watchpoint. The macros
954 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
955 are set to call these functions. @code{i386_insert_watchpoint} first
956 looks for a debug register which is already set to watch the same
957 region for the same access types; if found, it just increments the
958 reference count of that debug register, thus implementing debug
959 register sharing between watchpoints. If no such register is found,
960 the function looks for a vacant debug register, sets its mirrored
961 value to @var{addr}, sets the mirrored value of DR7 Debug Control
962 register as appropriate for the @var{len} and @var{type} parameters,
963 and then passes the new values of the debug register and DR7 to the
964 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
965 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
966 required to cover the given region, the above process is repeated for
969 @code{i386_remove_watchpoint} does the opposite: it resets the address
970 in the mirrored value of the debug register and its read/write and
971 length bits in the mirrored value of DR7, then passes these new
972 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
973 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
974 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
975 decrements the reference count, and only calls
976 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
977 the count goes to zero.
979 @findex i386_insert_hw_breakpoint
980 @findex i386_remove_hw_breakpoint
981 @item i386_insert_hw_breakpoint (@var{bp_tgt})
982 @itemx i386_remove_hw_breakpoint (@var{bp_tgt})
983 These functions insert and remove hardware-assisted breakpoints. The
984 macros @code{target_insert_hw_breakpoint} and
985 @code{target_remove_hw_breakpoint} are set to call these functions.
986 The argument is a @code{struct bp_target_info *}, as described in
987 the documentation for @code{target_insert_breakpoint}.
988 These functions work like @code{i386_insert_watchpoint} and
989 @code{i386_remove_watchpoint}, respectively, except that they set up
990 the debug registers to watch instruction execution, and each
991 hardware-assisted breakpoint always requires exactly one debug
994 @findex i386_stopped_by_hwbp
995 @item i386_stopped_by_hwbp (void)
996 This function returns non-zero if the inferior has some watchpoint or
997 hardware breakpoint that triggered. It works like
998 @code{i386_stopped_data_address}, except that it doesn't record the
999 address whose watchpoint triggered.
1001 @findex i386_cleanup_dregs
1002 @item i386_cleanup_dregs (void)
1003 This function clears all the reference counts, addresses, and control
1004 bits in the mirror images of the debug registers. It doesn't affect
1005 the actual debug registers in the inferior process.
1012 x86 processors support setting watchpoints on I/O reads or writes.
1013 However, since no target supports this (as of March 2001), and since
1014 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
1015 watchpoints, this feature is not yet available to @value{GDBN} running
1019 x86 processors can enable watchpoints locally, for the current task
1020 only, or globally, for all the tasks. For each debug register,
1021 there's a bit in the DR7 Debug Control register that determines
1022 whether the associated address is watched locally or globally. The
1023 current implementation of x86 watchpoint support in @value{GDBN}
1024 always sets watchpoints to be locally enabled, since global
1025 watchpoints might interfere with the underlying OS and are probably
1026 unavailable in many platforms.
1029 @section Checkpoints
1032 In the abstract, a checkpoint is a point in the execution history of
1033 the program, which the user may wish to return to at some later time.
1035 Internally, a checkpoint is a saved copy of the program state, including
1036 whatever information is required in order to restore the program to that
1037 state at a later time. This can be expected to include the state of
1038 registers and memory, and may include external state such as the state
1039 of open files and devices.
1041 There are a number of ways in which checkpoints may be implemented
1042 in gdb, e.g.@: as corefiles, as forked processes, and as some opaque
1043 method implemented on the target side.
1045 A corefile can be used to save an image of target memory and register
1046 state, which can in principle be restored later --- but corefiles do
1047 not typically include information about external entities such as
1048 open files. Currently this method is not implemented in gdb.
1050 A forked process can save the state of user memory and registers,
1051 as well as some subset of external (kernel) state. This method
1052 is used to implement checkpoints on Linux, and in principle might
1053 be used on other systems.
1055 Some targets, e.g.@: simulators, might have their own built-in
1056 method for saving checkpoints, and gdb might be able to take
1057 advantage of that capability without necessarily knowing any
1058 details of how it is done.
1061 @section Observing changes in @value{GDBN} internals
1062 @cindex observer pattern interface
1063 @cindex notifications about changes in internals
1065 In order to function properly, several modules need to be notified when
1066 some changes occur in the @value{GDBN} internals. Traditionally, these
1067 modules have relied on several paradigms, the most common ones being
1068 hooks and gdb-events. Unfortunately, none of these paradigms was
1069 versatile enough to become the standard notification mechanism in
1070 @value{GDBN}. The fact that they only supported one ``client'' was also
1071 a strong limitation.
1073 A new paradigm, based on the Observer pattern of the @cite{Design
1074 Patterns} book, has therefore been implemented. The goal was to provide
1075 a new interface overcoming the issues with the notification mechanisms
1076 previously available. This new interface needed to be strongly typed,
1077 easy to extend, and versatile enough to be used as the standard
1078 interface when adding new notifications.
1080 See @ref{GDB Observers} for a brief description of the observers
1081 currently implemented in GDB. The rationale for the current
1082 implementation is also briefly discussed.
1084 @node User Interface
1086 @chapter User Interface
1088 @value{GDBN} has several user interfaces, of which the traditional
1089 command-line interface is perhaps the most familiar.
1091 @section Command Interpreter
1093 @cindex command interpreter
1095 The command interpreter in @value{GDBN} is fairly simple. It is designed to
1096 allow for the set of commands to be augmented dynamically, and also
1097 has a recursive subcommand capability, where the first argument to
1098 a command may itself direct a lookup on a different command list.
1100 For instance, the @samp{set} command just starts a lookup on the
1101 @code{setlist} command list, while @samp{set thread} recurses
1102 to the @code{set_thread_cmd_list}.
1106 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
1107 the main command list, and should be used for those commands. The usual
1108 place to add commands is in the @code{_initialize_@var{xyz}} routines at
1109 the ends of most source files.
1111 @findex add_setshow_cmd
1112 @findex add_setshow_cmd_full
1113 To add paired @samp{set} and @samp{show} commands, use
1114 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
1115 a slightly simpler interface which is useful when you don't need to
1116 further modify the new command structures, while the latter returns
1117 the new command structures for manipulation.
1119 @cindex deprecating commands
1120 @findex deprecate_cmd
1121 Before removing commands from the command set it is a good idea to
1122 deprecate them for some time. Use @code{deprecate_cmd} on commands or
1123 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
1124 @code{struct cmd_list_element} as it's first argument. You can use the
1125 return value from @code{add_com} or @code{add_cmd} to deprecate the
1126 command immediately after it is created.
1128 The first time a command is used the user will be warned and offered a
1129 replacement (if one exists). Note that the replacement string passed to
1130 @code{deprecate_cmd} should be the full name of the command, i.e., the
1131 entire string the user should type at the command line.
1133 @anchor{UI-Independent Output}
1134 @section UI-Independent Output---the @code{ui_out} Functions
1135 @c This section is based on the documentation written by Fernando
1136 @c Nasser <fnasser@redhat.com>.
1138 @cindex @code{ui_out} functions
1139 The @code{ui_out} functions present an abstraction level for the
1140 @value{GDBN} output code. They hide the specifics of different user
1141 interfaces supported by @value{GDBN}, and thus free the programmer
1142 from the need to write several versions of the same code, one each for
1143 every UI, to produce output.
1145 @subsection Overview and Terminology
1147 In general, execution of each @value{GDBN} command produces some sort
1148 of output, and can even generate an input request.
1150 Output can be generated for the following purposes:
1154 to display a @emph{result} of an operation;
1157 to convey @emph{info} or produce side-effects of a requested
1161 to provide a @emph{notification} of an asynchronous event (including
1162 progress indication of a prolonged asynchronous operation);
1165 to display @emph{error messages} (including warnings);
1168 to show @emph{debug data};
1171 to @emph{query} or prompt a user for input (a special case).
1175 This section mainly concentrates on how to build result output,
1176 although some of it also applies to other kinds of output.
1178 Generation of output that displays the results of an operation
1179 involves one or more of the following:
1183 output of the actual data
1186 formatting the output as appropriate for console output, to make it
1187 easily readable by humans
1190 machine oriented formatting--a more terse formatting to allow for easy
1191 parsing by programs which read @value{GDBN}'s output
1194 annotation, whose purpose is to help legacy GUIs to identify interesting
1198 The @code{ui_out} routines take care of the first three aspects.
1199 Annotations are provided by separate annotation routines. Note that use
1200 of annotations for an interface between a GUI and @value{GDBN} is
1203 Output can be in the form of a single item, which we call a @dfn{field};
1204 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
1205 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
1206 header and a body. In a BNF-like form:
1209 @item <table> @expansion{}
1210 @code{<header> <body>}
1211 @item <header> @expansion{}
1212 @code{@{ <column> @}}
1213 @item <column> @expansion{}
1214 @code{<width> <alignment> <title>}
1215 @item <body> @expansion{}
1220 @subsection General Conventions
1222 Most @code{ui_out} routines are of type @code{void}, the exceptions are
1223 @code{ui_out_stream_new} (which returns a pointer to the newly created
1224 object) and the @code{make_cleanup} routines.
1226 The first parameter is always the @code{ui_out} vector object, a pointer
1227 to a @code{struct ui_out}.
1229 The @var{format} parameter is like in @code{printf} family of functions.
1230 When it is present, there must also be a variable list of arguments
1231 sufficient used to satisfy the @code{%} specifiers in the supplied
1234 When a character string argument is not used in a @code{ui_out} function
1235 call, a @code{NULL} pointer has to be supplied instead.
1238 @subsection Table, Tuple and List Functions
1240 @cindex list output functions
1241 @cindex table output functions
1242 @cindex tuple output functions
1243 This section introduces @code{ui_out} routines for building lists,
1244 tuples and tables. The routines to output the actual data items
1245 (fields) are presented in the next section.
1247 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
1248 containing information about an object; a @dfn{list} is a sequence of
1249 fields where each field describes an identical object.
1251 Use the @dfn{table} functions when your output consists of a list of
1252 rows (tuples) and the console output should include a heading. Use this
1253 even when you are listing just one object but you still want the header.
1255 @cindex nesting level in @code{ui_out} functions
1256 Tables can not be nested. Tuples and lists can be nested up to a
1257 maximum of five levels.
1259 The overall structure of the table output code is something like this:
1274 Here is the description of table-, tuple- and list-related @code{ui_out}
1277 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
1278 The function @code{ui_out_table_begin} marks the beginning of the output
1279 of a table. It should always be called before any other @code{ui_out}
1280 function for a given table. @var{nbrofcols} is the number of columns in
1281 the table. @var{nr_rows} is the number of rows in the table.
1282 @var{tblid} is an optional string identifying the table. The string
1283 pointed to by @var{tblid} is copied by the implementation of
1284 @code{ui_out_table_begin}, so the application can free the string if it
1285 was @code{malloc}ed.
1287 The companion function @code{ui_out_table_end}, described below, marks
1288 the end of the table's output.
1291 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
1292 @code{ui_out_table_header} provides the header information for a single
1293 table column. You call this function several times, one each for every
1294 column of the table, after @code{ui_out_table_begin}, but before
1295 @code{ui_out_table_body}.
1297 The value of @var{width} gives the column width in characters. The
1298 value of @var{alignment} is one of @code{left}, @code{center}, and
1299 @code{right}, and it specifies how to align the header: left-justify,
1300 center, or right-justify it. @var{colhdr} points to a string that
1301 specifies the column header; the implementation copies that string, so
1302 column header strings in @code{malloc}ed storage can be freed after the
1306 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
1307 This function delimits the table header from the table body.
1310 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
1311 This function signals the end of a table's output. It should be called
1312 after the table body has been produced by the list and field output
1315 There should be exactly one call to @code{ui_out_table_end} for each
1316 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
1317 will signal an internal error.
1320 The output of the tuples that represent the table rows must follow the
1321 call to @code{ui_out_table_body} and precede the call to
1322 @code{ui_out_table_end}. You build a tuple by calling
1323 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
1324 calls to functions which actually output fields between them.
1326 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
1327 This function marks the beginning of a tuple output. @var{id} points
1328 to an optional string that identifies the tuple; it is copied by the
1329 implementation, and so strings in @code{malloc}ed storage can be freed
1333 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
1334 This function signals an end of a tuple output. There should be exactly
1335 one call to @code{ui_out_tuple_end} for each call to
1336 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
1340 @deftypefun {struct cleanup *} make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1341 This function first opens the tuple and then establishes a cleanup
1342 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
1343 and correct implementation of the non-portable@footnote{The function
1344 cast is not portable ISO C.} code sequence:
1346 struct cleanup *old_cleanup;
1347 ui_out_tuple_begin (uiout, "...");
1348 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
1353 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
1354 This function marks the beginning of a list output. @var{id} points to
1355 an optional string that identifies the list; it is copied by the
1356 implementation, and so strings in @code{malloc}ed storage can be freed
1360 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
1361 This function signals an end of a list output. There should be exactly
1362 one call to @code{ui_out_list_end} for each call to
1363 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1367 @deftypefun {struct cleanup *} make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1368 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1369 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
1370 that will close the list.
1373 @subsection Item Output Functions
1375 @cindex item output functions
1376 @cindex field output functions
1378 The functions described below produce output for the actual data
1379 items, or fields, which contain information about the object.
1381 Choose the appropriate function accordingly to your particular needs.
1383 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1384 This is the most general output function. It produces the
1385 representation of the data in the variable-length argument list
1386 according to formatting specifications in @var{format}, a
1387 @code{printf}-like format string. The optional argument @var{fldname}
1388 supplies the name of the field. The data items themselves are
1389 supplied as additional arguments after @var{format}.
1391 This generic function should be used only when it is not possible to
1392 use one of the specialized versions (see below).
1395 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1396 This function outputs a value of an @code{int} variable. It uses the
1397 @code{"%d"} output conversion specification. @var{fldname} specifies
1398 the name of the field.
1401 @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})
1402 This function outputs a value of an @code{int} variable. It differs from
1403 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1404 @var{fldname} specifies
1405 the name of the field.
1408 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1409 This function outputs an address.
1412 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1413 This function outputs a string using the @code{"%s"} conversion
1417 Sometimes, there's a need to compose your output piece by piece using
1418 functions that operate on a stream, such as @code{value_print} or
1419 @code{fprintf_symbol_filtered}. These functions accept an argument of
1420 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1421 used to store the data stream used for the output. When you use one
1422 of these functions, you need a way to pass their results stored in a
1423 @code{ui_file} object to the @code{ui_out} functions. To this end,
1424 you first create a @code{ui_stream} object by calling
1425 @code{ui_out_stream_new}, pass the @code{stream} member of that
1426 @code{ui_stream} object to @code{value_print} and similar functions,
1427 and finally call @code{ui_out_field_stream} to output the field you
1428 constructed. When the @code{ui_stream} object is no longer needed,
1429 you should destroy it and free its memory by calling
1430 @code{ui_out_stream_delete}.
1432 @deftypefun {struct ui_stream *} ui_out_stream_new (struct ui_out *@var{uiout})
1433 This function creates a new @code{ui_stream} object which uses the
1434 same output methods as the @code{ui_out} object whose pointer is
1435 passed in @var{uiout}. It returns a pointer to the newly created
1436 @code{ui_stream} object.
1439 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1440 This functions destroys a @code{ui_stream} object specified by
1444 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1445 This function consumes all the data accumulated in
1446 @code{streambuf->stream} and outputs it like
1447 @code{ui_out_field_string} does. After a call to
1448 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1449 the stream is still valid and may be used for producing more fields.
1452 @strong{Important:} If there is any chance that your code could bail
1453 out before completing output generation and reaching the point where
1454 @code{ui_out_stream_delete} is called, it is necessary to set up a
1455 cleanup, to avoid leaking memory and other resources. Here's a
1456 skeleton code to do that:
1459 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1460 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1465 If the function already has the old cleanup chain set (for other kinds
1466 of cleanups), you just have to add your cleanup to it:
1469 mybuf = ui_out_stream_new (uiout);
1470 make_cleanup (ui_out_stream_delete, mybuf);
1473 Note that with cleanups in place, you should not call
1474 @code{ui_out_stream_delete} directly, or you would attempt to free the
1477 @subsection Utility Output Functions
1479 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1480 This function skips a field in a table. Use it if you have to leave
1481 an empty field without disrupting the table alignment. The argument
1482 @var{fldname} specifies a name for the (missing) filed.
1485 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1486 This function outputs the text in @var{string} in a way that makes it
1487 easy to be read by humans. For example, the console implementation of
1488 this method filters the text through a built-in pager, to prevent it
1489 from scrolling off the visible portion of the screen.
1491 Use this function for printing relatively long chunks of text around
1492 the actual field data: the text it produces is not aligned according
1493 to the table's format. Use @code{ui_out_field_string} to output a
1494 string field, and use @code{ui_out_message}, described below, to
1495 output short messages.
1498 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1499 This function outputs @var{nspaces} spaces. It is handy to align the
1500 text produced by @code{ui_out_text} with the rest of the table or
1504 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1505 This function produces a formatted message, provided that the current
1506 verbosity level is at least as large as given by @var{verbosity}. The
1507 current verbosity level is specified by the user with the @samp{set
1508 verbositylevel} command.@footnote{As of this writing (April 2001),
1509 setting verbosity level is not yet implemented, and is always returned
1510 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1511 argument more than zero will cause the message to never be printed.}
1514 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1515 This function gives the console output filter (a paging filter) a hint
1516 of where to break lines which are too long. Ignored for all other
1517 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1518 be printed to indent the wrapped text on the next line; it must remain
1519 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1520 explicit newline is produced by one of the other functions. If
1521 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1524 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1525 This function flushes whatever output has been accumulated so far, if
1526 the UI buffers output.
1530 @subsection Examples of Use of @code{ui_out} functions
1532 @cindex using @code{ui_out} functions
1533 @cindex @code{ui_out} functions, usage examples
1534 This section gives some practical examples of using the @code{ui_out}
1535 functions to generalize the old console-oriented code in
1536 @value{GDBN}. The examples all come from functions defined on the
1537 @file{breakpoints.c} file.
1539 This example, from the @code{breakpoint_1} function, shows how to
1542 The original code was:
1545 if (!found_a_breakpoint++)
1547 annotate_breakpoints_headers ();
1550 printf_filtered ("Num ");
1552 printf_filtered ("Type ");
1554 printf_filtered ("Disp ");
1556 printf_filtered ("Enb ");
1560 printf_filtered ("Address ");
1563 printf_filtered ("What\n");
1565 annotate_breakpoints_table ();
1569 Here's the new version:
1572 nr_printable_breakpoints = @dots{};
1575 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1577 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1579 if (nr_printable_breakpoints > 0)
1580 annotate_breakpoints_headers ();
1581 if (nr_printable_breakpoints > 0)
1583 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1584 if (nr_printable_breakpoints > 0)
1586 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1587 if (nr_printable_breakpoints > 0)
1589 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1590 if (nr_printable_breakpoints > 0)
1592 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1595 if (nr_printable_breakpoints > 0)
1597 if (gdbarch_addr_bit (current_gdbarch) <= 32)
1598 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1600 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1602 if (nr_printable_breakpoints > 0)
1604 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1605 ui_out_table_body (uiout);
1606 if (nr_printable_breakpoints > 0)
1607 annotate_breakpoints_table ();
1610 This example, from the @code{print_one_breakpoint} function, shows how
1611 to produce the actual data for the table whose structure was defined
1612 in the above example. The original code was:
1617 printf_filtered ("%-3d ", b->number);
1619 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1620 || ((int) b->type != bptypes[(int) b->type].type))
1621 internal_error ("bptypes table does not describe type #%d.",
1623 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1625 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1627 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1631 This is the new version:
1635 ui_out_tuple_begin (uiout, "bkpt");
1637 ui_out_field_int (uiout, "number", b->number);
1639 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1640 || ((int) b->type != bptypes[(int) b->type].type))
1641 internal_error ("bptypes table does not describe type #%d.",
1643 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1645 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1647 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1651 This example, also from @code{print_one_breakpoint}, shows how to
1652 produce a complicated output field using the @code{print_expression}
1653 functions which requires a stream to be passed. It also shows how to
1654 automate stream destruction with cleanups. The original code was:
1658 print_expression (b->exp, gdb_stdout);
1664 struct ui_stream *stb = ui_out_stream_new (uiout);
1665 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1668 print_expression (b->exp, stb->stream);
1669 ui_out_field_stream (uiout, "what", local_stream);
1672 This example, also from @code{print_one_breakpoint}, shows how to use
1673 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1678 if (b->dll_pathname == NULL)
1679 printf_filtered ("<any library> ");
1681 printf_filtered ("library \"%s\" ", b->dll_pathname);
1688 if (b->dll_pathname == NULL)
1690 ui_out_field_string (uiout, "what", "<any library>");
1691 ui_out_spaces (uiout, 1);
1695 ui_out_text (uiout, "library \"");
1696 ui_out_field_string (uiout, "what", b->dll_pathname);
1697 ui_out_text (uiout, "\" ");
1701 The following example from @code{print_one_breakpoint} shows how to
1702 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1707 if (b->forked_inferior_pid != 0)
1708 printf_filtered ("process %d ", b->forked_inferior_pid);
1715 if (b->forked_inferior_pid != 0)
1717 ui_out_text (uiout, "process ");
1718 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1719 ui_out_spaces (uiout, 1);
1723 Here's an example of using @code{ui_out_field_string}. The original
1728 if (b->exec_pathname != NULL)
1729 printf_filtered ("program \"%s\" ", b->exec_pathname);
1736 if (b->exec_pathname != NULL)
1738 ui_out_text (uiout, "program \"");
1739 ui_out_field_string (uiout, "what", b->exec_pathname);
1740 ui_out_text (uiout, "\" ");
1744 Finally, here's an example of printing an address. The original code:
1748 printf_filtered ("%s ",
1749 hex_string_custom ((unsigned long) b->address, 8));
1756 ui_out_field_core_addr (uiout, "Address", b->address);
1760 @section Console Printing
1769 @cindex @code{libgdb}
1770 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1771 to provide an API to @value{GDBN}'s functionality.
1774 @cindex @code{libgdb}
1775 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1776 better able to support graphical and other environments.
1778 Since @code{libgdb} development is on-going, its architecture is still
1779 evolving. The following components have so far been identified:
1783 Observer - @file{gdb-events.h}.
1785 Builder - @file{ui-out.h}
1787 Event Loop - @file{event-loop.h}
1789 Library - @file{gdb.h}
1792 The model that ties these components together is described below.
1794 @section The @code{libgdb} Model
1796 A client of @code{libgdb} interacts with the library in two ways.
1800 As an observer (using @file{gdb-events}) receiving notifications from
1801 @code{libgdb} of any internal state changes (break point changes, run
1804 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1805 obtain various status values from @value{GDBN}.
1808 Since @code{libgdb} could have multiple clients (e.g., a GUI supporting
1809 the existing @value{GDBN} CLI), those clients must co-operate when
1810 controlling @code{libgdb}. In particular, a client must ensure that
1811 @code{libgdb} is idle (i.e.@: no other client is using @code{libgdb})
1812 before responding to a @file{gdb-event} by making a query.
1814 @section CLI support
1816 At present @value{GDBN}'s CLI is very much entangled in with the core of
1817 @code{libgdb}. Consequently, a client wishing to include the CLI in
1818 their interface needs to carefully co-ordinate its own and the CLI's
1821 It is suggested that the client set @code{libgdb} up to be bi-modal
1822 (alternate between CLI and client query modes). The notes below sketch
1827 The client registers itself as an observer of @code{libgdb}.
1829 The client create and install @code{cli-out} builder using its own
1830 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1831 @code{gdb_stdout} streams.
1833 The client creates a separate custom @code{ui-out} builder that is only
1834 used while making direct queries to @code{libgdb}.
1837 When the client receives input intended for the CLI, it simply passes it
1838 along. Since the @code{cli-out} builder is installed by default, all
1839 the CLI output in response to that command is routed (pronounced rooted)
1840 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1841 At the same time, the client is kept abreast of internal changes by
1842 virtue of being a @code{libgdb} observer.
1844 The only restriction on the client is that it must wait until
1845 @code{libgdb} becomes idle before initiating any queries (using the
1846 client's custom builder).
1848 @section @code{libgdb} components
1850 @subheading Observer - @file{gdb-events.h}
1851 @file{gdb-events} provides the client with a very raw mechanism that can
1852 be used to implement an observer. At present it only allows for one
1853 observer and that observer must, internally, handle the need to delay
1854 the processing of any event notifications until after @code{libgdb} has
1855 finished the current command.
1857 @subheading Builder - @file{ui-out.h}
1858 @file{ui-out} provides the infrastructure necessary for a client to
1859 create a builder. That builder is then passed down to @code{libgdb}
1860 when doing any queries.
1862 @subheading Event Loop - @file{event-loop.h}
1863 @c There could be an entire section on the event-loop
1864 @file{event-loop}, currently non-re-entrant, provides a simple event
1865 loop. A client would need to either plug its self into this loop or,
1866 implement a new event-loop that @value{GDBN} would use.
1868 The event-loop will eventually be made re-entrant. This is so that
1869 @value{GDBN} can better handle the problem of some commands blocking
1870 instead of returning.
1872 @subheading Library - @file{gdb.h}
1873 @file{libgdb} is the most obvious component of this system. It provides
1874 the query interface. Each function is parameterized by a @code{ui-out}
1875 builder. The result of the query is constructed using that builder
1876 before the query function returns.
1883 @cindex @code{value} structure
1884 @value{GDBN} uses @code{struct value}, or @dfn{values}, as an internal
1885 abstraction for the representation of a variety of inferior objects
1886 and @value{GDBN} convenience objects.
1888 Values have an associated @code{struct type}, that describes a virtual
1889 view of the raw data or object stored in or accessed through the
1892 A value is in addition discriminated by its lvalue-ness, given its
1893 @code{enum lval_type} enumeration type:
1895 @cindex @code{lval_type} enumeration, for values.
1897 @item @code{not_lval}
1898 This value is not an lval. It can't be assigned to.
1900 @item @code{lval_memory}
1901 This value represents an object in memory.
1903 @item @code{lval_register}
1904 This value represents an object that lives in a register.
1906 @item @code{lval_internalvar}
1907 Represents the value of an internal variable.
1909 @item @code{lval_internalvar_component}
1910 Represents part of a @value{GDBN} internal variable. E.g., a
1913 @cindex computed values
1914 @item @code{lval_computed}
1915 These are ``computed'' values. They allow creating specialized value
1916 objects for specific purposes, all abstracted away from the core value
1917 support code. The creator of such a value writes specialized
1918 functions to handle the reading and writing to/from the value's
1919 backend data, and optionally, a ``copy operator'' and a
1922 Pointers to these functions are stored in a @code{struct lval_funcs}
1923 instance (declared in @file{value.h}), and passed to the
1924 @code{allocate_computed_value} function, as in the example below.
1928 nil_value_read (struct value *v)
1930 /* This callback reads data from some backend, and stores it in V.
1931 In this case, we always read null data. You'll want to fill in
1932 something more interesting. */
1934 memset (value_contents_all_raw (v),
1936 TYPE_LENGTH (value_type (v)));
1940 nil_value_write (struct value *v, struct value *fromval)
1942 /* Takes the data from FROMVAL and stores it in the backend of V. */
1944 to_oblivion (value_contents_all_raw (fromval),
1946 TYPE_LENGTH (value_type (fromval)));
1949 static struct lval_funcs nil_value_funcs =
1956 make_nil_value (void)
1961 type = make_nils_type ();
1962 v = allocate_computed_value (type, &nil_value_funcs, NULL);
1968 See the implementation of the @code{$_siginfo} convenience variable in
1969 @file{infrun.c} as a real example use of lval_computed.
1974 @chapter Stack Frames
1977 @cindex call stack frame
1978 A frame is a construct that @value{GDBN} uses to keep track of calling
1979 and called functions.
1981 @cindex unwind frame
1982 @value{GDBN}'s frame model, a fresh design, was implemented with the
1983 need to support @sc{dwarf}'s Call Frame Information in mind. In fact,
1984 the term ``unwind'' is taken directly from that specification.
1985 Developers wishing to learn more about unwinders, are encouraged to
1986 read the @sc{dwarf} specification, available from
1987 @url{http://www.dwarfstd.org}.
1989 @findex frame_register_unwind
1990 @findex get_frame_register
1991 @value{GDBN}'s model is that you find a frame's registers by
1992 ``unwinding'' them from the next younger frame. That is,
1993 @samp{get_frame_register} which returns the value of a register in
1994 frame #1 (the next-to-youngest frame), is implemented by calling frame
1995 #0's @code{frame_register_unwind} (the youngest frame). But then the
1996 obvious question is: how do you access the registers of the youngest
1999 @cindex sentinel frame
2000 @findex get_frame_type
2001 @vindex SENTINEL_FRAME
2002 To answer this question, @value{GDBN} has the @dfn{sentinel} frame, the
2003 ``-1st'' frame. Unwinding registers from the sentinel frame gives you
2004 the current values of the youngest real frame's registers. If @var{f}
2005 is a sentinel frame, then @code{get_frame_type (@var{f}) @equiv{}
2008 @section Selecting an Unwinder
2010 @findex frame_unwind_prepend_unwinder
2011 @findex frame_unwind_append_unwinder
2012 The architecture registers a list of frame unwinders (@code{struct
2013 frame_unwind}), using the functions
2014 @code{frame_unwind_prepend_unwinder} and
2015 @code{frame_unwind_append_unwinder}. Each unwinder includes a
2016 sniffer. Whenever @value{GDBN} needs to unwind a frame (to fetch the
2017 previous frame's registers or the current frame's ID), it calls
2018 registered sniffers in order to find one which recognizes the frame.
2019 The first time a sniffer returns non-zero, the corresponding unwinder
2020 is assigned to the frame.
2022 @section Unwinding the Frame ID
2025 Every frame has an associated ID, of type @code{struct frame_id}.
2026 The ID includes the stack base and function start address for
2027 the frame. The ID persists through the entire life of the frame,
2028 including while other called frames are running; it is used to
2029 locate an appropriate @code{struct frame_info} from the cache.
2031 Every time the inferior stops, and at various other times, the frame
2032 cache is flushed. Because of this, parts of @value{GDBN} which need
2033 to keep track of individual frames cannot use pointers to @code{struct
2034 frame_info}. A frame ID provides a stable reference to a frame, even
2035 when the unwinder must be run again to generate a new @code{struct
2036 frame_info} for the same frame.
2038 The frame's unwinder's @code{this_id} method is called to find the ID.
2039 Note that this is different from register unwinding, where the next
2040 frame's @code{prev_register} is called to unwind this frame's
2043 Both stack base and function address are required to identify the
2044 frame, because a recursive function has the same function address for
2045 two consecutive frames and a leaf function may have the same stack
2046 address as its caller. On some platforms, a third address is part of
2047 the ID to further disambiguate frames---for instance, on IA-64
2048 the separate register stack address is included in the ID.
2050 An invalid frame ID (@code{null_frame_id}) returned from the
2051 @code{this_id} method means to stop unwinding after this frame.
2053 @section Unwinding Registers
2055 Each unwinder includes a @code{prev_register} method. This method
2056 takes a frame, an associated cache pointer, and a register number.
2057 It returns a @code{struct value *} describing the requested register,
2058 as saved by this frame. This is the value of the register that is
2059 current in this frame's caller.
2061 The returned value must have the same type as the register. It may
2062 have any lvalue type. In most circumstances one of these routines
2063 will generate the appropriate value:
2066 @item frame_unwind_got_optimized
2067 @findex frame_unwind_got_optimized
2068 This register was not saved.
2070 @item frame_unwind_got_register
2071 @findex frame_unwind_got_register
2072 This register was copied into another register in this frame. This
2073 is also used for unchanged registers; they are ``copied'' into the
2076 @item frame_unwind_got_memory
2077 @findex frame_unwind_got_memory
2078 This register was saved in memory.
2080 @item frame_unwind_got_constant
2081 @findex frame_unwind_got_constant
2082 This register was not saved, but the unwinder can compute the previous
2083 value some other way.
2085 @item frame_unwind_got_address
2086 @findex frame_unwind_got_address
2087 Same as @code{frame_unwind_got_constant}, except that the value is a target
2088 address. This is frequently used for the stack pointer, which is not
2089 explicitly saved but has a known offset from this frame's stack
2090 pointer. For architectures with a flat unified address space, this is
2091 generally the same as @code{frame_unwind_got_constant}.
2094 @node Symbol Handling
2096 @chapter Symbol Handling
2098 Symbols are a key part of @value{GDBN}'s operation. Symbols include
2099 variables, functions, and types.
2101 Symbol information for a large program can be truly massive, and
2102 reading of symbol information is one of the major performance
2103 bottlenecks in @value{GDBN}; it can take many minutes to process it
2104 all. Studies have shown that nearly all the time spent is
2105 computational, rather than file reading.
2107 One of the ways for @value{GDBN} to provide a good user experience is
2108 to start up quickly, taking no more than a few seconds. It is simply
2109 not possible to process all of a program's debugging info in that
2110 time, and so we attempt to handle symbols incrementally. For instance,
2111 we create @dfn{partial symbol tables} consisting of only selected
2112 symbols, and only expand them to full symbol tables when necessary.
2114 @section Symbol Reading
2116 @cindex symbol reading
2117 @cindex reading of symbols
2118 @cindex symbol files
2119 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
2120 file is the file containing the program which @value{GDBN} is
2121 debugging. @value{GDBN} can be directed to use a different file for
2122 symbols (with the @samp{symbol-file} command), and it can also read
2123 more symbols via the @samp{add-file} and @samp{load} commands. In
2124 addition, it may bring in more symbols while loading shared
2127 @findex find_sym_fns
2128 Symbol files are initially opened by code in @file{symfile.c} using
2129 the BFD library (@pxref{Support Libraries}). BFD identifies the type
2130 of the file by examining its header. @code{find_sym_fns} then uses
2131 this identification to locate a set of symbol-reading functions.
2133 @findex add_symtab_fns
2134 @cindex @code{sym_fns} structure
2135 @cindex adding a symbol-reading module
2136 Symbol-reading modules identify themselves to @value{GDBN} by calling
2137 @code{add_symtab_fns} during their module initialization. The argument
2138 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
2139 name (or name prefix) of the symbol format, the length of the prefix,
2140 and pointers to four functions. These functions are called at various
2141 times to process symbol files whose identification matches the specified
2144 The functions supplied by each module are:
2147 @item @var{xyz}_symfile_init(struct sym_fns *sf)
2149 @cindex secondary symbol file
2150 Called from @code{symbol_file_add} when we are about to read a new
2151 symbol file. This function should clean up any internal state (possibly
2152 resulting from half-read previous files, for example) and prepare to
2153 read a new symbol file. Note that the symbol file which we are reading
2154 might be a new ``main'' symbol file, or might be a secondary symbol file
2155 whose symbols are being added to the existing symbol table.
2157 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
2158 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
2159 new symbol file being read. Its @code{private} field has been zeroed,
2160 and can be modified as desired. Typically, a struct of private
2161 information will be @code{malloc}'d, and a pointer to it will be placed
2162 in the @code{private} field.
2164 There is no result from @code{@var{xyz}_symfile_init}, but it can call
2165 @code{error} if it detects an unavoidable problem.
2167 @item @var{xyz}_new_init()
2169 Called from @code{symbol_file_add} when discarding existing symbols.
2170 This function needs only handle the symbol-reading module's internal
2171 state; the symbol table data structures visible to the rest of
2172 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
2173 arguments and no result. It may be called after
2174 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
2175 may be called alone if all symbols are simply being discarded.
2177 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
2179 Called from @code{symbol_file_add} to actually read the symbols from a
2180 symbol-file into a set of psymtabs or symtabs.
2182 @code{sf} points to the @code{struct sym_fns} originally passed to
2183 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
2184 the offset between the file's specified start address and its true
2185 address in memory. @code{mainline} is 1 if this is the main symbol
2186 table being read, and 0 if a secondary symbol file (e.g., shared library
2187 or dynamically loaded file) is being read.@refill
2190 In addition, if a symbol-reading module creates psymtabs when
2191 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
2192 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
2193 from any point in the @value{GDBN} symbol-handling code.
2196 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
2198 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
2199 the psymtab has not already been read in and had its @code{pst->symtab}
2200 pointer set. The argument is the psymtab to be fleshed-out into a
2201 symtab. Upon return, @code{pst->readin} should have been set to 1, and
2202 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
2203 zero if there were no symbols in that part of the symbol file.
2206 @section Partial Symbol Tables
2208 @value{GDBN} has three types of symbol tables:
2211 @cindex full symbol table
2214 Full symbol tables (@dfn{symtabs}). These contain the main
2215 information about symbols and addresses.
2219 Partial symbol tables (@dfn{psymtabs}). These contain enough
2220 information to know when to read the corresponding part of the full
2223 @cindex minimal symbol table
2226 Minimal symbol tables (@dfn{msymtabs}). These contain information
2227 gleaned from non-debugging symbols.
2230 @cindex partial symbol table
2231 This section describes partial symbol tables.
2233 A psymtab is constructed by doing a very quick pass over an executable
2234 file's debugging information. Small amounts of information are
2235 extracted---enough to identify which parts of the symbol table will
2236 need to be re-read and fully digested later, when the user needs the
2237 information. The speed of this pass causes @value{GDBN} to start up very
2238 quickly. Later, as the detailed rereading occurs, it occurs in small
2239 pieces, at various times, and the delay therefrom is mostly invisible to
2241 @c (@xref{Symbol Reading}.)
2243 The symbols that show up in a file's psymtab should be, roughly, those
2244 visible to the debugger's user when the program is not running code from
2245 that file. These include external symbols and types, static symbols and
2246 types, and @code{enum} values declared at file scope.
2248 The psymtab also contains the range of instruction addresses that the
2249 full symbol table would represent.
2251 @cindex finding a symbol
2252 @cindex symbol lookup
2253 The idea is that there are only two ways for the user (or much of the
2254 code in the debugger) to reference a symbol:
2257 @findex find_pc_function
2258 @findex find_pc_line
2260 By its address (e.g., execution stops at some address which is inside a
2261 function in this file). The address will be noticed to be in the
2262 range of this psymtab, and the full symtab will be read in.
2263 @code{find_pc_function}, @code{find_pc_line}, and other
2264 @code{find_pc_@dots{}} functions handle this.
2266 @cindex lookup_symbol
2269 (e.g., the user asks to print a variable, or set a breakpoint on a
2270 function). Global names and file-scope names will be found in the
2271 psymtab, which will cause the symtab to be pulled in. Local names will
2272 have to be qualified by a global name, or a file-scope name, in which
2273 case we will have already read in the symtab as we evaluated the
2274 qualifier. Or, a local symbol can be referenced when we are ``in'' a
2275 local scope, in which case the first case applies. @code{lookup_symbol}
2276 does most of the work here.
2279 The only reason that psymtabs exist is to cause a symtab to be read in
2280 at the right moment. Any symbol that can be elided from a psymtab,
2281 while still causing that to happen, should not appear in it. Since
2282 psymtabs don't have the idea of scope, you can't put local symbols in
2283 them anyway. Psymtabs don't have the idea of the type of a symbol,
2284 either, so types need not appear, unless they will be referenced by
2287 It is a bug for @value{GDBN} to behave one way when only a psymtab has
2288 been read, and another way if the corresponding symtab has been read
2289 in. Such bugs are typically caused by a psymtab that does not contain
2290 all the visible symbols, or which has the wrong instruction address
2293 The psymtab for a particular section of a symbol file (objfile) could be
2294 thrown away after the symtab has been read in. The symtab should always
2295 be searched before the psymtab, so the psymtab will never be used (in a
2296 bug-free environment). Currently, psymtabs are allocated on an obstack,
2297 and all the psymbols themselves are allocated in a pair of large arrays
2298 on an obstack, so there is little to be gained by trying to free them
2299 unless you want to do a lot more work.
2303 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
2305 @cindex fundamental types
2306 These are the fundamental types that @value{GDBN} uses internally. Fundamental
2307 types from the various debugging formats (stabs, ELF, etc) are mapped
2308 into one of these. They are basically a union of all fundamental types
2309 that @value{GDBN} knows about for all the languages that @value{GDBN}
2312 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
2315 Each time @value{GDBN} builds an internal type, it marks it with one
2316 of these types. The type may be a fundamental type, such as
2317 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
2318 which is a pointer to another type. Typically, several @code{FT_*}
2319 types map to one @code{TYPE_CODE_*} type, and are distinguished by
2320 other members of the type struct, such as whether the type is signed
2321 or unsigned, and how many bits it uses.
2323 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
2325 These are instances of type structs that roughly correspond to
2326 fundamental types and are created as global types for @value{GDBN} to
2327 use for various ugly historical reasons. We eventually want to
2328 eliminate these. Note for example that @code{builtin_type_int}
2329 initialized in @file{gdbtypes.c} is basically the same as a
2330 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
2331 an @code{FT_INTEGER} fundamental type. The difference is that the
2332 @code{builtin_type} is not associated with any particular objfile, and
2333 only one instance exists, while @file{c-lang.c} builds as many
2334 @code{TYPE_CODE_INT} types as needed, with each one associated with
2335 some particular objfile.
2337 @section Object File Formats
2338 @cindex object file formats
2342 @cindex @code{a.out} format
2343 The @code{a.out} format is the original file format for Unix. It
2344 consists of three sections: @code{text}, @code{data}, and @code{bss},
2345 which are for program code, initialized data, and uninitialized data,
2348 The @code{a.out} format is so simple that it doesn't have any reserved
2349 place for debugging information. (Hey, the original Unix hackers used
2350 @samp{adb}, which is a machine-language debugger!) The only debugging
2351 format for @code{a.out} is stabs, which is encoded as a set of normal
2352 symbols with distinctive attributes.
2354 The basic @code{a.out} reader is in @file{dbxread.c}.
2359 The COFF format was introduced with System V Release 3 (SVR3) Unix.
2360 COFF files may have multiple sections, each prefixed by a header. The
2361 number of sections is limited.
2363 The COFF specification includes support for debugging. Although this
2364 was a step forward, the debugging information was woefully limited.
2365 For instance, it was not possible to represent code that came from an
2366 included file. GNU's COFF-using configs often use stabs-type info,
2367 encapsulated in special sections.
2369 The COFF reader is in @file{coffread.c}.
2373 @cindex ECOFF format
2374 ECOFF is an extended COFF originally introduced for Mips and Alpha
2377 The basic ECOFF reader is in @file{mipsread.c}.
2381 @cindex XCOFF format
2382 The IBM RS/6000 running AIX uses an object file format called XCOFF.
2383 The COFF sections, symbols, and line numbers are used, but debugging
2384 symbols are @code{dbx}-style stabs whose strings are located in the
2385 @code{.debug} section (rather than the string table). For more
2386 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
2388 The shared library scheme has a clean interface for figuring out what
2389 shared libraries are in use, but the catch is that everything which
2390 refers to addresses (symbol tables and breakpoints at least) needs to be
2391 relocated for both shared libraries and the main executable. At least
2392 using the standard mechanism this can only be done once the program has
2393 been run (or the core file has been read).
2397 @cindex PE-COFF format
2398 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
2399 executables. PE is basically COFF with additional headers.
2401 While BFD includes special PE support, @value{GDBN} needs only the basic
2407 The ELF format came with System V Release 4 (SVR4) Unix. ELF is
2408 similar to COFF in being organized into a number of sections, but it
2409 removes many of COFF's limitations. Debugging info may be either stabs
2410 encapsulated in ELF sections, or more commonly these days, DWARF.
2412 The basic ELF reader is in @file{elfread.c}.
2417 SOM is HP's object file and debug format (not to be confused with IBM's
2418 SOM, which is a cross-language ABI).
2420 The SOM reader is in @file{somread.c}.
2422 @section Debugging File Formats
2424 This section describes characteristics of debugging information that
2425 are independent of the object file format.
2429 @cindex stabs debugging info
2430 @code{stabs} started out as special symbols within the @code{a.out}
2431 format. Since then, it has been encapsulated into other file
2432 formats, such as COFF and ELF.
2434 While @file{dbxread.c} does some of the basic stab processing,
2435 including for encapsulated versions, @file{stabsread.c} does
2440 @cindex COFF debugging info
2441 The basic COFF definition includes debugging information. The level
2442 of support is minimal and non-extensible, and is not often used.
2444 @subsection Mips debug (Third Eye)
2446 @cindex ECOFF debugging info
2447 ECOFF includes a definition of a special debug format.
2449 The file @file{mdebugread.c} implements reading for this format.
2451 @c mention DWARF 1 as a formerly-supported format
2455 @cindex DWARF 2 debugging info
2456 DWARF 2 is an improved but incompatible version of DWARF 1.
2458 The DWARF 2 reader is in @file{dwarf2read.c}.
2460 @subsection Compressed DWARF 2
2462 @cindex Compressed DWARF 2 debugging info
2463 Compressed DWARF 2 is not technically a separate debugging format, but
2464 merely DWARF 2 debug information that has been compressed. In this
2465 format, every object-file section holding DWARF 2 debugging
2466 information is compressed and prepended with a header. (The section
2467 is also typically renamed, so a section called @code{.debug_info} in a
2468 DWARF 2 binary would be called @code{.zdebug_info} in a compressed
2469 DWARF 2 binary.) The header is 12 bytes long:
2473 4 bytes: the literal string ``ZLIB''
2475 8 bytes: the uncompressed size of the section, in big-endian byte
2479 The same reader is used for both compressed an normal DWARF 2 info.
2480 Section decompression is done in @code{zlib_decompress_section} in
2481 @file{dwarf2read.c}.
2485 @cindex DWARF 3 debugging info
2486 DWARF 3 is an improved version of DWARF 2.
2490 @cindex SOM debugging info
2491 Like COFF, the SOM definition includes debugging information.
2493 @section Adding a New Symbol Reader to @value{GDBN}
2495 @cindex adding debugging info reader
2496 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
2497 there is probably little to be done.
2499 If you need to add a new object file format, you must first add it to
2500 BFD. This is beyond the scope of this document.
2502 You must then arrange for the BFD code to provide access to the
2503 debugging symbols. Generally @value{GDBN} will have to call swapping
2504 routines from BFD and a few other BFD internal routines to locate the
2505 debugging information. As much as possible, @value{GDBN} should not
2506 depend on the BFD internal data structures.
2508 For some targets (e.g., COFF), there is a special transfer vector used
2509 to call swapping routines, since the external data structures on various
2510 platforms have different sizes and layouts. Specialized routines that
2511 will only ever be implemented by one object file format may be called
2512 directly. This interface should be described in a file
2513 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
2515 @section Memory Management for Symbol Files
2517 Most memory associated with a loaded symbol file is stored on
2518 its @code{objfile_obstack}. This includes symbols, types,
2519 namespace data, and other information produced by the symbol readers.
2521 Because this data lives on the objfile's obstack, it is automatically
2522 released when the objfile is unloaded or reloaded. Therefore one
2523 objfile must not reference symbol or type data from another objfile;
2524 they could be unloaded at different times.
2526 User convenience variables, et cetera, have associated types. Normally
2527 these types live in the associated objfile. However, when the objfile
2528 is unloaded, those types are deep copied to global memory, so that
2529 the values of the user variables and history items are not lost.
2532 @node Language Support
2534 @chapter Language Support
2536 @cindex language support
2537 @value{GDBN}'s language support is mainly driven by the symbol reader,
2538 although it is possible for the user to set the source language
2541 @value{GDBN} chooses the source language by looking at the extension
2542 of the file recorded in the debug info; @file{.c} means C, @file{.f}
2543 means Fortran, etc. It may also use a special-purpose language
2544 identifier if the debug format supports it, like with DWARF.
2546 @section Adding a Source Language to @value{GDBN}
2548 @cindex adding source language
2549 To add other languages to @value{GDBN}'s expression parser, follow the
2553 @item Create the expression parser.
2555 @cindex expression parser
2556 This should reside in a file @file{@var{lang}-exp.y}. Routines for
2557 building parsed expressions into a @code{union exp_element} list are in
2560 @cindex language parser
2561 Since we can't depend upon everyone having Bison, and YACC produces
2562 parsers that define a bunch of global names, the following lines
2563 @strong{must} be included at the top of the YACC parser, to prevent the
2564 various parsers from defining the same global names:
2567 #define yyparse @var{lang}_parse
2568 #define yylex @var{lang}_lex
2569 #define yyerror @var{lang}_error
2570 #define yylval @var{lang}_lval
2571 #define yychar @var{lang}_char
2572 #define yydebug @var{lang}_debug
2573 #define yypact @var{lang}_pact
2574 #define yyr1 @var{lang}_r1
2575 #define yyr2 @var{lang}_r2
2576 #define yydef @var{lang}_def
2577 #define yychk @var{lang}_chk
2578 #define yypgo @var{lang}_pgo
2579 #define yyact @var{lang}_act
2580 #define yyexca @var{lang}_exca
2581 #define yyerrflag @var{lang}_errflag
2582 #define yynerrs @var{lang}_nerrs
2585 At the bottom of your parser, define a @code{struct language_defn} and
2586 initialize it with the right values for your language. Define an
2587 @code{initialize_@var{lang}} routine and have it call
2588 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
2589 that your language exists. You'll need some other supporting variables
2590 and functions, which will be used via pointers from your
2591 @code{@var{lang}_language_defn}. See the declaration of @code{struct
2592 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
2593 for more information.
2595 @item Add any evaluation routines, if necessary
2597 @cindex expression evaluation routines
2598 @findex evaluate_subexp
2599 @findex prefixify_subexp
2600 @findex length_of_subexp
2601 If you need new opcodes (that represent the operations of the language),
2602 add them to the enumerated type in @file{expression.h}. Add support
2603 code for these operations in the @code{evaluate_subexp} function
2604 defined in the file @file{eval.c}. Add cases
2605 for new opcodes in two functions from @file{parse.c}:
2606 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
2607 the number of @code{exp_element}s that a given operation takes up.
2609 @item Update some existing code
2611 Add an enumerated identifier for your language to the enumerated type
2612 @code{enum language} in @file{defs.h}.
2614 Update the routines in @file{language.c} so your language is included.
2615 These routines include type predicates and such, which (in some cases)
2616 are language dependent. If your language does not appear in the switch
2617 statement, an error is reported.
2619 @vindex current_language
2620 Also included in @file{language.c} is the code that updates the variable
2621 @code{current_language}, and the routines that translate the
2622 @code{language_@var{lang}} enumerated identifier into a printable
2625 @findex _initialize_language
2626 Update the function @code{_initialize_language} to include your
2627 language. This function picks the default language upon startup, so is
2628 dependent upon which languages that @value{GDBN} is built for.
2630 @findex allocate_symtab
2631 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2632 code so that the language of each symtab (source file) is set properly.
2633 This is used to determine the language to use at each stack frame level.
2634 Currently, the language is set based upon the extension of the source
2635 file. If the language can be better inferred from the symbol
2636 information, please set the language of the symtab in the symbol-reading
2639 @findex print_subexp
2640 @findex op_print_tab
2641 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2642 expression opcodes you have added to @file{expression.h}. Also, add the
2643 printed representations of your operators to @code{op_print_tab}.
2645 @item Add a place of call
2648 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2649 @code{parse_exp_1} (defined in @file{parse.c}).
2651 @item Edit @file{Makefile.in}
2653 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2654 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2655 not get linked in, or, worse yet, it may not get @code{tar}red into the
2660 @node Host Definition
2662 @chapter Host Definition
2664 With the advent of Autoconf, it's rarely necessary to have host
2665 definition machinery anymore. The following information is provided,
2666 mainly, as an historical reference.
2668 @section Adding a New Host
2670 @cindex adding a new host
2671 @cindex host, adding
2672 @value{GDBN}'s host configuration support normally happens via Autoconf.
2673 New host-specific definitions should not be needed. Older hosts
2674 @value{GDBN} still use the host-specific definitions and files listed
2675 below, but these mostly exist for historical reasons, and will
2676 eventually disappear.
2679 @item gdb/config/@var{arch}/@var{xyz}.mh
2680 This file is a Makefile fragment that once contained both host and
2681 native configuration information (@pxref{Native Debugging}) for the
2682 machine @var{xyz}. The host configuration information is now handled
2685 Host configuration information included definitions for @code{CC},
2686 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2687 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2689 New host-only configurations do not need this file.
2693 (Files named @file{gdb/config/@var{arch}/xm-@var{xyz}.h} were once
2694 used to define host-specific macros, but were no longer needed and
2695 have all been removed.)
2697 @subheading Generic Host Support Files
2699 @cindex generic host support
2700 There are some ``generic'' versions of routines that can be used by
2704 @cindex remote debugging support
2705 @cindex serial line support
2707 This contains serial line support for Unix systems. It is included by
2708 default on all Unix-like hosts.
2711 This contains serial pipe support for Unix systems. It is included by
2712 default on all Unix-like hosts.
2715 This contains serial line support for 32-bit programs running under
2716 Windows using MinGW.
2719 This contains serial line support for 32-bit programs running under DOS,
2720 using the DJGPP (a.k.a.@: GO32) execution environment.
2722 @cindex TCP remote support
2724 This contains generic TCP support using sockets. It is included by
2725 default on all Unix-like hosts and with MinGW.
2728 @section Host Conditionals
2730 When @value{GDBN} is configured and compiled, various macros are
2731 defined or left undefined, to control compilation based on the
2732 attributes of the host system. While formerly they could be set in
2733 host-specific header files, at present they can be changed only by
2734 setting @code{CFLAGS} when building, or by editing the source code.
2736 These macros and their meanings (or if the meaning is not documented
2737 here, then one of the source files where they are used is indicated)
2741 @item @value{GDBN}INIT_FILENAME
2742 The default name of @value{GDBN}'s initialization file (normally
2745 @item SIGWINCH_HANDLER
2746 If your host defines @code{SIGWINCH}, you can define this to be the name
2747 of a function to be called if @code{SIGWINCH} is received.
2749 @item SIGWINCH_HANDLER_BODY
2750 Define this to expand into code that will define the function named by
2751 the expansion of @code{SIGWINCH_HANDLER}.
2753 @item CRLF_SOURCE_FILES
2754 @cindex DOS text files
2755 Define this if host files use @code{\r\n} rather than @code{\n} as a
2756 line terminator. This will cause source file listings to omit @code{\r}
2757 characters when printing and it will allow @code{\r\n} line endings of files
2758 which are ``sourced'' by gdb. It must be possible to open files in binary
2759 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2761 @item DEFAULT_PROMPT
2763 The default value of the prompt string (normally @code{"(gdb) "}).
2766 @cindex terminal device
2767 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2770 Substitute for isatty, if not available.
2773 Define this if binary files are opened the same way as text files.
2775 @item CC_HAS_LONG_LONG
2776 @cindex @code{long long} data type
2777 Define this if the host C compiler supports @code{long long}. This is set
2778 by the @code{configure} script.
2780 @item PRINTF_HAS_LONG_LONG
2781 Define this if the host can handle printing of long long integers via
2782 the printf format conversion specifier @code{ll}. This is set by the
2783 @code{configure} script.
2785 @item LSEEK_NOT_LINEAR
2786 Define this if @code{lseek (n)} does not necessarily move to byte number
2787 @code{n} in the file. This is only used when reading source files. It
2788 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2791 If defined, this should be one or more tokens, such as @code{volatile},
2792 that can be used in both the declaration and definition of functions to
2793 indicate that they never return. The default is already set correctly
2794 if compiling with GCC. This will almost never need to be defined.
2797 If defined, this should be one or more tokens, such as
2798 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2799 of functions to indicate that they never return. The default is already
2800 set correctly if compiling with GCC. This will almost never need to be
2804 Define this to help placate @code{lint} in some situations.
2807 Define this to override the defaults of @code{__volatile__} or
2812 @node Target Architecture Definition
2814 @chapter Target Architecture Definition
2816 @cindex target architecture definition
2817 @value{GDBN}'s target architecture defines what sort of
2818 machine-language programs @value{GDBN} can work with, and how it works
2821 The target architecture object is implemented as the C structure
2822 @code{struct gdbarch *}. The structure, and its methods, are generated
2823 using the Bourne shell script @file{gdbarch.sh}.
2826 * OS ABI Variant Handling::
2827 * Initialize New Architecture::
2828 * Registers and Memory::
2829 * Pointers and Addresses::
2831 * Register Representation::
2832 * Frame Interpretation::
2833 * Inferior Call Setup::
2834 * Defining Other Architecture Features::
2835 * Adding a New Target::
2838 @node OS ABI Variant Handling
2839 @section Operating System ABI Variant Handling
2840 @cindex OS ABI variants
2842 @value{GDBN} provides a mechanism for handling variations in OS
2843 ABIs. An OS ABI variant may have influence over any number of
2844 variables in the target architecture definition. There are two major
2845 components in the OS ABI mechanism: sniffers and handlers.
2847 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2848 (the architecture may be wildcarded) in an attempt to determine the
2849 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2850 to be @dfn{generic}, while sniffers for a specific architecture are
2851 considered to be @dfn{specific}. A match from a specific sniffer
2852 overrides a match from a generic sniffer. Multiple sniffers for an
2853 architecture/flavour may exist, in order to differentiate between two
2854 different operating systems which use the same basic file format. The
2855 OS ABI framework provides a generic sniffer for ELF-format files which
2856 examines the @code{EI_OSABI} field of the ELF header, as well as note
2857 sections known to be used by several operating systems.
2859 @cindex fine-tuning @code{gdbarch} structure
2860 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2861 selected OS ABI. There may be only one handler for a given OS ABI
2862 for each BFD architecture.
2864 The following OS ABI variants are defined in @file{defs.h}:
2868 @findex GDB_OSABI_UNINITIALIZED
2869 @item GDB_OSABI_UNINITIALIZED
2870 Used for struct gdbarch_info if ABI is still uninitialized.
2872 @findex GDB_OSABI_UNKNOWN
2873 @item GDB_OSABI_UNKNOWN
2874 The ABI of the inferior is unknown. The default @code{gdbarch}
2875 settings for the architecture will be used.
2877 @findex GDB_OSABI_SVR4
2878 @item GDB_OSABI_SVR4
2879 UNIX System V Release 4.
2881 @findex GDB_OSABI_HURD
2882 @item GDB_OSABI_HURD
2883 GNU using the Hurd kernel.
2885 @findex GDB_OSABI_SOLARIS
2886 @item GDB_OSABI_SOLARIS
2889 @findex GDB_OSABI_OSF1
2890 @item GDB_OSABI_OSF1
2891 OSF/1, including Digital UNIX and Compaq Tru64 UNIX.
2893 @findex GDB_OSABI_LINUX
2894 @item GDB_OSABI_LINUX
2895 GNU using the Linux kernel.
2897 @findex GDB_OSABI_FREEBSD_AOUT
2898 @item GDB_OSABI_FREEBSD_AOUT
2899 FreeBSD using the @code{a.out} executable format.
2901 @findex GDB_OSABI_FREEBSD_ELF
2902 @item GDB_OSABI_FREEBSD_ELF
2903 FreeBSD using the ELF executable format.
2905 @findex GDB_OSABI_NETBSD_AOUT
2906 @item GDB_OSABI_NETBSD_AOUT
2907 NetBSD using the @code{a.out} executable format.
2909 @findex GDB_OSABI_NETBSD_ELF
2910 @item GDB_OSABI_NETBSD_ELF
2911 NetBSD using the ELF executable format.
2913 @findex GDB_OSABI_OPENBSD_ELF
2914 @item GDB_OSABI_OPENBSD_ELF
2915 OpenBSD using the ELF executable format.
2917 @findex GDB_OSABI_WINCE
2918 @item GDB_OSABI_WINCE
2921 @findex GDB_OSABI_GO32
2922 @item GDB_OSABI_GO32
2925 @findex GDB_OSABI_IRIX
2926 @item GDB_OSABI_IRIX
2929 @findex GDB_OSABI_INTERIX
2930 @item GDB_OSABI_INTERIX
2931 Interix (Posix layer for MS-Windows systems).
2933 @findex GDB_OSABI_HPUX_ELF
2934 @item GDB_OSABI_HPUX_ELF
2935 HP/UX using the ELF executable format.
2937 @findex GDB_OSABI_HPUX_SOM
2938 @item GDB_OSABI_HPUX_SOM
2939 HP/UX using the SOM executable format.
2941 @findex GDB_OSABI_QNXNTO
2942 @item GDB_OSABI_QNXNTO
2945 @findex GDB_OSABI_CYGWIN
2946 @item GDB_OSABI_CYGWIN
2949 @findex GDB_OSABI_AIX
2955 Here are the functions that make up the OS ABI framework:
2957 @deftypefun {const char *} gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2958 Return the name of the OS ABI corresponding to @var{osabi}.
2961 @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}))
2962 Register the OS ABI handler specified by @var{init_osabi} for the
2963 architecture, machine type and OS ABI specified by @var{arch},
2964 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2965 machine type, which implies the architecture's default machine type,
2969 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2970 Register the OS ABI file sniffer specified by @var{sniffer} for the
2971 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2972 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2973 be generic, and is allowed to examine @var{flavour}-flavoured files for
2977 @deftypefun {enum gdb_osabi} gdbarch_lookup_osabi (bfd *@var{abfd})
2978 Examine the file described by @var{abfd} to determine its OS ABI.
2979 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2983 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2984 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2985 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2986 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2987 architecture, a warning will be issued and the debugging session will continue
2988 with the defaults already established for @var{gdbarch}.
2991 @deftypefun void generic_elf_osabi_sniff_abi_tag_sections (bfd *@var{abfd}, asection *@var{sect}, void *@var{obj})
2992 Helper routine for ELF file sniffers. Examine the file described by
2993 @var{abfd} and look at ABI tag note sections to determine the OS ABI
2994 from the note. This function should be called via
2995 @code{bfd_map_over_sections}.
2998 @node Initialize New Architecture
2999 @section Initializing a New Architecture
3002 * How an Architecture is Represented::
3003 * Looking Up an Existing Architecture::
3004 * Creating a New Architecture::
3007 @node How an Architecture is Represented
3008 @subsection How an Architecture is Represented
3009 @cindex architecture representation
3010 @cindex representation of architecture
3012 Each @code{gdbarch} is associated with a single @sc{bfd} architecture,
3013 via a @code{bfd_arch_@var{arch}} in the @code{bfd_architecture}
3014 enumeration. The @code{gdbarch} is registered by a call to
3015 @code{register_gdbarch_init}, usually from the file's
3016 @code{_initialize_@var{filename}} routine, which will be automatically
3017 called during @value{GDBN} startup. The arguments are a @sc{bfd}
3018 architecture constant and an initialization function.
3020 @findex _initialize_@var{arch}_tdep
3021 @cindex @file{@var{arch}-tdep.c}
3022 A @value{GDBN} description for a new architecture, @var{arch} is created by
3023 defining a global function @code{_initialize_@var{arch}_tdep}, by
3024 convention in the source file @file{@var{arch}-tdep.c}. For example,
3025 in the case of the OpenRISC 1000, this function is called
3026 @code{_initialize_or1k_tdep} and is found in the file
3029 @cindex @file{configure.tgt}
3030 @cindex @code{gdbarch}
3031 @findex gdbarch_register
3032 The resulting object files containing the implementation of the
3033 @code{_initialize_@var{arch}_tdep} function are specified in the @value{GDBN}
3034 @file{configure.tgt} file, which includes a large case statement
3035 pattern matching against the @code{--target} option of the
3036 @code{configure} script. The new @code{struct gdbarch} is created
3037 within the @code{_initialize_@var{arch}_tdep} function by calling
3038 @code{gdbarch_register}:
3041 void gdbarch_register (enum bfd_architecture @var{architecture},
3042 gdbarch_init_ftype *@var{init_func},
3043 gdbarch_dump_tdep_ftype *@var{tdep_dump_func});
3046 The @var{architecture} will identify the unique @sc{bfd} to be
3047 associated with this @code{gdbarch}. The @var{init_func} funciton is
3048 called to create and return the new @code{struct gdbarch}. The
3049 @var{tdep_dump_func} function will dump the target specific details
3050 associated with this architecture.
3052 For example the function @code{_initialize_or1k_tdep} creates its
3053 architecture for 32-bit OpenRISC 1000 architectures by calling:
3056 gdbarch_register (bfd_arch_or32, or1k_gdbarch_init, or1k_dump_tdep);
3059 @node Looking Up an Existing Architecture
3060 @subsection Looking Up an Existing Architecture
3061 @cindex @code{gdbarch} lookup
3063 The initialization function has this prototype:
3066 static struct gdbarch *
3067 @var{arch}_gdbarch_init (struct gdbarch_info @var{info},
3068 struct gdbarch_list *@var{arches})
3071 The @var{info} argument contains parameters used to select the correct
3072 architecture, and @var{arches} is a list of architectures which
3073 have already been created with the same @code{bfd_arch_@var{arch}}
3076 The initialization function should first make sure that @var{info}
3077 is acceptable, and return @code{NULL} if it is not. Then, it should
3078 search through @var{arches} for an exact match to @var{info}, and
3079 return one if found. Lastly, if no exact match was found, it should
3080 create a new architecture based on @var{info} and return it.
3082 @findex gdbarch_list_lookup_by_info
3083 @cindex @code{gdbarch_info}
3084 The lookup is done using @code{gdbarch_list_lookup_by_info}. It is
3085 passed the list of existing architectures, @var{arches}, and the
3086 @code{struct gdbarch_info}, @var{info}, and returns the first matching
3087 architecture it finds, or @code{NULL} if none are found. If an
3088 architecture is found it can be returned as the result from the
3089 initialization function, otherwise a new @code{struct gdbach} will need
3092 The struct gdbarch_info has the following components:
3097 const struct bfd_arch_info *bfd_arch_info;
3100 struct gdbarch_tdep_info *tdep_info;
3101 enum gdb_osabi osabi;
3102 const struct target_desc *target_desc;
3106 @vindex bfd_arch_info
3107 The @code{bfd_arch_info} member holds the key details about the
3108 architecture. The @code{byte_order} member is a value in an
3109 enumeration indicating the endianism. The @code{abfd} member is a
3110 pointer to the full @sc{bfd}, the @code{tdep_info} member is
3111 additional custom target specific information, @code{osabi} identifies
3112 which (if any) of a number of operating specific ABIs are used by this
3113 architecture and the @code{target_desc} member is a set of name-value
3114 pairs with information about register usage in this target.
3116 When the @code{struct gdbarch} initialization function is called, not
3117 all the fields are provided---only those which can be deduced from the
3118 @sc{bfd}. The @code{struct gdbarch_info}, @var{info} is used as a
3119 look-up key with the list of existing architectures, @var{arches} to
3120 see if a suitable architecture already exists. The @var{tdep_info},
3121 @var{osabi} and @var{target_desc} fields may be added before this
3122 lookup to refine the search.
3124 Only information in @var{info} should be used to choose the new
3125 architecture. Historically, @var{info} could be sparse, and
3126 defaults would be collected from the first element on @var{arches}.
3127 However, @value{GDBN} now fills in @var{info} more thoroughly,
3128 so new @code{gdbarch} initialization functions should not take
3129 defaults from @var{arches}.
3131 @node Creating a New Architecture
3132 @subsection Creating a New Architecture
3133 @cindex @code{struct gdbarch} creation
3135 @findex gdbarch_alloc
3136 @cindex @code{gdbarch_tdep} when allocating new @code{gdbarch}
3137 If no architecture is found, then a new architecture must be created,
3138 by calling @code{gdbarch_alloc} using the supplied @code{@w{struct
3139 gdbarch_info}} and any additional custom target specific
3140 information in a @code{struct gdbarch_tdep}. The prototype for
3141 @code{gdbarch_alloc} is:
3144 struct gdbarch *gdbarch_alloc (const struct gdbarch_info *@var{info},
3145 struct gdbarch_tdep *@var{tdep});
3148 @cindex @code{set_gdbarch} functions
3149 @cindex @code{gdbarch} accessor functions
3150 The newly created struct gdbarch must then be populated. Although
3151 there are default values, in most cases they are not what is
3154 For each element, @var{X}, there is are a pair of corresponding accessor
3155 functions, one to set the value of that element,
3156 @code{set_gdbarch_@var{X}}, the second to either get the value of an
3157 element (if it is a variable) or to apply the element (if it is a
3158 function), @code{gdbarch_@var{X}}. Note that both accessor functions
3159 take a pointer to the @code{@w{struct gdbarch}} as first
3160 argument. Populating the new @code{gdbarch} should use the
3161 @code{set_gdbarch} functions.
3163 The following sections identify the main elements that should be set
3164 in this way. This is not the complete list, but represents the
3165 functions and elements that must commonly be specified for a new
3166 architecture. Many of the functions and variables are described in the
3167 header file @file{gdbarch.h}.
3169 This is the main work in defining a new architecture. Implementing the
3170 set of functions to populate the @code{struct gdbarch}.
3172 @cindex @code{gdbarch_tdep} definition
3173 @code{struct gdbarch_tdep} is not defined within @value{GDBN}---it is up
3174 to the user to define this struct if it is needed to hold custom target
3175 information that is not covered by the standard @code{@w{struct
3176 gdbarch}}. For example with the OpenRISC 1000 architecture it is used to
3177 hold the number of matchpoints available in the target (along with other
3180 If there is no additional target specific information, it can be set to
3183 @node Registers and Memory
3184 @section Registers and Memory
3186 @value{GDBN}'s model of the target machine is rather simple.
3187 @value{GDBN} assumes the machine includes a bank of registers and a
3188 block of memory. Each register may have a different size.
3190 @value{GDBN} does not have a magical way to match up with the
3191 compiler's idea of which registers are which; however, it is critical
3192 that they do match up accurately. The only way to make this work is
3193 to get accurate information about the order that the compiler uses,
3194 and to reflect that in the @code{gdbarch_register_name} and related functions.
3196 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
3198 @node Pointers and Addresses
3199 @section Pointers Are Not Always Addresses
3200 @cindex pointer representation
3201 @cindex address representation
3202 @cindex word-addressed machines
3203 @cindex separate data and code address spaces
3204 @cindex spaces, separate data and code address
3205 @cindex address spaces, separate data and code
3206 @cindex code pointers, word-addressed
3207 @cindex converting between pointers and addresses
3208 @cindex D10V addresses
3210 On almost all 32-bit architectures, the representation of a pointer is
3211 indistinguishable from the representation of some fixed-length number
3212 whose value is the byte address of the object pointed to. On such
3213 machines, the words ``pointer'' and ``address'' can be used interchangeably.
3214 However, architectures with smaller word sizes are often cramped for
3215 address space, so they may choose a pointer representation that breaks this
3216 identity, and allows a larger code address space.
3218 @c D10V is gone from sources - more current example?
3220 For example, the Renesas D10V is a 16-bit VLIW processor whose
3221 instructions are 32 bits long@footnote{Some D10V instructions are
3222 actually pairs of 16-bit sub-instructions. However, since you can't
3223 jump into the middle of such a pair, code addresses can only refer to
3224 full 32 bit instructions, which is what matters in this explanation.}.
3225 If the D10V used ordinary byte addresses to refer to code locations,
3226 then the processor would only be able to address 64kb of instructions.
3227 However, since instructions must be aligned on four-byte boundaries, the
3228 low two bits of any valid instruction's byte address are always
3229 zero---byte addresses waste two bits. So instead of byte addresses,
3230 the D10V uses word addresses---byte addresses shifted right two bits---to
3231 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
3234 However, this means that code pointers and data pointers have different
3235 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
3236 @code{0xC020} when used as a data address, but refers to byte address
3237 @code{0x30080} when used as a code address.
3239 (The D10V also uses separate code and data address spaces, which also
3240 affects the correspondence between pointers and addresses, but we're
3241 going to ignore that here; this example is already too long.)
3243 To cope with architectures like this---the D10V is not the only
3244 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
3245 byte numbers, and @dfn{pointers}, which are the target's representation
3246 of an address of a particular type of data. In the example above,
3247 @code{0xC020} is the pointer, which refers to one of the addresses
3248 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
3249 @value{GDBN} provides functions for turning a pointer into an address
3250 and vice versa, in the appropriate way for the current architecture.
3252 Unfortunately, since addresses and pointers are identical on almost all
3253 processors, this distinction tends to bit-rot pretty quickly. Thus,
3254 each time you port @value{GDBN} to an architecture which does
3255 distinguish between pointers and addresses, you'll probably need to
3256 clean up some architecture-independent code.
3258 Here are functions which convert between pointers and addresses:
3260 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
3261 Treat the bytes at @var{buf} as a pointer or reference of type
3262 @var{type}, and return the address it represents, in a manner
3263 appropriate for the current architecture. This yields an address
3264 @value{GDBN} can use to read target memory, disassemble, etc. Note that
3265 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
3268 For example, if the current architecture is the Intel x86, this function
3269 extracts a little-endian integer of the appropriate length from
3270 @var{buf} and returns it. However, if the current architecture is the
3271 D10V, this function will return a 16-bit integer extracted from
3272 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
3274 If @var{type} is not a pointer or reference type, then this function
3275 will signal an internal error.
3278 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
3279 Store the address @var{addr} in @var{buf}, in the proper format for a
3280 pointer of type @var{type} in the current architecture. Note that
3281 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
3284 For example, if the current architecture is the Intel x86, this function
3285 stores @var{addr} unmodified as a little-endian integer of the
3286 appropriate length in @var{buf}. However, if the current architecture
3287 is the D10V, this function divides @var{addr} by four if @var{type} is
3288 a pointer to a function, and then stores it in @var{buf}.
3290 If @var{type} is not a pointer or reference type, then this function
3291 will signal an internal error.
3294 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
3295 Assuming that @var{val} is a pointer, return the address it represents,
3296 as appropriate for the current architecture.
3298 This function actually works on integral values, as well as pointers.
3299 For pointers, it performs architecture-specific conversions as
3300 described above for @code{extract_typed_address}.
3303 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
3304 Create and return a value representing a pointer of type @var{type} to
3305 the address @var{addr}, as appropriate for the current architecture.
3306 This function performs architecture-specific conversions as described
3307 above for @code{store_typed_address}.
3310 Here are two functions which architectures can define to indicate the
3311 relationship between pointers and addresses. These have default
3312 definitions, appropriate for architectures on which all pointers are
3313 simple unsigned byte addresses.
3315 @deftypefun CORE_ADDR gdbarch_pointer_to_address (struct gdbarch *@var{current_gdbarch}, struct type *@var{type}, char *@var{buf})
3316 Assume that @var{buf} holds a pointer of type @var{type}, in the
3317 appropriate format for the current architecture. Return the byte
3318 address the pointer refers to.
3320 This function may safely assume that @var{type} is either a pointer or a
3321 C@t{++} reference type.
3324 @deftypefun void gdbarch_address_to_pointer (struct gdbarch *@var{current_gdbarch}, struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
3325 Store in @var{buf} a pointer of type @var{type} representing the address
3326 @var{addr}, in the appropriate format for the current architecture.
3328 This function may safely assume that @var{type} is either a pointer or a
3329 C@t{++} reference type.
3332 @node Address Classes
3333 @section Address Classes
3334 @cindex address classes
3335 @cindex DW_AT_byte_size
3336 @cindex DW_AT_address_class
3338 Sometimes information about different kinds of addresses is available
3339 via the debug information. For example, some programming environments
3340 define addresses of several different sizes. If the debug information
3341 distinguishes these kinds of address classes through either the size
3342 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
3343 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
3344 following macros should be defined in order to disambiguate these
3345 types within @value{GDBN} as well as provide the added information to
3346 a @value{GDBN} user when printing type expressions.
3348 @deftypefun int gdbarch_address_class_type_flags (struct gdbarch *@var{current_gdbarch}, int @var{byte_size}, int @var{dwarf2_addr_class})
3349 Returns the type flags needed to construct a pointer type whose size
3350 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
3351 This function is normally called from within a symbol reader. See
3352 @file{dwarf2read.c}.
3355 @deftypefun {char *} gdbarch_address_class_type_flags_to_name (struct gdbarch *@var{current_gdbarch}, int @var{type_flags})
3356 Given the type flags representing an address class qualifier, return
3359 @deftypefun int gdbarch_address_class_name_to_type_flags (struct gdbarch *@var{current_gdbarch}, int @var{name}, int *@var{type_flags_ptr})
3360 Given an address qualifier name, set the @code{int} referenced by @var{type_flags_ptr} to the type flags
3361 for that address class qualifier.
3364 Since the need for address classes is rather rare, none of
3365 the address class functions are defined by default. Predicate
3366 functions are provided to detect when they are defined.
3368 Consider a hypothetical architecture in which addresses are normally
3369 32-bits wide, but 16-bit addresses are also supported. Furthermore,
3370 suppose that the @w{DWARF 2} information for this architecture simply
3371 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
3372 of these "short" pointers. The following functions could be defined
3373 to implement the address class functions:
3376 somearch_address_class_type_flags (int byte_size,
3377 int dwarf2_addr_class)
3380 return TYPE_FLAG_ADDRESS_CLASS_1;
3386 somearch_address_class_type_flags_to_name (int type_flags)
3388 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
3395 somearch_address_class_name_to_type_flags (char *name,
3396 int *type_flags_ptr)
3398 if (strcmp (name, "short") == 0)
3400 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
3408 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
3409 to indicate the presence of one of these ``short'' pointers. For
3410 example if the debug information indicates that @code{short_ptr_var} is
3411 one of these short pointers, @value{GDBN} might show the following
3415 (gdb) ptype short_ptr_var
3416 type = int * @@short
3420 @node Register Representation
3421 @section Register Representation
3424 * Raw and Cooked Registers::
3425 * Register Architecture Functions & Variables::
3426 * Register Information Functions::
3427 * Register and Memory Data::
3428 * Register Caching::
3431 @node Raw and Cooked Registers
3432 @subsection Raw and Cooked Registers
3433 @cindex raw register representation
3434 @cindex cooked register representation
3435 @cindex representations, raw and cooked registers
3437 @value{GDBN} considers registers to be a set with members numbered
3438 linearly from 0 upwards. The first part of that set corresponds to real
3439 physical registers, the second part to any @dfn{pseudo-registers}.
3440 Pseudo-registers have no independent physical existence, but are useful
3441 representations of information within the architecture. For example the
3442 OpenRISC 1000 architecture has up to 32 general purpose registers, which
3443 are typically represented as 32-bit (or 64-bit) integers. However the
3444 GPRs are also used as operands to the floating point operations, and it
3445 could be convenient to define a set of pseudo-registers, to show the
3446 GPRs represented as floating point values.
3448 For any architecture, the implementer will decide on a mapping from
3449 hardware to @value{GDBN} register numbers. The registers corresponding to real
3450 hardware are referred to as @dfn{raw} registers, the remaining registers are
3451 @dfn{pseudo-registers}. The total register set (raw and pseudo) is called
3452 the @dfn{cooked} register set.
3455 @node Register Architecture Functions & Variables
3456 @subsection Functions and Variables Specifying the Register Architecture
3457 @cindex @code{gdbarch} register architecture functions
3459 These @code{struct gdbarch} functions and variables specify the number
3460 and type of registers in the architecture.
3462 @deftypefn {Architecture Function} CORE_ADDR read_pc (struct regcache *@var{regcache})
3464 @deftypefn {Architecture Function} void write_pc (struct regcache *@var{regcache}, CORE_ADDR @var{val})
3466 Read or write the program counter. The default value of both
3467 functions is @code{NULL} (no function available). If the program
3468 counter is just an ordinary register, it can be specified in
3469 @code{struct gdbarch} instead (see @code{pc_regnum} below) and it will
3470 be read or written using the standard routines to access registers. This
3471 function need only be specified if the program counter is not an
3474 Any register information can be obtained using the supplied register
3475 cache, @var{regcache}. @xref{Register Caching, , Register Caching}.
3479 @deftypefn {Architecture Function} void pseudo_register_read (struct gdbarch *@var{gdbarch}, struct regcache *@var{regcache}, int @var{regnum}, const gdb_byte *@var{buf})
3481 @deftypefn {Architecture Function} void pseudo_register_write (struct gdbarch *@var{gdbarch}, struct regcache *@var{regcache}, int @var{regnum}, const gdb_byte *@var{buf})
3483 These functions should be defined if there are any pseudo-registers.
3484 The default value is @code{NULL}. @var{regnum} is the number of the
3485 register to read or write (which will be a @dfn{cooked} register
3486 number) and @var{buf} is the buffer where the value read will be
3487 placed, or from which the value to be written will be taken. The
3488 value in the buffer may be converted to or from a signed or unsigned
3489 integral value using one of the utility functions (@pxref{Register and
3490 Memory Data, , Using Different Register and Memory Data
3493 The access should be for the specified architecture,
3494 @var{gdbarch}. Any register information can be obtained using the
3495 supplied register cache, @var{regcache}. @xref{Register Caching, ,
3500 @deftypevr {Architecture Variable} int sp_regnum
3502 @cindex stack pointer
3505 This specifies the register holding the stack pointer, which may be a
3506 raw or pseudo-register. It defaults to -1 (not defined), but it is an
3507 error for it not to be defined.
3509 The value of the stack pointer register can be accessed withing
3510 @value{GDBN} as the variable @kbd{$sp}.
3514 @deftypevr {Architecture Variable} int pc_regnum
3516 @cindex program counter
3519 This specifies the register holding the program counter, which may be a
3520 raw or pseudo-register. It defaults to -1 (not defined). If
3521 @code{pc_regnum} is not defined, then the functions @code{read_pc} and
3522 @code{write_pc} (see above) must be defined.
3524 The value of the program counter (whether defined as a register, or
3525 through @code{read_pc} and @code{write_pc}) can be accessed withing
3526 @value{GDBN} as the variable @kbd{$pc}.
3530 @deftypevr {Architecture Variable} int ps_regnum
3532 @cindex processor status register
3533 @cindex status register
3536 This specifies the register holding the processor status (often called
3537 the status register), which may be a raw or pseudo-register. It
3538 defaults to -1 (not defined).
3540 If defined, the value of this register can be accessed withing
3541 @value{GDBN} as the variable @kbd{$ps}.
3545 @deftypevr {Architecture Variable} int fp0_regnum
3547 @cindex first floating point register
3549 This specifies the first floating point register. It defaults to
3550 0. @code{fp0_regnum} is not needed unless the target offers support
3555 @node Register Information Functions
3556 @subsection Functions Giving Register Information
3557 @cindex @code{gdbarch} register information functions
3559 These functions return information about registers.
3561 @deftypefn {Architecture Function} {const char *} register_name (struct gdbarch *@var{gdbarch}, int @var{regnum})
3563 This function should convert a register number (raw or pseudo) to a
3564 register name (as a C @code{const char *}). This is used both to
3565 determine the name of a register for output and to work out the meaning
3566 of any register names used as input. The function may also return
3567 @code{NULL}, to indicate that @var{regnum} is not a valid register.
3569 For example with the OpenRISC 1000, @value{GDBN} registers 0-31 are the
3570 General Purpose Registers, register 32 is the program counter and
3571 register 33 is the supervision register (i.e.@: the processor status
3572 register), which map to the strings @code{"gpr00"} through
3573 @code{"gpr31"}, @code{"pc"} and @code{"sr"} respectively. This means
3574 that the @value{GDBN} command @kbd{print $gpr5} should print the value of
3575 the OR1K general purpose register 5@footnote{
3576 @cindex frame pointer
3578 Historically, @value{GDBN} always had a concept of a frame pointer
3579 register, which could be accessed via the @value{GDBN} variable,
3580 @kbd{$fp}. That concept is now deprecated, recognizing that not all
3581 architectures have a frame pointer. However if an architecture does
3582 have a frame pointer register, and defines a register or
3583 pseudo-register with the name @code{"fp"}, then that register will be
3584 used as the value of the @kbd{$fp} variable.}.
3586 The default value for this function is @code{NULL}, meaning
3587 undefined. It should always be defined.
3589 The access should be for the specified architecture, @var{gdbarch}.
3593 @deftypefn {Architecture Function} {struct type *} register_type (struct gdbarch *@var{gdbarch}, int @var{regnum})
3595 Given a register number, this function identifies the type of data it
3596 may be holding, specified as a @code{struct type}. @value{GDBN} allows
3597 creation of arbitrary types, but a number of built in types are
3598 provided (@code{builtin_type_void}, @code{builtin_type_int32} etc),
3599 together with functions to derive types from these.
3601 Typically the program counter will have a type of ``pointer to
3602 function'' (it points to code), the frame pointer and stack pointer
3603 will have types of ``pointer to void'' (they point to data on the stack)
3604 and all other integer registers will have a type of 32-bit integer or
3607 This information guides the formatting when displaying register
3608 information. The default value is @code{NULL} meaning no information is
3609 available to guide formatting when displaying registers.
3613 @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})
3615 Define this function to print out one or all of the registers for the
3616 @value{GDBN} @kbd{info registers} command. The default value is the
3617 function @code{default_print_registers_info}, which uses the register
3618 type information (see @code{register_type} above) to determine how each
3619 register should be printed. Define a custom version of this function
3620 for fuller control over how the registers are displayed.
3622 The access should be for the specified architecture, @var{gdbarch},
3623 with output to the the file specified by the User Interface
3624 Independent Output file handle, @var{file} (@pxref{UI-Independent
3625 Output, , UI-Independent Output---the @code{ui_out}
3628 The registers should show their values in the frame specified by
3629 @var{frame}. If @var{regnum} is -1 and @var{all} is zero, then all
3630 the ``significant'' registers should be shown (the implementer should
3631 decide which registers are ``significant''). Otherwise only the value of
3632 the register specified by @var{regnum} should be output. If
3633 @var{regnum} is -1 and @var{all} is non-zero (true), then the value of
3634 all registers should be shown.
3636 By default @code{default_print_registers_info} prints one register per
3637 line, and if @var{all} is zero omits floating-point registers.
3641 @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})
3643 Define this function to provide output about the floating point unit and
3644 registers for the @value{GDBN} @kbd{info float} command respectively.
3645 The default value is @code{NULL} (not defined), meaning no information
3648 The @var{gdbarch} and @var{file} and @var{frame} arguments have the same
3649 meaning as in the @code{print_registers_info} function above. The string
3650 @var{args} contains any supplementary arguments to the @kbd{info float}
3653 Define this function if the target supports floating point operations.
3657 @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})
3659 Define this function to provide output about the vector unit and
3660 registers for the @value{GDBN} @kbd{info vector} command respectively.
3661 The default value is @code{NULL} (not defined), meaning no information
3664 The @var{gdbarch}, @var{file} and @var{frame} arguments have the
3665 same meaning as in the @code{print_registers_info} function above. The
3666 string @var{args} contains any supplementary arguments to the @kbd{info
3669 Define this function if the target supports vector operations.
3673 @deftypefn {Architecture Function} int register_reggroup_p (struct gdbarch *@var{gdbarch}, int @var{regnum}, struct reggroup *@var{group})
3675 @value{GDBN} groups registers into different categories (general,
3676 vector, floating point etc). This function, given a register,
3677 @var{regnum}, and group, @var{group}, returns 1 (true) if the register
3678 is in the group and 0 (false) otherwise.
3680 The information should be for the specified architecture,
3683 The default value is the function @code{default_register_reggroup_p}
3684 which will do a reasonable job based on the type of the register (see
3685 the function @code{register_type} above), with groups for general
3686 purpose registers, floating point registers, vector registers and raw
3687 (i.e not pseudo) registers.
3691 @node Register and Memory Data
3692 @subsection Using Different Register and Memory Data Representations
3693 @cindex register representation
3694 @cindex memory representation
3695 @cindex representations, register and memory
3696 @cindex register data formats, converting
3697 @cindex @code{struct value}, converting register contents to
3699 Some architectures have different representations of data objects,
3700 depending whether the object is held in a register or memory. For
3706 The Alpha architecture can represent 32 bit integer values in
3707 floating-point registers.
3710 The x86 architecture supports 80-bit floating-point registers. The
3711 @code{long double} data type occupies 96 bits in memory but only 80
3712 bits when stored in a register.
3716 In general, the register representation of a data type is determined by
3717 the architecture, or @value{GDBN}'s interface to the architecture, while
3718 the memory representation is determined by the Application Binary
3721 For almost all data types on almost all architectures, the two
3722 representations are identical, and no special handling is needed.
3723 However, they do occasionally differ. An architecture may define the
3724 following @code{struct gdbarch} functions to request conversions
3725 between the register and memory representations of a data type:
3727 @deftypefn {Architecture Function} int gdbarch_convert_register_p (struct gdbarch *@var{gdbarch}, int @var{reg})
3729 Return non-zero (true) if the representation of a data value stored in
3730 this register may be different to the representation of that same data
3731 value when stored in memory. The default value is @code{NULL}
3734 If this function is defined and returns non-zero, the @code{struct
3735 gdbarch} functions @code{gdbarch_register_to_value} and
3736 @code{gdbarch_value_to_register} (see below) should be used to perform
3737 any necessary conversion.
3739 If defined, this function should return zero for the register's native
3740 type, when no conversion is necessary.
3743 @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})
3745 Convert the value of register number @var{reg} to a data object of
3746 type @var{type}. The buffer at @var{from} holds the register's value
3747 in raw format; the converted value should be placed in the buffer at
3751 @emph{Note:} @code{gdbarch_register_to_value} and
3752 @code{gdbarch_value_to_register} take their @var{reg} and @var{type}
3753 arguments in different orders.
3756 @code{gdbarch_register_to_value} should only be used with registers
3757 for which the @code{gdbarch_convert_register_p} function returns a
3762 @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})
3764 Convert a data value of type @var{type} to register number @var{reg}'
3768 @emph{Note:} @code{gdbarch_register_to_value} and
3769 @code{gdbarch_value_to_register} take their @var{reg} and @var{type}
3770 arguments in different orders.
3773 @code{gdbarch_value_to_register} should only be used with registers
3774 for which the @code{gdbarch_convert_register_p} function returns a
3779 @node Register Caching
3780 @subsection Register Caching
3781 @cindex register caching
3783 Caching of registers is used, so that the target does not need to be
3784 accessed and reanalyzed multiple times for each register in
3785 circumstances where the register value cannot have changed.
3787 @cindex @code{struct regcache}
3788 @value{GDBN} provides @code{struct regcache}, associated with a
3789 particular @code{struct gdbarch} to hold the cached values of the raw
3790 registers. A set of functions is provided to access both the raw
3791 registers (with @code{raw} in their name) and the full set of cooked
3792 registers (with @code{cooked} in their name). Functions are provided
3793 to ensure the register cache is kept synchronized with the values of
3794 the actual registers in the target.
3796 Accessing registers through the @code{struct regcache} routines will
3797 ensure that the appropriate @code{struct gdbarch} functions are called
3798 when necessary to access the underlying target architecture. In general
3799 users should use the @dfn{cooked} functions, since these will map to the
3800 @dfn{raw} functions automatically as appropriate.
3802 @findex regcache_cooked_read
3803 @findex regcache_cooked_write
3804 @cindex @code{gdb_byte}
3805 @findex regcache_cooked_read_signed
3806 @findex regcache_cooked_read_unsigned
3807 @findex regcache_cooked_write_signed
3808 @findex regcache_cooked_write_unsigned
3809 The two key functions are @code{regcache_cooked_read} and
3810 @code{regcache_cooked_write} which read or write a register from or to
3811 a byte buffer (type @code{gdb_byte *}). For convenience the wrapper
3812 functions @code{regcache_cooked_read_signed},
3813 @code{regcache_cooked_read_unsigned},
3814 @code{regcache_cooked_write_signed} and
3815 @code{regcache_cooked_write_unsigned} are provided, which read or
3816 write the value using the buffer and convert to or from an integral
3817 value as appropriate.
3819 @node Frame Interpretation
3820 @section Frame Interpretation
3823 * All About Stack Frames::
3824 * Frame Handling Terminology::
3826 * Functions and Variable to Analyze Frames::
3827 * Functions to Access Frame Data::
3828 * Analyzing Stacks---Frame Sniffers::
3831 @node All About Stack Frames
3832 @subsection All About Stack Frames
3834 @value{GDBN} needs to understand the stack on which local (automatic)
3835 variables are stored. The area of the stack containing all the local
3836 variables for a function invocation is known as the @dfn{stack frame}
3837 for that function (or colloquially just as the @dfn{frame}). In turn the
3838 function that called the function will have its stack frame, and so on
3839 back through the chain of functions that have been called.
3841 Almost all architectures have one register dedicated to point to the
3842 end of the stack (the @dfn{stack pointer}). Many have a second register
3843 which points to the start of the currently active stack frame (the
3844 @dfn{frame pointer}). The specific arrangements for an architecture are
3845 a key part of the ABI.
3847 A diagram helps to explain this. Here is a simple program to compute
3860 return n * fact (n - 1);
3868 for (i = 0; i < 10; i++)
3871 printf ("%d! = %d\n", i, f);
3876 Consider the state of the stack when the code reaches line 6 after the
3877 main program has called @code{fact@w{ }(3)}. The chain of function
3878 calls will be @code{main ()}, @code{fact@w{ }(3)}, @code{fact@w{
3879 }(2)}, @code{@w{fact (1)}} and @code{fact@w{ }(0)}.
3881 In this illustration the stack is falling (as used for example by the
3882 OpenRISC 1000 ABI). The stack pointer (SP) is at the end of the stack
3883 (lowest address) and the frame pointer (FP) is at the highest address
3884 in the current stack frame. The following diagram shows how the stack
3887 @center @image{stack_frame,14cm}
3889 In each stack frame, offset 0 from the stack pointer is the frame
3890 pointer of the previous frame and offset 4 (this is illustrating a
3891 32-bit architecture) from the stack pointer is the return address.
3892 Local variables are indexed from the frame pointer, with negative
3893 indexes. In the function @code{fact}, offset -4 from the frame
3894 pointer is the argument @var{n}. In the @code{main} function, offset
3895 -4 from the frame pointer is the local variable @var{i} and offset -8
3896 from the frame pointer is the local variable @var{f}@footnote{This is
3897 a simplified example for illustrative purposes only. Good optimizing
3898 compilers would not put anything on the stack for such simple
3899 functions. Indeed they might eliminate the recursion and use of the
3902 It is very easy to get confused when examining stacks. @value{GDBN}
3903 has terminology it uses rigorously throughout. The stack frame of the
3904 function currently executing, or where execution stopped is numbered
3905 zero. In this example frame #0 is the stack frame of the call to
3906 @code{fact@w{ }(0)}. The stack frame of its calling function
3907 (@code{fact@w{ }(1)} in this case) is numbered #1 and so on back
3908 through the chain of calls.
3910 The main @value{GDBN} data structure describing frames is
3911 @code{@w{struct frame_info}}. It is not used directly, but only via
3912 its accessor functions. @code{frame_info} includes information about
3913 the registers in the frame and a pointer to the code of the function
3914 with which the frame is associated. The entire stack is represented as
3915 a linked list of @code{frame_info} structs.
3917 @node Frame Handling Terminology
3918 @subsection Frame Handling Terminology
3920 It is easy to get confused when referencing stack frames. @value{GDBN}
3921 uses some precise terminology.
3927 @cindex stack frame, definition of THIS frame
3928 @cindex frame, definition of THIS frame
3929 @dfn{THIS} frame is the frame currently under consideration.
3933 @cindex stack frame, definition of NEXT frame
3934 @cindex frame, definition of NEXT frame
3935 The @dfn{NEXT} frame, also sometimes called the inner or newer frame is the
3936 frame of the function called by the function of THIS frame.
3939 @cindex PREVIOUS frame
3940 @cindex stack frame, definition of PREVIOUS frame
3941 @cindex frame, definition of PREVIOUS frame
3942 The @dfn{PREVIOUS} frame, also sometimes called the outer or older frame is
3943 the frame of the function which called the function of THIS frame.
3947 So in the example in the previous section (@pxref{All About Stack
3948 Frames, , All About Stack Frames}), if THIS frame is #3 (the call to
3949 @code{fact@w{ }(3)}), the NEXT frame is frame #2 (the call to
3950 @code{fact@w{ }(2)}) and the PREVIOUS frame is frame #4 (the call to
3951 @code{main@w{ }()}).
3953 @cindex innermost frame
3954 @cindex stack frame, definition of innermost frame
3955 @cindex frame, definition of innermost frame
3956 The @dfn{innermost} frame is the frame of the current executing
3957 function, or where the program stopped, in this example, in the middle
3958 of the call to @code{@w{fact (0))}}. It is always numbered frame #0.
3960 @cindex base of a frame
3961 @cindex stack frame, definition of base of a frame
3962 @cindex frame, definition of base of a frame
3963 The @dfn{base} of a frame is the address immediately before the start
3964 of the NEXT frame. For a stack which grows down in memory (a
3965 @dfn{falling} stack) this will be the lowest address and for a stack
3966 which grows up in memory (a @dfn{rising} stack) this will be the
3967 highest address in the frame.
3969 @value{GDBN} functions to analyze the stack are typically given a
3970 pointer to the NEXT frame to determine information about THIS
3971 frame. Information about THIS frame includes data on where the
3972 registers of the PREVIOUS frame are stored in this stack frame. In
3973 this example the frame pointer of the PREVIOUS frame is stored at
3974 offset 0 from the stack pointer of THIS frame.
3977 @cindex stack frame, definition of unwinding
3978 @cindex frame, definition of unwinding
3979 The process whereby a function is given a pointer to the NEXT
3980 frame to work out information about THIS frame is referred to as
3981 @dfn{unwinding}. The @value{GDBN} functions involved in this typically
3982 include unwind in their name.
3985 @cindex stack frame, definition of sniffing
3986 @cindex frame, definition of sniffing
3987 The process of analyzing a target to determine the information that
3988 should go in struct frame_info is called @dfn{sniffing}. The functions
3989 that carry this out are called sniffers and typically include sniffer
3990 in their name. More than one sniffer may be required to extract all
3991 the information for a particular frame.
3993 @cindex sentinel frame
3994 @cindex stack frame, definition of sentinel frame
3995 @cindex frame, definition of sentinel frame
3996 Because so many functions work using the NEXT frame, there is an issue
3997 about addressing the innermost frame---it has no NEXT frame. To solve
3998 this @value{GDBN} creates a dummy frame #-1, known as the
3999 @dfn{sentinel} frame.
4001 @node Prologue Caches
4002 @subsection Prologue Caches
4004 @cindex function prologue
4005 @cindex prologue of a function
4006 All the frame sniffing functions typically examine the code at the
4007 start of the corresponding function, to determine the state of
4008 registers. The ABI will save old values and set new values of key
4009 registers at the start of each function in what is known as the
4010 function @dfn{prologue}.
4012 @cindex prologue cache
4013 For any particular stack frame this data does not change, so all the
4014 standard unwinding functions, in addition to receiving a pointer to
4015 the NEXT frame as their first argument, receive a pointer to a
4016 @dfn{prologue cache} as their second argument. This can be used to store
4017 values associated with a particular frame, for reuse on subsequent
4018 calls involving the same frame.
4020 It is up to the user to define the structure used (it is a
4021 @code{void@w{ }*} pointer) and arrange allocation and deallocation of
4022 storage. However for general use, @value{GDBN} provides
4023 @code{@w{struct trad_frame_cache}}, with a set of accessor
4024 routines. This structure holds the stack and code address of
4025 THIS frame, the base address of the frame, a pointer to the
4026 struct @code{frame_info} for the NEXT frame and details of
4027 where the registers of the PREVIOUS frame may be found in THIS
4030 Typically the first time any sniffer function is called with NEXT
4031 frame, the prologue sniffer for THIS frame will be @code{NULL}. The
4032 sniffer will analyze the frame, allocate a prologue cache structure
4033 and populate it. Subsequent calls using the same NEXT frame will
4034 pass in this prologue cache, so the data can be returned with no
4035 additional analysis.
4037 @node Functions and Variable to Analyze Frames
4038 @subsection Functions and Variable to Analyze Frames
4040 These struct @code{gdbarch} functions and variable should be defined
4041 to provide analysis of the stack frame and allow it to be adjusted as
4044 @deftypefn {Architecture Function} CORE_ADDR skip_prologue (struct gdbarch *@var{gdbarch}, CORE_ADDR @var{pc})
4046 The prologue of a function is the code at the beginning of the
4047 function which sets up the stack frame, saves the return address
4048 etc. The code representing the behavior of the function starts after
4051 This function skips past the prologue of a function if the program
4052 counter, @var{pc}, is within the prologue of a function. The result is
4053 the program counter immediately after the prologue. With modern
4054 optimizing compilers, this may be a far from trivial exercise. However
4055 the required information may be within the binary as DWARF2 debugging
4056 information, making the job much easier.
4058 The default value is @code{NULL} (not defined). This function should always
4059 be provided, but can take advantage of DWARF2 debugging information,
4060 if that is available.
4064 @deftypefn {Architecture Function} int inner_than (CORE_ADDR @var{lhs}, CORE_ADDR @var{rhs})
4065 @findex core_addr_lessthan
4066 @findex core_addr_greaterthan
4068 Given two frame or stack pointers, return non-zero (true) if the first
4069 represents the @dfn{inner} stack frame and 0 (false) otherwise. This
4070 is used to determine whether the target has a stack which grows up in
4071 memory (rising stack) or grows down in memory (falling stack).
4072 @xref{All About Stack Frames, , All About Stack Frames}, for an
4073 explanation of @dfn{inner} frames.
4075 The default value of this function is @code{NULL} and it should always
4076 be defined. However for almost all architectures one of the built-in
4077 functions can be used: @code{core_addr_lessthan} (for stacks growing
4078 down in memory) or @code{core_addr_greaterthan} (for stacks growing up
4083 @anchor{frame_align}
4084 @deftypefn {Architecture Function} CORE_ADDR frame_align (struct gdbarch *@var{gdbarch}, CORE_ADDR @var{address})
4088 The architecture may have constraints on how its frames are
4089 aligned. For example the OpenRISC 1000 ABI requires stack frames to be
4090 double-word aligned, but 32-bit versions of the architecture allocate
4091 single-word values to the stack. Thus extra padding may be needed at
4092 the end of a stack frame.
4094 Given a proposed address for the stack pointer, this function
4095 returns a suitably aligned address (by expanding the stack frame).
4097 The default value is @code{NULL} (undefined). This function should be defined
4098 for any architecture where it is possible the stack could become
4099 misaligned. The utility functions @code{align_down} (for falling
4100 stacks) and @code{align_up} (for rising stacks) will facilitate the
4101 implementation of this function.
4105 @deftypevr {Architecture Variable} int frame_red_zone_size
4107 Some ABIs reserve space beyond the end of the stack for use by leaf
4108 functions without prologue or epilogue or by exception handlers (for
4109 example the OpenRISC 1000).
4111 This is known as a @dfn{red zone} (AMD terminology). The @sc{amd64}
4112 (nee x86-64) ABI documentation refers to the @dfn{red zone} when
4113 describing this scratch area.
4115 The default value is 0. Set this field if the architecture has such a
4116 red zone. The value must be aligned as required by the ABI (see
4117 @code{frame_align} above for an explanation of stack frame alignment).
4121 @node Functions to Access Frame Data
4122 @subsection Functions to Access Frame Data
4124 These functions provide access to key registers and arguments in the
4127 @deftypefn {Architecture Function} CORE_ADDR unwind_pc (struct gdbarch *@var{gdbarch}, struct frame_info *@var{next_frame})
4129 This function is given a pointer to the NEXT stack frame (@pxref{All
4130 About Stack Frames, , All About Stack Frames}, for how frames are
4131 represented) and returns the value of the program counter in the
4132 PREVIOUS frame (i.e.@: the frame of the function that called THIS
4133 one). This is commonly referred to as the @dfn{return address}.
4135 The implementation, which must be frame agnostic (work with any frame),
4136 is typically no more than:
4140 pc = frame_unwind_register_unsigned (next_frame, @var{ARCH}_PC_REGNUM);
4141 return gdbarch_addr_bits_remove (gdbarch, pc);
4146 @deftypefn {Architecture Function} CORE_ADDR unwind_sp (struct gdbarch *@var{gdbarch}, struct frame_info *@var{next_frame})
4148 This function is given a pointer to the NEXT stack frame
4149 (@pxref{All About Stack Frames, , All About Stack Frames} for how
4150 frames are represented) and returns the value of the stack pointer in
4151 the PREVIOUS frame (i.e.@: the frame of the function that called
4154 The implementation, which must be frame agnostic (work with any frame),
4155 is typically no more than:
4159 sp = frame_unwind_register_unsigned (next_frame, @var{ARCH}_SP_REGNUM);
4160 return gdbarch_addr_bits_remove (gdbarch, sp);
4165 @deftypefn {Architecture Function} int frame_num_args (struct gdbarch *@var{gdbarch}, struct frame_info *@var{this_frame})
4167 This function is given a pointer to THIS stack frame (@pxref{All
4168 About Stack Frames, , All About Stack Frames} for how frames are
4169 represented), and returns the number of arguments that are being
4170 passed, or -1 if not known.
4172 The default value is @code{NULL} (undefined), in which case the number of
4173 arguments passed on any stack frame is always unknown. For many
4174 architectures this will be a suitable default.
4178 @node Analyzing Stacks---Frame Sniffers
4179 @subsection Analyzing Stacks---Frame Sniffers
4181 When a program stops, @value{GDBN} needs to construct the chain of
4182 struct @code{frame_info} representing the state of the stack using
4183 appropriate @dfn{sniffers}.
4185 Each architecture requires appropriate sniffers, but they do not form
4186 entries in @code{@w{struct gdbarch}}, since more than one sniffer may
4187 be required and a sniffer may be suitable for more than one
4188 @code{@w{struct gdbarch}}. Instead sniffers are associated with
4189 architectures using the following functions.
4194 @findex frame_unwind_append_sniffer
4195 @code{frame_unwind_append_sniffer} is used to add a new sniffer to
4196 analyze THIS frame when given a pointer to the NEXT frame.
4199 @findex frame_base_append_sniffer
4200 @code{frame_base_append_sniffer} is used to add a new sniffer
4201 which can determine information about the base of a stack frame.
4204 @findex frame_base_set_default
4205 @code{frame_base_set_default} is used to specify the default base
4210 These functions all take a reference to @code{@w{struct gdbarch}}, so
4211 they are associated with a specific architecture. They are usually
4212 called in the @code{gdbarch} initialization function, after the
4213 @code{gdbarch} struct has been set up. Unless a default has been set, the
4214 most recently appended sniffer will be tried first.
4216 The main frame unwinding sniffer (as set by
4217 @code{frame_unwind_append_sniffer)} returns a structure specifying
4218 a set of sniffing functions:
4220 @cindex @code{frame_unwind}
4224 enum frame_type type;
4225 frame_this_id_ftype *this_id;
4226 frame_prev_register_ftype *prev_register;
4227 const struct frame_data *unwind_data;
4228 frame_sniffer_ftype *sniffer;
4229 frame_prev_pc_ftype *prev_pc;
4230 frame_dealloc_cache_ftype *dealloc_cache;
4234 The @code{type} field indicates the type of frame this sniffer can
4235 handle: normal, dummy (@pxref{Functions Creating Dummy Frames, ,
4236 Functions Creating Dummy Frames}), signal handler or sentinel. Signal
4237 handlers sometimes have their own simplified stack structure for
4238 efficiency, so may need their own handlers.
4240 The @code{unwind_data} field holds additional information which may be
4241 relevant to particular types of frame. For example it may hold
4242 additional information for signal handler frames.
4244 The remaining fields define functions that yield different types of
4245 information when given a pointer to the NEXT stack frame. Not all
4246 functions need be provided. If an entry is @code{NULL}, the next sniffer will
4252 @code{this_id} determines the stack pointer and function (code
4253 entry point) for THIS stack frame.
4256 @code{prev_register} determines where the values of registers for
4257 the PREVIOUS stack frame are stored in THIS stack frame.
4260 @code{sniffer} takes a look at THIS frame's registers to
4261 determine if this is the appropriate unwinder.
4264 @code{prev_pc} determines the program counter for THIS
4265 frame. Only needed if the program counter is not an ordinary register
4266 (@pxref{Register Architecture Functions & Variables,
4267 , Functions and Variables Specifying the Register Architecture}).
4270 @code{dealloc_cache} frees any additional memory associated with
4271 the prologue cache for this frame (@pxref{Prologue Caches, , Prologue
4276 In general it is only the @code{this_id} and @code{prev_register}
4277 fields that need be defined for custom sniffers.
4279 The frame base sniffer is much simpler. It is a @code{@w{struct
4280 frame_base}}, which refers to the corresponding @code{frame_unwind}
4281 struct and whose fields refer to functions yielding various addresses
4284 @cindex @code{frame_base}
4288 const struct frame_unwind *unwind;
4289 frame_this_base_ftype *this_base;
4290 frame_this_locals_ftype *this_locals;
4291 frame_this_args_ftype *this_args;
4295 All the functions referred to take a pointer to the NEXT frame as
4296 argument. The function referred to by @code{this_base} returns the
4297 base address of THIS frame, the function referred to by
4298 @code{this_locals} returns the base address of local variables in THIS
4299 frame and the function referred to by @code{this_args} returns the
4300 base address of the function arguments in this frame.
4302 As described above, the base address of a frame is the address
4303 immediately before the start of the NEXT frame. For a falling
4304 stack, this is the lowest address in the frame and for a rising stack
4305 it is the highest address in the frame. For most architectures the
4306 same address is also the base address for local variables and
4307 arguments, in which case the same function can be used for all three
4308 entries@footnote{It is worth noting that if it cannot be determined in any
4309 other way (for example by there being a register with the name
4310 @code{"fp"}), then the result of the @code{this_base} function will be
4311 used as the value of the frame pointer variable @kbd{$fp} in
4312 @value{GDBN}. This is very often not correct (for example with the
4313 OpenRISC 1000, this value is the stack pointer, @kbd{$sp}). In this
4314 case a register (raw or pseudo) with the name @code{"fp"} should be
4315 defined. It will be used in preference as the value of @kbd{$fp}.}.
4317 @node Inferior Call Setup
4318 @section Inferior Call Setup
4319 @cindex calls to the inferior
4322 * About Dummy Frames::
4323 * Functions Creating Dummy Frames::
4326 @node About Dummy Frames
4327 @subsection About Dummy Frames
4328 @cindex dummy frames
4330 @value{GDBN} can call functions in the target code (for example by
4331 using the @kbd{call} or @kbd{print} commands). These functions may be
4332 breakpointed, and it is essential that if a function does hit a
4333 breakpoint, commands like @kbd{backtrace} work correctly.
4335 This is achieved by making the stack look as though the function had
4336 been called from the point where @value{GDBN} had previously stopped.
4337 This requires that @value{GDBN} can set up stack frames appropriate for
4338 such function calls.
4340 @node Functions Creating Dummy Frames
4341 @subsection Functions Creating Dummy Frames
4343 The following functions provide the functionality to set up such
4344 @dfn{dummy} stack frames.
4346 @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})
4348 This function sets up a dummy stack frame for the function about to be
4349 called. @code{push_dummy_call} is given the arguments to be passed
4350 and must copy them into registers or push them on to the stack as
4351 appropriate for the ABI.
4353 @var{function} is a pointer to the function
4354 that will be called and @var{regcache} the register cache from which
4355 values should be obtained. @var{bp_addr} is the address to which the
4356 function should return (which is breakpointed, so @value{GDBN} can
4357 regain control, hence the name). @var{nargs} is the number of
4358 arguments to pass and @var{args} an array containing the argument
4359 values. @var{struct_return} is non-zero (true) if the function returns
4360 a structure, and if so @var{struct_addr} is the address in which the
4361 structure should be returned.
4363 After calling this function, @value{GDBN} will pass control to the
4364 target at the address of the function, which will find the stack and
4365 registers set up just as expected.
4367 The default value of this function is @code{NULL} (undefined). If the
4368 function is not defined, then @value{GDBN} will not allow the user to
4369 call functions within the target being debugged.
4373 @deftypefn {Architecture Function} {struct frame_id} unwind_dummy_id (struct gdbarch *@var{gdbarch}, struct frame_info *@var{next_frame})
4375 This is the inverse of @code{push_dummy_call} which restores the stack
4376 pointer and program counter after a call to evaluate a function using
4377 a dummy stack frame. The result is a @code{@w{struct frame_id}}, which
4378 contains the value of the stack pointer and program counter to be
4381 The NEXT frame pointer is provided as argument,
4382 @var{next_frame}. THIS frame is the frame of the dummy function,
4383 which can be unwound, to yield the required stack pointer and program
4384 counter from the PREVIOUS frame.
4386 The default value is @code{NULL} (undefined). If @code{push_dummy_call} is
4387 defined, then this function should also be defined.
4391 @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})
4393 If this function is not defined (its default value is @code{NULL}), a dummy
4394 call will use the entry point of the currently loaded code on the
4395 target as its return address. A temporary breakpoint will be set
4396 there, so the location must be writable and have room for a
4399 It is possible that this default is not suitable. It might not be
4400 writable (in ROM possibly), or the ABI might require code to be
4401 executed on return from a call to unwind the stack before the
4402 breakpoint is encountered.
4404 If either of these is the case, then push_dummy_code should be defined
4405 to push an instruction sequence onto the end of the stack to which the
4406 dummy call should return.
4408 The arguments are essentially the same as those to
4409 @code{push_dummy_call}. However the function is provided with the
4410 type of the function result, @var{value_type}, @var{bp_addr} is used
4411 to return a value (the address at which the breakpoint instruction
4412 should be inserted) and @var{real pc} is used to specify the resume
4413 address when starting the call sequence. The function should return
4414 the updated innermost stack address.
4417 @emph{Note:} This does require that code in the stack can be executed.
4418 Some Harvard architectures may not allow this.
4423 @node Defining Other Architecture Features
4424 @section Defining Other Architecture Features
4426 This section describes other functions and values in @code{gdbarch},
4427 together with some useful macros, that you can use to define the
4428 target architecture.
4432 @item CORE_ADDR gdbarch_addr_bits_remove (@var{gdbarch}, @var{addr})
4433 @findex gdbarch_addr_bits_remove
4434 If a raw machine instruction address includes any bits that are not
4435 really part of the address, then this function is used to zero those bits in
4436 @var{addr}. This is only used for addresses of instructions, and even then not
4439 For example, the two low-order bits of the PC on the Hewlett-Packard PA
4440 2.0 architecture contain the privilege level of the corresponding
4441 instruction. Since instructions must always be aligned on four-byte
4442 boundaries, the processor masks out these bits to generate the actual
4443 address of the instruction. @code{gdbarch_addr_bits_remove} would then for
4444 example look like that:
4446 arch_addr_bits_remove (CORE_ADDR addr)
4448 return (addr &= ~0x3);
4452 @item int address_class_name_to_type_flags (@var{gdbarch}, @var{name}, @var{type_flags_ptr})
4453 @findex address_class_name_to_type_flags
4454 If @var{name} is a valid address class qualifier name, set the @code{int}
4455 referenced by @var{type_flags_ptr} to the mask representing the qualifier
4456 and return 1. If @var{name} is not a valid address class qualifier name,
4459 The value for @var{type_flags_ptr} should be one of
4460 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
4461 possibly some combination of these values or'd together.
4462 @xref{Target Architecture Definition, , Address Classes}.
4464 @item int address_class_name_to_type_flags_p (@var{gdbarch})
4465 @findex address_class_name_to_type_flags_p
4466 Predicate which indicates whether @code{address_class_name_to_type_flags}
4469 @item int gdbarch_address_class_type_flags (@var{gdbarch}, @var{byte_size}, @var{dwarf2_addr_class})
4470 @findex gdbarch_address_class_type_flags
4471 Given a pointers byte size (as described by the debug information) and
4472 the possible @code{DW_AT_address_class} value, return the type flags
4473 used by @value{GDBN} to represent this address class. The value
4474 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
4475 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
4476 values or'd together.
4477 @xref{Target Architecture Definition, , Address Classes}.
4479 @item int gdbarch_address_class_type_flags_p (@var{gdbarch})
4480 @findex gdbarch_address_class_type_flags_p
4481 Predicate which indicates whether @code{gdbarch_address_class_type_flags_p} has
4484 @item const char *gdbarch_address_class_type_flags_to_name (@var{gdbarch}, @var{type_flags})
4485 @findex gdbarch_address_class_type_flags_to_name
4486 Return the name of the address class qualifier associated with the type
4487 flags given by @var{type_flags}.
4489 @item int gdbarch_address_class_type_flags_to_name_p (@var{gdbarch})
4490 @findex gdbarch_address_class_type_flags_to_name_p
4491 Predicate which indicates whether @code{gdbarch_address_class_type_flags_to_name} has been defined.
4492 @xref{Target Architecture Definition, , Address Classes}.
4494 @item void gdbarch_address_to_pointer (@var{gdbarch}, @var{type}, @var{buf}, @var{addr})
4495 @findex gdbarch_address_to_pointer
4496 Store in @var{buf} a pointer of type @var{type} representing the address
4497 @var{addr}, in the appropriate format for the current architecture.
4498 This function may safely assume that @var{type} is either a pointer or a
4499 C@t{++} reference type.
4500 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
4502 @item int gdbarch_believe_pcc_promotion (@var{gdbarch})
4503 @findex gdbarch_believe_pcc_promotion
4504 Used to notify if the compiler promotes a @code{short} or @code{char}
4505 parameter to an @code{int}, but still reports the parameter as its
4506 original type, rather than the promoted type.
4508 @item gdbarch_bits_big_endian (@var{gdbarch})
4509 @findex gdbarch_bits_big_endian
4510 This is used if the numbering of bits in the targets does @strong{not} match
4511 the endianism of the target byte order. A value of 1 means that the bits
4512 are numbered in a big-endian bit order, 0 means little-endian.
4514 @item set_gdbarch_bits_big_endian (@var{gdbarch}, @var{bits_big_endian})
4515 @findex set_gdbarch_bits_big_endian
4516 Calling set_gdbarch_bits_big_endian with a value of 1 indicates that the
4517 bits in the target are numbered in a big-endian bit order, 0 indicates
4522 This is the character array initializer for the bit pattern to put into
4523 memory where a breakpoint is set. Although it's common to use a trap
4524 instruction for a breakpoint, it's not required; for instance, the bit
4525 pattern could be an invalid instruction. The breakpoint must be no
4526 longer than the shortest instruction of the architecture.
4528 @code{BREAKPOINT} has been deprecated in favor of
4529 @code{gdbarch_breakpoint_from_pc}.
4531 @item BIG_BREAKPOINT
4532 @itemx LITTLE_BREAKPOINT
4533 @findex LITTLE_BREAKPOINT
4534 @findex BIG_BREAKPOINT
4535 Similar to BREAKPOINT, but used for bi-endian targets.
4537 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
4538 favor of @code{gdbarch_breakpoint_from_pc}.
4540 @item const gdb_byte *gdbarch_breakpoint_from_pc (@var{gdbarch}, @var{pcptr}, @var{lenptr})
4541 @findex gdbarch_breakpoint_from_pc
4542 @anchor{gdbarch_breakpoint_from_pc} Use the program counter to determine the
4543 contents and size of a breakpoint instruction. It returns a pointer to
4544 a static string of bytes that encode a breakpoint instruction, stores the
4545 length of the string to @code{*@var{lenptr}}, and adjusts the program
4546 counter (if necessary) to point to the actual memory location where the
4547 breakpoint should be inserted. May return @code{NULL} to indicate that
4548 software breakpoints are not supported.
4550 Although it is common to use a trap instruction for a breakpoint, it's
4551 not required; for instance, the bit pattern could be an invalid
4552 instruction. The breakpoint must be no longer than the shortest
4553 instruction of the architecture.
4555 Provided breakpoint bytes can be also used by @code{bp_loc_is_permanent} to
4556 detect permanent breakpoints. @code{gdbarch_breakpoint_from_pc} should return
4557 an unchanged memory copy if it was called for a location with permanent
4558 breakpoint as some architectures use breakpoint instructions containing
4559 arbitrary parameter value.
4561 Replaces all the other @var{BREAKPOINT} macros.
4563 @item int gdbarch_memory_insert_breakpoint (@var{gdbarch}, @var{bp_tgt})
4564 @itemx gdbarch_memory_remove_breakpoint (@var{gdbarch}, @var{bp_tgt})
4565 @findex gdbarch_memory_remove_breakpoint
4566 @findex gdbarch_memory_insert_breakpoint
4567 Insert or remove memory based breakpoints. Reasonable defaults
4568 (@code{default_memory_insert_breakpoint} and
4569 @code{default_memory_remove_breakpoint} respectively) have been
4570 provided so that it is not necessary to set these for most
4571 architectures. Architectures which may want to set
4572 @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
4573 conventional manner.
4575 It may also be desirable (from an efficiency standpoint) to define
4576 custom breakpoint insertion and removal routines if
4577 @code{gdbarch_breakpoint_from_pc} needs to read the target's memory for some
4580 @item CORE_ADDR gdbarch_adjust_breakpoint_address (@var{gdbarch}, @var{bpaddr})
4581 @findex gdbarch_adjust_breakpoint_address
4582 @cindex breakpoint address adjusted
4583 Given an address at which a breakpoint is desired, return a breakpoint
4584 address adjusted to account for architectural constraints on
4585 breakpoint placement. This method is not needed by most targets.
4587 The FR-V target (see @file{frv-tdep.c}) requires this method.
4588 The FR-V is a VLIW architecture in which a number of RISC-like
4589 instructions are grouped (packed) together into an aggregate
4590 instruction or instruction bundle. When the processor executes
4591 one of these bundles, the component instructions are executed
4594 In the course of optimization, the compiler may group instructions
4595 from distinct source statements into the same bundle. The line number
4596 information associated with one of the latter statements will likely
4597 refer to some instruction other than the first one in the bundle. So,
4598 if the user attempts to place a breakpoint on one of these latter
4599 statements, @value{GDBN} must be careful to @emph{not} place the break
4600 instruction on any instruction other than the first one in the bundle.
4601 (Remember though that the instructions within a bundle execute
4602 in parallel, so the @emph{first} instruction is the instruction
4603 at the lowest address and has nothing to do with execution order.)
4605 The FR-V's @code{gdbarch_adjust_breakpoint_address} method will adjust a
4606 breakpoint's address by scanning backwards for the beginning of
4607 the bundle, returning the address of the bundle.
4609 Since the adjustment of a breakpoint may significantly alter a user's
4610 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
4611 is initially set and each time that that breakpoint is hit.
4613 @item int gdbarch_call_dummy_location (@var{gdbarch})
4614 @findex gdbarch_call_dummy_location
4615 See the file @file{inferior.h}.
4617 This method has been replaced by @code{gdbarch_push_dummy_code}
4618 (@pxref{gdbarch_push_dummy_code}).
4620 @item int gdbarch_cannot_fetch_register (@var{gdbarch}, @var{regum})
4621 @findex gdbarch_cannot_fetch_register
4622 This function should return nonzero if @var{regno} cannot be fetched
4623 from an inferior process.
4625 @item int gdbarch_cannot_store_register (@var{gdbarch}, @var{regnum})
4626 @findex gdbarch_cannot_store_register
4627 This function should return nonzero if @var{regno} should not be
4628 written to the target. This is often the case for program counters,
4629 status words, and other special registers. This function returns 0 as
4630 default so that @value{GDBN} will assume that all registers may be written.
4632 @item int gdbarch_convert_register_p (@var{gdbarch}, @var{regnum}, struct type *@var{type})
4633 @findex gdbarch_convert_register_p
4634 Return non-zero if register @var{regnum} represents data values of type
4635 @var{type} in a non-standard form.
4636 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
4638 @item int gdbarch_fp0_regnum (@var{gdbarch})
4639 @findex gdbarch_fp0_regnum
4640 This function returns the number of the first floating point register,
4641 if the machine has such registers. Otherwise, it returns -1.
4643 @item CORE_ADDR gdbarch_decr_pc_after_break (@var{gdbarch})
4644 @findex gdbarch_decr_pc_after_break
4645 This function shall return the amount by which to decrement the PC after the
4646 program encounters a breakpoint. This is often the number of bytes in
4647 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
4649 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
4650 @findex DISABLE_UNSETTABLE_BREAK
4651 If defined, this should evaluate to 1 if @var{addr} is in a shared
4652 library in which breakpoints cannot be set and so should be disabled.
4654 @item int gdbarch_dwarf2_reg_to_regnum (@var{gdbarch}, @var{dwarf2_regnr})
4655 @findex gdbarch_dwarf2_reg_to_regnum
4656 Convert DWARF2 register number @var{dwarf2_regnr} into @value{GDBN} regnum.
4657 If not defined, no conversion will be performed.
4659 @item int gdbarch_ecoff_reg_to_regnum (@var{gdbarch}, @var{ecoff_regnr})
4660 @findex gdbarch_ecoff_reg_to_regnum
4661 Convert ECOFF register number @var{ecoff_regnr} into @value{GDBN} regnum. If
4662 not defined, no conversion will be performed.
4664 @item GCC_COMPILED_FLAG_SYMBOL
4665 @itemx GCC2_COMPILED_FLAG_SYMBOL
4666 @findex GCC2_COMPILED_FLAG_SYMBOL
4667 @findex GCC_COMPILED_FLAG_SYMBOL
4668 If defined, these are the names of the symbols that @value{GDBN} will
4669 look for to detect that GCC compiled the file. The default symbols
4670 are @code{gcc_compiled.} and @code{gcc2_compiled.},
4671 respectively. (Currently only defined for the Delta 68.)
4673 @item gdbarch_get_longjmp_target
4674 @findex gdbarch_get_longjmp_target
4675 This function determines the target PC address that @code{longjmp}
4676 will jump to, assuming that we have just stopped at a @code{longjmp}
4677 breakpoint. It takes a @code{CORE_ADDR *} as argument, and stores the
4678 target PC value through this pointer. It examines the current state
4679 of the machine as needed, typically by using a manually-determined
4680 offset into the @code{jmp_buf}. (While we might like to get the offset
4681 from the target's @file{jmpbuf.h}, that header file cannot be assumed
4682 to be available when building a cross-debugger.)
4684 @item DEPRECATED_IBM6000_TARGET
4685 @findex DEPRECATED_IBM6000_TARGET
4686 Shows that we are configured for an IBM RS/6000 system. This
4687 conditional should be eliminated (FIXME) and replaced by
4688 feature-specific macros. It was introduced in haste and we are
4689 repenting at leisure.
4691 @item I386_USE_GENERIC_WATCHPOINTS
4692 An x86-based target can define this to use the generic x86 watchpoint
4693 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4695 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{addr})
4696 @findex gdbarch_in_function_epilogue_p
4697 Returns non-zero if the given @var{addr} is in the epilogue of a function.
4698 The epilogue of a function is defined as the part of a function where
4699 the stack frame of the function already has been destroyed up to the
4700 final `return from function call' instruction.
4702 @item int gdbarch_in_solib_return_trampoline (@var{gdbarch}, @var{pc}, @var{name})
4703 @findex gdbarch_in_solib_return_trampoline
4704 Define this function to return nonzero if the program is stopped in the
4705 trampoline that returns from a shared library.
4707 @item target_so_ops.in_dynsym_resolve_code (@var{pc})
4708 @findex in_dynsym_resolve_code
4709 Define this to return nonzero if the program is stopped in the
4712 @item SKIP_SOLIB_RESOLVER (@var{pc})
4713 @findex SKIP_SOLIB_RESOLVER
4714 Define this to evaluate to the (nonzero) address at which execution
4715 should continue to get past the dynamic linker's symbol resolution
4716 function. A zero value indicates that it is not important or necessary
4717 to set a breakpoint to get through the dynamic linker and that single
4718 stepping will suffice.
4720 @item CORE_ADDR gdbarch_integer_to_address (@var{gdbarch}, @var{type}, @var{buf})
4721 @findex gdbarch_integer_to_address
4722 @cindex converting integers to addresses
4723 Define this when the architecture needs to handle non-pointer to address
4724 conversions specially. Converts that value to an address according to
4725 the current architectures conventions.
4727 @emph{Pragmatics: When the user copies a well defined expression from
4728 their source code and passes it, as a parameter, to @value{GDBN}'s
4729 @code{print} command, they should get the same value as would have been
4730 computed by the target program. Any deviation from this rule can cause
4731 major confusion and annoyance, and needs to be justified carefully. In
4732 other words, @value{GDBN} doesn't really have the freedom to do these
4733 conversions in clever and useful ways. It has, however, been pointed
4734 out that users aren't complaining about how @value{GDBN} casts integers
4735 to pointers; they are complaining that they can't take an address from a
4736 disassembly listing and give it to @code{x/i}. Adding an architecture
4737 method like @code{gdbarch_integer_to_address} certainly makes it possible for
4738 @value{GDBN} to ``get it right'' in all circumstances.}
4740 @xref{Target Architecture Definition, , Pointers Are Not Always
4743 @item CORE_ADDR gdbarch_pointer_to_address (@var{gdbarch}, @var{type}, @var{buf})
4744 @findex gdbarch_pointer_to_address
4745 Assume that @var{buf} holds a pointer of type @var{type}, in the
4746 appropriate format for the current architecture. Return the byte
4747 address the pointer refers to.
4748 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
4750 @item void gdbarch_register_to_value(@var{gdbarch}, @var{frame}, @var{regnum}, @var{type}, @var{fur})
4751 @findex gdbarch_register_to_value
4752 Convert the raw contents of register @var{regnum} into a value of type
4754 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
4756 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
4757 @findex REGISTER_CONVERT_TO_VIRTUAL
4758 Convert the value of register @var{reg} from its raw form to its virtual
4760 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
4762 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
4763 @findex REGISTER_CONVERT_TO_RAW
4764 Convert the value of register @var{reg} from its virtual form to its raw
4766 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
4768 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
4769 @findex regset_from_core_section
4770 Return the appropriate register set for a core file section with name
4771 @var{sect_name} and size @var{sect_size}.
4773 @item SOFTWARE_SINGLE_STEP_P()
4774 @findex SOFTWARE_SINGLE_STEP_P
4775 Define this as 1 if the target does not have a hardware single-step
4776 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
4778 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breakpoints_p})
4779 @findex SOFTWARE_SINGLE_STEP
4780 A function that inserts or removes (depending on
4781 @var{insert_breakpoints_p}) breakpoints at each possible destinations of
4782 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
4785 @item set_gdbarch_sofun_address_maybe_missing (@var{gdbarch}, @var{set})
4786 @findex set_gdbarch_sofun_address_maybe_missing
4787 Somebody clever observed that, the more actual addresses you have in the
4788 debug information, the more time the linker has to spend relocating
4789 them. So whenever there's some other way the debugger could find the
4790 address it needs, you should omit it from the debug info, to make
4793 Calling @code{set_gdbarch_sofun_address_maybe_missing} with a non-zero
4794 argument @var{set} indicates that a particular set of hacks of this sort
4795 are in use, affecting @code{N_SO} and @code{N_FUN} entries in stabs-format
4796 debugging information. @code{N_SO} stabs mark the beginning and ending
4797 addresses of compilation units in the text segment. @code{N_FUN} stabs
4798 mark the starts and ends of functions.
4800 In this case, @value{GDBN} assumes two things:
4804 @code{N_FUN} stabs have an address of zero. Instead of using those
4805 addresses, you should find the address where the function starts by
4806 taking the function name from the stab, and then looking that up in the
4807 minsyms (the linker/assembler symbol table). In other words, the stab
4808 has the name, and the linker/assembler symbol table is the only place
4809 that carries the address.
4812 @code{N_SO} stabs have an address of zero, too. You just look at the
4813 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab, and
4814 guess the starting and ending addresses of the compilation unit from them.
4817 @item int gdbarch_stabs_argument_has_addr (@var{gdbarch}, @var{type})
4818 @findex gdbarch_stabs_argument_has_addr
4819 @anchor{gdbarch_stabs_argument_has_addr} Define this function to return
4820 nonzero if a function argument of type @var{type} is passed by reference
4823 @item PROCESS_LINENUMBER_HOOK
4824 @findex PROCESS_LINENUMBER_HOOK
4825 A hook defined for XCOFF reading.
4827 @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})
4828 @findex gdbarch_push_dummy_call
4829 @anchor{gdbarch_push_dummy_call} Define this to push the dummy frame's call to
4830 the inferior function onto the stack. In addition to pushing @var{nargs}, the
4831 code should push @var{struct_addr} (when @var{struct_return} is non-zero), and
4832 the return address (@var{bp_addr}).
4834 @var{function} is a pointer to a @code{struct value}; on architectures that use
4835 function descriptors, this contains the function descriptor value.
4837 Returns the updated top-of-stack pointer.
4839 @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})
4840 @findex gdbarch_push_dummy_code
4841 @anchor{gdbarch_push_dummy_code} Given a stack based call dummy, push the
4842 instruction sequence (including space for a breakpoint) to which the
4843 called function should return.
4845 Set @var{bp_addr} to the address at which the breakpoint instruction
4846 should be inserted, @var{real_pc} to the resume address when starting
4847 the call sequence, and return the updated inner-most stack address.
4849 By default, the stack is grown sufficient to hold a frame-aligned
4850 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
4851 reserved for that breakpoint, and @var{real_pc} set to @var{funaddr}.
4853 This method replaces @w{@code{gdbarch_call_dummy_location (@var{gdbarch})}}.
4855 @item int gdbarch_sdb_reg_to_regnum (@var{gdbarch}, @var{sdb_regnr})
4856 @findex gdbarch_sdb_reg_to_regnum
4857 Use this function to convert sdb register @var{sdb_regnr} into @value{GDBN}
4858 regnum. If not defined, no conversion will be done.
4860 @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})
4861 @findex gdbarch_return_value
4862 @anchor{gdbarch_return_value} Given a function with a return-value of
4863 type @var{rettype}, return which return-value convention that function
4866 @value{GDBN} currently recognizes two function return-value conventions:
4867 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
4868 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
4869 value is found in memory and the address of that memory location is
4870 passed in as the function's first parameter.
4872 If the register convention is being used, and @var{writebuf} is
4873 non-@code{NULL}, also copy the return-value in @var{writebuf} into
4876 If the register convention is being used, and @var{readbuf} is
4877 non-@code{NULL}, also copy the return value from @var{regcache} into
4878 @var{readbuf} (@var{regcache} contains a copy of the registers from the
4879 just returned function).
4881 @emph{Maintainer note: This method replaces separate predicate, extract,
4882 store methods. By having only one method, the logic needed to determine
4883 the return-value convention need only be implemented in one place. If
4884 @value{GDBN} were written in an @sc{oo} language, this method would
4885 instead return an object that knew how to perform the register
4886 return-value extract and store.}
4888 @emph{Maintainer note: This method does not take a @var{gcc_p}
4889 parameter, and such a parameter should not be added. If an architecture
4890 that requires per-compiler or per-function information be identified,
4891 then the replacement of @var{rettype} with @code{struct value}
4892 @var{function} should be pursued.}
4894 @emph{Maintainer note: The @var{regcache} parameter limits this methods
4895 to the inner most frame. While replacing @var{regcache} with a
4896 @code{struct frame_info} @var{frame} parameter would remove that
4897 limitation there has yet to be a demonstrated need for such a change.}
4899 @item void gdbarch_skip_permanent_breakpoint (@var{gdbarch}, @var{regcache})
4900 @findex gdbarch_skip_permanent_breakpoint
4901 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
4902 steps over a breakpoint by removing it, stepping one instruction, and
4903 re-inserting the breakpoint. However, permanent breakpoints are
4904 hardwired into the inferior, and can't be removed, so this strategy
4905 doesn't work. Calling @code{gdbarch_skip_permanent_breakpoint} adjusts the
4906 processor's state so that execution will resume just after the breakpoint.
4907 This function does the right thing even when the breakpoint is in the delay slot
4908 of a branch or jump.
4910 @item CORE_ADDR gdbarch_skip_trampoline_code (@var{gdbarch}, @var{frame}, @var{pc})
4911 @findex gdbarch_skip_trampoline_code
4912 If the target machine has trampoline code that sits between callers and
4913 the functions being called, then define this function to return a new PC
4914 that is at the start of the real function.
4916 @item int gdbarch_deprecated_fp_regnum (@var{gdbarch})
4917 @findex gdbarch_deprecated_fp_regnum
4918 If the frame pointer is in a register, use this function to return the
4919 number of that register.
4921 @item int gdbarch_stab_reg_to_regnum (@var{gdbarch}, @var{stab_regnr})
4922 @findex gdbarch_stab_reg_to_regnum
4923 Use this function to convert stab register @var{stab_regnr} into @value{GDBN}
4924 regnum. If not defined, no conversion will be done.
4926 @item SYMBOL_RELOADING_DEFAULT
4927 @findex SYMBOL_RELOADING_DEFAULT
4928 The default value of the ``symbol-reloading'' variable. (Never defined in
4931 @item TARGET_CHAR_BIT
4932 @findex TARGET_CHAR_BIT
4933 Number of bits in a char; defaults to 8.
4935 @item int gdbarch_char_signed (@var{gdbarch})
4936 @findex gdbarch_char_signed
4937 Non-zero if @code{char} is normally signed on this architecture; zero if
4938 it should be unsigned.
4940 The ISO C standard requires the compiler to treat @code{char} as
4941 equivalent to either @code{signed char} or @code{unsigned char}; any
4942 character in the standard execution set is supposed to be positive.
4943 Most compilers treat @code{char} as signed, but @code{char} is unsigned
4944 on the IBM S/390, RS6000, and PowerPC targets.
4946 @item int gdbarch_double_bit (@var{gdbarch})
4947 @findex gdbarch_double_bit
4948 Number of bits in a double float; defaults to @w{@code{8 * TARGET_CHAR_BIT}}.
4950 @item int gdbarch_float_bit (@var{gdbarch})
4951 @findex gdbarch_float_bit
4952 Number of bits in a float; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4954 @item int gdbarch_int_bit (@var{gdbarch})
4955 @findex gdbarch_int_bit
4956 Number of bits in an integer; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4958 @item int gdbarch_long_bit (@var{gdbarch})
4959 @findex gdbarch_long_bit
4960 Number of bits in a long integer; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4962 @item int gdbarch_long_double_bit (@var{gdbarch})
4963 @findex gdbarch_long_double_bit
4964 Number of bits in a long double float;
4965 defaults to @w{@code{2 * gdbarch_double_bit (@var{gdbarch})}}.
4967 @item int gdbarch_long_long_bit (@var{gdbarch})
4968 @findex gdbarch_long_long_bit
4969 Number of bits in a long long integer; defaults to
4970 @w{@code{2 * gdbarch_long_bit (@var{gdbarch})}}.
4972 @item int gdbarch_ptr_bit (@var{gdbarch})
4973 @findex gdbarch_ptr_bit
4974 Number of bits in a pointer; defaults to
4975 @w{@code{gdbarch_int_bit (@var{gdbarch})}}.
4977 @item int gdbarch_short_bit (@var{gdbarch})
4978 @findex gdbarch_short_bit
4979 Number of bits in a short integer; defaults to @w{@code{2 * TARGET_CHAR_BIT}}.
4981 @item void gdbarch_virtual_frame_pointer (@var{gdbarch}, @var{pc}, @var{frame_regnum}, @var{frame_offset})
4982 @findex gdbarch_virtual_frame_pointer
4983 Returns a @code{(@var{register}, @var{offset})} pair representing the virtual
4984 frame pointer in use at the code address @var{pc}. If virtual frame
4985 pointers are not used, a default definition simply returns
4986 @code{gdbarch_deprecated_fp_regnum} (or @code{gdbarch_sp_regnum}, if
4987 no frame pointer is defined), with an offset of zero.
4989 @c need to explain virtual frame pointers, they are recorded in agent
4990 @c expressions for tracepoints
4992 @item TARGET_HAS_HARDWARE_WATCHPOINTS
4993 If non-zero, the target has support for hardware-assisted
4994 watchpoints. @xref{Algorithms, watchpoints}, for more details and
4995 other related macros.
4997 @item int gdbarch_print_insn (@var{gdbarch}, @var{vma}, @var{info})
4998 @findex gdbarch_print_insn
4999 This is the function used by @value{GDBN} to print an assembly
5000 instruction. It prints the instruction at address @var{vma} in
5001 debugged memory and returns the length of the instruction, in bytes.
5002 This usually points to a function in the @code{opcodes} library
5003 (@pxref{Support Libraries, ,Opcodes}). @var{info} is a structure (of
5004 type @code{disassemble_info}) defined in the header file
5005 @file{include/dis-asm.h}, and used to pass information to the
5006 instruction decoding routine.
5008 @item frame_id gdbarch_dummy_id (@var{gdbarch}, @var{frame})
5009 @findex gdbarch_dummy_id
5010 @anchor{gdbarch_dummy_id} Given @var{frame} return a @w{@code{struct
5011 frame_id}} that uniquely identifies an inferior function call's dummy
5012 frame. The value returned must match the dummy frame stack value
5013 previously saved by @code{call_function_by_hand}.
5015 @item void gdbarch_value_to_register (@var{gdbarch}, @var{frame}, @var{type}, @var{buf})
5016 @findex gdbarch_value_to_register
5017 Convert a value of type @var{type} into the raw contents of a register.
5018 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
5022 Motorola M68K target conditionals.
5026 Define this to be the 4-bit location of the breakpoint trap vector. If
5027 not defined, it will default to @code{0xf}.
5029 @item REMOTE_BPT_VECTOR
5030 Defaults to @code{1}.
5034 @node Adding a New Target
5035 @section Adding a New Target
5037 @cindex adding a target
5038 The following files add a target to @value{GDBN}:
5041 @cindex target dependent files
5043 @item gdb/@var{ttt}-tdep.c
5044 Contains any miscellaneous code required for this target machine. On
5045 some machines it doesn't exist at all.
5047 @item gdb/@var{arch}-tdep.c
5048 @itemx gdb/@var{arch}-tdep.h
5049 This is required to describe the basic layout of the target machine's
5050 processor chip (registers, stack, etc.). It can be shared among many
5051 targets that use the same processor architecture.
5055 (Target header files such as
5056 @file{gdb/config/@var{arch}/tm-@var{ttt}.h},
5057 @file{gdb/config/@var{arch}/tm-@var{arch}.h}, and
5058 @file{config/tm-@var{os}.h} are no longer used.)
5060 @findex _initialize_@var{arch}_tdep
5061 A @value{GDBN} description for a new architecture, arch is created by
5062 defining a global function @code{_initialize_@var{arch}_tdep}, by
5063 convention in the source file @file{@var{arch}-tdep.c}. For
5064 example, in the case of the OpenRISC 1000, this function is called
5065 @code{_initialize_or1k_tdep} and is found in the file
5068 The object file resulting from compiling this source file, which will
5069 contain the implementation of the
5070 @code{_initialize_@var{arch}_tdep} function is specified in the
5071 @value{GDBN} @file{configure.tgt} file, which includes a large case
5072 statement pattern matching against the @code{--target} option of the
5073 @kbd{configure} script.
5076 @emph{Note:} If the architecture requires multiple source files, the
5077 corresponding binaries should be included in
5078 @file{configure.tgt}. However if there are header files, the
5079 dependencies on these will not be picked up from the entries in
5080 @file{configure.tgt}. The @file{Makefile.in} file will need extending to
5081 show these dependencies.
5084 @findex gdbarch_register
5085 A new struct gdbarch, defining the new architecture, is created within
5086 the @code{_initialize_@var{arch}_tdep} function by calling
5087 @code{gdbarch_register}:
5090 void gdbarch_register (enum bfd_architecture architecture,
5091 gdbarch_init_ftype *init_func,
5092 gdbarch_dump_tdep_ftype *tdep_dump_func);
5095 This function has been described fully in an earlier
5096 section. @xref{How an Architecture is Represented, , How an
5097 Architecture is Represented}.
5099 The new @code{@w{struct gdbarch}} should contain implementations of
5100 the necessary functions (described in the previous sections) to
5101 describe the basic layout of the target machine's processor chip
5102 (registers, stack, etc.). It can be shared among many targets that use
5103 the same processor architecture.
5105 @node Target Descriptions
5106 @chapter Target Descriptions
5107 @cindex target descriptions
5109 The target architecture definition (@pxref{Target Architecture Definition})
5110 contains @value{GDBN}'s hard-coded knowledge about an architecture. For
5111 some platforms, it is handy to have more flexible knowledge about a specific
5112 instance of the architecture---for instance, a processor or development board.
5113 @dfn{Target descriptions} provide a mechanism for the user to tell @value{GDBN}
5114 more about what their target supports, or for the target to tell @value{GDBN}
5117 For details on writing, automatically supplying, and manually selecting
5118 target descriptions, see @ref{Target Descriptions, , , gdb,
5119 Debugging with @value{GDBN}}. This section will cover some related
5120 topics about the @value{GDBN} internals.
5123 * Target Descriptions Implementation::
5124 * Adding Target Described Register Support::
5127 @node Target Descriptions Implementation
5128 @section Target Descriptions Implementation
5129 @cindex target descriptions, implementation
5131 Before @value{GDBN} connects to a new target, or runs a new program on
5132 an existing target, it discards any existing target description and
5133 reverts to a default gdbarch. Then, after connecting, it looks for a
5134 new target description by calling @code{target_find_description}.
5136 A description may come from a user specified file (XML), the remote
5137 @samp{qXfer:features:read} packet (also XML), or from any custom
5138 @code{to_read_description} routine in the target vector. For instance,
5139 the remote target supports guessing whether a MIPS target is 32-bit or
5140 64-bit based on the size of the @samp{g} packet.
5142 If any target description is found, @value{GDBN} creates a new gdbarch
5143 incorporating the description by calling @code{gdbarch_update_p}. Any
5144 @samp{<architecture>} element is handled first, to determine which
5145 architecture's gdbarch initialization routine is called to create the
5146 new architecture. Then the initialization routine is called, and has
5147 a chance to adjust the constructed architecture based on the contents
5148 of the target description. For instance, it can recognize any
5149 properties set by a @code{to_read_description} routine. Also
5150 see @ref{Adding Target Described Register Support}.
5152 @node Adding Target Described Register Support
5153 @section Adding Target Described Register Support
5154 @cindex target descriptions, adding register support
5156 Target descriptions can report additional registers specific to an
5157 instance of the target. But it takes a little work in the architecture
5158 specific routines to support this.
5160 A target description must either have no registers or a complete
5161 set---this avoids complexity in trying to merge standard registers
5162 with the target defined registers. It is the architecture's
5163 responsibility to validate that a description with registers has
5164 everything it needs. To keep architecture code simple, the same
5165 mechanism is used to assign fixed internal register numbers to
5168 If @code{tdesc_has_registers} returns 1, the description contains
5169 registers. The architecture's @code{gdbarch_init} routine should:
5174 Call @code{tdesc_data_alloc} to allocate storage, early, before
5175 searching for a matching gdbarch or allocating a new one.
5178 Use @code{tdesc_find_feature} to locate standard features by name.
5181 Use @code{tdesc_numbered_register} and @code{tdesc_numbered_register_choices}
5182 to locate the expected registers in the standard features.
5185 Return @code{NULL} if a required feature is missing, or if any standard
5186 feature is missing expected registers. This will produce a warning that
5187 the description was incomplete.
5190 Free the allocated data before returning, unless @code{tdesc_use_registers}
5194 Call @code{set_gdbarch_num_regs} as usual, with a number higher than any
5195 fixed number passed to @code{tdesc_numbered_register}.
5198 Call @code{tdesc_use_registers} after creating a new gdbarch, before
5203 After @code{tdesc_use_registers} has been called, the architecture's
5204 @code{register_name}, @code{register_type}, and @code{register_reggroup_p}
5205 routines will not be called; that information will be taken from
5206 the target description. @code{num_regs} may be increased to account
5207 for any additional registers in the description.
5209 Pseudo-registers require some extra care:
5214 Using @code{tdesc_numbered_register} allows the architecture to give
5215 constant register numbers to standard architectural registers, e.g.@:
5216 as an @code{enum} in @file{@var{arch}-tdep.h}. But because
5217 pseudo-registers are always numbered above @code{num_regs},
5218 which may be increased by the description, constant numbers
5219 can not be used for pseudos. They must be numbered relative to
5220 @code{num_regs} instead.
5223 The description will not describe pseudo-registers, so the
5224 architecture must call @code{set_tdesc_pseudo_register_name},
5225 @code{set_tdesc_pseudo_register_type}, and
5226 @code{set_tdesc_pseudo_register_reggroup_p} to supply routines
5227 describing pseudo registers. These routines will be passed
5228 internal register numbers, so the same routines used for the
5229 gdbarch equivalents are usually suitable.
5234 @node Target Vector Definition
5236 @chapter Target Vector Definition
5237 @cindex target vector
5239 The target vector defines the interface between @value{GDBN}'s
5240 abstract handling of target systems, and the nitty-gritty code that
5241 actually exercises control over a process or a serial port.
5242 @value{GDBN} includes some 30-40 different target vectors; however,
5243 each configuration of @value{GDBN} includes only a few of them.
5246 * Managing Execution State::
5247 * Existing Targets::
5250 @node Managing Execution State
5251 @section Managing Execution State
5252 @cindex execution state
5254 A target vector can be completely inactive (not pushed on the target
5255 stack), active but not running (pushed, but not connected to a fully
5256 manifested inferior), or completely active (pushed, with an accessible
5257 inferior). Most targets are only completely inactive or completely
5258 active, but some support persistent connections to a target even
5259 when the target has exited or not yet started.
5261 For example, connecting to the simulator using @code{target sim} does
5262 not create a running program. Neither registers nor memory are
5263 accessible until @code{run}. Similarly, after @code{kill}, the
5264 program can not continue executing. But in both cases @value{GDBN}
5265 remains connected to the simulator, and target-specific commands
5266 are directed to the simulator.
5268 A target which only supports complete activation should push itself
5269 onto the stack in its @code{to_open} routine (by calling
5270 @code{push_target}), and unpush itself from the stack in its
5271 @code{to_mourn_inferior} routine (by calling @code{unpush_target}).
5273 A target which supports both partial and complete activation should
5274 still call @code{push_target} in @code{to_open}, but not call
5275 @code{unpush_target} in @code{to_mourn_inferior}. Instead, it should
5276 call either @code{target_mark_running} or @code{target_mark_exited}
5277 in its @code{to_open}, depending on whether the target is fully active
5278 after connection. It should also call @code{target_mark_running} any
5279 time the inferior becomes fully active (e.g.@: in
5280 @code{to_create_inferior} and @code{to_attach}), and
5281 @code{target_mark_exited} when the inferior becomes inactive (in
5282 @code{to_mourn_inferior}). The target should also make sure to call
5283 @code{target_mourn_inferior} from its @code{to_kill}, to return the
5284 target to inactive state.
5286 @node Existing Targets
5287 @section Existing Targets
5290 @subsection File Targets
5292 Both executables and core files have target vectors.
5294 @subsection Standard Protocol and Remote Stubs
5296 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code that
5297 runs in the target system. @value{GDBN} provides several sample
5298 @dfn{stubs} that can be integrated into target programs or operating
5299 systems for this purpose; they are named @file{@var{cpu}-stub.c}. Many
5300 operating systems, embedded targets, emulators, and simulators already
5301 have a @value{GDBN} stub built into them, and maintenance of the remote
5302 protocol must be careful to preserve compatibility.
5304 The @value{GDBN} user's manual describes how to put such a stub into
5305 your target code. What follows is a discussion of integrating the
5306 SPARC stub into a complicated operating system (rather than a simple
5307 program), by Stu Grossman, the author of this stub.
5309 The trap handling code in the stub assumes the following upon entry to
5314 %l1 and %l2 contain pc and npc respectively at the time of the trap;
5320 you are in the correct trap window.
5323 As long as your trap handler can guarantee those conditions, then there
5324 is no reason why you shouldn't be able to ``share'' traps with the stub.
5325 The stub has no requirement that it be jumped to directly from the
5326 hardware trap vector. That is why it calls @code{exceptionHandler()},
5327 which is provided by the external environment. For instance, this could
5328 set up the hardware traps to actually execute code which calls the stub
5329 first, and then transfers to its own trap handler.
5331 For the most point, there probably won't be much of an issue with
5332 ``sharing'' traps, as the traps we use are usually not used by the kernel,
5333 and often indicate unrecoverable error conditions. Anyway, this is all
5334 controlled by a table, and is trivial to modify. The most important
5335 trap for us is for @code{ta 1}. Without that, we can't single step or
5336 do breakpoints. Everything else is unnecessary for the proper operation
5337 of the debugger/stub.
5339 From reading the stub, it's probably not obvious how breakpoints work.
5340 They are simply done by deposit/examine operations from @value{GDBN}.
5342 @subsection ROM Monitor Interface
5344 @subsection Custom Protocols
5346 @subsection Transport Layer
5348 @subsection Builtin Simulator
5351 @node Native Debugging
5353 @chapter Native Debugging
5354 @cindex native debugging
5356 Several files control @value{GDBN}'s configuration for native support:
5360 @item gdb/config/@var{arch}/@var{xyz}.mh
5361 Specifies Makefile fragments needed by a @emph{native} configuration on
5362 machine @var{xyz}. In particular, this lists the required
5363 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
5364 Also specifies the header file which describes native support on
5365 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
5366 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
5367 @samp{NAT_CDEPS}, @samp{NAT_GENERATED_FILES}, etc.; see @file{Makefile.in}.
5369 @emph{Maintainer's note: The @file{.mh} suffix is because this file
5370 originally contained @file{Makefile} fragments for hosting @value{GDBN}
5371 on machine @var{xyz}. While the file is no longer used for this
5372 purpose, the @file{.mh} suffix remains. Perhaps someone will
5373 eventually rename these fragments so that they have a @file{.mn}
5376 @item gdb/config/@var{arch}/nm-@var{xyz}.h
5377 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
5378 macro definitions describing the native system environment, such as
5379 child process control and core file support.
5381 @item gdb/@var{xyz}-nat.c
5382 Contains any miscellaneous C code required for this native support of
5383 this machine. On some machines it doesn't exist at all.
5386 There are some ``generic'' versions of routines that can be used by
5387 various systems. These can be customized in various ways by macros
5388 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
5389 the @var{xyz} host, you can just include the generic file's name (with
5390 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
5392 Otherwise, if your machine needs custom support routines, you will need
5393 to write routines that perform the same functions as the generic file.
5394 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
5395 into @code{NATDEPFILES}.
5399 This contains the @emph{target_ops vector} that supports Unix child
5400 processes on systems which use ptrace and wait to control the child.
5403 This contains the @emph{target_ops vector} that supports Unix child
5404 processes on systems which use /proc to control the child.
5407 This does the low-level grunge that uses Unix system calls to do a ``fork
5408 and exec'' to start up a child process.
5411 This is the low level interface to inferior processes for systems using
5412 the Unix @code{ptrace} call in a vanilla way.
5415 @section Native core file Support
5416 @cindex native core files
5419 @findex fetch_core_registers
5420 @item core-aout.c::fetch_core_registers()
5421 Support for reading registers out of a core file. This routine calls
5422 @code{register_addr()}, see below. Now that BFD is used to read core
5423 files, virtually all machines should use @code{core-aout.c}, and should
5424 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
5425 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
5427 @item core-aout.c::register_addr()
5428 If your @code{nm-@var{xyz}.h} file defines the macro
5429 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
5430 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
5431 register number @code{regno}. @code{blockend} is the offset within the
5432 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
5433 @file{core-aout.c} will define the @code{register_addr()} function and
5434 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
5435 you are using the standard @code{fetch_core_registers()}, you will need
5436 to define your own version of @code{register_addr()}, put it into your
5437 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
5438 the @code{NATDEPFILES} list. If you have your own
5439 @code{fetch_core_registers()}, you may not need a separate
5440 @code{register_addr()}. Many custom @code{fetch_core_registers()}
5441 implementations simply locate the registers themselves.@refill
5444 When making @value{GDBN} run native on a new operating system, to make it
5445 possible to debug core files, you will need to either write specific
5446 code for parsing your OS's core files, or customize
5447 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
5448 machine uses to define the struct of registers that is accessible
5449 (possibly in the u-area) in a core file (rather than
5450 @file{machine/reg.h}), and an include file that defines whatever header
5451 exists on a core file (e.g., the u-area or a @code{struct core}). Then
5452 modify @code{trad_unix_core_file_p} to use these values to set up the
5453 section information for the data segment, stack segment, any other
5454 segments in the core file (perhaps shared library contents or control
5455 information), ``registers'' segment, and if there are two discontiguous
5456 sets of registers (e.g., integer and float), the ``reg2'' segment. This
5457 section information basically delimits areas in the core file in a
5458 standard way, which the section-reading routines in BFD know how to seek
5461 Then back in @value{GDBN}, you need a matching routine called
5462 @code{fetch_core_registers}. If you can use the generic one, it's in
5463 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
5464 It will be passed a char pointer to the entire ``registers'' segment,
5465 its length, and a zero; or a char pointer to the entire ``regs2''
5466 segment, its length, and a 2. The routine should suck out the supplied
5467 register values and install them into @value{GDBN}'s ``registers'' array.
5469 If your system uses @file{/proc} to control processes, and uses ELF
5470 format core files, then you may be able to use the same routines for
5471 reading the registers out of processes and out of core files.
5479 @section shared libraries
5481 @section Native Conditionals
5482 @cindex native conditionals
5484 When @value{GDBN} is configured and compiled, various macros are
5485 defined or left undefined, to control compilation when the host and
5486 target systems are the same. These macros should be defined (or left
5487 undefined) in @file{nm-@var{system}.h}.
5491 @item I386_USE_GENERIC_WATCHPOINTS
5492 An x86-based machine can define this to use the generic x86 watchpoint
5493 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
5495 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
5497 Define this to expand into an expression that will cause the symbols in
5498 @var{filename} to be added to @value{GDBN}'s symbol table. If
5499 @var{readsyms} is zero symbols are not read but any necessary low level
5500 processing for @var{filename} is still done.
5502 @item SOLIB_CREATE_INFERIOR_HOOK
5503 @findex SOLIB_CREATE_INFERIOR_HOOK
5504 Define this to expand into any shared-library-relocation code that you
5505 want to be run just after the child process has been forked.
5507 @item START_INFERIOR_TRAPS_EXPECTED
5508 @findex START_INFERIOR_TRAPS_EXPECTED
5509 When starting an inferior, @value{GDBN} normally expects to trap
5511 the shell execs, and once when the program itself execs. If the actual
5512 number of traps is something other than 2, then define this macro to
5513 expand into the number expected.
5517 @node Support Libraries
5519 @chapter Support Libraries
5524 BFD provides support for @value{GDBN} in several ways:
5527 @item identifying executable and core files
5528 BFD will identify a variety of file types, including a.out, coff, and
5529 several variants thereof, as well as several kinds of core files.
5531 @item access to sections of files
5532 BFD parses the file headers to determine the names, virtual addresses,
5533 sizes, and file locations of all the various named sections in files
5534 (such as the text section or the data section). @value{GDBN} simply
5535 calls BFD to read or write section @var{x} at byte offset @var{y} for
5538 @item specialized core file support
5539 BFD provides routines to determine the failing command name stored in a
5540 core file, the signal with which the program failed, and whether a core
5541 file matches (i.e.@: could be a core dump of) a particular executable
5544 @item locating the symbol information
5545 @value{GDBN} uses an internal interface of BFD to determine where to find the
5546 symbol information in an executable file or symbol-file. @value{GDBN} itself
5547 handles the reading of symbols, since BFD does not ``understand'' debug
5548 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
5553 @cindex opcodes library
5555 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
5556 library because it's also used in binutils, for @file{objdump}).
5559 @cindex readline library
5560 The @code{readline} library provides a set of functions for use by applications
5561 that allow users to edit command lines as they are typed in.
5564 @cindex @code{libiberty} library
5566 The @code{libiberty} library provides a set of functions and features
5567 that integrate and improve on functionality found in modern operating
5568 systems. Broadly speaking, such features can be divided into three
5569 groups: supplemental functions (functions that may be missing in some
5570 environments and operating systems), replacement functions (providing
5571 a uniform and easier to use interface for commonly used standard
5572 functions), and extensions (which provide additional functionality
5573 beyond standard functions).
5575 @value{GDBN} uses various features provided by the @code{libiberty}
5576 library, for instance the C@t{++} demangler, the @acronym{IEEE}
5577 floating format support functions, the input options parser
5578 @samp{getopt}, the @samp{obstack} extension, and other functions.
5580 @subsection @code{obstacks} in @value{GDBN}
5581 @cindex @code{obstacks}
5583 The obstack mechanism provides a convenient way to allocate and free
5584 chunks of memory. Each obstack is a pool of memory that is managed
5585 like a stack. Objects (of any nature, size and alignment) are
5586 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
5587 @code{libiberty}'s documentation for a more detailed explanation of
5590 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
5591 object files. There is an obstack associated with each internal
5592 representation of an object file. Lots of things get allocated on
5593 these @code{obstacks}: dictionary entries, blocks, blockvectors,
5594 symbols, minimal symbols, types, vectors of fundamental types, class
5595 fields of types, object files section lists, object files section
5596 offset lists, line tables, symbol tables, partial symbol tables,
5597 string tables, symbol table private data, macros tables, debug
5598 information sections and entries, import and export lists (som),
5599 unwind information (hppa), dwarf2 location expressions data. Plus
5600 various strings such as directory names strings, debug format strings,
5603 An essential and convenient property of all data on @code{obstacks} is
5604 that memory for it gets allocated (with @code{obstack_alloc}) at
5605 various times during a debugging session, but it is released all at
5606 once using the @code{obstack_free} function. The @code{obstack_free}
5607 function takes a pointer to where in the stack it must start the
5608 deletion from (much like the cleanup chains have a pointer to where to
5609 start the cleanups). Because of the stack like structure of the
5610 @code{obstacks}, this allows to free only a top portion of the
5611 obstack. There are a few instances in @value{GDBN} where such thing
5612 happens. Calls to @code{obstack_free} are done after some local data
5613 is allocated to the obstack. Only the local data is deleted from the
5614 obstack. Of course this assumes that nothing between the
5615 @code{obstack_alloc} and the @code{obstack_free} allocates anything
5616 else on the same obstack. For this reason it is best and safest to
5617 use temporary @code{obstacks}.
5619 Releasing the whole obstack is also not safe per se. It is safe only
5620 under the condition that we know the @code{obstacks} memory is no
5621 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
5622 when we get rid of the whole objfile(s), for instance upon reading a
5626 @cindex regular expressions library
5637 @item SIGN_EXTEND_CHAR
5639 @item SWITCH_ENUM_BUG
5648 @section Array Containers
5649 @cindex Array Containers
5652 Often it is necessary to manipulate a dynamic array of a set of
5653 objects. C forces some bookkeeping on this, which can get cumbersome
5654 and repetitive. The @file{vec.h} file contains macros for defining
5655 and using a typesafe vector type. The functions defined will be
5656 inlined when compiling, and so the abstraction cost should be zero.
5657 Domain checks are added to detect programming errors.
5659 An example use would be an array of symbols or section information.
5660 The array can be grown as symbols are read in (or preallocated), and
5661 the accessor macros provided keep care of all the necessary
5662 bookkeeping. Because the arrays are type safe, there is no danger of
5663 accidentally mixing up the contents. Think of these as C++ templates,
5664 but implemented in C.
5666 Because of the different behavior of structure objects, scalar objects
5667 and of pointers, there are three flavors of vector, one for each of
5668 these variants. Both the structure object and pointer variants pass
5669 pointers to objects around --- in the former case the pointers are
5670 stored into the vector and in the latter case the pointers are
5671 dereferenced and the objects copied into the vector. The scalar
5672 object variant is suitable for @code{int}-like objects, and the vector
5673 elements are returned by value.
5675 There are both @code{index} and @code{iterate} accessors. The iterator
5676 returns a boolean iteration condition and updates the iteration
5677 variable passed by reference. Because the iterator will be inlined,
5678 the address-of can be optimized away.
5680 The vectors are implemented using the trailing array idiom, thus they
5681 are not resizeable without changing the address of the vector object
5682 itself. This means you cannot have variables or fields of vector type
5683 --- always use a pointer to a vector. The one exception is the final
5684 field of a structure, which could be a vector type. You will have to
5685 use the @code{embedded_size} & @code{embedded_init} calls to create
5686 such objects, and they will probably not be resizeable (so don't use
5687 the @dfn{safe} allocation variants). The trailing array idiom is used
5688 (rather than a pointer to an array of data), because, if we allow
5689 @code{NULL} to also represent an empty vector, empty vectors occupy
5690 minimal space in the structure containing them.
5692 Each operation that increases the number of active elements is
5693 available in @dfn{quick} and @dfn{safe} variants. The former presumes
5694 that there is sufficient allocated space for the operation to succeed
5695 (it dies if there is not). The latter will reallocate the vector, if
5696 needed. Reallocation causes an exponential increase in vector size.
5697 If you know you will be adding N elements, it would be more efficient
5698 to use the reserve operation before adding the elements with the
5699 @dfn{quick} operation. This will ensure there are at least as many
5700 elements as you ask for, it will exponentially increase if there are
5701 too few spare slots. If you want reserve a specific number of slots,
5702 but do not want the exponential increase (for instance, you know this
5703 is the last allocation), use a negative number for reservation. You
5704 can also create a vector of a specific size from the get go.
5706 You should prefer the push and pop operations, as they append and
5707 remove from the end of the vector. If you need to remove several items
5708 in one go, use the truncate operation. The insert and remove
5709 operations allow you to change elements in the middle of the vector.
5710 There are two remove operations, one which preserves the element
5711 ordering @code{ordered_remove}, and one which does not
5712 @code{unordered_remove}. The latter function copies the end element
5713 into the removed slot, rather than invoke a memmove operation. The
5714 @code{lower_bound} function will determine where to place an item in
5715 the array using insert that will maintain sorted order.
5717 If you need to directly manipulate a vector, then the @code{address}
5718 accessor will return the address of the start of the vector. Also the
5719 @code{space} predicate will tell you whether there is spare capacity in the
5720 vector. You will not normally need to use these two functions.
5722 Vector types are defined using a
5723 @code{DEF_VEC_@{O,P,I@}(@var{typename})} macro. Variables of vector
5724 type are declared using a @code{VEC(@var{typename})} macro. The
5725 characters @code{O}, @code{P} and @code{I} indicate whether
5726 @var{typename} is an object (@code{O}), pointer (@code{P}) or integral
5727 (@code{I}) type. Be careful to pick the correct one, as you'll get an
5728 awkward and inefficient API if you use the wrong one. There is a
5729 check, which results in a compile-time warning, for the @code{P} and
5730 @code{I} versions, but there is no check for the @code{O} versions, as
5731 that is not possible in plain C.
5733 An example of their use would be,
5736 DEF_VEC_P(tree); // non-managed tree vector.
5739 VEC(tree) *v; // A (pointer to) a vector of tree pointers.
5742 struct my_struct *s;
5744 if (VEC_length(tree, s->v)) @{ we have some contents @}
5745 VEC_safe_push(tree, s->v, decl); // append some decl onto the end
5746 for (ix = 0; VEC_iterate(tree, s->v, ix, elt); ix++)
5747 @{ do something with elt @}
5751 The @file{vec.h} file provides details on how to invoke the various
5752 accessors provided. They are enumerated here:
5756 Return the number of items in the array,
5759 Return true if the array has no elements.
5763 Return the last or arbitrary item in the array.
5766 Access an array element and indicate whether the array has been
5771 Create and destroy an array.
5773 @item VEC_embedded_size
5774 @itemx VEC_embedded_init
5775 Helpers for embedding an array as the final element of another struct.
5781 Return the amount of free space in an array.
5784 Ensure a certain amount of free space.
5786 @item VEC_quick_push
5787 @itemx VEC_safe_push
5788 Append to an array, either assuming the space is available, or making
5792 Remove the last item from an array.
5795 Remove several items from the end of an array.
5798 Add several items to the end of an array.
5801 Overwrite an item in the array.
5803 @item VEC_quick_insert
5804 @itemx VEC_safe_insert
5805 Insert an item into the middle of the array. Either the space must
5806 already exist, or the space is created.
5808 @item VEC_ordered_remove
5809 @itemx VEC_unordered_remove
5810 Remove an item from the array, preserving order or not.
5812 @item VEC_block_remove
5813 Remove a set of items from the array.
5816 Provide the address of the first element.
5818 @item VEC_lower_bound
5819 Binary search the array.
5829 This chapter covers topics that are lower-level than the major
5830 algorithms of @value{GDBN}.
5835 Cleanups are a structured way to deal with things that need to be done
5838 When your code does something (e.g., @code{xmalloc} some memory, or
5839 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
5840 the memory or @code{close} the file), it can make a cleanup. The
5841 cleanup will be done at some future point: when the command is finished
5842 and control returns to the top level; when an error occurs and the stack
5843 is unwound; or when your code decides it's time to explicitly perform
5844 cleanups. Alternatively you can elect to discard the cleanups you
5850 @item struct cleanup *@var{old_chain};
5851 Declare a variable which will hold a cleanup chain handle.
5853 @findex make_cleanup
5854 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
5855 Make a cleanup which will cause @var{function} to be called with
5856 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
5857 handle that can later be passed to @code{do_cleanups} or
5858 @code{discard_cleanups}. Unless you are going to call
5859 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
5860 from @code{make_cleanup}.
5863 @item do_cleanups (@var{old_chain});
5864 Do all cleanups added to the chain since the corresponding
5865 @code{make_cleanup} call was made.
5867 @findex discard_cleanups
5868 @item discard_cleanups (@var{old_chain});
5869 Same as @code{do_cleanups} except that it just removes the cleanups from
5870 the chain and does not call the specified functions.
5873 Cleanups are implemented as a chain. The handle returned by
5874 @code{make_cleanups} includes the cleanup passed to the call and any
5875 later cleanups appended to the chain (but not yet discarded or
5879 make_cleanup (a, 0);
5881 struct cleanup *old = make_cleanup (b, 0);
5889 will call @code{c()} and @code{b()} but will not call @code{a()}. The
5890 cleanup that calls @code{a()} will remain in the cleanup chain, and will
5891 be done later unless otherwise discarded.@refill
5893 Your function should explicitly do or discard the cleanups it creates.
5894 Failing to do this leads to non-deterministic behavior since the caller
5895 will arbitrarily do or discard your functions cleanups. This need leads
5896 to two common cleanup styles.
5898 The first style is try/finally. Before it exits, your code-block calls
5899 @code{do_cleanups} with the old cleanup chain and thus ensures that your
5900 code-block's cleanups are always performed. For instance, the following
5901 code-segment avoids a memory leak problem (even when @code{error} is
5902 called and a forced stack unwind occurs) by ensuring that the
5903 @code{xfree} will always be called:
5906 struct cleanup *old = make_cleanup (null_cleanup, 0);
5907 data = xmalloc (sizeof blah);
5908 make_cleanup (xfree, data);
5913 The second style is try/except. Before it exits, your code-block calls
5914 @code{discard_cleanups} with the old cleanup chain and thus ensures that
5915 any created cleanups are not performed. For instance, the following
5916 code segment, ensures that the file will be closed but only if there is
5920 FILE *file = fopen ("afile", "r");
5921 struct cleanup *old = make_cleanup (close_file, file);
5923 discard_cleanups (old);
5927 Some functions, e.g., @code{fputs_filtered()} or @code{error()}, specify
5928 that they ``should not be called when cleanups are not in place''. This
5929 means that any actions you need to reverse in the case of an error or
5930 interruption must be on the cleanup chain before you call these
5931 functions, since they might never return to your code (they
5932 @samp{longjmp} instead).
5934 @section Per-architecture module data
5935 @cindex per-architecture module data
5936 @cindex multi-arch data
5937 @cindex data-pointer, per-architecture/per-module
5939 The multi-arch framework includes a mechanism for adding module
5940 specific per-architecture data-pointers to the @code{struct gdbarch}
5941 architecture object.
5943 A module registers one or more per-architecture data-pointers using:
5945 @deftypefn {Architecture Function} {struct gdbarch_data *} gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
5946 @var{pre_init} is used to, on-demand, allocate an initial value for a
5947 per-architecture data-pointer using the architecture's obstack (passed
5948 in as a parameter). Since @var{pre_init} can be called during
5949 architecture creation, it is not parameterized with the architecture.
5950 and must not call modules that use per-architecture data.
5953 @deftypefn {Architecture Function} {struct gdbarch_data *} gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
5954 @var{post_init} is used to obtain an initial value for a
5955 per-architecture data-pointer @emph{after}. Since @var{post_init} is
5956 always called after architecture creation, it both receives the fully
5957 initialized architecture and is free to call modules that use
5958 per-architecture data (care needs to be taken to ensure that those
5959 other modules do not try to call back to this module as that will
5960 create in cycles in the initialization call graph).
5963 These functions return a @code{struct gdbarch_data} that is used to
5964 identify the per-architecture data-pointer added for that module.
5966 The per-architecture data-pointer is accessed using the function:
5968 @deftypefn {Architecture Function} {void *} gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
5969 Given the architecture @var{arch} and module data handle
5970 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
5971 or @code{gdbarch_data_register_post_init}), this function returns the
5972 current value of the per-architecture data-pointer. If the data
5973 pointer is @code{NULL}, it is first initialized by calling the
5974 corresponding @var{pre_init} or @var{post_init} method.
5977 The examples below assume the following definitions:
5980 struct nozel @{ int total; @};
5981 static struct gdbarch_data *nozel_handle;
5984 A module can extend the architecture vector, adding additional
5985 per-architecture data, using the @var{pre_init} method. The module's
5986 per-architecture data is then initialized during architecture
5989 In the below, the module's per-architecture @emph{nozel} is added. An
5990 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
5991 from @code{gdbarch_init}.
5995 nozel_pre_init (struct obstack *obstack)
5997 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
6004 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
6006 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
6007 data->total = nozel;
6011 A module can on-demand create architecture dependent data structures
6012 using @code{post_init}.
6014 In the below, the nozel's total is computed on-demand by
6015 @code{nozel_post_init} using information obtained from the
6020 nozel_post_init (struct gdbarch *gdbarch)
6022 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
6023 nozel->total = gdbarch@dots{} (gdbarch);
6030 nozel_total (struct gdbarch *gdbarch)
6032 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
6037 @section Wrapping Output Lines
6038 @cindex line wrap in output
6041 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
6042 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
6043 added in places that would be good breaking points. The utility
6044 routines will take care of actually wrapping if the line width is
6047 The argument to @code{wrap_here} is an indentation string which is
6048 printed @emph{only} if the line breaks there. This argument is saved
6049 away and used later. It must remain valid until the next call to
6050 @code{wrap_here} or until a newline has been printed through the
6051 @code{*_filtered} functions. Don't pass in a local variable and then
6054 It is usually best to call @code{wrap_here} after printing a comma or
6055 space. If you call it before printing a space, make sure that your
6056 indentation properly accounts for the leading space that will print if
6057 the line wraps there.
6059 Any function or set of functions that produce filtered output must
6060 finish by printing a newline, to flush the wrap buffer, before switching
6061 to unfiltered (@code{printf}) output. Symbol reading routines that
6062 print warnings are a good example.
6064 @section @value{GDBN} Coding Standards
6065 @cindex coding standards
6067 @value{GDBN} follows the GNU coding standards, as described in
6068 @file{etc/standards.texi}. This file is also available for anonymous
6069 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
6070 of the standard; in general, when the GNU standard recommends a practice
6071 but does not require it, @value{GDBN} requires it.
6073 @value{GDBN} follows an additional set of coding standards specific to
6074 @value{GDBN}, as described in the following sections.
6079 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
6082 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
6085 @subsection Memory Management
6087 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
6088 @code{calloc}, @code{free} and @code{asprintf}.
6090 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
6091 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
6092 these functions do not return when the memory pool is empty. Instead,
6093 they unwind the stack using cleanups. These functions return
6094 @code{NULL} when requested to allocate a chunk of memory of size zero.
6096 @emph{Pragmatics: By using these functions, the need to check every
6097 memory allocation is removed. These functions provide portable
6100 @value{GDBN} does not use the function @code{free}.
6102 @value{GDBN} uses the function @code{xfree} to return memory to the
6103 memory pool. Consistent with ISO-C, this function ignores a request to
6104 free a @code{NULL} pointer.
6106 @emph{Pragmatics: On some systems @code{free} fails when passed a
6107 @code{NULL} pointer.}
6109 @value{GDBN} can use the non-portable function @code{alloca} for the
6110 allocation of small temporary values (such as strings).
6112 @emph{Pragmatics: This function is very non-portable. Some systems
6113 restrict the memory being allocated to no more than a few kilobytes.}
6115 @value{GDBN} uses the string function @code{xstrdup} and the print
6116 function @code{xstrprintf}.
6118 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
6119 functions such as @code{sprintf} are very prone to buffer overflow
6123 @subsection Compiler Warnings
6124 @cindex compiler warnings
6126 With few exceptions, developers should avoid the configuration option
6127 @samp{--disable-werror} when building @value{GDBN}. The exceptions
6128 are listed in the file @file{gdb/MAINTAINERS}. The default, when
6129 building with @sc{gcc}, is @samp{--enable-werror}.
6131 This option causes @value{GDBN} (when built using GCC) to be compiled
6132 with a carefully selected list of compiler warning flags. Any warnings
6133 from those flags are treated as errors.
6135 The current list of warning flags includes:
6139 Recommended @sc{gcc} warnings.
6141 @item -Wdeclaration-after-statement
6143 @sc{gcc} 3.x (and later) and @sc{c99} allow declarations mixed with
6144 code, but @sc{gcc} 2.x and @sc{c89} do not.
6146 @item -Wpointer-arith
6148 @item -Wformat-nonliteral
6149 Non-literal format strings, with a few exceptions, are bugs - they
6150 might contain unintended user-supplied format specifiers.
6151 Since @value{GDBN} uses the @code{format printf} attribute on all
6152 @code{printf} like functions this checks not just @code{printf} calls
6153 but also calls to functions such as @code{fprintf_unfiltered}.
6155 @item -Wno-pointer-sign
6156 In version 4.0, GCC began warning about pointer argument passing or
6157 assignment even when the source and destination differed only in
6158 signedness. However, most @value{GDBN} code doesn't distinguish
6159 carefully between @code{char} and @code{unsigned char}. In early 2006
6160 the @value{GDBN} developers decided correcting these warnings wasn't
6161 worth the time it would take.
6163 @item -Wno-unused-parameter
6164 Due to the way that @value{GDBN} is implemented many functions have
6165 unused parameters. Consequently this warning is avoided. The macro
6166 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
6167 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
6172 @itemx -Wno-char-subscripts
6173 These are warnings which might be useful for @value{GDBN}, but are
6174 currently too noisy to enable with @samp{-Werror}.
6178 @subsection Formatting
6180 @cindex source code formatting
6181 The standard GNU recommendations for formatting must be followed
6184 A function declaration should not have its name in column zero. A
6185 function definition should have its name in column zero.
6189 static void foo (void);
6197 @emph{Pragmatics: This simplifies scripting. Function definitions can
6198 be found using @samp{^function-name}.}
6200 There must be a space between a function or macro name and the opening
6201 parenthesis of its argument list (except for macro definitions, as
6202 required by C). There must not be a space after an open paren/bracket
6203 or before a close paren/bracket.
6205 While additional whitespace is generally helpful for reading, do not use
6206 more than one blank line to separate blocks, and avoid adding whitespace
6207 after the end of a program line (as of 1/99, some 600 lines had
6208 whitespace after the semicolon). Excess whitespace causes difficulties
6209 for @code{diff} and @code{patch} utilities.
6211 Pointers are declared using the traditional K&R C style:
6225 @subsection Comments
6227 @cindex comment formatting
6228 The standard GNU requirements on comments must be followed strictly.
6230 Block comments must appear in the following form, with no @code{/*}- or
6231 @code{*/}-only lines, and no leading @code{*}:
6234 /* Wait for control to return from inferior to debugger. If inferior
6235 gets a signal, we may decide to start it up again instead of
6236 returning. That is why there is a loop in this function. When
6237 this function actually returns it means the inferior should be left
6238 stopped and @value{GDBN} should read more commands. */
6241 (Note that this format is encouraged by Emacs; tabbing for a multi-line
6242 comment works correctly, and @kbd{M-q} fills the block consistently.)
6244 Put a blank line between the block comments preceding function or
6245 variable definitions, and the definition itself.
6247 In general, put function-body comments on lines by themselves, rather
6248 than trying to fit them into the 20 characters left at the end of a
6249 line, since either the comment or the code will inevitably get longer
6250 than will fit, and then somebody will have to move it anyhow.
6254 @cindex C data types
6255 Code must not depend on the sizes of C data types, the format of the
6256 host's floating point numbers, the alignment of anything, or the order
6257 of evaluation of expressions.
6259 @cindex function usage
6260 Use functions freely. There are only a handful of compute-bound areas
6261 in @value{GDBN} that might be affected by the overhead of a function
6262 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
6263 limited by the target interface (whether serial line or system call).
6265 However, use functions with moderation. A thousand one-line functions
6266 are just as hard to understand as a single thousand-line function.
6268 @emph{Macros are bad, M'kay.}
6269 (But if you have to use a macro, make sure that the macro arguments are
6270 protected with parentheses.)
6274 Declarations like @samp{struct foo *} should be used in preference to
6275 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
6278 @subsection Function Prototypes
6279 @cindex function prototypes
6281 Prototypes must be used when both @emph{declaring} and @emph{defining}
6282 a function. Prototypes for @value{GDBN} functions must include both the
6283 argument type and name, with the name matching that used in the actual
6284 function definition.
6286 All external functions should have a declaration in a header file that
6287 callers include, except for @code{_initialize_*} functions, which must
6288 be external so that @file{init.c} construction works, but shouldn't be
6289 visible to random source files.
6291 Where a source file needs a forward declaration of a static function,
6292 that declaration must appear in a block near the top of the source file.
6295 @subsection Internal Error Recovery
6297 During its execution, @value{GDBN} can encounter two types of errors.
6298 User errors and internal errors. User errors include not only a user
6299 entering an incorrect command but also problems arising from corrupt
6300 object files and system errors when interacting with the target.
6301 Internal errors include situations where @value{GDBN} has detected, at
6302 run time, a corrupt or erroneous situation.
6304 When reporting an internal error, @value{GDBN} uses
6305 @code{internal_error} and @code{gdb_assert}.
6307 @value{GDBN} must not call @code{abort} or @code{assert}.
6309 @emph{Pragmatics: There is no @code{internal_warning} function. Either
6310 the code detected a user error, recovered from it and issued a
6311 @code{warning} or the code failed to correctly recover from the user
6312 error and issued an @code{internal_error}.}
6314 @subsection File Names
6316 Any file used when building the core of @value{GDBN} must be in lower
6317 case. Any file used when building the core of @value{GDBN} must be 8.3
6318 unique. These requirements apply to both source and generated files.
6320 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
6321 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
6322 is introduced to the build process both @file{Makefile.in} and
6323 @file{configure.in} need to be modified accordingly. Compare the
6324 convoluted conversion process needed to transform @file{COPYING} into
6325 @file{copying.c} with the conversion needed to transform
6326 @file{version.in} into @file{version.c}.}
6328 Any file non 8.3 compliant file (that is not used when building the core
6329 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
6331 @emph{Pragmatics: This is clearly a compromise.}
6333 When @value{GDBN} has a local version of a system header file (ex
6334 @file{string.h}) the file name based on the POSIX header prefixed with
6335 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
6336 independent: they should use only macros defined by @file{configure},
6337 the compiler, or the host; they should include only system headers; they
6338 should refer only to system types. They may be shared between multiple
6339 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
6341 For other files @samp{-} is used as the separator.
6344 @subsection Include Files
6346 A @file{.c} file should include @file{defs.h} first.
6348 A @file{.c} file should directly include the @code{.h} file of every
6349 declaration and/or definition it directly refers to. It cannot rely on
6352 A @file{.h} file should directly include the @code{.h} file of every
6353 declaration and/or definition it directly refers to. It cannot rely on
6354 indirect inclusion. Exception: The file @file{defs.h} does not need to
6355 be directly included.
6357 An external declaration should only appear in one include file.
6359 An external declaration should never appear in a @code{.c} file.
6360 Exception: a declaration for the @code{_initialize} function that
6361 pacifies @option{-Wmissing-declaration}.
6363 A @code{typedef} definition should only appear in one include file.
6365 An opaque @code{struct} declaration can appear in multiple @file{.h}
6366 files. Where possible, a @file{.h} file should use an opaque
6367 @code{struct} declaration instead of an include.
6369 All @file{.h} files should be wrapped in:
6372 #ifndef INCLUDE_FILE_NAME_H
6373 #define INCLUDE_FILE_NAME_H
6379 @subsection Clean Design and Portable Implementation
6382 In addition to getting the syntax right, there's the little question of
6383 semantics. Some things are done in certain ways in @value{GDBN} because long
6384 experience has shown that the more obvious ways caused various kinds of
6387 @cindex assumptions about targets
6388 You can't assume the byte order of anything that comes from a target
6389 (including @var{value}s, object files, and instructions). Such things
6390 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
6391 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
6392 such as @code{bfd_get_32}.
6394 You can't assume that you know what interface is being used to talk to
6395 the target system. All references to the target must go through the
6396 current @code{target_ops} vector.
6398 You can't assume that the host and target machines are the same machine
6399 (except in the ``native'' support modules). In particular, you can't
6400 assume that the target machine's header files will be available on the
6401 host machine. Target code must bring along its own header files --
6402 written from scratch or explicitly donated by their owner, to avoid
6406 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
6407 to write the code portably than to conditionalize it for various
6410 @cindex system dependencies
6411 New @code{#ifdef}'s which test for specific compilers or manufacturers
6412 or operating systems are unacceptable. All @code{#ifdef}'s should test
6413 for features. The information about which configurations contain which
6414 features should be segregated into the configuration files. Experience
6415 has proven far too often that a feature unique to one particular system
6416 often creeps into other systems; and that a conditional based on some
6417 predefined macro for your current system will become worthless over
6418 time, as new versions of your system come out that behave differently
6419 with regard to this feature.
6421 Adding code that handles specific architectures, operating systems,
6422 target interfaces, or hosts, is not acceptable in generic code.
6424 @cindex portable file name handling
6425 @cindex file names, portability
6426 One particularly notorious area where system dependencies tend to
6427 creep in is handling of file names. The mainline @value{GDBN} code
6428 assumes Posix semantics of file names: absolute file names begin with
6429 a forward slash @file{/}, slashes are used to separate leading
6430 directories, case-sensitive file names. These assumptions are not
6431 necessarily true on non-Posix systems such as MS-Windows. To avoid
6432 system-dependent code where you need to take apart or construct a file
6433 name, use the following portable macros:
6436 @findex HAVE_DOS_BASED_FILE_SYSTEM
6437 @item HAVE_DOS_BASED_FILE_SYSTEM
6438 This preprocessing symbol is defined to a non-zero value on hosts
6439 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
6440 symbol to write conditional code which should only be compiled for
6443 @findex IS_DIR_SEPARATOR
6444 @item IS_DIR_SEPARATOR (@var{c})
6445 Evaluates to a non-zero value if @var{c} is a directory separator
6446 character. On Unix and GNU/Linux systems, only a slash @file{/} is
6447 such a character, but on Windows, both @file{/} and @file{\} will
6450 @findex IS_ABSOLUTE_PATH
6451 @item IS_ABSOLUTE_PATH (@var{file})
6452 Evaluates to a non-zero value if @var{file} is an absolute file name.
6453 For Unix and GNU/Linux hosts, a name which begins with a slash
6454 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
6455 @file{x:\bar} are also absolute file names.
6457 @findex FILENAME_CMP
6458 @item FILENAME_CMP (@var{f1}, @var{f2})
6459 Calls a function which compares file names @var{f1} and @var{f2} as
6460 appropriate for the underlying host filesystem. For Posix systems,
6461 this simply calls @code{strcmp}; on case-insensitive filesystems it
6462 will call @code{strcasecmp} instead.
6464 @findex DIRNAME_SEPARATOR
6465 @item DIRNAME_SEPARATOR
6466 Evaluates to a character which separates directories in
6467 @code{PATH}-style lists, typically held in environment variables.
6468 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
6470 @findex SLASH_STRING
6472 This evaluates to a constant string you should use to produce an
6473 absolute filename from leading directories and the file's basename.
6474 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
6475 @code{"\\"} for some Windows-based ports.
6478 In addition to using these macros, be sure to use portable library
6479 functions whenever possible. For example, to extract a directory or a
6480 basename part from a file name, use the @code{dirname} and
6481 @code{basename} library functions (available in @code{libiberty} for
6482 platforms which don't provide them), instead of searching for a slash
6483 with @code{strrchr}.
6485 Another way to generalize @value{GDBN} along a particular interface is with an
6486 attribute struct. For example, @value{GDBN} has been generalized to handle
6487 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
6488 by defining the @code{target_ops} structure and having a current target (as
6489 well as a stack of targets below it, for memory references). Whenever
6490 something needs to be done that depends on which remote interface we are
6491 using, a flag in the current target_ops structure is tested (e.g.,
6492 @code{target_has_stack}), or a function is called through a pointer in the
6493 current target_ops structure. In this way, when a new remote interface
6494 is added, only one module needs to be touched---the one that actually
6495 implements the new remote interface. Other examples of
6496 attribute-structs are BFD access to multiple kinds of object file
6497 formats, or @value{GDBN}'s access to multiple source languages.
6499 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
6500 the code interfacing between @code{ptrace} and the rest of
6501 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
6502 something was very painful. In @value{GDBN} 4.x, these have all been
6503 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
6504 with variations between systems the same way any system-independent
6505 file would (hooks, @code{#if defined}, etc.), and machines which are
6506 radically different don't need to use @file{infptrace.c} at all.
6508 All debugging code must be controllable using the @samp{set debug
6509 @var{module}} command. Do not use @code{printf} to print trace
6510 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
6511 @code{#ifdef DEBUG}.
6516 @chapter Porting @value{GDBN}
6517 @cindex porting to new machines
6519 Most of the work in making @value{GDBN} compile on a new machine is in
6520 specifying the configuration of the machine. Porting a new
6521 architecture to @value{GDBN} can be broken into a number of steps.
6526 Ensure a @sc{bfd} exists for executables of the target architecture in
6527 the @file{bfd} directory. If one does not exist, create one by
6528 modifying an existing similar one.
6531 Implement a disassembler for the target architecture in the @file{opcodes}
6535 Define the target architecture in the @file{gdb} directory
6536 (@pxref{Adding a New Target, , Adding a New Target}). Add the pattern
6537 for the new target to @file{configure.tgt} with the names of the files
6538 that contain the code. By convention the target architecture
6539 definition for an architecture @var{arch} is placed in
6540 @file{@var{arch}-tdep.c}.
6542 Within @file{@var{arch}-tdep.c} define the function
6543 @code{_initialize_@var{arch}_tdep} which calls
6544 @code{gdbarch_register} to create the new @code{@w{struct
6545 gdbarch}} for the architecture.
6548 If a new remote target is needed, consider adding a new remote target
6549 by defining a function
6550 @code{_initialize_remote_@var{arch}}. However if at all possible
6551 use the @value{GDBN} @emph{Remote Serial Protocol} for this and implement
6552 the server side protocol independently with the target.
6555 If desired implement a simulator in the @file{sim} directory. This
6556 should create the library @file{libsim.a} implementing the interface
6557 in @file{remote-sim.h} (found in the @file{include} directory).
6560 Build and test. If desired, lobby the @sc{gdb} steering group to
6561 have the new port included in the main distribution!
6564 Add a description of the new architecture to the main @value{GDBN} user
6565 guide (@pxref{Configuration Specific Information, , Configuration
6566 Specific Information, gdb, Debugging with @value{GDBN}}).
6570 @node Versions and Branches
6571 @chapter Versions and Branches
6575 @value{GDBN}'s version is determined by the file
6576 @file{gdb/version.in} and takes one of the following forms:
6579 @item @var{major}.@var{minor}
6580 @itemx @var{major}.@var{minor}.@var{patchlevel}
6581 an official release (e.g., 6.2 or 6.2.1)
6582 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}
6583 a snapshot taken at @var{YYYY}-@var{MM}-@var{DD}-gmt (e.g.,
6584 6.1.50.20020302, 6.1.90.20020304, or 6.1.0.20020308)
6585 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}-cvs
6586 a @sc{cvs} check out drawn on @var{YYYY}-@var{MM}-@var{DD} (e.g.,
6587 6.1.50.20020302-cvs, 6.1.90.20020304-cvs, or 6.1.0.20020308-cvs)
6588 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD} (@var{vendor})
6589 a vendor specific release of @value{GDBN}, that while based on@*
6590 @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD},
6591 may include additional changes
6594 @value{GDBN}'s mainline uses the @var{major} and @var{minor} version
6595 numbers from the most recent release branch, with a @var{patchlevel}
6596 of 50. At the time each new release branch is created, the mainline's
6597 @var{major} and @var{minor} version numbers are updated.
6599 @value{GDBN}'s release branch is similar. When the branch is cut, the
6600 @var{patchlevel} is changed from 50 to 90. As draft releases are
6601 drawn from the branch, the @var{patchlevel} is incremented. Once the
6602 first release (@var{major}.@var{minor}) has been made, the
6603 @var{patchlevel} is set to 0 and updates have an incremented
6606 For snapshots, and @sc{cvs} check outs, it is also possible to
6607 identify the @sc{cvs} origin:
6610 @item @var{major}.@var{minor}.50.@var{YYYY}@var{MM}@var{DD}
6611 drawn from the @sc{head} of mainline @sc{cvs} (e.g., 6.1.50.20020302)
6612 @item @var{major}.@var{minor}.90.@var{YYYY}@var{MM}@var{DD}
6613 @itemx @var{major}.@var{minor}.91.@var{YYYY}@var{MM}@var{DD} @dots{}
6614 drawn from a release branch prior to the release (e.g.,
6616 @item @var{major}.@var{minor}.0.@var{YYYY}@var{MM}@var{DD}
6617 @itemx @var{major}.@var{minor}.1.@var{YYYY}@var{MM}@var{DD} @dots{}
6618 drawn from a release branch after the release (e.g., 6.2.0.20020308)
6621 If the previous @value{GDBN} version is 6.1 and the current version is
6622 6.2, then, substituting 6 for @var{major} and 1 or 2 for @var{minor},
6623 here's an illustration of a typical sequence:
6630 +--------------------------.
6633 6.2.50.20020303-cvs 6.1.90 (draft #1)
6635 6.2.50.20020304-cvs 6.1.90.20020304-cvs
6637 6.2.50.20020305-cvs 6.1.91 (draft #2)
6639 6.2.50.20020306-cvs 6.1.91.20020306-cvs
6641 6.2.50.20020307-cvs 6.2 (release)
6643 6.2.50.20020308-cvs 6.2.0.20020308-cvs
6645 6.2.50.20020309-cvs 6.2.1 (update)
6647 6.2.50.20020310-cvs <branch closed>
6651 +--------------------------.
6654 6.3.50.20020312-cvs 6.2.90 (draft #1)
6658 @section Release Branches
6659 @cindex Release Branches
6661 @value{GDBN} draws a release series (6.2, 6.2.1, @dots{}) from a
6662 single release branch, and identifies that branch using the @sc{cvs}
6666 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-branchpoint
6667 gdb_@var{major}_@var{minor}-branch
6668 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-release
6671 @emph{Pragmatics: To help identify the date at which a branch or
6672 release is made, both the branchpoint and release tags include the
6673 date that they are cut (@var{YYYY}@var{MM}@var{DD}) in the tag. The
6674 branch tag, denoting the head of the branch, does not need this.}
6676 @section Vendor Branches
6677 @cindex vendor branches
6679 To avoid version conflicts, vendors are expected to modify the file
6680 @file{gdb/version.in} to include a vendor unique alphabetic identifier
6681 (an official @value{GDBN} release never uses alphabetic characters in
6682 its version identifier). E.g., @samp{6.2widgit2}, or @samp{6.2 (Widgit
6685 @section Experimental Branches
6686 @cindex experimental branches
6688 @subsection Guidelines
6690 @value{GDBN} permits the creation of branches, cut from the @sc{cvs}
6691 repository, for experimental development. Branches make it possible
6692 for developers to share preliminary work, and maintainers to examine
6693 significant new developments.
6695 The following are a set of guidelines for creating such branches:
6699 @item a branch has an owner
6700 The owner can set further policy for a branch, but may not change the
6701 ground rules. In particular, they can set a policy for commits (be it
6702 adding more reviewers or deciding who can commit).
6704 @item all commits are posted
6705 All changes committed to a branch shall also be posted to
6706 @email{gdb-patches@@sourceware.org, the @value{GDBN} patches
6707 mailing list}. While commentary on such changes are encouraged, people
6708 should remember that the changes only apply to a branch.
6710 @item all commits are covered by an assignment
6711 This ensures that all changes belong to the Free Software Foundation,
6712 and avoids the possibility that the branch may become contaminated.
6714 @item a branch is focused
6715 A focused branch has a single objective or goal, and does not contain
6716 unnecessary or irrelevant changes. Cleanups, where identified, being
6717 be pushed into the mainline as soon as possible.
6719 @item a branch tracks mainline
6720 This keeps the level of divergence under control. It also keeps the
6721 pressure on developers to push cleanups and other stuff into the
6724 @item a branch shall contain the entire @value{GDBN} module
6725 The @value{GDBN} module @code{gdb} should be specified when creating a
6726 branch (branches of individual files should be avoided). @xref{Tags}.
6728 @item a branch shall be branded using @file{version.in}
6729 The file @file{gdb/version.in} shall be modified so that it identifies
6730 the branch @var{owner} and branch @var{name}, e.g.,
6731 @samp{6.2.50.20030303_owner_name} or @samp{6.2 (Owner Name)}.
6738 To simplify the identification of @value{GDBN} branches, the following
6739 branch tagging convention is strongly recommended:
6743 @item @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
6744 @itemx @var{owner}_@var{name}-@var{YYYYMMDD}-branch
6745 The branch point and corresponding branch tag. @var{YYYYMMDD} is the
6746 date that the branch was created. A branch is created using the
6747 sequence: @anchor{experimental branch tags}
6749 cvs rtag @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint gdb
6750 cvs rtag -b -r @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint \
6751 @var{owner}_@var{name}-@var{YYYYMMDD}-branch gdb
6754 @item @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
6755 The tagged point, on the mainline, that was used when merging the branch
6756 on @var{yyyymmdd}. To merge in all changes since the branch was cut,
6757 use a command sequence like:
6759 cvs rtag @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint gdb
6761 -j@var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
6762 -j@var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
6765 Similar sequences can be used to just merge in changes since the last
6771 For further information on @sc{cvs}, see
6772 @uref{http://www.gnu.org/software/cvs/, Concurrent Versions System}.
6774 @node Start of New Year Procedure
6775 @chapter Start of New Year Procedure
6776 @cindex new year procedure
6778 At the start of each new year, the following actions should be performed:
6782 Rotate the ChangeLog file
6784 The current @file{ChangeLog} file should be renamed into
6785 @file{ChangeLog-YYYY} where YYYY is the year that has just passed.
6786 A new @file{ChangeLog} file should be created, and its contents should
6787 contain a reference to the previous ChangeLog. The following should
6788 also be preserved at the end of the new ChangeLog, in order to provide
6789 the appropriate settings when editing this file with Emacs:
6795 version-control: never
6801 Add an entry for the newly created ChangeLog file (@file{ChangeLog-YYYY})
6802 in @file{gdb/config/djgpp/fnchange.lst}.
6805 Update the copyright year in the startup message
6807 Update the copyright year in:
6809 @item file @file{top.c}, function @code{print_gdb_version}
6810 @item file @file{gdbserver/server.c}, function @code{gdbserver_version}
6811 @item file @file{gdbserver/gdbreplay.c}, function @code{gdbreplay_version}
6815 Add the new year in the copyright notices of all source and documentation
6816 files. This can be done semi-automatically by running the @code{copyright.sh}
6817 script. This script requires Emacs 22 or later to be installed.
6823 @chapter Releasing @value{GDBN}
6824 @cindex making a new release of gdb
6826 @section Branch Commit Policy
6828 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
6829 5.1 and 5.2 all used the below:
6833 The @file{gdb/MAINTAINERS} file still holds.
6835 Don't fix something on the branch unless/until it is also fixed in the
6836 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
6837 file is better than committing a hack.
6839 When considering a patch for the branch, suggested criteria include:
6840 Does it fix a build? Does it fix the sequence @kbd{break main; run}
6841 when debugging a static binary?
6843 The further a change is from the core of @value{GDBN}, the less likely
6844 the change will worry anyone (e.g., target specific code).
6846 Only post a proposal to change the core of @value{GDBN} after you've
6847 sent individual bribes to all the people listed in the
6848 @file{MAINTAINERS} file @t{;-)}
6851 @emph{Pragmatics: Provided updates are restricted to non-core
6852 functionality there is little chance that a broken change will be fatal.
6853 This means that changes such as adding a new architectures or (within
6854 reason) support for a new host are considered acceptable.}
6857 @section Obsoleting code
6859 Before anything else, poke the other developers (and around the source
6860 code) to see if there is anything that can be removed from @value{GDBN}
6861 (an old target, an unused file).
6863 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
6864 line. Doing this means that it is easy to identify something that has
6865 been obsoleted when greping through the sources.
6867 The process is done in stages --- this is mainly to ensure that the
6868 wider @value{GDBN} community has a reasonable opportunity to respond.
6869 Remember, everything on the Internet takes a week.
6873 Post the proposal on @email{gdb@@sourceware.org, the GDB mailing
6874 list} Creating a bug report to track the task's state, is also highly
6879 Post the proposal on @email{gdb-announce@@sourceware.org, the GDB
6880 Announcement mailing list}.
6884 Go through and edit all relevant files and lines so that they are
6885 prefixed with the word @code{OBSOLETE}.
6887 Wait until the next GDB version, containing this obsolete code, has been
6890 Remove the obsolete code.
6894 @emph{Maintainer note: While removing old code is regrettable it is
6895 hopefully better for @value{GDBN}'s long term development. Firstly it
6896 helps the developers by removing code that is either no longer relevant
6897 or simply wrong. Secondly since it removes any history associated with
6898 the file (effectively clearing the slate) the developer has a much freer
6899 hand when it comes to fixing broken files.}
6903 @section Before the Branch
6905 The most important objective at this stage is to find and fix simple
6906 changes that become a pain to track once the branch is created. For
6907 instance, configuration problems that stop @value{GDBN} from even
6908 building. If you can't get the problem fixed, document it in the
6909 @file{gdb/PROBLEMS} file.
6911 @subheading Prompt for @file{gdb/NEWS}
6913 People always forget. Send a post reminding them but also if you know
6914 something interesting happened add it yourself. The @code{schedule}
6915 script will mention this in its e-mail.
6917 @subheading Review @file{gdb/README}
6919 Grab one of the nightly snapshots and then walk through the
6920 @file{gdb/README} looking for anything that can be improved. The
6921 @code{schedule} script will mention this in its e-mail.
6923 @subheading Refresh any imported files.
6925 A number of files are taken from external repositories. They include:
6929 @file{texinfo/texinfo.tex}
6931 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
6934 @file{etc/standards.texi}, @file{etc/make-stds.texi}
6937 @subheading Check the ARI
6939 @uref{http://sourceware.org/gdb/ari,,A.R.I.} is an @code{awk} script
6940 (Awk Regression Index ;-) that checks for a number of errors and coding
6941 conventions. The checks include things like using @code{malloc} instead
6942 of @code{xmalloc} and file naming problems. There shouldn't be any
6945 @subsection Review the bug data base
6947 Close anything obviously fixed.
6949 @subsection Check all cross targets build
6951 The targets are listed in @file{gdb/MAINTAINERS}.
6954 @section Cut the Branch
6956 @subheading Create the branch
6961 $ V=`echo $v | sed 's/\./_/g'`
6962 $ D=`date -u +%Y-%m-%d`
6965 $ echo cvs -f -d :ext:sourceware.org:/cvs/src rtag \
6966 -D $D-gmt gdb_$V-$D-branchpoint insight
6967 cvs -f -d :ext:sourceware.org:/cvs/src rtag
6968 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight
6971 $ echo cvs -f -d :ext:sourceware.org:/cvs/src rtag \
6972 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight
6973 cvs -f -d :ext:sourceware.org:/cvs/src rtag \
6974 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight
6982 By using @kbd{-D YYYY-MM-DD-gmt}, the branch is forced to an exact
6985 The trunk is first tagged so that the branch point can easily be found.
6987 Insight, which includes @value{GDBN}, is tagged at the same time.
6989 @file{version.in} gets bumped to avoid version number conflicts.
6991 The reading of @file{.cvsrc} is disabled using @file{-f}.
6994 @subheading Update @file{version.in}
6999 $ V=`echo $v | sed 's/\./_/g'`
7003 $ echo cvs -f -d :ext:sourceware.org:/cvs/src co \
7004 -r gdb_$V-branch src/gdb/version.in
7005 cvs -f -d :ext:sourceware.org:/cvs/src co
7006 -r gdb_5_2-branch src/gdb/version.in
7008 U src/gdb/version.in
7010 $ echo $u.90-0000-00-00-cvs > version.in
7012 5.1.90-0000-00-00-cvs
7013 $ cvs -f commit version.in
7018 @file{0000-00-00} is used as a date to pump prime the version.in update
7021 @file{.90} and the previous branch version are used as fairly arbitrary
7022 initial branch version number.
7026 @subheading Update the web and news pages
7030 @subheading Tweak cron to track the new branch
7032 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
7033 This file needs to be updated so that:
7037 A daily timestamp is added to the file @file{version.in}.
7039 The new branch is included in the snapshot process.
7043 See the file @file{gdbadmin/cron/README} for how to install the updated
7046 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
7047 any changes. That file is copied to both the branch/ and current/
7048 snapshot directories.
7051 @subheading Update the NEWS and README files
7053 The @file{NEWS} file needs to be updated so that on the branch it refers
7054 to @emph{changes in the current release} while on the trunk it also
7055 refers to @emph{changes since the current release}.
7057 The @file{README} file needs to be updated so that it refers to the
7060 @subheading Post the branch info
7062 Send an announcement to the mailing lists:
7066 @email{gdb-announce@@sourceware.org, GDB Announcement mailing list}
7068 @email{gdb@@sourceware.org, GDB Discussion mailing list} and
7069 @email{gdb-testers@@sourceware.org, GDB Testers mailing list}
7072 @emph{Pragmatics: The branch creation is sent to the announce list to
7073 ensure that people people not subscribed to the higher volume discussion
7076 The announcement should include:
7082 How to check out the branch using CVS.
7084 The date/number of weeks until the release.
7086 The branch commit policy still holds.
7089 @section Stabilize the branch
7091 Something goes here.
7093 @section Create a Release
7095 The process of creating and then making available a release is broken
7096 down into a number of stages. The first part addresses the technical
7097 process of creating a releasable tar ball. The later stages address the
7098 process of releasing that tar ball.
7100 When making a release candidate just the first section is needed.
7102 @subsection Create a release candidate
7104 The objective at this stage is to create a set of tar balls that can be
7105 made available as a formal release (or as a less formal release
7108 @subsubheading Freeze the branch
7110 Send out an e-mail notifying everyone that the branch is frozen to
7111 @email{gdb-patches@@sourceware.org}.
7113 @subsubheading Establish a few defaults.
7118 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
7120 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
7124 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
7126 /home/gdbadmin/bin/autoconf
7135 Check the @code{autoconf} version carefully. You want to be using the
7136 version taken from the @file{binutils} snapshot directory, which can be
7137 found at @uref{ftp://sourceware.org/pub/binutils/}. It is very
7138 unlikely that a system installed version of @code{autoconf} (e.g.,
7139 @file{/usr/bin/autoconf}) is correct.
7142 @subsubheading Check out the relevant modules:
7145 $ for m in gdb insight
7147 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
7157 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
7158 any confusion between what is written here and what your local
7159 @code{cvs} really does.
7162 @subsubheading Update relevant files.
7168 Major releases get their comments added as part of the mainline. Minor
7169 releases should probably mention any significant bugs that were fixed.
7171 Don't forget to include the @file{ChangeLog} entry.
7174 $ emacs gdb/src/gdb/NEWS
7179 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
7180 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
7185 You'll need to update:
7197 $ emacs gdb/src/gdb/README
7202 $ cp gdb/src/gdb/README insight/src/gdb/README
7203 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
7206 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
7207 before the initial branch was cut so just a simple substitute is needed
7210 @emph{Maintainer note: Other projects generate @file{README} and
7211 @file{INSTALL} from the core documentation. This might be worth
7214 @item gdb/version.in
7217 $ echo $v > gdb/src/gdb/version.in
7218 $ cat gdb/src/gdb/version.in
7220 $ emacs gdb/src/gdb/version.in
7223 ... Bump to version ...
7225 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
7226 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
7231 @subsubheading Do the dirty work
7233 This is identical to the process used to create the daily snapshot.
7236 $ for m in gdb insight
7238 ( cd $m/src && gmake -f src-release $m.tar )
7242 If the top level source directory does not have @file{src-release}
7243 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
7246 $ for m in gdb insight
7248 ( cd $m/src && gmake -f Makefile.in $m.tar )
7252 @subsubheading Check the source files
7254 You're looking for files that have mysteriously disappeared.
7255 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
7256 for the @file{version.in} update @kbd{cronjob}.
7259 $ ( cd gdb/src && cvs -f -q -n update )
7263 @dots{} lots of generated files @dots{}
7268 @dots{} lots of generated files @dots{}
7273 @emph{Don't worry about the @file{gdb.info-??} or
7274 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
7275 was also generated only something strange with CVS means that they
7276 didn't get suppressed). Fixing it would be nice though.}
7278 @subsubheading Create compressed versions of the release
7284 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
7285 $ for m in gdb insight
7287 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
7288 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
7298 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
7299 in that mode, @code{gzip} does not know the name of the file and, hence,
7300 can not include it in the compressed file. This is also why the release
7301 process runs @code{tar} and @code{bzip2} as separate passes.
7304 @subsection Sanity check the tar ball
7306 Pick a popular machine (Solaris/PPC?) and try the build on that.
7309 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
7314 $ ./gdb/gdb ./gdb/gdb
7318 Breakpoint 1 at 0x80732bc: file main.c, line 734.
7320 Starting program: /tmp/gdb-5.2/gdb/gdb
7322 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
7323 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
7325 $1 = @{argc = 136426532, argv = 0x821b7f0@}
7329 @subsection Make a release candidate available
7331 If this is a release candidate then the only remaining steps are:
7335 Commit @file{version.in} and @file{ChangeLog}
7337 Tweak @file{version.in} (and @file{ChangeLog} to read
7338 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
7339 process can restart.
7341 Make the release candidate available in
7342 @uref{ftp://sourceware.org/pub/gdb/snapshots/branch}
7344 Notify the relevant mailing lists ( @email{gdb@@sourceware.org} and
7345 @email{gdb-testers@@sourceware.org} that the candidate is available.
7348 @subsection Make a formal release available
7350 (And you thought all that was required was to post an e-mail.)
7352 @subsubheading Install on sware
7354 Copy the new files to both the release and the old release directory:
7357 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
7358 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
7362 Clean up the releases directory so that only the most recent releases
7363 are available (e.g.@: keep 5.2 and 5.2.1 but remove 5.1):
7366 $ cd ~ftp/pub/gdb/releases
7371 Update the file @file{README} and @file{.message} in the releases
7378 $ ln README .message
7381 @subsubheading Update the web pages.
7385 @item htdocs/download/ANNOUNCEMENT
7386 This file, which is posted as the official announcement, includes:
7389 General announcement.
7391 News. If making an @var{M}.@var{N}.1 release, retain the news from
7392 earlier @var{M}.@var{N} release.
7397 @item htdocs/index.html
7398 @itemx htdocs/news/index.html
7399 @itemx htdocs/download/index.html
7400 These files include:
7403 Announcement of the most recent release.
7405 News entry (remember to update both the top level and the news directory).
7407 These pages also need to be regenerate using @code{index.sh}.
7409 @item download/onlinedocs/
7410 You need to find the magic command that is used to generate the online
7411 docs from the @file{.tar.bz2}. The best way is to look in the output
7412 from one of the nightly @code{cron} jobs and then just edit accordingly.
7416 $ ~/ss/update-web-docs \
7417 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
7419 /www/sourceware/htdocs/gdb/download/onlinedocs \
7424 Just like the online documentation. Something like:
7427 $ /bin/sh ~/ss/update-web-ari \
7428 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
7430 /www/sourceware/htdocs/gdb/download/ari \
7436 @subsubheading Shadow the pages onto gnu
7438 Something goes here.
7441 @subsubheading Install the @value{GDBN} tar ball on GNU
7443 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
7444 @file{~ftp/gnu/gdb}.
7446 @subsubheading Make the @file{ANNOUNCEMENT}
7448 Post the @file{ANNOUNCEMENT} file you created above to:
7452 @email{gdb-announce@@sourceware.org, GDB Announcement mailing list}
7454 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
7455 day or so to let things get out)
7457 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
7462 The release is out but you're still not finished.
7464 @subsubheading Commit outstanding changes
7466 In particular you'll need to commit any changes to:
7470 @file{gdb/ChangeLog}
7472 @file{gdb/version.in}
7479 @subsubheading Tag the release
7484 $ d=`date -u +%Y-%m-%d`
7487 $ ( cd insight/src/gdb && cvs -f -q update )
7488 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
7491 Insight is used since that contains more of the release than
7494 @subsubheading Mention the release on the trunk
7496 Just put something in the @file{ChangeLog} so that the trunk also
7497 indicates when the release was made.
7499 @subsubheading Restart @file{gdb/version.in}
7501 If @file{gdb/version.in} does not contain an ISO date such as
7502 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
7503 committed all the release changes it can be set to
7504 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
7505 is important - it affects the snapshot process).
7507 Don't forget the @file{ChangeLog}.
7509 @subsubheading Merge into trunk
7511 The files committed to the branch may also need changes merged into the
7514 @subsubheading Revise the release schedule
7516 Post a revised release schedule to @email{gdb@@sourceware.org, GDB
7517 Discussion List} with an updated announcement. The schedule can be
7518 generated by running:
7521 $ ~/ss/schedule `date +%s` schedule
7525 The first parameter is approximate date/time in seconds (from the epoch)
7526 of the most recent release.
7528 Also update the schedule @code{cronjob}.
7530 @section Post release
7532 Remove any @code{OBSOLETE} code.
7539 The testsuite is an important component of the @value{GDBN} package.
7540 While it is always worthwhile to encourage user testing, in practice
7541 this is rarely sufficient; users typically use only a small subset of
7542 the available commands, and it has proven all too common for a change
7543 to cause a significant regression that went unnoticed for some time.
7545 The @value{GDBN} testsuite uses the DejaGNU testing framework. The
7546 tests themselves are calls to various @code{Tcl} procs; the framework
7547 runs all the procs and summarizes the passes and fails.
7549 @section Using the Testsuite
7551 @cindex running the test suite
7552 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
7553 testsuite's objdir) and type @code{make check}. This just sets up some
7554 environment variables and invokes DejaGNU's @code{runtest} script. While
7555 the testsuite is running, you'll get mentions of which test file is in use,
7556 and a mention of any unexpected passes or fails. When the testsuite is
7557 finished, you'll get a summary that looks like this:
7562 # of expected passes 6016
7563 # of unexpected failures 58
7564 # of unexpected successes 5
7565 # of expected failures 183
7566 # of unresolved testcases 3
7567 # of untested testcases 5
7570 To run a specific test script, type:
7572 make check RUNTESTFLAGS='@var{tests}'
7574 where @var{tests} is a list of test script file names, separated by
7577 The ideal test run consists of expected passes only; however, reality
7578 conspires to keep us from this ideal. Unexpected failures indicate
7579 real problems, whether in @value{GDBN} or in the testsuite. Expected
7580 failures are still failures, but ones which have been decided are too
7581 hard to deal with at the time; for instance, a test case might work
7582 everywhere except on AIX, and there is no prospect of the AIX case
7583 being fixed in the near future. Expected failures should not be added
7584 lightly, since you may be masking serious bugs in @value{GDBN}.
7585 Unexpected successes are expected fails that are passing for some
7586 reason, while unresolved and untested cases often indicate some minor
7587 catastrophe, such as the compiler being unable to deal with a test
7590 When making any significant change to @value{GDBN}, you should run the
7591 testsuite before and after the change, to confirm that there are no
7592 regressions. Note that truly complete testing would require that you
7593 run the testsuite with all supported configurations and a variety of
7594 compilers; however this is more than really necessary. In many cases
7595 testing with a single configuration is sufficient. Other useful
7596 options are to test one big-endian (Sparc) and one little-endian (x86)
7597 host, a cross config with a builtin simulator (powerpc-eabi,
7598 mips-elf), or a 64-bit host (Alpha).
7600 If you add new functionality to @value{GDBN}, please consider adding
7601 tests for it as well; this way future @value{GDBN} hackers can detect
7602 and fix their changes that break the functionality you added.
7603 Similarly, if you fix a bug that was not previously reported as a test
7604 failure, please add a test case for it. Some cases are extremely
7605 difficult to test, such as code that handles host OS failures or bugs
7606 in particular versions of compilers, and it's OK not to try to write
7607 tests for all of those.
7609 DejaGNU supports separate build, host, and target machines. However,
7610 some @value{GDBN} test scripts do not work if the build machine and
7611 the host machine are not the same. In such an environment, these scripts
7612 will give a result of ``UNRESOLVED'', like this:
7615 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
7618 @section Testsuite Organization
7620 @cindex test suite organization
7621 The testsuite is entirely contained in @file{gdb/testsuite}. While the
7622 testsuite includes some makefiles and configury, these are very minimal,
7623 and used for little besides cleaning up, since the tests themselves
7624 handle the compilation of the programs that @value{GDBN} will run. The file
7625 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
7626 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
7627 configuration-specific files, typically used for special-purpose
7628 definitions of procs like @code{gdb_load} and @code{gdb_start}.
7630 The tests themselves are to be found in @file{testsuite/gdb.*} and
7631 subdirectories of those. The names of the test files must always end
7632 with @file{.exp}. DejaGNU collects the test files by wildcarding
7633 in the test directories, so both subdirectories and individual files
7634 get chosen and run in alphabetical order.
7636 The following table lists the main types of subdirectories and what they
7637 are for. Since DejaGNU finds test files no matter where they are
7638 located, and since each test file sets up its own compilation and
7639 execution environment, this organization is simply for convenience and
7644 This is the base testsuite. The tests in it should apply to all
7645 configurations of @value{GDBN} (but generic native-only tests may live here).
7646 The test programs should be in the subset of C that is valid K&R,
7647 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
7650 @item gdb.@var{lang}
7651 Language-specific tests for any language @var{lang} besides C. Examples are
7652 @file{gdb.cp} and @file{gdb.java}.
7654 @item gdb.@var{platform}
7655 Non-portable tests. The tests are specific to a specific configuration
7656 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
7659 @item gdb.@var{compiler}
7660 Tests specific to a particular compiler. As of this writing (June
7661 1999), there aren't currently any groups of tests in this category that
7662 couldn't just as sensibly be made platform-specific, but one could
7663 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
7666 @item gdb.@var{subsystem}
7667 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
7668 instance, @file{gdb.disasm} exercises various disassemblers, while
7669 @file{gdb.stabs} tests pathways through the stabs symbol reader.
7672 @section Writing Tests
7673 @cindex writing tests
7675 In many areas, the @value{GDBN} tests are already quite comprehensive; you
7676 should be able to copy existing tests to handle new cases.
7678 You should try to use @code{gdb_test} whenever possible, since it
7679 includes cases to handle all the unexpected errors that might happen.
7680 However, it doesn't cost anything to add new test procedures; for
7681 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
7682 calls @code{gdb_test} multiple times.
7684 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
7685 necessary. Even if @value{GDBN} has several valid responses to
7686 a command, you can use @code{gdb_test_multiple}. Like @code{gdb_test},
7687 @code{gdb_test_multiple} recognizes internal errors and unexpected
7690 Do not write tests which expect a literal tab character from @value{GDBN}.
7691 On some operating systems (e.g.@: OpenBSD) the TTY layer expands tabs to
7692 spaces, so by the time @value{GDBN}'s output reaches expect the tab is gone.
7694 The source language programs do @emph{not} need to be in a consistent
7695 style. Since @value{GDBN} is used to debug programs written in many different
7696 styles, it's worth having a mix of styles in the testsuite; for
7697 instance, some @value{GDBN} bugs involving the display of source lines would
7698 never manifest themselves if the programs used GNU coding style
7705 Check the @file{README} file, it often has useful information that does not
7706 appear anywhere else in the directory.
7709 * Getting Started:: Getting started working on @value{GDBN}
7710 * Debugging GDB:: Debugging @value{GDBN} with itself
7713 @node Getting Started,,, Hints
7715 @section Getting Started
7717 @value{GDBN} is a large and complicated program, and if you first starting to
7718 work on it, it can be hard to know where to start. Fortunately, if you
7719 know how to go about it, there are ways to figure out what is going on.
7721 This manual, the @value{GDBN} Internals manual, has information which applies
7722 generally to many parts of @value{GDBN}.
7724 Information about particular functions or data structures are located in
7725 comments with those functions or data structures. If you run across a
7726 function or a global variable which does not have a comment correctly
7727 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
7728 free to submit a bug report, with a suggested comment if you can figure
7729 out what the comment should say. If you find a comment which is
7730 actually wrong, be especially sure to report that.
7732 Comments explaining the function of macros defined in host, target, or
7733 native dependent files can be in several places. Sometimes they are
7734 repeated every place the macro is defined. Sometimes they are where the
7735 macro is used. Sometimes there is a header file which supplies a
7736 default definition of the macro, and the comment is there. This manual
7737 also documents all the available macros.
7738 @c (@pxref{Host Conditionals}, @pxref{Target
7739 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
7742 Start with the header files. Once you have some idea of how
7743 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
7744 @file{gdbtypes.h}), you will find it much easier to understand the
7745 code which uses and creates those symbol tables.
7747 You may wish to process the information you are getting somehow, to
7748 enhance your understanding of it. Summarize it, translate it to another
7749 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
7750 the code to predict what a test case would do and write the test case
7751 and verify your prediction, etc. If you are reading code and your eyes
7752 are starting to glaze over, this is a sign you need to use a more active
7755 Once you have a part of @value{GDBN} to start with, you can find more
7756 specifically the part you are looking for by stepping through each
7757 function with the @code{next} command. Do not use @code{step} or you
7758 will quickly get distracted; when the function you are stepping through
7759 calls another function try only to get a big-picture understanding
7760 (perhaps using the comment at the beginning of the function being
7761 called) of what it does. This way you can identify which of the
7762 functions being called by the function you are stepping through is the
7763 one which you are interested in. You may need to examine the data
7764 structures generated at each stage, with reference to the comments in
7765 the header files explaining what the data structures are supposed to
7768 Of course, this same technique can be used if you are just reading the
7769 code, rather than actually stepping through it. The same general
7770 principle applies---when the code you are looking at calls something
7771 else, just try to understand generally what the code being called does,
7772 rather than worrying about all its details.
7774 @cindex command implementation
7775 A good place to start when tracking down some particular area is with
7776 a command which invokes that feature. Suppose you want to know how
7777 single-stepping works. As a @value{GDBN} user, you know that the
7778 @code{step} command invokes single-stepping. The command is invoked
7779 via command tables (see @file{command.h}); by convention the function
7780 which actually performs the command is formed by taking the name of
7781 the command and adding @samp{_command}, or in the case of an
7782 @code{info} subcommand, @samp{_info}. For example, the @code{step}
7783 command invokes the @code{step_command} function and the @code{info
7784 display} command invokes @code{display_info}. When this convention is
7785 not followed, you might have to use @code{grep} or @kbd{M-x
7786 tags-search} in emacs, or run @value{GDBN} on itself and set a
7787 breakpoint in @code{execute_command}.
7789 @cindex @code{bug-gdb} mailing list
7790 If all of the above fail, it may be appropriate to ask for information
7791 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
7792 wondering if anyone could give me some tips about understanding
7793 @value{GDBN}''---if we had some magic secret we would put it in this manual.
7794 Suggestions for improving the manual are always welcome, of course.
7796 @node Debugging GDB,,,Hints
7798 @section Debugging @value{GDBN} with itself
7799 @cindex debugging @value{GDBN}
7801 If @value{GDBN} is limping on your machine, this is the preferred way to get it
7802 fully functional. Be warned that in some ancient Unix systems, like
7803 Ultrix 4.2, a program can't be running in one process while it is being
7804 debugged in another. Rather than typing the command @kbd{@w{./gdb
7805 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
7806 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
7808 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
7809 @file{.gdbinit} file that sets up some simple things to make debugging
7810 gdb easier. The @code{info} command, when executed without a subcommand
7811 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
7812 gdb. See @file{.gdbinit} for details.
7814 If you use emacs, you will probably want to do a @code{make TAGS} after
7815 you configure your distribution; this will put the machine dependent
7816 routines for your local machine where they will be accessed first by
7819 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
7820 have run @code{fixincludes} if you are compiling with gcc.
7822 @section Submitting Patches
7824 @cindex submitting patches
7825 Thanks for thinking of offering your changes back to the community of
7826 @value{GDBN} users. In general we like to get well designed enhancements.
7827 Thanks also for checking in advance about the best way to transfer the
7830 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
7831 This manual summarizes what we believe to be clean design for @value{GDBN}.
7833 If the maintainers don't have time to put the patch in when it arrives,
7834 or if there is any question about a patch, it goes into a large queue
7835 with everyone else's patches and bug reports.
7837 @cindex legal papers for code contributions
7838 The legal issue is that to incorporate substantial changes requires a
7839 copyright assignment from you and/or your employer, granting ownership
7840 of the changes to the Free Software Foundation. You can get the
7841 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
7842 and asking for it. We recommend that people write in "All programs
7843 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
7844 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
7846 contributed with only one piece of legalese pushed through the
7847 bureaucracy and filed with the FSF. We can't start merging changes until
7848 this paperwork is received by the FSF (their rules, which we follow
7849 since we maintain it for them).
7851 Technically, the easiest way to receive changes is to receive each
7852 feature as a small context diff or unidiff, suitable for @code{patch}.
7853 Each message sent to me should include the changes to C code and
7854 header files for a single feature, plus @file{ChangeLog} entries for
7855 each directory where files were modified, and diffs for any changes
7856 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
7857 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
7858 single feature, they can be split down into multiple messages.
7860 In this way, if we read and like the feature, we can add it to the
7861 sources with a single patch command, do some testing, and check it in.
7862 If you leave out the @file{ChangeLog}, we have to write one. If you leave
7863 out the doc, we have to puzzle out what needs documenting. Etc., etc.
7865 The reason to send each change in a separate message is that we will not
7866 install some of the changes. They'll be returned to you with questions
7867 or comments. If we're doing our job correctly, the message back to you
7868 will say what you have to fix in order to make the change acceptable.
7869 The reason to have separate messages for separate features is so that
7870 the acceptable changes can be installed while one or more changes are
7871 being reworked. If multiple features are sent in a single message, we
7872 tend to not put in the effort to sort out the acceptable changes from
7873 the unacceptable, so none of the features get installed until all are
7876 If this sounds painful or authoritarian, well, it is. But we get a lot
7877 of bug reports and a lot of patches, and many of them don't get
7878 installed because we don't have the time to finish the job that the bug
7879 reporter or the contributor could have done. Patches that arrive
7880 complete, working, and well designed, tend to get installed on the day
7881 they arrive. The others go into a queue and get installed as time
7882 permits, which, since the maintainers have many demands to meet, may not
7883 be for quite some time.
7885 Please send patches directly to
7886 @email{gdb-patches@@sourceware.org, the @value{GDBN} maintainers}.
7888 @section Build Script
7890 @cindex build script
7892 The script @file{gdb_buildall.sh} builds @value{GDBN} with flag
7893 @option{--enable-targets=all} set. This builds @value{GDBN} with all supported
7894 targets activated. This helps testing @value{GDBN} when doing changes that
7895 affect more than one architecture and is much faster than using
7896 @file{gdb_mbuild.sh}.
7898 After building @value{GDBN} the script checks which architectures are
7899 supported and then switches the current architecture to each of those to get
7900 information about the architecture. The test results are stored in log files
7901 in the directory the script was called from.
7903 @include observer.texi