1 \input texinfo @c -*- texinfo -*-
2 @setfilename gdbint.info
4 @settitle @value{GDBN} Internals
6 @dircategory Software development
8 * Gdb-Internals: (gdbint). The GNU debugger's internals.
12 Copyright @copyright{} 1990-2013 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 1.3 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}.
33 @title @value{GDBN} Internals
34 @subtitle A guide to the internals of the GNU debugger
36 @author Cygnus Solutions
37 @author Second Edition:
39 @author Cygnus Solutions
42 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
43 \xdef\manvers{\$Revision$} % For use in headers, footers too
45 \hfill Cygnus Solutions\par
47 \hfill \TeX{}info \texinfoversion\par
51 @vskip 0pt plus 1filll
58 @c Perhaps this should be the title of the document (but only for info,
59 @c not for TeX). Existing GNU manuals seem inconsistent on this point.
60 @top Scope of this Document
62 This document documents the internals of the GNU debugger, @value{GDBN}. It
63 includes description of @value{GDBN}'s key algorithms and operations, as well
64 as the mechanisms that adapt @value{GDBN} to specific hosts and targets.
77 * Target Architecture Definition::
78 * Target Descriptions::
79 * Target Vector Definition::
85 * Versions and Branches::
86 * Start of New Year Procedure::
91 * GDB Observers:: @value{GDBN} Currently available observers
92 * GNU Free Documentation License:: The license for this documentation
94 * Function and Variable Index::
106 @section Requirements
107 @cindex requirements for @value{GDBN}
109 Before diving into the internals, you should understand the formal
110 requirements and other expectations for @value{GDBN}. Although some
111 of these may seem obvious, there have been proposals for @value{GDBN}
112 that have run counter to these requirements.
114 First of all, @value{GDBN} is a debugger. It's not designed to be a
115 front panel for embedded systems. It's not a text editor. It's not a
116 shell. It's not a programming environment.
118 @value{GDBN} is an interactive tool. Although a batch mode is
119 available, @value{GDBN}'s primary role is to interact with a human
122 @value{GDBN} should be responsive to the user. A programmer hot on
123 the trail of a nasty bug, and operating under a looming deadline, is
124 going to be very impatient of everything, including the response time
125 to debugger commands.
127 @value{GDBN} should be relatively permissive, such as for expressions.
128 While the compiler should be picky (or have the option to be made
129 picky), since source code lives for a long time usually, the
130 programmer doing debugging shouldn't be spending time figuring out to
131 mollify the debugger.
133 @value{GDBN} will be called upon to deal with really large programs.
134 Executable sizes of 50 to 100 megabytes occur regularly, and we've
135 heard reports of programs approaching 1 gigabyte in size.
137 @value{GDBN} should be able to run everywhere. No other debugger is
138 available for even half as many configurations as @value{GDBN}
142 @section Contributors
144 The first edition of this document was written by John Gilmore of
145 Cygnus Solutions. The current second edition was written by Stan Shebs
146 of Cygnus Solutions, who continues to update the manual.
148 Over the years, many others have made additions and changes to this
149 document. This section attempts to record the significant contributors
150 to that effort. One of the virtues of free software is that everyone
151 is free to contribute to it; with regret, we cannot actually
152 acknowledge everyone here.
155 @emph{Plea:} This section has only been added relatively recently (four
156 years after publication of the second edition). Additions to this
157 section are particularly welcome. If you or your friends (or enemies,
158 to be evenhanded) have been unfairly omitted from this list, we would
159 like to add your names!
162 A document such as this relies on being kept up to date by numerous
163 small updates by contributing engineers as they make changes to the
164 code base. The file @file{ChangeLog} in the @value{GDBN} distribution
165 approximates a blow-by-blow account. The most prolific contributors to
166 this important, but low profile task are Andrew Cagney (responsible
167 for over half the entries), Daniel Jacobowitz, Mark Kettenis, Jim
168 Blandy and Eli Zaretskii.
170 Eli Zaretskii and Daniel Jacobowitz wrote the sections documenting
173 Jeremy Bennett updated the sections on initializing a new architecture
174 and register representation, and added the section on Frame Interpretation.
177 @node Overall Structure
179 @chapter Overall Structure
181 @value{GDBN} consists of three major subsystems: user interface,
182 symbol handling (the @dfn{symbol side}), and target system handling (the
185 The user interface consists of several actual interfaces, plus
188 The symbol side consists of object file readers, debugging info
189 interpreters, symbol table management, source language expression
190 parsing, type and value printing.
192 The target side consists of execution control, stack frame analysis, and
193 physical target manipulation.
195 The target side/symbol side division is not formal, and there are a
196 number of exceptions. For instance, core file support involves symbolic
197 elements (the basic core file reader is in BFD) and target elements (it
198 supplies the contents of memory and the values of registers). Instead,
199 this division is useful for understanding how the minor subsystems
202 @section The Symbol Side
204 The symbolic side of @value{GDBN} can be thought of as ``everything
205 you can do in @value{GDBN} without having a live program running''.
206 For instance, you can look at the types of variables, and evaluate
207 many kinds of expressions.
209 @section The Target Side
211 The target side of @value{GDBN} is the ``bits and bytes manipulator''.
212 Although it may make reference to symbolic info here and there, most
213 of the target side will run with only a stripped executable
214 available---or even no executable at all, in remote debugging cases.
216 Operations such as disassembly, stack frame crawls, and register
217 display, are able to work with no symbolic info at all. In some cases,
218 such as disassembly, @value{GDBN} will use symbolic info to present addresses
219 relative to symbols rather than as raw numbers, but it will work either
222 @section Configurations
226 @dfn{Host} refers to attributes of the system where @value{GDBN} runs.
227 @dfn{Target} refers to the system where the program being debugged
228 executes. In most cases they are the same machine, in which case a
229 third type of @dfn{Native} attributes come into play.
231 Defines and include files needed to build on the host are host
232 support. Examples are tty support, system defined types, host byte
233 order, host float format. These are all calculated by @code{autoconf}
234 when the debugger is built.
236 Defines and information needed to handle the target format are target
237 dependent. Examples are the stack frame format, instruction set,
238 breakpoint instruction, registers, and how to set up and tear down the stack
241 Information that is only needed when the host and target are the same,
242 is native dependent. One example is Unix child process support; if the
243 host and target are not the same, calling @code{fork} to start the target
244 process is a bad idea. The various macros needed for finding the
245 registers in the @code{upage}, running @code{ptrace}, and such are all
246 in the native-dependent files.
248 Another example of native-dependent code is support for features that
249 are really part of the target environment, but which require
250 @code{#include} files that are only available on the host system. Core
251 file handling and @code{setjmp} handling are two common cases.
253 When you want to make @value{GDBN} work as the traditional native debugger
254 on a system, you will need to supply both target and native information.
256 @section Source Tree Structure
257 @cindex @value{GDBN} source tree structure
259 The @value{GDBN} source directory has a mostly flat structure---there
260 are only a few subdirectories. A file's name usually gives a hint as
261 to what it does; for example, @file{stabsread.c} reads stabs,
262 @file{dwarf2read.c} reads @sc{DWARF 2}, etc.
264 Files that are related to some common task have names that share
265 common substrings. For example, @file{*-thread.c} files deal with
266 debugging threads on various platforms; @file{*read.c} files deal with
267 reading various kinds of symbol and object files; @file{inf*.c} files
268 deal with direct control of the @dfn{inferior program} (@value{GDBN}
269 parlance for the program being debugged).
271 There are several dozens of files in the @file{*-tdep.c} family.
272 @samp{tdep} stands for @dfn{target-dependent code}---each of these
273 files implements debug support for a specific target architecture
274 (sparc, mips, etc). Usually, only one of these will be used in a
275 specific @value{GDBN} configuration (sometimes two, closely related).
277 Similarly, there are many @file{*-nat.c} files, each one for native
278 debugging on a specific system (e.g., @file{sparc-linux-nat.c} is for
279 native debugging of Sparc machines running the Linux kernel).
281 The few subdirectories of the source tree are:
285 Code that implements @dfn{CLI}, the @value{GDBN} Command-Line
286 Interpreter. @xref{User Interface, Command Interpreter}.
289 Code for the @value{GDBN} remote server.
292 Code for Insight, the @value{GDBN} TK-based GUI front-end.
295 The @dfn{GDB/MI}, the @value{GDBN} Machine Interface interpreter.
298 Target signal translation code.
301 Code for @dfn{TUI}, the @value{GDBN} Text-mode full-screen User
302 Interface. @xref{User Interface, TUI}.
310 @value{GDBN} uses a number of debugging-specific algorithms. They are
311 often not very complicated, but get lost in the thicket of special
312 cases and real-world issues. This chapter describes the basic
313 algorithms and mentions some of the specific target definitions that
316 @section Prologue Analysis
318 @cindex prologue analysis
319 @cindex call frame information
320 @cindex CFI (call frame information)
321 To produce a backtrace and allow the user to manipulate older frames'
322 variables and arguments, @value{GDBN} needs to find the base addresses
323 of older frames, and discover where those frames' registers have been
324 saved. Since a frame's ``callee-saves'' registers get saved by
325 younger frames if and when they're reused, a frame's registers may be
326 scattered unpredictably across younger frames. This means that
327 changing the value of a register-allocated variable in an older frame
328 may actually entail writing to a save slot in some younger frame.
330 Modern versions of GCC emit Dwarf call frame information (``CFI''),
331 which describes how to find frame base addresses and saved registers.
332 But CFI is not always available, so as a fallback @value{GDBN} uses a
333 technique called @dfn{prologue analysis} to find frame sizes and saved
334 registers. A prologue analyzer disassembles the function's machine
335 code starting from its entry point, and looks for instructions that
336 allocate frame space, save the stack pointer in a frame pointer
337 register, save registers, and so on. Obviously, this can't be done
338 accurately in general, but it's tractable to do well enough to be very
339 helpful. Prologue analysis predates the GNU toolchain's support for
340 CFI; at one time, prologue analysis was the only mechanism
341 @value{GDBN} used for stack unwinding at all, when the function
342 calling conventions didn't specify a fixed frame layout.
344 In the olden days, function prologues were generated by hand-written,
345 target-specific code in GCC, and treated as opaque and untouchable by
346 optimizers. Looking at this code, it was usually straightforward to
347 write a prologue analyzer for @value{GDBN} that would accurately
348 understand all the prologues GCC would generate. However, over time
349 GCC became more aggressive about instruction scheduling, and began to
350 understand more about the semantics of the prologue instructions
351 themselves; in response, @value{GDBN}'s analyzers became more complex
352 and fragile. Keeping the prologue analyzers working as GCC (and the
353 instruction sets themselves) evolved became a substantial task.
355 @cindex @file{prologue-value.c}
356 @cindex abstract interpretation of function prologues
357 @cindex pseudo-evaluation of function prologues
358 To try to address this problem, the code in @file{prologue-value.h}
359 and @file{prologue-value.c} provides a general framework for writing
360 prologue analyzers that are simpler and more robust than ad-hoc
361 analyzers. When we analyze a prologue using the prologue-value
362 framework, we're really doing ``abstract interpretation'' or
363 ``pseudo-evaluation'': running the function's code in simulation, but
364 using conservative approximations of the values registers and memory
365 would hold when the code actually runs. For example, if our function
366 starts with the instruction:
369 addi r1, 42 # add 42 to r1
372 we don't know exactly what value will be in @code{r1} after executing
373 this instruction, but we do know it'll be 42 greater than its original
376 If we then see an instruction like:
379 addi r1, 22 # add 22 to r1
382 we still don't know what @code{r1's} value is, but again, we can say
383 it is now 64 greater than its original value.
385 If the next instruction were:
388 mov r2, r1 # set r2 to r1's value
391 then we can say that @code{r2's} value is now the original value of
394 It's common for prologues to save registers on the stack, so we'll
395 need to track the values of stack frame slots, as well as the
396 registers. So after an instruction like this:
402 then we'd know that the stack slot four bytes above the frame pointer
403 holds the original value of @code{r1} plus 64.
407 Of course, this can only go so far before it gets unreasonable. If we
408 wanted to be able to say anything about the value of @code{r1} after
412 xor r1, r3 # exclusive-or r1 and r3, place result in r1
415 then things would get pretty complex. But remember, we're just doing
416 a conservative approximation; if exclusive-or instructions aren't
417 relevant to prologues, we can just say @code{r1}'s value is now
418 ``unknown''. We can ignore things that are too complex, if that loss of
419 information is acceptable for our application.
421 So when we say ``conservative approximation'' here, what we mean is an
422 approximation that is either accurate, or marked ``unknown'', but
425 Using this framework, a prologue analyzer is simply an interpreter for
426 machine code, but one that uses conservative approximations for the
427 contents of registers and memory instead of actual values. Starting
428 from the function's entry point, you simulate instructions up to the
429 current PC, or an instruction that you don't know how to simulate.
430 Now you can examine the state of the registers and stack slots you've
436 To see how large your stack frame is, just check the value of the
437 stack pointer register; if it's the original value of the SP
438 minus a constant, then that constant is the stack frame's size.
439 If the SP's value has been marked as ``unknown'', then that means
440 the prologue has done something too complex for us to track, and
441 we don't know the frame size.
444 To see where we've saved the previous frame's registers, we just
445 search the values we've tracked --- stack slots, usually, but
446 registers, too, if you want --- for something equal to the register's
447 original value. If the calling conventions suggest a standard place
448 to save a given register, then we can check there first, but really,
449 anything that will get us back the original value will probably work.
452 This does take some work. But prologue analyzers aren't
453 quick-and-simple pattern patching to recognize a few fixed prologue
454 forms any more; they're big, hairy functions. Along with inferior
455 function calls, prologue analysis accounts for a substantial portion
456 of the time needed to stabilize a @value{GDBN} port. So it's
457 worthwhile to look for an approach that will be easier to understand
458 and maintain. In the approach described above:
463 It's easier to see that the analyzer is correct: you just see
464 whether the analyzer properly (albeit conservatively) simulates
465 the effect of each instruction.
468 It's easier to extend the analyzer: you can add support for new
469 instructions, and know that you haven't broken anything that
470 wasn't already broken before.
473 It's orthogonal: to gather new information, you don't need to
474 complicate the code for each instruction. As long as your domain
475 of conservative values is already detailed enough to tell you
476 what you need, then all the existing instruction simulations are
477 already gathering the right data for you.
481 The file @file{prologue-value.h} contains detailed comments explaining
482 the framework and how to use it.
485 @section Breakpoint Handling
488 In general, a breakpoint is a user-designated location in the program
489 where the user wants to regain control if program execution ever reaches
492 There are two main ways to implement breakpoints; either as ``hardware''
493 breakpoints or as ``software'' breakpoints.
495 @cindex hardware breakpoints
496 @cindex program counter
497 Hardware breakpoints are sometimes available as a builtin debugging
498 features with some chips. Typically these work by having dedicated
499 register into which the breakpoint address may be stored. If the PC
500 (shorthand for @dfn{program counter})
501 ever matches a value in a breakpoint registers, the CPU raises an
502 exception and reports it to @value{GDBN}.
504 Another possibility is when an emulator is in use; many emulators
505 include circuitry that watches the address lines coming out from the
506 processor, and force it to stop if the address matches a breakpoint's
509 A third possibility is that the target already has the ability to do
510 breakpoints somehow; for instance, a ROM monitor may do its own
511 software breakpoints. So although these are not literally ``hardware
512 breakpoints'', from @value{GDBN}'s point of view they work the same;
513 @value{GDBN} need not do anything more than set the breakpoint and wait
514 for something to happen.
516 Since they depend on hardware resources, hardware breakpoints may be
517 limited in number; when the user asks for more, @value{GDBN} will
518 start trying to set software breakpoints. (On some architectures,
519 notably the 32-bit x86 platforms, @value{GDBN} cannot always know
520 whether there's enough hardware resources to insert all the hardware
521 breakpoints and watchpoints. On those platforms, @value{GDBN} prints
522 an error message only when the program being debugged is continued.)
524 @cindex software breakpoints
525 Software breakpoints require @value{GDBN} to do somewhat more work.
526 The basic theory is that @value{GDBN} will replace a program
527 instruction with a trap, illegal divide, or some other instruction
528 that will cause an exception, and then when it's encountered,
529 @value{GDBN} will take the exception and stop the program. When the
530 user says to continue, @value{GDBN} will restore the original
531 instruction, single-step, re-insert the trap, and continue on.
533 Since it literally overwrites the program being tested, the program area
534 must be writable, so this technique won't work on programs in ROM. It
535 can also distort the behavior of programs that examine themselves,
536 although such a situation would be highly unusual.
538 Also, the software breakpoint instruction should be the smallest size of
539 instruction, so it doesn't overwrite an instruction that might be a jump
540 target, and cause disaster when the program jumps into the middle of the
541 breakpoint instruction. (Strictly speaking, the breakpoint must be no
542 larger than the smallest interval between instructions that may be jump
543 targets; perhaps there is an architecture where only even-numbered
544 instructions may jumped to.) Note that it's possible for an instruction
545 set not to have any instructions usable for a software breakpoint,
546 although in practice only the ARC has failed to define such an
549 Basic breakpoint object handling is in @file{breakpoint.c}. However,
550 much of the interesting breakpoint action is in @file{infrun.c}.
553 @cindex insert or remove software breakpoint
554 @findex target_remove_breakpoint
555 @findex target_insert_breakpoint
556 @item target_remove_breakpoint (@var{bp_tgt})
557 @itemx target_insert_breakpoint (@var{bp_tgt})
558 Insert or remove a software breakpoint at address
559 @code{@var{bp_tgt}->placed_address}. Returns zero for success,
560 non-zero for failure. On input, @var{bp_tgt} contains the address of the
561 breakpoint, and is otherwise initialized to zero. The fields of the
562 @code{struct bp_target_info} pointed to by @var{bp_tgt} are updated
563 to contain other information about the breakpoint on output. The field
564 @code{placed_address} may be updated if the breakpoint was placed at a
565 related address; the field @code{shadow_contents} contains the real
566 contents of the bytes where the breakpoint has been inserted,
567 if reading memory would return the breakpoint instead of the
568 underlying memory; the field @code{shadow_len} is the length of
569 memory cached in @code{shadow_contents}, if any; and the field
570 @code{placed_size} is optionally set and used by the target, if
571 it could differ from @code{shadow_len}.
573 For example, the remote target @samp{Z0} packet does not require
574 shadowing memory, so @code{shadow_len} is left at zero. However,
575 the length reported by @code{gdbarch_breakpoint_from_pc} is cached in
576 @code{placed_size}, so that a matching @samp{z0} packet can be
577 used to remove the breakpoint.
579 @cindex insert or remove hardware breakpoint
580 @findex target_remove_hw_breakpoint
581 @findex target_insert_hw_breakpoint
582 @item target_remove_hw_breakpoint (@var{bp_tgt})
583 @itemx target_insert_hw_breakpoint (@var{bp_tgt})
584 Insert or remove a hardware-assisted breakpoint at address
585 @code{@var{bp_tgt}->placed_address}. Returns zero for success,
586 non-zero for failure. See @code{target_insert_breakpoint} for
587 a description of the @code{struct bp_target_info} pointed to by
588 @var{bp_tgt}; the @code{shadow_contents} and
589 @code{shadow_len} members are not used for hardware breakpoints,
590 but @code{placed_size} may be.
593 @section Single Stepping
595 @section Signal Handling
597 @section Thread Handling
599 @section Inferior Function Calls
601 @section Longjmp Support
603 @cindex @code{longjmp} debugging
604 @value{GDBN} has support for figuring out that the target is doing a
605 @code{longjmp} and for stopping at the target of the jump, if we are
606 stepping. This is done with a few specialized internal breakpoints,
607 which are visible in the output of the @samp{maint info breakpoint}
610 @findex gdbarch_get_longjmp_target
611 To make this work, you need to define a function called
612 @code{gdbarch_get_longjmp_target}, which will examine the
613 @code{jmp_buf} structure and extract the @code{longjmp} target address.
614 Since @code{jmp_buf} is target specific and typically defined in a
615 target header not available to @value{GDBN}, you will need to
616 determine the offset of the PC manually and return that; many targets
617 define a @code{jb_pc_offset} field in the tdep structure to save the
618 value once calculated.
623 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
624 breakpoints}) which break when data is accessed rather than when some
625 instruction is executed. When you have data which changes without
626 your knowing what code does that, watchpoints are the silver bullet to
627 hunt down and kill such bugs.
629 @cindex hardware watchpoints
630 @cindex software watchpoints
631 Watchpoints can be either hardware-assisted or not; the latter type is
632 known as ``software watchpoints.'' @value{GDBN} always uses
633 hardware-assisted watchpoints if they are available, and falls back on
634 software watchpoints otherwise. Typical situations where @value{GDBN}
635 will use software watchpoints are:
639 The watched memory region is too large for the underlying hardware
640 watchpoint support. For example, each x86 debug register can watch up
641 to 4 bytes of memory, so trying to watch data structures whose size is
642 more than 16 bytes will cause @value{GDBN} to use software
646 The value of the expression to be watched depends on data held in
647 registers (as opposed to memory).
650 Too many different watchpoints requested. (On some architectures,
651 this situation is impossible to detect until the debugged program is
652 resumed.) Note that x86 debug registers are used both for hardware
653 breakpoints and for watchpoints, so setting too many hardware
654 breakpoints might cause watchpoint insertion to fail.
657 No hardware-assisted watchpoints provided by the target
661 Software watchpoints are very slow, since @value{GDBN} needs to
662 single-step the program being debugged and test the value of the
663 watched expression(s) after each instruction. The rest of this
664 section is mostly irrelevant for software watchpoints.
666 When the inferior stops, @value{GDBN} tries to establish, among other
667 possible reasons, whether it stopped due to a watchpoint being hit.
668 It first uses @code{STOPPED_BY_WATCHPOINT} to see if any watchpoint
669 was hit. If not, all watchpoint checking is skipped.
671 Then @value{GDBN} calls @code{target_stopped_data_address} exactly
672 once. This method returns the address of the watchpoint which
673 triggered, if the target can determine it. If the triggered address
674 is available, @value{GDBN} compares the address returned by this
675 method with each watched memory address in each active watchpoint.
676 For data-read and data-access watchpoints, @value{GDBN} announces
677 every watchpoint that watches the triggered address as being hit.
678 For this reason, data-read and data-access watchpoints
679 @emph{require} that the triggered address be available; if not, read
680 and access watchpoints will never be considered hit. For data-write
681 watchpoints, if the triggered address is available, @value{GDBN}
682 considers only those watchpoints which match that address;
683 otherwise, @value{GDBN} considers all data-write watchpoints. For
684 each data-write watchpoint that @value{GDBN} considers, it evaluates
685 the expression whose value is being watched, and tests whether the
686 watched value has changed. Watchpoints whose watched values have
687 changed are announced as hit.
689 @c FIXME move these to the main lists of target/native defns
691 @value{GDBN} uses several macros and primitives to support hardware
695 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
696 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
697 Return the number of hardware watchpoints of type @var{type} that are
698 possible to be set. The value is positive if @var{count} watchpoints
699 of this type can be set, zero if setting watchpoints of this type is
700 not supported, and negative if @var{count} is more than the maximum
701 number of watchpoints of type @var{type} that can be set. @var{other}
702 is non-zero if other types of watchpoints are currently enabled (there
703 are architectures which cannot set watchpoints of different types at
706 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
707 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
708 Return non-zero if hardware watchpoints can be used to watch a region
709 whose address is @var{addr} and whose length in bytes is @var{len}.
711 @cindex insert or remove hardware watchpoint
712 @findex target_insert_watchpoint
713 @findex target_remove_watchpoint
714 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
715 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
716 Insert or remove a hardware watchpoint starting at @var{addr}, for
717 @var{len} bytes. @var{type} is the watchpoint type, one of the
718 possible values of the enumerated data type @code{target_hw_bp_type},
719 defined by @file{breakpoint.h} as follows:
722 enum target_hw_bp_type
724 hw_write = 0, /* Common (write) HW watchpoint */
725 hw_read = 1, /* Read HW watchpoint */
726 hw_access = 2, /* Access (read or write) HW watchpoint */
727 hw_execute = 3 /* Execute HW breakpoint */
732 These two macros should return 0 for success, non-zero for failure.
734 @findex target_stopped_data_address
735 @item target_stopped_data_address (@var{addr_p})
736 If the inferior has some watchpoint that triggered, place the address
737 associated with the watchpoint at the location pointed to by
738 @var{addr_p} and return non-zero. Otherwise, return zero. This
739 is required for data-read and data-access watchpoints. It is
740 not required for data-write watchpoints, but @value{GDBN} uses
741 it to improve handling of those also.
743 @value{GDBN} will only call this method once per watchpoint stop,
744 immediately after calling @code{STOPPED_BY_WATCHPOINT}. If the
745 target's watchpoint indication is sticky, i.e., stays set after
746 resuming, this method should clear it. For instance, the x86 debug
747 control register has sticky triggered flags.
749 @findex target_watchpoint_addr_within_range
750 @item target_watchpoint_addr_within_range (@var{target}, @var{addr}, @var{start}, @var{length})
751 Check whether @var{addr} (as returned by @code{target_stopped_data_address})
752 lies within the hardware-defined watchpoint region described by
753 @var{start} and @var{length}. This only needs to be provided if the
754 granularity of a watchpoint is greater than one byte, i.e., if the
755 watchpoint can also trigger on nearby addresses outside of the watched
758 @findex HAVE_STEPPABLE_WATCHPOINT
759 @item HAVE_STEPPABLE_WATCHPOINT
760 If defined to a non-zero value, it is not necessary to disable a
761 watchpoint to step over it. Like @code{gdbarch_have_nonsteppable_watchpoint},
762 this is usually set when watchpoints trigger at the instruction
763 which will perform an interesting read or write. It should be
764 set if there is a temporary disable bit which allows the processor
765 to step over the interesting instruction without raising the
766 watchpoint exception again.
768 @findex gdbarch_have_nonsteppable_watchpoint
769 @item int gdbarch_have_nonsteppable_watchpoint (@var{gdbarch})
770 If it returns a non-zero value, @value{GDBN} should disable a
771 watchpoint to step the inferior over it. This is usually set when
772 watchpoints trigger at the instruction which will perform an
773 interesting read or write.
775 @findex HAVE_CONTINUABLE_WATCHPOINT
776 @item HAVE_CONTINUABLE_WATCHPOINT
777 If defined to a non-zero value, it is possible to continue the
778 inferior after a watchpoint has been hit. This is usually set
779 when watchpoints trigger at the instruction following an interesting
782 @findex STOPPED_BY_WATCHPOINT
783 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
784 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
785 the type @code{struct target_waitstatus}, defined by @file{target.h}.
786 Normally, this macro is defined to invoke the function pointed to by
787 the @code{to_stopped_by_watchpoint} member of the structure (of the
788 type @code{target_ops}, defined on @file{target.h}) that describes the
789 target-specific operations; @code{to_stopped_by_watchpoint} ignores
790 the @var{wait_status} argument.
792 @value{GDBN} does not require the non-zero value returned by
793 @code{STOPPED_BY_WATCHPOINT} to be 100% correct, so if a target cannot
794 determine for sure whether the inferior stopped due to a watchpoint,
795 it could return non-zero ``just in case''.
798 @subsection Watchpoints and Threads
799 @cindex watchpoints, with threads
801 @value{GDBN} only supports process-wide watchpoints, which trigger
802 in all threads. @value{GDBN} uses the thread ID to make watchpoints
803 act as if they were thread-specific, but it cannot set hardware
804 watchpoints that only trigger in a specific thread. Therefore, even
805 if the target supports threads, per-thread debug registers, and
806 watchpoints which only affect a single thread, it should set the
807 per-thread debug registers for all threads to the same value. On
808 @sc{gnu}/Linux native targets, this is accomplished by using
809 @code{ALL_LWPS} in @code{target_insert_watchpoint} and
810 @code{target_remove_watchpoint} and by using
811 @code{linux_set_new_thread} to register a handler for newly created
814 @value{GDBN}'s @sc{gnu}/Linux support only reports a single event
815 at a time, although multiple events can trigger simultaneously for
816 multi-threaded programs. When multiple events occur, @file{linux-nat.c}
817 queues subsequent events and returns them the next time the program
818 is resumed. This means that @code{STOPPED_BY_WATCHPOINT} and
819 @code{target_stopped_data_address} only need to consult the current
820 thread's state---the thread indicated by @code{inferior_ptid}. If
821 two threads have hit watchpoints simultaneously, those routines
822 will be called a second time for the second thread.
824 @subsection x86 Watchpoints
825 @cindex x86 debug registers
826 @cindex watchpoints, on x86
828 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
829 registers designed to facilitate debugging. @value{GDBN} provides a
830 generic library of functions that x86-based ports can use to implement
831 support for watchpoints and hardware-assisted breakpoints. This
832 subsection documents the x86 watchpoint facilities in @value{GDBN}.
834 (At present, the library functions read and write debug registers directly, and are
835 thus only available for native configurations.)
837 To use the generic x86 watchpoint support, a port should do the
841 @findex I386_USE_GENERIC_WATCHPOINTS
843 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
844 target-dependent headers.
847 Include the @file{config/i386/nm-i386.h} header file @emph{after}
848 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
851 Add @file{i386-nat.o} to the value of the Make variable
852 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}).
855 Provide implementations for the @code{I386_DR_LOW_*} macros described
856 below. Typically, each macro should call a target-specific function
857 which does the real work.
860 The x86 watchpoint support works by maintaining mirror images of the
861 debug registers. Values are copied between the mirror images and the
862 real debug registers via a set of macros which each target needs to
866 @findex I386_DR_LOW_SET_CONTROL
867 @item I386_DR_LOW_SET_CONTROL (@var{val})
868 Set the Debug Control (DR7) register to the value @var{val}.
870 @findex I386_DR_LOW_SET_ADDR
871 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
872 Put the address @var{addr} into the debug register number @var{idx}.
874 @findex I386_DR_LOW_RESET_ADDR
875 @item I386_DR_LOW_RESET_ADDR (@var{idx})
876 Reset (i.e.@: zero out) the address stored in the debug register
879 @findex I386_DR_LOW_GET_STATUS
880 @item I386_DR_LOW_GET_STATUS
881 Return the value of the Debug Status (DR6) register. This value is
882 used immediately after it is returned by
883 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
887 For each one of the 4 debug registers (whose indices are from 0 to 3)
888 that store addresses, a reference count is maintained by @value{GDBN},
889 to allow sharing of debug registers by several watchpoints. This
890 allows users to define several watchpoints that watch the same
891 expression, but with different conditions and/or commands, without
892 wasting debug registers which are in short supply. @value{GDBN}
893 maintains the reference counts internally, targets don't have to do
894 anything to use this feature.
896 The x86 debug registers can each watch a region that is 1, 2, or 4
897 bytes long. The ia32 architecture requires that each watched region
898 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
899 region on 4-byte boundary. However, the x86 watchpoint support in
900 @value{GDBN} can watch unaligned regions and regions larger than 4
901 bytes (up to 16 bytes) by allocating several debug registers to watch
902 a single region. This allocation of several registers per a watched
903 region is also done automatically without target code intervention.
905 The generic x86 watchpoint support provides the following API for the
906 @value{GDBN}'s application code:
909 @findex i386_region_ok_for_watchpoint
910 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
911 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
912 this function. It counts the number of debug registers required to
913 watch a given region, and returns a non-zero value if that number is
914 less than 4, the number of debug registers available to x86
917 @findex i386_stopped_data_address
918 @item i386_stopped_data_address (@var{addr_p})
920 @code{target_stopped_data_address} is set to call this function.
922 function examines the breakpoint condition bits in the DR6 Debug
923 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
924 macro, and returns the address associated with the first bit that is
927 @findex i386_stopped_by_watchpoint
928 @item i386_stopped_by_watchpoint (void)
929 The macro @code{STOPPED_BY_WATCHPOINT}
930 is set to call this function. The
931 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
932 function examines the breakpoint condition bits in the DR6 Debug
933 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
934 macro, and returns true if any bit is set. Otherwise, false is
937 @findex i386_insert_watchpoint
938 @findex i386_remove_watchpoint
939 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
940 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
941 Insert or remove a watchpoint. The macros
942 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
943 are set to call these functions. @code{i386_insert_watchpoint} first
944 looks for a debug register which is already set to watch the same
945 region for the same access types; if found, it just increments the
946 reference count of that debug register, thus implementing debug
947 register sharing between watchpoints. If no such register is found,
948 the function looks for a vacant debug register, sets its mirrored
949 value to @var{addr}, sets the mirrored value of DR7 Debug Control
950 register as appropriate for the @var{len} and @var{type} parameters,
951 and then passes the new values of the debug register and DR7 to the
952 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
953 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
954 required to cover the given region, the above process is repeated for
957 @code{i386_remove_watchpoint} does the opposite: it resets the address
958 in the mirrored value of the debug register and its read/write and
959 length bits in the mirrored value of DR7, then passes these new
960 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
961 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
962 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
963 decrements the reference count, and only calls
964 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
965 the count goes to zero.
967 @findex i386_insert_hw_breakpoint
968 @findex i386_remove_hw_breakpoint
969 @item i386_insert_hw_breakpoint (@var{bp_tgt})
970 @itemx i386_remove_hw_breakpoint (@var{bp_tgt})
971 These functions insert and remove hardware-assisted breakpoints. The
972 macros @code{target_insert_hw_breakpoint} and
973 @code{target_remove_hw_breakpoint} are set to call these functions.
974 The argument is a @code{struct bp_target_info *}, as described in
975 the documentation for @code{target_insert_breakpoint}.
976 These functions work like @code{i386_insert_watchpoint} and
977 @code{i386_remove_watchpoint}, respectively, except that they set up
978 the debug registers to watch instruction execution, and each
979 hardware-assisted breakpoint always requires exactly one debug
982 @findex i386_cleanup_dregs
983 @item i386_cleanup_dregs (void)
984 This function clears all the reference counts, addresses, and control
985 bits in the mirror images of the debug registers. It doesn't affect
986 the actual debug registers in the inferior process.
993 x86 processors support setting watchpoints on I/O reads or writes.
994 However, since no target supports this (as of March 2001), and since
995 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
996 watchpoints, this feature is not yet available to @value{GDBN} running
1000 x86 processors can enable watchpoints locally, for the current task
1001 only, or globally, for all the tasks. For each debug register,
1002 there's a bit in the DR7 Debug Control register that determines
1003 whether the associated address is watched locally or globally. The
1004 current implementation of x86 watchpoint support in @value{GDBN}
1005 always sets watchpoints to be locally enabled, since global
1006 watchpoints might interfere with the underlying OS and are probably
1007 unavailable in many platforms.
1010 @section Checkpoints
1013 In the abstract, a checkpoint is a point in the execution history of
1014 the program, which the user may wish to return to at some later time.
1016 Internally, a checkpoint is a saved copy of the program state, including
1017 whatever information is required in order to restore the program to that
1018 state at a later time. This can be expected to include the state of
1019 registers and memory, and may include external state such as the state
1020 of open files and devices.
1022 There are a number of ways in which checkpoints may be implemented
1023 in gdb, e.g.@: as corefiles, as forked processes, and as some opaque
1024 method implemented on the target side.
1026 A corefile can be used to save an image of target memory and register
1027 state, which can in principle be restored later --- but corefiles do
1028 not typically include information about external entities such as
1029 open files. Currently this method is not implemented in gdb.
1031 A forked process can save the state of user memory and registers,
1032 as well as some subset of external (kernel) state. This method
1033 is used to implement checkpoints on Linux, and in principle might
1034 be used on other systems.
1036 Some targets, e.g.@: simulators, might have their own built-in
1037 method for saving checkpoints, and gdb might be able to take
1038 advantage of that capability without necessarily knowing any
1039 details of how it is done.
1042 @section Observing changes in @value{GDBN} internals
1043 @cindex observer pattern interface
1044 @cindex notifications about changes in internals
1046 In order to function properly, several modules need to be notified when
1047 some changes occur in the @value{GDBN} internals. Traditionally, these
1048 modules have relied on several paradigms, the most common ones being
1049 hooks and gdb-events. Unfortunately, none of these paradigms was
1050 versatile enough to become the standard notification mechanism in
1051 @value{GDBN}. The fact that they only supported one ``client'' was also
1052 a strong limitation.
1054 A new paradigm, based on the Observer pattern of the @cite{Design
1055 Patterns} book, has therefore been implemented. The goal was to provide
1056 a new interface overcoming the issues with the notification mechanisms
1057 previously available. This new interface needed to be strongly typed,
1058 easy to extend, and versatile enough to be used as the standard
1059 interface when adding new notifications.
1061 See @ref{GDB Observers} for a brief description of the observers
1062 currently implemented in GDB. The rationale for the current
1063 implementation is also briefly discussed.
1065 @node User Interface
1067 @chapter User Interface
1069 @value{GDBN} has several user interfaces, of which the traditional
1070 command-line interface is perhaps the most familiar.
1072 @section Command Interpreter
1074 @cindex command interpreter
1076 The command interpreter in @value{GDBN} is fairly simple. It is designed to
1077 allow for the set of commands to be augmented dynamically, and also
1078 has a recursive subcommand capability, where the first argument to
1079 a command may itself direct a lookup on a different command list.
1081 For instance, the @samp{set} command just starts a lookup on the
1082 @code{setlist} command list, while @samp{set thread} recurses
1083 to the @code{set_thread_cmd_list}.
1087 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
1088 the main command list, and should be used for those commands. The usual
1089 place to add commands is in the @code{_initialize_@var{xyz}} routines at
1090 the ends of most source files.
1092 @findex add_setshow_cmd
1093 @findex add_setshow_cmd_full
1094 To add paired @samp{set} and @samp{show} commands, use
1095 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
1096 a slightly simpler interface which is useful when you don't need to
1097 further modify the new command structures, while the latter returns
1098 the new command structures for manipulation.
1100 @cindex deprecating commands
1101 @findex deprecate_cmd
1102 Before removing commands from the command set it is a good idea to
1103 deprecate them for some time. Use @code{deprecate_cmd} on commands or
1104 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
1105 @code{struct cmd_list_element} as it's first argument. You can use the
1106 return value from @code{add_com} or @code{add_cmd} to deprecate the
1107 command immediately after it is created.
1109 The first time a command is used the user will be warned and offered a
1110 replacement (if one exists). Note that the replacement string passed to
1111 @code{deprecate_cmd} should be the full name of the command, i.e., the
1112 entire string the user should type at the command line.
1114 @anchor{UI-Independent Output}
1115 @section UI-Independent Output---the @code{ui_out} Functions
1116 @c This section is based on the documentation written by Fernando
1117 @c Nasser <fnasser@redhat.com>.
1119 @cindex @code{ui_out} functions
1120 The @code{ui_out} functions present an abstraction level for the
1121 @value{GDBN} output code. They hide the specifics of different user
1122 interfaces supported by @value{GDBN}, and thus free the programmer
1123 from the need to write several versions of the same code, one each for
1124 every UI, to produce output.
1126 @subsection Overview and Terminology
1128 In general, execution of each @value{GDBN} command produces some sort
1129 of output, and can even generate an input request.
1131 Output can be generated for the following purposes:
1135 to display a @emph{result} of an operation;
1138 to convey @emph{info} or produce side-effects of a requested
1142 to provide a @emph{notification} of an asynchronous event (including
1143 progress indication of a prolonged asynchronous operation);
1146 to display @emph{error messages} (including warnings);
1149 to show @emph{debug data};
1152 to @emph{query} or prompt a user for input (a special case).
1156 This section mainly concentrates on how to build result output,
1157 although some of it also applies to other kinds of output.
1159 Generation of output that displays the results of an operation
1160 involves one or more of the following:
1164 output of the actual data
1167 formatting the output as appropriate for console output, to make it
1168 easily readable by humans
1171 machine oriented formatting--a more terse formatting to allow for easy
1172 parsing by programs which read @value{GDBN}'s output
1175 annotation, whose purpose is to help legacy GUIs to identify interesting
1179 The @code{ui_out} routines take care of the first three aspects.
1180 Annotations are provided by separate annotation routines. Note that use
1181 of annotations for an interface between a GUI and @value{GDBN} is
1184 Output can be in the form of a single item, which we call a @dfn{field};
1185 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
1186 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
1187 header and a body. In a BNF-like form:
1190 @item <table> @expansion{}
1191 @code{<header> <body>}
1192 @item <header> @expansion{}
1193 @code{@{ <column> @}}
1194 @item <column> @expansion{}
1195 @code{<width> <alignment> <title>}
1196 @item <body> @expansion{}
1201 @subsection General Conventions
1203 Most @code{ui_out} routines are of type @code{void}, the exceptions are
1204 @code{ui_out_stream_new} (which returns a pointer to the newly created
1205 object) and the @code{make_cleanup} routines.
1207 The first parameter is always the @code{ui_out} vector object, a pointer
1208 to a @code{struct ui_out}.
1210 The @var{format} parameter is like in @code{printf} family of functions.
1211 When it is present, there must also be a variable list of arguments
1212 sufficient used to satisfy the @code{%} specifiers in the supplied
1215 When a character string argument is not used in a @code{ui_out} function
1216 call, a @code{NULL} pointer has to be supplied instead.
1219 @subsection Table, Tuple and List Functions
1221 @cindex list output functions
1222 @cindex table output functions
1223 @cindex tuple output functions
1224 This section introduces @code{ui_out} routines for building lists,
1225 tuples and tables. The routines to output the actual data items
1226 (fields) are presented in the next section.
1228 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
1229 containing information about an object; a @dfn{list} is a sequence of
1230 fields where each field describes an identical object.
1232 Use the @dfn{table} functions when your output consists of a list of
1233 rows (tuples) and the console output should include a heading. Use this
1234 even when you are listing just one object but you still want the header.
1236 @cindex nesting level in @code{ui_out} functions
1237 Tables can not be nested. Tuples and lists can be nested up to a
1238 maximum of five levels.
1240 The overall structure of the table output code is something like this:
1255 Here is the description of table-, tuple- and list-related @code{ui_out}
1258 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
1259 The function @code{ui_out_table_begin} marks the beginning of the output
1260 of a table. It should always be called before any other @code{ui_out}
1261 function for a given table. @var{nbrofcols} is the number of columns in
1262 the table. @var{nr_rows} is the number of rows in the table.
1263 @var{tblid} is an optional string identifying the table. The string
1264 pointed to by @var{tblid} is copied by the implementation of
1265 @code{ui_out_table_begin}, so the application can free the string if it
1266 was @code{malloc}ed.
1268 The companion function @code{ui_out_table_end}, described below, marks
1269 the end of the table's output.
1272 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
1273 @code{ui_out_table_header} provides the header information for a single
1274 table column. You call this function several times, one each for every
1275 column of the table, after @code{ui_out_table_begin}, but before
1276 @code{ui_out_table_body}.
1278 The value of @var{width} gives the column width in characters. The
1279 value of @var{alignment} is one of @code{left}, @code{center}, and
1280 @code{right}, and it specifies how to align the header: left-justify,
1281 center, or right-justify it. @var{colhdr} points to a string that
1282 specifies the column header; the implementation copies that string, so
1283 column header strings in @code{malloc}ed storage can be freed after the
1287 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
1288 This function delimits the table header from the table body.
1291 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
1292 This function signals the end of a table's output. It should be called
1293 after the table body has been produced by the list and field output
1296 There should be exactly one call to @code{ui_out_table_end} for each
1297 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
1298 will signal an internal error.
1301 The output of the tuples that represent the table rows must follow the
1302 call to @code{ui_out_table_body} and precede the call to
1303 @code{ui_out_table_end}. You build a tuple by calling
1304 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
1305 calls to functions which actually output fields between them.
1307 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
1308 This function marks the beginning of a tuple output. @var{id} points
1309 to an optional string that identifies the tuple; it is copied by the
1310 implementation, and so strings in @code{malloc}ed storage can be freed
1314 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
1315 This function signals an end of a tuple output. There should be exactly
1316 one call to @code{ui_out_tuple_end} for each call to
1317 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
1321 @deftypefun {struct cleanup *} make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1322 This function first opens the tuple and then establishes a cleanup
1323 (@pxref{Misc Guidelines, Cleanups}) to close the tuple.
1324 It provides a convenient and correct implementation of the
1325 non-portable@footnote{The function cast is not portable ISO C.} code sequence:
1327 struct cleanup *old_cleanup;
1328 ui_out_tuple_begin (uiout, "...");
1329 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
1334 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
1335 This function marks the beginning of a list output. @var{id} points to
1336 an optional string that identifies the list; it is copied by the
1337 implementation, and so strings in @code{malloc}ed storage can be freed
1341 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
1342 This function signals an end of a list output. There should be exactly
1343 one call to @code{ui_out_list_end} for each call to
1344 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1348 @deftypefun {struct cleanup *} make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1349 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1350 opens a list and then establishes cleanup
1351 (@pxref{Misc Guidelines, Cleanups})
1352 that will close the list.
1355 @subsection Item Output Functions
1357 @cindex item output functions
1358 @cindex field output functions
1360 The functions described below produce output for the actual data
1361 items, or fields, which contain information about the object.
1363 Choose the appropriate function accordingly to your particular needs.
1365 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1366 This is the most general output function. It produces the
1367 representation of the data in the variable-length argument list
1368 according to formatting specifications in @var{format}, a
1369 @code{printf}-like format string. The optional argument @var{fldname}
1370 supplies the name of the field. The data items themselves are
1371 supplied as additional arguments after @var{format}.
1373 This generic function should be used only when it is not possible to
1374 use one of the specialized versions (see below).
1377 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1378 This function outputs a value of an @code{int} variable. It uses the
1379 @code{"%d"} output conversion specification. @var{fldname} specifies
1380 the name of the field.
1383 @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})
1384 This function outputs a value of an @code{int} variable. It differs from
1385 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1386 @var{fldname} specifies
1387 the name of the field.
1390 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, struct gdbarch *@var{gdbarch}, CORE_ADDR @var{address})
1391 This function outputs an address as appropriate for @var{gdbarch}.
1394 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1395 This function outputs a string using the @code{"%s"} conversion
1399 Sometimes, there's a need to compose your output piece by piece using
1400 functions that operate on a stream, such as @code{value_print} or
1401 @code{fprintf_symbol_filtered}. These functions accept an argument of
1402 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1403 used to store the data stream used for the output. When you use one
1404 of these functions, you need a way to pass their results stored in a
1405 @code{ui_file} object to the @code{ui_out} functions. To this end,
1406 you first create a @code{ui_stream} object by calling
1407 @code{ui_out_stream_new}, pass the @code{stream} member of that
1408 @code{ui_stream} object to @code{value_print} and similar functions,
1409 and finally call @code{ui_out_field_stream} to output the field you
1410 constructed. When the @code{ui_stream} object is no longer needed,
1411 you should destroy it and free its memory by calling
1412 @code{ui_out_stream_delete}.
1414 @deftypefun {struct ui_stream *} ui_out_stream_new (struct ui_out *@var{uiout})
1415 This function creates a new @code{ui_stream} object which uses the
1416 same output methods as the @code{ui_out} object whose pointer is
1417 passed in @var{uiout}. It returns a pointer to the newly created
1418 @code{ui_stream} object.
1421 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1422 This functions destroys a @code{ui_stream} object specified by
1426 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1427 This function consumes all the data accumulated in
1428 @code{streambuf->stream} and outputs it like
1429 @code{ui_out_field_string} does. After a call to
1430 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1431 the stream is still valid and may be used for producing more fields.
1434 @strong{Important:} If there is any chance that your code could bail
1435 out before completing output generation and reaching the point where
1436 @code{ui_out_stream_delete} is called, it is necessary to set up a
1437 cleanup, to avoid leaking memory and other resources. Here's a
1438 skeleton code to do that:
1441 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1442 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1447 If the function already has the old cleanup chain set (for other kinds
1448 of cleanups), you just have to add your cleanup to it:
1451 mybuf = ui_out_stream_new (uiout);
1452 make_cleanup (ui_out_stream_delete, mybuf);
1455 Note that with cleanups in place, you should not call
1456 @code{ui_out_stream_delete} directly, or you would attempt to free the
1459 @subsection Utility Output Functions
1461 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1462 This function skips a field in a table. Use it if you have to leave
1463 an empty field without disrupting the table alignment. The argument
1464 @var{fldname} specifies a name for the (missing) filed.
1467 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1468 This function outputs the text in @var{string} in a way that makes it
1469 easy to be read by humans. For example, the console implementation of
1470 this method filters the text through a built-in pager, to prevent it
1471 from scrolling off the visible portion of the screen.
1473 Use this function for printing relatively long chunks of text around
1474 the actual field data: the text it produces is not aligned according
1475 to the table's format. Use @code{ui_out_field_string} to output a
1476 string field, and use @code{ui_out_message}, described below, to
1477 output short messages.
1480 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1481 This function outputs @var{nspaces} spaces. It is handy to align the
1482 text produced by @code{ui_out_text} with the rest of the table or
1486 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1487 This function produces a formatted message, provided that the current
1488 verbosity level is at least as large as given by @var{verbosity}. The
1489 current verbosity level is specified by the user with the @samp{set
1490 verbositylevel} command.@footnote{As of this writing (April 2001),
1491 setting verbosity level is not yet implemented, and is always returned
1492 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1493 argument more than zero will cause the message to never be printed.}
1496 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1497 This function gives the console output filter (a paging filter) a hint
1498 of where to break lines which are too long. Ignored for all other
1499 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1500 be printed to indent the wrapped text on the next line; it must remain
1501 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1502 explicit newline is produced by one of the other functions. If
1503 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1506 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1507 This function flushes whatever output has been accumulated so far, if
1508 the UI buffers output.
1512 @subsection Examples of Use of @code{ui_out} functions
1514 @cindex using @code{ui_out} functions
1515 @cindex @code{ui_out} functions, usage examples
1516 This section gives some practical examples of using the @code{ui_out}
1517 functions to generalize the old console-oriented code in
1518 @value{GDBN}. The examples all come from functions defined on the
1519 @file{breakpoints.c} file.
1521 This example, from the @code{breakpoint_1} function, shows how to
1524 The original code was:
1527 if (!found_a_breakpoint++)
1529 annotate_breakpoints_headers ();
1532 printf_filtered ("Num ");
1534 printf_filtered ("Type ");
1536 printf_filtered ("Disp ");
1538 printf_filtered ("Enb ");
1542 printf_filtered ("Address ");
1545 printf_filtered ("What\n");
1547 annotate_breakpoints_table ();
1551 Here's the new version:
1554 nr_printable_breakpoints = @dots{};
1557 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1559 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1561 if (nr_printable_breakpoints > 0)
1562 annotate_breakpoints_headers ();
1563 if (nr_printable_breakpoints > 0)
1565 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1566 if (nr_printable_breakpoints > 0)
1568 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1569 if (nr_printable_breakpoints > 0)
1571 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1572 if (nr_printable_breakpoints > 0)
1574 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1577 if (nr_printable_breakpoints > 0)
1579 if (print_address_bits <= 32)
1580 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1582 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1584 if (nr_printable_breakpoints > 0)
1586 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1587 ui_out_table_body (uiout);
1588 if (nr_printable_breakpoints > 0)
1589 annotate_breakpoints_table ();
1592 This example, from the @code{print_one_breakpoint} function, shows how
1593 to produce the actual data for the table whose structure was defined
1594 in the above example. The original code was:
1599 printf_filtered ("%-3d ", b->number);
1601 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1602 || ((int) b->type != bptypes[(int) b->type].type))
1603 internal_error ("bptypes table does not describe type #%d.",
1605 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1607 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1609 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1613 This is the new version:
1617 ui_out_tuple_begin (uiout, "bkpt");
1619 ui_out_field_int (uiout, "number", b->number);
1621 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1622 || ((int) b->type != bptypes[(int) b->type].type))
1623 internal_error ("bptypes table does not describe type #%d.",
1625 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1627 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1629 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1633 This example, also from @code{print_one_breakpoint}, shows how to
1634 produce a complicated output field using the @code{print_expression}
1635 functions which requires a stream to be passed. It also shows how to
1636 automate stream destruction with cleanups. The original code was:
1640 print_expression (b->exp, gdb_stdout);
1646 struct ui_stream *stb = ui_out_stream_new (uiout);
1647 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1650 print_expression (b->exp, stb->stream);
1651 ui_out_field_stream (uiout, "what", local_stream);
1654 This example, also from @code{print_one_breakpoint}, shows how to use
1655 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1660 if (b->dll_pathname == NULL)
1661 printf_filtered ("<any library> ");
1663 printf_filtered ("library \"%s\" ", b->dll_pathname);
1670 if (b->dll_pathname == NULL)
1672 ui_out_field_string (uiout, "what", "<any library>");
1673 ui_out_spaces (uiout, 1);
1677 ui_out_text (uiout, "library \"");
1678 ui_out_field_string (uiout, "what", b->dll_pathname);
1679 ui_out_text (uiout, "\" ");
1683 The following example from @code{print_one_breakpoint} shows how to
1684 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1689 if (b->forked_inferior_pid != 0)
1690 printf_filtered ("process %d ", b->forked_inferior_pid);
1697 if (b->forked_inferior_pid != 0)
1699 ui_out_text (uiout, "process ");
1700 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1701 ui_out_spaces (uiout, 1);
1705 Here's an example of using @code{ui_out_field_string}. The original
1710 if (b->exec_pathname != NULL)
1711 printf_filtered ("program \"%s\" ", b->exec_pathname);
1718 if (b->exec_pathname != NULL)
1720 ui_out_text (uiout, "program \"");
1721 ui_out_field_string (uiout, "what", b->exec_pathname);
1722 ui_out_text (uiout, "\" ");
1726 Finally, here's an example of printing an address. The original code:
1730 printf_filtered ("%s ",
1731 hex_string_custom ((unsigned long) b->address, 8));
1738 ui_out_field_core_addr (uiout, "Address", b->address);
1742 @section Console Printing
1751 @cindex @code{libgdb}
1752 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1753 to provide an API to @value{GDBN}'s functionality.
1756 @cindex @code{libgdb}
1757 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1758 better able to support graphical and other environments.
1760 Since @code{libgdb} development is on-going, its architecture is still
1761 evolving. The following components have so far been identified:
1765 Observer - @file{gdb-events.h}.
1767 Builder - @file{ui-out.h}
1769 Event Loop - @file{event-loop.h}
1771 Library - @file{gdb.h}
1774 The model that ties these components together is described below.
1776 @section The @code{libgdb} Model
1778 A client of @code{libgdb} interacts with the library in two ways.
1782 As an observer (using @file{gdb-events}) receiving notifications from
1783 @code{libgdb} of any internal state changes (break point changes, run
1786 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1787 obtain various status values from @value{GDBN}.
1790 Since @code{libgdb} could have multiple clients (e.g., a GUI supporting
1791 the existing @value{GDBN} CLI), those clients must co-operate when
1792 controlling @code{libgdb}. In particular, a client must ensure that
1793 @code{libgdb} is idle (i.e.@: no other client is using @code{libgdb})
1794 before responding to a @file{gdb-event} by making a query.
1796 @section CLI support
1798 At present @value{GDBN}'s CLI is very much entangled in with the core of
1799 @code{libgdb}. Consequently, a client wishing to include the CLI in
1800 their interface needs to carefully co-ordinate its own and the CLI's
1803 It is suggested that the client set @code{libgdb} up to be bi-modal
1804 (alternate between CLI and client query modes). The notes below sketch
1809 The client registers itself as an observer of @code{libgdb}.
1811 The client create and install @code{cli-out} builder using its own
1812 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1813 @code{gdb_stdout} streams.
1815 The client creates a separate custom @code{ui-out} builder that is only
1816 used while making direct queries to @code{libgdb}.
1819 When the client receives input intended for the CLI, it simply passes it
1820 along. Since the @code{cli-out} builder is installed by default, all
1821 the CLI output in response to that command is routed (pronounced rooted)
1822 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1823 At the same time, the client is kept abreast of internal changes by
1824 virtue of being a @code{libgdb} observer.
1826 The only restriction on the client is that it must wait until
1827 @code{libgdb} becomes idle before initiating any queries (using the
1828 client's custom builder).
1830 @section @code{libgdb} components
1832 @subheading Observer - @file{gdb-events.h}
1833 @file{gdb-events} provides the client with a very raw mechanism that can
1834 be used to implement an observer. At present it only allows for one
1835 observer and that observer must, internally, handle the need to delay
1836 the processing of any event notifications until after @code{libgdb} has
1837 finished the current command.
1839 @subheading Builder - @file{ui-out.h}
1840 @file{ui-out} provides the infrastructure necessary for a client to
1841 create a builder. That builder is then passed down to @code{libgdb}
1842 when doing any queries.
1844 @subheading Event Loop - @file{event-loop.h}
1845 @c There could be an entire section on the event-loop
1846 @file{event-loop}, currently non-re-entrant, provides a simple event
1847 loop. A client would need to either plug its self into this loop or,
1848 implement a new event-loop that @value{GDBN} would use.
1850 The event-loop will eventually be made re-entrant. This is so that
1851 @value{GDBN} can better handle the problem of some commands blocking
1852 instead of returning.
1854 @subheading Library - @file{gdb.h}
1855 @file{libgdb} is the most obvious component of this system. It provides
1856 the query interface. Each function is parameterized by a @code{ui-out}
1857 builder. The result of the query is constructed using that builder
1858 before the query function returns.
1865 @cindex @code{value} structure
1866 @value{GDBN} uses @code{struct value}, or @dfn{values}, as an internal
1867 abstraction for the representation of a variety of inferior objects
1868 and @value{GDBN} convenience objects.
1870 Values have an associated @code{struct type}, that describes a virtual
1871 view of the raw data or object stored in or accessed through the
1874 A value is in addition discriminated by its lvalue-ness, given its
1875 @code{enum lval_type} enumeration type:
1877 @cindex @code{lval_type} enumeration, for values.
1879 @item @code{not_lval}
1880 This value is not an lval. It can't be assigned to.
1882 @item @code{lval_memory}
1883 This value represents an object in memory.
1885 @item @code{lval_register}
1886 This value represents an object that lives in a register.
1888 @item @code{lval_internalvar}
1889 Represents the value of an internal variable.
1891 @item @code{lval_internalvar_component}
1892 Represents part of a @value{GDBN} internal variable. E.g., a
1895 @cindex computed values
1896 @item @code{lval_computed}
1897 These are ``computed'' values. They allow creating specialized value
1898 objects for specific purposes, all abstracted away from the core value
1899 support code. The creator of such a value writes specialized
1900 functions to handle the reading and writing to/from the value's
1901 backend data, and optionally, a ``copy operator'' and a
1904 Pointers to these functions are stored in a @code{struct lval_funcs}
1905 instance (declared in @file{value.h}), and passed to the
1906 @code{allocate_computed_value} function, as in the example below.
1910 nil_value_read (struct value *v)
1912 /* This callback reads data from some backend, and stores it in V.
1913 In this case, we always read null data. You'll want to fill in
1914 something more interesting. */
1916 memset (value_contents_all_raw (v),
1918 TYPE_LENGTH (value_type (v)));
1922 nil_value_write (struct value *v, struct value *fromval)
1924 /* Takes the data from FROMVAL and stores it in the backend of V. */
1926 to_oblivion (value_contents_all_raw (fromval),
1928 TYPE_LENGTH (value_type (fromval)));
1931 static struct lval_funcs nil_value_funcs =
1938 make_nil_value (void)
1943 type = make_nils_type ();
1944 v = allocate_computed_value (type, &nil_value_funcs, NULL);
1950 See the implementation of the @code{$_siginfo} convenience variable in
1951 @file{infrun.c} as a real example use of lval_computed.
1956 @chapter Stack Frames
1959 @cindex call stack frame
1960 A frame is a construct that @value{GDBN} uses to keep track of calling
1961 and called functions.
1963 @cindex unwind frame
1964 @value{GDBN}'s frame model, a fresh design, was implemented with the
1965 need to support @sc{dwarf}'s Call Frame Information in mind. In fact,
1966 the term ``unwind'' is taken directly from that specification.
1967 Developers wishing to learn more about unwinders, are encouraged to
1968 read the @sc{dwarf} specification, available from
1969 @url{http://www.dwarfstd.org}.
1971 @findex frame_register_unwind
1972 @findex get_frame_register
1973 @value{GDBN}'s model is that you find a frame's registers by
1974 ``unwinding'' them from the next younger frame. That is,
1975 @samp{get_frame_register} which returns the value of a register in
1976 frame #1 (the next-to-youngest frame), is implemented by calling frame
1977 #0's @code{frame_register_unwind} (the youngest frame). But then the
1978 obvious question is: how do you access the registers of the youngest
1981 @cindex sentinel frame
1982 @findex get_frame_type
1983 @vindex SENTINEL_FRAME
1984 To answer this question, @value{GDBN} has the @dfn{sentinel} frame, the
1985 ``-1st'' frame. Unwinding registers from the sentinel frame gives you
1986 the current values of the youngest real frame's registers. If @var{f}
1987 is a sentinel frame, then @code{get_frame_type (@var{f}) @equiv{}
1990 @section Selecting an Unwinder
1992 @findex frame_unwind_prepend_unwinder
1993 @findex frame_unwind_append_unwinder
1994 The architecture registers a list of frame unwinders (@code{struct
1995 frame_unwind}), using the functions
1996 @code{frame_unwind_prepend_unwinder} and
1997 @code{frame_unwind_append_unwinder}. Each unwinder includes a
1998 sniffer. Whenever @value{GDBN} needs to unwind a frame (to fetch the
1999 previous frame's registers or the current frame's ID), it calls
2000 registered sniffers in order to find one which recognizes the frame.
2001 The first time a sniffer returns non-zero, the corresponding unwinder
2002 is assigned to the frame.
2004 @section Unwinding the Frame ID
2007 Every frame has an associated ID, of type @code{struct frame_id}.
2008 The ID includes the stack base and function start address for
2009 the frame. The ID persists through the entire life of the frame,
2010 including while other called frames are running; it is used to
2011 locate an appropriate @code{struct frame_info} from the cache.
2013 Every time the inferior stops, and at various other times, the frame
2014 cache is flushed. Because of this, parts of @value{GDBN} which need
2015 to keep track of individual frames cannot use pointers to @code{struct
2016 frame_info}. A frame ID provides a stable reference to a frame, even
2017 when the unwinder must be run again to generate a new @code{struct
2018 frame_info} for the same frame.
2020 The frame's unwinder's @code{this_id} method is called to find the ID.
2021 Note that this is different from register unwinding, where the next
2022 frame's @code{prev_register} is called to unwind this frame's
2025 Both stack base and function address are required to identify the
2026 frame, because a recursive function has the same function address for
2027 two consecutive frames and a leaf function may have the same stack
2028 address as its caller. On some platforms, a third address is part of
2029 the ID to further disambiguate frames---for instance, on IA-64
2030 the separate register stack address is included in the ID.
2032 An invalid frame ID (@code{outer_frame_id}) returned from the
2033 @code{this_id} method means to stop unwinding after this frame.
2035 @code{null_frame_id} is another invalid frame ID which should be used
2036 when there is no frame. For instance, certain breakpoints are attached
2037 to a specific frame, and that frame is identified through its frame ID
2038 (we use this to implement the "finish" command). Using
2039 @code{null_frame_id} as the frame ID for a given breakpoint means
2040 that the breakpoint is not specific to any frame. The @code{this_id}
2041 method should never return @code{null_frame_id}.
2043 @section Unwinding Registers
2045 Each unwinder includes a @code{prev_register} method. This method
2046 takes a frame, an associated cache pointer, and a register number.
2047 It returns a @code{struct value *} describing the requested register,
2048 as saved by this frame. This is the value of the register that is
2049 current in this frame's caller.
2051 The returned value must have the same type as the register. It may
2052 have any lvalue type. In most circumstances one of these routines
2053 will generate the appropriate value:
2056 @item frame_unwind_got_optimized
2057 @findex frame_unwind_got_optimized
2058 This register was not saved.
2060 @item frame_unwind_got_register
2061 @findex frame_unwind_got_register
2062 This register was copied into another register in this frame. This
2063 is also used for unchanged registers; they are ``copied'' into the
2066 @item frame_unwind_got_memory
2067 @findex frame_unwind_got_memory
2068 This register was saved in memory.
2070 @item frame_unwind_got_constant
2071 @findex frame_unwind_got_constant
2072 This register was not saved, but the unwinder can compute the previous
2073 value some other way.
2075 @item frame_unwind_got_address
2076 @findex frame_unwind_got_address
2077 Same as @code{frame_unwind_got_constant}, except that the value is a target
2078 address. This is frequently used for the stack pointer, which is not
2079 explicitly saved but has a known offset from this frame's stack
2080 pointer. For architectures with a flat unified address space, this is
2081 generally the same as @code{frame_unwind_got_constant}.
2084 @node Symbol Handling
2086 @chapter Symbol Handling
2088 Symbols are a key part of @value{GDBN}'s operation. Symbols include
2089 variables, functions, and types.
2091 Symbol information for a large program can be truly massive, and
2092 reading of symbol information is one of the major performance
2093 bottlenecks in @value{GDBN}; it can take many minutes to process it
2094 all. Studies have shown that nearly all the time spent is
2095 computational, rather than file reading.
2097 One of the ways for @value{GDBN} to provide a good user experience is
2098 to start up quickly, taking no more than a few seconds. It is simply
2099 not possible to process all of a program's debugging info in that
2100 time, and so we attempt to handle symbols incrementally. For instance,
2101 we create @dfn{partial symbol tables} consisting of only selected
2102 symbols, and only expand them to full symbol tables when necessary.
2104 @section Symbol Reading
2106 @cindex symbol reading
2107 @cindex reading of symbols
2108 @cindex symbol files
2109 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
2110 file is the file containing the program which @value{GDBN} is
2111 debugging. @value{GDBN} can be directed to use a different file for
2112 symbols (with the @samp{symbol-file} command), and it can also read
2113 more symbols via the @samp{add-file} and @samp{load} commands. In
2114 addition, it may bring in more symbols while loading shared
2117 @findex find_sym_fns
2118 Symbol files are initially opened by code in @file{symfile.c} using
2119 the BFD library (@pxref{Support Libraries}). BFD identifies the type
2120 of the file by examining its header. @code{find_sym_fns} then uses
2121 this identification to locate a set of symbol-reading functions.
2123 @findex add_symtab_fns
2124 @cindex @code{sym_fns} structure
2125 @cindex adding a symbol-reading module
2126 Symbol-reading modules identify themselves to @value{GDBN} by calling
2127 @code{add_symtab_fns} during their module initialization. The argument
2128 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
2129 name (or name prefix) of the symbol format, the length of the prefix,
2130 and pointers to four functions. These functions are called at various
2131 times to process symbol files whose identification matches the specified
2134 The functions supplied by each module are:
2137 @item @var{xyz}_symfile_init(struct sym_fns *sf)
2139 @cindex secondary symbol file
2140 Called from @code{symbol_file_add} when we are about to read a new
2141 symbol file. This function should clean up any internal state (possibly
2142 resulting from half-read previous files, for example) and prepare to
2143 read a new symbol file. Note that the symbol file which we are reading
2144 might be a new ``main'' symbol file, or might be a secondary symbol file
2145 whose symbols are being added to the existing symbol table.
2147 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
2148 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
2149 new symbol file being read. Its @code{private} field has been zeroed,
2150 and can be modified as desired. Typically, a struct of private
2151 information will be @code{malloc}'d, and a pointer to it will be placed
2152 in the @code{private} field.
2154 There is no result from @code{@var{xyz}_symfile_init}, but it can call
2155 @code{error} if it detects an unavoidable problem.
2157 @item @var{xyz}_new_init()
2159 Called from @code{symbol_file_add} when discarding existing symbols.
2160 This function needs only handle the symbol-reading module's internal
2161 state; the symbol table data structures visible to the rest of
2162 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
2163 arguments and no result. It may be called after
2164 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
2165 may be called alone if all symbols are simply being discarded.
2167 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
2169 Called from @code{symbol_file_add} to actually read the symbols from a
2170 symbol-file into a set of psymtabs or symtabs.
2172 @code{sf} points to the @code{struct sym_fns} originally passed to
2173 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
2174 the offset between the file's specified start address and its true
2175 address in memory. @code{mainline} is 1 if this is the main symbol
2176 table being read, and 0 if a secondary symbol file (e.g., shared library
2177 or dynamically loaded file) is being read.@refill
2180 In addition, if a symbol-reading module creates psymtabs when
2181 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
2182 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
2183 from any point in the @value{GDBN} symbol-handling code.
2186 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
2188 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
2189 the psymtab has not already been read in and had its @code{pst->symtab}
2190 pointer set. The argument is the psymtab to be fleshed-out into a
2191 symtab. Upon return, @code{pst->readin} should have been set to 1, and
2192 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
2193 zero if there were no symbols in that part of the symbol file.
2196 @section Partial Symbol Tables
2198 @value{GDBN} has three types of symbol tables:
2201 @cindex full symbol table
2204 Full symbol tables (@dfn{symtabs}). These contain the main
2205 information about symbols and addresses.
2209 Partial symbol tables (@dfn{psymtabs}). These contain enough
2210 information to know when to read the corresponding part of the full
2213 @cindex minimal symbol table
2216 Minimal symbol tables (@dfn{msymtabs}). These contain information
2217 gleaned from non-debugging symbols.
2220 @cindex partial symbol table
2221 This section describes partial symbol tables.
2223 A psymtab is constructed by doing a very quick pass over an executable
2224 file's debugging information. Small amounts of information are
2225 extracted---enough to identify which parts of the symbol table will
2226 need to be re-read and fully digested later, when the user needs the
2227 information. The speed of this pass causes @value{GDBN} to start up very
2228 quickly. Later, as the detailed rereading occurs, it occurs in small
2229 pieces, at various times, and the delay therefrom is mostly invisible to
2231 @c (@xref{Symbol Reading}.)
2233 The symbols that show up in a file's psymtab should be, roughly, those
2234 visible to the debugger's user when the program is not running code from
2235 that file. These include external symbols and types, static symbols and
2236 types, and @code{enum} values declared at file scope.
2238 The psymtab also contains the range of instruction addresses that the
2239 full symbol table would represent.
2241 @cindex finding a symbol
2242 @cindex symbol lookup
2243 The idea is that there are only two ways for the user (or much of the
2244 code in the debugger) to reference a symbol:
2247 @findex find_pc_function
2248 @findex find_pc_line
2250 By its address (e.g., execution stops at some address which is inside a
2251 function in this file). The address will be noticed to be in the
2252 range of this psymtab, and the full symtab will be read in.
2253 @code{find_pc_function}, @code{find_pc_line}, and other
2254 @code{find_pc_@dots{}} functions handle this.
2256 @cindex lookup_symbol
2259 (e.g., the user asks to print a variable, or set a breakpoint on a
2260 function). Global names and file-scope names will be found in the
2261 psymtab, which will cause the symtab to be pulled in. Local names will
2262 have to be qualified by a global name, or a file-scope name, in which
2263 case we will have already read in the symtab as we evaluated the
2264 qualifier. Or, a local symbol can be referenced when we are ``in'' a
2265 local scope, in which case the first case applies. @code{lookup_symbol}
2266 does most of the work here.
2269 The only reason that psymtabs exist is to cause a symtab to be read in
2270 at the right moment. Any symbol that can be elided from a psymtab,
2271 while still causing that to happen, should not appear in it. Since
2272 psymtabs don't have the idea of scope, you can't put local symbols in
2273 them anyway. Psymtabs don't have the idea of the type of a symbol,
2274 either, so types need not appear, unless they will be referenced by
2277 It is a bug for @value{GDBN} to behave one way when only a psymtab has
2278 been read, and another way if the corresponding symtab has been read
2279 in. Such bugs are typically caused by a psymtab that does not contain
2280 all the visible symbols, or which has the wrong instruction address
2283 The psymtab for a particular section of a symbol file (objfile) could be
2284 thrown away after the symtab has been read in. The symtab should always
2285 be searched before the psymtab, so the psymtab will never be used (in a
2286 bug-free environment). Currently, psymtabs are allocated on an obstack,
2287 and all the psymbols themselves are allocated in a pair of large arrays
2288 on an obstack, so there is little to be gained by trying to free them
2289 unless you want to do a lot more work.
2291 Whether or not psymtabs are created depends on the objfile's symbol
2292 reader. The core of @value{GDBN} hides the details of partial symbols
2293 and partial symbol tables behind a set of function pointers known as
2294 the @dfn{quick symbol functions}. These are documented in
2299 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
2301 @cindex fundamental types
2302 These are the fundamental types that @value{GDBN} uses internally. Fundamental
2303 types from the various debugging formats (stabs, ELF, etc) are mapped
2304 into one of these. They are basically a union of all fundamental types
2305 that @value{GDBN} knows about for all the languages that @value{GDBN}
2308 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
2311 Each time @value{GDBN} builds an internal type, it marks it with one
2312 of these types. The type may be a fundamental type, such as
2313 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
2314 which is a pointer to another type. Typically, several @code{FT_*}
2315 types map to one @code{TYPE_CODE_*} type, and are distinguished by
2316 other members of the type struct, such as whether the type is signed
2317 or unsigned, and how many bits it uses.
2319 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
2321 These are instances of type structs that roughly correspond to
2322 fundamental types and are created as global types for @value{GDBN} to
2323 use for various ugly historical reasons. We eventually want to
2324 eliminate these. Note for example that @code{builtin_type_int}
2325 initialized in @file{gdbtypes.c} is basically the same as a
2326 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
2327 an @code{FT_INTEGER} fundamental type. The difference is that the
2328 @code{builtin_type} is not associated with any particular objfile, and
2329 only one instance exists, while @file{c-lang.c} builds as many
2330 @code{TYPE_CODE_INT} types as needed, with each one associated with
2331 some particular objfile.
2333 @section Object File Formats
2334 @cindex object file formats
2338 @cindex @code{a.out} format
2339 The @code{a.out} format is the original file format for Unix. It
2340 consists of three sections: @code{text}, @code{data}, and @code{bss},
2341 which are for program code, initialized data, and uninitialized data,
2344 The @code{a.out} format is so simple that it doesn't have any reserved
2345 place for debugging information. (Hey, the original Unix hackers used
2346 @samp{adb}, which is a machine-language debugger!) The only debugging
2347 format for @code{a.out} is stabs, which is encoded as a set of normal
2348 symbols with distinctive attributes.
2350 The basic @code{a.out} reader is in @file{dbxread.c}.
2355 The COFF format was introduced with System V Release 3 (SVR3) Unix.
2356 COFF files may have multiple sections, each prefixed by a header. The
2357 number of sections is limited.
2359 The COFF specification includes support for debugging. Although this
2360 was a step forward, the debugging information was woefully limited.
2361 For instance, it was not possible to represent code that came from an
2362 included file. GNU's COFF-using configs often use stabs-type info,
2363 encapsulated in special sections.
2365 The COFF reader is in @file{coffread.c}.
2369 @cindex ECOFF format
2370 ECOFF is an extended COFF originally introduced for Mips and Alpha
2373 The basic ECOFF reader is in @file{mipsread.c}.
2377 @cindex XCOFF format
2378 The IBM RS/6000 running AIX uses an object file format called XCOFF.
2379 The COFF sections, symbols, and line numbers are used, but debugging
2380 symbols are @code{dbx}-style stabs whose strings are located in the
2381 @code{.debug} section (rather than the string table). For more
2382 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
2384 The shared library scheme has a clean interface for figuring out what
2385 shared libraries are in use, but the catch is that everything which
2386 refers to addresses (symbol tables and breakpoints at least) needs to be
2387 relocated for both shared libraries and the main executable. At least
2388 using the standard mechanism this can only be done once the program has
2389 been run (or the core file has been read).
2393 @cindex PE-COFF format
2394 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
2395 executables. PE is basically COFF with additional headers.
2397 While BFD includes special PE support, @value{GDBN} needs only the basic
2403 The ELF format came with System V Release 4 (SVR4) Unix. ELF is
2404 similar to COFF in being organized into a number of sections, but it
2405 removes many of COFF's limitations. Debugging info may be either stabs
2406 encapsulated in ELF sections, or more commonly these days, DWARF.
2408 The basic ELF reader is in @file{elfread.c}.
2413 SOM is HP's object file and debug format (not to be confused with IBM's
2414 SOM, which is a cross-language ABI).
2416 The SOM reader is in @file{somread.c}.
2418 @section Debugging File Formats
2420 This section describes characteristics of debugging information that
2421 are independent of the object file format.
2425 @cindex stabs debugging info
2426 @code{stabs} started out as special symbols within the @code{a.out}
2427 format. Since then, it has been encapsulated into other file
2428 formats, such as COFF and ELF.
2430 While @file{dbxread.c} does some of the basic stab processing,
2431 including for encapsulated versions, @file{stabsread.c} does
2436 @cindex COFF debugging info
2437 The basic COFF definition includes debugging information. The level
2438 of support is minimal and non-extensible, and is not often used.
2440 @subsection Mips debug (Third Eye)
2442 @cindex ECOFF debugging info
2443 ECOFF includes a definition of a special debug format.
2445 The file @file{mdebugread.c} implements reading for this format.
2447 @c mention DWARF 1 as a formerly-supported format
2451 @cindex DWARF 2 debugging info
2452 DWARF 2 is an improved but incompatible version of DWARF 1.
2454 The DWARF 2 reader is in @file{dwarf2read.c}.
2456 @subsection Compressed DWARF 2
2458 @cindex Compressed DWARF 2 debugging info
2459 Compressed DWARF 2 is not technically a separate debugging format, but
2460 merely DWARF 2 debug information that has been compressed. In this
2461 format, every object-file section holding DWARF 2 debugging
2462 information is compressed and prepended with a header. (The section
2463 is also typically renamed, so a section called @code{.debug_info} in a
2464 DWARF 2 binary would be called @code{.zdebug_info} in a compressed
2465 DWARF 2 binary.) The header is 12 bytes long:
2469 4 bytes: the literal string ``ZLIB''
2471 8 bytes: the uncompressed size of the section, in big-endian byte
2475 The same reader is used for both compressed an normal DWARF 2 info.
2476 Section decompression is done in @code{zlib_decompress_section} in
2477 @file{dwarf2read.c}.
2481 @cindex DWARF 3 debugging info
2482 DWARF 3 is an improved version of DWARF 2.
2486 @cindex SOM debugging info
2487 Like COFF, the SOM definition includes debugging information.
2489 @section Adding a New Symbol Reader to @value{GDBN}
2491 @cindex adding debugging info reader
2492 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
2493 there is probably little to be done.
2495 If you need to add a new object file format, you must first add it to
2496 BFD. This is beyond the scope of this document.
2498 You must then arrange for the BFD code to provide access to the
2499 debugging symbols. Generally @value{GDBN} will have to call swapping
2500 routines from BFD and a few other BFD internal routines to locate the
2501 debugging information. As much as possible, @value{GDBN} should not
2502 depend on the BFD internal data structures.
2504 For some targets (e.g., COFF), there is a special transfer vector used
2505 to call swapping routines, since the external data structures on various
2506 platforms have different sizes and layouts. Specialized routines that
2507 will only ever be implemented by one object file format may be called
2508 directly. This interface should be described in a file
2509 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
2511 @section Memory Management for Symbol Files
2513 Most memory associated with a loaded symbol file is stored on
2514 its @code{objfile_obstack}. This includes symbols, types,
2515 namespace data, and other information produced by the symbol readers.
2517 Because this data lives on the objfile's obstack, it is automatically
2518 released when the objfile is unloaded or reloaded. Therefore one
2519 objfile must not reference symbol or type data from another objfile;
2520 they could be unloaded at different times.
2522 User convenience variables, et cetera, have associated types. Normally
2523 these types live in the associated objfile. However, when the objfile
2524 is unloaded, those types are deep copied to global memory, so that
2525 the values of the user variables and history items are not lost.
2528 @node Language Support
2530 @chapter Language Support
2532 @cindex language support
2533 @value{GDBN}'s language support is mainly driven by the symbol reader,
2534 although it is possible for the user to set the source language
2537 @value{GDBN} chooses the source language by looking at the extension
2538 of the file recorded in the debug info; @file{.c} means C, @file{.f}
2539 means Fortran, etc. It may also use a special-purpose language
2540 identifier if the debug format supports it, like with DWARF.
2542 @section Adding a Source Language to @value{GDBN}
2544 @cindex adding source language
2545 To add other languages to @value{GDBN}'s expression parser, follow the
2549 @item Create the expression parser.
2551 @cindex expression parser
2552 This should reside in a file @file{@var{lang}-exp.y}. Routines for
2553 building parsed expressions into a @code{union exp_element} list are in
2556 @cindex language parser
2557 Since we can't depend upon everyone having Bison, and YACC produces
2558 parsers that define a bunch of global names, the following lines
2559 @strong{must} be included at the top of the YACC parser, to prevent the
2560 various parsers from defining the same global names:
2563 #define yyparse @var{lang}_parse
2564 #define yylex @var{lang}_lex
2565 #define yyerror @var{lang}_error
2566 #define yylval @var{lang}_lval
2567 #define yychar @var{lang}_char
2568 #define yydebug @var{lang}_debug
2569 #define yypact @var{lang}_pact
2570 #define yyr1 @var{lang}_r1
2571 #define yyr2 @var{lang}_r2
2572 #define yydef @var{lang}_def
2573 #define yychk @var{lang}_chk
2574 #define yypgo @var{lang}_pgo
2575 #define yyact @var{lang}_act
2576 #define yyexca @var{lang}_exca
2577 #define yyerrflag @var{lang}_errflag
2578 #define yynerrs @var{lang}_nerrs
2581 At the bottom of your parser, define a @code{struct language_defn} and
2582 initialize it with the right values for your language. Define an
2583 @code{initialize_@var{lang}} routine and have it call
2584 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
2585 that your language exists. You'll need some other supporting variables
2586 and functions, which will be used via pointers from your
2587 @code{@var{lang}_language_defn}. See the declaration of @code{struct
2588 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
2589 for more information.
2591 @item Add any evaluation routines, if necessary
2593 @cindex expression evaluation routines
2594 @findex evaluate_subexp
2595 @findex prefixify_subexp
2596 @findex length_of_subexp
2597 If you need new opcodes (that represent the operations of the language),
2598 add them to the enumerated type in @file{expression.h}. Add support
2599 code for these operations in the @code{evaluate_subexp} function
2600 defined in the file @file{eval.c}. Add cases
2601 for new opcodes in two functions from @file{parse.c}:
2602 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
2603 the number of @code{exp_element}s that a given operation takes up.
2605 @item Update some existing code
2607 Add an enumerated identifier for your language to the enumerated type
2608 @code{enum language} in @file{defs.h}.
2610 Update the routines in @file{language.c} so your language is included.
2611 These routines include type predicates and such, which (in some cases)
2612 are language dependent. If your language does not appear in the switch
2613 statement, an error is reported.
2615 @vindex current_language
2616 Also included in @file{language.c} is the code that updates the variable
2617 @code{current_language}, and the routines that translate the
2618 @code{language_@var{lang}} enumerated identifier into a printable
2621 @findex _initialize_language
2622 Update the function @code{_initialize_language} to include your
2623 language. This function picks the default language upon startup, so is
2624 dependent upon which languages that @value{GDBN} is built for.
2626 @findex allocate_symtab
2627 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2628 code so that the language of each symtab (source file) is set properly.
2629 This is used to determine the language to use at each stack frame level.
2630 Currently, the language is set based upon the extension of the source
2631 file. If the language can be better inferred from the symbol
2632 information, please set the language of the symtab in the symbol-reading
2635 @findex print_subexp
2636 @findex op_print_tab
2637 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2638 expression opcodes you have added to @file{expression.h}. Also, add the
2639 printed representations of your operators to @code{op_print_tab}.
2641 @item Add a place of call
2644 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2645 @code{parse_exp_1} (defined in @file{parse.c}).
2647 @item Edit @file{Makefile.in}
2649 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2650 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2651 not get linked in, or, worse yet, it may not get @code{tar}red into the
2656 @node Host Definition
2658 @chapter Host Definition
2660 With the advent of Autoconf, it's rarely necessary to have host
2661 definition machinery anymore. The following information is provided,
2662 mainly, as an historical reference.
2664 @section Adding a New Host
2666 @cindex adding a new host
2667 @cindex host, adding
2668 @value{GDBN}'s host configuration support normally happens via Autoconf.
2669 New host-specific definitions should not be needed. Older hosts
2670 @value{GDBN} still use the host-specific definitions and files listed
2671 below, but these mostly exist for historical reasons, and will
2672 eventually disappear.
2675 @item gdb/config/@var{arch}/@var{xyz}.mh
2676 This file is a Makefile fragment that once contained both host and
2677 native configuration information (@pxref{Native Debugging}) for the
2678 machine @var{xyz}. The host configuration information is now handled
2681 Host configuration information included definitions for @code{CC},
2682 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2683 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2685 New host-only configurations do not need this file.
2689 (Files named @file{gdb/config/@var{arch}/xm-@var{xyz}.h} were once
2690 used to define host-specific macros, but were no longer needed and
2691 have all been removed.)
2693 @subheading Generic Host Support Files
2695 @cindex generic host support
2696 There are some ``generic'' versions of routines that can be used by
2700 @cindex remote debugging support
2701 @cindex serial line support
2703 This contains serial line support for Unix systems. It is included by
2704 default on all Unix-like hosts.
2707 This contains serial pipe support for Unix systems. It is included by
2708 default on all Unix-like hosts.
2711 This contains serial line support for 32-bit programs running under
2712 Windows using MinGW.
2715 This contains serial line support for 32-bit programs running under DOS,
2716 using the DJGPP (a.k.a.@: GO32) execution environment.
2718 @cindex TCP remote support
2720 This contains generic TCP support using sockets. It is included by
2721 default on all Unix-like hosts and with MinGW.
2724 @section Host Conditionals
2726 When @value{GDBN} is configured and compiled, various macros are
2727 defined or left undefined, to control compilation based on the
2728 attributes of the host system. While formerly they could be set in
2729 host-specific header files, at present they can be changed only by
2730 setting @code{CFLAGS} when building, or by editing the source code.
2732 These macros and their meanings (or if the meaning is not documented
2733 here, then one of the source files where they are used is indicated)
2737 @item @value{GDBN}INIT_FILENAME
2738 The default name of @value{GDBN}'s initialization file (normally
2741 @item CRLF_SOURCE_FILES
2742 @cindex DOS text files
2743 Define this if host files use @code{\r\n} rather than @code{\n} as a
2744 line terminator. This will cause source file listings to omit @code{\r}
2745 characters when printing and it will allow @code{\r\n} line endings of files
2746 which are ``sourced'' by gdb. It must be possible to open files in binary
2747 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2749 @item DEFAULT_PROMPT
2751 The default value of the prompt string (normally @code{"(gdb) "}).
2754 @cindex terminal device
2755 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2758 Substitute for isatty, if not available.
2761 Define this if binary files are opened the same way as text files.
2763 @item PRINTF_HAS_LONG_LONG
2764 Define this if the host can handle printing of long long integers via
2765 the printf format conversion specifier @code{ll}. This is set by the
2766 @code{configure} script.
2768 @item LSEEK_NOT_LINEAR
2769 Define this if @code{lseek (n)} does not necessarily move to byte number
2770 @code{n} in the file. This is only used when reading source files. It
2771 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2774 Define this to help placate @code{lint} in some situations.
2777 Define this to override the defaults of @code{__volatile__} or
2782 @node Target Architecture Definition
2784 @chapter Target Architecture Definition
2786 @cindex target architecture definition
2787 @value{GDBN}'s target architecture defines what sort of
2788 machine-language programs @value{GDBN} can work with, and how it works
2791 The target architecture object is implemented as the C structure
2792 @code{struct gdbarch *}. The structure, and its methods, are generated
2793 using the Bourne shell script @file{gdbarch.sh}.
2796 * OS ABI Variant Handling::
2797 * Initialize New Architecture::
2798 * Registers and Memory::
2799 * Pointers and Addresses::
2801 * Register Representation::
2802 * Frame Interpretation::
2803 * Inferior Call Setup::
2804 * Adding support for debugging core files::
2805 * Defining Other Architecture Features::
2806 * Adding a New Target::
2809 @node OS ABI Variant Handling
2810 @section Operating System ABI Variant Handling
2811 @cindex OS ABI variants
2813 @value{GDBN} provides a mechanism for handling variations in OS
2814 ABIs. An OS ABI variant may have influence over any number of
2815 variables in the target architecture definition. There are two major
2816 components in the OS ABI mechanism: sniffers and handlers.
2818 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2819 (the architecture may be wildcarded) in an attempt to determine the
2820 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2821 to be @dfn{generic}, while sniffers for a specific architecture are
2822 considered to be @dfn{specific}. A match from a specific sniffer
2823 overrides a match from a generic sniffer. Multiple sniffers for an
2824 architecture/flavour may exist, in order to differentiate between two
2825 different operating systems which use the same basic file format. The
2826 OS ABI framework provides a generic sniffer for ELF-format files which
2827 examines the @code{EI_OSABI} field of the ELF header, as well as note
2828 sections known to be used by several operating systems.
2830 @cindex fine-tuning @code{gdbarch} structure
2831 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2832 selected OS ABI. There may be only one handler for a given OS ABI
2833 for each BFD architecture.
2835 The following OS ABI variants are defined in @file{defs.h}:
2839 @findex GDB_OSABI_UNINITIALIZED
2840 @item GDB_OSABI_UNINITIALIZED
2841 Used for struct gdbarch_info if ABI is still uninitialized.
2843 @findex GDB_OSABI_UNKNOWN
2844 @item GDB_OSABI_UNKNOWN
2845 The ABI of the inferior is unknown. The default @code{gdbarch}
2846 settings for the architecture will be used.
2848 @findex GDB_OSABI_SVR4
2849 @item GDB_OSABI_SVR4
2850 UNIX System V Release 4.
2852 @findex GDB_OSABI_HURD
2853 @item GDB_OSABI_HURD
2854 GNU using the Hurd kernel.
2856 @findex GDB_OSABI_SOLARIS
2857 @item GDB_OSABI_SOLARIS
2860 @findex GDB_OSABI_OSF1
2861 @item GDB_OSABI_OSF1
2862 OSF/1, including Digital UNIX and Compaq Tru64 UNIX.
2864 @findex GDB_OSABI_LINUX
2865 @item GDB_OSABI_LINUX
2866 GNU using the Linux kernel.
2868 @findex GDB_OSABI_FREEBSD_AOUT
2869 @item GDB_OSABI_FREEBSD_AOUT
2870 FreeBSD using the @code{a.out} executable format.
2872 @findex GDB_OSABI_FREEBSD_ELF
2873 @item GDB_OSABI_FREEBSD_ELF
2874 FreeBSD using the ELF executable format.
2876 @findex GDB_OSABI_NETBSD_AOUT
2877 @item GDB_OSABI_NETBSD_AOUT
2878 NetBSD using the @code{a.out} executable format.
2880 @findex GDB_OSABI_NETBSD_ELF
2881 @item GDB_OSABI_NETBSD_ELF
2882 NetBSD using the ELF executable format.
2884 @findex GDB_OSABI_OPENBSD_ELF
2885 @item GDB_OSABI_OPENBSD_ELF
2886 OpenBSD using the ELF executable format.
2888 @findex GDB_OSABI_WINCE
2889 @item GDB_OSABI_WINCE
2892 @findex GDB_OSABI_GO32
2893 @item GDB_OSABI_GO32
2896 @findex GDB_OSABI_IRIX
2897 @item GDB_OSABI_IRIX
2900 @findex GDB_OSABI_INTERIX
2901 @item GDB_OSABI_INTERIX
2902 Interix (Posix layer for MS-Windows systems).
2904 @findex GDB_OSABI_HPUX_ELF
2905 @item GDB_OSABI_HPUX_ELF
2906 HP/UX using the ELF executable format.
2908 @findex GDB_OSABI_HPUX_SOM
2909 @item GDB_OSABI_HPUX_SOM
2910 HP/UX using the SOM executable format.
2912 @findex GDB_OSABI_QNXNTO
2913 @item GDB_OSABI_QNXNTO
2916 @findex GDB_OSABI_CYGWIN
2917 @item GDB_OSABI_CYGWIN
2920 @findex GDB_OSABI_AIX
2926 Here are the functions that make up the OS ABI framework:
2928 @deftypefun {const char *} gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2929 Return the name of the OS ABI corresponding to @var{osabi}.
2932 @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}))
2933 Register the OS ABI handler specified by @var{init_osabi} for the
2934 architecture, machine type and OS ABI specified by @var{arch},
2935 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2936 machine type, which implies the architecture's default machine type,
2940 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2941 Register the OS ABI file sniffer specified by @var{sniffer} for the
2942 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2943 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2944 be generic, and is allowed to examine @var{flavour}-flavoured files for
2948 @deftypefun {enum gdb_osabi} gdbarch_lookup_osabi (bfd *@var{abfd})
2949 Examine the file described by @var{abfd} to determine its OS ABI.
2950 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2954 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2955 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2956 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2957 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2958 architecture, a warning will be issued and the debugging session will continue
2959 with the defaults already established for @var{gdbarch}.
2962 @deftypefun void generic_elf_osabi_sniff_abi_tag_sections (bfd *@var{abfd}, asection *@var{sect}, void *@var{obj})
2963 Helper routine for ELF file sniffers. Examine the file described by
2964 @var{abfd} and look at ABI tag note sections to determine the OS ABI
2965 from the note. This function should be called via
2966 @code{bfd_map_over_sections}.
2969 @node Initialize New Architecture
2970 @section Initializing a New Architecture
2973 * How an Architecture is Represented::
2974 * Looking Up an Existing Architecture::
2975 * Creating a New Architecture::
2978 @node How an Architecture is Represented
2979 @subsection How an Architecture is Represented
2980 @cindex architecture representation
2981 @cindex representation of architecture
2983 Each @code{gdbarch} is associated with a single @sc{bfd} architecture,
2984 via a @code{bfd_arch_@var{arch}} in the @code{bfd_architecture}
2985 enumeration. The @code{gdbarch} is registered by a call to
2986 @code{register_gdbarch_init}, usually from the file's
2987 @code{_initialize_@var{filename}} routine, which will be automatically
2988 called during @value{GDBN} startup. The arguments are a @sc{bfd}
2989 architecture constant and an initialization function.
2991 @findex _initialize_@var{arch}_tdep
2992 @cindex @file{@var{arch}-tdep.c}
2993 A @value{GDBN} description for a new architecture, @var{arch} is created by
2994 defining a global function @code{_initialize_@var{arch}_tdep}, by
2995 convention in the source file @file{@var{arch}-tdep.c}. For example,
2996 in the case of the OpenRISC 1000, this function is called
2997 @code{_initialize_or1k_tdep} and is found in the file
3000 @cindex @file{configure.tgt}
3001 @cindex @code{gdbarch}
3002 @findex gdbarch_register
3003 The resulting object files containing the implementation of the
3004 @code{_initialize_@var{arch}_tdep} function are specified in the @value{GDBN}
3005 @file{configure.tgt} file, which includes a large case statement
3006 pattern matching against the @code{--target} option of the
3007 @code{configure} script. The new @code{struct gdbarch} is created
3008 within the @code{_initialize_@var{arch}_tdep} function by calling
3009 @code{gdbarch_register}:
3012 void gdbarch_register (enum bfd_architecture @var{architecture},
3013 gdbarch_init_ftype *@var{init_func},
3014 gdbarch_dump_tdep_ftype *@var{tdep_dump_func});
3017 The @var{architecture} will identify the unique @sc{bfd} to be
3018 associated with this @code{gdbarch}. The @var{init_func} funciton is
3019 called to create and return the new @code{struct gdbarch}. The
3020 @var{tdep_dump_func} function will dump the target specific details
3021 associated with this architecture.
3023 For example the function @code{_initialize_or1k_tdep} creates its
3024 architecture for 32-bit OpenRISC 1000 architectures by calling:
3027 gdbarch_register (bfd_arch_or32, or1k_gdbarch_init, or1k_dump_tdep);
3030 @node Looking Up an Existing Architecture
3031 @subsection Looking Up an Existing Architecture
3032 @cindex @code{gdbarch} lookup
3034 The initialization function has this prototype:
3037 static struct gdbarch *
3038 @var{arch}_gdbarch_init (struct gdbarch_info @var{info},
3039 struct gdbarch_list *@var{arches})
3042 The @var{info} argument contains parameters used to select the correct
3043 architecture, and @var{arches} is a list of architectures which
3044 have already been created with the same @code{bfd_arch_@var{arch}}
3047 The initialization function should first make sure that @var{info}
3048 is acceptable, and return @code{NULL} if it is not. Then, it should
3049 search through @var{arches} for an exact match to @var{info}, and
3050 return one if found. Lastly, if no exact match was found, it should
3051 create a new architecture based on @var{info} and return it.
3053 @findex gdbarch_list_lookup_by_info
3054 @cindex @code{gdbarch_info}
3055 The lookup is done using @code{gdbarch_list_lookup_by_info}. It is
3056 passed the list of existing architectures, @var{arches}, and the
3057 @code{struct gdbarch_info}, @var{info}, and returns the first matching
3058 architecture it finds, or @code{NULL} if none are found. If an
3059 architecture is found it can be returned as the result from the
3060 initialization function, otherwise a new @code{struct gdbach} will need
3063 The struct gdbarch_info has the following components:
3068 const struct bfd_arch_info *bfd_arch_info;
3071 struct gdbarch_tdep_info *tdep_info;
3072 enum gdb_osabi osabi;
3073 const struct target_desc *target_desc;
3077 @vindex bfd_arch_info
3078 The @code{bfd_arch_info} member holds the key details about the
3079 architecture. The @code{byte_order} member is a value in an
3080 enumeration indicating the endianism. The @code{abfd} member is a
3081 pointer to the full @sc{bfd}, the @code{tdep_info} member is
3082 additional custom target specific information, @code{osabi} identifies
3083 which (if any) of a number of operating specific ABIs are used by this
3084 architecture and the @code{target_desc} member is a set of name-value
3085 pairs with information about register usage in this target.
3087 When the @code{struct gdbarch} initialization function is called, not
3088 all the fields are provided---only those which can be deduced from the
3089 @sc{bfd}. The @code{struct gdbarch_info}, @var{info} is used as a
3090 look-up key with the list of existing architectures, @var{arches} to
3091 see if a suitable architecture already exists. The @var{tdep_info},
3092 @var{osabi} and @var{target_desc} fields may be added before this
3093 lookup to refine the search.
3095 Only information in @var{info} should be used to choose the new
3096 architecture. Historically, @var{info} could be sparse, and
3097 defaults would be collected from the first element on @var{arches}.
3098 However, @value{GDBN} now fills in @var{info} more thoroughly,
3099 so new @code{gdbarch} initialization functions should not take
3100 defaults from @var{arches}.
3102 @node Creating a New Architecture
3103 @subsection Creating a New Architecture
3104 @cindex @code{struct gdbarch} creation
3106 @findex gdbarch_alloc
3107 @cindex @code{gdbarch_tdep} when allocating new @code{gdbarch}
3108 If no architecture is found, then a new architecture must be created,
3109 by calling @code{gdbarch_alloc} using the supplied @code{@w{struct
3110 gdbarch_info}} and any additional custom target specific
3111 information in a @code{struct gdbarch_tdep}. The prototype for
3112 @code{gdbarch_alloc} is:
3115 struct gdbarch *gdbarch_alloc (const struct gdbarch_info *@var{info},
3116 struct gdbarch_tdep *@var{tdep});
3119 @cindex @code{set_gdbarch} functions
3120 @cindex @code{gdbarch} accessor functions
3121 The newly created struct gdbarch must then be populated. Although
3122 there are default values, in most cases they are not what is
3125 For each element, @var{X}, there is are a pair of corresponding accessor
3126 functions, one to set the value of that element,
3127 @code{set_gdbarch_@var{X}}, the second to either get the value of an
3128 element (if it is a variable) or to apply the element (if it is a
3129 function), @code{gdbarch_@var{X}}. Note that both accessor functions
3130 take a pointer to the @code{@w{struct gdbarch}} as first
3131 argument. Populating the new @code{gdbarch} should use the
3132 @code{set_gdbarch} functions.
3134 The following sections identify the main elements that should be set
3135 in this way. This is not the complete list, but represents the
3136 functions and elements that must commonly be specified for a new
3137 architecture. Many of the functions and variables are described in the
3138 header file @file{gdbarch.h}.
3140 This is the main work in defining a new architecture. Implementing the
3141 set of functions to populate the @code{struct gdbarch}.
3143 @cindex @code{gdbarch_tdep} definition
3144 @code{struct gdbarch_tdep} is not defined within @value{GDBN}---it is up
3145 to the user to define this struct if it is needed to hold custom target
3146 information that is not covered by the standard @code{@w{struct
3147 gdbarch}}. For example with the OpenRISC 1000 architecture it is used to
3148 hold the number of matchpoints available in the target (along with other
3151 If there is no additional target specific information, it can be set to
3154 @node Registers and Memory
3155 @section Registers and Memory
3157 @value{GDBN}'s model of the target machine is rather simple.
3158 @value{GDBN} assumes the machine includes a bank of registers and a
3159 block of memory. Each register may have a different size.
3161 @value{GDBN} does not have a magical way to match up with the
3162 compiler's idea of which registers are which; however, it is critical
3163 that they do match up accurately. The only way to make this work is
3164 to get accurate information about the order that the compiler uses,
3165 and to reflect that in the @code{gdbarch_register_name} and related functions.
3167 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
3169 @node Pointers and Addresses
3170 @section Pointers Are Not Always Addresses
3171 @cindex pointer representation
3172 @cindex address representation
3173 @cindex word-addressed machines
3174 @cindex separate data and code address spaces
3175 @cindex spaces, separate data and code address
3176 @cindex address spaces, separate data and code
3177 @cindex code pointers, word-addressed
3178 @cindex converting between pointers and addresses
3179 @cindex D10V addresses
3181 On almost all 32-bit architectures, the representation of a pointer is
3182 indistinguishable from the representation of some fixed-length number
3183 whose value is the byte address of the object pointed to. On such
3184 machines, the words ``pointer'' and ``address'' can be used interchangeably.
3185 However, architectures with smaller word sizes are often cramped for
3186 address space, so they may choose a pointer representation that breaks this
3187 identity, and allows a larger code address space.
3189 @c D10V is gone from sources - more current example?
3191 For example, the Renesas D10V is a 16-bit VLIW processor whose
3192 instructions are 32 bits long@footnote{Some D10V instructions are
3193 actually pairs of 16-bit sub-instructions. However, since you can't
3194 jump into the middle of such a pair, code addresses can only refer to
3195 full 32 bit instructions, which is what matters in this explanation.}.
3196 If the D10V used ordinary byte addresses to refer to code locations,
3197 then the processor would only be able to address 64kb of instructions.
3198 However, since instructions must be aligned on four-byte boundaries, the
3199 low two bits of any valid instruction's byte address are always
3200 zero---byte addresses waste two bits. So instead of byte addresses,
3201 the D10V uses word addresses---byte addresses shifted right two bits---to
3202 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
3205 However, this means that code pointers and data pointers have different
3206 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
3207 @code{0xC020} when used as a data address, but refers to byte address
3208 @code{0x30080} when used as a code address.
3210 (The D10V also uses separate code and data address spaces, which also
3211 affects the correspondence between pointers and addresses, but we're
3212 going to ignore that here; this example is already too long.)
3214 To cope with architectures like this---the D10V is not the only
3215 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
3216 byte numbers, and @dfn{pointers}, which are the target's representation
3217 of an address of a particular type of data. In the example above,
3218 @code{0xC020} is the pointer, which refers to one of the addresses
3219 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
3220 @value{GDBN} provides functions for turning a pointer into an address
3221 and vice versa, in the appropriate way for the current architecture.
3223 Unfortunately, since addresses and pointers are identical on almost all
3224 processors, this distinction tends to bit-rot pretty quickly. Thus,
3225 each time you port @value{GDBN} to an architecture which does
3226 distinguish between pointers and addresses, you'll probably need to
3227 clean up some architecture-independent code.
3229 Here are functions which convert between pointers and addresses:
3231 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
3232 Treat the bytes at @var{buf} as a pointer or reference of type
3233 @var{type}, and return the address it represents, in a manner
3234 appropriate for the current architecture. This yields an address
3235 @value{GDBN} can use to read target memory, disassemble, etc. Note that
3236 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
3239 For example, if the current architecture is the Intel x86, this function
3240 extracts a little-endian integer of the appropriate length from
3241 @var{buf} and returns it. However, if the current architecture is the
3242 D10V, this function will return a 16-bit integer extracted from
3243 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
3245 If @var{type} is not a pointer or reference type, then this function
3246 will signal an internal error.
3249 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
3250 Store the address @var{addr} in @var{buf}, in the proper format for a
3251 pointer of type @var{type} in the current architecture. Note that
3252 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
3255 For example, if the current architecture is the Intel x86, this function
3256 stores @var{addr} unmodified as a little-endian integer of the
3257 appropriate length in @var{buf}. However, if the current architecture
3258 is the D10V, this function divides @var{addr} by four if @var{type} is
3259 a pointer to a function, and then stores it in @var{buf}.
3261 If @var{type} is not a pointer or reference type, then this function
3262 will signal an internal error.
3265 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
3266 Assuming that @var{val} is a pointer, return the address it represents,
3267 as appropriate for the current architecture.
3269 This function actually works on integral values, as well as pointers.
3270 For pointers, it performs architecture-specific conversions as
3271 described above for @code{extract_typed_address}.
3274 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
3275 Create and return a value representing a pointer of type @var{type} to
3276 the address @var{addr}, as appropriate for the current architecture.
3277 This function performs architecture-specific conversions as described
3278 above for @code{store_typed_address}.
3281 Here are two functions which architectures can define to indicate the
3282 relationship between pointers and addresses. These have default
3283 definitions, appropriate for architectures on which all pointers are
3284 simple unsigned byte addresses.
3286 @deftypefun CORE_ADDR gdbarch_pointer_to_address (struct gdbarch *@var{gdbarch}, struct type *@var{type}, char *@var{buf})
3287 Assume that @var{buf} holds a pointer of type @var{type}, in the
3288 appropriate format for the current architecture. Return the byte
3289 address the pointer refers to.
3291 This function may safely assume that @var{type} is either a pointer or a
3292 C@t{++} reference type.
3295 @deftypefun void gdbarch_address_to_pointer (struct gdbarch *@var{gdbarch}, struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
3296 Store in @var{buf} a pointer of type @var{type} representing the address
3297 @var{addr}, in the appropriate format for the current architecture.
3299 This function may safely assume that @var{type} is either a pointer or a
3300 C@t{++} reference type.
3303 @node Address Classes
3304 @section Address Classes
3305 @cindex address classes
3306 @cindex DW_AT_byte_size
3307 @cindex DW_AT_address_class
3309 Sometimes information about different kinds of addresses is available
3310 via the debug information. For example, some programming environments
3311 define addresses of several different sizes. If the debug information
3312 distinguishes these kinds of address classes through either the size
3313 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
3314 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
3315 following macros should be defined in order to disambiguate these
3316 types within @value{GDBN} as well as provide the added information to
3317 a @value{GDBN} user when printing type expressions.
3319 @deftypefun int gdbarch_address_class_type_flags (struct gdbarch *@var{gdbarch}, int @var{byte_size}, int @var{dwarf2_addr_class})
3320 Returns the type flags needed to construct a pointer type whose size
3321 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
3322 This function is normally called from within a symbol reader. See
3323 @file{dwarf2read.c}.
3326 @deftypefun {char *} gdbarch_address_class_type_flags_to_name (struct gdbarch *@var{gdbarch}, int @var{type_flags})
3327 Given the type flags representing an address class qualifier, return
3330 @deftypefun int gdbarch_address_class_name_to_type_flags (struct gdbarch *@var{gdbarch}, int @var{name}, int *@var{type_flags_ptr})
3331 Given an address qualifier name, set the @code{int} referenced by @var{type_flags_ptr} to the type flags
3332 for that address class qualifier.
3335 Since the need for address classes is rather rare, none of
3336 the address class functions are defined by default. Predicate
3337 functions are provided to detect when they are defined.
3339 Consider a hypothetical architecture in which addresses are normally
3340 32-bits wide, but 16-bit addresses are also supported. Furthermore,
3341 suppose that the @w{DWARF 2} information for this architecture simply
3342 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
3343 of these "short" pointers. The following functions could be defined
3344 to implement the address class functions:
3347 somearch_address_class_type_flags (int byte_size,
3348 int dwarf2_addr_class)
3351 return TYPE_FLAG_ADDRESS_CLASS_1;
3357 somearch_address_class_type_flags_to_name (int type_flags)
3359 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
3366 somearch_address_class_name_to_type_flags (char *name,
3367 int *type_flags_ptr)
3369 if (strcmp (name, "short") == 0)
3371 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
3379 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
3380 to indicate the presence of one of these ``short'' pointers. For
3381 example if the debug information indicates that @code{short_ptr_var} is
3382 one of these short pointers, @value{GDBN} might show the following
3386 (gdb) ptype short_ptr_var
3387 type = int * @@short
3391 @node Register Representation
3392 @section Register Representation
3395 * Raw and Cooked Registers::
3396 * Register Architecture Functions & Variables::
3397 * Register Information Functions::
3398 * Register and Memory Data::
3399 * Register Caching::
3402 @node Raw and Cooked Registers
3403 @subsection Raw and Cooked Registers
3404 @cindex raw register representation
3405 @cindex cooked register representation
3406 @cindex representations, raw and cooked registers
3408 @value{GDBN} considers registers to be a set with members numbered
3409 linearly from 0 upwards. The first part of that set corresponds to real
3410 physical registers, the second part to any @dfn{pseudo-registers}.
3411 Pseudo-registers have no independent physical existence, but are useful
3412 representations of information within the architecture. For example the
3413 OpenRISC 1000 architecture has up to 32 general purpose registers, which
3414 are typically represented as 32-bit (or 64-bit) integers. However the
3415 GPRs are also used as operands to the floating point operations, and it
3416 could be convenient to define a set of pseudo-registers, to show the
3417 GPRs represented as floating point values.
3419 For any architecture, the implementer will decide on a mapping from
3420 hardware to @value{GDBN} register numbers. The registers corresponding to real
3421 hardware are referred to as @dfn{raw} registers, the remaining registers are
3422 @dfn{pseudo-registers}. The total register set (raw and pseudo) is called
3423 the @dfn{cooked} register set.
3426 @node Register Architecture Functions & Variables
3427 @subsection Functions and Variables Specifying the Register Architecture
3428 @cindex @code{gdbarch} register architecture functions
3430 These @code{struct gdbarch} functions and variables specify the number
3431 and type of registers in the architecture.
3433 @deftypefn {Architecture Function} CORE_ADDR read_pc (struct regcache *@var{regcache})
3435 @deftypefn {Architecture Function} void write_pc (struct regcache *@var{regcache}, CORE_ADDR @var{val})
3437 Read or write the program counter. The default value of both
3438 functions is @code{NULL} (no function available). If the program
3439 counter is just an ordinary register, it can be specified in
3440 @code{struct gdbarch} instead (see @code{pc_regnum} below) and it will
3441 be read or written using the standard routines to access registers. This
3442 function need only be specified if the program counter is not an
3445 Any register information can be obtained using the supplied register
3446 cache, @var{regcache}. @xref{Register Caching, , Register Caching}.
3450 @deftypefn {Architecture Function} void pseudo_register_read (struct gdbarch *@var{gdbarch}, struct regcache *@var{regcache}, int @var{regnum}, const gdb_byte *@var{buf})
3452 @deftypefn {Architecture Function} void pseudo_register_write (struct gdbarch *@var{gdbarch}, struct regcache *@var{regcache}, int @var{regnum}, const gdb_byte *@var{buf})
3454 These functions should be defined if there are any pseudo-registers.
3455 The default value is @code{NULL}. @var{regnum} is the number of the
3456 register to read or write (which will be a @dfn{cooked} register
3457 number) and @var{buf} is the buffer where the value read will be
3458 placed, or from which the value to be written will be taken. The
3459 value in the buffer may be converted to or from a signed or unsigned
3460 integral value using one of the utility functions (@pxref{Register and
3461 Memory Data, , Using Different Register and Memory Data
3464 The access should be for the specified architecture,
3465 @var{gdbarch}. Any register information can be obtained using the
3466 supplied register cache, @var{regcache}. @xref{Register Caching, ,
3471 @deftypevr {Architecture Variable} int sp_regnum
3473 @cindex stack pointer
3476 This specifies the register holding the stack pointer, which may be a
3477 raw or pseudo-register. It defaults to -1 (not defined), but it is an
3478 error for it not to be defined.
3480 The value of the stack pointer register can be accessed withing
3481 @value{GDBN} as the variable @kbd{$sp}.
3485 @deftypevr {Architecture Variable} int pc_regnum
3487 @cindex program counter
3490 This specifies the register holding the program counter, which may be a
3491 raw or pseudo-register. It defaults to -1 (not defined). If
3492 @code{pc_regnum} is not defined, then the functions @code{read_pc} and
3493 @code{write_pc} (see above) must be defined.
3495 The value of the program counter (whether defined as a register, or
3496 through @code{read_pc} and @code{write_pc}) can be accessed withing
3497 @value{GDBN} as the variable @kbd{$pc}.
3501 @deftypevr {Architecture Variable} int ps_regnum
3503 @cindex processor status register
3504 @cindex status register
3507 This specifies the register holding the processor status (often called
3508 the status register), which may be a raw or pseudo-register. It
3509 defaults to -1 (not defined).
3511 If defined, the value of this register can be accessed withing
3512 @value{GDBN} as the variable @kbd{$ps}.
3516 @deftypevr {Architecture Variable} int fp0_regnum
3518 @cindex first floating point register
3520 This specifies the first floating point register. It defaults to
3521 0. @code{fp0_regnum} is not needed unless the target offers support
3526 @node Register Information Functions
3527 @subsection Functions Giving Register Information
3528 @cindex @code{gdbarch} register information functions
3530 These functions return information about registers.
3532 @deftypefn {Architecture Function} {const char *} register_name (struct gdbarch *@var{gdbarch}, int @var{regnum})
3534 This function should convert a register number (raw or pseudo) to a
3535 register name (as a C @code{const char *}). This is used both to
3536 determine the name of a register for output and to work out the meaning
3537 of any register names used as input. The function may also return
3538 @code{NULL}, to indicate that @var{regnum} is not a valid register.
3540 For example with the OpenRISC 1000, @value{GDBN} registers 0-31 are the
3541 General Purpose Registers, register 32 is the program counter and
3542 register 33 is the supervision register (i.e.@: the processor status
3543 register), which map to the strings @code{"gpr00"} through
3544 @code{"gpr31"}, @code{"pc"} and @code{"sr"} respectively. This means
3545 that the @value{GDBN} command @kbd{print $gpr5} should print the value of
3546 the OR1K general purpose register 5@footnote{
3547 @cindex frame pointer
3549 Historically, @value{GDBN} always had a concept of a frame pointer
3550 register, which could be accessed via the @value{GDBN} variable,
3551 @kbd{$fp}. That concept is now deprecated, recognizing that not all
3552 architectures have a frame pointer. However if an architecture does
3553 have a frame pointer register, and defines a register or
3554 pseudo-register with the name @code{"fp"}, then that register will be
3555 used as the value of the @kbd{$fp} variable.}.
3557 The default value for this function is @code{NULL}, meaning
3558 undefined. It should always be defined.
3560 The access should be for the specified architecture, @var{gdbarch}.
3564 @deftypefn {Architecture Function} {struct type *} register_type (struct gdbarch *@var{gdbarch}, int @var{regnum})
3566 Given a register number, this function identifies the type of data it
3567 may be holding, specified as a @code{struct type}. @value{GDBN} allows
3568 creation of arbitrary types, but a number of built in types are
3569 provided (@code{builtin_type_void}, @code{builtin_type_int32} etc),
3570 together with functions to derive types from these.
3572 Typically the program counter will have a type of ``pointer to
3573 function'' (it points to code), the frame pointer and stack pointer
3574 will have types of ``pointer to void'' (they point to data on the stack)
3575 and all other integer registers will have a type of 32-bit integer or
3578 This information guides the formatting when displaying register
3579 information. The default value is @code{NULL} meaning no information is
3580 available to guide formatting when displaying registers.
3584 @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})
3586 Define this function to print out one or all of the registers for the
3587 @value{GDBN} @kbd{info registers} command. The default value is the
3588 function @code{default_print_registers_info}, which uses the register
3589 type information (see @code{register_type} above) to determine how each
3590 register should be printed. Define a custom version of this function
3591 for fuller control over how the registers are displayed.
3593 The access should be for the specified architecture, @var{gdbarch},
3594 with output to the file specified by the User Interface
3595 Independent Output file handle, @var{file} (@pxref{UI-Independent
3596 Output, , UI-Independent Output---the @code{ui_out}
3599 The registers should show their values in the frame specified by
3600 @var{frame}. If @var{regnum} is -1 and @var{all} is zero, then all
3601 the ``significant'' registers should be shown (the implementer should
3602 decide which registers are ``significant''). Otherwise only the value of
3603 the register specified by @var{regnum} should be output. If
3604 @var{regnum} is -1 and @var{all} is non-zero (true), then the value of
3605 all registers should be shown.
3607 By default @code{default_print_registers_info} prints one register per
3608 line, and if @var{all} is zero omits floating-point registers.
3612 @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})
3614 Define this function to provide output about the floating point unit and
3615 registers for the @value{GDBN} @kbd{info float} command respectively.
3616 The default value is @code{NULL} (not defined), meaning no information
3619 The @var{gdbarch} and @var{file} and @var{frame} arguments have the same
3620 meaning as in the @code{print_registers_info} function above. The string
3621 @var{args} contains any supplementary arguments to the @kbd{info float}
3624 Define this function if the target supports floating point operations.
3628 @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})
3630 Define this function to provide output about the vector unit and
3631 registers for the @value{GDBN} @kbd{info vector} command respectively.
3632 The default value is @code{NULL} (not defined), meaning no information
3635 The @var{gdbarch}, @var{file} and @var{frame} arguments have the
3636 same meaning as in the @code{print_registers_info} function above. The
3637 string @var{args} contains any supplementary arguments to the @kbd{info
3640 Define this function if the target supports vector operations.
3644 @deftypefn {Architecture Function} int register_reggroup_p (struct gdbarch *@var{gdbarch}, int @var{regnum}, struct reggroup *@var{group})
3646 @value{GDBN} groups registers into different categories (general,
3647 vector, floating point etc). This function, given a register,
3648 @var{regnum}, and group, @var{group}, returns 1 (true) if the register
3649 is in the group and 0 (false) otherwise.
3651 The information should be for the specified architecture,
3654 The default value is the function @code{default_register_reggroup_p}
3655 which will do a reasonable job based on the type of the register (see
3656 the function @code{register_type} above), with groups for general
3657 purpose registers, floating point registers, vector registers and raw
3658 (i.e not pseudo) registers.
3662 @node Register and Memory Data
3663 @subsection Using Different Register and Memory Data Representations
3664 @cindex register representation
3665 @cindex memory representation
3666 @cindex representations, register and memory
3667 @cindex register data formats, converting
3668 @cindex @code{struct value}, converting register contents to
3670 Some architectures have different representations of data objects,
3671 depending whether the object is held in a register or memory. For
3677 The Alpha architecture can represent 32 bit integer values in
3678 floating-point registers.
3681 The x86 architecture supports 80-bit floating-point registers. The
3682 @code{long double} data type occupies 96 bits in memory but only 80
3683 bits when stored in a register.
3687 In general, the register representation of a data type is determined by
3688 the architecture, or @value{GDBN}'s interface to the architecture, while
3689 the memory representation is determined by the Application Binary
3692 For almost all data types on almost all architectures, the two
3693 representations are identical, and no special handling is needed.
3694 However, they do occasionally differ. An architecture may define the
3695 following @code{struct gdbarch} functions to request conversions
3696 between the register and memory representations of a data type:
3698 @deftypefn {Architecture Function} int gdbarch_convert_register_p (struct gdbarch *@var{gdbarch}, int @var{reg})
3700 Return non-zero (true) if the representation of a data value stored in
3701 this register may be different to the representation of that same data
3702 value when stored in memory. The default value is @code{NULL}
3705 If this function is defined and returns non-zero, the @code{struct
3706 gdbarch} functions @code{gdbarch_register_to_value} and
3707 @code{gdbarch_value_to_register} (see below) should be used to perform
3708 any necessary conversion.
3710 If defined, this function should return zero for the register's native
3711 type, when no conversion is necessary.
3714 @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})
3716 Convert the value of register number @var{reg} to a data object of
3717 type @var{type}. The buffer at @var{from} holds the register's value
3718 in raw format; the converted value should be placed in the buffer at
3722 @emph{Note:} @code{gdbarch_register_to_value} and
3723 @code{gdbarch_value_to_register} take their @var{reg} and @var{type}
3724 arguments in different orders.
3727 @code{gdbarch_register_to_value} should only be used with registers
3728 for which the @code{gdbarch_convert_register_p} function returns a
3733 @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})
3735 Convert a data value of type @var{type} to register number @var{reg}'
3739 @emph{Note:} @code{gdbarch_register_to_value} and
3740 @code{gdbarch_value_to_register} take their @var{reg} and @var{type}
3741 arguments in different orders.
3744 @code{gdbarch_value_to_register} should only be used with registers
3745 for which the @code{gdbarch_convert_register_p} function returns a
3750 @node Register Caching
3751 @subsection Register Caching
3752 @cindex register caching
3754 Caching of registers is used, so that the target does not need to be
3755 accessed and reanalyzed multiple times for each register in
3756 circumstances where the register value cannot have changed.
3758 @cindex @code{struct regcache}
3759 @value{GDBN} provides @code{struct regcache}, associated with a
3760 particular @code{struct gdbarch} to hold the cached values of the raw
3761 registers. A set of functions is provided to access both the raw
3762 registers (with @code{raw} in their name) and the full set of cooked
3763 registers (with @code{cooked} in their name). Functions are provided
3764 to ensure the register cache is kept synchronized with the values of
3765 the actual registers in the target.
3767 Accessing registers through the @code{struct regcache} routines will
3768 ensure that the appropriate @code{struct gdbarch} functions are called
3769 when necessary to access the underlying target architecture. In general
3770 users should use the @dfn{cooked} functions, since these will map to the
3771 @dfn{raw} functions automatically as appropriate.
3773 @findex regcache_cooked_read
3774 @findex regcache_cooked_write
3775 @cindex @code{gdb_byte}
3776 @findex regcache_cooked_read_signed
3777 @findex regcache_cooked_read_unsigned
3778 @findex regcache_cooked_write_signed
3779 @findex regcache_cooked_write_unsigned
3780 The two key functions are @code{regcache_cooked_read} and
3781 @code{regcache_cooked_write} which read or write a register from or to
3782 a byte buffer (type @code{gdb_byte *}). For convenience the wrapper
3783 functions @code{regcache_cooked_read_signed},
3784 @code{regcache_cooked_read_unsigned},
3785 @code{regcache_cooked_write_signed} and
3786 @code{regcache_cooked_write_unsigned} are provided, which read or
3787 write the value using the buffer and convert to or from an integral
3788 value as appropriate.
3790 @node Frame Interpretation
3791 @section Frame Interpretation
3794 * All About Stack Frames::
3795 * Frame Handling Terminology::
3797 * Functions and Variable to Analyze Frames::
3798 * Functions to Access Frame Data::
3799 * Analyzing Stacks---Frame Sniffers::
3802 @node All About Stack Frames
3803 @subsection All About Stack Frames
3805 @value{GDBN} needs to understand the stack on which local (automatic)
3806 variables are stored. The area of the stack containing all the local
3807 variables for a function invocation is known as the @dfn{stack frame}
3808 for that function (or colloquially just as the @dfn{frame}). In turn the
3809 function that called the function will have its stack frame, and so on
3810 back through the chain of functions that have been called.
3812 Almost all architectures have one register dedicated to point to the
3813 end of the stack (the @dfn{stack pointer}). Many have a second register
3814 which points to the start of the currently active stack frame (the
3815 @dfn{frame pointer}). The specific arrangements for an architecture are
3816 a key part of the ABI.
3818 A diagram helps to explain this. Here is a simple program to compute
3831 return n * fact (n - 1);
3839 for (i = 0; i < 10; i++)
3842 printf ("%d! = %d\n", i, f);
3847 Consider the state of the stack when the code reaches line 6 after the
3848 main program has called @code{fact@w{ }(3)}. The chain of function
3849 calls will be @code{main ()}, @code{fact@w{ }(3)}, @code{fact@w{
3850 }(2)}, @code{@w{fact (1)}} and @code{fact@w{ }(0)}.
3852 In this illustration the stack is falling (as used for example by the
3853 OpenRISC 1000 ABI). The stack pointer (SP) is at the end of the stack
3854 (lowest address) and the frame pointer (FP) is at the highest address
3855 in the current stack frame. The following diagram shows how the stack
3858 @center @image{stack_frame,14cm}
3860 In each stack frame, offset 0 from the stack pointer is the frame
3861 pointer of the previous frame and offset 4 (this is illustrating a
3862 32-bit architecture) from the stack pointer is the return address.
3863 Local variables are indexed from the frame pointer, with negative
3864 indexes. In the function @code{fact}, offset -4 from the frame
3865 pointer is the argument @var{n}. In the @code{main} function, offset
3866 -4 from the frame pointer is the local variable @var{i} and offset -8
3867 from the frame pointer is the local variable @var{f}@footnote{This is
3868 a simplified example for illustrative purposes only. Good optimizing
3869 compilers would not put anything on the stack for such simple
3870 functions. Indeed they might eliminate the recursion and use of the
3873 It is very easy to get confused when examining stacks. @value{GDBN}
3874 has terminology it uses rigorously throughout. The stack frame of the
3875 function currently executing, or where execution stopped is numbered
3876 zero. In this example frame #0 is the stack frame of the call to
3877 @code{fact@w{ }(0)}. The stack frame of its calling function
3878 (@code{fact@w{ }(1)} in this case) is numbered #1 and so on back
3879 through the chain of calls.
3881 The main @value{GDBN} data structure describing frames is
3882 @code{@w{struct frame_info}}. It is not used directly, but only via
3883 its accessor functions. @code{frame_info} includes information about
3884 the registers in the frame and a pointer to the code of the function
3885 with which the frame is associated. The entire stack is represented as
3886 a linked list of @code{frame_info} structs.
3888 @node Frame Handling Terminology
3889 @subsection Frame Handling Terminology
3891 It is easy to get confused when referencing stack frames. @value{GDBN}
3892 uses some precise terminology.
3898 @cindex stack frame, definition of THIS frame
3899 @cindex frame, definition of THIS frame
3900 @dfn{THIS} frame is the frame currently under consideration.
3904 @cindex stack frame, definition of NEXT frame
3905 @cindex frame, definition of NEXT frame
3906 The @dfn{NEXT} frame, also sometimes called the inner or newer frame is the
3907 frame of the function called by the function of THIS frame.
3910 @cindex PREVIOUS frame
3911 @cindex stack frame, definition of PREVIOUS frame
3912 @cindex frame, definition of PREVIOUS frame
3913 The @dfn{PREVIOUS} frame, also sometimes called the outer or older frame is
3914 the frame of the function which called the function of THIS frame.
3918 So in the example in the previous section (@pxref{All About Stack
3919 Frames, , All About Stack Frames}), if THIS frame is #3 (the call to
3920 @code{fact@w{ }(3)}), the NEXT frame is frame #2 (the call to
3921 @code{fact@w{ }(2)}) and the PREVIOUS frame is frame #4 (the call to
3922 @code{main@w{ }()}).
3924 @cindex innermost frame
3925 @cindex stack frame, definition of innermost frame
3926 @cindex frame, definition of innermost frame
3927 The @dfn{innermost} frame is the frame of the current executing
3928 function, or where the program stopped, in this example, in the middle
3929 of the call to @code{@w{fact (0))}}. It is always numbered frame #0.
3931 @cindex base of a frame
3932 @cindex stack frame, definition of base of a frame
3933 @cindex frame, definition of base of a frame
3934 The @dfn{base} of a frame is the address immediately before the start
3935 of the NEXT frame. For a stack which grows down in memory (a
3936 @dfn{falling} stack) this will be the lowest address and for a stack
3937 which grows up in memory (a @dfn{rising} stack) this will be the
3938 highest address in the frame.
3940 @value{GDBN} functions to analyze the stack are typically given a
3941 pointer to the NEXT frame to determine information about THIS
3942 frame. Information about THIS frame includes data on where the
3943 registers of the PREVIOUS frame are stored in this stack frame. In
3944 this example the frame pointer of the PREVIOUS frame is stored at
3945 offset 0 from the stack pointer of THIS frame.
3948 @cindex stack frame, definition of unwinding
3949 @cindex frame, definition of unwinding
3950 The process whereby a function is given a pointer to the NEXT
3951 frame to work out information about THIS frame is referred to as
3952 @dfn{unwinding}. The @value{GDBN} functions involved in this typically
3953 include unwind in their name.
3956 @cindex stack frame, definition of sniffing
3957 @cindex frame, definition of sniffing
3958 The process of analyzing a target to determine the information that
3959 should go in struct frame_info is called @dfn{sniffing}. The functions
3960 that carry this out are called sniffers and typically include sniffer
3961 in their name. More than one sniffer may be required to extract all
3962 the information for a particular frame.
3964 @cindex sentinel frame
3965 @cindex stack frame, definition of sentinel frame
3966 @cindex frame, definition of sentinel frame
3967 Because so many functions work using the NEXT frame, there is an issue
3968 about addressing the innermost frame---it has no NEXT frame. To solve
3969 this @value{GDBN} creates a dummy frame #-1, known as the
3970 @dfn{sentinel} frame.
3972 @node Prologue Caches
3973 @subsection Prologue Caches
3975 @cindex function prologue
3976 @cindex prologue of a function
3977 All the frame sniffing functions typically examine the code at the
3978 start of the corresponding function, to determine the state of
3979 registers. The ABI will save old values and set new values of key
3980 registers at the start of each function in what is known as the
3981 function @dfn{prologue}.
3983 @cindex prologue cache
3984 For any particular stack frame this data does not change, so all the
3985 standard unwinding functions, in addition to receiving a pointer to
3986 the NEXT frame as their first argument, receive a pointer to a
3987 @dfn{prologue cache} as their second argument. This can be used to store
3988 values associated with a particular frame, for reuse on subsequent
3989 calls involving the same frame.
3991 It is up to the user to define the structure used (it is a
3992 @code{void@w{ }*} pointer) and arrange allocation and deallocation of
3993 storage. However for general use, @value{GDBN} provides
3994 @code{@w{struct trad_frame_cache}}, with a set of accessor
3995 routines. This structure holds the stack and code address of
3996 THIS frame, the base address of the frame, a pointer to the
3997 struct @code{frame_info} for the NEXT frame and details of
3998 where the registers of the PREVIOUS frame may be found in THIS
4001 Typically the first time any sniffer function is called with NEXT
4002 frame, the prologue sniffer for THIS frame will be @code{NULL}. The
4003 sniffer will analyze the frame, allocate a prologue cache structure
4004 and populate it. Subsequent calls using the same NEXT frame will
4005 pass in this prologue cache, so the data can be returned with no
4006 additional analysis.
4008 @node Functions and Variable to Analyze Frames
4009 @subsection Functions and Variable to Analyze Frames
4011 These struct @code{gdbarch} functions and variable should be defined
4012 to provide analysis of the stack frame and allow it to be adjusted as
4015 @deftypefn {Architecture Function} CORE_ADDR skip_prologue (struct gdbarch *@var{gdbarch}, CORE_ADDR @var{pc})
4017 The prologue of a function is the code at the beginning of the
4018 function which sets up the stack frame, saves the return address
4019 etc. The code representing the behavior of the function starts after
4022 This function skips past the prologue of a function if the program
4023 counter, @var{pc}, is within the prologue of a function. The result is
4024 the program counter immediately after the prologue. With modern
4025 optimizing compilers, this may be a far from trivial exercise. However
4026 the required information may be within the binary as DWARF2 debugging
4027 information, making the job much easier.
4029 The default value is @code{NULL} (not defined). This function should always
4030 be provided, but can take advantage of DWARF2 debugging information,
4031 if that is available.
4035 @deftypefn {Architecture Function} int inner_than (CORE_ADDR @var{lhs}, CORE_ADDR @var{rhs})
4036 @findex core_addr_lessthan
4037 @findex core_addr_greaterthan
4039 Given two frame or stack pointers, return non-zero (true) if the first
4040 represents the @dfn{inner} stack frame and 0 (false) otherwise. This
4041 is used to determine whether the target has a stack which grows up in
4042 memory (rising stack) or grows down in memory (falling stack).
4043 @xref{All About Stack Frames, , All About Stack Frames}, for an
4044 explanation of @dfn{inner} frames.
4046 The default value of this function is @code{NULL} and it should always
4047 be defined. However for almost all architectures one of the built-in
4048 functions can be used: @code{core_addr_lessthan} (for stacks growing
4049 down in memory) or @code{core_addr_greaterthan} (for stacks growing up
4054 @anchor{frame_align}
4055 @deftypefn {Architecture Function} CORE_ADDR frame_align (struct gdbarch *@var{gdbarch}, CORE_ADDR @var{address})
4059 The architecture may have constraints on how its frames are
4060 aligned. For example the OpenRISC 1000 ABI requires stack frames to be
4061 double-word aligned, but 32-bit versions of the architecture allocate
4062 single-word values to the stack. Thus extra padding may be needed at
4063 the end of a stack frame.
4065 Given a proposed address for the stack pointer, this function
4066 returns a suitably aligned address (by expanding the stack frame).
4068 The default value is @code{NULL} (undefined). This function should be defined
4069 for any architecture where it is possible the stack could become
4070 misaligned. The utility functions @code{align_down} (for falling
4071 stacks) and @code{align_up} (for rising stacks) will facilitate the
4072 implementation of this function.
4076 @deftypevr {Architecture Variable} int frame_red_zone_size
4078 Some ABIs reserve space beyond the end of the stack for use by leaf
4079 functions without prologue or epilogue or by exception handlers (for
4080 example the OpenRISC 1000).
4082 This is known as a @dfn{red zone} (AMD terminology). The @sc{amd64}
4083 (nee x86-64) ABI documentation refers to the @dfn{red zone} when
4084 describing this scratch area.
4086 The default value is 0. Set this field if the architecture has such a
4087 red zone. The value must be aligned as required by the ABI (see
4088 @code{frame_align} above for an explanation of stack frame alignment).
4092 @node Functions to Access Frame Data
4093 @subsection Functions to Access Frame Data
4095 These functions provide access to key registers and arguments in the
4098 @deftypefn {Architecture Function} CORE_ADDR unwind_pc (struct gdbarch *@var{gdbarch}, struct frame_info *@var{next_frame})
4100 This function is given a pointer to the NEXT stack frame (@pxref{All
4101 About Stack Frames, , All About Stack Frames}, for how frames are
4102 represented) and returns the value of the program counter in the
4103 PREVIOUS frame (i.e.@: the frame of the function that called THIS
4104 one). This is commonly referred to as the @dfn{return address}.
4106 The implementation, which must be frame agnostic (work with any frame),
4107 is typically no more than:
4111 pc = frame_unwind_register_unsigned (next_frame, @var{ARCH}_PC_REGNUM);
4112 return gdbarch_addr_bits_remove (gdbarch, pc);
4117 @deftypefn {Architecture Function} CORE_ADDR unwind_sp (struct gdbarch *@var{gdbarch}, struct frame_info *@var{next_frame})
4119 This function is given a pointer to the NEXT stack frame
4120 (@pxref{All About Stack Frames, , All About Stack Frames} for how
4121 frames are represented) and returns the value of the stack pointer in
4122 the PREVIOUS frame (i.e.@: the frame of the function that called
4125 The implementation, which must be frame agnostic (work with any frame),
4126 is typically no more than:
4130 sp = frame_unwind_register_unsigned (next_frame, @var{ARCH}_SP_REGNUM);
4131 return gdbarch_addr_bits_remove (gdbarch, sp);
4136 @deftypefn {Architecture Function} int frame_num_args (struct gdbarch *@var{gdbarch}, struct frame_info *@var{this_frame})
4138 This function is given a pointer to THIS stack frame (@pxref{All
4139 About Stack Frames, , All About Stack Frames} for how frames are
4140 represented), and returns the number of arguments that are being
4141 passed, or -1 if not known.
4143 The default value is @code{NULL} (undefined), in which case the number of
4144 arguments passed on any stack frame is always unknown. For many
4145 architectures this will be a suitable default.
4149 @node Analyzing Stacks---Frame Sniffers
4150 @subsection Analyzing Stacks---Frame Sniffers
4152 When a program stops, @value{GDBN} needs to construct the chain of
4153 struct @code{frame_info} representing the state of the stack using
4154 appropriate @dfn{sniffers}.
4156 Each architecture requires appropriate sniffers, but they do not form
4157 entries in @code{@w{struct gdbarch}}, since more than one sniffer may
4158 be required and a sniffer may be suitable for more than one
4159 @code{@w{struct gdbarch}}. Instead sniffers are associated with
4160 architectures using the following functions.
4165 @findex frame_unwind_append_sniffer
4166 @code{frame_unwind_append_sniffer} is used to add a new sniffer to
4167 analyze THIS frame when given a pointer to the NEXT frame.
4170 @findex frame_base_append_sniffer
4171 @code{frame_base_append_sniffer} is used to add a new sniffer
4172 which can determine information about the base of a stack frame.
4175 @findex frame_base_set_default
4176 @code{frame_base_set_default} is used to specify the default base
4181 These functions all take a reference to @code{@w{struct gdbarch}}, so
4182 they are associated with a specific architecture. They are usually
4183 called in the @code{gdbarch} initialization function, after the
4184 @code{gdbarch} struct has been set up. Unless a default has been set, the
4185 most recently appended sniffer will be tried first.
4187 The main frame unwinding sniffer (as set by
4188 @code{frame_unwind_append_sniffer)} returns a structure specifying
4189 a set of sniffing functions:
4191 @cindex @code{frame_unwind}
4195 enum frame_type type;
4196 frame_this_id_ftype *this_id;
4197 frame_prev_register_ftype *prev_register;
4198 const struct frame_data *unwind_data;
4199 frame_sniffer_ftype *sniffer;
4200 frame_prev_pc_ftype *prev_pc;
4201 frame_dealloc_cache_ftype *dealloc_cache;
4205 The @code{type} field indicates the type of frame this sniffer can
4206 handle: normal, dummy (@pxref{Functions Creating Dummy Frames, ,
4207 Functions Creating Dummy Frames}), signal handler or sentinel. Signal
4208 handlers sometimes have their own simplified stack structure for
4209 efficiency, so may need their own handlers.
4211 The @code{unwind_data} field holds additional information which may be
4212 relevant to particular types of frame. For example it may hold
4213 additional information for signal handler frames.
4215 The remaining fields define functions that yield different types of
4216 information when given a pointer to the NEXT stack frame. Not all
4217 functions need be provided. If an entry is @code{NULL}, the next sniffer will
4223 @code{this_id} determines the stack pointer and function (code
4224 entry point) for THIS stack frame.
4227 @code{prev_register} determines where the values of registers for
4228 the PREVIOUS stack frame are stored in THIS stack frame.
4231 @code{sniffer} takes a look at THIS frame's registers to
4232 determine if this is the appropriate unwinder.
4235 @code{prev_pc} determines the program counter for THIS
4236 frame. Only needed if the program counter is not an ordinary register
4237 (@pxref{Register Architecture Functions & Variables,
4238 , Functions and Variables Specifying the Register Architecture}).
4241 @code{dealloc_cache} frees any additional memory associated with
4242 the prologue cache for this frame (@pxref{Prologue Caches, , Prologue
4247 In general it is only the @code{this_id} and @code{prev_register}
4248 fields that need be defined for custom sniffers.
4250 The frame base sniffer is much simpler. It is a @code{@w{struct
4251 frame_base}}, which refers to the corresponding @code{frame_unwind}
4252 struct and whose fields refer to functions yielding various addresses
4255 @cindex @code{frame_base}
4259 const struct frame_unwind *unwind;
4260 frame_this_base_ftype *this_base;
4261 frame_this_locals_ftype *this_locals;
4262 frame_this_args_ftype *this_args;
4266 All the functions referred to take a pointer to the NEXT frame as
4267 argument. The function referred to by @code{this_base} returns the
4268 base address of THIS frame, the function referred to by
4269 @code{this_locals} returns the base address of local variables in THIS
4270 frame and the function referred to by @code{this_args} returns the
4271 base address of the function arguments in this frame.
4273 As described above, the base address of a frame is the address
4274 immediately before the start of the NEXT frame. For a falling
4275 stack, this is the lowest address in the frame and for a rising stack
4276 it is the highest address in the frame. For most architectures the
4277 same address is also the base address for local variables and
4278 arguments, in which case the same function can be used for all three
4279 entries@footnote{It is worth noting that if it cannot be determined in any
4280 other way (for example by there being a register with the name
4281 @code{"fp"}), then the result of the @code{this_base} function will be
4282 used as the value of the frame pointer variable @kbd{$fp} in
4283 @value{GDBN}. This is very often not correct (for example with the
4284 OpenRISC 1000, this value is the stack pointer, @kbd{$sp}). In this
4285 case a register (raw or pseudo) with the name @code{"fp"} should be
4286 defined. It will be used in preference as the value of @kbd{$fp}.}.
4288 @node Inferior Call Setup
4289 @section Inferior Call Setup
4290 @cindex calls to the inferior
4293 * About Dummy Frames::
4294 * Functions Creating Dummy Frames::
4297 @node About Dummy Frames
4298 @subsection About Dummy Frames
4299 @cindex dummy frames
4301 @value{GDBN} can call functions in the target code (for example by
4302 using the @kbd{call} or @kbd{print} commands). These functions may be
4303 breakpointed, and it is essential that if a function does hit a
4304 breakpoint, commands like @kbd{backtrace} work correctly.
4306 This is achieved by making the stack look as though the function had
4307 been called from the point where @value{GDBN} had previously stopped.
4308 This requires that @value{GDBN} can set up stack frames appropriate for
4309 such function calls.
4311 @node Functions Creating Dummy Frames
4312 @subsection Functions Creating Dummy Frames
4314 The following functions provide the functionality to set up such
4315 @dfn{dummy} stack frames.
4317 @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})
4319 This function sets up a dummy stack frame for the function about to be
4320 called. @code{push_dummy_call} is given the arguments to be passed
4321 and must copy them into registers or push them on to the stack as
4322 appropriate for the ABI.
4324 @var{function} is a pointer to the function
4325 that will be called and @var{regcache} the register cache from which
4326 values should be obtained. @var{bp_addr} is the address to which the
4327 function should return (which is breakpointed, so @value{GDBN} can
4328 regain control, hence the name). @var{nargs} is the number of
4329 arguments to pass and @var{args} an array containing the argument
4330 values. @var{struct_return} is non-zero (true) if the function returns
4331 a structure, and if so @var{struct_addr} is the address in which the
4332 structure should be returned.
4334 After calling this function, @value{GDBN} will pass control to the
4335 target at the address of the function, which will find the stack and
4336 registers set up just as expected.
4338 The default value of this function is @code{NULL} (undefined). If the
4339 function is not defined, then @value{GDBN} will not allow the user to
4340 call functions within the target being debugged.
4344 @deftypefn {Architecture Function} {struct frame_id} unwind_dummy_id (struct gdbarch *@var{gdbarch}, struct frame_info *@var{next_frame})
4346 This is the inverse of @code{push_dummy_call} which restores the stack
4347 pointer and program counter after a call to evaluate a function using
4348 a dummy stack frame. The result is a @code{@w{struct frame_id}}, which
4349 contains the value of the stack pointer and program counter to be
4352 The NEXT frame pointer is provided as argument,
4353 @var{next_frame}. THIS frame is the frame of the dummy function,
4354 which can be unwound, to yield the required stack pointer and program
4355 counter from the PREVIOUS frame.
4357 The default value is @code{NULL} (undefined). If @code{push_dummy_call} is
4358 defined, then this function should also be defined.
4362 @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})
4364 If this function is not defined (its default value is @code{NULL}), a dummy
4365 call will use the entry point of the currently loaded code on the
4366 target as its return address. A temporary breakpoint will be set
4367 there, so the location must be writable and have room for a
4370 It is possible that this default is not suitable. It might not be
4371 writable (in ROM possibly), or the ABI might require code to be
4372 executed on return from a call to unwind the stack before the
4373 breakpoint is encountered.
4375 If either of these is the case, then push_dummy_code should be defined
4376 to push an instruction sequence onto the end of the stack to which the
4377 dummy call should return.
4379 The arguments are essentially the same as those to
4380 @code{push_dummy_call}. However the function is provided with the
4381 type of the function result, @var{value_type}, @var{bp_addr} is used
4382 to return a value (the address at which the breakpoint instruction
4383 should be inserted) and @var{real pc} is used to specify the resume
4384 address when starting the call sequence. The function should return
4385 the updated innermost stack address.
4388 @emph{Note:} This does require that code in the stack can be executed.
4389 Some Harvard architectures may not allow this.
4394 @node Adding support for debugging core files
4395 @section Adding support for debugging core files
4398 The prerequisite for adding core file support in @value{GDBN} is to have
4399 core file support in BFD.
4401 Once BFD support is available, writing the apropriate
4402 @code{regset_from_core_section} architecture function should be all
4403 that is needed in order to add support for core files in @value{GDBN}.
4405 @node Defining Other Architecture Features
4406 @section Defining Other Architecture Features
4408 This section describes other functions and values in @code{gdbarch},
4409 together with some useful macros, that you can use to define the
4410 target architecture.
4414 @item CORE_ADDR gdbarch_addr_bits_remove (@var{gdbarch}, @var{addr})
4415 @findex gdbarch_addr_bits_remove
4416 If a raw machine instruction address includes any bits that are not
4417 really part of the address, then this function is used to zero those bits in
4418 @var{addr}. This is only used for addresses of instructions, and even then not
4421 For example, the two low-order bits of the PC on the Hewlett-Packard PA
4422 2.0 architecture contain the privilege level of the corresponding
4423 instruction. Since instructions must always be aligned on four-byte
4424 boundaries, the processor masks out these bits to generate the actual
4425 address of the instruction. @code{gdbarch_addr_bits_remove} would then for
4426 example look like that:
4428 arch_addr_bits_remove (CORE_ADDR addr)
4430 return (addr &= ~0x3);
4434 @item int address_class_name_to_type_flags (@var{gdbarch}, @var{name}, @var{type_flags_ptr})
4435 @findex address_class_name_to_type_flags
4436 If @var{name} is a valid address class qualifier name, set the @code{int}
4437 referenced by @var{type_flags_ptr} to the mask representing the qualifier
4438 and return 1. If @var{name} is not a valid address class qualifier name,
4441 The value for @var{type_flags_ptr} should be one of
4442 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
4443 possibly some combination of these values or'd together.
4444 @xref{Target Architecture Definition, , Address Classes}.
4446 @item int address_class_name_to_type_flags_p (@var{gdbarch})
4447 @findex address_class_name_to_type_flags_p
4448 Predicate which indicates whether @code{address_class_name_to_type_flags}
4451 @item int gdbarch_address_class_type_flags (@var{gdbarch}, @var{byte_size}, @var{dwarf2_addr_class})
4452 @findex gdbarch_address_class_type_flags
4453 Given a pointers byte size (as described by the debug information) and
4454 the possible @code{DW_AT_address_class} value, return the type flags
4455 used by @value{GDBN} to represent this address class. The value
4456 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
4457 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
4458 values or'd together.
4459 @xref{Target Architecture Definition, , Address Classes}.
4461 @item int gdbarch_address_class_type_flags_p (@var{gdbarch})
4462 @findex gdbarch_address_class_type_flags_p
4463 Predicate which indicates whether @code{gdbarch_address_class_type_flags_p} has
4466 @item const char *gdbarch_address_class_type_flags_to_name (@var{gdbarch}, @var{type_flags})
4467 @findex gdbarch_address_class_type_flags_to_name
4468 Return the name of the address class qualifier associated with the type
4469 flags given by @var{type_flags}.
4471 @item int gdbarch_address_class_type_flags_to_name_p (@var{gdbarch})
4472 @findex gdbarch_address_class_type_flags_to_name_p
4473 Predicate which indicates whether @code{gdbarch_address_class_type_flags_to_name} has been defined.
4474 @xref{Target Architecture Definition, , Address Classes}.
4476 @item void gdbarch_address_to_pointer (@var{gdbarch}, @var{type}, @var{buf}, @var{addr})
4477 @findex gdbarch_address_to_pointer
4478 Store in @var{buf} a pointer of type @var{type} representing the address
4479 @var{addr}, in the appropriate format for the current architecture.
4480 This function may safely assume that @var{type} is either a pointer or a
4481 C@t{++} reference type.
4482 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
4484 @item int gdbarch_believe_pcc_promotion (@var{gdbarch})
4485 @findex gdbarch_believe_pcc_promotion
4486 Used to notify if the compiler promotes a @code{short} or @code{char}
4487 parameter to an @code{int}, but still reports the parameter as its
4488 original type, rather than the promoted type.
4490 @item gdbarch_bits_big_endian (@var{gdbarch})
4491 @findex gdbarch_bits_big_endian
4492 This is used if the numbering of bits in the targets does @strong{not} match
4493 the endianism of the target byte order. A value of 1 means that the bits
4494 are numbered in a big-endian bit order, 0 means little-endian.
4496 @item set_gdbarch_bits_big_endian (@var{gdbarch}, @var{bits_big_endian})
4497 @findex set_gdbarch_bits_big_endian
4498 Calling set_gdbarch_bits_big_endian with a value of 1 indicates that the
4499 bits in the target are numbered in a big-endian bit order, 0 indicates
4504 This is the character array initializer for the bit pattern to put into
4505 memory where a breakpoint is set. Although it's common to use a trap
4506 instruction for a breakpoint, it's not required; for instance, the bit
4507 pattern could be an invalid instruction. The breakpoint must be no
4508 longer than the shortest instruction of the architecture.
4510 @code{BREAKPOINT} has been deprecated in favor of
4511 @code{gdbarch_breakpoint_from_pc}.
4513 @item BIG_BREAKPOINT
4514 @itemx LITTLE_BREAKPOINT
4515 @findex LITTLE_BREAKPOINT
4516 @findex BIG_BREAKPOINT
4517 Similar to BREAKPOINT, but used for bi-endian targets.
4519 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
4520 favor of @code{gdbarch_breakpoint_from_pc}.
4522 @item const gdb_byte *gdbarch_breakpoint_from_pc (@var{gdbarch}, @var{pcptr}, @var{lenptr})
4523 @findex gdbarch_breakpoint_from_pc
4524 @anchor{gdbarch_breakpoint_from_pc} Use the program counter to determine the
4525 contents and size of a breakpoint instruction. It returns a pointer to
4526 a static string of bytes that encode a breakpoint instruction, stores the
4527 length of the string to @code{*@var{lenptr}}, and adjusts the program
4528 counter (if necessary) to point to the actual memory location where the
4529 breakpoint should be inserted. On input, the program counter
4530 (@code{*@var{pcptr}} is the encoded inferior's PC register. If software
4531 breakpoints are supported, the function sets this argument to the PC's
4532 plain address. If software breakpoints are not supported, the function
4533 returns NULL instead of the encoded breakpoint instruction.
4535 Although it is common to use a trap instruction for a breakpoint, it's
4536 not required; for instance, the bit pattern could be an invalid
4537 instruction. The breakpoint must be no longer than the shortest
4538 instruction of the architecture.
4540 Provided breakpoint bytes can be also used by @code{bp_loc_is_permanent} to
4541 detect permanent breakpoints. @code{gdbarch_breakpoint_from_pc} should return
4542 an unchanged memory copy if it was called for a location with permanent
4543 breakpoint as some architectures use breakpoint instructions containing
4544 arbitrary parameter value.
4546 Replaces all the other @var{BREAKPOINT} macros.
4548 @item int gdbarch_memory_insert_breakpoint (@var{gdbarch}, @var{bp_tgt})
4549 @itemx gdbarch_memory_remove_breakpoint (@var{gdbarch}, @var{bp_tgt})
4550 @findex gdbarch_memory_remove_breakpoint
4551 @findex gdbarch_memory_insert_breakpoint
4552 Insert or remove memory based breakpoints. Reasonable defaults
4553 (@code{default_memory_insert_breakpoint} and
4554 @code{default_memory_remove_breakpoint} respectively) have been
4555 provided so that it is not necessary to set these for most
4556 architectures. Architectures which may want to set
4557 @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
4558 conventional manner.
4560 It may also be desirable (from an efficiency standpoint) to define
4561 custom breakpoint insertion and removal routines if
4562 @code{gdbarch_breakpoint_from_pc} needs to read the target's memory for some
4565 @item CORE_ADDR gdbarch_adjust_breakpoint_address (@var{gdbarch}, @var{bpaddr})
4566 @findex gdbarch_adjust_breakpoint_address
4567 @cindex breakpoint address adjusted
4568 Given an address at which a breakpoint is desired, return a breakpoint
4569 address adjusted to account for architectural constraints on
4570 breakpoint placement. This method is not needed by most targets.
4572 The FR-V target (see @file{frv-tdep.c}) requires this method.
4573 The FR-V is a VLIW architecture in which a number of RISC-like
4574 instructions are grouped (packed) together into an aggregate
4575 instruction or instruction bundle. When the processor executes
4576 one of these bundles, the component instructions are executed
4579 In the course of optimization, the compiler may group instructions
4580 from distinct source statements into the same bundle. The line number
4581 information associated with one of the latter statements will likely
4582 refer to some instruction other than the first one in the bundle. So,
4583 if the user attempts to place a breakpoint on one of these latter
4584 statements, @value{GDBN} must be careful to @emph{not} place the break
4585 instruction on any instruction other than the first one in the bundle.
4586 (Remember though that the instructions within a bundle execute
4587 in parallel, so the @emph{first} instruction is the instruction
4588 at the lowest address and has nothing to do with execution order.)
4590 The FR-V's @code{gdbarch_adjust_breakpoint_address} method will adjust a
4591 breakpoint's address by scanning backwards for the beginning of
4592 the bundle, returning the address of the bundle.
4594 Since the adjustment of a breakpoint may significantly alter a user's
4595 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
4596 is initially set and each time that that breakpoint is hit.
4598 @item int gdbarch_call_dummy_location (@var{gdbarch})
4599 @findex gdbarch_call_dummy_location
4600 See the file @file{inferior.h}.
4602 This method has been replaced by @code{gdbarch_push_dummy_code}
4603 (@pxref{gdbarch_push_dummy_code}).
4605 @item int gdbarch_cannot_fetch_register (@var{gdbarch}, @var{regum})
4606 @findex gdbarch_cannot_fetch_register
4607 This function should return nonzero if @var{regno} cannot be fetched
4608 from an inferior process.
4610 @item int gdbarch_cannot_store_register (@var{gdbarch}, @var{regnum})
4611 @findex gdbarch_cannot_store_register
4612 This function should return nonzero if @var{regno} should not be
4613 written to the target. This is often the case for program counters,
4614 status words, and other special registers. This function returns 0 as
4615 default so that @value{GDBN} will assume that all registers may be written.
4617 @item int gdbarch_convert_register_p (@var{gdbarch}, @var{regnum}, struct type *@var{type})
4618 @findex gdbarch_convert_register_p
4619 Return non-zero if register @var{regnum} represents data values of type
4620 @var{type} in a non-standard form.
4621 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
4623 @item int gdbarch_fp0_regnum (@var{gdbarch})
4624 @findex gdbarch_fp0_regnum
4625 This function returns the number of the first floating point register,
4626 if the machine has such registers. Otherwise, it returns -1.
4628 @item CORE_ADDR gdbarch_decr_pc_after_break (@var{gdbarch})
4629 @findex gdbarch_decr_pc_after_break
4630 This function shall return the amount by which to decrement the PC after the
4631 program encounters a breakpoint. This is often the number of bytes in
4632 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
4634 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
4635 @findex DISABLE_UNSETTABLE_BREAK
4636 If defined, this should evaluate to 1 if @var{addr} is in a shared
4637 library in which breakpoints cannot be set and so should be disabled.
4639 @item int gdbarch_dwarf2_reg_to_regnum (@var{gdbarch}, @var{dwarf2_regnr})
4640 @findex gdbarch_dwarf2_reg_to_regnum
4641 Convert DWARF2 register number @var{dwarf2_regnr} into @value{GDBN} regnum.
4642 If not defined, no conversion will be performed.
4644 @item int gdbarch_ecoff_reg_to_regnum (@var{gdbarch}, @var{ecoff_regnr})
4645 @findex gdbarch_ecoff_reg_to_regnum
4646 Convert ECOFF register number @var{ecoff_regnr} into @value{GDBN} regnum. If
4647 not defined, no conversion will be performed.
4649 @item GCC_COMPILED_FLAG_SYMBOL
4650 @itemx GCC2_COMPILED_FLAG_SYMBOL
4651 @findex GCC2_COMPILED_FLAG_SYMBOL
4652 @findex GCC_COMPILED_FLAG_SYMBOL
4653 If defined, these are the names of the symbols that @value{GDBN} will
4654 look for to detect that GCC compiled the file. The default symbols
4655 are @code{gcc_compiled.} and @code{gcc2_compiled.},
4656 respectively. (Currently only defined for the Delta 68.)
4658 @item gdbarch_get_longjmp_target
4659 @findex gdbarch_get_longjmp_target
4660 This function determines the target PC address that @code{longjmp}
4661 will jump to, assuming that we have just stopped at a @code{longjmp}
4662 breakpoint. It takes a @code{CORE_ADDR *} as argument, and stores the
4663 target PC value through this pointer. It examines the current state
4664 of the machine as needed, typically by using a manually-determined
4665 offset into the @code{jmp_buf}. (While we might like to get the offset
4666 from the target's @file{jmpbuf.h}, that header file cannot be assumed
4667 to be available when building a cross-debugger.)
4669 @item DEPRECATED_IBM6000_TARGET
4670 @findex DEPRECATED_IBM6000_TARGET
4671 Shows that we are configured for an IBM RS/6000 system. This
4672 conditional should be eliminated (FIXME) and replaced by
4673 feature-specific macros. It was introduced in haste and we are
4674 repenting at leisure.
4676 @item I386_USE_GENERIC_WATCHPOINTS
4677 An x86-based target can define this to use the generic x86 watchpoint
4678 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4680 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{addr})
4681 @findex gdbarch_in_function_epilogue_p
4682 Returns non-zero if the given @var{addr} is in the epilogue of a function.
4683 The epilogue of a function is defined as the part of a function where
4684 the stack frame of the function already has been destroyed up to the
4685 final `return from function call' instruction.
4687 @item int gdbarch_in_solib_return_trampoline (@var{gdbarch}, @var{pc}, @var{name})
4688 @findex gdbarch_in_solib_return_trampoline
4689 Define this function to return nonzero if the program is stopped in the
4690 trampoline that returns from a shared library.
4692 @item target_so_ops.in_dynsym_resolve_code (@var{pc})
4693 @findex in_dynsym_resolve_code
4694 Define this to return nonzero if the program is stopped in the
4697 @item SKIP_SOLIB_RESOLVER (@var{pc})
4698 @findex SKIP_SOLIB_RESOLVER
4699 Define this to evaluate to the (nonzero) address at which execution
4700 should continue to get past the dynamic linker's symbol resolution
4701 function. A zero value indicates that it is not important or necessary
4702 to set a breakpoint to get through the dynamic linker and that single
4703 stepping will suffice.
4705 @item CORE_ADDR gdbarch_integer_to_address (@var{gdbarch}, @var{type}, @var{buf})
4706 @findex gdbarch_integer_to_address
4707 @cindex converting integers to addresses
4708 Define this when the architecture needs to handle non-pointer to address
4709 conversions specially. Converts that value to an address according to
4710 the current architectures conventions.
4712 @emph{Pragmatics: When the user copies a well defined expression from
4713 their source code and passes it, as a parameter, to @value{GDBN}'s
4714 @code{print} command, they should get the same value as would have been
4715 computed by the target program. Any deviation from this rule can cause
4716 major confusion and annoyance, and needs to be justified carefully. In
4717 other words, @value{GDBN} doesn't really have the freedom to do these
4718 conversions in clever and useful ways. It has, however, been pointed
4719 out that users aren't complaining about how @value{GDBN} casts integers
4720 to pointers; they are complaining that they can't take an address from a
4721 disassembly listing and give it to @code{x/i}. Adding an architecture
4722 method like @code{gdbarch_integer_to_address} certainly makes it possible for
4723 @value{GDBN} to ``get it right'' in all circumstances.}
4725 @xref{Target Architecture Definition, , Pointers Are Not Always
4728 @item CORE_ADDR gdbarch_pointer_to_address (@var{gdbarch}, @var{type}, @var{buf})
4729 @findex gdbarch_pointer_to_address
4730 Assume that @var{buf} holds a pointer of type @var{type}, in the
4731 appropriate format for the current architecture. Return the byte
4732 address the pointer refers to.
4733 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
4735 @item void gdbarch_register_to_value(@var{gdbarch}, @var{frame}, @var{regnum}, @var{type}, @var{fur})
4736 @findex gdbarch_register_to_value
4737 Convert the raw contents of register @var{regnum} into a value of type
4739 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
4741 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
4742 @findex REGISTER_CONVERT_TO_VIRTUAL
4743 Convert the value of register @var{reg} from its raw form to its virtual
4745 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
4747 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
4748 @findex REGISTER_CONVERT_TO_RAW
4749 Convert the value of register @var{reg} from its virtual form to its raw
4751 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
4753 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
4754 @findex regset_from_core_section
4755 Return the appropriate register set for a core file section with name
4756 @var{sect_name} and size @var{sect_size}.
4758 @item SOFTWARE_SINGLE_STEP_P()
4759 @findex SOFTWARE_SINGLE_STEP_P
4760 Define this as 1 if the target does not have a hardware single-step
4761 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
4763 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breakpoints_p})
4764 @findex SOFTWARE_SINGLE_STEP
4765 A function that inserts or removes (depending on
4766 @var{insert_breakpoints_p}) breakpoints at each possible destinations of
4767 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
4770 @item set_gdbarch_sofun_address_maybe_missing (@var{gdbarch}, @var{set})
4771 @findex set_gdbarch_sofun_address_maybe_missing
4772 Somebody clever observed that, the more actual addresses you have in the
4773 debug information, the more time the linker has to spend relocating
4774 them. So whenever there's some other way the debugger could find the
4775 address it needs, you should omit it from the debug info, to make
4778 Calling @code{set_gdbarch_sofun_address_maybe_missing} with a non-zero
4779 argument @var{set} indicates that a particular set of hacks of this sort
4780 are in use, affecting @code{N_SO} and @code{N_FUN} entries in stabs-format
4781 debugging information. @code{N_SO} stabs mark the beginning and ending
4782 addresses of compilation units in the text segment. @code{N_FUN} stabs
4783 mark the starts and ends of functions.
4785 In this case, @value{GDBN} assumes two things:
4789 @code{N_FUN} stabs have an address of zero. Instead of using those
4790 addresses, you should find the address where the function starts by
4791 taking the function name from the stab, and then looking that up in the
4792 minsyms (the linker/assembler symbol table). In other words, the stab
4793 has the name, and the linker/assembler symbol table is the only place
4794 that carries the address.
4797 @code{N_SO} stabs have an address of zero, too. You just look at the
4798 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab, and
4799 guess the starting and ending addresses of the compilation unit from them.
4802 @item int gdbarch_stabs_argument_has_addr (@var{gdbarch}, @var{type})
4803 @findex gdbarch_stabs_argument_has_addr
4804 @anchor{gdbarch_stabs_argument_has_addr} Define this function to return
4805 nonzero if a function argument of type @var{type} is passed by reference
4808 @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})
4809 @findex gdbarch_push_dummy_call
4810 @anchor{gdbarch_push_dummy_call} Define this to push the dummy frame's call to
4811 the inferior function onto the stack. In addition to pushing @var{nargs}, the
4812 code should push @var{struct_addr} (when @var{struct_return} is non-zero), and
4813 the return address (@var{bp_addr}, in inferior's PC register encoding).
4815 @var{function} is a pointer to a @code{struct value}; on architectures that use
4816 function descriptors, this contains the function descriptor value.
4818 Returns the updated top-of-stack pointer.
4820 @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})
4821 @findex gdbarch_push_dummy_code
4822 @anchor{gdbarch_push_dummy_code} Given a stack based call dummy, push the
4823 instruction sequence (including space for a breakpoint) to which the
4824 called function should return.
4826 Set @var{bp_addr} to the address at which the breakpoint instruction
4827 should be inserted (in inferior's PC register encoding), @var{real_pc} to the
4828 resume address when starting the call sequence, and return the updated
4829 inner-most stack address.
4831 By default, the stack is grown sufficient to hold a frame-aligned
4832 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
4833 reserved for that breakpoint (in inferior's PC register encoding), and
4834 @var{real_pc} set to @var{funaddr}.
4836 This method replaces @w{@code{gdbarch_call_dummy_location (@var{gdbarch})}}.
4838 @item int gdbarch_sdb_reg_to_regnum (@var{gdbarch}, @var{sdb_regnr})
4839 @findex gdbarch_sdb_reg_to_regnum
4840 Use this function to convert sdb register @var{sdb_regnr} into @value{GDBN}
4841 regnum. If not defined, no conversion will be done.
4843 @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})
4844 @findex gdbarch_return_value
4845 @anchor{gdbarch_return_value} Given a function with a return-value of
4846 type @var{rettype}, return which return-value convention that function
4849 @value{GDBN} currently recognizes two function return-value conventions:
4850 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
4851 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
4852 value is found in memory and the address of that memory location is
4853 passed in as the function's first parameter.
4855 If the register convention is being used, and @var{writebuf} is
4856 non-@code{NULL}, also copy the return-value in @var{writebuf} into
4859 If the register convention is being used, and @var{readbuf} is
4860 non-@code{NULL}, also copy the return value from @var{regcache} into
4861 @var{readbuf} (@var{regcache} contains a copy of the registers from the
4862 just returned function).
4864 @emph{Maintainer note: This method replaces separate predicate, extract,
4865 store methods. By having only one method, the logic needed to determine
4866 the return-value convention need only be implemented in one place. If
4867 @value{GDBN} were written in an @sc{oo} language, this method would
4868 instead return an object that knew how to perform the register
4869 return-value extract and store.}
4871 @emph{Maintainer note: This method does not take a @var{gcc_p}
4872 parameter, and such a parameter should not be added. If an architecture
4873 that requires per-compiler or per-function information be identified,
4874 then the replacement of @var{rettype} with @code{struct value}
4875 @var{function} should be pursued.}
4877 @emph{Maintainer note: The @var{regcache} parameter limits this methods
4878 to the inner most frame. While replacing @var{regcache} with a
4879 @code{struct frame_info} @var{frame} parameter would remove that
4880 limitation there has yet to be a demonstrated need for such a change.}
4882 @item void gdbarch_skip_permanent_breakpoint (@var{gdbarch}, @var{regcache})
4883 @findex gdbarch_skip_permanent_breakpoint
4884 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
4885 steps over a breakpoint by removing it, stepping one instruction, and
4886 re-inserting the breakpoint. However, permanent breakpoints are
4887 hardwired into the inferior, and can't be removed, so this strategy
4888 doesn't work. Calling @code{gdbarch_skip_permanent_breakpoint} adjusts the
4889 processor's state so that execution will resume just after the breakpoint.
4890 This function does the right thing even when the breakpoint is in the delay slot
4891 of a branch or jump.
4893 @item CORE_ADDR gdbarch_skip_trampoline_code (@var{gdbarch}, @var{frame}, @var{pc})
4894 @findex gdbarch_skip_trampoline_code
4895 If the target machine has trampoline code that sits between callers and
4896 the functions being called, then define this function to return a new PC
4897 that is at the start of the real function.
4899 @item int gdbarch_deprecated_fp_regnum (@var{gdbarch})
4900 @findex gdbarch_deprecated_fp_regnum
4901 If the frame pointer is in a register, use this function to return the
4902 number of that register.
4904 @item int gdbarch_stab_reg_to_regnum (@var{gdbarch}, @var{stab_regnr})
4905 @findex gdbarch_stab_reg_to_regnum
4906 Use this function to convert stab register @var{stab_regnr} into @value{GDBN}
4907 regnum. If not defined, no conversion will be done.
4909 @item TARGET_CHAR_BIT
4910 @findex TARGET_CHAR_BIT
4911 Number of bits in a char; defaults to 8.
4913 @item int gdbarch_char_signed (@var{gdbarch})
4914 @findex gdbarch_char_signed
4915 Non-zero if @code{char} is normally signed on this architecture; zero if
4916 it should be unsigned.
4918 The ISO C standard requires the compiler to treat @code{char} as
4919 equivalent to either @code{signed char} or @code{unsigned char}; any
4920 character in the standard execution set is supposed to be positive.
4921 Most compilers treat @code{char} as signed, but @code{char} is unsigned
4922 on the IBM S/390, RS6000, and PowerPC targets.
4924 @item int gdbarch_double_bit (@var{gdbarch})
4925 @findex gdbarch_double_bit
4926 Number of bits in a double float; defaults to @w{@code{8 * TARGET_CHAR_BIT}}.
4928 @item int gdbarch_float_bit (@var{gdbarch})
4929 @findex gdbarch_float_bit
4930 Number of bits in a float; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4932 @item int gdbarch_int_bit (@var{gdbarch})
4933 @findex gdbarch_int_bit
4934 Number of bits in an integer; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4936 @item int gdbarch_long_bit (@var{gdbarch})
4937 @findex gdbarch_long_bit
4938 Number of bits in a long integer; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4940 @item int gdbarch_long_double_bit (@var{gdbarch})
4941 @findex gdbarch_long_double_bit
4942 Number of bits in a long double float;
4943 defaults to @w{@code{2 * gdbarch_double_bit (@var{gdbarch})}}.
4945 @item int gdbarch_long_long_bit (@var{gdbarch})
4946 @findex gdbarch_long_long_bit
4947 Number of bits in a long long integer; defaults to
4948 @w{@code{2 * gdbarch_long_bit (@var{gdbarch})}}.
4950 @item int gdbarch_ptr_bit (@var{gdbarch})
4951 @findex gdbarch_ptr_bit
4952 Number of bits in a pointer; defaults to
4953 @w{@code{gdbarch_int_bit (@var{gdbarch})}}.
4955 @item int gdbarch_short_bit (@var{gdbarch})
4956 @findex gdbarch_short_bit
4957 Number of bits in a short integer; defaults to @w{@code{2 * TARGET_CHAR_BIT}}.
4959 @item void gdbarch_virtual_frame_pointer (@var{gdbarch}, @var{pc}, @var{frame_regnum}, @var{frame_offset})
4960 @findex gdbarch_virtual_frame_pointer
4961 Returns a @code{(@var{register}, @var{offset})} pair representing the virtual
4962 frame pointer in use at the code address @var{pc}. If virtual frame
4963 pointers are not used, a default definition simply returns
4964 @code{gdbarch_deprecated_fp_regnum} (or @code{gdbarch_sp_regnum}, if
4965 no frame pointer is defined), with an offset of zero.
4967 @c need to explain virtual frame pointers, they are recorded in agent
4968 @c expressions for tracepoints
4970 @item TARGET_HAS_HARDWARE_WATCHPOINTS
4971 If non-zero, the target has support for hardware-assisted
4972 watchpoints. @xref{Algorithms, watchpoints}, for more details and
4973 other related macros.
4975 @item int gdbarch_print_insn (@var{gdbarch}, @var{vma}, @var{info})
4976 @findex gdbarch_print_insn
4977 This is the function used by @value{GDBN} to print an assembly
4978 instruction. It prints the instruction at address @var{vma} in
4979 debugged memory and returns the length of the instruction, in bytes.
4980 This usually points to a function in the @code{opcodes} library
4981 (@pxref{Support Libraries, ,Opcodes}). @var{info} is a structure (of
4982 type @code{disassemble_info}) defined in the header file
4983 @file{include/dis-asm.h}, and used to pass information to the
4984 instruction decoding routine.
4986 @item frame_id gdbarch_dummy_id (@var{gdbarch}, @var{frame})
4987 @findex gdbarch_dummy_id
4988 @anchor{gdbarch_dummy_id} Given @var{frame} return a @w{@code{struct
4989 frame_id}} that uniquely identifies an inferior function call's dummy
4990 frame. The value returned must match the dummy frame stack value
4991 previously saved by @code{call_function_by_hand}.
4993 @item void gdbarch_value_to_register (@var{gdbarch}, @var{frame}, @var{type}, @var{buf})
4994 @findex gdbarch_value_to_register
4995 Convert a value of type @var{type} into the raw contents of a register.
4996 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
5000 Motorola M68K target conditionals.
5004 Define this to be the 4-bit location of the breakpoint trap vector. If
5005 not defined, it will default to @code{0xf}.
5007 @item REMOTE_BPT_VECTOR
5008 Defaults to @code{1}.
5012 @node Adding a New Target
5013 @section Adding a New Target
5015 @cindex adding a target
5016 The following files add a target to @value{GDBN}:
5019 @cindex target dependent files
5021 @item gdb/@var{ttt}-tdep.c
5022 Contains any miscellaneous code required for this target machine. On
5023 some machines it doesn't exist at all.
5025 @item gdb/@var{arch}-tdep.c
5026 @itemx gdb/@var{arch}-tdep.h
5027 This is required to describe the basic layout of the target machine's
5028 processor chip (registers, stack, etc.). It can be shared among many
5029 targets that use the same processor architecture.
5033 (Target header files such as
5034 @file{gdb/config/@var{arch}/tm-@var{ttt}.h},
5035 @file{gdb/config/@var{arch}/tm-@var{arch}.h}, and
5036 @file{config/tm-@var{os}.h} are no longer used.)
5038 @findex _initialize_@var{arch}_tdep
5039 A @value{GDBN} description for a new architecture, arch is created by
5040 defining a global function @code{_initialize_@var{arch}_tdep}, by
5041 convention in the source file @file{@var{arch}-tdep.c}. For
5042 example, in the case of the OpenRISC 1000, this function is called
5043 @code{_initialize_or1k_tdep} and is found in the file
5046 The object file resulting from compiling this source file, which will
5047 contain the implementation of the
5048 @code{_initialize_@var{arch}_tdep} function is specified in the
5049 @value{GDBN} @file{configure.tgt} file, which includes a large case
5050 statement pattern matching against the @code{--target} option of the
5051 @kbd{configure} script.
5054 @emph{Note:} If the architecture requires multiple source files, the
5055 corresponding binaries should be included in
5056 @file{configure.tgt}. However if there are header files, the
5057 dependencies on these will not be picked up from the entries in
5058 @file{configure.tgt}. The @file{Makefile.in} file will need extending to
5059 show these dependencies.
5062 @findex gdbarch_register
5063 A new struct gdbarch, defining the new architecture, is created within
5064 the @code{_initialize_@var{arch}_tdep} function by calling
5065 @code{gdbarch_register}:
5068 void gdbarch_register (enum bfd_architecture architecture,
5069 gdbarch_init_ftype *init_func,
5070 gdbarch_dump_tdep_ftype *tdep_dump_func);
5073 This function has been described fully in an earlier
5074 section. @xref{How an Architecture is Represented, , How an
5075 Architecture is Represented}.
5077 The new @code{@w{struct gdbarch}} should contain implementations of
5078 the necessary functions (described in the previous sections) to
5079 describe the basic layout of the target machine's processor chip
5080 (registers, stack, etc.). It can be shared among many targets that use
5081 the same processor architecture.
5083 @node Target Descriptions
5084 @chapter Target Descriptions
5085 @cindex target descriptions
5087 The target architecture definition (@pxref{Target Architecture Definition})
5088 contains @value{GDBN}'s hard-coded knowledge about an architecture. For
5089 some platforms, it is handy to have more flexible knowledge about a specific
5090 instance of the architecture---for instance, a processor or development board.
5091 @dfn{Target descriptions} provide a mechanism for the user to tell @value{GDBN}
5092 more about what their target supports, or for the target to tell @value{GDBN}
5095 For details on writing, automatically supplying, and manually selecting
5096 target descriptions, see @ref{Target Descriptions, , , gdb,
5097 Debugging with @value{GDBN}}. This section will cover some related
5098 topics about the @value{GDBN} internals.
5101 * Target Descriptions Implementation::
5102 * Adding Target Described Register Support::
5105 @node Target Descriptions Implementation
5106 @section Target Descriptions Implementation
5107 @cindex target descriptions, implementation
5109 Before @value{GDBN} connects to a new target, or runs a new program on
5110 an existing target, it discards any existing target description and
5111 reverts to a default gdbarch. Then, after connecting, it looks for a
5112 new target description by calling @code{target_find_description}.
5114 A description may come from a user specified file (XML), the remote
5115 @samp{qXfer:features:read} packet (also XML), or from any custom
5116 @code{to_read_description} routine in the target vector. For instance,
5117 the remote target supports guessing whether a MIPS target is 32-bit or
5118 64-bit based on the size of the @samp{g} packet.
5120 If any target description is found, @value{GDBN} creates a new gdbarch
5121 incorporating the description by calling @code{gdbarch_update_p}. Any
5122 @samp{<architecture>} element is handled first, to determine which
5123 architecture's gdbarch initialization routine is called to create the
5124 new architecture. Then the initialization routine is called, and has
5125 a chance to adjust the constructed architecture based on the contents
5126 of the target description. For instance, it can recognize any
5127 properties set by a @code{to_read_description} routine. Also
5128 see @ref{Adding Target Described Register Support}.
5130 @node Adding Target Described Register Support
5131 @section Adding Target Described Register Support
5132 @cindex target descriptions, adding register support
5134 Target descriptions can report additional registers specific to an
5135 instance of the target. But it takes a little work in the architecture
5136 specific routines to support this.
5138 A target description must either have no registers or a complete
5139 set---this avoids complexity in trying to merge standard registers
5140 with the target defined registers. It is the architecture's
5141 responsibility to validate that a description with registers has
5142 everything it needs. To keep architecture code simple, the same
5143 mechanism is used to assign fixed internal register numbers to
5146 If @code{tdesc_has_registers} returns 1, the description contains
5147 registers. The architecture's @code{gdbarch_init} routine should:
5152 Call @code{tdesc_data_alloc} to allocate storage, early, before
5153 searching for a matching gdbarch or allocating a new one.
5156 Use @code{tdesc_find_feature} to locate standard features by name.
5159 Use @code{tdesc_numbered_register} and @code{tdesc_numbered_register_choices}
5160 to locate the expected registers in the standard features.
5163 Return @code{NULL} if a required feature is missing, or if any standard
5164 feature is missing expected registers. This will produce a warning that
5165 the description was incomplete.
5168 Free the allocated data before returning, unless @code{tdesc_use_registers}
5172 Call @code{set_gdbarch_num_regs} as usual, with a number higher than any
5173 fixed number passed to @code{tdesc_numbered_register}.
5176 Call @code{tdesc_use_registers} after creating a new gdbarch, before
5181 After @code{tdesc_use_registers} has been called, the architecture's
5182 @code{register_name}, @code{register_type}, and @code{register_reggroup_p}
5183 routines will not be called; that information will be taken from
5184 the target description. @code{num_regs} may be increased to account
5185 for any additional registers in the description.
5187 Pseudo-registers require some extra care:
5192 Using @code{tdesc_numbered_register} allows the architecture to give
5193 constant register numbers to standard architectural registers, e.g.@:
5194 as an @code{enum} in @file{@var{arch}-tdep.h}. But because
5195 pseudo-registers are always numbered above @code{num_regs},
5196 which may be increased by the description, constant numbers
5197 can not be used for pseudos. They must be numbered relative to
5198 @code{num_regs} instead.
5201 The description will not describe pseudo-registers, so the
5202 architecture must call @code{set_tdesc_pseudo_register_name},
5203 @code{set_tdesc_pseudo_register_type}, and
5204 @code{set_tdesc_pseudo_register_reggroup_p} to supply routines
5205 describing pseudo registers. These routines will be passed
5206 internal register numbers, so the same routines used for the
5207 gdbarch equivalents are usually suitable.
5212 @node Target Vector Definition
5214 @chapter Target Vector Definition
5215 @cindex target vector
5217 The target vector defines the interface between @value{GDBN}'s
5218 abstract handling of target systems, and the nitty-gritty code that
5219 actually exercises control over a process or a serial port.
5220 @value{GDBN} includes some 30-40 different target vectors; however,
5221 each configuration of @value{GDBN} includes only a few of them.
5224 * Managing Execution State::
5225 * Existing Targets::
5228 @node Managing Execution State
5229 @section Managing Execution State
5230 @cindex execution state
5232 A target vector can be completely inactive (not pushed on the target
5233 stack), active but not running (pushed, but not connected to a fully
5234 manifested inferior), or completely active (pushed, with an accessible
5235 inferior). Most targets are only completely inactive or completely
5236 active, but some support persistent connections to a target even
5237 when the target has exited or not yet started.
5239 For example, connecting to the simulator using @code{target sim} does
5240 not create a running program. Neither registers nor memory are
5241 accessible until @code{run}. Similarly, after @code{kill}, the
5242 program can not continue executing. But in both cases @value{GDBN}
5243 remains connected to the simulator, and target-specific commands
5244 are directed to the simulator.
5246 A target which only supports complete activation should push itself
5247 onto the stack in its @code{to_open} routine (by calling
5248 @code{push_target}), and unpush itself from the stack in its
5249 @code{to_mourn_inferior} routine (by calling @code{unpush_target}).
5251 A target which supports both partial and complete activation should
5252 still call @code{push_target} in @code{to_open}, but not call
5253 @code{unpush_target} in @code{to_mourn_inferior}. Instead, it should
5254 call either @code{target_mark_running} or @code{target_mark_exited}
5255 in its @code{to_open}, depending on whether the target is fully active
5256 after connection. It should also call @code{target_mark_running} any
5257 time the inferior becomes fully active (e.g.@: in
5258 @code{to_create_inferior} and @code{to_attach}), and
5259 @code{target_mark_exited} when the inferior becomes inactive (in
5260 @code{to_mourn_inferior}). The target should also make sure to call
5261 @code{target_mourn_inferior} from its @code{to_kill}, to return the
5262 target to inactive state.
5264 @node Existing Targets
5265 @section Existing Targets
5268 @subsection File Targets
5270 Both executables and core files have target vectors.
5272 @subsection Standard Protocol and Remote Stubs
5274 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code that
5275 runs in the target system. @value{GDBN} provides several sample
5276 @dfn{stubs} that can be integrated into target programs or operating
5277 systems for this purpose; they are named @file{@var{cpu}-stub.c}. Many
5278 operating systems, embedded targets, emulators, and simulators already
5279 have a @value{GDBN} stub built into them, and maintenance of the remote
5280 protocol must be careful to preserve compatibility.
5282 The @value{GDBN} user's manual describes how to put such a stub into
5283 your target code. What follows is a discussion of integrating the
5284 SPARC stub into a complicated operating system (rather than a simple
5285 program), by Stu Grossman, the author of this stub.
5287 The trap handling code in the stub assumes the following upon entry to
5292 %l1 and %l2 contain pc and npc respectively at the time of the trap;
5298 you are in the correct trap window.
5301 As long as your trap handler can guarantee those conditions, then there
5302 is no reason why you shouldn't be able to ``share'' traps with the stub.
5303 The stub has no requirement that it be jumped to directly from the
5304 hardware trap vector. That is why it calls @code{exceptionHandler()},
5305 which is provided by the external environment. For instance, this could
5306 set up the hardware traps to actually execute code which calls the stub
5307 first, and then transfers to its own trap handler.
5309 For the most point, there probably won't be much of an issue with
5310 ``sharing'' traps, as the traps we use are usually not used by the kernel,
5311 and often indicate unrecoverable error conditions. Anyway, this is all
5312 controlled by a table, and is trivial to modify. The most important
5313 trap for us is for @code{ta 1}. Without that, we can't single step or
5314 do breakpoints. Everything else is unnecessary for the proper operation
5315 of the debugger/stub.
5317 From reading the stub, it's probably not obvious how breakpoints work.
5318 They are simply done by deposit/examine operations from @value{GDBN}.
5320 @subsection ROM Monitor Interface
5322 @subsection Custom Protocols
5324 @subsection Transport Layer
5326 @subsection Builtin Simulator
5329 @node Native Debugging
5331 @chapter Native Debugging
5332 @cindex native debugging
5334 Several files control @value{GDBN}'s configuration for native support:
5338 @item gdb/config/@var{arch}/@var{xyz}.mh
5339 Specifies Makefile fragments needed by a @emph{native} configuration on
5340 machine @var{xyz}. In particular, this lists the required
5341 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
5342 Also specifies the header file which describes native support on
5343 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
5344 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
5345 @samp{NAT_CDEPS}, @samp{NAT_GENERATED_FILES}, etc.; see @file{Makefile.in}.
5347 @emph{Maintainer's note: The @file{.mh} suffix is because this file
5348 originally contained @file{Makefile} fragments for hosting @value{GDBN}
5349 on machine @var{xyz}. While the file is no longer used for this
5350 purpose, the @file{.mh} suffix remains. Perhaps someone will
5351 eventually rename these fragments so that they have a @file{.mn}
5354 @item gdb/config/@var{arch}/nm-@var{xyz}.h
5355 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
5356 macro definitions describing the native system environment, such as
5357 child process control and core file support.
5359 @item gdb/@var{xyz}-nat.c
5360 Contains any miscellaneous C code required for this native support of
5361 this machine. On some machines it doesn't exist at all.
5364 There are some ``generic'' versions of routines that can be used by
5365 various systems. These can be customized in various ways by macros
5366 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
5367 the @var{xyz} host, you can just include the generic file's name (with
5368 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
5370 Otherwise, if your machine needs custom support routines, you will need
5371 to write routines that perform the same functions as the generic file.
5372 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
5373 into @code{NATDEPFILES}.
5377 This contains the @emph{target_ops vector} that supports Unix child
5378 processes on systems which use ptrace and wait to control the child.
5381 This contains the @emph{target_ops vector} that supports Unix child
5382 processes on systems which use /proc to control the child.
5385 This does the low-level grunge that uses Unix system calls to do a ``fork
5386 and exec'' to start up a child process.
5389 This is the low level interface to inferior processes for systems using
5390 the Unix @code{ptrace} call in a vanilla way.
5399 @section shared libraries
5401 @section Native Conditionals
5402 @cindex native conditionals
5404 When @value{GDBN} is configured and compiled, various macros are
5405 defined or left undefined, to control compilation when the host and
5406 target systems are the same. These macros should be defined (or left
5407 undefined) in @file{nm-@var{system}.h}.
5411 @item I386_USE_GENERIC_WATCHPOINTS
5412 An x86-based machine can define this to use the generic x86 watchpoint
5413 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
5415 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
5417 Define this to expand into an expression that will cause the symbols in
5418 @var{filename} to be added to @value{GDBN}'s symbol table. If
5419 @var{readsyms} is zero symbols are not read but any necessary low level
5420 processing for @var{filename} is still done.
5422 @item SOLIB_CREATE_INFERIOR_HOOK
5423 @findex SOLIB_CREATE_INFERIOR_HOOK
5424 Define this to expand into any shared-library-relocation code that you
5425 want to be run just after the child process has been forked.
5427 @item START_INFERIOR_TRAPS_EXPECTED
5428 @findex START_INFERIOR_TRAPS_EXPECTED
5429 When starting an inferior, @value{GDBN} normally expects to trap
5431 the shell execs, and once when the program itself execs. If the actual
5432 number of traps is something other than 2, then define this macro to
5433 expand into the number expected.
5437 @node Support Libraries
5439 @chapter Support Libraries
5444 BFD provides support for @value{GDBN} in several ways:
5447 @item identifying executable and core files
5448 BFD will identify a variety of file types, including a.out, coff, and
5449 several variants thereof, as well as several kinds of core files.
5451 @item access to sections of files
5452 BFD parses the file headers to determine the names, virtual addresses,
5453 sizes, and file locations of all the various named sections in files
5454 (such as the text section or the data section). @value{GDBN} simply
5455 calls BFD to read or write section @var{x} at byte offset @var{y} for
5458 @item specialized core file support
5459 BFD provides routines to determine the failing command name stored in a
5460 core file, the signal with which the program failed, and whether a core
5461 file matches (i.e.@: could be a core dump of) a particular executable
5464 @item locating the symbol information
5465 @value{GDBN} uses an internal interface of BFD to determine where to find the
5466 symbol information in an executable file or symbol-file. @value{GDBN} itself
5467 handles the reading of symbols, since BFD does not ``understand'' debug
5468 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
5473 @cindex opcodes library
5475 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
5476 library because it's also used in binutils, for @file{objdump}).
5479 @cindex readline library
5480 The @code{readline} library provides a set of functions for use by applications
5481 that allow users to edit command lines as they are typed in.
5484 @cindex @code{libiberty} library
5486 The @code{libiberty} library provides a set of functions and features
5487 that integrate and improve on functionality found in modern operating
5488 systems. Broadly speaking, such features can be divided into three
5489 groups: supplemental functions (functions that may be missing in some
5490 environments and operating systems), replacement functions (providing
5491 a uniform and easier to use interface for commonly used standard
5492 functions), and extensions (which provide additional functionality
5493 beyond standard functions).
5495 @value{GDBN} uses various features provided by the @code{libiberty}
5496 library, for instance the C@t{++} demangler, the @acronym{IEEE}
5497 floating format support functions, the input options parser
5498 @samp{getopt}, the @samp{obstack} extension, and other functions.
5500 @subsection @code{obstacks} in @value{GDBN}
5501 @cindex @code{obstacks}
5503 The obstack mechanism provides a convenient way to allocate and free
5504 chunks of memory. Each obstack is a pool of memory that is managed
5505 like a stack. Objects (of any nature, size and alignment) are
5506 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
5507 @code{libiberty}'s documentation for a more detailed explanation of
5510 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
5511 object files. There is an obstack associated with each internal
5512 representation of an object file. Lots of things get allocated on
5513 these @code{obstacks}: dictionary entries, blocks, blockvectors,
5514 symbols, minimal symbols, types, vectors of fundamental types, class
5515 fields of types, object files section lists, object files section
5516 offset lists, line tables, symbol tables, partial symbol tables,
5517 string tables, symbol table private data, macros tables, debug
5518 information sections and entries, import and export lists (som),
5519 unwind information (hppa), dwarf2 location expressions data. Plus
5520 various strings such as directory names strings, debug format strings,
5523 An essential and convenient property of all data on @code{obstacks} is
5524 that memory for it gets allocated (with @code{obstack_alloc}) at
5525 various times during a debugging session, but it is released all at
5526 once using the @code{obstack_free} function. The @code{obstack_free}
5527 function takes a pointer to where in the stack it must start the
5528 deletion from (much like the cleanup chains have a pointer to where to
5529 start the cleanups). Because of the stack like structure of the
5530 @code{obstacks}, this allows to free only a top portion of the
5531 obstack. There are a few instances in @value{GDBN} where such thing
5532 happens. Calls to @code{obstack_free} are done after some local data
5533 is allocated to the obstack. Only the local data is deleted from the
5534 obstack. Of course this assumes that nothing between the
5535 @code{obstack_alloc} and the @code{obstack_free} allocates anything
5536 else on the same obstack. For this reason it is best and safest to
5537 use temporary @code{obstacks}.
5539 Releasing the whole obstack is also not safe per se. It is safe only
5540 under the condition that we know the @code{obstacks} memory is no
5541 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
5542 when we get rid of the whole objfile(s), for instance upon reading a
5546 @cindex regular expressions library
5557 @item SIGN_EXTEND_CHAR
5559 @item SWITCH_ENUM_BUG
5568 @section Array Containers
5569 @cindex Array Containers
5572 Often it is necessary to manipulate a dynamic array of a set of
5573 objects. C forces some bookkeeping on this, which can get cumbersome
5574 and repetitive. The @file{vec.h} file contains macros for defining
5575 and using a typesafe vector type. The functions defined will be
5576 inlined when compiling, and so the abstraction cost should be zero.
5577 Domain checks are added to detect programming errors.
5579 An example use would be an array of symbols or section information.
5580 The array can be grown as symbols are read in (or preallocated), and
5581 the accessor macros provided keep care of all the necessary
5582 bookkeeping. Because the arrays are type safe, there is no danger of
5583 accidentally mixing up the contents. Think of these as C++ templates,
5584 but implemented in C.
5586 Because of the different behavior of structure objects, scalar objects
5587 and of pointers, there are three flavors of vector, one for each of
5588 these variants. Both the structure object and pointer variants pass
5589 pointers to objects around --- in the former case the pointers are
5590 stored into the vector and in the latter case the pointers are
5591 dereferenced and the objects copied into the vector. The scalar
5592 object variant is suitable for @code{int}-like objects, and the vector
5593 elements are returned by value.
5595 There are both @code{index} and @code{iterate} accessors. The iterator
5596 returns a boolean iteration condition and updates the iteration
5597 variable passed by reference. Because the iterator will be inlined,
5598 the address-of can be optimized away.
5600 The vectors are implemented using the trailing array idiom, thus they
5601 are not resizeable without changing the address of the vector object
5602 itself. This means you cannot have variables or fields of vector type
5603 --- always use a pointer to a vector. The one exception is the final
5604 field of a structure, which could be a vector type. You will have to
5605 use the @code{embedded_size} & @code{embedded_init} calls to create
5606 such objects, and they will probably not be resizeable (so don't use
5607 the @dfn{safe} allocation variants). The trailing array idiom is used
5608 (rather than a pointer to an array of data), because, if we allow
5609 @code{NULL} to also represent an empty vector, empty vectors occupy
5610 minimal space in the structure containing them.
5612 Each operation that increases the number of active elements is
5613 available in @dfn{quick} and @dfn{safe} variants. The former presumes
5614 that there is sufficient allocated space for the operation to succeed
5615 (it dies if there is not). The latter will reallocate the vector, if
5616 needed. Reallocation causes an exponential increase in vector size.
5617 If you know you will be adding N elements, it would be more efficient
5618 to use the reserve operation before adding the elements with the
5619 @dfn{quick} operation. This will ensure there are at least as many
5620 elements as you ask for, it will exponentially increase if there are
5621 too few spare slots. If you want reserve a specific number of slots,
5622 but do not want the exponential increase (for instance, you know this
5623 is the last allocation), use a negative number for reservation. You
5624 can also create a vector of a specific size from the get go.
5626 You should prefer the push and pop operations, as they append and
5627 remove from the end of the vector. If you need to remove several items
5628 in one go, use the truncate operation. The insert and remove
5629 operations allow you to change elements in the middle of the vector.
5630 There are two remove operations, one which preserves the element
5631 ordering @code{ordered_remove}, and one which does not
5632 @code{unordered_remove}. The latter function copies the end element
5633 into the removed slot, rather than invoke a memmove operation. The
5634 @code{lower_bound} function will determine where to place an item in
5635 the array using insert that will maintain sorted order.
5637 If you need to directly manipulate a vector, then the @code{address}
5638 accessor will return the address of the start of the vector. Also the
5639 @code{space} predicate will tell you whether there is spare capacity in the
5640 vector. You will not normally need to use these two functions.
5642 Vector types are defined using a
5643 @code{DEF_VEC_@{O,P,I@}(@var{typename})} macro. Variables of vector
5644 type are declared using a @code{VEC(@var{typename})} macro. The
5645 characters @code{O}, @code{P} and @code{I} indicate whether
5646 @var{typename} is an object (@code{O}), pointer (@code{P}) or integral
5647 (@code{I}) type. Be careful to pick the correct one, as you'll get an
5648 awkward and inefficient API if you use the wrong one. There is a
5649 check, which results in a compile-time warning, for the @code{P} and
5650 @code{I} versions, but there is no check for the @code{O} versions, as
5651 that is not possible in plain C.
5653 An example of their use would be,
5656 DEF_VEC_P(tree); // non-managed tree vector.
5659 VEC(tree) *v; // A (pointer to) a vector of tree pointers.
5662 struct my_struct *s;
5664 if (VEC_length(tree, s->v)) @{ we have some contents @}
5665 VEC_safe_push(tree, s->v, decl); // append some decl onto the end
5666 for (ix = 0; VEC_iterate(tree, s->v, ix, elt); ix++)
5667 @{ do something with elt @}
5671 The @file{vec.h} file provides details on how to invoke the various
5672 accessors provided. They are enumerated here:
5676 Return the number of items in the array,
5679 Return true if the array has no elements.
5683 Return the last or arbitrary item in the array.
5686 Access an array element and indicate whether the array has been
5691 Create and destroy an array.
5693 @item VEC_embedded_size
5694 @itemx VEC_embedded_init
5695 Helpers for embedding an array as the final element of another struct.
5701 Return the amount of free space in an array.
5704 Ensure a certain amount of free space.
5706 @item VEC_quick_push
5707 @itemx VEC_safe_push
5708 Append to an array, either assuming the space is available, or making
5712 Remove the last item from an array.
5715 Remove several items from the end of an array.
5718 Add several items to the end of an array.
5721 Overwrite an item in the array.
5723 @item VEC_quick_insert
5724 @itemx VEC_safe_insert
5725 Insert an item into the middle of the array. Either the space must
5726 already exist, or the space is created.
5728 @item VEC_ordered_remove
5729 @itemx VEC_unordered_remove
5730 Remove an item from the array, preserving order or not.
5732 @item VEC_block_remove
5733 Remove a set of items from the array.
5736 Provide the address of the first element.
5738 @item VEC_lower_bound
5739 Binary search the array.
5745 @node Coding Standards
5747 @chapter Coding Standards
5748 @cindex coding standards
5750 @section @value{GDBN} C Coding Standards
5752 @value{GDBN} follows the GNU coding standards, as described in
5753 @file{etc/standards.texi}. This file is also available for anonymous
5754 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
5755 of the standard; in general, when the GNU standard recommends a practice
5756 but does not require it, @value{GDBN} requires it.
5758 @value{GDBN} follows an additional set of coding standards specific to
5759 @value{GDBN}, as described in the following sections.
5763 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
5766 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
5768 @subsection Formatting
5770 @cindex source code formatting
5771 The standard GNU recommendations for formatting must be followed
5772 strictly. Any @value{GDBN}-specific deviation from GNU
5773 recomendations is described below.
5775 A function declaration should not have its name in column zero. A
5776 function definition should have its name in column zero.
5780 static void foo (void);
5788 @emph{Pragmatics: This simplifies scripting. Function definitions can
5789 be found using @samp{^function-name}.}
5791 There must be a space between a function or macro name and the opening
5792 parenthesis of its argument list (except for macro definitions, as
5793 required by C). There must not be a space after an open paren/bracket
5794 or before a close paren/bracket.
5796 While additional whitespace is generally helpful for reading, do not use
5797 more than one blank line to separate blocks, and avoid adding whitespace
5798 after the end of a program line (as of 1/99, some 600 lines had
5799 whitespace after the semicolon). Excess whitespace causes difficulties
5800 for @code{diff} and @code{patch} utilities.
5802 Pointers are declared using the traditional K&R C style:
5816 In addition, whitespace around casts and unary operators should follow
5817 the following guidelines:
5819 @multitable @columnfractions .2 .2 .8
5820 @item Use... @tab ...instead of @tab
5829 @item @code{(foo) x}
5834 @tab (pointer dereference)
5837 Any two or more lines in code should be wrapped in braces, even if
5838 they are comments, as they look like separate statements:
5843 /* Return success. */
5853 /* Return success. */
5857 @subsection Comments
5859 @cindex comment formatting
5860 The standard GNU requirements on comments must be followed strictly.
5862 Block comments must appear in the following form, with no @code{/*}- or
5863 @code{*/}-only lines, and no leading @code{*}:
5866 /* Wait for control to return from inferior to debugger. If inferior
5867 gets a signal, we may decide to start it up again instead of
5868 returning. That is why there is a loop in this function. When
5869 this function actually returns it means the inferior should be left
5870 stopped and @value{GDBN} should read more commands. */
5873 (Note that this format is encouraged by Emacs; tabbing for a multi-line
5874 comment works correctly, and @kbd{M-q} fills the block consistently.)
5876 Put a blank line between the block comments preceding function or
5877 variable definitions, and the definition itself.
5879 In general, put function-body comments on lines by themselves, rather
5880 than trying to fit them into the 20 characters left at the end of a
5881 line, since either the comment or the code will inevitably get longer
5882 than will fit, and then somebody will have to move it anyhow.
5886 @cindex C data types
5887 Code must not depend on the sizes of C data types, the format of the
5888 host's floating point numbers, the alignment of anything, or the order
5889 of evaluation of expressions.
5891 @cindex function usage
5892 Use functions freely. There are only a handful of compute-bound areas
5893 in @value{GDBN} that might be affected by the overhead of a function
5894 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
5895 limited by the target interface (whether serial line or system call).
5897 However, use functions with moderation. A thousand one-line functions
5898 are just as hard to understand as a single thousand-line function.
5900 @emph{Macros are bad, M'kay.}
5901 (But if you have to use a macro, make sure that the macro arguments are
5902 protected with parentheses.)
5906 Declarations like @samp{struct foo *} should be used in preference to
5907 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
5909 Zero constant (@code{0}) is not interchangeable with a null pointer
5910 constant (@code{NULL}) anywhere. @sc{gcc} does not give a warning for
5911 such interchange. Specifically:
5913 @multitable @columnfractions .2 .5
5915 @tab @code{if (pointervar) @{@}}
5917 @tab @code{if (!pointervar) @{@}}
5919 @tab @code{if (pointervar != 0) @{@}}
5921 @tab @code{if (pointervar == 0) @{@}}
5923 @tab @code{if (pointervar != NULL) @{@}}
5925 @tab @code{if (pointervar == NULL) @{@}}
5928 @subsection Function Prototypes
5929 @cindex function prototypes
5931 Prototypes must be used when both @emph{declaring} and @emph{defining}
5932 a function. Prototypes for @value{GDBN} functions must include both the
5933 argument type and name, with the name matching that used in the actual
5934 function definition.
5936 All external functions should have a declaration in a header file that
5937 callers include, that declaration should use the @code{extern} modifier.
5938 The only exception concerns @code{_initialize_*} functions, which must
5939 be external so that @file{init.c} construction works, but shouldn't be
5940 visible to random source files.
5942 Where a source file needs a forward declaration of a static function,
5943 that declaration must appear in a block near the top of the source file.
5945 @subsection File Names
5947 Any file used when building the core of @value{GDBN} must be in lower
5948 case. Any file used when building the core of @value{GDBN} must be 8.3
5949 unique. These requirements apply to both source and generated files.
5951 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
5952 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
5953 is introduced to the build process both @file{Makefile.in} and
5954 @file{configure.in} need to be modified accordingly. Compare the
5955 convoluted conversion process needed to transform @file{COPYING} into
5956 @file{copying.c} with the conversion needed to transform
5957 @file{version.in} into @file{version.c}.}
5959 Any file non 8.3 compliant file (that is not used when building the core
5960 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
5962 @emph{Pragmatics: This is clearly a compromise.}
5964 When @value{GDBN} has a local version of a system header file (ex
5965 @file{string.h}) the file name based on the POSIX header prefixed with
5966 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
5967 independent: they should use only macros defined by @file{configure},
5968 the compiler, or the host; they should include only system headers; they
5969 should refer only to system types. They may be shared between multiple
5970 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
5972 For other files @samp{-} is used as the separator.
5974 @subsection Include Files
5976 A @file{.c} file should include @file{defs.h} first.
5978 A @file{.c} file should directly include the @code{.h} file of every
5979 declaration and/or definition it directly refers to. It cannot rely on
5982 A @file{.h} file should directly include the @code{.h} file of every
5983 declaration and/or definition it directly refers to. It cannot rely on
5984 indirect inclusion. Exception: The file @file{defs.h} does not need to
5985 be directly included.
5987 An external declaration should only appear in one include file.
5989 An external declaration should never appear in a @code{.c} file.
5990 Exception: a declaration for the @code{_initialize} function that
5991 pacifies @option{-Wmissing-declaration}.
5993 A @code{typedef} definition should only appear in one include file.
5995 An opaque @code{struct} declaration can appear in multiple @file{.h}
5996 files. Where possible, a @file{.h} file should use an opaque
5997 @code{struct} declaration instead of an include.
5999 All @file{.h} files should be wrapped in:
6002 #ifndef INCLUDE_FILE_NAME_H
6003 #define INCLUDE_FILE_NAME_H
6008 @section @value{GDBN} Python Coding Standards
6010 @value{GDBN} follows the published @code{Python} coding standards in
6011 @uref{http://www.python.org/dev/peps/pep-0008/, @code{PEP008}}.
6013 In addition, the guidelines in the
6014 @uref{http://google-styleguide.googlecode.com/svn/trunk/pyguide.html,
6015 Google Python Style Guide} are also followed where they do not
6016 conflict with @code{PEP008}.
6018 @subsection @value{GDBN}-specific exceptions
6020 There are a few exceptions to the published standards.
6021 They exist mainly for consistency with the @code{C} standards.
6023 @c It is expected that there are a few more exceptions,
6024 @c so we use itemize here.
6029 Use @code{FIXME} instead of @code{TODO}.
6033 @node Misc Guidelines
6035 @chapter Misc Guidelines
6037 This chapter covers topics that are lower-level than the major
6038 algorithms of @value{GDBN}.
6043 Cleanups are a structured way to deal with things that need to be done
6046 When your code does something (e.g., @code{xmalloc} some memory, or
6047 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
6048 the memory or @code{close} the file), it can make a cleanup. The
6049 cleanup will be done at some future point: when the command is finished
6050 and control returns to the top level; when an error occurs and the stack
6051 is unwound; or when your code decides it's time to explicitly perform
6052 cleanups. Alternatively you can elect to discard the cleanups you
6058 @item struct cleanup *@var{old_chain};
6059 Declare a variable which will hold a cleanup chain handle.
6061 @findex make_cleanup
6062 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
6063 Make a cleanup which will cause @var{function} to be called with
6064 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
6065 handle that can later be passed to @code{do_cleanups} or
6066 @code{discard_cleanups}. Unless you are going to call
6067 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
6068 from @code{make_cleanup}.
6071 @item do_cleanups (@var{old_chain});
6072 Do all cleanups added to the chain since the corresponding
6073 @code{make_cleanup} call was made.
6075 @findex discard_cleanups
6076 @item discard_cleanups (@var{old_chain});
6077 Same as @code{do_cleanups} except that it just removes the cleanups from
6078 the chain and does not call the specified functions.
6081 Cleanups are implemented as a chain. The handle returned by
6082 @code{make_cleanups} includes the cleanup passed to the call and any
6083 later cleanups appended to the chain (but not yet discarded or
6087 make_cleanup (a, 0);
6089 struct cleanup *old = make_cleanup (b, 0);
6097 will call @code{c()} and @code{b()} but will not call @code{a()}. The
6098 cleanup that calls @code{a()} will remain in the cleanup chain, and will
6099 be done later unless otherwise discarded.@refill
6101 Your function should explicitly do or discard the cleanups it creates.
6102 Failing to do this leads to non-deterministic behavior since the caller
6103 will arbitrarily do or discard your functions cleanups. This need leads
6104 to two common cleanup styles.
6106 The first style is try/finally. Before it exits, your code-block calls
6107 @code{do_cleanups} with the old cleanup chain and thus ensures that your
6108 code-block's cleanups are always performed. For instance, the following
6109 code-segment avoids a memory leak problem (even when @code{error} is
6110 called and a forced stack unwind occurs) by ensuring that the
6111 @code{xfree} will always be called:
6114 struct cleanup *old = make_cleanup (null_cleanup, 0);
6115 data = xmalloc (sizeof blah);
6116 make_cleanup (xfree, data);
6121 The second style is try/except. Before it exits, your code-block calls
6122 @code{discard_cleanups} with the old cleanup chain and thus ensures that
6123 any created cleanups are not performed. For instance, the following
6124 code segment, ensures that the file will be closed but only if there is
6128 FILE *file = fopen ("afile", "r");
6129 struct cleanup *old = make_cleanup (close_file, file);
6131 discard_cleanups (old);
6135 Some functions, e.g., @code{fputs_filtered()} or @code{error()}, specify
6136 that they ``should not be called when cleanups are not in place''. This
6137 means that any actions you need to reverse in the case of an error or
6138 interruption must be on the cleanup chain before you call these
6139 functions, since they might never return to your code (they
6140 @samp{longjmp} instead).
6142 @section Per-architecture module data
6143 @cindex per-architecture module data
6144 @cindex multi-arch data
6145 @cindex data-pointer, per-architecture/per-module
6147 The multi-arch framework includes a mechanism for adding module
6148 specific per-architecture data-pointers to the @code{struct gdbarch}
6149 architecture object.
6151 A module registers one or more per-architecture data-pointers using:
6153 @deftypefn {Architecture Function} {struct gdbarch_data *} gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
6154 @var{pre_init} is used to, on-demand, allocate an initial value for a
6155 per-architecture data-pointer using the architecture's obstack (passed
6156 in as a parameter). Since @var{pre_init} can be called during
6157 architecture creation, it is not parameterized with the architecture.
6158 and must not call modules that use per-architecture data.
6161 @deftypefn {Architecture Function} {struct gdbarch_data *} gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
6162 @var{post_init} is used to obtain an initial value for a
6163 per-architecture data-pointer @emph{after}. Since @var{post_init} is
6164 always called after architecture creation, it both receives the fully
6165 initialized architecture and is free to call modules that use
6166 per-architecture data (care needs to be taken to ensure that those
6167 other modules do not try to call back to this module as that will
6168 create in cycles in the initialization call graph).
6171 These functions return a @code{struct gdbarch_data} that is used to
6172 identify the per-architecture data-pointer added for that module.
6174 The per-architecture data-pointer is accessed using the function:
6176 @deftypefn {Architecture Function} {void *} gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
6177 Given the architecture @var{arch} and module data handle
6178 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
6179 or @code{gdbarch_data_register_post_init}), this function returns the
6180 current value of the per-architecture data-pointer. If the data
6181 pointer is @code{NULL}, it is first initialized by calling the
6182 corresponding @var{pre_init} or @var{post_init} method.
6185 The examples below assume the following definitions:
6188 struct nozel @{ int total; @};
6189 static struct gdbarch_data *nozel_handle;
6192 A module can extend the architecture vector, adding additional
6193 per-architecture data, using the @var{pre_init} method. The module's
6194 per-architecture data is then initialized during architecture
6197 In the below, the module's per-architecture @emph{nozel} is added. An
6198 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
6199 from @code{gdbarch_init}.
6203 nozel_pre_init (struct obstack *obstack)
6205 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
6212 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
6214 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
6215 data->total = nozel;
6219 A module can on-demand create architecture dependent data structures
6220 using @code{post_init}.
6222 In the below, the nozel's total is computed on-demand by
6223 @code{nozel_post_init} using information obtained from the
6228 nozel_post_init (struct gdbarch *gdbarch)
6230 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
6231 nozel->total = gdbarch@dots{} (gdbarch);
6238 nozel_total (struct gdbarch *gdbarch)
6240 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
6245 @section Wrapping Output Lines
6246 @cindex line wrap in output
6249 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
6250 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
6251 added in places that would be good breaking points. The utility
6252 routines will take care of actually wrapping if the line width is
6255 The argument to @code{wrap_here} is an indentation string which is
6256 printed @emph{only} if the line breaks there. This argument is saved
6257 away and used later. It must remain valid until the next call to
6258 @code{wrap_here} or until a newline has been printed through the
6259 @code{*_filtered} functions. Don't pass in a local variable and then
6262 It is usually best to call @code{wrap_here} after printing a comma or
6263 space. If you call it before printing a space, make sure that your
6264 indentation properly accounts for the leading space that will print if
6265 the line wraps there.
6267 Any function or set of functions that produce filtered output must
6268 finish by printing a newline, to flush the wrap buffer, before switching
6269 to unfiltered (@code{printf}) output. Symbol reading routines that
6270 print warnings are a good example.
6272 @section Memory Management
6274 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
6275 @code{calloc}, @code{free} and @code{asprintf}.
6277 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
6278 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
6279 these functions do not return when the memory pool is empty. Instead,
6280 they unwind the stack using cleanups. These functions return
6281 @code{NULL} when requested to allocate a chunk of memory of size zero.
6283 @emph{Pragmatics: By using these functions, the need to check every
6284 memory allocation is removed. These functions provide portable
6287 @value{GDBN} does not use the function @code{free}.
6289 @value{GDBN} uses the function @code{xfree} to return memory to the
6290 memory pool. Consistent with ISO-C, this function ignores a request to
6291 free a @code{NULL} pointer.
6293 @emph{Pragmatics: On some systems @code{free} fails when passed a
6294 @code{NULL} pointer.}
6296 @value{GDBN} can use the non-portable function @code{alloca} for the
6297 allocation of small temporary values (such as strings).
6299 @emph{Pragmatics: This function is very non-portable. Some systems
6300 restrict the memory being allocated to no more than a few kilobytes.}
6302 @value{GDBN} uses the string function @code{xstrdup} and the print
6303 function @code{xstrprintf}.
6305 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
6306 functions such as @code{sprintf} are very prone to buffer overflow
6310 @section Compiler Warnings
6311 @cindex compiler warnings
6313 With few exceptions, developers should avoid the configuration option
6314 @samp{--disable-werror} when building @value{GDBN}. The exceptions
6315 are listed in the file @file{gdb/MAINTAINERS}. The default, when
6316 building with @sc{gcc}, is @samp{--enable-werror}.
6318 This option causes @value{GDBN} (when built using GCC) to be compiled
6319 with a carefully selected list of compiler warning flags. Any warnings
6320 from those flags are treated as errors.
6322 The current list of warning flags includes:
6326 Recommended @sc{gcc} warnings.
6328 @item -Wdeclaration-after-statement
6330 @sc{gcc} 3.x (and later) and @sc{c99} allow declarations mixed with
6331 code, but @sc{gcc} 2.x and @sc{c89} do not.
6333 @item -Wpointer-arith
6335 @item -Wformat-nonliteral
6336 Non-literal format strings, with a few exceptions, are bugs - they
6337 might contain unintended user-supplied format specifiers.
6338 Since @value{GDBN} uses the @code{format printf} attribute on all
6339 @code{printf} like functions this checks not just @code{printf} calls
6340 but also calls to functions such as @code{fprintf_unfiltered}.
6342 @item -Wpointer-sign
6343 This helps make sure @value{GDBN} code uses @code{gdb_byte} which is
6344 really @code{unsigned char} for raw bytes instead of @code{char},
6345 whose signedness is host-dependent. @sc{gcc} enables this with
6346 @code{-Wall} since version 4.0. We enable it explicitly too to be
6347 decoupled from future @sc{gcc} (or other compiler)'s defaults.
6349 @item -Wno-unused-parameter
6350 Due to the way that @value{GDBN} is implemented many functions have
6351 unused parameters. Consequently this warning is avoided. The macro
6352 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
6353 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
6358 @itemx -Wno-char-subscripts
6359 These are warnings which might be useful for @value{GDBN}, but are
6360 currently too noisy to enable with @samp{-Werror}.
6364 @section Internal Error Recovery
6366 During its execution, @value{GDBN} can encounter two types of errors.
6367 User errors and internal errors. User errors include not only a user
6368 entering an incorrect command but also problems arising from corrupt
6369 object files and system errors when interacting with the target.
6370 Internal errors include situations where @value{GDBN} has detected, at
6371 run time, a corrupt or erroneous situation.
6373 When reporting an internal error, @value{GDBN} uses
6374 @code{internal_error} and @code{gdb_assert}.
6376 @value{GDBN} must not call @code{abort} or @code{assert}.
6378 @emph{Pragmatics: There is no @code{internal_warning} function. Either
6379 the code detected a user error, recovered from it and issued a
6380 @code{warning} or the code failed to correctly recover from the user
6381 error and issued an @code{internal_error}.}
6383 @section Command Names
6385 GDB U/I commands are written @samp{foo-bar}, not @samp{foo_bar}.
6387 @section Clean Design and Portable Implementation
6390 In addition to getting the syntax right, there's the little question of
6391 semantics. Some things are done in certain ways in @value{GDBN} because long
6392 experience has shown that the more obvious ways caused various kinds of
6395 @cindex assumptions about targets
6396 You can't assume the byte order of anything that comes from a target
6397 (including @var{value}s, object files, and instructions). Such things
6398 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
6399 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
6400 such as @code{bfd_get_32}.
6402 You can't assume that you know what interface is being used to talk to
6403 the target system. All references to the target must go through the
6404 current @code{target_ops} vector.
6406 You can't assume that the host and target machines are the same machine
6407 (except in the ``native'' support modules). In particular, you can't
6408 assume that the target machine's header files will be available on the
6409 host machine. Target code must bring along its own header files --
6410 written from scratch or explicitly donated by their owner, to avoid
6414 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
6415 to write the code portably than to conditionalize it for various
6418 @cindex system dependencies
6419 New @code{#ifdef}'s which test for specific compilers or manufacturers
6420 or operating systems are unacceptable. All @code{#ifdef}'s should test
6421 for features. The information about which configurations contain which
6422 features should be segregated into the configuration files. Experience
6423 has proven far too often that a feature unique to one particular system
6424 often creeps into other systems; and that a conditional based on some
6425 predefined macro for your current system will become worthless over
6426 time, as new versions of your system come out that behave differently
6427 with regard to this feature.
6429 Adding code that handles specific architectures, operating systems,
6430 target interfaces, or hosts, is not acceptable in generic code.
6432 @cindex portable file name handling
6433 @cindex file names, portability
6434 One particularly notorious area where system dependencies tend to
6435 creep in is handling of file names. The mainline @value{GDBN} code
6436 assumes Posix semantics of file names: absolute file names begin with
6437 a forward slash @file{/}, slashes are used to separate leading
6438 directories, case-sensitive file names. These assumptions are not
6439 necessarily true on non-Posix systems such as MS-Windows. To avoid
6440 system-dependent code where you need to take apart or construct a file
6441 name, use the following portable macros:
6444 @findex HAVE_DOS_BASED_FILE_SYSTEM
6445 @item HAVE_DOS_BASED_FILE_SYSTEM
6446 This preprocessing symbol is defined to a non-zero value on hosts
6447 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
6448 symbol to write conditional code which should only be compiled for
6451 @findex IS_DIR_SEPARATOR
6452 @item IS_DIR_SEPARATOR (@var{c})
6453 Evaluates to a non-zero value if @var{c} is a directory separator
6454 character. On Unix and GNU/Linux systems, only a slash @file{/} is
6455 such a character, but on Windows, both @file{/} and @file{\} will
6458 @findex IS_ABSOLUTE_PATH
6459 @item IS_ABSOLUTE_PATH (@var{file})
6460 Evaluates to a non-zero value if @var{file} is an absolute file name.
6461 For Unix and GNU/Linux hosts, a name which begins with a slash
6462 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
6463 @file{x:\bar} are also absolute file names.
6465 @findex FILENAME_CMP
6466 @item FILENAME_CMP (@var{f1}, @var{f2})
6467 Calls a function which compares file names @var{f1} and @var{f2} as
6468 appropriate for the underlying host filesystem. For Posix systems,
6469 this simply calls @code{strcmp}; on case-insensitive filesystems it
6470 will call @code{strcasecmp} instead.
6472 @findex DIRNAME_SEPARATOR
6473 @item DIRNAME_SEPARATOR
6474 Evaluates to a character which separates directories in
6475 @code{PATH}-style lists, typically held in environment variables.
6476 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
6478 @findex SLASH_STRING
6480 This evaluates to a constant string you should use to produce an
6481 absolute filename from leading directories and the file's basename.
6482 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
6483 @code{"\\"} for some Windows-based ports.
6486 In addition to using these macros, be sure to use portable library
6487 functions whenever possible. For example, to extract a directory or a
6488 basename part from a file name, use the @code{dirname} and
6489 @code{basename} library functions (available in @code{libiberty} for
6490 platforms which don't provide them), instead of searching for a slash
6491 with @code{strrchr}.
6493 Another way to generalize @value{GDBN} along a particular interface is with an
6494 attribute struct. For example, @value{GDBN} has been generalized to handle
6495 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
6496 by defining the @code{target_ops} structure and having a current target (as
6497 well as a stack of targets below it, for memory references). Whenever
6498 something needs to be done that depends on which remote interface we are
6499 using, a flag in the current target_ops structure is tested (e.g.,
6500 @code{target_has_stack}), or a function is called through a pointer in the
6501 current target_ops structure. In this way, when a new remote interface
6502 is added, only one module needs to be touched---the one that actually
6503 implements the new remote interface. Other examples of
6504 attribute-structs are BFD access to multiple kinds of object file
6505 formats, or @value{GDBN}'s access to multiple source languages.
6507 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
6508 the code interfacing between @code{ptrace} and the rest of
6509 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
6510 something was very painful. In @value{GDBN} 4.x, these have all been
6511 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
6512 with variations between systems the same way any system-independent
6513 file would (hooks, @code{#if defined}, etc.), and machines which are
6514 radically different don't need to use @file{infptrace.c} at all.
6516 All debugging code must be controllable using the @samp{set debug
6517 @var{module}} command. Do not use @code{printf} to print trace
6518 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
6519 @code{#ifdef DEBUG}.
6523 @chapter Porting @value{GDBN}
6524 @cindex porting to new machines
6526 Most of the work in making @value{GDBN} compile on a new machine is in
6527 specifying the configuration of the machine. Porting a new
6528 architecture to @value{GDBN} can be broken into a number of steps.
6533 Ensure a @sc{bfd} exists for executables of the target architecture in
6534 the @file{bfd} directory. If one does not exist, create one by
6535 modifying an existing similar one.
6538 Implement a disassembler for the target architecture in the @file{opcodes}
6542 Define the target architecture in the @file{gdb} directory
6543 (@pxref{Adding a New Target, , Adding a New Target}). Add the pattern
6544 for the new target to @file{configure.tgt} with the names of the files
6545 that contain the code. By convention the target architecture
6546 definition for an architecture @var{arch} is placed in
6547 @file{@var{arch}-tdep.c}.
6549 Within @file{@var{arch}-tdep.c} define the function
6550 @code{_initialize_@var{arch}_tdep} which calls
6551 @code{gdbarch_register} to create the new @code{@w{struct
6552 gdbarch}} for the architecture.
6555 If a new remote target is needed, consider adding a new remote target
6556 by defining a function
6557 @code{_initialize_remote_@var{arch}}. However if at all possible
6558 use the @value{GDBN} @emph{Remote Serial Protocol} for this and implement
6559 the server side protocol independently with the target.
6562 If desired implement a simulator in the @file{sim} directory. This
6563 should create the library @file{libsim.a} implementing the interface
6564 in @file{remote-sim.h} (found in the @file{include} directory).
6567 Build and test. If desired, lobby the @sc{gdb} steering group to
6568 have the new port included in the main distribution!
6571 Add a description of the new architecture to the main @value{GDBN} user
6572 guide (@pxref{Configuration Specific Information, , Configuration
6573 Specific Information, gdb, Debugging with @value{GDBN}}).
6577 @node Versions and Branches
6578 @chapter Versions and Branches
6582 @value{GDBN}'s version is determined by the file
6583 @file{gdb/version.in} and takes one of the following forms:
6586 @item @var{major}.@var{minor}
6587 @itemx @var{major}.@var{minor}.@var{patchlevel}
6588 an official release (e.g., 6.2 or 6.2.1)
6589 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}
6590 a snapshot taken at @var{YYYY}-@var{MM}-@var{DD}-gmt (e.g.,
6591 6.1.50.20020302, 6.1.90.20020304, or 6.1.0.20020308)
6592 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}-cvs
6593 a @sc{cvs} check out drawn on @var{YYYY}-@var{MM}-@var{DD} (e.g.,
6594 6.1.50.20020302-cvs, 6.1.90.20020304-cvs, or 6.1.0.20020308-cvs)
6595 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD} (@var{vendor})
6596 a vendor specific release of @value{GDBN}, that while based on@*
6597 @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD},
6598 may include additional changes
6601 @value{GDBN}'s mainline uses the @var{major} and @var{minor} version
6602 numbers from the most recent release branch, with a @var{patchlevel}
6603 of 50. At the time each new release branch is created, the mainline's
6604 @var{major} and @var{minor} version numbers are updated.
6606 @value{GDBN}'s release branch is similar. When the branch is cut, the
6607 @var{patchlevel} is changed from 50 to 90. As draft releases are
6608 drawn from the branch, the @var{patchlevel} is incremented. Once the
6609 first release (@var{major}.@var{minor}) has been made, the
6610 @var{patchlevel} is set to 0 and updates have an incremented
6613 For snapshots, and @sc{cvs} check outs, it is also possible to
6614 identify the @sc{cvs} origin:
6617 @item @var{major}.@var{minor}.50.@var{YYYY}@var{MM}@var{DD}
6618 drawn from the @sc{head} of mainline @sc{cvs} (e.g., 6.1.50.20020302)
6619 @item @var{major}.@var{minor}.90.@var{YYYY}@var{MM}@var{DD}
6620 @itemx @var{major}.@var{minor}.91.@var{YYYY}@var{MM}@var{DD} @dots{}
6621 drawn from a release branch prior to the release (e.g.,
6623 @item @var{major}.@var{minor}.0.@var{YYYY}@var{MM}@var{DD}
6624 @itemx @var{major}.@var{minor}.1.@var{YYYY}@var{MM}@var{DD} @dots{}
6625 drawn from a release branch after the release (e.g., 6.2.0.20020308)
6628 If the previous @value{GDBN} version is 6.1 and the current version is
6629 6.2, then, substituting 6 for @var{major} and 1 or 2 for @var{minor},
6630 here's an illustration of a typical sequence:
6637 +--------------------------.
6640 6.2.50.20020303-cvs 6.1.90 (draft #1)
6642 6.2.50.20020304-cvs 6.1.90.20020304-cvs
6644 6.2.50.20020305-cvs 6.1.91 (draft #2)
6646 6.2.50.20020306-cvs 6.1.91.20020306-cvs
6648 6.2.50.20020307-cvs 6.2 (release)
6650 6.2.50.20020308-cvs 6.2.0.20020308-cvs
6652 6.2.50.20020309-cvs 6.2.1 (update)
6654 6.2.50.20020310-cvs <branch closed>
6658 +--------------------------.
6661 6.3.50.20020312-cvs 6.2.90 (draft #1)
6665 @section Release Branches
6666 @cindex Release Branches
6668 @value{GDBN} draws a release series (6.2, 6.2.1, @dots{}) from a
6669 single release branch, and identifies that branch using the @sc{cvs}
6673 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-branchpoint
6674 gdb_@var{major}_@var{minor}-branch
6675 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-release
6678 @emph{Pragmatics: To help identify the date at which a branch or
6679 release is made, both the branchpoint and release tags include the
6680 date that they are cut (@var{YYYY}@var{MM}@var{DD}) in the tag. The
6681 branch tag, denoting the head of the branch, does not need this.}
6683 @section Vendor Branches
6684 @cindex vendor branches
6686 To avoid version conflicts, vendors are expected to modify the file
6687 @file{gdb/version.in} to include a vendor unique alphabetic identifier
6688 (an official @value{GDBN} release never uses alphabetic characters in
6689 its version identifier). E.g., @samp{6.2widgit2}, or @samp{6.2 (Widgit
6692 @section Experimental Branches
6693 @cindex experimental branches
6695 @subsection Guidelines
6697 @value{GDBN} permits the creation of branches, cut from the @sc{cvs}
6698 repository, for experimental development. Branches make it possible
6699 for developers to share preliminary work, and maintainers to examine
6700 significant new developments.
6702 The following are a set of guidelines for creating such branches:
6706 @item a branch has an owner
6707 The owner can set further policy for a branch, but may not change the
6708 ground rules. In particular, they can set a policy for commits (be it
6709 adding more reviewers or deciding who can commit).
6711 @item all commits are posted
6712 All changes committed to a branch shall also be posted to
6713 @email{gdb-patches@@sourceware.org, the @value{GDBN} patches
6714 mailing list}. While commentary on such changes are encouraged, people
6715 should remember that the changes only apply to a branch.
6717 @item all commits are covered by an assignment
6718 This ensures that all changes belong to the Free Software Foundation,
6719 and avoids the possibility that the branch may become contaminated.
6721 @item a branch is focused
6722 A focused branch has a single objective or goal, and does not contain
6723 unnecessary or irrelevant changes. Cleanups, where identified, being
6724 be pushed into the mainline as soon as possible.
6726 @item a branch tracks mainline
6727 This keeps the level of divergence under control. It also keeps the
6728 pressure on developers to push cleanups and other stuff into the
6731 @item a branch shall contain the entire @value{GDBN} module
6732 The @value{GDBN} module @code{gdb} should be specified when creating a
6733 branch (branches of individual files should be avoided). @xref{Tags}.
6735 @item a branch shall be branded using @file{version.in}
6736 The file @file{gdb/version.in} shall be modified so that it identifies
6737 the branch @var{owner} and branch @var{name}, e.g.,
6738 @samp{6.2.50.20030303_owner_name} or @samp{6.2 (Owner Name)}.
6745 To simplify the identification of @value{GDBN} branches, the following
6746 branch tagging convention is strongly recommended:
6750 @item @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
6751 @itemx @var{owner}_@var{name}-@var{YYYYMMDD}-branch
6752 The branch point and corresponding branch tag. @var{YYYYMMDD} is the
6753 date that the branch was created. A branch is created using the
6754 sequence: @anchor{experimental branch tags}
6756 cvs rtag @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint gdb
6757 cvs rtag -b -r @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint \
6758 @var{owner}_@var{name}-@var{YYYYMMDD}-branch gdb
6761 @item @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
6762 The tagged point, on the mainline, that was used when merging the branch
6763 on @var{yyyymmdd}. To merge in all changes since the branch was cut,
6764 use a command sequence like:
6766 cvs rtag @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint gdb
6768 -j@var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
6769 -j@var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
6772 Similar sequences can be used to just merge in changes since the last
6778 For further information on @sc{cvs}, see
6779 @uref{http://www.gnu.org/software/cvs/, Concurrent Versions System}.
6781 @node Start of New Year Procedure
6782 @chapter Start of New Year Procedure
6783 @cindex new year procedure
6785 At the start of each new year, the following actions should be performed:
6789 Rotate the ChangeLog file
6791 The current @file{ChangeLog} file should be renamed into
6792 @file{ChangeLog-YYYY} where YYYY is the year that has just passed.
6793 A new @file{ChangeLog} file should be created, and its contents should
6794 contain a reference to the previous ChangeLog. The following should
6795 also be preserved at the end of the new ChangeLog, in order to provide
6796 the appropriate settings when editing this file with Emacs:
6802 version-control: never
6808 Add an entry for the newly created ChangeLog file (@file{ChangeLog-YYYY})
6809 in @file{gdb/config/djgpp/fnchange.lst}.
6812 Update the copyright year in the startup message
6814 Update the copyright year in:
6817 file @file{top.c}, function @code{print_gdb_version}
6819 file @file{gdbserver/server.c}, function @code{gdbserver_version}
6821 file @file{gdbserver/gdbreplay.c}, function @code{gdbreplay_version}
6825 Run the @file{copyright.py} Python script to add the new year in the copyright
6826 notices of most source files. This script has been tested with Python
6833 @chapter Releasing @value{GDBN}
6834 @cindex making a new release of gdb
6836 @section Branch Commit Policy
6838 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
6839 5.1 and 5.2 all used the below:
6843 The @file{gdb/MAINTAINERS} file still holds.
6845 Don't fix something on the branch unless/until it is also fixed in the
6846 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
6847 file is better than committing a hack.
6849 When considering a patch for the branch, suggested criteria include:
6850 Does it fix a build? Does it fix the sequence @kbd{break main; run}
6851 when debugging a static binary?
6853 The further a change is from the core of @value{GDBN}, the less likely
6854 the change will worry anyone (e.g., target specific code).
6856 Only post a proposal to change the core of @value{GDBN} after you've
6857 sent individual bribes to all the people listed in the
6858 @file{MAINTAINERS} file @t{;-)}
6861 @emph{Pragmatics: Provided updates are restricted to non-core
6862 functionality there is little chance that a broken change will be fatal.
6863 This means that changes such as adding a new architectures or (within
6864 reason) support for a new host are considered acceptable.}
6867 @section Obsoleting code
6869 Before anything else, poke the other developers (and around the source
6870 code) to see if there is anything that can be removed from @value{GDBN}
6871 (an old target, an unused file).
6873 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
6874 line. Doing this means that it is easy to identify something that has
6875 been obsoleted when greping through the sources.
6877 The process is done in stages --- this is mainly to ensure that the
6878 wider @value{GDBN} community has a reasonable opportunity to respond.
6879 Remember, everything on the Internet takes a week.
6883 Post the proposal on @email{gdb@@sourceware.org, the GDB mailing
6884 list} Creating a bug report to track the task's state, is also highly
6889 Post the proposal on @email{gdb-announce@@sourceware.org, the GDB
6890 Announcement mailing list}.
6894 Go through and edit all relevant files and lines so that they are
6895 prefixed with the word @code{OBSOLETE}.
6897 Wait until the next GDB version, containing this obsolete code, has been
6900 Remove the obsolete code.
6904 @emph{Maintainer note: While removing old code is regrettable it is
6905 hopefully better for @value{GDBN}'s long term development. Firstly it
6906 helps the developers by removing code that is either no longer relevant
6907 or simply wrong. Secondly since it removes any history associated with
6908 the file (effectively clearing the slate) the developer has a much freer
6909 hand when it comes to fixing broken files.}
6913 @section Before the Branch
6915 The most important objective at this stage is to find and fix simple
6916 changes that become a pain to track once the branch is created. For
6917 instance, configuration problems that stop @value{GDBN} from even
6918 building. If you can't get the problem fixed, document it in the
6919 @file{gdb/PROBLEMS} file.
6921 @subheading Prompt for @file{gdb/NEWS}
6923 People always forget. Send a post reminding them but also if you know
6924 something interesting happened add it yourself. The @code{schedule}
6925 script will mention this in its e-mail.
6927 @subheading Review @file{gdb/README}
6929 Grab one of the nightly snapshots and then walk through the
6930 @file{gdb/README} looking for anything that can be improved. The
6931 @code{schedule} script will mention this in its e-mail.
6933 @subheading Refresh any imported files.
6935 A number of files are taken from external repositories. They include:
6939 @file{texinfo/texinfo.tex}
6941 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
6944 @file{etc/standards.texi}, @file{etc/make-stds.texi}
6947 @subheading Check the ARI
6949 @uref{http://sourceware.org/gdb/ari,,A.R.I.} is an @code{awk} script
6950 (Awk Regression Index ;-) that checks for a number of errors and coding
6951 conventions. The checks include things like using @code{malloc} instead
6952 of @code{xmalloc} and file naming problems. There shouldn't be any
6955 @subsection Review the bug data base
6957 Close anything obviously fixed.
6959 @subsection Check all cross targets build
6961 The targets are listed in @file{gdb/MAINTAINERS}.
6964 @section Cut the Branch
6966 @subheading Create the branch
6971 $ V=`echo $v | sed 's/\./_/g'`
6972 $ D=`date -u +%Y-%m-%d`
6975 $ echo cvs -f -d :ext:sourceware.org:/cvs/src rtag \
6976 -D $D-gmt gdb_$V-$D-branchpoint insight
6977 cvs -f -d :ext:sourceware.org:/cvs/src rtag
6978 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight
6981 $ echo cvs -f -d :ext:sourceware.org:/cvs/src rtag \
6982 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight
6983 cvs -f -d :ext:sourceware.org:/cvs/src rtag \
6984 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight
6992 By using @kbd{-D YYYY-MM-DD-gmt}, the branch is forced to an exact
6995 The trunk is first tagged so that the branch point can easily be found.
6997 Insight, which includes @value{GDBN}, is tagged at the same time.
6999 @file{version.in} gets bumped to avoid version number conflicts.
7001 The reading of @file{.cvsrc} is disabled using @file{-f}.
7004 @subheading Update @file{version.in}
7009 $ V=`echo $v | sed 's/\./_/g'`
7013 $ echo cvs -f -d :ext:sourceware.org:/cvs/src co \
7014 -r gdb_$V-branch src/gdb/version.in
7015 cvs -f -d :ext:sourceware.org:/cvs/src co
7016 -r gdb_5_2-branch src/gdb/version.in
7018 U src/gdb/version.in
7020 $ echo $u.90-0000-00-00-cvs > version.in
7022 5.1.90-0000-00-00-cvs
7023 $ cvs -f commit version.in
7028 @file{0000-00-00} is used as a date to pump prime the version.in update
7031 @file{.90} and the previous branch version are used as fairly arbitrary
7032 initial branch version number.
7036 @subheading Update the web and news pages
7040 @subheading Tweak cron to track the new branch
7042 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
7043 This file needs to be updated so that:
7047 A daily timestamp is added to the file @file{version.in}.
7049 The new branch is included in the snapshot process.
7053 See the file @file{gdbadmin/cron/README} for how to install the updated
7056 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
7057 any changes. That file is copied to both the branch/ and current/
7058 snapshot directories.
7061 @subheading Update the NEWS and README files
7063 The @file{NEWS} file needs to be updated so that on the branch it refers
7064 to @emph{changes in the current release} while on the trunk it also
7065 refers to @emph{changes since the current release}.
7067 The @file{README} file needs to be updated so that it refers to the
7070 @subheading Post the branch info
7072 Send an announcement to the mailing lists:
7076 @email{gdb-announce@@sourceware.org, GDB Announcement mailing list}
7078 @email{gdb@@sourceware.org, GDB Discussion mailing list} and
7079 @email{gdb-testers@@sourceware.org, GDB Testers mailing list}
7082 @emph{Pragmatics: The branch creation is sent to the announce list to
7083 ensure that people people not subscribed to the higher volume discussion
7086 The announcement should include:
7092 How to check out the branch using CVS.
7094 The date/number of weeks until the release.
7096 The branch commit policy still holds.
7099 @section Stabilize the branch
7101 Something goes here.
7103 @section Create a Release
7105 The process of creating and then making available a release is broken
7106 down into a number of stages. The first part addresses the technical
7107 process of creating a releasable tar ball. The later stages address the
7108 process of releasing that tar ball.
7110 When making a release candidate just the first section is needed.
7112 @subsection Create a release candidate
7114 The objective at this stage is to create a set of tar balls that can be
7115 made available as a formal release (or as a less formal release
7118 @subsubheading Freeze the branch
7120 Send out an e-mail notifying everyone that the branch is frozen to
7121 @email{gdb-patches@@sourceware.org}.
7123 @subsubheading Establish a few defaults.
7128 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
7130 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
7134 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
7136 /home/gdbadmin/bin/autoconf
7145 Check the @code{autoconf} version carefully. You want to be using the
7146 version documented in the toplevel @file{README-maintainer-mode} file.
7147 It is very unlikely that the version of @code{autoconf} installed in
7148 system directories (e.g., @file{/usr/bin/autoconf}) is correct.
7151 @subsubheading Check out the relevant modules:
7154 $ for m in gdb insight
7156 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
7166 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
7167 any confusion between what is written here and what your local
7168 @code{cvs} really does.
7171 @subsubheading Update relevant files.
7177 Major releases get their comments added as part of the mainline. Minor
7178 releases should probably mention any significant bugs that were fixed.
7180 Don't forget to include the @file{ChangeLog} entry.
7183 $ emacs gdb/src/gdb/NEWS
7188 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
7189 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
7194 You'll need to update:
7206 $ emacs gdb/src/gdb/README
7211 $ cp gdb/src/gdb/README insight/src/gdb/README
7212 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
7215 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
7216 before the initial branch was cut so just a simple substitute is needed
7219 @emph{Maintainer note: Other projects generate @file{README} and
7220 @file{INSTALL} from the core documentation. This might be worth
7223 @item gdb/version.in
7226 $ echo $v > gdb/src/gdb/version.in
7227 $ cat gdb/src/gdb/version.in
7229 $ emacs gdb/src/gdb/version.in
7232 ... Bump to version ...
7234 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
7235 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
7240 @subsubheading Do the dirty work
7242 This is identical to the process used to create the daily snapshot.
7245 $ for m in gdb insight
7247 ( cd $m/src && gmake -f src-release $m.tar )
7251 If the top level source directory does not have @file{src-release}
7252 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
7255 $ for m in gdb insight
7257 ( cd $m/src && gmake -f Makefile.in $m.tar )
7261 @subsubheading Check the source files
7263 You're looking for files that have mysteriously disappeared.
7264 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
7265 for the @file{version.in} update @kbd{cronjob}.
7268 $ ( cd gdb/src && cvs -f -q -n update )
7272 @dots{} lots of generated files @dots{}
7277 @dots{} lots of generated files @dots{}
7282 @emph{Don't worry about the @file{gdb.info-??} or
7283 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
7284 was also generated only something strange with CVS means that they
7285 didn't get suppressed). Fixing it would be nice though.}
7287 @subsubheading Create compressed versions of the release
7293 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
7294 $ for m in gdb insight
7296 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
7297 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
7307 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
7308 in that mode, @code{gzip} does not know the name of the file and, hence,
7309 can not include it in the compressed file. This is also why the release
7310 process runs @code{tar} and @code{bzip2} as separate passes.
7313 @subsection Sanity check the tar ball
7315 Pick a popular machine (Solaris/PPC?) and try the build on that.
7318 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
7323 $ ./gdb/gdb ./gdb/gdb
7327 Breakpoint 1 at 0x80732bc: file main.c, line 734.
7329 Starting program: /tmp/gdb-5.2/gdb/gdb
7331 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
7332 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
7334 $1 = @{argc = 136426532, argv = 0x821b7f0@}
7338 @subsection Make a release candidate available
7340 If this is a release candidate then the only remaining steps are:
7344 Commit @file{version.in} and @file{ChangeLog}
7346 Tweak @file{version.in} (and @file{ChangeLog} to read
7347 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
7348 process can restart.
7350 Make the release candidate available in
7351 @uref{ftp://sourceware.org/pub/gdb/snapshots/branch}
7353 Notify the relevant mailing lists ( @email{gdb@@sourceware.org} and
7354 @email{gdb-testers@@sourceware.org} that the candidate is available.
7357 @subsection Make a formal release available
7359 (And you thought all that was required was to post an e-mail.)
7361 @subsubheading Install on sware
7363 Copy the new files to both the release and the old release directory:
7366 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
7367 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
7371 Clean up the releases directory so that only the most recent releases
7372 are available (e.g.@: keep 5.2 and 5.2.1 but remove 5.1):
7375 $ cd ~ftp/pub/gdb/releases
7380 Update the file @file{README} and @file{.message} in the releases
7387 $ ln README .message
7390 @subsubheading Update the web pages.
7394 @item htdocs/download/ANNOUNCEMENT
7395 This file, which is posted as the official announcement, includes:
7398 General announcement.
7400 News. If making an @var{M}.@var{N}.1 release, retain the news from
7401 earlier @var{M}.@var{N} release.
7406 @item htdocs/index.html
7407 @itemx htdocs/news/index.html
7408 @itemx htdocs/download/index.html
7409 These files include:
7412 Announcement of the most recent release.
7414 News entry (remember to update both the top level and the news directory).
7416 These pages also need to be regenerate using @code{index.sh}.
7418 @item download/onlinedocs/
7419 You need to find the magic command that is used to generate the online
7420 docs from the @file{.tar.bz2}. The best way is to look in the output
7421 from one of the nightly @code{cron} jobs and then just edit accordingly.
7425 $ ~/ss/update-web-docs \
7426 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
7428 /www/sourceware/htdocs/gdb/download/onlinedocs \
7433 Just like the online documentation. Something like:
7436 $ /bin/sh ~/ss/update-web-ari \
7437 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
7439 /www/sourceware/htdocs/gdb/download/ari \
7445 @subsubheading Shadow the pages onto gnu
7447 Something goes here.
7450 @subsubheading Install the @value{GDBN} tar ball on GNU
7452 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
7453 @file{~ftp/gnu/gdb}.
7455 @subsubheading Make the @file{ANNOUNCEMENT}
7457 Post the @file{ANNOUNCEMENT} file you created above to:
7461 @email{gdb-announce@@sourceware.org, GDB Announcement mailing list}
7463 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
7464 day or so to let things get out)
7466 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
7471 The release is out but you're still not finished.
7473 @subsubheading Commit outstanding changes
7475 In particular you'll need to commit any changes to:
7479 @file{gdb/ChangeLog}
7481 @file{gdb/version.in}
7488 @subsubheading Tag the release
7493 $ d=`date -u +%Y-%m-%d`
7496 $ ( cd insight/src/gdb && cvs -f -q update )
7497 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
7500 Insight is used since that contains more of the release than
7503 @subsubheading Mention the release on the trunk
7505 Just put something in the @file{ChangeLog} so that the trunk also
7506 indicates when the release was made.
7508 @subsubheading Restart @file{gdb/version.in}
7510 If @file{gdb/version.in} does not contain an ISO date such as
7511 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
7512 committed all the release changes it can be set to
7513 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
7514 is important - it affects the snapshot process).
7516 Don't forget the @file{ChangeLog}.
7518 @subsubheading Merge into trunk
7520 The files committed to the branch may also need changes merged into the
7523 @subsubheading Revise the release schedule
7525 Post a revised release schedule to @email{gdb@@sourceware.org, GDB
7526 Discussion List} with an updated announcement. The schedule can be
7527 generated by running:
7530 $ ~/ss/schedule `date +%s` schedule
7534 The first parameter is approximate date/time in seconds (from the epoch)
7535 of the most recent release.
7537 Also update the schedule @code{cronjob}.
7539 @section Post release
7541 Remove any @code{OBSOLETE} code.
7548 The testsuite is an important component of the @value{GDBN} package.
7549 While it is always worthwhile to encourage user testing, in practice
7550 this is rarely sufficient; users typically use only a small subset of
7551 the available commands, and it has proven all too common for a change
7552 to cause a significant regression that went unnoticed for some time.
7554 The @value{GDBN} testsuite uses the DejaGNU testing framework. The
7555 tests themselves are calls to various @code{Tcl} procs; the framework
7556 runs all the procs and summarizes the passes and fails.
7558 @section Using the Testsuite
7560 @cindex running the test suite
7561 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
7562 testsuite's objdir) and type @code{make check}. This just sets up some
7563 environment variables and invokes DejaGNU's @code{runtest} script. While
7564 the testsuite is running, you'll get mentions of which test file is in use,
7565 and a mention of any unexpected passes or fails. When the testsuite is
7566 finished, you'll get a summary that looks like this:
7571 # of expected passes 6016
7572 # of unexpected failures 58
7573 # of unexpected successes 5
7574 # of expected failures 183
7575 # of unresolved testcases 3
7576 # of untested testcases 5
7579 To run a specific test script, type:
7581 make check RUNTESTFLAGS='@var{tests}'
7583 where @var{tests} is a list of test script file names, separated by
7586 If you use GNU make, you can use its @option{-j} option to run the
7587 testsuite in parallel. This can greatly reduce the amount of time it
7588 takes for the testsuite to run. In this case, if you set
7589 @code{RUNTESTFLAGS} then, by default, the tests will be run serially
7590 even under @option{-j}. You can override this and force a parallel run
7591 by setting the @code{make} variable @code{FORCE_PARALLEL} to any
7592 non-empty value. Note that the parallel @kbd{make check} assumes
7593 that you want to run the entire testsuite, so it is not compatible
7594 with some dejagnu options, like @option{--directory}.
7596 The ideal test run consists of expected passes only; however, reality
7597 conspires to keep us from this ideal. Unexpected failures indicate
7598 real problems, whether in @value{GDBN} or in the testsuite. Expected
7599 failures are still failures, but ones which have been decided are too
7600 hard to deal with at the time; for instance, a test case might work
7601 everywhere except on AIX, and there is no prospect of the AIX case
7602 being fixed in the near future. Expected failures should not be added
7603 lightly, since you may be masking serious bugs in @value{GDBN}.
7604 Unexpected successes are expected fails that are passing for some
7605 reason, while unresolved and untested cases often indicate some minor
7606 catastrophe, such as the compiler being unable to deal with a test
7609 When making any significant change to @value{GDBN}, you should run the
7610 testsuite before and after the change, to confirm that there are no
7611 regressions. Note that truly complete testing would require that you
7612 run the testsuite with all supported configurations and a variety of
7613 compilers; however this is more than really necessary. In many cases
7614 testing with a single configuration is sufficient. Other useful
7615 options are to test one big-endian (Sparc) and one little-endian (x86)
7616 host, a cross config with a builtin simulator (powerpc-eabi,
7617 mips-elf), or a 64-bit host (Alpha).
7619 If you add new functionality to @value{GDBN}, please consider adding
7620 tests for it as well; this way future @value{GDBN} hackers can detect
7621 and fix their changes that break the functionality you added.
7622 Similarly, if you fix a bug that was not previously reported as a test
7623 failure, please add a test case for it. Some cases are extremely
7624 difficult to test, such as code that handles host OS failures or bugs
7625 in particular versions of compilers, and it's OK not to try to write
7626 tests for all of those.
7628 DejaGNU supports separate build, host, and target machines. However,
7629 some @value{GDBN} test scripts do not work if the build machine and
7630 the host machine are not the same. In such an environment, these scripts
7631 will give a result of ``UNRESOLVED'', like this:
7634 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
7637 @section Testsuite Parameters
7639 Several variables exist to modify the behavior of the testsuite.
7643 @item @code{TRANSCRIPT}
7645 Sometimes it is convenient to get a transcript of the commands which
7646 the testsuite sends to @value{GDBN}. For example, if @value{GDBN}
7647 crashes during testing, a transcript can be used to more easily
7648 reconstruct the failure when running @value{GDBN} under @value{GDBN}.
7650 You can instruct the @value{GDBN} testsuite to write transcripts by
7651 setting the DejaGNU variable @code{TRANSCRIPT} (to any value)
7652 before invoking @code{runtest} or @kbd{make check}. The transcripts
7653 will be written into DejaGNU's output directory. One transcript will
7654 be made for each invocation of @value{GDBN}; they will be named
7655 @file{transcript.@var{n}}, where @var{n} is an integer. The first
7656 line of the transcript file will show how @value{GDBN} was invoked;
7657 each subsequent line is a command sent as input to @value{GDBN}.
7660 make check RUNTESTFLAGS=TRANSCRIPT=y
7663 Note that the transcript is not always complete. In particular, tests
7664 of completion can yield partial command lines.
7668 Sometimes one wishes to test a different @value{GDBN} than the one in the build
7669 directory. For example, one may wish to run the testsuite on
7670 @file{/usr/bin/gdb}.
7673 make check RUNTESTFLAGS=GDB=/usr/bin/gdb
7676 @item @code{GDBSERVER}
7678 When testing a different @value{GDBN}, it is often useful to also test a
7679 different gdbserver.
7682 make check RUNTESTFLAGS="GDB=/usr/bin/gdb GDBSERVER=/usr/bin/gdbserver"
7685 @item @code{INTERNAL_GDBFLAGS}
7687 When running the testsuite normally one doesn't want whatever is in
7688 @file{~/.gdbinit} to interfere with the tests, therefore the test harness
7689 passes @option{-nx} to @value{GDBN}. One also doesn't want any windowed
7690 version of @value{GDBN}, e.g., @samp{gdb -tui}, to run.
7691 This is achieved via @code{INTERNAL_GDBFLAGS}.
7694 set INTERNAL_GDBFLAGS "-nw -nx"
7697 This is all well and good, except when testing an installed @value{GDBN}
7698 that has been configured with @option{--with-system-gdbinit}. Here one
7699 does not want @file{~/.gdbinit} loaded but one may want the system
7700 @file{.gdbinit} file loaded. This can be achieved by pointing @code{$HOME}
7701 at a directory without a @file{.gdbinit} and by overriding
7702 @code{INTERNAL_GDBFLAGS} and removing @option{-nx}.
7706 HOME=`pwd` runtest \
7708 GDBSERVER=/usr/bin/gdbserver \
7709 INTERNAL_GDBFLAGS=-nw
7714 There are two ways to run the testsuite and pass additional parameters
7715 to DejaGnu. The first is with @kbd{make check} and specifying the
7716 makefile variable @samp{RUNTESTFLAGS}.
7719 make check RUNTESTFLAGS=TRANSCRIPT=y
7722 The second is to cd to the @file{testsuite} directory and invoke the DejaGnu
7723 @command{runtest} command directly.
7728 runtest TRANSCRIPT=y
7731 @section Testsuite Configuration
7732 @cindex Testsuite Configuration
7734 It is possible to adjust the behavior of the testsuite by defining
7735 the global variables listed below, either in a @file{site.exp} file,
7740 @item @code{gdb_test_timeout}
7742 Defining this variable changes the default timeout duration used during
7743 communication with @value{GDBN}. More specifically, the global variable
7744 used during testing is @code{timeout}, but this variable gets reset to
7745 @code{gdb_test_timeout} at the beginning of each testcase, making sure
7746 that any local change to @code{timeout} in a testcase does not affect
7747 subsequent testcases.
7749 This global variable comes in handy when the debugger is slower than
7750 normal due to the testing environment, triggering unexpected @code{TIMEOUT}
7751 test failures. Examples include when testing on a remote machine, or
7752 against a system where communications are slow.
7754 If not specifically defined, this variable gets automatically defined
7755 to the same value as @code{timeout} during the testsuite initialization.
7756 The default value of the timeout is defined in the file
7757 @file{gdb/testsuite/config/unix.exp} that is part of the @value{GDBN}
7758 test suite@footnote{If you are using a board file, it could override
7759 the test-suite default; search the board file for "timeout".}.
7763 @section Testsuite Organization
7765 @cindex test suite organization
7766 The testsuite is entirely contained in @file{gdb/testsuite}. While the
7767 testsuite includes some makefiles and configury, these are very minimal,
7768 and used for little besides cleaning up, since the tests themselves
7769 handle the compilation of the programs that @value{GDBN} will run. The file
7770 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
7771 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
7772 configuration-specific files, typically used for special-purpose
7773 definitions of procs like @code{gdb_load} and @code{gdb_start}.
7775 The tests themselves are to be found in @file{testsuite/gdb.*} and
7776 subdirectories of those. The names of the test files must always end
7777 with @file{.exp}. DejaGNU collects the test files by wildcarding
7778 in the test directories, so both subdirectories and individual files
7779 get chosen and run in alphabetical order.
7781 The following table lists the main types of subdirectories and what they
7782 are for. Since DejaGNU finds test files no matter where they are
7783 located, and since each test file sets up its own compilation and
7784 execution environment, this organization is simply for convenience and
7789 This is the base testsuite. The tests in it should apply to all
7790 configurations of @value{GDBN} (but generic native-only tests may live here).
7791 The test programs should be in the subset of C that is valid K&R,
7792 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
7795 @item gdb.@var{lang}
7796 Language-specific tests for any language @var{lang} besides C. Examples are
7797 @file{gdb.cp} and @file{gdb.java}.
7799 @item gdb.@var{platform}
7800 Non-portable tests. The tests are specific to a specific configuration
7801 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
7804 @item gdb.@var{compiler}
7805 Tests specific to a particular compiler. As of this writing (June
7806 1999), there aren't currently any groups of tests in this category that
7807 couldn't just as sensibly be made platform-specific, but one could
7808 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
7811 @item gdb.@var{subsystem}
7812 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
7813 instance, @file{gdb.disasm} exercises various disassemblers, while
7814 @file{gdb.stabs} tests pathways through the stabs symbol reader.
7817 @section Writing Tests
7818 @cindex writing tests
7820 In many areas, the @value{GDBN} tests are already quite comprehensive; you
7821 should be able to copy existing tests to handle new cases.
7823 You should try to use @code{gdb_test} whenever possible, since it
7824 includes cases to handle all the unexpected errors that might happen.
7825 However, it doesn't cost anything to add new test procedures; for
7826 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
7827 calls @code{gdb_test} multiple times.
7829 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
7830 necessary. Even if @value{GDBN} has several valid responses to
7831 a command, you can use @code{gdb_test_multiple}. Like @code{gdb_test},
7832 @code{gdb_test_multiple} recognizes internal errors and unexpected
7835 Do not write tests which expect a literal tab character from @value{GDBN}.
7836 On some operating systems (e.g.@: OpenBSD) the TTY layer expands tabs to
7837 spaces, so by the time @value{GDBN}'s output reaches expect the tab is gone.
7839 The source language programs do @emph{not} need to be in a consistent
7840 style. Since @value{GDBN} is used to debug programs written in many different
7841 styles, it's worth having a mix of styles in the testsuite; for
7842 instance, some @value{GDBN} bugs involving the display of source lines would
7843 never manifest themselves if the programs used GNU coding style
7846 Some testcase results need more detailed explanation:
7850 Known problem of @value{GDBN} itself. You must specify the @value{GDBN} bug
7851 report number like in these sample tests:
7853 kfail "gdb/13392" "continue to marker 2"
7857 setup_kfail gdb/13392 "*-*-*"
7858 kfail "continue to marker 2"
7862 Known problem of environment. This typically includes @value{NGCC} but it
7863 includes also many other system components which cannot be fixed in the
7864 @value{GDBN} project. Sample test with sanity check not knowing the specific
7865 cause of the problem:
7867 # On x86_64 it is commonly about 4MB.
7868 if @{$stub_size > 25000000@} @{
7869 xfail "stub size $stub_size is too large"
7874 You should provide bug report number for the failing component of the
7875 environment, if such bug report is available:
7877 if @{[test_compiler_info @{gcc-[0-3]-*@}]
7878 || [test_compiler_info @{gcc-4-[0-5]-*@}]@} @{
7879 setup_xfail "gcc/46955" *-*-*
7881 gdb_test "python print ttype.template_argument(2)" "&C::c"
7885 @section Board settings
7886 In @value{GDBN} testsuite, the tests can be configured or customized in the board
7887 file by means of @dfn{Board Settings}. Each setting should be consulted by
7888 test cases that depend on the corresponding feature.
7890 Here are the supported board settings:
7894 @item gdb,cannot_call_functions
7895 The board does not support inferior call, that is, invoking inferior functions
7897 @item gdb,can_reverse
7898 The board supports reverse execution.
7899 @item gdb,no_hardware_watchpoints
7900 The board does not support hardware watchpoints.
7902 @value{GDBN} is unable to intercept target file operations in remote and perform
7904 @item gdb,noinferiorio
7905 The board is unable to provide I/O capability to the inferior.
7906 @c @item gdb,noresults
7909 The board does not support signals.
7910 @item gdb,skip_huge_test
7911 Skip time-consuming tests on the board with slow connection.
7912 @item gdb,skip_float_tests
7913 Skip tests related to float points on target board.
7914 @item gdb,use_precord
7915 The board supports process record.
7916 @item gdb_server_prog
7917 The location of GDBserver. If GDBserver somewhere other than its default
7918 location is used in test, specify the location of GDBserver in this variable.
7919 The location is a file name of GDBserver that can be either absolute or
7920 relative to testsuite subdirectory in build directory.
7922 The location of in-process agent. If in-process agent other than its default
7923 location is used in test, specify the location of in-process agent in
7924 this variable. The location is a file name of in-process agent that can be
7925 either absolute or relative to testsuite subdirectory in build directory.
7927 @value{GDBN} does not support argument passing for inferior.
7929 The board does not support type @code{long long}.
7933 The tests are running with gdb stub.
7940 Check the @file{README} file, it often has useful information that does not
7941 appear anywhere else in the directory.
7944 * Getting Started:: Getting started working on @value{GDBN}
7945 * Debugging GDB:: Debugging @value{GDBN} with itself
7948 @node Getting Started
7950 @section Getting Started
7952 @value{GDBN} is a large and complicated program, and if you first starting to
7953 work on it, it can be hard to know where to start. Fortunately, if you
7954 know how to go about it, there are ways to figure out what is going on.
7956 This manual, the @value{GDBN} Internals manual, has information which applies
7957 generally to many parts of @value{GDBN}.
7959 Information about particular functions or data structures are located in
7960 comments with those functions or data structures. If you run across a
7961 function or a global variable which does not have a comment correctly
7962 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
7963 free to submit a bug report, with a suggested comment if you can figure
7964 out what the comment should say. If you find a comment which is
7965 actually wrong, be especially sure to report that.
7967 Comments explaining the function of macros defined in host, target, or
7968 native dependent files can be in several places. Sometimes they are
7969 repeated every place the macro is defined. Sometimes they are where the
7970 macro is used. Sometimes there is a header file which supplies a
7971 default definition of the macro, and the comment is there. This manual
7972 also documents all the available macros.
7973 @c (@pxref{Host Conditionals}, @pxref{Target
7974 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
7977 Start with the header files. Once you have some idea of how
7978 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
7979 @file{gdbtypes.h}), you will find it much easier to understand the
7980 code which uses and creates those symbol tables.
7982 You may wish to process the information you are getting somehow, to
7983 enhance your understanding of it. Summarize it, translate it to another
7984 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
7985 the code to predict what a test case would do and write the test case
7986 and verify your prediction, etc. If you are reading code and your eyes
7987 are starting to glaze over, this is a sign you need to use a more active
7990 Once you have a part of @value{GDBN} to start with, you can find more
7991 specifically the part you are looking for by stepping through each
7992 function with the @code{next} command. Do not use @code{step} or you
7993 will quickly get distracted; when the function you are stepping through
7994 calls another function try only to get a big-picture understanding
7995 (perhaps using the comment at the beginning of the function being
7996 called) of what it does. This way you can identify which of the
7997 functions being called by the function you are stepping through is the
7998 one which you are interested in. You may need to examine the data
7999 structures generated at each stage, with reference to the comments in
8000 the header files explaining what the data structures are supposed to
8003 Of course, this same technique can be used if you are just reading the
8004 code, rather than actually stepping through it. The same general
8005 principle applies---when the code you are looking at calls something
8006 else, just try to understand generally what the code being called does,
8007 rather than worrying about all its details.
8009 @cindex command implementation
8010 A good place to start when tracking down some particular area is with
8011 a command which invokes that feature. Suppose you want to know how
8012 single-stepping works. As a @value{GDBN} user, you know that the
8013 @code{step} command invokes single-stepping. The command is invoked
8014 via command tables (see @file{command.h}); by convention the function
8015 which actually performs the command is formed by taking the name of
8016 the command and adding @samp{_command}, or in the case of an
8017 @code{info} subcommand, @samp{_info}. For example, the @code{step}
8018 command invokes the @code{step_command} function and the @code{info
8019 display} command invokes @code{display_info}. When this convention is
8020 not followed, you might have to use @code{grep} or @kbd{M-x
8021 tags-search} in emacs, or run @value{GDBN} on itself and set a
8022 breakpoint in @code{execute_command}.
8024 @cindex @code{bug-gdb} mailing list
8025 If all of the above fail, it may be appropriate to ask for information
8026 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
8027 wondering if anyone could give me some tips about understanding
8028 @value{GDBN}''---if we had some magic secret we would put it in this manual.
8029 Suggestions for improving the manual are always welcome, of course.
8033 @section Debugging @value{GDBN} with itself
8034 @cindex debugging @value{GDBN}
8036 If @value{GDBN} is limping on your machine, this is the preferred way to get it
8037 fully functional. Be warned that in some ancient Unix systems, like
8038 Ultrix 4.2, a program can't be running in one process while it is being
8039 debugged in another. Rather than typing the command @kbd{@w{./gdb
8040 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
8041 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
8043 When you run @value{GDBN} in the @value{GDBN} source directory, it will read
8044 @file{gdb-gdb.gdb} file (plus possibly @file{gdb-gdb.py} file) that sets up
8045 some simple things to make debugging gdb easier. The @code{info} command, when
8046 executed without a subcommand in a @value{GDBN} being debugged by gdb, will pop
8047 you back up to the top level gdb. See @file{gdb-gdb.gdb} for details.
8049 If you use emacs, you will probably want to do a @code{make TAGS} after
8050 you configure your distribution; this will put the machine dependent
8051 routines for your local machine where they will be accessed first by
8054 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
8055 have run @code{fixincludes} if you are compiling with gcc.
8057 @section Submitting Patches
8059 @cindex submitting patches
8060 Thanks for thinking of offering your changes back to the community of
8061 @value{GDBN} users. In general we like to get well designed enhancements.
8062 Thanks also for checking in advance about the best way to transfer the
8065 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
8066 This manual summarizes what we believe to be clean design for @value{GDBN}.
8068 If the maintainers don't have time to put the patch in when it arrives,
8069 or if there is any question about a patch, it goes into a large queue
8070 with everyone else's patches and bug reports.
8072 @cindex legal papers for code contributions
8073 The legal issue is that to incorporate substantial changes requires a
8074 copyright assignment from you and/or your employer, granting ownership
8075 of the changes to the Free Software Foundation. You can get the
8076 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
8077 and asking for it. We recommend that people write in "All programs
8078 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
8079 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
8081 contributed with only one piece of legalese pushed through the
8082 bureaucracy and filed with the FSF. We can't start merging changes until
8083 this paperwork is received by the FSF (their rules, which we follow
8084 since we maintain it for them).
8086 Technically, the easiest way to receive changes is to receive each
8087 feature as a small context diff or unidiff, suitable for @code{patch}.
8088 Each message sent to me should include the changes to C code and
8089 header files for a single feature, plus @file{ChangeLog} entries for
8090 each directory where files were modified, and diffs for any changes
8091 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
8092 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
8093 single feature, they can be split down into multiple messages.
8095 In this way, if we read and like the feature, we can add it to the
8096 sources with a single patch command, do some testing, and check it in.
8097 If you leave out the @file{ChangeLog}, we have to write one. If you leave
8098 out the doc, we have to puzzle out what needs documenting. Etc., etc.
8100 The reason to send each change in a separate message is that we will not
8101 install some of the changes. They'll be returned to you with questions
8102 or comments. If we're doing our job correctly, the message back to you
8103 will say what you have to fix in order to make the change acceptable.
8104 The reason to have separate messages for separate features is so that
8105 the acceptable changes can be installed while one or more changes are
8106 being reworked. If multiple features are sent in a single message, we
8107 tend to not put in the effort to sort out the acceptable changes from
8108 the unacceptable, so none of the features get installed until all are
8111 If this sounds painful or authoritarian, well, it is. But we get a lot
8112 of bug reports and a lot of patches, and many of them don't get
8113 installed because we don't have the time to finish the job that the bug
8114 reporter or the contributor could have done. Patches that arrive
8115 complete, working, and well designed, tend to get installed on the day
8116 they arrive. The others go into a queue and get installed as time
8117 permits, which, since the maintainers have many demands to meet, may not
8118 be for quite some time.
8120 Please send patches directly to
8121 @email{gdb-patches@@sourceware.org, the @value{GDBN} maintainers}.
8123 @section Build Script
8125 @cindex build script
8127 The script @file{gdb_buildall.sh} builds @value{GDBN} with flag
8128 @option{--enable-targets=all} set. This builds @value{GDBN} with all supported
8129 targets activated. This helps testing @value{GDBN} when doing changes that
8130 affect more than one architecture and is much faster than using
8131 @file{gdb_mbuild.sh}.
8133 After building @value{GDBN} the script checks which architectures are
8134 supported and then switches the current architecture to each of those to get
8135 information about the architecture. The test results are stored in log files
8136 in the directory the script was called from.
8138 @include observer.texi
8140 @node GNU Free Documentation License
8141 @appendix GNU Free Documentation License
8145 @unnumbered Concept Index
8149 @node Function and Variable Index
8150 @unnumbered Function and Variable Index