1 \input texinfo @c -*- texinfo -*-
2 @setfilename gdbint.info
4 @dircategory Software development
6 * Gdb-Internals: (gdbint). The GNU debugger's internals.
10 This file documents the internals of the GNU debugger @value{GDBN}.
11 Copyright (C) 1990, 1991, 1992, 1993, 1994, 1996, 1998, 1999, 2000, 2001,
12 2002, 2003, 2004, 2005, 2006
13 Free Software Foundation, Inc.
14 Contributed by Cygnus Solutions. Written by John Gilmore.
15 Second Edition by Stan Shebs.
17 Permission is granted to copy, distribute and/or modify this document
18 under the terms of the GNU Free Documentation License, Version 1.1 or
19 any later version published by the Free Software Foundation; with no
20 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
21 Texts. A copy of the license is included in the section entitled ``GNU
22 Free Documentation License''.
25 @setchapternewpage off
26 @settitle @value{GDBN} Internals
32 @title @value{GDBN} Internals
33 @subtitle{A guide to the internals of the GNU debugger}
35 @author Cygnus Solutions
36 @author Second Edition:
38 @author Cygnus Solutions
41 \def\$#1${{#1}} % Kluge: collect RCS revision info without $...$
42 \xdef\manvers{\$Revision$} % For use in headers, footers too
44 \hfill Cygnus Solutions\par
46 \hfill \TeX{}info \texinfoversion\par
50 @vskip 0pt plus 1filll
51 Copyright @copyright{} 1990,1991,1992,1993,1994,1996,1998,1999,2000,2001,
52 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
54 Permission is granted to copy, distribute and/or modify this document
55 under the terms of the GNU Free Documentation License, Version 1.1 or
56 any later version published by the Free Software Foundation; with no
57 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
58 Texts. A copy of the license is included in the section entitled ``GNU
59 Free Documentation License''.
65 @c Perhaps this should be the title of the document (but only for info,
66 @c not for TeX). Existing GNU manuals seem inconsistent on this point.
67 @top Scope of this Document
69 This document documents the internals of the GNU debugger, @value{GDBN}. It
70 includes description of @value{GDBN}'s key algorithms and operations, as well
71 as the mechanisms that adapt @value{GDBN} to specific hosts and targets.
82 * Target Architecture Definition::
83 * Target Vector Definition::
88 * Versions and Branches::
89 * Start of New Year Procedure::
94 * GDB Observers:: @value{GDBN} Currently available observers
95 * GNU Free Documentation License:: The license for this documentation
101 @chapter Requirements
102 @cindex requirements for @value{GDBN}
104 Before diving into the internals, you should understand the formal
105 requirements and other expectations for @value{GDBN}. Although some
106 of these may seem obvious, there have been proposals for @value{GDBN}
107 that have run counter to these requirements.
109 First of all, @value{GDBN} is a debugger. It's not designed to be a
110 front panel for embedded systems. It's not a text editor. It's not a
111 shell. It's not a programming environment.
113 @value{GDBN} is an interactive tool. Although a batch mode is
114 available, @value{GDBN}'s primary role is to interact with a human
117 @value{GDBN} should be responsive to the user. A programmer hot on
118 the trail of a nasty bug, and operating under a looming deadline, is
119 going to be very impatient of everything, including the response time
120 to debugger commands.
122 @value{GDBN} should be relatively permissive, such as for expressions.
123 While the compiler should be picky (or have the option to be made
124 picky), since source code lives for a long time usually, the
125 programmer doing debugging shouldn't be spending time figuring out to
126 mollify the debugger.
128 @value{GDBN} will be called upon to deal with really large programs.
129 Executable sizes of 50 to 100 megabytes occur regularly, and we've
130 heard reports of programs approaching 1 gigabyte in size.
132 @value{GDBN} should be able to run everywhere. No other debugger is
133 available for even half as many configurations as @value{GDBN}
137 @node Overall Structure
139 @chapter Overall Structure
141 @value{GDBN} consists of three major subsystems: user interface,
142 symbol handling (the @dfn{symbol side}), and target system handling (the
145 The user interface consists of several actual interfaces, plus
148 The symbol side consists of object file readers, debugging info
149 interpreters, symbol table management, source language expression
150 parsing, type and value printing.
152 The target side consists of execution control, stack frame analysis, and
153 physical target manipulation.
155 The target side/symbol side division is not formal, and there are a
156 number of exceptions. For instance, core file support involves symbolic
157 elements (the basic core file reader is in BFD) and target elements (it
158 supplies the contents of memory and the values of registers). Instead,
159 this division is useful for understanding how the minor subsystems
162 @section The Symbol Side
164 The symbolic side of @value{GDBN} can be thought of as ``everything
165 you can do in @value{GDBN} without having a live program running''.
166 For instance, you can look at the types of variables, and evaluate
167 many kinds of expressions.
169 @section The Target Side
171 The target side of @value{GDBN} is the ``bits and bytes manipulator''.
172 Although it may make reference to symbolic info here and there, most
173 of the target side will run with only a stripped executable
174 available---or even no executable at all, in remote debugging cases.
176 Operations such as disassembly, stack frame crawls, and register
177 display, are able to work with no symbolic info at all. In some cases,
178 such as disassembly, @value{GDBN} will use symbolic info to present addresses
179 relative to symbols rather than as raw numbers, but it will work either
182 @section Configurations
186 @dfn{Host} refers to attributes of the system where @value{GDBN} runs.
187 @dfn{Target} refers to the system where the program being debugged
188 executes. In most cases they are the same machine, in which case a
189 third type of @dfn{Native} attributes come into play.
191 Defines and include files needed to build on the host are host support.
192 Examples are tty support, system defined types, host byte order, host
195 Defines and information needed to handle the target format are target
196 dependent. Examples are the stack frame format, instruction set,
197 breakpoint instruction, registers, and how to set up and tear down the stack
200 Information that is only needed when the host and target are the same,
201 is native dependent. One example is Unix child process support; if the
202 host and target are not the same, doing a fork to start the target
203 process is a bad idea. The various macros needed for finding the
204 registers in the @code{upage}, running @code{ptrace}, and such are all
205 in the native-dependent files.
207 Another example of native-dependent code is support for features that
208 are really part of the target environment, but which require
209 @code{#include} files that are only available on the host system. Core
210 file handling and @code{setjmp} handling are two common cases.
212 When you want to make @value{GDBN} work ``native'' on a particular machine, you
213 have to include all three kinds of information.
215 @section Source Tree Structure
216 @cindex @value{GDBN} source tree structure
218 The @value{GDBN} source directory has a mostly flat structure---there
219 are only a few subdirectories. A file's name usually gives a hint as
220 to what it does; for example, @file{stabsread.c} reads stabs,
221 @file{dwarfread.c} reads DWARF, etc.
223 Files that are related to some common task have names that share
224 common substrings. For example, @file{*-thread.c} files deal with
225 debugging threads on various platforms; @file{*read.c} files deal with
226 reading various kinds of symbol and object files; @file{inf*.c} files
227 deal with direct control of the @dfn{inferior program} (@value{GDBN}
228 parlance for the program being debugged).
230 There are several dozens of files in the @file{*-tdep.c} family.
231 @samp{tdep} stands for @dfn{target-dependent code}---each of these
232 files implements debug support for a specific target architecture
233 (sparc, mips, etc). Usually, only one of these will be used in a
234 specific @value{GDBN} configuration (sometimes two, closely related).
236 Similarly, there are many @file{*-nat.c} files, each one for native
237 debugging on a specific system (e.g., @file{sparc-linux-nat.c} is for
238 native debugging of Sparc machines running the Linux kernel).
240 The few subdirectories of the source tree are:
244 Code that implements @dfn{CLI}, the @value{GDBN} Command-Line
245 Interpreter. @xref{User Interface, Command Interpreter}.
248 Code for the @value{GDBN} remote server.
251 Code for Insight, the @value{GDBN} TK-based GUI front-end.
254 The @dfn{GDB/MI}, the @value{GDBN} Machine Interface interpreter.
257 Target signal translation code.
260 Code for @dfn{TUI}, the @value{GDBN} Text-mode full-screen User
261 Interface. @xref{User Interface, TUI}.
269 @value{GDBN} uses a number of debugging-specific algorithms. They are
270 often not very complicated, but get lost in the thicket of special
271 cases and real-world issues. This chapter describes the basic
272 algorithms and mentions some of the specific target definitions that
278 @cindex call stack frame
279 A frame is a construct that @value{GDBN} uses to keep track of calling
280 and called functions.
282 @cindex frame, unwind
283 @value{GDBN}'s frame model, a fresh design, was implemented with the
284 need to support @sc{dwarf}'s Call Frame Information in mind. In fact,
285 the term ``unwind'' is taken directly from that specification.
286 Developers wishing to learn more about unwinders, are encouraged to
287 read the the @sc{dwarf} specification.
289 @findex frame_register_unwind
290 @findex get_frame_register
291 @value{GDBN}'s model is that you find a frame's registers by
292 ``unwinding'' them from the next younger frame. That is,
293 @samp{get_frame_register} which returns the value of a register in
294 frame #1 (the next-to-youngest frame), is implemented by calling frame
295 #0's @code{frame_register_unwind} (the youngest frame). But then the
296 obvious question is: how do you access the registers of the youngest
299 @cindex sentinel frame
300 @findex get_frame_type
301 @vindex SENTINEL_FRAME
302 To answer this question, GDB has the @dfn{sentinel} frame, the
303 ``-1st'' frame. Unwinding registers from the sentinel frame gives you
304 the current values of the youngest real frame's registers. If @var{f}
305 is a sentinel frame, then @code{get_frame_type (@var{f}) ==
308 @section Prologue Analysis
310 @cindex prologue analysis
311 @cindex call frame information
312 @cindex CFI (call frame information)
313 To produce a backtrace and allow the user to manipulate older frames'
314 variables and arguments, @value{GDBN} needs to find the base addresses
315 of older frames, and discover where those frames' registers have been
316 saved. Since a frame's ``callee-saves'' registers get saved by
317 younger frames if and when they're reused, a frame's registers may be
318 scattered unpredictably across younger frames. This means that
319 changing the value of a register-allocated variable in an older frame
320 may actually entail writing to a save slot in some younger frame.
322 Modern versions of GCC emit Dwarf call frame information (``CFI''),
323 which describes how to find frame base addresses and saved registers.
324 But CFI is not always available, so as a fallback @value{GDBN} uses a
325 technique called @dfn{prologue analysis} to find frame sizes and saved
326 registers. A prologue analyzer disassembles the function's machine
327 code starting from its entry point, and looks for instructions that
328 allocate frame space, save the stack pointer in a frame pointer
329 register, save registers, and so on. Obviously, this can't be done
330 accurately in general, but it's tractable to do well enough to be very
331 helpful. Prologue analysis predates the GNU toolchain's support for
332 CFI; at one time, prologue analysis was the only mechanism
333 @value{GDBN} used for stack unwinding at all, when the function
334 calling conventions didn't specify a fixed frame layout.
336 In the olden days, function prologues were generated by hand-written,
337 target-specific code in GCC, and treated as opaque and untouchable by
338 optimizers. Looking at this code, it was usually straightforward to
339 write a prologue analyzer for @value{GDBN} that would accurately
340 understand all the prologues GCC would generate. However, over time
341 GCC became more aggressive about instruction scheduling, and began to
342 understand more about the semantics of the prologue instructions
343 themselves; in response, @value{GDBN}'s analyzers became more complex
344 and fragile. Keeping the prologue analyzers working as GCC (and the
345 instruction sets themselves) evolved became a substantial task.
347 @cindex @file{prologue-value.c}
348 @cindex abstract interpretation of function prologues
349 @cindex pseudo-evaluation of function prologues
350 To try to address this problem, the code in @file{prologue-value.h}
351 and @file{prologue-value.c} provides a general framework for writing
352 prologue analyzers that are simpler and more robust than ad-hoc
353 analyzers. When we analyze a prologue using the prologue-value
354 framework, we're really doing ``abstract interpretation'' or
355 ``pseudo-evaluation'': running the function's code in simulation, but
356 using conservative approximations of the values registers and memory
357 would hold when the code actually runs. For example, if our function
358 starts with the instruction:
361 addi r1, 42 # add 42 to r1
364 we don't know exactly what value will be in @code{r1} after executing
365 this instruction, but we do know it'll be 42 greater than its original
368 If we then see an instruction like:
371 addi r1, 22 # add 22 to r1
374 we still don't know what @code{r1's} value is, but again, we can say
375 it is now 64 greater than its original value.
377 If the next instruction were:
380 mov r2, r1 # set r2 to r1's value
383 then we can say that @code{r2's} value is now the original value of
386 It's common for prologues to save registers on the stack, so we'll
387 need to track the values of stack frame slots, as well as the
388 registers. So after an instruction like this:
394 then we'd know that the stack slot four bytes above the frame pointer
395 holds the original value of @code{r1} plus 64.
399 Of course, this can only go so far before it gets unreasonable. If we
400 wanted to be able to say anything about the value of @code{r1} after
404 xor r1, r3 # exclusive-or r1 and r3, place result in r1
407 then things would get pretty complex. But remember, we're just doing
408 a conservative approximation; if exclusive-or instructions aren't
409 relevant to prologues, we can just say @code{r1}'s value is now
410 ``unknown''. We can ignore things that are too complex, if that loss of
411 information is acceptable for our application.
413 So when we say ``conservative approximation'' here, what we mean is an
414 approximation that is either accurate, or marked ``unknown'', but
417 Using this framework, a prologue analyzer is simply an interpreter for
418 machine code, but one that uses conservative approximations for the
419 contents of registers and memory instead of actual values. Starting
420 from the function's entry point, you simulate instructions up to the
421 current PC, or an instruction that you don't know how to simulate.
422 Now you can examine the state of the registers and stack slots you've
428 To see how large your stack frame is, just check the value of the
429 stack pointer register; if it's the original value of the SP
430 minus a constant, then that constant is the stack frame's size.
431 If the SP's value has been marked as ``unknown'', then that means
432 the prologue has done something too complex for us to track, and
433 we don't know the frame size.
436 To see where we've saved the previous frame's registers, we just
437 search the values we've tracked --- stack slots, usually, but
438 registers, too, if you want --- for something equal to the register's
439 original value. If the calling conventions suggest a standard place
440 to save a given register, then we can check there first, but really,
441 anything that will get us back the original value will probably work.
444 This does take some work. But prologue analyzers aren't
445 quick-and-simple pattern patching to recognize a few fixed prologue
446 forms any more; they're big, hairy functions. Along with inferior
447 function calls, prologue analysis accounts for a substantial portion
448 of the time needed to stabilize a @value{GDBN} port. So it's
449 worthwhile to look for an approach that will be easier to understand
450 and maintain. In the approach described above:
455 It's easier to see that the analyzer is correct: you just see
456 whether the analyzer properly (albeit conservatively) simulates
457 the effect of each instruction.
460 It's easier to extend the analyzer: you can add support for new
461 instructions, and know that you haven't broken anything that
462 wasn't already broken before.
465 It's orthogonal: to gather new information, you don't need to
466 complicate the code for each instruction. As long as your domain
467 of conservative values is already detailed enough to tell you
468 what you need, then all the existing instruction simulations are
469 already gathering the right data for you.
473 The file @file{prologue-value.h} contains detailed comments explaining
474 the framework and how to use it.
477 @section Breakpoint Handling
480 In general, a breakpoint is a user-designated location in the program
481 where the user wants to regain control if program execution ever reaches
484 There are two main ways to implement breakpoints; either as ``hardware''
485 breakpoints or as ``software'' breakpoints.
487 @cindex hardware breakpoints
488 @cindex program counter
489 Hardware breakpoints are sometimes available as a builtin debugging
490 features with some chips. Typically these work by having dedicated
491 register into which the breakpoint address may be stored. If the PC
492 (shorthand for @dfn{program counter})
493 ever matches a value in a breakpoint registers, the CPU raises an
494 exception and reports it to @value{GDBN}.
496 Another possibility is when an emulator is in use; many emulators
497 include circuitry that watches the address lines coming out from the
498 processor, and force it to stop if the address matches a breakpoint's
501 A third possibility is that the target already has the ability to do
502 breakpoints somehow; for instance, a ROM monitor may do its own
503 software breakpoints. So although these are not literally ``hardware
504 breakpoints'', from @value{GDBN}'s point of view they work the same;
505 @value{GDBN} need not do anything more than set the breakpoint and wait
506 for something to happen.
508 Since they depend on hardware resources, hardware breakpoints may be
509 limited in number; when the user asks for more, @value{GDBN} will
510 start trying to set software breakpoints. (On some architectures,
511 notably the 32-bit x86 platforms, @value{GDBN} cannot always know
512 whether there's enough hardware resources to insert all the hardware
513 breakpoints and watchpoints. On those platforms, @value{GDBN} prints
514 an error message only when the program being debugged is continued.)
516 @cindex software breakpoints
517 Software breakpoints require @value{GDBN} to do somewhat more work.
518 The basic theory is that @value{GDBN} will replace a program
519 instruction with a trap, illegal divide, or some other instruction
520 that will cause an exception, and then when it's encountered,
521 @value{GDBN} will take the exception and stop the program. When the
522 user says to continue, @value{GDBN} will restore the original
523 instruction, single-step, re-insert the trap, and continue on.
525 Since it literally overwrites the program being tested, the program area
526 must be writable, so this technique won't work on programs in ROM. It
527 can also distort the behavior of programs that examine themselves,
528 although such a situation would be highly unusual.
530 Also, the software breakpoint instruction should be the smallest size of
531 instruction, so it doesn't overwrite an instruction that might be a jump
532 target, and cause disaster when the program jumps into the middle of the
533 breakpoint instruction. (Strictly speaking, the breakpoint must be no
534 larger than the smallest interval between instructions that may be jump
535 targets; perhaps there is an architecture where only even-numbered
536 instructions may jumped to.) Note that it's possible for an instruction
537 set not to have any instructions usable for a software breakpoint,
538 although in practice only the ARC has failed to define such an
542 The basic definition of the software breakpoint is the macro
545 Basic breakpoint object handling is in @file{breakpoint.c}. However,
546 much of the interesting breakpoint action is in @file{infrun.c}.
549 @cindex insert or remove software breakpoint
550 @findex target_remove_breakpoint
551 @findex target_insert_breakpoint
552 @item target_remove_breakpoint (@var{bp_tgt})
553 @itemx target_insert_breakpoint (@var{bp_tgt})
554 Insert or remove a software breakpoint at address
555 @code{@var{bp_tgt}->placed_address}. Returns zero for success,
556 non-zero for failure. On input, @var{bp_tgt} contains the address of the
557 breakpoint, and is otherwise initialized to zero. The fields of the
558 @code{struct bp_target_info} pointed to by @var{bp_tgt} are updated
559 to contain other information about the breakpoint on output. The field
560 @code{placed_address} may be updated if the breakpoint was placed at a
561 related address; the field @code{shadow_contents} contains the real
562 contents of the bytes where the breakpoint has been inserted,
563 if reading memory would return the breakpoint instead of the
564 underlying memory; the field @code{shadow_len} is the length of
565 memory cached in @code{shadow_contents}, if any; and the field
566 @code{placed_size} is optionally set and used by the target, if
567 it could differ from @code{shadow_len}.
569 For example, the remote target @samp{Z0} packet does not require
570 shadowing memory, so @code{shadow_len} is left at zero. However,
571 the length reported by @code{BREAKPOINT_FROM_PC} is cached in
572 @code{placed_size}, so that a matching @samp{z0} packet can be
573 used to remove the breakpoint.
575 @cindex insert or remove hardware breakpoint
576 @findex target_remove_hw_breakpoint
577 @findex target_insert_hw_breakpoint
578 @item target_remove_hw_breakpoint (@var{bp_tgt})
579 @itemx target_insert_hw_breakpoint (@var{bp_tgt})
580 Insert or remove a hardware-assisted breakpoint at address
581 @code{@var{bp_tgt}->placed_address}. Returns zero for success,
582 non-zero for failure. See @code{target_insert_breakpoint} for
583 a description of the @code{struct bp_target_info} pointed to by
584 @var{bp_tgt}; the @code{shadow_contents} and
585 @code{shadow_len} members are not used for hardware breakpoints,
586 but @code{placed_size} may be.
589 @section Single Stepping
591 @section Signal Handling
593 @section Thread Handling
595 @section Inferior Function Calls
597 @section Longjmp Support
599 @cindex @code{longjmp} debugging
600 @value{GDBN} has support for figuring out that the target is doing a
601 @code{longjmp} and for stopping at the target of the jump, if we are
602 stepping. This is done with a few specialized internal breakpoints,
603 which are visible in the output of the @samp{maint info breakpoint}
606 @findex GET_LONGJMP_TARGET
607 To make this work, you need to define a macro called
608 @code{GET_LONGJMP_TARGET}, which will examine the @code{jmp_buf}
609 structure and extract the longjmp target address. Since @code{jmp_buf}
610 is target specific, you will need to define it in the appropriate
611 @file{tm-@var{target}.h} file. Look in @file{tm-sun4os4.h} and
612 @file{sparc-tdep.c} for examples of how to do this.
617 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
618 breakpoints}) which break when data is accessed rather than when some
619 instruction is executed. When you have data which changes without
620 your knowing what code does that, watchpoints are the silver bullet to
621 hunt down and kill such bugs.
623 @cindex hardware watchpoints
624 @cindex software watchpoints
625 Watchpoints can be either hardware-assisted or not; the latter type is
626 known as ``software watchpoints.'' @value{GDBN} always uses
627 hardware-assisted watchpoints if they are available, and falls back on
628 software watchpoints otherwise. Typical situations where @value{GDBN}
629 will use software watchpoints are:
633 The watched memory region is too large for the underlying hardware
634 watchpoint support. For example, each x86 debug register can watch up
635 to 4 bytes of memory, so trying to watch data structures whose size is
636 more than 16 bytes will cause @value{GDBN} to use software
640 The value of the expression to be watched depends on data held in
641 registers (as opposed to memory).
644 Too many different watchpoints requested. (On some architectures,
645 this situation is impossible to detect until the debugged program is
646 resumed.) Note that x86 debug registers are used both for hardware
647 breakpoints and for watchpoints, so setting too many hardware
648 breakpoints might cause watchpoint insertion to fail.
651 No hardware-assisted watchpoints provided by the target
655 Software watchpoints are very slow, since @value{GDBN} needs to
656 single-step the program being debugged and test the value of the
657 watched expression(s) after each instruction. The rest of this
658 section is mostly irrelevant for software watchpoints.
660 When the inferior stops, @value{GDBN} tries to establish, among other
661 possible reasons, whether it stopped due to a watchpoint being hit.
662 For a data-write watchpoint, it does so by evaluating, for each
663 watchpoint, the expression whose value is being watched, and testing
664 whether the watched value has changed. For data-read and data-access
665 watchpoints, @value{GDBN} needs the target to supply a primitive that
666 returns the address of the data that was accessed or read (see the
667 description of @code{target_stopped_data_address} below): if this
668 primitive returns a valid address, @value{GDBN} infers that a
669 watchpoint triggered if it watches an expression whose evaluation uses
672 @value{GDBN} uses several macros and primitives to support hardware
676 @findex TARGET_HAS_HARDWARE_WATCHPOINTS
677 @item TARGET_HAS_HARDWARE_WATCHPOINTS
678 If defined, the target supports hardware watchpoints.
680 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
681 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
682 Return the number of hardware watchpoints of type @var{type} that are
683 possible to be set. The value is positive if @var{count} watchpoints
684 of this type can be set, zero if setting watchpoints of this type is
685 not supported, and negative if @var{count} is more than the maximum
686 number of watchpoints of type @var{type} that can be set. @var{other}
687 is non-zero if other types of watchpoints are currently enabled (there
688 are architectures which cannot set watchpoints of different types at
691 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
692 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
693 Return non-zero if hardware watchpoints can be used to watch a region
694 whose address is @var{addr} and whose length in bytes is @var{len}.
696 @cindex insert or remove hardware watchpoint
697 @findex target_insert_watchpoint
698 @findex target_remove_watchpoint
699 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
700 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
701 Insert or remove a hardware watchpoint starting at @var{addr}, for
702 @var{len} bytes. @var{type} is the watchpoint type, one of the
703 possible values of the enumerated data type @code{target_hw_bp_type},
704 defined by @file{breakpoint.h} as follows:
707 enum target_hw_bp_type
709 hw_write = 0, /* Common (write) HW watchpoint */
710 hw_read = 1, /* Read HW watchpoint */
711 hw_access = 2, /* Access (read or write) HW watchpoint */
712 hw_execute = 3 /* Execute HW breakpoint */
717 These two macros should return 0 for success, non-zero for failure.
719 @findex target_stopped_data_address
720 @item target_stopped_data_address (@var{addr_p})
721 If the inferior has some watchpoint that triggered, place the address
722 associated with the watchpoint at the location pointed to by
723 @var{addr_p} and return non-zero. Otherwise, return zero. Note that
724 this primitive is used by @value{GDBN} only on targets that support
725 data-read or data-access type watchpoints, so targets that have
726 support only for data-write watchpoints need not implement these
729 @findex HAVE_STEPPABLE_WATCHPOINT
730 @item HAVE_STEPPABLE_WATCHPOINT
731 If defined to a non-zero value, it is not necessary to disable a
732 watchpoint to step over it.
734 @findex HAVE_NONSTEPPABLE_WATCHPOINT
735 @item HAVE_NONSTEPPABLE_WATCHPOINT
736 If defined to a non-zero value, @value{GDBN} should disable a
737 watchpoint to step the inferior over it.
739 @findex HAVE_CONTINUABLE_WATCHPOINT
740 @item HAVE_CONTINUABLE_WATCHPOINT
741 If defined to a non-zero value, it is possible to continue the
742 inferior after a watchpoint has been hit.
744 @findex CANNOT_STEP_HW_WATCHPOINTS
745 @item CANNOT_STEP_HW_WATCHPOINTS
746 If this is defined to a non-zero value, @value{GDBN} will remove all
747 watchpoints before stepping the inferior.
749 @findex STOPPED_BY_WATCHPOINT
750 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
751 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
752 the type @code{struct target_waitstatus}, defined by @file{target.h}.
753 Normally, this macro is defined to invoke the function pointed to by
754 the @code{to_stopped_by_watchpoint} member of the structure (of the
755 type @code{target_ops}, defined on @file{target.h}) that describes the
756 target-specific operations; @code{to_stopped_by_watchpoint} ignores
757 the @var{wait_status} argument.
759 @value{GDBN} does not require the non-zero value returned by
760 @code{STOPPED_BY_WATCHPOINT} to be 100% correct, so if a target cannot
761 determine for sure whether the inferior stopped due to a watchpoint,
762 it could return non-zero ``just in case''.
765 @subsection x86 Watchpoints
766 @cindex x86 debug registers
767 @cindex watchpoints, on x86
769 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
770 registers designed to facilitate debugging. @value{GDBN} provides a
771 generic library of functions that x86-based ports can use to implement
772 support for watchpoints and hardware-assisted breakpoints. This
773 subsection documents the x86 watchpoint facilities in @value{GDBN}.
775 To use the generic x86 watchpoint support, a port should do the
779 @findex I386_USE_GENERIC_WATCHPOINTS
781 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
782 target-dependent headers.
785 Include the @file{config/i386/nm-i386.h} header file @emph{after}
786 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
789 Add @file{i386-nat.o} to the value of the Make variable
790 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}) or
791 @code{TDEPFILES} (@pxref{Target Architecture Definition, TDEPFILES}).
794 Provide implementations for the @code{I386_DR_LOW_*} macros described
795 below. Typically, each macro should call a target-specific function
796 which does the real work.
799 The x86 watchpoint support works by maintaining mirror images of the
800 debug registers. Values are copied between the mirror images and the
801 real debug registers via a set of macros which each target needs to
805 @findex I386_DR_LOW_SET_CONTROL
806 @item I386_DR_LOW_SET_CONTROL (@var{val})
807 Set the Debug Control (DR7) register to the value @var{val}.
809 @findex I386_DR_LOW_SET_ADDR
810 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
811 Put the address @var{addr} into the debug register number @var{idx}.
813 @findex I386_DR_LOW_RESET_ADDR
814 @item I386_DR_LOW_RESET_ADDR (@var{idx})
815 Reset (i.e.@: zero out) the address stored in the debug register
818 @findex I386_DR_LOW_GET_STATUS
819 @item I386_DR_LOW_GET_STATUS
820 Return the value of the Debug Status (DR6) register. This value is
821 used immediately after it is returned by
822 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
826 For each one of the 4 debug registers (whose indices are from 0 to 3)
827 that store addresses, a reference count is maintained by @value{GDBN},
828 to allow sharing of debug registers by several watchpoints. This
829 allows users to define several watchpoints that watch the same
830 expression, but with different conditions and/or commands, without
831 wasting debug registers which are in short supply. @value{GDBN}
832 maintains the reference counts internally, targets don't have to do
833 anything to use this feature.
835 The x86 debug registers can each watch a region that is 1, 2, or 4
836 bytes long. The ia32 architecture requires that each watched region
837 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
838 region on 4-byte boundary. However, the x86 watchpoint support in
839 @value{GDBN} can watch unaligned regions and regions larger than 4
840 bytes (up to 16 bytes) by allocating several debug registers to watch
841 a single region. This allocation of several registers per a watched
842 region is also done automatically without target code intervention.
844 The generic x86 watchpoint support provides the following API for the
845 @value{GDBN}'s application code:
848 @findex i386_region_ok_for_watchpoint
849 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
850 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
851 this function. It counts the number of debug registers required to
852 watch a given region, and returns a non-zero value if that number is
853 less than 4, the number of debug registers available to x86
856 @findex i386_stopped_data_address
857 @item i386_stopped_data_address (@var{addr_p})
859 @code{target_stopped_data_address} is set to call this function.
861 function examines the breakpoint condition bits in the DR6 Debug
862 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
863 macro, and returns the address associated with the first bit that is
866 @findex i386_stopped_by_watchpoint
867 @item i386_stopped_by_watchpoint (void)
868 The macro @code{STOPPED_BY_WATCHPOINT}
869 is set to call this function. The
870 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
871 function examines the breakpoint condition bits in the DR6 Debug
872 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
873 macro, and returns true if any bit is set. Otherwise, false is
876 @findex i386_insert_watchpoint
877 @findex i386_remove_watchpoint
878 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
879 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
880 Insert or remove a watchpoint. The macros
881 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
882 are set to call these functions. @code{i386_insert_watchpoint} first
883 looks for a debug register which is already set to watch the same
884 region for the same access types; if found, it just increments the
885 reference count of that debug register, thus implementing debug
886 register sharing between watchpoints. If no such register is found,
887 the function looks for a vacant debug register, sets its mirrored
888 value to @var{addr}, sets the mirrored value of DR7 Debug Control
889 register as appropriate for the @var{len} and @var{type} parameters,
890 and then passes the new values of the debug register and DR7 to the
891 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
892 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
893 required to cover the given region, the above process is repeated for
896 @code{i386_remove_watchpoint} does the opposite: it resets the address
897 in the mirrored value of the debug register and its read/write and
898 length bits in the mirrored value of DR7, then passes these new
899 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
900 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
901 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
902 decrements the reference count, and only calls
903 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
904 the count goes to zero.
906 @findex i386_insert_hw_breakpoint
907 @findex i386_remove_hw_breakpoint
908 @item i386_insert_hw_breakpoint (@var{bp_tgt})
909 @itemx i386_remove_hw_breakpoint (@var{bp_tgt})
910 These functions insert and remove hardware-assisted breakpoints. The
911 macros @code{target_insert_hw_breakpoint} and
912 @code{target_remove_hw_breakpoint} are set to call these functions.
913 The argument is a @code{struct bp_target_info *}, as described in
914 the documentation for @code{target_insert_breakpoint}.
915 These functions work like @code{i386_insert_watchpoint} and
916 @code{i386_remove_watchpoint}, respectively, except that they set up
917 the debug registers to watch instruction execution, and each
918 hardware-assisted breakpoint always requires exactly one debug
921 @findex i386_stopped_by_hwbp
922 @item i386_stopped_by_hwbp (void)
923 This function returns non-zero if the inferior has some watchpoint or
924 hardware breakpoint that triggered. It works like
925 @code{i386_stopped_data_address}, except that it doesn't record the
926 address whose watchpoint triggered.
928 @findex i386_cleanup_dregs
929 @item i386_cleanup_dregs (void)
930 This function clears all the reference counts, addresses, and control
931 bits in the mirror images of the debug registers. It doesn't affect
932 the actual debug registers in the inferior process.
939 x86 processors support setting watchpoints on I/O reads or writes.
940 However, since no target supports this (as of March 2001), and since
941 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
942 watchpoints, this feature is not yet available to @value{GDBN} running
946 x86 processors can enable watchpoints locally, for the current task
947 only, or globally, for all the tasks. For each debug register,
948 there's a bit in the DR7 Debug Control register that determines
949 whether the associated address is watched locally or globally. The
950 current implementation of x86 watchpoint support in @value{GDBN}
951 always sets watchpoints to be locally enabled, since global
952 watchpoints might interfere with the underlying OS and are probably
953 unavailable in many platforms.
959 In the abstract, a checkpoint is a point in the execution history of
960 the program, which the user may wish to return to at some later time.
962 Internally, a checkpoint is a saved copy of the program state, including
963 whatever information is required in order to restore the program to that
964 state at a later time. This can be expected to include the state of
965 registers and memory, and may include external state such as the state
966 of open files and devices.
968 There are a number of ways in which checkpoints may be implemented
969 in gdb, e.g.@: as corefiles, as forked processes, and as some opaque
970 method implemented on the target side.
972 A corefile can be used to save an image of target memory and register
973 state, which can in principle be restored later --- but corefiles do
974 not typically include information about external entities such as
975 open files. Currently this method is not implemented in gdb.
977 A forked process can save the state of user memory and registers,
978 as well as some subset of external (kernel) state. This method
979 is used to implement checkpoints on Linux, and in principle might
980 be used on other systems.
982 Some targets, e.g.@: simulators, might have their own built-in
983 method for saving checkpoints, and gdb might be able to take
984 advantage of that capability without necessarily knowing any
985 details of how it is done.
988 @section Observing changes in @value{GDBN} internals
989 @cindex observer pattern interface
990 @cindex notifications about changes in internals
992 In order to function properly, several modules need to be notified when
993 some changes occur in the @value{GDBN} internals. Traditionally, these
994 modules have relied on several paradigms, the most common ones being
995 hooks and gdb-events. Unfortunately, none of these paradigms was
996 versatile enough to become the standard notification mechanism in
997 @value{GDBN}. The fact that they only supported one ``client'' was also
1000 A new paradigm, based on the Observer pattern of the @cite{Design
1001 Patterns} book, has therefore been implemented. The goal was to provide
1002 a new interface overcoming the issues with the notification mechanisms
1003 previously available. This new interface needed to be strongly typed,
1004 easy to extend, and versatile enough to be used as the standard
1005 interface when adding new notifications.
1007 See @ref{GDB Observers} for a brief description of the observers
1008 currently implemented in GDB. The rationale for the current
1009 implementation is also briefly discussed.
1011 @node User Interface
1013 @chapter User Interface
1015 @value{GDBN} has several user interfaces. Although the command-line interface
1016 is the most common and most familiar, there are others.
1018 @section Command Interpreter
1020 @cindex command interpreter
1022 The command interpreter in @value{GDBN} is fairly simple. It is designed to
1023 allow for the set of commands to be augmented dynamically, and also
1024 has a recursive subcommand capability, where the first argument to
1025 a command may itself direct a lookup on a different command list.
1027 For instance, the @samp{set} command just starts a lookup on the
1028 @code{setlist} command list, while @samp{set thread} recurses
1029 to the @code{set_thread_cmd_list}.
1033 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
1034 the main command list, and should be used for those commands. The usual
1035 place to add commands is in the @code{_initialize_@var{xyz}} routines at
1036 the ends of most source files.
1038 @findex add_setshow_cmd
1039 @findex add_setshow_cmd_full
1040 To add paired @samp{set} and @samp{show} commands, use
1041 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
1042 a slightly simpler interface which is useful when you don't need to
1043 further modify the new command structures, while the latter returns
1044 the new command structures for manipulation.
1046 @cindex deprecating commands
1047 @findex deprecate_cmd
1048 Before removing commands from the command set it is a good idea to
1049 deprecate them for some time. Use @code{deprecate_cmd} on commands or
1050 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
1051 @code{struct cmd_list_element} as it's first argument. You can use the
1052 return value from @code{add_com} or @code{add_cmd} to deprecate the
1053 command immediately after it is created.
1055 The first time a command is used the user will be warned and offered a
1056 replacement (if one exists). Note that the replacement string passed to
1057 @code{deprecate_cmd} should be the full name of the command, i.e. the
1058 entire string the user should type at the command line.
1060 @section UI-Independent Output---the @code{ui_out} Functions
1061 @c This section is based on the documentation written by Fernando
1062 @c Nasser <fnasser@redhat.com>.
1064 @cindex @code{ui_out} functions
1065 The @code{ui_out} functions present an abstraction level for the
1066 @value{GDBN} output code. They hide the specifics of different user
1067 interfaces supported by @value{GDBN}, and thus free the programmer
1068 from the need to write several versions of the same code, one each for
1069 every UI, to produce output.
1071 @subsection Overview and Terminology
1073 In general, execution of each @value{GDBN} command produces some sort
1074 of output, and can even generate an input request.
1076 Output can be generated for the following purposes:
1080 to display a @emph{result} of an operation;
1083 to convey @emph{info} or produce side-effects of a requested
1087 to provide a @emph{notification} of an asynchronous event (including
1088 progress indication of a prolonged asynchronous operation);
1091 to display @emph{error messages} (including warnings);
1094 to show @emph{debug data};
1097 to @emph{query} or prompt a user for input (a special case).
1101 This section mainly concentrates on how to build result output,
1102 although some of it also applies to other kinds of output.
1104 Generation of output that displays the results of an operation
1105 involves one or more of the following:
1109 output of the actual data
1112 formatting the output as appropriate for console output, to make it
1113 easily readable by humans
1116 machine oriented formatting--a more terse formatting to allow for easy
1117 parsing by programs which read @value{GDBN}'s output
1120 annotation, whose purpose is to help legacy GUIs to identify interesting
1124 The @code{ui_out} routines take care of the first three aspects.
1125 Annotations are provided by separate annotation routines. Note that use
1126 of annotations for an interface between a GUI and @value{GDBN} is
1129 Output can be in the form of a single item, which we call a @dfn{field};
1130 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
1131 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
1132 header and a body. In a BNF-like form:
1135 @item <table> @expansion{}
1136 @code{<header> <body>}
1137 @item <header> @expansion{}
1138 @code{@{ <column> @}}
1139 @item <column> @expansion{}
1140 @code{<width> <alignment> <title>}
1141 @item <body> @expansion{}
1146 @subsection General Conventions
1148 Most @code{ui_out} routines are of type @code{void}, the exceptions are
1149 @code{ui_out_stream_new} (which returns a pointer to the newly created
1150 object) and the @code{make_cleanup} routines.
1152 The first parameter is always the @code{ui_out} vector object, a pointer
1153 to a @code{struct ui_out}.
1155 The @var{format} parameter is like in @code{printf} family of functions.
1156 When it is present, there must also be a variable list of arguments
1157 sufficient used to satisfy the @code{%} specifiers in the supplied
1160 When a character string argument is not used in a @code{ui_out} function
1161 call, a @code{NULL} pointer has to be supplied instead.
1164 @subsection Table, Tuple and List Functions
1166 @cindex list output functions
1167 @cindex table output functions
1168 @cindex tuple output functions
1169 This section introduces @code{ui_out} routines for building lists,
1170 tuples and tables. The routines to output the actual data items
1171 (fields) are presented in the next section.
1173 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
1174 containing information about an object; a @dfn{list} is a sequence of
1175 fields where each field describes an identical object.
1177 Use the @dfn{table} functions when your output consists of a list of
1178 rows (tuples) and the console output should include a heading. Use this
1179 even when you are listing just one object but you still want the header.
1181 @cindex nesting level in @code{ui_out} functions
1182 Tables can not be nested. Tuples and lists can be nested up to a
1183 maximum of five levels.
1185 The overall structure of the table output code is something like this:
1200 Here is the description of table-, tuple- and list-related @code{ui_out}
1203 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
1204 The function @code{ui_out_table_begin} marks the beginning of the output
1205 of a table. It should always be called before any other @code{ui_out}
1206 function for a given table. @var{nbrofcols} is the number of columns in
1207 the table. @var{nr_rows} is the number of rows in the table.
1208 @var{tblid} is an optional string identifying the table. The string
1209 pointed to by @var{tblid} is copied by the implementation of
1210 @code{ui_out_table_begin}, so the application can free the string if it
1211 was @code{malloc}ed.
1213 The companion function @code{ui_out_table_end}, described below, marks
1214 the end of the table's output.
1217 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
1218 @code{ui_out_table_header} provides the header information for a single
1219 table column. You call this function several times, one each for every
1220 column of the table, after @code{ui_out_table_begin}, but before
1221 @code{ui_out_table_body}.
1223 The value of @var{width} gives the column width in characters. The
1224 value of @var{alignment} is one of @code{left}, @code{center}, and
1225 @code{right}, and it specifies how to align the header: left-justify,
1226 center, or right-justify it. @var{colhdr} points to a string that
1227 specifies the column header; the implementation copies that string, so
1228 column header strings in @code{malloc}ed storage can be freed after the
1232 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
1233 This function delimits the table header from the table body.
1236 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
1237 This function signals the end of a table's output. It should be called
1238 after the table body has been produced by the list and field output
1241 There should be exactly one call to @code{ui_out_table_end} for each
1242 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
1243 will signal an internal error.
1246 The output of the tuples that represent the table rows must follow the
1247 call to @code{ui_out_table_body} and precede the call to
1248 @code{ui_out_table_end}. You build a tuple by calling
1249 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
1250 calls to functions which actually output fields between them.
1252 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
1253 This function marks the beginning of a tuple output. @var{id} points
1254 to an optional string that identifies the tuple; it is copied by the
1255 implementation, and so strings in @code{malloc}ed storage can be freed
1259 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
1260 This function signals an end of a tuple output. There should be exactly
1261 one call to @code{ui_out_tuple_end} for each call to
1262 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
1266 @deftypefun struct cleanup *make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1267 This function first opens the tuple and then establishes a cleanup
1268 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
1269 and correct implementation of the non-portable@footnote{The function
1270 cast is not portable ISO C.} code sequence:
1272 struct cleanup *old_cleanup;
1273 ui_out_tuple_begin (uiout, "...");
1274 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
1279 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
1280 This function marks the beginning of a list output. @var{id} points to
1281 an optional string that identifies the list; it is copied by the
1282 implementation, and so strings in @code{malloc}ed storage can be freed
1286 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
1287 This function signals an end of a list output. There should be exactly
1288 one call to @code{ui_out_list_end} for each call to
1289 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1293 @deftypefun struct cleanup *make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1294 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1295 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
1296 that will close the list.list.
1299 @subsection Item Output Functions
1301 @cindex item output functions
1302 @cindex field output functions
1304 The functions described below produce output for the actual data
1305 items, or fields, which contain information about the object.
1307 Choose the appropriate function accordingly to your particular needs.
1309 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1310 This is the most general output function. It produces the
1311 representation of the data in the variable-length argument list
1312 according to formatting specifications in @var{format}, a
1313 @code{printf}-like format string. The optional argument @var{fldname}
1314 supplies the name of the field. The data items themselves are
1315 supplied as additional arguments after @var{format}.
1317 This generic function should be used only when it is not possible to
1318 use one of the specialized versions (see below).
1321 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1322 This function outputs a value of an @code{int} variable. It uses the
1323 @code{"%d"} output conversion specification. @var{fldname} specifies
1324 the name of the field.
1327 @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})
1328 This function outputs a value of an @code{int} variable. It differs from
1329 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1330 @var{fldname} specifies
1331 the name of the field.
1334 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1335 This function outputs an address.
1338 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1339 This function outputs a string using the @code{"%s"} conversion
1343 Sometimes, there's a need to compose your output piece by piece using
1344 functions that operate on a stream, such as @code{value_print} or
1345 @code{fprintf_symbol_filtered}. These functions accept an argument of
1346 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1347 used to store the data stream used for the output. When you use one
1348 of these functions, you need a way to pass their results stored in a
1349 @code{ui_file} object to the @code{ui_out} functions. To this end,
1350 you first create a @code{ui_stream} object by calling
1351 @code{ui_out_stream_new}, pass the @code{stream} member of that
1352 @code{ui_stream} object to @code{value_print} and similar functions,
1353 and finally call @code{ui_out_field_stream} to output the field you
1354 constructed. When the @code{ui_stream} object is no longer needed,
1355 you should destroy it and free its memory by calling
1356 @code{ui_out_stream_delete}.
1358 @deftypefun struct ui_stream *ui_out_stream_new (struct ui_out *@var{uiout})
1359 This function creates a new @code{ui_stream} object which uses the
1360 same output methods as the @code{ui_out} object whose pointer is
1361 passed in @var{uiout}. It returns a pointer to the newly created
1362 @code{ui_stream} object.
1365 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1366 This functions destroys a @code{ui_stream} object specified by
1370 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1371 This function consumes all the data accumulated in
1372 @code{streambuf->stream} and outputs it like
1373 @code{ui_out_field_string} does. After a call to
1374 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1375 the stream is still valid and may be used for producing more fields.
1378 @strong{Important:} If there is any chance that your code could bail
1379 out before completing output generation and reaching the point where
1380 @code{ui_out_stream_delete} is called, it is necessary to set up a
1381 cleanup, to avoid leaking memory and other resources. Here's a
1382 skeleton code to do that:
1385 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1386 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1391 If the function already has the old cleanup chain set (for other kinds
1392 of cleanups), you just have to add your cleanup to it:
1395 mybuf = ui_out_stream_new (uiout);
1396 make_cleanup (ui_out_stream_delete, mybuf);
1399 Note that with cleanups in place, you should not call
1400 @code{ui_out_stream_delete} directly, or you would attempt to free the
1403 @subsection Utility Output Functions
1405 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1406 This function skips a field in a table. Use it if you have to leave
1407 an empty field without disrupting the table alignment. The argument
1408 @var{fldname} specifies a name for the (missing) filed.
1411 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1412 This function outputs the text in @var{string} in a way that makes it
1413 easy to be read by humans. For example, the console implementation of
1414 this method filters the text through a built-in pager, to prevent it
1415 from scrolling off the visible portion of the screen.
1417 Use this function for printing relatively long chunks of text around
1418 the actual field data: the text it produces is not aligned according
1419 to the table's format. Use @code{ui_out_field_string} to output a
1420 string field, and use @code{ui_out_message}, described below, to
1421 output short messages.
1424 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1425 This function outputs @var{nspaces} spaces. It is handy to align the
1426 text produced by @code{ui_out_text} with the rest of the table or
1430 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1431 This function produces a formatted message, provided that the current
1432 verbosity level is at least as large as given by @var{verbosity}. The
1433 current verbosity level is specified by the user with the @samp{set
1434 verbositylevel} command.@footnote{As of this writing (April 2001),
1435 setting verbosity level is not yet implemented, and is always returned
1436 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1437 argument more than zero will cause the message to never be printed.}
1440 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1441 This function gives the console output filter (a paging filter) a hint
1442 of where to break lines which are too long. Ignored for all other
1443 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1444 be printed to indent the wrapped text on the next line; it must remain
1445 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1446 explicit newline is produced by one of the other functions. If
1447 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1450 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1451 This function flushes whatever output has been accumulated so far, if
1452 the UI buffers output.
1456 @subsection Examples of Use of @code{ui_out} functions
1458 @cindex using @code{ui_out} functions
1459 @cindex @code{ui_out} functions, usage examples
1460 This section gives some practical examples of using the @code{ui_out}
1461 functions to generalize the old console-oriented code in
1462 @value{GDBN}. The examples all come from functions defined on the
1463 @file{breakpoints.c} file.
1465 This example, from the @code{breakpoint_1} function, shows how to
1468 The original code was:
1471 if (!found_a_breakpoint++)
1473 annotate_breakpoints_headers ();
1476 printf_filtered ("Num ");
1478 printf_filtered ("Type ");
1480 printf_filtered ("Disp ");
1482 printf_filtered ("Enb ");
1486 printf_filtered ("Address ");
1489 printf_filtered ("What\n");
1491 annotate_breakpoints_table ();
1495 Here's the new version:
1498 nr_printable_breakpoints = @dots{};
1501 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1503 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1505 if (nr_printable_breakpoints > 0)
1506 annotate_breakpoints_headers ();
1507 if (nr_printable_breakpoints > 0)
1509 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1510 if (nr_printable_breakpoints > 0)
1512 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1513 if (nr_printable_breakpoints > 0)
1515 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1516 if (nr_printable_breakpoints > 0)
1518 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1521 if (nr_printable_breakpoints > 0)
1523 if (TARGET_ADDR_BIT <= 32)
1524 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1526 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1528 if (nr_printable_breakpoints > 0)
1530 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1531 ui_out_table_body (uiout);
1532 if (nr_printable_breakpoints > 0)
1533 annotate_breakpoints_table ();
1536 This example, from the @code{print_one_breakpoint} function, shows how
1537 to produce the actual data for the table whose structure was defined
1538 in the above example. The original code was:
1543 printf_filtered ("%-3d ", b->number);
1545 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1546 || ((int) b->type != bptypes[(int) b->type].type))
1547 internal_error ("bptypes table does not describe type #%d.",
1549 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1551 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1553 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1557 This is the new version:
1561 ui_out_tuple_begin (uiout, "bkpt");
1563 ui_out_field_int (uiout, "number", b->number);
1565 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1566 || ((int) b->type != bptypes[(int) b->type].type))
1567 internal_error ("bptypes table does not describe type #%d.",
1569 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1571 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1573 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1577 This example, also from @code{print_one_breakpoint}, shows how to
1578 produce a complicated output field using the @code{print_expression}
1579 functions which requires a stream to be passed. It also shows how to
1580 automate stream destruction with cleanups. The original code was:
1584 print_expression (b->exp, gdb_stdout);
1590 struct ui_stream *stb = ui_out_stream_new (uiout);
1591 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1594 print_expression (b->exp, stb->stream);
1595 ui_out_field_stream (uiout, "what", local_stream);
1598 This example, also from @code{print_one_breakpoint}, shows how to use
1599 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1604 if (b->dll_pathname == NULL)
1605 printf_filtered ("<any library> ");
1607 printf_filtered ("library \"%s\" ", b->dll_pathname);
1614 if (b->dll_pathname == NULL)
1616 ui_out_field_string (uiout, "what", "<any library>");
1617 ui_out_spaces (uiout, 1);
1621 ui_out_text (uiout, "library \"");
1622 ui_out_field_string (uiout, "what", b->dll_pathname);
1623 ui_out_text (uiout, "\" ");
1627 The following example from @code{print_one_breakpoint} shows how to
1628 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1633 if (b->forked_inferior_pid != 0)
1634 printf_filtered ("process %d ", b->forked_inferior_pid);
1641 if (b->forked_inferior_pid != 0)
1643 ui_out_text (uiout, "process ");
1644 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1645 ui_out_spaces (uiout, 1);
1649 Here's an example of using @code{ui_out_field_string}. The original
1654 if (b->exec_pathname != NULL)
1655 printf_filtered ("program \"%s\" ", b->exec_pathname);
1662 if (b->exec_pathname != NULL)
1664 ui_out_text (uiout, "program \"");
1665 ui_out_field_string (uiout, "what", b->exec_pathname);
1666 ui_out_text (uiout, "\" ");
1670 Finally, here's an example of printing an address. The original code:
1674 printf_filtered ("%s ",
1675 hex_string_custom ((unsigned long) b->address, 8));
1682 ui_out_field_core_addr (uiout, "Address", b->address);
1686 @section Console Printing
1695 @cindex @code{libgdb}
1696 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1697 to provide an API to @value{GDBN}'s functionality.
1700 @cindex @code{libgdb}
1701 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1702 better able to support graphical and other environments.
1704 Since @code{libgdb} development is on-going, its architecture is still
1705 evolving. The following components have so far been identified:
1709 Observer - @file{gdb-events.h}.
1711 Builder - @file{ui-out.h}
1713 Event Loop - @file{event-loop.h}
1715 Library - @file{gdb.h}
1718 The model that ties these components together is described below.
1720 @section The @code{libgdb} Model
1722 A client of @code{libgdb} interacts with the library in two ways.
1726 As an observer (using @file{gdb-events}) receiving notifications from
1727 @code{libgdb} of any internal state changes (break point changes, run
1730 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1731 obtain various status values from @value{GDBN}.
1734 Since @code{libgdb} could have multiple clients (e.g., a GUI supporting
1735 the existing @value{GDBN} CLI), those clients must co-operate when
1736 controlling @code{libgdb}. In particular, a client must ensure that
1737 @code{libgdb} is idle (i.e. no other client is using @code{libgdb})
1738 before responding to a @file{gdb-event} by making a query.
1740 @section CLI support
1742 At present @value{GDBN}'s CLI is very much entangled in with the core of
1743 @code{libgdb}. Consequently, a client wishing to include the CLI in
1744 their interface needs to carefully co-ordinate its own and the CLI's
1747 It is suggested that the client set @code{libgdb} up to be bi-modal
1748 (alternate between CLI and client query modes). The notes below sketch
1753 The client registers itself as an observer of @code{libgdb}.
1755 The client create and install @code{cli-out} builder using its own
1756 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1757 @code{gdb_stdout} streams.
1759 The client creates a separate custom @code{ui-out} builder that is only
1760 used while making direct queries to @code{libgdb}.
1763 When the client receives input intended for the CLI, it simply passes it
1764 along. Since the @code{cli-out} builder is installed by default, all
1765 the CLI output in response to that command is routed (pronounced rooted)
1766 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1767 At the same time, the client is kept abreast of internal changes by
1768 virtue of being a @code{libgdb} observer.
1770 The only restriction on the client is that it must wait until
1771 @code{libgdb} becomes idle before initiating any queries (using the
1772 client's custom builder).
1774 @section @code{libgdb} components
1776 @subheading Observer - @file{gdb-events.h}
1777 @file{gdb-events} provides the client with a very raw mechanism that can
1778 be used to implement an observer. At present it only allows for one
1779 observer and that observer must, internally, handle the need to delay
1780 the processing of any event notifications until after @code{libgdb} has
1781 finished the current command.
1783 @subheading Builder - @file{ui-out.h}
1784 @file{ui-out} provides the infrastructure necessary for a client to
1785 create a builder. That builder is then passed down to @code{libgdb}
1786 when doing any queries.
1788 @subheading Event Loop - @file{event-loop.h}
1789 @c There could be an entire section on the event-loop
1790 @file{event-loop}, currently non-re-entrant, provides a simple event
1791 loop. A client would need to either plug its self into this loop or,
1792 implement a new event-loop that GDB would use.
1794 The event-loop will eventually be made re-entrant. This is so that
1795 @value{GDBN} can better handle the problem of some commands blocking
1796 instead of returning.
1798 @subheading Library - @file{gdb.h}
1799 @file{libgdb} is the most obvious component of this system. It provides
1800 the query interface. Each function is parameterized by a @code{ui-out}
1801 builder. The result of the query is constructed using that builder
1802 before the query function returns.
1804 @node Symbol Handling
1806 @chapter Symbol Handling
1808 Symbols are a key part of @value{GDBN}'s operation. Symbols include variables,
1809 functions, and types.
1811 @section Symbol Reading
1813 @cindex symbol reading
1814 @cindex reading of symbols
1815 @cindex symbol files
1816 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
1817 file is the file containing the program which @value{GDBN} is
1818 debugging. @value{GDBN} can be directed to use a different file for
1819 symbols (with the @samp{symbol-file} command), and it can also read
1820 more symbols via the @samp{add-file} and @samp{load} commands, or while
1821 reading symbols from shared libraries.
1823 @findex find_sym_fns
1824 Symbol files are initially opened by code in @file{symfile.c} using
1825 the BFD library (@pxref{Support Libraries}). BFD identifies the type
1826 of the file by examining its header. @code{find_sym_fns} then uses
1827 this identification to locate a set of symbol-reading functions.
1829 @findex add_symtab_fns
1830 @cindex @code{sym_fns} structure
1831 @cindex adding a symbol-reading module
1832 Symbol-reading modules identify themselves to @value{GDBN} by calling
1833 @code{add_symtab_fns} during their module initialization. The argument
1834 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
1835 name (or name prefix) of the symbol format, the length of the prefix,
1836 and pointers to four functions. These functions are called at various
1837 times to process symbol files whose identification matches the specified
1840 The functions supplied by each module are:
1843 @item @var{xyz}_symfile_init(struct sym_fns *sf)
1845 @cindex secondary symbol file
1846 Called from @code{symbol_file_add} when we are about to read a new
1847 symbol file. This function should clean up any internal state (possibly
1848 resulting from half-read previous files, for example) and prepare to
1849 read a new symbol file. Note that the symbol file which we are reading
1850 might be a new ``main'' symbol file, or might be a secondary symbol file
1851 whose symbols are being added to the existing symbol table.
1853 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
1854 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
1855 new symbol file being read. Its @code{private} field has been zeroed,
1856 and can be modified as desired. Typically, a struct of private
1857 information will be @code{malloc}'d, and a pointer to it will be placed
1858 in the @code{private} field.
1860 There is no result from @code{@var{xyz}_symfile_init}, but it can call
1861 @code{error} if it detects an unavoidable problem.
1863 @item @var{xyz}_new_init()
1865 Called from @code{symbol_file_add} when discarding existing symbols.
1866 This function needs only handle the symbol-reading module's internal
1867 state; the symbol table data structures visible to the rest of
1868 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
1869 arguments and no result. It may be called after
1870 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
1871 may be called alone if all symbols are simply being discarded.
1873 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
1875 Called from @code{symbol_file_add} to actually read the symbols from a
1876 symbol-file into a set of psymtabs or symtabs.
1878 @code{sf} points to the @code{struct sym_fns} originally passed to
1879 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
1880 the offset between the file's specified start address and its true
1881 address in memory. @code{mainline} is 1 if this is the main symbol
1882 table being read, and 0 if a secondary symbol file (e.g., shared library
1883 or dynamically loaded file) is being read.@refill
1886 In addition, if a symbol-reading module creates psymtabs when
1887 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
1888 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
1889 from any point in the @value{GDBN} symbol-handling code.
1892 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
1894 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
1895 the psymtab has not already been read in and had its @code{pst->symtab}
1896 pointer set. The argument is the psymtab to be fleshed-out into a
1897 symtab. Upon return, @code{pst->readin} should have been set to 1, and
1898 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
1899 zero if there were no symbols in that part of the symbol file.
1902 @section Partial Symbol Tables
1904 @value{GDBN} has three types of symbol tables:
1907 @cindex full symbol table
1910 Full symbol tables (@dfn{symtabs}). These contain the main
1911 information about symbols and addresses.
1915 Partial symbol tables (@dfn{psymtabs}). These contain enough
1916 information to know when to read the corresponding part of the full
1919 @cindex minimal symbol table
1922 Minimal symbol tables (@dfn{msymtabs}). These contain information
1923 gleaned from non-debugging symbols.
1926 @cindex partial symbol table
1927 This section describes partial symbol tables.
1929 A psymtab is constructed by doing a very quick pass over an executable
1930 file's debugging information. Small amounts of information are
1931 extracted---enough to identify which parts of the symbol table will
1932 need to be re-read and fully digested later, when the user needs the
1933 information. The speed of this pass causes @value{GDBN} to start up very
1934 quickly. Later, as the detailed rereading occurs, it occurs in small
1935 pieces, at various times, and the delay therefrom is mostly invisible to
1937 @c (@xref{Symbol Reading}.)
1939 The symbols that show up in a file's psymtab should be, roughly, those
1940 visible to the debugger's user when the program is not running code from
1941 that file. These include external symbols and types, static symbols and
1942 types, and @code{enum} values declared at file scope.
1944 The psymtab also contains the range of instruction addresses that the
1945 full symbol table would represent.
1947 @cindex finding a symbol
1948 @cindex symbol lookup
1949 The idea is that there are only two ways for the user (or much of the
1950 code in the debugger) to reference a symbol:
1953 @findex find_pc_function
1954 @findex find_pc_line
1956 By its address (e.g., execution stops at some address which is inside a
1957 function in this file). The address will be noticed to be in the
1958 range of this psymtab, and the full symtab will be read in.
1959 @code{find_pc_function}, @code{find_pc_line}, and other
1960 @code{find_pc_@dots{}} functions handle this.
1962 @cindex lookup_symbol
1965 (e.g., the user asks to print a variable, or set a breakpoint on a
1966 function). Global names and file-scope names will be found in the
1967 psymtab, which will cause the symtab to be pulled in. Local names will
1968 have to be qualified by a global name, or a file-scope name, in which
1969 case we will have already read in the symtab as we evaluated the
1970 qualifier. Or, a local symbol can be referenced when we are ``in'' a
1971 local scope, in which case the first case applies. @code{lookup_symbol}
1972 does most of the work here.
1975 The only reason that psymtabs exist is to cause a symtab to be read in
1976 at the right moment. Any symbol that can be elided from a psymtab,
1977 while still causing that to happen, should not appear in it. Since
1978 psymtabs don't have the idea of scope, you can't put local symbols in
1979 them anyway. Psymtabs don't have the idea of the type of a symbol,
1980 either, so types need not appear, unless they will be referenced by
1983 It is a bug for @value{GDBN} to behave one way when only a psymtab has
1984 been read, and another way if the corresponding symtab has been read
1985 in. Such bugs are typically caused by a psymtab that does not contain
1986 all the visible symbols, or which has the wrong instruction address
1989 The psymtab for a particular section of a symbol file (objfile) could be
1990 thrown away after the symtab has been read in. The symtab should always
1991 be searched before the psymtab, so the psymtab will never be used (in a
1992 bug-free environment). Currently, psymtabs are allocated on an obstack,
1993 and all the psymbols themselves are allocated in a pair of large arrays
1994 on an obstack, so there is little to be gained by trying to free them
1995 unless you want to do a lot more work.
1999 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
2001 @cindex fundamental types
2002 These are the fundamental types that @value{GDBN} uses internally. Fundamental
2003 types from the various debugging formats (stabs, ELF, etc) are mapped
2004 into one of these. They are basically a union of all fundamental types
2005 that @value{GDBN} knows about for all the languages that @value{GDBN}
2008 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
2011 Each time @value{GDBN} builds an internal type, it marks it with one
2012 of these types. The type may be a fundamental type, such as
2013 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
2014 which is a pointer to another type. Typically, several @code{FT_*}
2015 types map to one @code{TYPE_CODE_*} type, and are distinguished by
2016 other members of the type struct, such as whether the type is signed
2017 or unsigned, and how many bits it uses.
2019 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
2021 These are instances of type structs that roughly correspond to
2022 fundamental types and are created as global types for @value{GDBN} to
2023 use for various ugly historical reasons. We eventually want to
2024 eliminate these. Note for example that @code{builtin_type_int}
2025 initialized in @file{gdbtypes.c} is basically the same as a
2026 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
2027 an @code{FT_INTEGER} fundamental type. The difference is that the
2028 @code{builtin_type} is not associated with any particular objfile, and
2029 only one instance exists, while @file{c-lang.c} builds as many
2030 @code{TYPE_CODE_INT} types as needed, with each one associated with
2031 some particular objfile.
2033 @section Object File Formats
2034 @cindex object file formats
2038 @cindex @code{a.out} format
2039 The @code{a.out} format is the original file format for Unix. It
2040 consists of three sections: @code{text}, @code{data}, and @code{bss},
2041 which are for program code, initialized data, and uninitialized data,
2044 The @code{a.out} format is so simple that it doesn't have any reserved
2045 place for debugging information. (Hey, the original Unix hackers used
2046 @samp{adb}, which is a machine-language debugger!) The only debugging
2047 format for @code{a.out} is stabs, which is encoded as a set of normal
2048 symbols with distinctive attributes.
2050 The basic @code{a.out} reader is in @file{dbxread.c}.
2055 The COFF format was introduced with System V Release 3 (SVR3) Unix.
2056 COFF files may have multiple sections, each prefixed by a header. The
2057 number of sections is limited.
2059 The COFF specification includes support for debugging. Although this
2060 was a step forward, the debugging information was woefully limited. For
2061 instance, it was not possible to represent code that came from an
2064 The COFF reader is in @file{coffread.c}.
2068 @cindex ECOFF format
2069 ECOFF is an extended COFF originally introduced for Mips and Alpha
2072 The basic ECOFF reader is in @file{mipsread.c}.
2076 @cindex XCOFF format
2077 The IBM RS/6000 running AIX uses an object file format called XCOFF.
2078 The COFF sections, symbols, and line numbers are used, but debugging
2079 symbols are @code{dbx}-style stabs whose strings are located in the
2080 @code{.debug} section (rather than the string table). For more
2081 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
2083 The shared library scheme has a clean interface for figuring out what
2084 shared libraries are in use, but the catch is that everything which
2085 refers to addresses (symbol tables and breakpoints at least) needs to be
2086 relocated for both shared libraries and the main executable. At least
2087 using the standard mechanism this can only be done once the program has
2088 been run (or the core file has been read).
2092 @cindex PE-COFF format
2093 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
2094 executables. PE is basically COFF with additional headers.
2096 While BFD includes special PE support, @value{GDBN} needs only the basic
2102 The ELF format came with System V Release 4 (SVR4) Unix. ELF is similar
2103 to COFF in being organized into a number of sections, but it removes
2104 many of COFF's limitations.
2106 The basic ELF reader is in @file{elfread.c}.
2111 SOM is HP's object file and debug format (not to be confused with IBM's
2112 SOM, which is a cross-language ABI).
2114 The SOM reader is in @file{hpread.c}.
2116 @subsection Other File Formats
2118 @cindex Netware Loadable Module format
2119 Other file formats that have been supported by @value{GDBN} include Netware
2120 Loadable Modules (@file{nlmread.c}).
2122 @section Debugging File Formats
2124 This section describes characteristics of debugging information that
2125 are independent of the object file format.
2129 @cindex stabs debugging info
2130 @code{stabs} started out as special symbols within the @code{a.out}
2131 format. Since then, it has been encapsulated into other file
2132 formats, such as COFF and ELF.
2134 While @file{dbxread.c} does some of the basic stab processing,
2135 including for encapsulated versions, @file{stabsread.c} does
2140 @cindex COFF debugging info
2141 The basic COFF definition includes debugging information. The level
2142 of support is minimal and non-extensible, and is not often used.
2144 @subsection Mips debug (Third Eye)
2146 @cindex ECOFF debugging info
2147 ECOFF includes a definition of a special debug format.
2149 The file @file{mdebugread.c} implements reading for this format.
2153 @cindex DWARF 1 debugging info
2154 DWARF 1 is a debugging format that was originally designed to be
2155 used with ELF in SVR4 systems.
2160 @c If defined, these are the producer strings in a DWARF 1 file. All of
2161 @c these have reasonable defaults already.
2163 The DWARF 1 reader is in @file{dwarfread.c}.
2167 @cindex DWARF 2 debugging info
2168 DWARF 2 is an improved but incompatible version of DWARF 1.
2170 The DWARF 2 reader is in @file{dwarf2read.c}.
2174 @cindex SOM debugging info
2175 Like COFF, the SOM definition includes debugging information.
2177 @section Adding a New Symbol Reader to @value{GDBN}
2179 @cindex adding debugging info reader
2180 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
2181 there is probably little to be done.
2183 If you need to add a new object file format, you must first add it to
2184 BFD. This is beyond the scope of this document.
2186 You must then arrange for the BFD code to provide access to the
2187 debugging symbols. Generally @value{GDBN} will have to call swapping routines
2188 from BFD and a few other BFD internal routines to locate the debugging
2189 information. As much as possible, @value{GDBN} should not depend on the BFD
2190 internal data structures.
2192 For some targets (e.g., COFF), there is a special transfer vector used
2193 to call swapping routines, since the external data structures on various
2194 platforms have different sizes and layouts. Specialized routines that
2195 will only ever be implemented by one object file format may be called
2196 directly. This interface should be described in a file
2197 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
2199 @section Memory Management for Symbol Files
2201 Most memory associated with a loaded symbol file is stored on
2202 its @code{objfile_obstack}. This includes symbols, types,
2203 namespace data, and other information produced by the symbol readers.
2205 Because this data lives on the objfile's obstack, it is automatically
2206 released when the objfile is unloaded or reloaded. Therefore one
2207 objfile must not reference symbol or type data from another objfile;
2208 they could be unloaded at different times.
2210 User convenience variables, et cetera, have associated types. Normally
2211 these types live in the associated objfile. However, when the objfile
2212 is unloaded, those types are deep copied to global memory, so that
2213 the values of the user variables and history items are not lost.
2216 @node Language Support
2218 @chapter Language Support
2220 @cindex language support
2221 @value{GDBN}'s language support is mainly driven by the symbol reader,
2222 although it is possible for the user to set the source language
2225 @value{GDBN} chooses the source language by looking at the extension
2226 of the file recorded in the debug info; @file{.c} means C, @file{.f}
2227 means Fortran, etc. It may also use a special-purpose language
2228 identifier if the debug format supports it, like with DWARF.
2230 @section Adding a Source Language to @value{GDBN}
2232 @cindex adding source language
2233 To add other languages to @value{GDBN}'s expression parser, follow the
2237 @item Create the expression parser.
2239 @cindex expression parser
2240 This should reside in a file @file{@var{lang}-exp.y}. Routines for
2241 building parsed expressions into a @code{union exp_element} list are in
2244 @cindex language parser
2245 Since we can't depend upon everyone having Bison, and YACC produces
2246 parsers that define a bunch of global names, the following lines
2247 @strong{must} be included at the top of the YACC parser, to prevent the
2248 various parsers from defining the same global names:
2251 #define yyparse @var{lang}_parse
2252 #define yylex @var{lang}_lex
2253 #define yyerror @var{lang}_error
2254 #define yylval @var{lang}_lval
2255 #define yychar @var{lang}_char
2256 #define yydebug @var{lang}_debug
2257 #define yypact @var{lang}_pact
2258 #define yyr1 @var{lang}_r1
2259 #define yyr2 @var{lang}_r2
2260 #define yydef @var{lang}_def
2261 #define yychk @var{lang}_chk
2262 #define yypgo @var{lang}_pgo
2263 #define yyact @var{lang}_act
2264 #define yyexca @var{lang}_exca
2265 #define yyerrflag @var{lang}_errflag
2266 #define yynerrs @var{lang}_nerrs
2269 At the bottom of your parser, define a @code{struct language_defn} and
2270 initialize it with the right values for your language. Define an
2271 @code{initialize_@var{lang}} routine and have it call
2272 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
2273 that your language exists. You'll need some other supporting variables
2274 and functions, which will be used via pointers from your
2275 @code{@var{lang}_language_defn}. See the declaration of @code{struct
2276 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
2277 for more information.
2279 @item Add any evaluation routines, if necessary
2281 @cindex expression evaluation routines
2282 @findex evaluate_subexp
2283 @findex prefixify_subexp
2284 @findex length_of_subexp
2285 If you need new opcodes (that represent the operations of the language),
2286 add them to the enumerated type in @file{expression.h}. Add support
2287 code for these operations in the @code{evaluate_subexp} function
2288 defined in the file @file{eval.c}. Add cases
2289 for new opcodes in two functions from @file{parse.c}:
2290 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
2291 the number of @code{exp_element}s that a given operation takes up.
2293 @item Update some existing code
2295 Add an enumerated identifier for your language to the enumerated type
2296 @code{enum language} in @file{defs.h}.
2298 Update the routines in @file{language.c} so your language is included.
2299 These routines include type predicates and such, which (in some cases)
2300 are language dependent. If your language does not appear in the switch
2301 statement, an error is reported.
2303 @vindex current_language
2304 Also included in @file{language.c} is the code that updates the variable
2305 @code{current_language}, and the routines that translate the
2306 @code{language_@var{lang}} enumerated identifier into a printable
2309 @findex _initialize_language
2310 Update the function @code{_initialize_language} to include your
2311 language. This function picks the default language upon startup, so is
2312 dependent upon which languages that @value{GDBN} is built for.
2314 @findex allocate_symtab
2315 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2316 code so that the language of each symtab (source file) is set properly.
2317 This is used to determine the language to use at each stack frame level.
2318 Currently, the language is set based upon the extension of the source
2319 file. If the language can be better inferred from the symbol
2320 information, please set the language of the symtab in the symbol-reading
2323 @findex print_subexp
2324 @findex op_print_tab
2325 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2326 expression opcodes you have added to @file{expression.h}. Also, add the
2327 printed representations of your operators to @code{op_print_tab}.
2329 @item Add a place of call
2332 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2333 @code{parse_exp_1} (defined in @file{parse.c}).
2335 @item Use macros to trim code
2337 @cindex trimming language-dependent code
2338 The user has the option of building @value{GDBN} for some or all of the
2339 languages. If the user decides to build @value{GDBN} for the language
2340 @var{lang}, then every file dependent on @file{language.h} will have the
2341 macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to
2342 leave out large routines that the user won't need if he or she is not
2343 using your language.
2345 Note that you do not need to do this in your YACC parser, since if @value{GDBN}
2346 is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the
2347 compiled form of your parser) is not linked into @value{GDBN} at all.
2349 See the file @file{configure.in} for how @value{GDBN} is configured
2350 for different languages.
2352 @item Edit @file{Makefile.in}
2354 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2355 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2356 not get linked in, or, worse yet, it may not get @code{tar}red into the
2361 @node Host Definition
2363 @chapter Host Definition
2365 With the advent of Autoconf, it's rarely necessary to have host
2366 definition machinery anymore. The following information is provided,
2367 mainly, as an historical reference.
2369 @section Adding a New Host
2371 @cindex adding a new host
2372 @cindex host, adding
2373 @value{GDBN}'s host configuration support normally happens via Autoconf.
2374 New host-specific definitions should not be needed. Older hosts
2375 @value{GDBN} still use the host-specific definitions and files listed
2376 below, but these mostly exist for historical reasons, and will
2377 eventually disappear.
2380 @item gdb/config/@var{arch}/@var{xyz}.mh
2381 This file once contained both host and native configuration information
2382 (@pxref{Native Debugging}) for the machine @var{xyz}. The host
2383 configuration information is now handed by Autoconf.
2385 Host configuration information included a definition of
2386 @code{XM_FILE=xm-@var{xyz}.h} and possibly definitions for @code{CC},
2387 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2388 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2390 New host only configurations do not need this file.
2392 @item gdb/config/@var{arch}/xm-@var{xyz}.h
2393 This file once contained definitions and includes required when hosting
2394 gdb on machine @var{xyz}. Those definitions and includes are now
2395 handled by Autoconf.
2397 New host and native configurations do not need this file.
2399 @emph{Maintainer's note: Some hosts continue to use the @file{xm-xyz.h}
2400 file to define the macros @var{HOST_FLOAT_FORMAT},
2401 @var{HOST_DOUBLE_FORMAT} and @var{HOST_LONG_DOUBLE_FORMAT}. That code
2402 also needs to be replaced with either an Autoconf or run-time test.}
2406 @subheading Generic Host Support Files
2408 @cindex generic host support
2409 There are some ``generic'' versions of routines that can be used by
2410 various systems. These can be customized in various ways by macros
2411 defined in your @file{xm-@var{xyz}.h} file. If these routines work for
2412 the @var{xyz} host, you can just include the generic file's name (with
2413 @samp{.o}, not @samp{.c}) in @code{XDEPFILES}.
2415 Otherwise, if your machine needs custom support routines, you will need
2416 to write routines that perform the same functions as the generic file.
2417 Put them into @code{@var{xyz}-xdep.c}, and put @code{@var{xyz}-xdep.o}
2418 into @code{XDEPFILES}.
2421 @cindex remote debugging support
2422 @cindex serial line support
2424 This contains serial line support for Unix systems. This is always
2425 included, via the makefile variable @code{SER_HARDWIRE}; override this
2426 variable in the @file{.mh} file to avoid it.
2429 This contains serial line support for 32-bit programs running under DOS,
2430 using the DJGPP (a.k.a.@: GO32) execution environment.
2432 @cindex TCP remote support
2434 This contains generic TCP support using sockets.
2437 @section Host Conditionals
2439 When @value{GDBN} is configured and compiled, various macros are
2440 defined or left undefined, to control compilation based on the
2441 attributes of the host system. These macros and their meanings (or if
2442 the meaning is not documented here, then one of the source files where
2443 they are used is indicated) are:
2446 @item @value{GDBN}INIT_FILENAME
2447 The default name of @value{GDBN}'s initialization file (normally
2451 This macro is deprecated.
2453 @item SIGWINCH_HANDLER
2454 If your host defines @code{SIGWINCH}, you can define this to be the name
2455 of a function to be called if @code{SIGWINCH} is received.
2457 @item SIGWINCH_HANDLER_BODY
2458 Define this to expand into code that will define the function named by
2459 the expansion of @code{SIGWINCH_HANDLER}.
2461 @item ALIGN_STACK_ON_STARTUP
2462 @cindex stack alignment
2463 Define this if your system is of a sort that will crash in
2464 @code{tgetent} if the stack happens not to be longword-aligned when
2465 @code{main} is called. This is a rare situation, but is known to occur
2466 on several different types of systems.
2468 @item CRLF_SOURCE_FILES
2469 @cindex DOS text files
2470 Define this if host files use @code{\r\n} rather than @code{\n} as a
2471 line terminator. This will cause source file listings to omit @code{\r}
2472 characters when printing and it will allow @code{\r\n} line endings of files
2473 which are ``sourced'' by gdb. It must be possible to open files in binary
2474 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2476 @item DEFAULT_PROMPT
2478 The default value of the prompt string (normally @code{"(gdb) "}).
2481 @cindex terminal device
2482 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2485 Define this if binary files are opened the same way as text files.
2489 In some cases, use the system call @code{mmap} for reading symbol
2490 tables. For some machines this allows for sharing and quick updates.
2493 Define this if the host system has @code{termio.h}.
2500 Values for host-side constants.
2503 Substitute for isatty, if not available.
2506 This is the longest integer type available on the host. If not defined,
2507 it will default to @code{long long} or @code{long}, depending on
2508 @code{CC_HAS_LONG_LONG}.
2510 @item CC_HAS_LONG_LONG
2511 @cindex @code{long long} data type
2512 Define this if the host C compiler supports @code{long long}. This is set
2513 by the @code{configure} script.
2515 @item PRINTF_HAS_LONG_LONG
2516 Define this if the host can handle printing of long long integers via
2517 the printf format conversion specifier @code{ll}. This is set by the
2518 @code{configure} script.
2520 @item HAVE_LONG_DOUBLE
2521 Define this if the host C compiler supports @code{long double}. This is
2522 set by the @code{configure} script.
2524 @item PRINTF_HAS_LONG_DOUBLE
2525 Define this if the host can handle printing of long double float-point
2526 numbers via the printf format conversion specifier @code{Lg}. This is
2527 set by the @code{configure} script.
2529 @item SCANF_HAS_LONG_DOUBLE
2530 Define this if the host can handle the parsing of long double
2531 float-point numbers via the scanf format conversion specifier
2532 @code{Lg}. This is set by the @code{configure} script.
2534 @item LSEEK_NOT_LINEAR
2535 Define this if @code{lseek (n)} does not necessarily move to byte number
2536 @code{n} in the file. This is only used when reading source files. It
2537 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2540 This macro is used as the argument to @code{lseek} (or, most commonly,
2541 @code{bfd_seek}). FIXME, should be replaced by SEEK_SET instead,
2542 which is the POSIX equivalent.
2545 If defined, this should be one or more tokens, such as @code{volatile},
2546 that can be used in both the declaration and definition of functions to
2547 indicate that they never return. The default is already set correctly
2548 if compiling with GCC. This will almost never need to be defined.
2551 If defined, this should be one or more tokens, such as
2552 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2553 of functions to indicate that they never return. The default is already
2554 set correctly if compiling with GCC. This will almost never need to be
2559 Define these to appropriate value for the system @code{lseek}, if not already
2563 This is the signal for stopping @value{GDBN}. Defaults to
2564 @code{SIGTSTP}. (Only redefined for the Convex.)
2567 Means that System V (prior to SVR4) include files are in use. (FIXME:
2568 This symbol is abused in @file{infrun.c}, @file{regex.c}, and
2569 @file{utils.c} for other things, at the moment.)
2572 Define this to help placate @code{lint} in some situations.
2575 Define this to override the defaults of @code{__volatile__} or
2580 @node Target Architecture Definition
2582 @chapter Target Architecture Definition
2584 @cindex target architecture definition
2585 @value{GDBN}'s target architecture defines what sort of
2586 machine-language programs @value{GDBN} can work with, and how it works
2589 The target architecture object is implemented as the C structure
2590 @code{struct gdbarch *}. The structure, and its methods, are generated
2591 using the Bourne shell script @file{gdbarch.sh}.
2593 @section Operating System ABI Variant Handling
2594 @cindex OS ABI variants
2596 @value{GDBN} provides a mechanism for handling variations in OS
2597 ABIs. An OS ABI variant may have influence over any number of
2598 variables in the target architecture definition. There are two major
2599 components in the OS ABI mechanism: sniffers and handlers.
2601 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2602 (the architecture may be wildcarded) in an attempt to determine the
2603 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2604 to be @dfn{generic}, while sniffers for a specific architecture are
2605 considered to be @dfn{specific}. A match from a specific sniffer
2606 overrides a match from a generic sniffer. Multiple sniffers for an
2607 architecture/flavour may exist, in order to differentiate between two
2608 different operating systems which use the same basic file format. The
2609 OS ABI framework provides a generic sniffer for ELF-format files which
2610 examines the @code{EI_OSABI} field of the ELF header, as well as note
2611 sections known to be used by several operating systems.
2613 @cindex fine-tuning @code{gdbarch} structure
2614 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2615 selected OS ABI. There may be only one handler for a given OS ABI
2616 for each BFD architecture.
2618 The following OS ABI variants are defined in @file{osabi.h}:
2622 @findex GDB_OSABI_UNKNOWN
2623 @item GDB_OSABI_UNKNOWN
2624 The ABI of the inferior is unknown. The default @code{gdbarch}
2625 settings for the architecture will be used.
2627 @findex GDB_OSABI_SVR4
2628 @item GDB_OSABI_SVR4
2629 UNIX System V Release 4
2631 @findex GDB_OSABI_HURD
2632 @item GDB_OSABI_HURD
2633 GNU using the Hurd kernel
2635 @findex GDB_OSABI_SOLARIS
2636 @item GDB_OSABI_SOLARIS
2639 @findex GDB_OSABI_OSF1
2640 @item GDB_OSABI_OSF1
2641 OSF/1, including Digital UNIX and Compaq Tru64 UNIX
2643 @findex GDB_OSABI_LINUX
2644 @item GDB_OSABI_LINUX
2645 GNU using the Linux kernel
2647 @findex GDB_OSABI_FREEBSD_AOUT
2648 @item GDB_OSABI_FREEBSD_AOUT
2649 FreeBSD using the a.out executable format
2651 @findex GDB_OSABI_FREEBSD_ELF
2652 @item GDB_OSABI_FREEBSD_ELF
2653 FreeBSD using the ELF executable format
2655 @findex GDB_OSABI_NETBSD_AOUT
2656 @item GDB_OSABI_NETBSD_AOUT
2657 NetBSD using the a.out executable format
2659 @findex GDB_OSABI_NETBSD_ELF
2660 @item GDB_OSABI_NETBSD_ELF
2661 NetBSD using the ELF executable format
2663 @findex GDB_OSABI_WINCE
2664 @item GDB_OSABI_WINCE
2667 @findex GDB_OSABI_GO32
2668 @item GDB_OSABI_GO32
2671 @findex GDB_OSABI_NETWARE
2672 @item GDB_OSABI_NETWARE
2675 @findex GDB_OSABI_ARM_EABI_V1
2676 @item GDB_OSABI_ARM_EABI_V1
2677 ARM Embedded ABI version 1
2679 @findex GDB_OSABI_ARM_EABI_V2
2680 @item GDB_OSABI_ARM_EABI_V2
2681 ARM Embedded ABI version 2
2683 @findex GDB_OSABI_ARM_APCS
2684 @item GDB_OSABI_ARM_APCS
2685 Generic ARM Procedure Call Standard
2689 Here are the functions that make up the OS ABI framework:
2691 @deftypefun const char *gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2692 Return the name of the OS ABI corresponding to @var{osabi}.
2695 @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}))
2696 Register the OS ABI handler specified by @var{init_osabi} for the
2697 architecture, machine type and OS ABI specified by @var{arch},
2698 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2699 machine type, which implies the architecture's default machine type,
2703 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2704 Register the OS ABI file sniffer specified by @var{sniffer} for the
2705 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2706 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2707 be generic, and is allowed to examine @var{flavour}-flavoured files for
2711 @deftypefun enum gdb_osabi gdbarch_lookup_osabi (bfd *@var{abfd})
2712 Examine the file described by @var{abfd} to determine its OS ABI.
2713 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2717 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2718 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2719 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2720 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2721 architecture, a warning will be issued and the debugging session will continue
2722 with the defaults already established for @var{gdbarch}.
2725 @section Initializing a New Architecture
2727 Each @code{gdbarch} is associated with a single @sc{bfd} architecture,
2728 via a @code{bfd_arch_@var{arch}} constant. The @code{gdbarch} is
2729 registered by a call to @code{register_gdbarch_init}, usually from
2730 the file's @code{_initialize_@var{filename}} routine, which will
2731 be automatically called during @value{GDBN} startup. The arguments
2732 are a @sc{bfd} architecture constant and an initialization function.
2734 The initialization function has this type:
2737 static struct gdbarch *
2738 @var{arch}_gdbarch_init (struct gdbarch_info @var{info},
2739 struct gdbarch_list *@var{arches})
2742 The @var{info} argument contains parameters used to select the correct
2743 architecture, and @var{arches} is a list of architectures which
2744 have already been created with the same @code{bfd_arch_@var{arch}}
2747 The initialization function should first make sure that @var{info}
2748 is acceptable, and return @code{NULL} if it is not. Then, it should
2749 search through @var{arches} for an exact match to @var{info}, and
2750 return one if found. Lastly, if no exact match was found, it should
2751 create a new architecture based on @var{info} and return it.
2753 Only information in @var{info} should be used to choose the new
2754 architecture. Historically, @var{info} could be sparse, and
2755 defaults would be collected from the first element on @var{arches}.
2756 However, @value{GDBN} now fills in @var{info} more thoroughly,
2757 so new @code{gdbarch} initialization functions should not take
2758 defaults from @var{arches}.
2760 @section Registers and Memory
2762 @value{GDBN}'s model of the target machine is rather simple.
2763 @value{GDBN} assumes the machine includes a bank of registers and a
2764 block of memory. Each register may have a different size.
2766 @value{GDBN} does not have a magical way to match up with the
2767 compiler's idea of which registers are which; however, it is critical
2768 that they do match up accurately. The only way to make this work is
2769 to get accurate information about the order that the compiler uses,
2770 and to reflect that in the @code{REGISTER_NAME} and related macros.
2772 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
2774 @section Pointers Are Not Always Addresses
2775 @cindex pointer representation
2776 @cindex address representation
2777 @cindex word-addressed machines
2778 @cindex separate data and code address spaces
2779 @cindex spaces, separate data and code address
2780 @cindex address spaces, separate data and code
2781 @cindex code pointers, word-addressed
2782 @cindex converting between pointers and addresses
2783 @cindex D10V addresses
2785 On almost all 32-bit architectures, the representation of a pointer is
2786 indistinguishable from the representation of some fixed-length number
2787 whose value is the byte address of the object pointed to. On such
2788 machines, the words ``pointer'' and ``address'' can be used interchangeably.
2789 However, architectures with smaller word sizes are often cramped for
2790 address space, so they may choose a pointer representation that breaks this
2791 identity, and allows a larger code address space.
2793 For example, the Renesas D10V is a 16-bit VLIW processor whose
2794 instructions are 32 bits long@footnote{Some D10V instructions are
2795 actually pairs of 16-bit sub-instructions. However, since you can't
2796 jump into the middle of such a pair, code addresses can only refer to
2797 full 32 bit instructions, which is what matters in this explanation.}.
2798 If the D10V used ordinary byte addresses to refer to code locations,
2799 then the processor would only be able to address 64kb of instructions.
2800 However, since instructions must be aligned on four-byte boundaries, the
2801 low two bits of any valid instruction's byte address are always
2802 zero---byte addresses waste two bits. So instead of byte addresses,
2803 the D10V uses word addresses---byte addresses shifted right two bits---to
2804 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
2807 However, this means that code pointers and data pointers have different
2808 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
2809 @code{0xC020} when used as a data address, but refers to byte address
2810 @code{0x30080} when used as a code address.
2812 (The D10V also uses separate code and data address spaces, which also
2813 affects the correspondence between pointers and addresses, but we're
2814 going to ignore that here; this example is already too long.)
2816 To cope with architectures like this---the D10V is not the only
2817 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
2818 byte numbers, and @dfn{pointers}, which are the target's representation
2819 of an address of a particular type of data. In the example above,
2820 @code{0xC020} is the pointer, which refers to one of the addresses
2821 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
2822 @value{GDBN} provides functions for turning a pointer into an address
2823 and vice versa, in the appropriate way for the current architecture.
2825 Unfortunately, since addresses and pointers are identical on almost all
2826 processors, this distinction tends to bit-rot pretty quickly. Thus,
2827 each time you port @value{GDBN} to an architecture which does
2828 distinguish between pointers and addresses, you'll probably need to
2829 clean up some architecture-independent code.
2831 Here are functions which convert between pointers and addresses:
2833 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
2834 Treat the bytes at @var{buf} as a pointer or reference of type
2835 @var{type}, and return the address it represents, in a manner
2836 appropriate for the current architecture. This yields an address
2837 @value{GDBN} can use to read target memory, disassemble, etc. Note that
2838 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2841 For example, if the current architecture is the Intel x86, this function
2842 extracts a little-endian integer of the appropriate length from
2843 @var{buf} and returns it. However, if the current architecture is the
2844 D10V, this function will return a 16-bit integer extracted from
2845 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
2847 If @var{type} is not a pointer or reference type, then this function
2848 will signal an internal error.
2851 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
2852 Store the address @var{addr} in @var{buf}, in the proper format for a
2853 pointer of type @var{type} in the current architecture. Note that
2854 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2857 For example, if the current architecture is the Intel x86, this function
2858 stores @var{addr} unmodified as a little-endian integer of the
2859 appropriate length in @var{buf}. However, if the current architecture
2860 is the D10V, this function divides @var{addr} by four if @var{type} is
2861 a pointer to a function, and then stores it in @var{buf}.
2863 If @var{type} is not a pointer or reference type, then this function
2864 will signal an internal error.
2867 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
2868 Assuming that @var{val} is a pointer, return the address it represents,
2869 as appropriate for the current architecture.
2871 This function actually works on integral values, as well as pointers.
2872 For pointers, it performs architecture-specific conversions as
2873 described above for @code{extract_typed_address}.
2876 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
2877 Create and return a value representing a pointer of type @var{type} to
2878 the address @var{addr}, as appropriate for the current architecture.
2879 This function performs architecture-specific conversions as described
2880 above for @code{store_typed_address}.
2883 Here are some macros which architectures can define to indicate the
2884 relationship between pointers and addresses. These have default
2885 definitions, appropriate for architectures on which all pointers are
2886 simple unsigned byte addresses.
2888 @deftypefn {Target Macro} CORE_ADDR POINTER_TO_ADDRESS (struct type *@var{type}, char *@var{buf})
2889 Assume that @var{buf} holds a pointer of type @var{type}, in the
2890 appropriate format for the current architecture. Return the byte
2891 address the pointer refers to.
2893 This function may safely assume that @var{type} is either a pointer or a
2894 C@t{++} reference type.
2897 @deftypefn {Target Macro} void ADDRESS_TO_POINTER (struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
2898 Store in @var{buf} a pointer of type @var{type} representing the address
2899 @var{addr}, in the appropriate format for the current architecture.
2901 This function may safely assume that @var{type} is either a pointer or a
2902 C@t{++} reference type.
2905 @section Address Classes
2906 @cindex address classes
2907 @cindex DW_AT_byte_size
2908 @cindex DW_AT_address_class
2910 Sometimes information about different kinds of addresses is available
2911 via the debug information. For example, some programming environments
2912 define addresses of several different sizes. If the debug information
2913 distinguishes these kinds of address classes through either the size
2914 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
2915 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
2916 following macros should be defined in order to disambiguate these
2917 types within @value{GDBN} as well as provide the added information to
2918 a @value{GDBN} user when printing type expressions.
2920 @deftypefn {Target Macro} int ADDRESS_CLASS_TYPE_FLAGS (int @var{byte_size}, int @var{dwarf2_addr_class})
2921 Returns the type flags needed to construct a pointer type whose size
2922 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
2923 This function is normally called from within a symbol reader. See
2924 @file{dwarf2read.c}.
2927 @deftypefn {Target Macro} char *ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (int @var{type_flags})
2928 Given the type flags representing an address class qualifier, return
2931 @deftypefn {Target Macro} int ADDRESS_CLASS_NAME_to_TYPE_FLAGS (int @var{name}, int *var{type_flags_ptr})
2932 Given an address qualifier name, set the @code{int} refererenced by @var{type_flags_ptr} to the type flags
2933 for that address class qualifier.
2936 Since the need for address classes is rather rare, none of
2937 the address class macros defined by default. Predicate
2938 macros are provided to detect when they are defined.
2940 Consider a hypothetical architecture in which addresses are normally
2941 32-bits wide, but 16-bit addresses are also supported. Furthermore,
2942 suppose that the @w{DWARF 2} information for this architecture simply
2943 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
2944 of these "short" pointers. The following functions could be defined
2945 to implement the address class macros:
2948 somearch_address_class_type_flags (int byte_size,
2949 int dwarf2_addr_class)
2952 return TYPE_FLAG_ADDRESS_CLASS_1;
2958 somearch_address_class_type_flags_to_name (int type_flags)
2960 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
2967 somearch_address_class_name_to_type_flags (char *name,
2968 int *type_flags_ptr)
2970 if (strcmp (name, "short") == 0)
2972 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
2980 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
2981 to indicate the presence of one of these "short" pointers. E.g, if
2982 the debug information indicates that @code{short_ptr_var} is one of these
2983 short pointers, @value{GDBN} might show the following behavior:
2986 (gdb) ptype short_ptr_var
2987 type = int * @@short
2991 @section Raw and Virtual Register Representations
2992 @cindex raw register representation
2993 @cindex virtual register representation
2994 @cindex representations, raw and virtual registers
2996 @emph{Maintainer note: This section is pretty much obsolete. The
2997 functionality described here has largely been replaced by
2998 pseudo-registers and the mechanisms described in @ref{Target
2999 Architecture Definition, , Using Different Register and Memory Data
3000 Representations}. See also @uref{http://www.gnu.org/software/gdb/bugs/,
3001 Bug Tracking Database} and
3002 @uref{http://sources.redhat.com/gdb/current/ari/, ARI Index} for more
3003 up-to-date information.}
3005 Some architectures use one representation for a value when it lives in a
3006 register, but use a different representation when it lives in memory.
3007 In @value{GDBN}'s terminology, the @dfn{raw} representation is the one used in
3008 the target registers, and the @dfn{virtual} representation is the one
3009 used in memory, and within @value{GDBN} @code{struct value} objects.
3011 @emph{Maintainer note: Notice that the same mechanism is being used to
3012 both convert a register to a @code{struct value} and alternative
3015 For almost all data types on almost all architectures, the virtual and
3016 raw representations are identical, and no special handling is needed.
3017 However, they do occasionally differ. For example:
3021 The x86 architecture supports an 80-bit @code{long double} type. However, when
3022 we store those values in memory, they occupy twelve bytes: the
3023 floating-point number occupies the first ten, and the final two bytes
3024 are unused. This keeps the values aligned on four-byte boundaries,
3025 allowing more efficient access. Thus, the x86 80-bit floating-point
3026 type is the raw representation, and the twelve-byte loosely-packed
3027 arrangement is the virtual representation.
3030 Some 64-bit MIPS targets present 32-bit registers to @value{GDBN} as 64-bit
3031 registers, with garbage in their upper bits. @value{GDBN} ignores the top 32
3032 bits. Thus, the 64-bit form, with garbage in the upper 32 bits, is the
3033 raw representation, and the trimmed 32-bit representation is the
3034 virtual representation.
3037 In general, the raw representation is determined by the architecture, or
3038 @value{GDBN}'s interface to the architecture, while the virtual representation
3039 can be chosen for @value{GDBN}'s convenience. @value{GDBN}'s register file,
3040 @code{registers}, holds the register contents in raw format, and the
3041 @value{GDBN} remote protocol transmits register values in raw format.
3043 Your architecture may define the following macros to request
3044 conversions between the raw and virtual format:
3046 @deftypefn {Target Macro} int REGISTER_CONVERTIBLE (int @var{reg})
3047 Return non-zero if register number @var{reg}'s value needs different raw
3048 and virtual formats.
3050 You should not use @code{REGISTER_CONVERT_TO_VIRTUAL} for a register
3051 unless this macro returns a non-zero value for that register.
3054 @deftypefn {Target Macro} int DEPRECATED_REGISTER_RAW_SIZE (int @var{reg})
3055 The size of register number @var{reg}'s raw value. This is the number
3056 of bytes the register will occupy in @code{registers}, or in a @value{GDBN}
3057 remote protocol packet.
3060 @deftypefn {Target Macro} int DEPRECATED_REGISTER_VIRTUAL_SIZE (int @var{reg})
3061 The size of register number @var{reg}'s value, in its virtual format.
3062 This is the size a @code{struct value}'s buffer will have, holding that
3066 @deftypefn {Target Macro} struct type *DEPRECATED_REGISTER_VIRTUAL_TYPE (int @var{reg})
3067 This is the type of the virtual representation of register number
3068 @var{reg}. Note that there is no need for a macro giving a type for the
3069 register's raw form; once the register's value has been obtained, @value{GDBN}
3070 always uses the virtual form.
3073 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_VIRTUAL (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
3074 Convert the value of register number @var{reg} to @var{type}, which
3075 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
3076 at @var{from} holds the register's value in raw format; the macro should
3077 convert the value to virtual format, and place it at @var{to}.
3079 Note that @code{REGISTER_CONVERT_TO_VIRTUAL} and
3080 @code{REGISTER_CONVERT_TO_RAW} take their @var{reg} and @var{type}
3081 arguments in different orders.
3083 You should only use @code{REGISTER_CONVERT_TO_VIRTUAL} with registers
3084 for which the @code{REGISTER_CONVERTIBLE} macro returns a non-zero
3088 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_RAW (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3089 Convert the value of register number @var{reg} to @var{type}, which
3090 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
3091 at @var{from} holds the register's value in raw format; the macro should
3092 convert the value to virtual format, and place it at @var{to}.
3094 Note that REGISTER_CONVERT_TO_VIRTUAL and REGISTER_CONVERT_TO_RAW take
3095 their @var{reg} and @var{type} arguments in different orders.
3099 @section Using Different Register and Memory Data Representations
3100 @cindex register representation
3101 @cindex memory representation
3102 @cindex representations, register and memory
3103 @cindex register data formats, converting
3104 @cindex @code{struct value}, converting register contents to
3106 @emph{Maintainer's note: The way GDB manipulates registers is undergoing
3107 significant change. Many of the macros and functions refered to in this
3108 section are likely to be subject to further revision. See
3109 @uref{http://sources.redhat.com/gdb/current/ari/, A.R. Index} and
3110 @uref{http://www.gnu.org/software/gdb/bugs, Bug Tracking Database} for
3111 further information. cagney/2002-05-06.}
3113 Some architectures can represent a data object in a register using a
3114 form that is different to the objects more normal memory representation.
3120 The Alpha architecture can represent 32 bit integer values in
3121 floating-point registers.
3124 The x86 architecture supports 80-bit floating-point registers. The
3125 @code{long double} data type occupies 96 bits in memory but only 80 bits
3126 when stored in a register.
3130 In general, the register representation of a data type is determined by
3131 the architecture, or @value{GDBN}'s interface to the architecture, while
3132 the memory representation is determined by the Application Binary
3135 For almost all data types on almost all architectures, the two
3136 representations are identical, and no special handling is needed.
3137 However, they do occasionally differ. Your architecture may define the
3138 following macros to request conversions between the register and memory
3139 representations of a data type:
3141 @deftypefn {Target Macro} int CONVERT_REGISTER_P (int @var{reg})
3142 Return non-zero if the representation of a data value stored in this
3143 register may be different to the representation of that same data value
3144 when stored in memory.
3146 When non-zero, the macros @code{REGISTER_TO_VALUE} and
3147 @code{VALUE_TO_REGISTER} are used to perform any necessary conversion.
3150 @deftypefn {Target Macro} void REGISTER_TO_VALUE (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
3151 Convert the value of register number @var{reg} to a data object of type
3152 @var{type}. The buffer at @var{from} holds the register's value in raw
3153 format; the converted value should be placed in the buffer at @var{to}.
3155 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
3156 their @var{reg} and @var{type} arguments in different orders.
3158 You should only use @code{REGISTER_TO_VALUE} with registers for which
3159 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
3162 @deftypefn {Target Macro} void VALUE_TO_REGISTER (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3163 Convert a data value of type @var{type} to register number @var{reg}'
3166 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
3167 their @var{reg} and @var{type} arguments in different orders.
3169 You should only use @code{VALUE_TO_REGISTER} with registers for which
3170 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
3173 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_TYPE (int @var{regnum}, struct type *@var{type}, char *@var{buf})
3174 See @file{mips-tdep.c}. It does not do what you want.
3178 @section Frame Interpretation
3180 @section Inferior Call Setup
3182 @section Compiler Characteristics
3184 @section Target Conditionals
3186 This section describes the macros that you can use to define the target
3191 @item ADDR_BITS_REMOVE (addr)
3192 @findex ADDR_BITS_REMOVE
3193 If a raw machine instruction address includes any bits that are not
3194 really part of the address, then define this macro to expand into an
3195 expression that zeroes those bits in @var{addr}. This is only used for
3196 addresses of instructions, and even then not in all contexts.
3198 For example, the two low-order bits of the PC on the Hewlett-Packard PA
3199 2.0 architecture contain the privilege level of the corresponding
3200 instruction. Since instructions must always be aligned on four-byte
3201 boundaries, the processor masks out these bits to generate the actual
3202 address of the instruction. ADDR_BITS_REMOVE should filter out these
3203 bits with an expression such as @code{((addr) & ~3)}.
3205 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS (@var{name}, @var{type_flags_ptr})
3206 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS
3207 If @var{name} is a valid address class qualifier name, set the @code{int}
3208 referenced by @var{type_flags_ptr} to the mask representing the qualifier
3209 and return 1. If @var{name} is not a valid address class qualifier name,
3212 The value for @var{type_flags_ptr} should be one of
3213 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
3214 possibly some combination of these values or'd together.
3215 @xref{Target Architecture Definition, , Address Classes}.
3217 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P ()
3218 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P
3219 Predicate which indicates whether @code{ADDRESS_CLASS_NAME_TO_TYPE_FLAGS}
3222 @item ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
3223 @findex ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
3224 Given a pointers byte size (as described by the debug information) and
3225 the possible @code{DW_AT_address_class} value, return the type flags
3226 used by @value{GDBN} to represent this address class. The value
3227 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
3228 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
3229 values or'd together.
3230 @xref{Target Architecture Definition, , Address Classes}.
3232 @item ADDRESS_CLASS_TYPE_FLAGS_P ()
3233 @findex ADDRESS_CLASS_TYPE_FLAGS_P
3234 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS} has
3237 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (@var{type_flags})
3238 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME
3239 Return the name of the address class qualifier associated with the type
3240 flags given by @var{type_flags}.
3242 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P ()
3243 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P
3244 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS_TO_NAME} has
3246 @xref{Target Architecture Definition, , Address Classes}.
3248 @item ADDRESS_TO_POINTER (@var{type}, @var{buf}, @var{addr})
3249 @findex ADDRESS_TO_POINTER
3250 Store in @var{buf} a pointer of type @var{type} representing the address
3251 @var{addr}, in the appropriate format for the current architecture.
3252 This macro may safely assume that @var{type} is either a pointer or a
3253 C@t{++} reference type.
3254 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3256 @item BELIEVE_PCC_PROMOTION
3257 @findex BELIEVE_PCC_PROMOTION
3258 Define if the compiler promotes a @code{short} or @code{char}
3259 parameter to an @code{int}, but still reports the parameter as its
3260 original type, rather than the promoted type.
3262 @item BITS_BIG_ENDIAN
3263 @findex BITS_BIG_ENDIAN
3264 Define this if the numbering of bits in the targets does @strong{not} match the
3265 endianness of the target byte order. A value of 1 means that the bits
3266 are numbered in a big-endian bit order, 0 means little-endian.
3270 This is the character array initializer for the bit pattern to put into
3271 memory where a breakpoint is set. Although it's common to use a trap
3272 instruction for a breakpoint, it's not required; for instance, the bit
3273 pattern could be an invalid instruction. The breakpoint must be no
3274 longer than the shortest instruction of the architecture.
3276 @code{BREAKPOINT} has been deprecated in favor of
3277 @code{BREAKPOINT_FROM_PC}.
3279 @item BIG_BREAKPOINT
3280 @itemx LITTLE_BREAKPOINT
3281 @findex LITTLE_BREAKPOINT
3282 @findex BIG_BREAKPOINT
3283 Similar to BREAKPOINT, but used for bi-endian targets.
3285 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
3286 favor of @code{BREAKPOINT_FROM_PC}.
3288 @item DEPRECATED_REMOTE_BREAKPOINT
3289 @itemx DEPRECATED_LITTLE_REMOTE_BREAKPOINT
3290 @itemx DEPRECATED_BIG_REMOTE_BREAKPOINT
3291 @findex DEPRECATED_BIG_REMOTE_BREAKPOINT
3292 @findex DEPRECATED_LITTLE_REMOTE_BREAKPOINT
3293 @findex DEPRECATED_REMOTE_BREAKPOINT
3294 Specify the breakpoint instruction sequence for a remote target.
3295 @code{DEPRECATED_REMOTE_BREAKPOINT},
3296 @code{DEPRECATED_BIG_REMOTE_BREAKPOINT} and
3297 @code{DEPRECATED_LITTLE_REMOTE_BREAKPOINT} have been deprecated in
3298 favor of @code{BREAKPOINT_FROM_PC} (@pxref{BREAKPOINT_FROM_PC}).
3300 @item BREAKPOINT_FROM_PC (@var{pcptr}, @var{lenptr})
3301 @findex BREAKPOINT_FROM_PC
3302 @anchor{BREAKPOINT_FROM_PC} Use the program counter to determine the
3303 contents and size of a breakpoint instruction. It returns a pointer to
3304 a string of bytes that encode a breakpoint instruction, stores the
3305 length of the string to @code{*@var{lenptr}}, and adjusts the program
3306 counter (if necessary) to point to the actual memory location where the
3307 breakpoint should be inserted.
3309 Although it is common to use a trap instruction for a breakpoint, it's
3310 not required; for instance, the bit pattern could be an invalid
3311 instruction. The breakpoint must be no longer than the shortest
3312 instruction of the architecture.
3314 Replaces all the other @var{BREAKPOINT} macros.
3316 @item MEMORY_INSERT_BREAKPOINT (@var{bp_tgt})
3317 @itemx MEMORY_REMOVE_BREAKPOINT (@var{bp_tgt})
3318 @findex MEMORY_REMOVE_BREAKPOINT
3319 @findex MEMORY_INSERT_BREAKPOINT
3320 Insert or remove memory based breakpoints. Reasonable defaults
3321 (@code{default_memory_insert_breakpoint} and
3322 @code{default_memory_remove_breakpoint} respectively) have been
3323 provided so that it is not necessary to define these for most
3324 architectures. Architectures which may want to define
3325 @code{MEMORY_INSERT_BREAKPOINT} and @code{MEMORY_REMOVE_BREAKPOINT} will
3326 likely have instructions that are oddly sized or are not stored in a
3327 conventional manner.
3329 It may also be desirable (from an efficiency standpoint) to define
3330 custom breakpoint insertion and removal routines if
3331 @code{BREAKPOINT_FROM_PC} needs to read the target's memory for some
3334 @item ADJUST_BREAKPOINT_ADDRESS (@var{address})
3335 @findex ADJUST_BREAKPOINT_ADDRESS
3336 @cindex breakpoint address adjusted
3337 Given an address at which a breakpoint is desired, return a breakpoint
3338 address adjusted to account for architectural constraints on
3339 breakpoint placement. This method is not needed by most targets.
3341 The FR-V target (see @file{frv-tdep.c}) requires this method.
3342 The FR-V is a VLIW architecture in which a number of RISC-like
3343 instructions are grouped (packed) together into an aggregate
3344 instruction or instruction bundle. When the processor executes
3345 one of these bundles, the component instructions are executed
3348 In the course of optimization, the compiler may group instructions
3349 from distinct source statements into the same bundle. The line number
3350 information associated with one of the latter statements will likely
3351 refer to some instruction other than the first one in the bundle. So,
3352 if the user attempts to place a breakpoint on one of these latter
3353 statements, @value{GDBN} must be careful to @emph{not} place the break
3354 instruction on any instruction other than the first one in the bundle.
3355 (Remember though that the instructions within a bundle execute
3356 in parallel, so the @emph{first} instruction is the instruction
3357 at the lowest address and has nothing to do with execution order.)
3359 The FR-V's @code{ADJUST_BREAKPOINT_ADDRESS} method will adjust a
3360 breakpoint's address by scanning backwards for the beginning of
3361 the bundle, returning the address of the bundle.
3363 Since the adjustment of a breakpoint may significantly alter a user's
3364 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
3365 is initially set and each time that that breakpoint is hit.
3367 @item CALL_DUMMY_LOCATION
3368 @findex CALL_DUMMY_LOCATION
3369 See the file @file{inferior.h}.
3371 This method has been replaced by @code{push_dummy_code}
3372 (@pxref{push_dummy_code}).
3374 @item CANNOT_FETCH_REGISTER (@var{regno})
3375 @findex CANNOT_FETCH_REGISTER
3376 A C expression that should be nonzero if @var{regno} cannot be fetched
3377 from an inferior process. This is only relevant if
3378 @code{FETCH_INFERIOR_REGISTERS} is not defined.
3380 @item CANNOT_STORE_REGISTER (@var{regno})
3381 @findex CANNOT_STORE_REGISTER
3382 A C expression that should be nonzero if @var{regno} should not be
3383 written to the target. This is often the case for program counters,
3384 status words, and other special registers. If this is not defined,
3385 @value{GDBN} will assume that all registers may be written.
3387 @item int CONVERT_REGISTER_P(@var{regnum})
3388 @findex CONVERT_REGISTER_P
3389 Return non-zero if register @var{regnum} can represent data values in a
3391 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3393 @item DECR_PC_AFTER_BREAK
3394 @findex DECR_PC_AFTER_BREAK
3395 Define this to be the amount by which to decrement the PC after the
3396 program encounters a breakpoint. This is often the number of bytes in
3397 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
3399 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
3400 @findex DISABLE_UNSETTABLE_BREAK
3401 If defined, this should evaluate to 1 if @var{addr} is in a shared
3402 library in which breakpoints cannot be set and so should be disabled.
3404 @item PRINT_FLOAT_INFO()
3405 @findex PRINT_FLOAT_INFO
3406 If defined, then the @samp{info float} command will print information about
3407 the processor's floating point unit.
3409 @item print_registers_info (@var{gdbarch}, @var{frame}, @var{regnum}, @var{all})
3410 @findex print_registers_info
3411 If defined, pretty print the value of the register @var{regnum} for the
3412 specified @var{frame}. If the value of @var{regnum} is -1, pretty print
3413 either all registers (@var{all} is non zero) or a select subset of
3414 registers (@var{all} is zero).
3416 The default method prints one register per line, and if @var{all} is
3417 zero omits floating-point registers.
3419 @item PRINT_VECTOR_INFO()
3420 @findex PRINT_VECTOR_INFO
3421 If defined, then the @samp{info vector} command will call this function
3422 to print information about the processor's vector unit.
3424 By default, the @samp{info vector} command will print all vector
3425 registers (the register's type having the vector attribute).
3427 @item DWARF_REG_TO_REGNUM
3428 @findex DWARF_REG_TO_REGNUM
3429 Convert DWARF register number into @value{GDBN} regnum. If not defined,
3430 no conversion will be performed.
3432 @item DWARF2_REG_TO_REGNUM
3433 @findex DWARF2_REG_TO_REGNUM
3434 Convert DWARF2 register number into @value{GDBN} regnum. If not
3435 defined, no conversion will be performed.
3437 @item ECOFF_REG_TO_REGNUM
3438 @findex ECOFF_REG_TO_REGNUM
3439 Convert ECOFF register number into @value{GDBN} regnum. If not defined,
3440 no conversion will be performed.
3442 @item END_OF_TEXT_DEFAULT
3443 @findex END_OF_TEXT_DEFAULT
3444 This is an expression that should designate the end of the text section.
3447 @item EXTRACT_RETURN_VALUE(@var{type}, @var{regbuf}, @var{valbuf})
3448 @findex EXTRACT_RETURN_VALUE
3449 Define this to extract a function's return value of type @var{type} from
3450 the raw register state @var{regbuf} and copy that, in virtual format,
3453 This method has been deprecated in favour of @code{gdbarch_return_value}
3454 (@pxref{gdbarch_return_value}).
3456 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS(@var{regbuf})
3457 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS
3458 @anchor{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}
3459 When defined, extract from the array @var{regbuf} (containing the raw
3460 register state) the @code{CORE_ADDR} at which a function should return
3461 its structure value.
3463 @xref{gdbarch_return_value}.
3465 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P()
3466 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P
3467 Predicate for @code{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}.
3469 @item DEPRECATED_FP_REGNUM
3470 @findex DEPRECATED_FP_REGNUM
3471 If the virtual frame pointer is kept in a register, then define this
3472 macro to be the number (greater than or equal to zero) of that register.
3474 This should only need to be defined if @code{DEPRECATED_TARGET_READ_FP}
3477 @item DEPRECATED_FRAMELESS_FUNCTION_INVOCATION(@var{fi})
3478 @findex DEPRECATED_FRAMELESS_FUNCTION_INVOCATION
3479 Define this to an expression that returns 1 if the function invocation
3480 represented by @var{fi} does not have a stack frame associated with it.
3483 @item frame_align (@var{address})
3484 @anchor{frame_align}
3486 Define this to adjust @var{address} so that it meets the alignment
3487 requirements for the start of a new stack frame. A stack frame's
3488 alignment requirements are typically stronger than a target processors
3489 stack alignment requirements (@pxref{DEPRECATED_STACK_ALIGN}).
3491 This function is used to ensure that, when creating a dummy frame, both
3492 the initial stack pointer and (if needed) the address of the return
3493 value are correctly aligned.
3495 Unlike @code{DEPRECATED_STACK_ALIGN}, this function always adjusts the
3496 address in the direction of stack growth.
3498 By default, no frame based stack alignment is performed.
3500 @item int frame_red_zone_size
3502 The number of bytes, beyond the innermost-stack-address, reserved by the
3503 @sc{abi}. A function is permitted to use this scratch area (instead of
3504 allocating extra stack space).
3506 When performing an inferior function call, to ensure that it does not
3507 modify this area, @value{GDBN} adjusts the innermost-stack-address by
3508 @var{frame_red_zone_size} bytes before pushing parameters onto the
3511 By default, zero bytes are allocated. The value must be aligned
3512 (@pxref{frame_align}).
3514 The @sc{amd64} (nee x86-64) @sc{abi} documentation refers to the
3515 @emph{red zone} when describing this scratch area.
3518 @item DEPRECATED_FRAME_CHAIN(@var{frame})
3519 @findex DEPRECATED_FRAME_CHAIN
3520 Given @var{frame}, return a pointer to the calling frame.
3522 @item DEPRECATED_FRAME_CHAIN_VALID(@var{chain}, @var{thisframe})
3523 @findex DEPRECATED_FRAME_CHAIN_VALID
3524 Define this to be an expression that returns zero if the given frame is an
3525 outermost frame, with no caller, and nonzero otherwise. Most normal
3526 situations can be handled without defining this macro, including @code{NULL}
3527 chain pointers, dummy frames, and frames whose PC values are inside the
3528 startup file (e.g.@: @file{crt0.o}), inside @code{main}, or inside
3531 @item DEPRECATED_FRAME_INIT_SAVED_REGS(@var{frame})
3532 @findex DEPRECATED_FRAME_INIT_SAVED_REGS
3533 See @file{frame.h}. Determines the address of all registers in the
3534 current stack frame storing each in @code{frame->saved_regs}. Space for
3535 @code{frame->saved_regs} shall be allocated by
3536 @code{DEPRECATED_FRAME_INIT_SAVED_REGS} using
3537 @code{frame_saved_regs_zalloc}.
3539 @code{FRAME_FIND_SAVED_REGS} is deprecated.
3541 @item FRAME_NUM_ARGS (@var{fi})
3542 @findex FRAME_NUM_ARGS
3543 For the frame described by @var{fi} return the number of arguments that
3544 are being passed. If the number of arguments is not known, return
3547 @item DEPRECATED_FRAME_SAVED_PC(@var{frame})
3548 @findex DEPRECATED_FRAME_SAVED_PC
3549 @anchor{DEPRECATED_FRAME_SAVED_PC} Given @var{frame}, return the pc
3550 saved there. This is the return address.
3552 This method is deprecated. @xref{unwind_pc}.
3554 @item CORE_ADDR unwind_pc (struct frame_info *@var{this_frame})
3556 @anchor{unwind_pc} Return the instruction address, in @var{this_frame}'s
3557 caller, at which execution will resume after @var{this_frame} returns.
3558 This is commonly refered to as the return address.
3560 The implementation, which must be frame agnostic (work with any frame),
3561 is typically no more than:
3565 frame_unwind_unsigned_register (this_frame, D10V_PC_REGNUM, &pc);
3566 return d10v_make_iaddr (pc);
3570 @xref{DEPRECATED_FRAME_SAVED_PC}, which this method replaces.
3572 @item CORE_ADDR unwind_sp (struct frame_info *@var{this_frame})
3574 @anchor{unwind_sp} Return the frame's inner most stack address. This is
3575 commonly refered to as the frame's @dfn{stack pointer}.
3577 The implementation, which must be frame agnostic (work with any frame),
3578 is typically no more than:
3582 frame_unwind_unsigned_register (this_frame, D10V_SP_REGNUM, &sp);
3583 return d10v_make_daddr (sp);
3587 @xref{TARGET_READ_SP}, which this method replaces.
3589 @item FUNCTION_EPILOGUE_SIZE
3590 @findex FUNCTION_EPILOGUE_SIZE
3591 For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the
3592 function end symbol is 0. For such targets, you must define
3593 @code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size of a
3594 function's epilogue.
3596 @item DEPRECATED_FUNCTION_START_OFFSET
3597 @findex DEPRECATED_FUNCTION_START_OFFSET
3598 An integer, giving the offset in bytes from a function's address (as
3599 used in the values of symbols, function pointers, etc.), and the
3600 function's first genuine instruction.
3602 This is zero on almost all machines: the function's address is usually
3603 the address of its first instruction. However, on the VAX, for
3604 example, each function starts with two bytes containing a bitmask
3605 indicating which registers to save upon entry to the function. The
3606 VAX @code{call} instructions check this value, and save the
3607 appropriate registers automatically. Thus, since the offset from the
3608 function's address to its first instruction is two bytes,
3609 @code{DEPRECATED_FUNCTION_START_OFFSET} would be 2 on the VAX.
3611 @item GCC_COMPILED_FLAG_SYMBOL
3612 @itemx GCC2_COMPILED_FLAG_SYMBOL
3613 @findex GCC2_COMPILED_FLAG_SYMBOL
3614 @findex GCC_COMPILED_FLAG_SYMBOL
3615 If defined, these are the names of the symbols that @value{GDBN} will
3616 look for to detect that GCC compiled the file. The default symbols
3617 are @code{gcc_compiled.} and @code{gcc2_compiled.},
3618 respectively. (Currently only defined for the Delta 68.)
3620 @item @value{GDBN}_MULTI_ARCH
3621 @findex @value{GDBN}_MULTI_ARCH
3622 If defined and non-zero, enables support for multiple architectures
3623 within @value{GDBN}.
3625 This support can be enabled at two levels. At level one, only
3626 definitions for previously undefined macros are provided; at level two,
3627 a multi-arch definition of all architecture dependent macros will be
3630 @item @value{GDBN}_TARGET_IS_HPPA
3631 @findex @value{GDBN}_TARGET_IS_HPPA
3632 This determines whether horrible kludge code in @file{dbxread.c} and
3633 @file{partial-stab.h} is used to mangle multiple-symbol-table files from
3634 HPPA's. This should all be ripped out, and a scheme like @file{elfread.c}
3637 @item GET_LONGJMP_TARGET
3638 @findex GET_LONGJMP_TARGET
3639 For most machines, this is a target-dependent parameter. On the
3640 DECstation and the Iris, this is a native-dependent parameter, since
3641 the header file @file{setjmp.h} is needed to define it.
3643 This macro determines the target PC address that @code{longjmp} will jump to,
3644 assuming that we have just stopped at a @code{longjmp} breakpoint. It takes a
3645 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3646 pointer. It examines the current state of the machine as needed.
3648 @item DEPRECATED_GET_SAVED_REGISTER
3649 @findex DEPRECATED_GET_SAVED_REGISTER
3650 Define this if you need to supply your own definition for the function
3651 @code{DEPRECATED_GET_SAVED_REGISTER}.
3653 @item DEPRECATED_IBM6000_TARGET
3654 @findex DEPRECATED_IBM6000_TARGET
3655 Shows that we are configured for an IBM RS/6000 system. This
3656 conditional should be eliminated (FIXME) and replaced by
3657 feature-specific macros. It was introduced in a haste and we are
3658 repenting at leisure.
3660 @item I386_USE_GENERIC_WATCHPOINTS
3661 An x86-based target can define this to use the generic x86 watchpoint
3662 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3664 @item SYMBOLS_CAN_START_WITH_DOLLAR
3665 @findex SYMBOLS_CAN_START_WITH_DOLLAR
3666 Some systems have routines whose names start with @samp{$}. Giving this
3667 macro a non-zero value tells @value{GDBN}'s expression parser to check for such
3668 routines when parsing tokens that begin with @samp{$}.
3670 On HP-UX, certain system routines (millicode) have names beginning with
3671 @samp{$} or @samp{$$}. For example, @code{$$dyncall} is a millicode
3672 routine that handles inter-space procedure calls on PA-RISC.
3674 @item DEPRECATED_INIT_EXTRA_FRAME_INFO (@var{fromleaf}, @var{frame})
3675 @findex DEPRECATED_INIT_EXTRA_FRAME_INFO
3676 If additional information about the frame is required this should be
3677 stored in @code{frame->extra_info}. Space for @code{frame->extra_info}
3678 is allocated using @code{frame_extra_info_zalloc}.
3680 @item DEPRECATED_INIT_FRAME_PC (@var{fromleaf}, @var{prev})
3681 @findex DEPRECATED_INIT_FRAME_PC
3682 This is a C statement that sets the pc of the frame pointed to by
3683 @var{prev}. [By default...]
3685 @item INNER_THAN (@var{lhs}, @var{rhs})
3687 Returns non-zero if stack address @var{lhs} is inner than (nearer to the
3688 stack top) stack address @var{rhs}. Define this as @code{lhs < rhs} if
3689 the target's stack grows downward in memory, or @code{lhs > rsh} if the
3692 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{pc})
3693 @findex gdbarch_in_function_epilogue_p
3694 Returns non-zero if the given @var{pc} is in the epilogue of a function.
3695 The epilogue of a function is defined as the part of a function where
3696 the stack frame of the function already has been destroyed up to the
3697 final `return from function call' instruction.
3699 @item DEPRECATED_SIGTRAMP_START (@var{pc})
3700 @findex DEPRECATED_SIGTRAMP_START
3701 @itemx DEPRECATED_SIGTRAMP_END (@var{pc})
3702 @findex DEPRECATED_SIGTRAMP_END
3703 Define these to be the start and end address of the @code{sigtramp} for the
3704 given @var{pc}. On machines where the address is just a compile time
3705 constant, the macro expansion will typically just ignore the supplied
3708 @item IN_SOLIB_CALL_TRAMPOLINE (@var{pc}, @var{name})
3709 @findex IN_SOLIB_CALL_TRAMPOLINE
3710 Define this to evaluate to nonzero if the program is stopped in the
3711 trampoline that connects to a shared library.
3713 @item IN_SOLIB_RETURN_TRAMPOLINE (@var{pc}, @var{name})
3714 @findex IN_SOLIB_RETURN_TRAMPOLINE
3715 Define this to evaluate to nonzero if the program is stopped in the
3716 trampoline that returns from a shared library.
3718 @item IN_SOLIB_DYNSYM_RESOLVE_CODE (@var{pc})
3719 @findex IN_SOLIB_DYNSYM_RESOLVE_CODE
3720 Define this to evaluate to nonzero if the program is stopped in the
3723 @item SKIP_SOLIB_RESOLVER (@var{pc})
3724 @findex SKIP_SOLIB_RESOLVER
3725 Define this to evaluate to the (nonzero) address at which execution
3726 should continue to get past the dynamic linker's symbol resolution
3727 function. A zero value indicates that it is not important or necessary
3728 to set a breakpoint to get through the dynamic linker and that single
3729 stepping will suffice.
3731 @item INTEGER_TO_ADDRESS (@var{type}, @var{buf})
3732 @findex INTEGER_TO_ADDRESS
3733 @cindex converting integers to addresses
3734 Define this when the architecture needs to handle non-pointer to address
3735 conversions specially. Converts that value to an address according to
3736 the current architectures conventions.
3738 @emph{Pragmatics: When the user copies a well defined expression from
3739 their source code and passes it, as a parameter, to @value{GDBN}'s
3740 @code{print} command, they should get the same value as would have been
3741 computed by the target program. Any deviation from this rule can cause
3742 major confusion and annoyance, and needs to be justified carefully. In
3743 other words, @value{GDBN} doesn't really have the freedom to do these
3744 conversions in clever and useful ways. It has, however, been pointed
3745 out that users aren't complaining about how @value{GDBN} casts integers
3746 to pointers; they are complaining that they can't take an address from a
3747 disassembly listing and give it to @code{x/i}. Adding an architecture
3748 method like @code{INTEGER_TO_ADDRESS} certainly makes it possible for
3749 @value{GDBN} to ``get it right'' in all circumstances.}
3751 @xref{Target Architecture Definition, , Pointers Are Not Always
3754 @item NO_HIF_SUPPORT
3755 @findex NO_HIF_SUPPORT
3756 (Specific to the a29k.)
3758 @item POINTER_TO_ADDRESS (@var{type}, @var{buf})
3759 @findex POINTER_TO_ADDRESS
3760 Assume that @var{buf} holds a pointer of type @var{type}, in the
3761 appropriate format for the current architecture. Return the byte
3762 address the pointer refers to.
3763 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3765 @item REGISTER_CONVERTIBLE (@var{reg})
3766 @findex REGISTER_CONVERTIBLE
3767 Return non-zero if @var{reg} uses different raw and virtual formats.
3768 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3770 @item REGISTER_TO_VALUE(@var{regnum}, @var{type}, @var{from}, @var{to})
3771 @findex REGISTER_TO_VALUE
3772 Convert the raw contents of register @var{regnum} into a value of type
3774 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3776 @item DEPRECATED_REGISTER_RAW_SIZE (@var{reg})
3777 @findex DEPRECATED_REGISTER_RAW_SIZE
3778 Return the raw size of @var{reg}; defaults to the size of the register's
3780 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3782 @item register_reggroup_p (@var{gdbarch}, @var{regnum}, @var{reggroup})
3783 @findex register_reggroup_p
3784 @cindex register groups
3785 Return non-zero if register @var{regnum} is a member of the register
3786 group @var{reggroup}.
3788 By default, registers are grouped as follows:
3791 @item float_reggroup
3792 Any register with a valid name and a floating-point type.
3793 @item vector_reggroup
3794 Any register with a valid name and a vector type.
3795 @item general_reggroup
3796 Any register with a valid name and a type other than vector or
3797 floating-point. @samp{float_reggroup}.
3799 @itemx restore_reggroup
3801 Any register with a valid name.
3804 @item DEPRECATED_REGISTER_VIRTUAL_SIZE (@var{reg})
3805 @findex DEPRECATED_REGISTER_VIRTUAL_SIZE
3806 Return the virtual size of @var{reg}; defaults to the size of the
3807 register's virtual type.
3808 Return the virtual size of @var{reg}.
3809 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3811 @item DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})
3812 @findex REGISTER_VIRTUAL_TYPE
3813 Return the virtual type of @var{reg}.
3814 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3816 @item struct type *register_type (@var{gdbarch}, @var{reg})
3817 @findex register_type
3818 If defined, return the type of register @var{reg}. This function
3819 superseeds @code{DEPRECATED_REGISTER_VIRTUAL_TYPE}. @xref{Target Architecture
3820 Definition, , Raw and Virtual Register Representations}.
3822 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
3823 @findex REGISTER_CONVERT_TO_VIRTUAL
3824 Convert the value of register @var{reg} from its raw form to its virtual
3826 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3828 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
3829 @findex REGISTER_CONVERT_TO_RAW
3830 Convert the value of register @var{reg} from its virtual form to its raw
3832 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3834 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
3835 @findex regset_from_core_section
3836 Return the appropriate register set for a core file section with name
3837 @var{sect_name} and size @var{sect_size}.
3839 @item SOFTWARE_SINGLE_STEP_P()
3840 @findex SOFTWARE_SINGLE_STEP_P
3841 Define this as 1 if the target does not have a hardware single-step
3842 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
3844 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breapoints_p})
3845 @findex SOFTWARE_SINGLE_STEP
3846 A function that inserts or removes (depending on
3847 @var{insert_breapoints_p}) breakpoints at each possible destinations of
3848 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
3851 @item SOFUN_ADDRESS_MAYBE_MISSING
3852 @findex SOFUN_ADDRESS_MAYBE_MISSING
3853 Somebody clever observed that, the more actual addresses you have in the
3854 debug information, the more time the linker has to spend relocating
3855 them. So whenever there's some other way the debugger could find the
3856 address it needs, you should omit it from the debug info, to make
3859 @code{SOFUN_ADDRESS_MAYBE_MISSING} indicates that a particular set of
3860 hacks of this sort are in use, affecting @code{N_SO} and @code{N_FUN}
3861 entries in stabs-format debugging information. @code{N_SO} stabs mark
3862 the beginning and ending addresses of compilation units in the text
3863 segment. @code{N_FUN} stabs mark the starts and ends of functions.
3865 @code{SOFUN_ADDRESS_MAYBE_MISSING} means two things:
3869 @code{N_FUN} stabs have an address of zero. Instead, you should find the
3870 addresses where the function starts by taking the function name from
3871 the stab, and then looking that up in the minsyms (the
3872 linker/assembler symbol table). In other words, the stab has the
3873 name, and the linker/assembler symbol table is the only place that carries
3877 @code{N_SO} stabs have an address of zero, too. You just look at the
3878 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab,
3879 and guess the starting and ending addresses of the compilation unit from
3883 @item PC_LOAD_SEGMENT
3884 @findex PC_LOAD_SEGMENT
3885 If defined, print information about the load segment for the program
3886 counter. (Defined only for the RS/6000.)
3890 If the program counter is kept in a register, then define this macro to
3891 be the number (greater than or equal to zero) of that register.
3893 This should only need to be defined if @code{TARGET_READ_PC} and
3894 @code{TARGET_WRITE_PC} are not defined.
3897 @findex PARM_BOUNDARY
3898 If non-zero, round arguments to a boundary of this many bits before
3899 pushing them on the stack.
3901 @item stabs_argument_has_addr (@var{gdbarch}, @var{type})
3902 @findex stabs_argument_has_addr
3903 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3904 @anchor{stabs_argument_has_addr} Define this to return nonzero if a
3905 function argument of type @var{type} is passed by reference instead of
3908 This method replaces @code{DEPRECATED_REG_STRUCT_HAS_ADDR}
3909 (@pxref{DEPRECATED_REG_STRUCT_HAS_ADDR}).
3911 @item PROCESS_LINENUMBER_HOOK
3912 @findex PROCESS_LINENUMBER_HOOK
3913 A hook defined for XCOFF reading.
3915 @item PROLOGUE_FIRSTLINE_OVERLAP
3916 @findex PROLOGUE_FIRSTLINE_OVERLAP
3917 (Only used in unsupported Convex configuration.)
3921 If defined, this is the number of the processor status register. (This
3922 definition is only used in generic code when parsing "$ps".)
3924 @item DEPRECATED_POP_FRAME
3925 @findex DEPRECATED_POP_FRAME
3927 If defined, used by @code{frame_pop} to remove a stack frame. This
3928 method has been superseeded by generic code.
3930 @item push_dummy_call (@var{gdbarch}, @var{function}, @var{regcache}, @var{pc_addr}, @var{nargs}, @var{args}, @var{sp}, @var{struct_return}, @var{struct_addr})
3931 @findex push_dummy_call
3932 @findex DEPRECATED_PUSH_ARGUMENTS.
3933 @anchor{push_dummy_call} Define this to push the dummy frame's call to
3934 the inferior function onto the stack. In addition to pushing
3935 @var{nargs}, the code should push @var{struct_addr} (when
3936 @var{struct_return}), and the return address (@var{bp_addr}).
3938 @var{function} is a pointer to a @code{struct value}; on architectures that use
3939 function descriptors, this contains the function descriptor value.
3941 Returns the updated top-of-stack pointer.
3943 This method replaces @code{DEPRECATED_PUSH_ARGUMENTS}.
3945 @item CORE_ADDR 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})
3946 @findex push_dummy_code
3947 @anchor{push_dummy_code} Given a stack based call dummy, push the
3948 instruction sequence (including space for a breakpoint) to which the
3949 called function should return.
3951 Set @var{bp_addr} to the address at which the breakpoint instruction
3952 should be inserted, @var{real_pc} to the resume address when starting
3953 the call sequence, and return the updated inner-most stack address.
3955 By default, the stack is grown sufficient to hold a frame-aligned
3956 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
3957 reserved for that breakpoint, and @var{real_pc} set to @var{funaddr}.
3959 This method replaces @code{CALL_DUMMY_LOCATION},
3960 @code{DEPRECATED_REGISTER_SIZE}.
3962 @item REGISTER_NAME(@var{i})
3963 @findex REGISTER_NAME
3964 Return the name of register @var{i} as a string. May return @code{NULL}
3965 or @code{NUL} to indicate that register @var{i} is not valid.
3967 @item DEPRECATED_REG_STRUCT_HAS_ADDR (@var{gcc_p}, @var{type})
3968 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3969 @anchor{DEPRECATED_REG_STRUCT_HAS_ADDR}Define this to return 1 if the
3970 given type will be passed by pointer rather than directly.
3972 This method has been replaced by @code{stabs_argument_has_addr}
3973 (@pxref{stabs_argument_has_addr}).
3975 @item SAVE_DUMMY_FRAME_TOS (@var{sp})
3976 @findex SAVE_DUMMY_FRAME_TOS
3977 @anchor{SAVE_DUMMY_FRAME_TOS} Used in @samp{call_function_by_hand} to
3978 notify the target dependent code of the top-of-stack value that will be
3979 passed to the the inferior code. This is the value of the @code{SP}
3980 after both the dummy frame and space for parameters/results have been
3981 allocated on the stack. @xref{unwind_dummy_id}.
3983 @item SDB_REG_TO_REGNUM
3984 @findex SDB_REG_TO_REGNUM
3985 Define this to convert sdb register numbers into @value{GDBN} regnums. If not
3986 defined, no conversion will be done.
3988 @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})
3989 @findex gdbarch_return_value
3990 @anchor{gdbarch_return_value} Given a function with a return-value of
3991 type @var{rettype}, return which return-value convention that function
3994 @value{GDBN} currently recognizes two function return-value conventions:
3995 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
3996 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
3997 value is found in memory and the address of that memory location is
3998 passed in as the function's first parameter.
4000 If the register convention is being used, and @var{writebuf} is
4001 non-@code{NULL}, also copy the return-value in @var{writebuf} into
4004 If the register convention is being used, and @var{readbuf} is
4005 non-@code{NULL}, also copy the return value from @var{regcache} into
4006 @var{readbuf} (@var{regcache} contains a copy of the registers from the
4007 just returned function).
4009 @xref{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}, for a description of how
4010 return-values that use the struct convention are handled.
4012 @emph{Maintainer note: This method replaces separate predicate, extract,
4013 store methods. By having only one method, the logic needed to determine
4014 the return-value convention need only be implemented in one place. If
4015 @value{GDBN} were written in an @sc{oo} language, this method would
4016 instead return an object that knew how to perform the register
4017 return-value extract and store.}
4019 @emph{Maintainer note: This method does not take a @var{gcc_p}
4020 parameter, and such a parameter should not be added. If an architecture
4021 that requires per-compiler or per-function information be identified,
4022 then the replacement of @var{rettype} with @code{struct value}
4023 @var{function} should be persued.}
4025 @emph{Maintainer note: The @var{regcache} parameter limits this methods
4026 to the inner most frame. While replacing @var{regcache} with a
4027 @code{struct frame_info} @var{frame} parameter would remove that
4028 limitation there has yet to be a demonstrated need for such a change.}
4030 @item SKIP_PERMANENT_BREAKPOINT
4031 @findex SKIP_PERMANENT_BREAKPOINT
4032 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
4033 steps over a breakpoint by removing it, stepping one instruction, and
4034 re-inserting the breakpoint. However, permanent breakpoints are
4035 hardwired into the inferior, and can't be removed, so this strategy
4036 doesn't work. Calling @code{SKIP_PERMANENT_BREAKPOINT} adjusts the processor's
4037 state so that execution will resume just after the breakpoint. This
4038 macro does the right thing even when the breakpoint is in the delay slot
4039 of a branch or jump.
4041 @item SKIP_PROLOGUE (@var{pc})
4042 @findex SKIP_PROLOGUE
4043 A C expression that returns the address of the ``real'' code beyond the
4044 function entry prologue found at @var{pc}.
4046 @item SKIP_TRAMPOLINE_CODE (@var{pc})
4047 @findex SKIP_TRAMPOLINE_CODE
4048 If the target machine has trampoline code that sits between callers and
4049 the functions being called, then define this macro to return a new PC
4050 that is at the start of the real function.
4054 If the stack-pointer is kept in a register, then define this macro to be
4055 the number (greater than or equal to zero) of that register, or -1 if
4056 there is no such register.
4058 @item STAB_REG_TO_REGNUM
4059 @findex STAB_REG_TO_REGNUM
4060 Define this to convert stab register numbers (as gotten from `r'
4061 declarations) into @value{GDBN} regnums. If not defined, no conversion will be
4064 @item DEPRECATED_STACK_ALIGN (@var{addr})
4065 @anchor{DEPRECATED_STACK_ALIGN}
4066 @findex DEPRECATED_STACK_ALIGN
4067 Define this to increase @var{addr} so that it meets the alignment
4068 requirements for the processor's stack.
4070 Unlike @ref{frame_align}, this function always adjusts @var{addr}
4073 By default, no stack alignment is performed.
4075 @item STEP_SKIPS_DELAY (@var{addr})
4076 @findex STEP_SKIPS_DELAY
4077 Define this to return true if the address is of an instruction with a
4078 delay slot. If a breakpoint has been placed in the instruction's delay
4079 slot, @value{GDBN} will single-step over that instruction before resuming
4080 normally. Currently only defined for the Mips.
4082 @item STORE_RETURN_VALUE (@var{type}, @var{regcache}, @var{valbuf})
4083 @findex STORE_RETURN_VALUE
4084 A C expression that writes the function return value, found in
4085 @var{valbuf}, into the @var{regcache}. @var{type} is the type of the
4086 value that is to be returned.
4088 This method has been deprecated in favour of @code{gdbarch_return_value}
4089 (@pxref{gdbarch_return_value}).
4091 @item SYMBOL_RELOADING_DEFAULT
4092 @findex SYMBOL_RELOADING_DEFAULT
4093 The default value of the ``symbol-reloading'' variable. (Never defined in
4096 @item TARGET_CHAR_BIT
4097 @findex TARGET_CHAR_BIT
4098 Number of bits in a char; defaults to 8.
4100 @item TARGET_CHAR_SIGNED
4101 @findex TARGET_CHAR_SIGNED
4102 Non-zero if @code{char} is normally signed on this architecture; zero if
4103 it should be unsigned.
4105 The ISO C standard requires the compiler to treat @code{char} as
4106 equivalent to either @code{signed char} or @code{unsigned char}; any
4107 character in the standard execution set is supposed to be positive.
4108 Most compilers treat @code{char} as signed, but @code{char} is unsigned
4109 on the IBM S/390, RS6000, and PowerPC targets.
4111 @item TARGET_COMPLEX_BIT
4112 @findex TARGET_COMPLEX_BIT
4113 Number of bits in a complex number; defaults to @code{2 * TARGET_FLOAT_BIT}.
4115 At present this macro is not used.
4117 @item TARGET_DOUBLE_BIT
4118 @findex TARGET_DOUBLE_BIT
4119 Number of bits in a double float; defaults to @code{8 * TARGET_CHAR_BIT}.
4121 @item TARGET_DOUBLE_COMPLEX_BIT
4122 @findex TARGET_DOUBLE_COMPLEX_BIT
4123 Number of bits in a double complex; defaults to @code{2 * TARGET_DOUBLE_BIT}.
4125 At present this macro is not used.
4127 @item TARGET_FLOAT_BIT
4128 @findex TARGET_FLOAT_BIT
4129 Number of bits in a float; defaults to @code{4 * TARGET_CHAR_BIT}.
4131 @item TARGET_INT_BIT
4132 @findex TARGET_INT_BIT
4133 Number of bits in an integer; defaults to @code{4 * TARGET_CHAR_BIT}.
4135 @item TARGET_LONG_BIT
4136 @findex TARGET_LONG_BIT
4137 Number of bits in a long integer; defaults to @code{4 * TARGET_CHAR_BIT}.
4139 @item TARGET_LONG_DOUBLE_BIT
4140 @findex TARGET_LONG_DOUBLE_BIT
4141 Number of bits in a long double float;
4142 defaults to @code{2 * TARGET_DOUBLE_BIT}.
4144 @item TARGET_LONG_LONG_BIT
4145 @findex TARGET_LONG_LONG_BIT
4146 Number of bits in a long long integer; defaults to @code{2 * TARGET_LONG_BIT}.
4148 @item TARGET_PTR_BIT
4149 @findex TARGET_PTR_BIT
4150 Number of bits in a pointer; defaults to @code{TARGET_INT_BIT}.
4152 @item TARGET_SHORT_BIT
4153 @findex TARGET_SHORT_BIT
4154 Number of bits in a short integer; defaults to @code{2 * TARGET_CHAR_BIT}.
4156 @item TARGET_READ_PC
4157 @findex TARGET_READ_PC
4158 @itemx TARGET_WRITE_PC (@var{val}, @var{pid})
4159 @findex TARGET_WRITE_PC
4160 @anchor{TARGET_WRITE_PC}
4161 @itemx TARGET_READ_SP
4162 @findex TARGET_READ_SP
4163 @itemx TARGET_READ_FP
4164 @findex TARGET_READ_FP
4169 @anchor{TARGET_READ_SP} These change the behavior of @code{read_pc},
4170 @code{write_pc}, and @code{read_sp}. For most targets, these may be
4171 left undefined. @value{GDBN} will call the read and write register
4172 functions with the relevant @code{_REGNUM} argument.
4174 These macros are useful when a target keeps one of these registers in a
4175 hard to get at place; for example, part in a segment register and part
4176 in an ordinary register.
4178 @xref{unwind_sp}, which replaces @code{TARGET_READ_SP}.
4180 @item TARGET_VIRTUAL_FRAME_POINTER(@var{pc}, @var{regp}, @var{offsetp})
4181 @findex TARGET_VIRTUAL_FRAME_POINTER
4182 Returns a @code{(register, offset)} pair representing the virtual frame
4183 pointer in use at the code address @var{pc}. If virtual frame pointers
4184 are not used, a default definition simply returns
4185 @code{DEPRECATED_FP_REGNUM}, with an offset of zero.
4187 @item TARGET_HAS_HARDWARE_WATCHPOINTS
4188 If non-zero, the target has support for hardware-assisted
4189 watchpoints. @xref{Algorithms, watchpoints}, for more details and
4190 other related macros.
4192 @item TARGET_PRINT_INSN (@var{addr}, @var{info})
4193 @findex TARGET_PRINT_INSN
4194 This is the function used by @value{GDBN} to print an assembly
4195 instruction. It prints the instruction at address @var{addr} in
4196 debugged memory and returns the length of the instruction, in bytes. If
4197 a target doesn't define its own printing routine, it defaults to an
4198 accessor function for the global pointer
4199 @code{deprecated_tm_print_insn}. This usually points to a function in
4200 the @code{opcodes} library (@pxref{Support Libraries, ,Opcodes}).
4201 @var{info} is a structure (of type @code{disassemble_info}) defined in
4202 @file{include/dis-asm.h} used to pass information to the instruction
4205 @item struct frame_id unwind_dummy_id (struct frame_info *@var{frame})
4206 @findex unwind_dummy_id
4207 @anchor{unwind_dummy_id} Given @var{frame} return a @code{struct
4208 frame_id} that uniquely identifies an inferior function call's dummy
4209 frame. The value returned must match the dummy frame stack value
4210 previously saved using @code{SAVE_DUMMY_FRAME_TOS}.
4211 @xref{SAVE_DUMMY_FRAME_TOS}.
4213 @item DEPRECATED_USE_STRUCT_CONVENTION (@var{gcc_p}, @var{type})
4214 @findex DEPRECATED_USE_STRUCT_CONVENTION
4215 If defined, this must be an expression that is nonzero if a value of the
4216 given @var{type} being returned from a function must have space
4217 allocated for it on the stack. @var{gcc_p} is true if the function
4218 being considered is known to have been compiled by GCC; this is helpful
4219 for systems where GCC is known to use different calling convention than
4222 This method has been deprecated in favour of @code{gdbarch_return_value}
4223 (@pxref{gdbarch_return_value}).
4225 @item VALUE_TO_REGISTER(@var{type}, @var{regnum}, @var{from}, @var{to})
4226 @findex VALUE_TO_REGISTER
4227 Convert a value of type @var{type} into the raw contents of register
4229 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
4231 @item VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
4232 @findex VARIABLES_INSIDE_BLOCK
4233 For dbx-style debugging information, if the compiler puts variable
4234 declarations inside LBRAC/RBRAC blocks, this should be defined to be
4235 nonzero. @var{desc} is the value of @code{n_desc} from the
4236 @code{N_RBRAC} symbol, and @var{gcc_p} is true if @value{GDBN} has noticed the
4237 presence of either the @code{GCC_COMPILED_SYMBOL} or the
4238 @code{GCC2_COMPILED_SYMBOL}. By default, this is 0.
4240 @item OS9K_VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
4241 @findex OS9K_VARIABLES_INSIDE_BLOCK
4242 Similarly, for OS/9000. Defaults to 1.
4245 Motorola M68K target conditionals.
4249 Define this to be the 4-bit location of the breakpoint trap vector. If
4250 not defined, it will default to @code{0xf}.
4252 @item REMOTE_BPT_VECTOR
4253 Defaults to @code{1}.
4255 @item NAME_OF_MALLOC
4256 @findex NAME_OF_MALLOC
4257 A string containing the name of the function to call in order to
4258 allocate some memory in the inferior. The default value is "malloc".
4262 @section Adding a New Target
4264 @cindex adding a target
4265 The following files add a target to @value{GDBN}:
4269 @item gdb/config/@var{arch}/@var{ttt}.mt
4270 Contains a Makefile fragment specific to this target. Specifies what
4271 object files are needed for target @var{ttt}, by defining
4272 @samp{TDEPFILES=@dots{}} and @samp{TDEPLIBS=@dots{}}. Also specifies
4273 the header file which describes @var{ttt}, by defining @samp{TM_FILE=
4276 You can also define @samp{TM_CFLAGS}, @samp{TM_CLIBS}, @samp{TM_CDEPS},
4277 but these are now deprecated, replaced by autoconf, and may go away in
4278 future versions of @value{GDBN}.
4280 @item gdb/@var{ttt}-tdep.c
4281 Contains any miscellaneous code required for this target machine. On
4282 some machines it doesn't exist at all. Sometimes the macros in
4283 @file{tm-@var{ttt}.h} become very complicated, so they are implemented
4284 as functions here instead, and the macro is simply defined to call the
4285 function. This is vastly preferable, since it is easier to understand
4288 @item gdb/@var{arch}-tdep.c
4289 @itemx gdb/@var{arch}-tdep.h
4290 This often exists to describe the basic layout of the target machine's
4291 processor chip (registers, stack, etc.). If used, it is included by
4292 @file{@var{ttt}-tdep.h}. It can be shared among many targets that use
4295 @item gdb/config/@var{arch}/tm-@var{ttt}.h
4296 (@file{tm.h} is a link to this file, created by @code{configure}). Contains
4297 macro definitions about the target machine's registers, stack frame
4298 format and instructions.
4300 New targets do not need this file and should not create it.
4302 @item gdb/config/@var{arch}/tm-@var{arch}.h
4303 This often exists to describe the basic layout of the target machine's
4304 processor chip (registers, stack, etc.). If used, it is included by
4305 @file{tm-@var{ttt}.h}. It can be shared among many targets that use the
4308 New targets do not need this file and should not create it.
4312 If you are adding a new operating system for an existing CPU chip, add a
4313 @file{config/tm-@var{os}.h} file that describes the operating system
4314 facilities that are unusual (extra symbol table info; the breakpoint
4315 instruction needed; etc.). Then write a @file{@var{arch}/tm-@var{os}.h}
4316 that just @code{#include}s @file{tm-@var{arch}.h} and
4317 @file{config/tm-@var{os}.h}.
4320 @section Converting an existing Target Architecture to Multi-arch
4321 @cindex converting targets to multi-arch
4323 This section describes the current accepted best practice for converting
4324 an existing target architecture to the multi-arch framework.
4326 The process consists of generating, testing, posting and committing a
4327 sequence of patches. Each patch must contain a single change, for
4333 Directly convert a group of functions into macros (the conversion does
4334 not change the behavior of any of the functions).
4337 Replace a non-multi-arch with a multi-arch mechanism (e.g.,
4341 Enable multi-arch level one.
4344 Delete one or more files.
4349 There isn't a size limit on a patch, however, a developer is strongly
4350 encouraged to keep the patch size down.
4352 Since each patch is well defined, and since each change has been tested
4353 and shows no regressions, the patches are considered @emph{fairly}
4354 obvious. Such patches, when submitted by developers listed in the
4355 @file{MAINTAINERS} file, do not need approval. Occasional steps in the
4356 process may be more complicated and less clear. The developer is
4357 expected to use their judgment and is encouraged to seek advice as
4360 @subsection Preparation
4362 The first step is to establish control. Build (with @option{-Werror}
4363 enabled) and test the target so that there is a baseline against which
4364 the debugger can be compared.
4366 At no stage can the test results regress or @value{GDBN} stop compiling
4367 with @option{-Werror}.
4369 @subsection Add the multi-arch initialization code
4371 The objective of this step is to establish the basic multi-arch
4372 framework. It involves
4377 The addition of a @code{@var{arch}_gdbarch_init} function@footnote{The
4378 above is from the original example and uses K&R C. @value{GDBN}
4379 has since converted to ISO C but lets ignore that.} that creates
4382 static struct gdbarch *
4383 d10v_gdbarch_init (info, arches)
4384 struct gdbarch_info info;
4385 struct gdbarch_list *arches;
4387 struct gdbarch *gdbarch;
4388 /* there is only one d10v architecture */
4390 return arches->gdbarch;
4391 gdbarch = gdbarch_alloc (&info, NULL);
4399 A per-architecture dump function to print any architecture specific
4403 mips_dump_tdep (struct gdbarch *current_gdbarch,
4404 struct ui_file *file)
4406 @dots{} code to print architecture specific info @dots{}
4411 A change to @code{_initialize_@var{arch}_tdep} to register this new
4415 _initialize_mips_tdep (void)
4417 gdbarch_register (bfd_arch_mips, mips_gdbarch_init,
4422 Add the macro @code{GDB_MULTI_ARCH}, defined as 0 (zero), to the file@*
4423 @file{config/@var{arch}/tm-@var{arch}.h}.
4427 @subsection Update multi-arch incompatible mechanisms
4429 Some mechanisms do not work with multi-arch. They include:
4432 @item FRAME_FIND_SAVED_REGS
4433 Replaced with @code{DEPRECATED_FRAME_INIT_SAVED_REGS}
4437 At this stage you could also consider converting the macros into
4440 @subsection Prepare for multi-arch level to one
4442 Temporally set @code{GDB_MULTI_ARCH} to @code{GDB_MULTI_ARCH_PARTIAL}
4443 and then build and start @value{GDBN} (the change should not be
4444 committed). @value{GDBN} may not build, and once built, it may die with
4445 an internal error listing the architecture methods that must be
4448 Fix any build problems (patch(es)).
4450 Convert all the architecture methods listed, which are only macros, into
4451 functions (patch(es)).
4453 Update @code{@var{arch}_gdbarch_init} to set all the missing
4454 architecture methods and wrap the corresponding macros in @code{#if
4455 !GDB_MULTI_ARCH} (patch(es)).
4457 @subsection Set multi-arch level one
4459 Change the value of @code{GDB_MULTI_ARCH} to GDB_MULTI_ARCH_PARTIAL (a
4462 Any problems with throwing ``the switch'' should have been fixed
4465 @subsection Convert remaining macros
4467 Suggest converting macros into functions (and setting the corresponding
4468 architecture method) in small batches.
4470 @subsection Set multi-arch level to two
4472 This should go smoothly.
4474 @subsection Delete the TM file
4476 The @file{tm-@var{arch}.h} can be deleted. @file{@var{arch}.mt} and
4477 @file{configure.in} updated.
4480 @node Target Vector Definition
4482 @chapter Target Vector Definition
4483 @cindex target vector
4485 The target vector defines the interface between @value{GDBN}'s
4486 abstract handling of target systems, and the nitty-gritty code that
4487 actually exercises control over a process or a serial port.
4488 @value{GDBN} includes some 30-40 different target vectors; however,
4489 each configuration of @value{GDBN} includes only a few of them.
4492 * Managing Execution State::
4493 * Existing Targets::
4496 @node Managing Execution State
4497 @section Managing Execution State
4498 @cindex execution state
4500 A target vector can be completely inactive (not pushed on the target
4501 stack), active but not running (pushed, but not connected to a fully
4502 manifested inferior), or completely active (pushed, with an accessible
4503 inferior). Most targets are only completely inactive or completely
4504 active, but some support persistant connections to a target even
4505 when the target has exited or not yet started.
4507 For example, connecting to the simulator using @code{target sim} does
4508 not create a running program. Neither registers nor memory are
4509 accessible until @code{run}. Similarly, after @code{kill}, the
4510 program can not continue executing. But in both cases @value{GDBN}
4511 remains connected to the simulator, and target-specific commands
4512 are directed to the simulator.
4514 A target which only supports complete activation should push itself
4515 onto the stack in its @code{to_open} routine (by calling
4516 @code{push_target}), and unpush itself from the stack in its
4517 @code{to_mourn_inferior} routine (by calling @code{unpush_target}).
4519 A target which supports both partial and complete activation should
4520 still call @code{push_target} in @code{to_open}, but not call
4521 @code{unpush_target} in @code{to_mourn_inferior}. Instead, it should
4522 call either @code{target_mark_running} or @code{target_mark_exited}
4523 in its @code{to_open}, depending on whether the target is fully active
4524 after connection. It should also call @code{target_mark_running} any
4525 time the inferior becomes fully active (e.g.@: in
4526 @code{to_create_inferior} and @code{to_attach}), and
4527 @code{target_mark_exited} when the inferior becomes inactive (in
4528 @code{to_mourn_inferior}). The target should also make sure to call
4529 @code{target_mourn_inferior} from its @code{to_kill}, to return the
4530 target to inactive state.
4532 @node Existing Targets
4533 @section Existing Targets
4536 @subsection File Targets
4538 Both executables and core files have target vectors.
4540 @subsection Standard Protocol and Remote Stubs
4542 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code
4543 that runs in the target system. @value{GDBN} provides several sample
4544 @dfn{stubs} that can be integrated into target programs or operating
4545 systems for this purpose; they are named @file{*-stub.c}.
4547 The @value{GDBN} user's manual describes how to put such a stub into
4548 your target code. What follows is a discussion of integrating the
4549 SPARC stub into a complicated operating system (rather than a simple
4550 program), by Stu Grossman, the author of this stub.
4552 The trap handling code in the stub assumes the following upon entry to
4557 %l1 and %l2 contain pc and npc respectively at the time of the trap;
4563 you are in the correct trap window.
4566 As long as your trap handler can guarantee those conditions, then there
4567 is no reason why you shouldn't be able to ``share'' traps with the stub.
4568 The stub has no requirement that it be jumped to directly from the
4569 hardware trap vector. That is why it calls @code{exceptionHandler()},
4570 which is provided by the external environment. For instance, this could
4571 set up the hardware traps to actually execute code which calls the stub
4572 first, and then transfers to its own trap handler.
4574 For the most point, there probably won't be much of an issue with
4575 ``sharing'' traps, as the traps we use are usually not used by the kernel,
4576 and often indicate unrecoverable error conditions. Anyway, this is all
4577 controlled by a table, and is trivial to modify. The most important
4578 trap for us is for @code{ta 1}. Without that, we can't single step or
4579 do breakpoints. Everything else is unnecessary for the proper operation
4580 of the debugger/stub.
4582 From reading the stub, it's probably not obvious how breakpoints work.
4583 They are simply done by deposit/examine operations from @value{GDBN}.
4585 @subsection ROM Monitor Interface
4587 @subsection Custom Protocols
4589 @subsection Transport Layer
4591 @subsection Builtin Simulator
4594 @node Native Debugging
4596 @chapter Native Debugging
4597 @cindex native debugging
4599 Several files control @value{GDBN}'s configuration for native support:
4603 @item gdb/config/@var{arch}/@var{xyz}.mh
4604 Specifies Makefile fragments needed by a @emph{native} configuration on
4605 machine @var{xyz}. In particular, this lists the required
4606 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
4607 Also specifies the header file which describes native support on
4608 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
4609 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
4610 @samp{NAT_CDEPS}, etc.; see @file{Makefile.in}.
4612 @emph{Maintainer's note: The @file{.mh} suffix is because this file
4613 originally contained @file{Makefile} fragments for hosting @value{GDBN}
4614 on machine @var{xyz}. While the file is no longer used for this
4615 purpose, the @file{.mh} suffix remains. Perhaps someone will
4616 eventually rename these fragments so that they have a @file{.mn}
4619 @item gdb/config/@var{arch}/nm-@var{xyz}.h
4620 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
4621 macro definitions describing the native system environment, such as
4622 child process control and core file support.
4624 @item gdb/@var{xyz}-nat.c
4625 Contains any miscellaneous C code required for this native support of
4626 this machine. On some machines it doesn't exist at all.
4629 There are some ``generic'' versions of routines that can be used by
4630 various systems. These can be customized in various ways by macros
4631 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
4632 the @var{xyz} host, you can just include the generic file's name (with
4633 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
4635 Otherwise, if your machine needs custom support routines, you will need
4636 to write routines that perform the same functions as the generic file.
4637 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
4638 into @code{NATDEPFILES}.
4642 This contains the @emph{target_ops vector} that supports Unix child
4643 processes on systems which use ptrace and wait to control the child.
4646 This contains the @emph{target_ops vector} that supports Unix child
4647 processes on systems which use /proc to control the child.
4650 This does the low-level grunge that uses Unix system calls to do a ``fork
4651 and exec'' to start up a child process.
4654 This is the low level interface to inferior processes for systems using
4655 the Unix @code{ptrace} call in a vanilla way.
4658 @section Native core file Support
4659 @cindex native core files
4662 @findex fetch_core_registers
4663 @item core-aout.c::fetch_core_registers()
4664 Support for reading registers out of a core file. This routine calls
4665 @code{register_addr()}, see below. Now that BFD is used to read core
4666 files, virtually all machines should use @code{core-aout.c}, and should
4667 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
4668 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
4670 @item core-aout.c::register_addr()
4671 If your @code{nm-@var{xyz}.h} file defines the macro
4672 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
4673 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
4674 register number @code{regno}. @code{blockend} is the offset within the
4675 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
4676 @file{core-aout.c} will define the @code{register_addr()} function and
4677 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
4678 you are using the standard @code{fetch_core_registers()}, you will need
4679 to define your own version of @code{register_addr()}, put it into your
4680 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
4681 the @code{NATDEPFILES} list. If you have your own
4682 @code{fetch_core_registers()}, you may not need a separate
4683 @code{register_addr()}. Many custom @code{fetch_core_registers()}
4684 implementations simply locate the registers themselves.@refill
4687 When making @value{GDBN} run native on a new operating system, to make it
4688 possible to debug core files, you will need to either write specific
4689 code for parsing your OS's core files, or customize
4690 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
4691 machine uses to define the struct of registers that is accessible
4692 (possibly in the u-area) in a core file (rather than
4693 @file{machine/reg.h}), and an include file that defines whatever header
4694 exists on a core file (e.g., the u-area or a @code{struct core}). Then
4695 modify @code{trad_unix_core_file_p} to use these values to set up the
4696 section information for the data segment, stack segment, any other
4697 segments in the core file (perhaps shared library contents or control
4698 information), ``registers'' segment, and if there are two discontiguous
4699 sets of registers (e.g., integer and float), the ``reg2'' segment. This
4700 section information basically delimits areas in the core file in a
4701 standard way, which the section-reading routines in BFD know how to seek
4704 Then back in @value{GDBN}, you need a matching routine called
4705 @code{fetch_core_registers}. If you can use the generic one, it's in
4706 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
4707 It will be passed a char pointer to the entire ``registers'' segment,
4708 its length, and a zero; or a char pointer to the entire ``regs2''
4709 segment, its length, and a 2. The routine should suck out the supplied
4710 register values and install them into @value{GDBN}'s ``registers'' array.
4712 If your system uses @file{/proc} to control processes, and uses ELF
4713 format core files, then you may be able to use the same routines for
4714 reading the registers out of processes and out of core files.
4722 @section shared libraries
4724 @section Native Conditionals
4725 @cindex native conditionals
4727 When @value{GDBN} is configured and compiled, various macros are
4728 defined or left undefined, to control compilation when the host and
4729 target systems are the same. These macros should be defined (or left
4730 undefined) in @file{nm-@var{system}.h}.
4734 @item CHILD_PREPARE_TO_STORE
4735 @findex CHILD_PREPARE_TO_STORE
4736 If the machine stores all registers at once in the child process, then
4737 define this to ensure that all values are correct. This usually entails
4738 a read from the child.
4740 [Note that this is incorrectly defined in @file{xm-@var{system}.h} files
4743 @item FETCH_INFERIOR_REGISTERS
4744 @findex FETCH_INFERIOR_REGISTERS
4745 Define this if the native-dependent code will provide its own routines
4746 @code{fetch_inferior_registers} and @code{store_inferior_registers} in
4747 @file{@var{host}-nat.c}. If this symbol is @emph{not} defined, and
4748 @file{infptrace.c} is included in this configuration, the default
4749 routines in @file{infptrace.c} are used for these functions.
4753 This macro is normally defined to be the number of the first floating
4754 point register, if the machine has such registers. As such, it would
4755 appear only in target-specific code. However, @file{/proc} support uses this
4756 to decide whether floats are in use on this target.
4758 @item GET_LONGJMP_TARGET
4759 @findex GET_LONGJMP_TARGET
4760 For most machines, this is a target-dependent parameter. On the
4761 DECstation and the Iris, this is a native-dependent parameter, since
4762 @file{setjmp.h} is needed to define it.
4764 This macro determines the target PC address that @code{longjmp} will jump to,
4765 assuming that we have just stopped at a longjmp breakpoint. It takes a
4766 @code{CORE_ADDR *} as argument, and stores the target PC value through this
4767 pointer. It examines the current state of the machine as needed.
4769 @item I386_USE_GENERIC_WATCHPOINTS
4770 An x86-based machine can define this to use the generic x86 watchpoint
4771 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4774 @findex KERNEL_U_ADDR
4775 Define this to the address of the @code{u} structure (the ``user
4776 struct'', also known as the ``u-page'') in kernel virtual memory. @value{GDBN}
4777 needs to know this so that it can subtract this address from absolute
4778 addresses in the upage, that are obtained via ptrace or from core files.
4779 On systems that don't need this value, set it to zero.
4781 @item KERNEL_U_ADDR_HPUX
4782 @findex KERNEL_U_ADDR_HPUX
4783 Define this to cause @value{GDBN} to determine the address of @code{u} at
4784 runtime, by using HP-style @code{nlist} on the kernel's image in the
4787 @item ONE_PROCESS_WRITETEXT
4788 @findex ONE_PROCESS_WRITETEXT
4789 Define this to be able to, when a breakpoint insertion fails, warn the
4790 user that another process may be running with the same executable.
4793 @findex PROC_NAME_FMT
4794 Defines the format for the name of a @file{/proc} device. Should be
4795 defined in @file{nm.h} @emph{only} in order to override the default
4796 definition in @file{procfs.c}.
4798 @item PTRACE_ARG3_TYPE
4799 @findex PTRACE_ARG3_TYPE
4800 The type of the third argument to the @code{ptrace} system call, if it
4801 exists and is different from @code{int}.
4803 @item REGISTER_U_ADDR
4804 @findex REGISTER_U_ADDR
4805 Defines the offset of the registers in the ``u area''.
4807 @item SHELL_COMMAND_CONCAT
4808 @findex SHELL_COMMAND_CONCAT
4809 If defined, is a string to prefix on the shell command used to start the
4814 If defined, this is the name of the shell to use to run the inferior.
4815 Defaults to @code{"/bin/sh"}.
4817 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
4819 Define this to expand into an expression that will cause the symbols in
4820 @var{filename} to be added to @value{GDBN}'s symbol table. If
4821 @var{readsyms} is zero symbols are not read but any necessary low level
4822 processing for @var{filename} is still done.
4824 @item SOLIB_CREATE_INFERIOR_HOOK
4825 @findex SOLIB_CREATE_INFERIOR_HOOK
4826 Define this to expand into any shared-library-relocation code that you
4827 want to be run just after the child process has been forked.
4829 @item START_INFERIOR_TRAPS_EXPECTED
4830 @findex START_INFERIOR_TRAPS_EXPECTED
4831 When starting an inferior, @value{GDBN} normally expects to trap
4833 the shell execs, and once when the program itself execs. If the actual
4834 number of traps is something other than 2, then define this macro to
4835 expand into the number expected.
4839 This determines whether small routines in @file{*-tdep.c}, which
4840 translate register values between @value{GDBN}'s internal
4841 representation and the @file{/proc} representation, are compiled.
4844 @findex U_REGS_OFFSET
4845 This is the offset of the registers in the upage. It need only be
4846 defined if the generic ptrace register access routines in
4847 @file{infptrace.c} are being used (that is, @file{infptrace.c} is
4848 configured in, and @code{FETCH_INFERIOR_REGISTERS} is not defined). If
4849 the default value from @file{infptrace.c} is good enough, leave it
4852 The default value means that u.u_ar0 @emph{points to} the location of
4853 the registers. I'm guessing that @code{#define U_REGS_OFFSET 0} means
4854 that @code{u.u_ar0} @emph{is} the location of the registers.
4858 See @file{objfiles.c}.
4861 @findex DEBUG_PTRACE
4862 Define this to debug @code{ptrace} calls.
4866 @node Support Libraries
4868 @chapter Support Libraries
4873 BFD provides support for @value{GDBN} in several ways:
4876 @item identifying executable and core files
4877 BFD will identify a variety of file types, including a.out, coff, and
4878 several variants thereof, as well as several kinds of core files.
4880 @item access to sections of files
4881 BFD parses the file headers to determine the names, virtual addresses,
4882 sizes, and file locations of all the various named sections in files
4883 (such as the text section or the data section). @value{GDBN} simply
4884 calls BFD to read or write section @var{x} at byte offset @var{y} for
4887 @item specialized core file support
4888 BFD provides routines to determine the failing command name stored in a
4889 core file, the signal with which the program failed, and whether a core
4890 file matches (i.e.@: could be a core dump of) a particular executable
4893 @item locating the symbol information
4894 @value{GDBN} uses an internal interface of BFD to determine where to find the
4895 symbol information in an executable file or symbol-file. @value{GDBN} itself
4896 handles the reading of symbols, since BFD does not ``understand'' debug
4897 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
4902 @cindex opcodes library
4904 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
4905 library because it's also used in binutils, for @file{objdump}).
4912 @cindex @code{libiberty} library
4914 The @code{libiberty} library provides a set of functions and features
4915 that integrate and improve on functionality found in modern operating
4916 systems. Broadly speaking, such features can be divided into three
4917 groups: supplemental functions (functions that may be missing in some
4918 environments and operating systems), replacement functions (providing
4919 a uniform and easier to use interface for commonly used standard
4920 functions), and extensions (which provide additional functionality
4921 beyond standard functions).
4923 @value{GDBN} uses various features provided by the @code{libiberty}
4924 library, for instance the C@t{++} demangler, the @acronym{IEEE}
4925 floating format support functions, the input options parser
4926 @samp{getopt}, the @samp{obstack} extension, and other functions.
4928 @subsection @code{obstacks} in @value{GDBN}
4929 @cindex @code{obstacks}
4931 The obstack mechanism provides a convenient way to allocate and free
4932 chunks of memory. Each obstack is a pool of memory that is managed
4933 like a stack. Objects (of any nature, size and alignment) are
4934 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
4935 @code{libiberty}'s documenatation for a more detailed explanation of
4938 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
4939 object files. There is an obstack associated with each internal
4940 representation of an object file. Lots of things get allocated on
4941 these @code{obstacks}: dictionary entries, blocks, blockvectors,
4942 symbols, minimal symbols, types, vectors of fundamental types, class
4943 fields of types, object files section lists, object files section
4944 offets lists, line tables, symbol tables, partial symbol tables,
4945 string tables, symbol table private data, macros tables, debug
4946 information sections and entries, import and export lists (som),
4947 unwind information (hppa), dwarf2 location expressions data. Plus
4948 various strings such as directory names strings, debug format strings,
4951 An essential and convenient property of all data on @code{obstacks} is
4952 that memory for it gets allocated (with @code{obstack_alloc}) at
4953 various times during a debugging sesssion, but it is released all at
4954 once using the @code{obstack_free} function. The @code{obstack_free}
4955 function takes a pointer to where in the stack it must start the
4956 deletion from (much like the cleanup chains have a pointer to where to
4957 start the cleanups). Because of the stack like structure of the
4958 @code{obstacks}, this allows to free only a top portion of the
4959 obstack. There are a few instances in @value{GDBN} where such thing
4960 happens. Calls to @code{obstack_free} are done after some local data
4961 is allocated to the obstack. Only the local data is deleted from the
4962 obstack. Of course this assumes that nothing between the
4963 @code{obstack_alloc} and the @code{obstack_free} allocates anything
4964 else on the same obstack. For this reason it is best and safest to
4965 use temporary @code{obstacks}.
4967 Releasing the whole obstack is also not safe per se. It is safe only
4968 under the condition that we know the @code{obstacks} memory is no
4969 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
4970 when we get rid of the whole objfile(s), for instance upon reading a
4974 @cindex regular expressions library
4985 @item SIGN_EXTEND_CHAR
4987 @item SWITCH_ENUM_BUG
4996 @section Array Containers
4997 @cindex Array Containers
5000 Often it is necessary to manipulate a dynamic array of a set of
5001 objects. C forces some bookkeeping on this, which can get cumbersome
5002 and repetative. The @file{vec.h} file contains macros for defining
5003 and using a typesafe vector type. The functions defined will be
5004 inlined when compiling, and so the abstraction cost should be zero.
5005 Domain checks are added to detect programming errors.
5007 An example use would be an array of symbols or section information.
5008 The array can be grown as symbols are read in (or preallocated), and
5009 the accessor macros provided keep care of all the necessary
5010 bookkeeping. Because the arrays are type safe, there is no danger of
5011 accidentally mixing up the contents. Think of these as C++ templates,
5012 but implemented in C.
5014 Because of the different behavior of structure objects, scalar objects
5015 and of pointers, there are three flavors of vector, one for each of
5016 these variants. Both the structure object and pointer variants pass
5017 pointers to objects around --- in the former case the pointers are
5018 stored into the vector and in the latter case the pointers are
5019 dereferenced and the objects copied into the vector. The scalar
5020 object variant is suitable for @code{int}-like objects, and the vector
5021 elements are returned by value.
5023 There are both @code{index} and @code{iterate} accessors. The iterator
5024 returns a boolean iteration condition and updates the iteration
5025 variable passed by reference. Because the iterator will be inlined,
5026 the address-of can be optimized away.
5028 The vectors are implemented using the trailing array idiom, thus they
5029 are not resizeable without changing the address of the vector object
5030 itself. This means you cannot have variables or fields of vector type
5031 --- always use a pointer to a vector. The one exception is the final
5032 field of a structure, which could be a vector type. You will have to
5033 use the @code{embedded_size} & @code{embedded_init} calls to create
5034 such objects, and they will probably not be resizeable (so don't use
5035 the @dfn{safe} allocation variants). The trailing array idiom is used
5036 (rather than a pointer to an array of data), because, if we allow
5037 @code{NULL} to also represent an empty vector, empty vectors occupy
5038 minimal space in the structure containing them.
5040 Each operation that increases the number of active elements is
5041 available in @dfn{quick} and @dfn{safe} variants. The former presumes
5042 that there is sufficient allocated space for the operation to succeed
5043 (it dies if there is not). The latter will reallocate the vector, if
5044 needed. Reallocation causes an exponential increase in vector size.
5045 If you know you will be adding N elements, it would be more efficient
5046 to use the reserve operation before adding the elements with the
5047 @dfn{quick} operation. This will ensure there are at least as many
5048 elements as you ask for, it will exponentially increase if there are
5049 too few spare slots. If you want reserve a specific number of slots,
5050 but do not want the exponential increase (for instance, you know this
5051 is the last allocation), use a negative number for reservation. You
5052 can also create a vector of a specific size from the get go.
5054 You should prefer the push and pop operations, as they append and
5055 remove from the end of the vector. If you need to remove several items
5056 in one go, use the truncate operation. The insert and remove
5057 operations allow you to change elements in the middle of the vector.
5058 There are two remove operations, one which preserves the element
5059 ordering @code{ordered_remove}, and one which does not
5060 @code{unordered_remove}. The latter function copies the end element
5061 into the removed slot, rather than invoke a memmove operation. The
5062 @code{lower_bound} function will determine where to place an item in
5063 the array using insert that will maintain sorted order.
5065 If you need to directly manipulate a vector, then the @code{address}
5066 accessor will return the address of the start of the vector. Also the
5067 @code{space} predicate will tell you whether there is spare capacity in the
5068 vector. You will not normally need to use these two functions.
5070 Vector types are defined using a
5071 @code{DEF_VEC_@{O,P,I@}(@var{typename})} macro. Variables of vector
5072 type are declared using a @code{VEC(@var{typename})} macro. The
5073 characters @code{O}, @code{P} and @code{I} indicate whether
5074 @var{typename} is an object (@code{O}), pointer (@code{P}) or integral
5075 (@code{I}) type. Be careful to pick the correct one, as you'll get an
5076 awkward and inefficient API if you use the wrong one. There is a
5077 check, which results in a compile-time warning, for the @code{P} and
5078 @code{I} versions, but there is no check for the @code{O} versions, as
5079 that is not possible in plain C.
5081 An example of their use would be,
5084 DEF_VEC_P(tree); // non-managed tree vector.
5087 VEC(tree) *v; // A (pointer to) a vector of tree pointers.
5090 struct my_struct *s;
5092 if (VEC_length(tree, s->v)) @{ we have some contents @}
5093 VEC_safe_push(tree, s->v, decl); // append some decl onto the end
5094 for (ix = 0; VEC_iterate(tree, s->v, ix, elt); ix++)
5095 @{ do something with elt @}
5099 The @file{vec.h} file provides details on how to invoke the various
5100 accessors provided. They are enumerated here:
5104 Return the number of items in the array,
5107 Return true if the array has no elements.
5111 Return the last or arbitrary item in the array.
5114 Access an array element and indicate whether the array has been
5119 Create and destroy an array.
5121 @item VEC_embedded_size
5122 @itemx VEC_embedded_init
5123 Helpers for embedding an array as the final element of another struct.
5129 Return the amount of free space in an array.
5132 Ensure a certain amount of free space.
5134 @item VEC_quick_push
5135 @itemx VEC_safe_push
5136 Append to an array, either assuming the space is available, or making
5140 Remove the last item from an array.
5143 Remove several items from the end of an array.
5146 Add several items to the end of an array.
5149 Overwrite an item in the array.
5151 @item VEC_quick_insert
5152 @itemx VEC_safe_insert
5153 Insert an item into the middle of the array. Either the space must
5154 already exist, or the space is created.
5156 @item VEC_ordered_remove
5157 @itemx VEC_unordered_remove
5158 Remove an item from the array, preserving order or not.
5160 @item VEC_block_remove
5161 Remove a set of items from the array.
5164 Provide the address of the first element.
5166 @item VEC_lower_bound
5167 Binary search the array.
5177 This chapter covers topics that are lower-level than the major
5178 algorithms of @value{GDBN}.
5183 Cleanups are a structured way to deal with things that need to be done
5186 When your code does something (e.g., @code{xmalloc} some memory, or
5187 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
5188 the memory or @code{close} the file), it can make a cleanup. The
5189 cleanup will be done at some future point: when the command is finished
5190 and control returns to the top level; when an error occurs and the stack
5191 is unwound; or when your code decides it's time to explicitly perform
5192 cleanups. Alternatively you can elect to discard the cleanups you
5198 @item struct cleanup *@var{old_chain};
5199 Declare a variable which will hold a cleanup chain handle.
5201 @findex make_cleanup
5202 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
5203 Make a cleanup which will cause @var{function} to be called with
5204 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
5205 handle that can later be passed to @code{do_cleanups} or
5206 @code{discard_cleanups}. Unless you are going to call
5207 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
5208 from @code{make_cleanup}.
5211 @item do_cleanups (@var{old_chain});
5212 Do all cleanups added to the chain since the corresponding
5213 @code{make_cleanup} call was made.
5215 @findex discard_cleanups
5216 @item discard_cleanups (@var{old_chain});
5217 Same as @code{do_cleanups} except that it just removes the cleanups from
5218 the chain and does not call the specified functions.
5221 Cleanups are implemented as a chain. The handle returned by
5222 @code{make_cleanups} includes the cleanup passed to the call and any
5223 later cleanups appended to the chain (but not yet discarded or
5227 make_cleanup (a, 0);
5229 struct cleanup *old = make_cleanup (b, 0);
5237 will call @code{c()} and @code{b()} but will not call @code{a()}. The
5238 cleanup that calls @code{a()} will remain in the cleanup chain, and will
5239 be done later unless otherwise discarded.@refill
5241 Your function should explicitly do or discard the cleanups it creates.
5242 Failing to do this leads to non-deterministic behavior since the caller
5243 will arbitrarily do or discard your functions cleanups. This need leads
5244 to two common cleanup styles.
5246 The first style is try/finally. Before it exits, your code-block calls
5247 @code{do_cleanups} with the old cleanup chain and thus ensures that your
5248 code-block's cleanups are always performed. For instance, the following
5249 code-segment avoids a memory leak problem (even when @code{error} is
5250 called and a forced stack unwind occurs) by ensuring that the
5251 @code{xfree} will always be called:
5254 struct cleanup *old = make_cleanup (null_cleanup, 0);
5255 data = xmalloc (sizeof blah);
5256 make_cleanup (xfree, data);
5261 The second style is try/except. Before it exits, your code-block calls
5262 @code{discard_cleanups} with the old cleanup chain and thus ensures that
5263 any created cleanups are not performed. For instance, the following
5264 code segment, ensures that the file will be closed but only if there is
5268 FILE *file = fopen ("afile", "r");
5269 struct cleanup *old = make_cleanup (close_file, file);
5271 discard_cleanups (old);
5275 Some functions, e.g., @code{fputs_filtered()} or @code{error()}, specify
5276 that they ``should not be called when cleanups are not in place''. This
5277 means that any actions you need to reverse in the case of an error or
5278 interruption must be on the cleanup chain before you call these
5279 functions, since they might never return to your code (they
5280 @samp{longjmp} instead).
5282 @section Per-architecture module data
5283 @cindex per-architecture module data
5284 @cindex multi-arch data
5285 @cindex data-pointer, per-architecture/per-module
5287 The multi-arch framework includes a mechanism for adding module
5288 specific per-architecture data-pointers to the @code{struct gdbarch}
5289 architecture object.
5291 A module registers one or more per-architecture data-pointers using:
5293 @deftypefun struct gdbarch_data *gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
5294 @var{pre_init} is used to, on-demand, allocate an initial value for a
5295 per-architecture data-pointer using the architecture's obstack (passed
5296 in as a parameter). Since @var{pre_init} can be called during
5297 architecture creation, it is not parameterized with the architecture.
5298 and must not call modules that use per-architecture data.
5301 @deftypefun struct gdbarch_data *gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
5302 @var{post_init} is used to obtain an initial value for a
5303 per-architecture data-pointer @emph{after}. Since @var{post_init} is
5304 always called after architecture creation, it both receives the fully
5305 initialized architecture and is free to call modules that use
5306 per-architecture data (care needs to be taken to ensure that those
5307 other modules do not try to call back to this module as that will
5308 create in cycles in the initialization call graph).
5311 These functions return a @code{struct gdbarch_data} that is used to
5312 identify the per-architecture data-pointer added for that module.
5314 The per-architecture data-pointer is accessed using the function:
5316 @deftypefun void *gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
5317 Given the architecture @var{arch} and module data handle
5318 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
5319 or @code{gdbarch_data_register_post_init}), this function returns the
5320 current value of the per-architecture data-pointer. If the data
5321 pointer is @code{NULL}, it is first initialized by calling the
5322 corresponding @var{pre_init} or @var{post_init} method.
5325 The examples below assume the following definitions:
5328 struct nozel @{ int total; @};
5329 static struct gdbarch_data *nozel_handle;
5332 A module can extend the architecture vector, adding additional
5333 per-architecture data, using the @var{pre_init} method. The module's
5334 per-architecture data is then initialized during architecture
5337 In the below, the module's per-architecture @emph{nozel} is added. An
5338 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
5339 from @code{gdbarch_init}.
5343 nozel_pre_init (struct obstack *obstack)
5345 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
5352 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
5354 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5355 data->total = nozel;
5359 A module can on-demand create architecture dependant data structures
5360 using @code{post_init}.
5362 In the below, the nozel's total is computed on-demand by
5363 @code{nozel_post_init} using information obtained from the
5368 nozel_post_init (struct gdbarch *gdbarch)
5370 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
5371 nozel->total = gdbarch@dots{} (gdbarch);
5378 nozel_total (struct gdbarch *gdbarch)
5380 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5385 @section Wrapping Output Lines
5386 @cindex line wrap in output
5389 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
5390 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
5391 added in places that would be good breaking points. The utility
5392 routines will take care of actually wrapping if the line width is
5395 The argument to @code{wrap_here} is an indentation string which is
5396 printed @emph{only} if the line breaks there. This argument is saved
5397 away and used later. It must remain valid until the next call to
5398 @code{wrap_here} or until a newline has been printed through the
5399 @code{*_filtered} functions. Don't pass in a local variable and then
5402 It is usually best to call @code{wrap_here} after printing a comma or
5403 space. If you call it before printing a space, make sure that your
5404 indentation properly accounts for the leading space that will print if
5405 the line wraps there.
5407 Any function or set of functions that produce filtered output must
5408 finish by printing a newline, to flush the wrap buffer, before switching
5409 to unfiltered (@code{printf}) output. Symbol reading routines that
5410 print warnings are a good example.
5412 @section @value{GDBN} Coding Standards
5413 @cindex coding standards
5415 @value{GDBN} follows the GNU coding standards, as described in
5416 @file{etc/standards.texi}. This file is also available for anonymous
5417 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
5418 of the standard; in general, when the GNU standard recommends a practice
5419 but does not require it, @value{GDBN} requires it.
5421 @value{GDBN} follows an additional set of coding standards specific to
5422 @value{GDBN}, as described in the following sections.
5427 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
5430 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
5433 @subsection Memory Management
5435 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
5436 @code{calloc}, @code{free} and @code{asprintf}.
5438 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
5439 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
5440 these functions do not return when the memory pool is empty. Instead,
5441 they unwind the stack using cleanups. These functions return
5442 @code{NULL} when requested to allocate a chunk of memory of size zero.
5444 @emph{Pragmatics: By using these functions, the need to check every
5445 memory allocation is removed. These functions provide portable
5448 @value{GDBN} does not use the function @code{free}.
5450 @value{GDBN} uses the function @code{xfree} to return memory to the
5451 memory pool. Consistent with ISO-C, this function ignores a request to
5452 free a @code{NULL} pointer.
5454 @emph{Pragmatics: On some systems @code{free} fails when passed a
5455 @code{NULL} pointer.}
5457 @value{GDBN} can use the non-portable function @code{alloca} for the
5458 allocation of small temporary values (such as strings).
5460 @emph{Pragmatics: This function is very non-portable. Some systems
5461 restrict the memory being allocated to no more than a few kilobytes.}
5463 @value{GDBN} uses the string function @code{xstrdup} and the print
5464 function @code{xstrprintf}.
5466 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
5467 functions such as @code{sprintf} are very prone to buffer overflow
5471 @subsection Compiler Warnings
5472 @cindex compiler warnings
5474 With few exceptions, developers should avoid the configuration option
5475 @samp{--disable-werror} when building @value{GDBN}. The exceptions
5476 are listed in the file @file{gdb/MAINTAINERS}. The default, when
5477 building with @sc{gcc}, is @samp{--enable-werror}.
5479 This option causes @value{GDBN} (when built using GCC) to be compiled
5480 with a carefully selected list of compiler warning flags. Any warnings
5481 from those flags are treated as errors.
5483 The current list of warning flags includes:
5487 Recommended @sc{gcc} warnings.
5489 @item -Wdeclaration-after-statement
5491 @sc{gcc} 3.x (and later) and @sc{c99} allow declarations mixed with
5492 code, but @sc{gcc} 2.x and @sc{c89} do not.
5494 @item -Wpointer-arith
5496 @item -Wformat-nonliteral
5497 Non-literal format strings, with a few exceptions, are bugs - they
5498 might contain unintented user-supplied format specifiers.
5499 Since @value{GDBN} uses the @code{format printf} attribute on all
5500 @code{printf} like functions this checks not just @code{printf} calls
5501 but also calls to functions such as @code{fprintf_unfiltered}.
5503 @item -Wno-pointer-sign
5504 In version 4.0, GCC began warning about pointer argument passing or
5505 assignment even when the source and destination differed only in
5506 signedness. However, most @value{GDBN} code doesn't distinguish
5507 carefully between @code{char} and @code{unsigned char}. In early 2006
5508 the @value{GDBN} developers decided correcting these warnings wasn't
5509 worth the time it would take.
5511 @item -Wno-unused-parameter
5512 Due to the way that @value{GDBN} is implemented many functions have
5513 unused parameters. Consequently this warning is avoided. The macro
5514 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
5515 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
5520 These are warnings which might be useful for @value{GDBN}, but are
5521 currently too noisy to enable with @samp{-Werror}.
5525 @subsection Formatting
5527 @cindex source code formatting
5528 The standard GNU recommendations for formatting must be followed
5531 A function declaration should not have its name in column zero. A
5532 function definition should have its name in column zero.
5536 static void foo (void);
5544 @emph{Pragmatics: This simplifies scripting. Function definitions can
5545 be found using @samp{^function-name}.}
5547 There must be a space between a function or macro name and the opening
5548 parenthesis of its argument list (except for macro definitions, as
5549 required by C). There must not be a space after an open paren/bracket
5550 or before a close paren/bracket.
5552 While additional whitespace is generally helpful for reading, do not use
5553 more than one blank line to separate blocks, and avoid adding whitespace
5554 after the end of a program line (as of 1/99, some 600 lines had
5555 whitespace after the semicolon). Excess whitespace causes difficulties
5556 for @code{diff} and @code{patch} utilities.
5558 Pointers are declared using the traditional K&R C style:
5572 @subsection Comments
5574 @cindex comment formatting
5575 The standard GNU requirements on comments must be followed strictly.
5577 Block comments must appear in the following form, with no @code{/*}- or
5578 @code{*/}-only lines, and no leading @code{*}:
5581 /* Wait for control to return from inferior to debugger. If inferior
5582 gets a signal, we may decide to start it up again instead of
5583 returning. That is why there is a loop in this function. When
5584 this function actually returns it means the inferior should be left
5585 stopped and @value{GDBN} should read more commands. */
5588 (Note that this format is encouraged by Emacs; tabbing for a multi-line
5589 comment works correctly, and @kbd{M-q} fills the block consistently.)
5591 Put a blank line between the block comments preceding function or
5592 variable definitions, and the definition itself.
5594 In general, put function-body comments on lines by themselves, rather
5595 than trying to fit them into the 20 characters left at the end of a
5596 line, since either the comment or the code will inevitably get longer
5597 than will fit, and then somebody will have to move it anyhow.
5601 @cindex C data types
5602 Code must not depend on the sizes of C data types, the format of the
5603 host's floating point numbers, the alignment of anything, or the order
5604 of evaluation of expressions.
5606 @cindex function usage
5607 Use functions freely. There are only a handful of compute-bound areas
5608 in @value{GDBN} that might be affected by the overhead of a function
5609 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
5610 limited by the target interface (whether serial line or system call).
5612 However, use functions with moderation. A thousand one-line functions
5613 are just as hard to understand as a single thousand-line function.
5615 @emph{Macros are bad, M'kay.}
5616 (But if you have to use a macro, make sure that the macro arguments are
5617 protected with parentheses.)
5621 Declarations like @samp{struct foo *} should be used in preference to
5622 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
5625 @subsection Function Prototypes
5626 @cindex function prototypes
5628 Prototypes must be used when both @emph{declaring} and @emph{defining}
5629 a function. Prototypes for @value{GDBN} functions must include both the
5630 argument type and name, with the name matching that used in the actual
5631 function definition.
5633 All external functions should have a declaration in a header file that
5634 callers include, except for @code{_initialize_*} functions, which must
5635 be external so that @file{init.c} construction works, but shouldn't be
5636 visible to random source files.
5638 Where a source file needs a forward declaration of a static function,
5639 that declaration must appear in a block near the top of the source file.
5642 @subsection Internal Error Recovery
5644 During its execution, @value{GDBN} can encounter two types of errors.
5645 User errors and internal errors. User errors include not only a user
5646 entering an incorrect command but also problems arising from corrupt
5647 object files and system errors when interacting with the target.
5648 Internal errors include situations where @value{GDBN} has detected, at
5649 run time, a corrupt or erroneous situation.
5651 When reporting an internal error, @value{GDBN} uses
5652 @code{internal_error} and @code{gdb_assert}.
5654 @value{GDBN} must not call @code{abort} or @code{assert}.
5656 @emph{Pragmatics: There is no @code{internal_warning} function. Either
5657 the code detected a user error, recovered from it and issued a
5658 @code{warning} or the code failed to correctly recover from the user
5659 error and issued an @code{internal_error}.}
5661 @subsection File Names
5663 Any file used when building the core of @value{GDBN} must be in lower
5664 case. Any file used when building the core of @value{GDBN} must be 8.3
5665 unique. These requirements apply to both source and generated files.
5667 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
5668 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
5669 is introduced to the build process both @file{Makefile.in} and
5670 @file{configure.in} need to be modified accordingly. Compare the
5671 convoluted conversion process needed to transform @file{COPYING} into
5672 @file{copying.c} with the conversion needed to transform
5673 @file{version.in} into @file{version.c}.}
5675 Any file non 8.3 compliant file (that is not used when building the core
5676 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
5678 @emph{Pragmatics: This is clearly a compromise.}
5680 When @value{GDBN} has a local version of a system header file (ex
5681 @file{string.h}) the file name based on the POSIX header prefixed with
5682 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
5683 independent: they should use only macros defined by @file{configure},
5684 the compiler, or the host; they should include only system headers; they
5685 should refer only to system types. They may be shared between multiple
5686 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
5688 For other files @samp{-} is used as the separator.
5691 @subsection Include Files
5693 A @file{.c} file should include @file{defs.h} first.
5695 A @file{.c} file should directly include the @code{.h} file of every
5696 declaration and/or definition it directly refers to. It cannot rely on
5699 A @file{.h} file should directly include the @code{.h} file of every
5700 declaration and/or definition it directly refers to. It cannot rely on
5701 indirect inclusion. Exception: The file @file{defs.h} does not need to
5702 be directly included.
5704 An external declaration should only appear in one include file.
5706 An external declaration should never appear in a @code{.c} file.
5707 Exception: a declaration for the @code{_initialize} function that
5708 pacifies @option{-Wmissing-declaration}.
5710 A @code{typedef} definition should only appear in one include file.
5712 An opaque @code{struct} declaration can appear in multiple @file{.h}
5713 files. Where possible, a @file{.h} file should use an opaque
5714 @code{struct} declaration instead of an include.
5716 All @file{.h} files should be wrapped in:
5719 #ifndef INCLUDE_FILE_NAME_H
5720 #define INCLUDE_FILE_NAME_H
5726 @subsection Clean Design and Portable Implementation
5729 In addition to getting the syntax right, there's the little question of
5730 semantics. Some things are done in certain ways in @value{GDBN} because long
5731 experience has shown that the more obvious ways caused various kinds of
5734 @cindex assumptions about targets
5735 You can't assume the byte order of anything that comes from a target
5736 (including @var{value}s, object files, and instructions). Such things
5737 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
5738 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
5739 such as @code{bfd_get_32}.
5741 You can't assume that you know what interface is being used to talk to
5742 the target system. All references to the target must go through the
5743 current @code{target_ops} vector.
5745 You can't assume that the host and target machines are the same machine
5746 (except in the ``native'' support modules). In particular, you can't
5747 assume that the target machine's header files will be available on the
5748 host machine. Target code must bring along its own header files --
5749 written from scratch or explicitly donated by their owner, to avoid
5753 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
5754 to write the code portably than to conditionalize it for various
5757 @cindex system dependencies
5758 New @code{#ifdef}'s which test for specific compilers or manufacturers
5759 or operating systems are unacceptable. All @code{#ifdef}'s should test
5760 for features. The information about which configurations contain which
5761 features should be segregated into the configuration files. Experience
5762 has proven far too often that a feature unique to one particular system
5763 often creeps into other systems; and that a conditional based on some
5764 predefined macro for your current system will become worthless over
5765 time, as new versions of your system come out that behave differently
5766 with regard to this feature.
5768 Adding code that handles specific architectures, operating systems,
5769 target interfaces, or hosts, is not acceptable in generic code.
5771 @cindex portable file name handling
5772 @cindex file names, portability
5773 One particularly notorious area where system dependencies tend to
5774 creep in is handling of file names. The mainline @value{GDBN} code
5775 assumes Posix semantics of file names: absolute file names begin with
5776 a forward slash @file{/}, slashes are used to separate leading
5777 directories, case-sensitive file names. These assumptions are not
5778 necessarily true on non-Posix systems such as MS-Windows. To avoid
5779 system-dependent code where you need to take apart or construct a file
5780 name, use the following portable macros:
5783 @findex HAVE_DOS_BASED_FILE_SYSTEM
5784 @item HAVE_DOS_BASED_FILE_SYSTEM
5785 This preprocessing symbol is defined to a non-zero value on hosts
5786 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
5787 symbol to write conditional code which should only be compiled for
5790 @findex IS_DIR_SEPARATOR
5791 @item IS_DIR_SEPARATOR (@var{c})
5792 Evaluates to a non-zero value if @var{c} is a directory separator
5793 character. On Unix and GNU/Linux systems, only a slash @file{/} is
5794 such a character, but on Windows, both @file{/} and @file{\} will
5797 @findex IS_ABSOLUTE_PATH
5798 @item IS_ABSOLUTE_PATH (@var{file})
5799 Evaluates to a non-zero value if @var{file} is an absolute file name.
5800 For Unix and GNU/Linux hosts, a name which begins with a slash
5801 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
5802 @file{x:\bar} are also absolute file names.
5804 @findex FILENAME_CMP
5805 @item FILENAME_CMP (@var{f1}, @var{f2})
5806 Calls a function which compares file names @var{f1} and @var{f2} as
5807 appropriate for the underlying host filesystem. For Posix systems,
5808 this simply calls @code{strcmp}; on case-insensitive filesystems it
5809 will call @code{strcasecmp} instead.
5811 @findex DIRNAME_SEPARATOR
5812 @item DIRNAME_SEPARATOR
5813 Evaluates to a character which separates directories in
5814 @code{PATH}-style lists, typically held in environment variables.
5815 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
5817 @findex SLASH_STRING
5819 This evaluates to a constant string you should use to produce an
5820 absolute filename from leading directories and the file's basename.
5821 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
5822 @code{"\\"} for some Windows-based ports.
5825 In addition to using these macros, be sure to use portable library
5826 functions whenever possible. For example, to extract a directory or a
5827 basename part from a file name, use the @code{dirname} and
5828 @code{basename} library functions (available in @code{libiberty} for
5829 platforms which don't provide them), instead of searching for a slash
5830 with @code{strrchr}.
5832 Another way to generalize @value{GDBN} along a particular interface is with an
5833 attribute struct. For example, @value{GDBN} has been generalized to handle
5834 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
5835 by defining the @code{target_ops} structure and having a current target (as
5836 well as a stack of targets below it, for memory references). Whenever
5837 something needs to be done that depends on which remote interface we are
5838 using, a flag in the current target_ops structure is tested (e.g.,
5839 @code{target_has_stack}), or a function is called through a pointer in the
5840 current target_ops structure. In this way, when a new remote interface
5841 is added, only one module needs to be touched---the one that actually
5842 implements the new remote interface. Other examples of
5843 attribute-structs are BFD access to multiple kinds of object file
5844 formats, or @value{GDBN}'s access to multiple source languages.
5846 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
5847 the code interfacing between @code{ptrace} and the rest of
5848 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
5849 something was very painful. In @value{GDBN} 4.x, these have all been
5850 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
5851 with variations between systems the same way any system-independent
5852 file would (hooks, @code{#if defined}, etc.), and machines which are
5853 radically different don't need to use @file{infptrace.c} at all.
5855 All debugging code must be controllable using the @samp{set debug
5856 @var{module}} command. Do not use @code{printf} to print trace
5857 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
5858 @code{#ifdef DEBUG}.
5863 @chapter Porting @value{GDBN}
5864 @cindex porting to new machines
5866 Most of the work in making @value{GDBN} compile on a new machine is in
5867 specifying the configuration of the machine. This is done in a
5868 dizzying variety of header files and configuration scripts, which we
5869 hope to make more sensible soon. Let's say your new host is called an
5870 @var{xyz} (e.g., @samp{sun4}), and its full three-part configuration
5871 name is @code{@var{arch}-@var{xvend}-@var{xos}} (e.g.,
5872 @samp{sparc-sun-sunos4}). In particular:
5876 In the top level directory, edit @file{config.sub} and add @var{arch},
5877 @var{xvend}, and @var{xos} to the lists of supported architectures,
5878 vendors, and operating systems near the bottom of the file. Also, add
5879 @var{xyz} as an alias that maps to
5880 @code{@var{arch}-@var{xvend}-@var{xos}}. You can test your changes by
5884 ./config.sub @var{xyz}
5891 ./config.sub @code{@var{arch}-@var{xvend}-@var{xos}}
5895 which should both respond with @code{@var{arch}-@var{xvend}-@var{xos}}
5896 and no error messages.
5899 You need to port BFD, if that hasn't been done already. Porting BFD is
5900 beyond the scope of this manual.
5903 To configure @value{GDBN} itself, edit @file{gdb/configure.host} to recognize
5904 your system and set @code{gdb_host} to @var{xyz}, and (unless your
5905 desired target is already available) also edit @file{gdb/configure.tgt},
5906 setting @code{gdb_target} to something appropriate (for instance,
5909 @emph{Maintainer's note: Work in progress. The file
5910 @file{gdb/configure.host} originally needed to be modified when either a
5911 new native target or a new host machine was being added to @value{GDBN}.
5912 Recent changes have removed this requirement. The file now only needs
5913 to be modified when adding a new native configuration. This will likely
5914 changed again in the future.}
5917 Finally, you'll need to specify and define @value{GDBN}'s host-, native-, and
5918 target-dependent @file{.h} and @file{.c} files used for your
5922 @node Versions and Branches
5923 @chapter Versions and Branches
5927 @value{GDBN}'s version is determined by the file
5928 @file{gdb/version.in} and takes one of the following forms:
5931 @item @var{major}.@var{minor}
5932 @itemx @var{major}.@var{minor}.@var{patchlevel}
5933 an official release (e.g., 6.2 or 6.2.1)
5934 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}
5935 a snapshot taken at @var{YYYY}-@var{MM}-@var{DD}-gmt (e.g.,
5936 6.1.50.20020302, 6.1.90.20020304, or 6.1.0.20020308)
5937 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}-cvs
5938 a @sc{cvs} check out drawn on @var{YYYY}-@var{MM}-@var{DD} (e.g.,
5939 6.1.50.20020302-cvs, 6.1.90.20020304-cvs, or 6.1.0.20020308-cvs)
5940 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD} (@var{vendor})
5941 a vendor specific release of @value{GDBN}, that while based on@*
5942 @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD},
5943 may include additional changes
5946 @value{GDBN}'s mainline uses the @var{major} and @var{minor} version
5947 numbers from the most recent release branch, with a @var{patchlevel}
5948 of 50. At the time each new release branch is created, the mainline's
5949 @var{major} and @var{minor} version numbers are updated.
5951 @value{GDBN}'s release branch is similar. When the branch is cut, the
5952 @var{patchlevel} is changed from 50 to 90. As draft releases are
5953 drawn from the branch, the @var{patchlevel} is incremented. Once the
5954 first release (@var{major}.@var{minor}) has been made, the
5955 @var{patchlevel} is set to 0 and updates have an incremented
5958 For snapshots, and @sc{cvs} check outs, it is also possible to
5959 identify the @sc{cvs} origin:
5962 @item @var{major}.@var{minor}.50.@var{YYYY}@var{MM}@var{DD}
5963 drawn from the @sc{head} of mainline @sc{cvs} (e.g., 6.1.50.20020302)
5964 @item @var{major}.@var{minor}.90.@var{YYYY}@var{MM}@var{DD}
5965 @itemx @var{major}.@var{minor}.91.@var{YYYY}@var{MM}@var{DD} @dots{}
5966 drawn from a release branch prior to the release (e.g.,
5968 @item @var{major}.@var{minor}.0.@var{YYYY}@var{MM}@var{DD}
5969 @itemx @var{major}.@var{minor}.1.@var{YYYY}@var{MM}@var{DD} @dots{}
5970 drawn from a release branch after the release (e.g., 6.2.0.20020308)
5973 If the previous @value{GDBN} version is 6.1 and the current version is
5974 6.2, then, substituting 6 for @var{major} and 1 or 2 for @var{minor},
5975 here's an illustration of a typical sequence:
5982 +--------------------------.
5985 6.2.50.20020303-cvs 6.1.90 (draft #1)
5987 6.2.50.20020304-cvs 6.1.90.20020304-cvs
5989 6.2.50.20020305-cvs 6.1.91 (draft #2)
5991 6.2.50.20020306-cvs 6.1.91.20020306-cvs
5993 6.2.50.20020307-cvs 6.2 (release)
5995 6.2.50.20020308-cvs 6.2.0.20020308-cvs
5997 6.2.50.20020309-cvs 6.2.1 (update)
5999 6.2.50.20020310-cvs <branch closed>
6003 +--------------------------.
6006 6.3.50.20020312-cvs 6.2.90 (draft #1)
6010 @section Release Branches
6011 @cindex Release Branches
6013 @value{GDBN} draws a release series (6.2, 6.2.1, @dots{}) from a
6014 single release branch, and identifies that branch using the @sc{cvs}
6018 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-branchpoint
6019 gdb_@var{major}_@var{minor}-branch
6020 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-release
6023 @emph{Pragmatics: To help identify the date at which a branch or
6024 release is made, both the branchpoint and release tags include the
6025 date that they are cut (@var{YYYY}@var{MM}@var{DD}) in the tag. The
6026 branch tag, denoting the head of the branch, does not need this.}
6028 @section Vendor Branches
6029 @cindex vendor branches
6031 To avoid version conflicts, vendors are expected to modify the file
6032 @file{gdb/version.in} to include a vendor unique alphabetic identifier
6033 (an official @value{GDBN} release never uses alphabetic characters in
6034 its version identifer). E.g., @samp{6.2widgit2}, or @samp{6.2 (Widgit
6037 @section Experimental Branches
6038 @cindex experimental branches
6040 @subsection Guidelines
6042 @value{GDBN} permits the creation of branches, cut from the @sc{cvs}
6043 repository, for experimental development. Branches make it possible
6044 for developers to share preliminary work, and maintainers to examine
6045 significant new developments.
6047 The following are a set of guidelines for creating such branches:
6051 @item a branch has an owner
6052 The owner can set further policy for a branch, but may not change the
6053 ground rules. In particular, they can set a policy for commits (be it
6054 adding more reviewers or deciding who can commit).
6056 @item all commits are posted
6057 All changes committed to a branch shall also be posted to
6058 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} patches
6059 mailing list}. While commentary on such changes are encouraged, people
6060 should remember that the changes only apply to a branch.
6062 @item all commits are covered by an assignment
6063 This ensures that all changes belong to the Free Software Foundation,
6064 and avoids the possibility that the branch may become contaminated.
6066 @item a branch is focused
6067 A focused branch has a single objective or goal, and does not contain
6068 unnecessary or irrelevant changes. Cleanups, where identified, being
6069 be pushed into the mainline as soon as possible.
6071 @item a branch tracks mainline
6072 This keeps the level of divergence under control. It also keeps the
6073 pressure on developers to push cleanups and other stuff into the
6076 @item a branch shall contain the entire @value{GDBN} module
6077 The @value{GDBN} module @code{gdb} should be specified when creating a
6078 branch (branches of individual files should be avoided). @xref{Tags}.
6080 @item a branch shall be branded using @file{version.in}
6081 The file @file{gdb/version.in} shall be modified so that it identifies
6082 the branch @var{owner} and branch @var{name}, e.g.,
6083 @samp{6.2.50.20030303_owner_name} or @samp{6.2 (Owner Name)}.
6090 To simplify the identification of @value{GDBN} branches, the following
6091 branch tagging convention is strongly recommended:
6095 @item @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
6096 @itemx @var{owner}_@var{name}-@var{YYYYMMDD}-branch
6097 The branch point and corresponding branch tag. @var{YYYYMMDD} is the
6098 date that the branch was created. A branch is created using the
6099 sequence: @anchor{experimental branch tags}
6101 cvs rtag @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint gdb
6102 cvs rtag -b -r @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint \
6103 @var{owner}_@var{name}-@var{YYYYMMDD}-branch gdb
6106 @item @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
6107 The tagged point, on the mainline, that was used when merging the branch
6108 on @var{yyyymmdd}. To merge in all changes since the branch was cut,
6109 use a command sequence like:
6111 cvs rtag @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint gdb
6113 -j@var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
6114 -j@var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
6117 Similar sequences can be used to just merge in changes since the last
6123 For further information on @sc{cvs}, see
6124 @uref{http://www.gnu.org/software/cvs/, Concurrent Versions System}.
6126 @node Start of New Year Procedure
6127 @chapter Start of New Year Procedure
6128 @cindex new year procedure
6130 At the start of each new year, the following actions should be performed:
6134 Rotate the ChangeLog file
6136 The current @file{ChangeLog} file should be renamed into
6137 @file{ChangeLog-YYYY} where YYYY is the year that has just passed.
6138 A new @file{ChangeLog} file should be created, and its contents should
6139 contain a reference to the previous ChangeLog. The following should
6140 also be preserved at the end of the new ChangeLog, in order to provide
6141 the appropriate settings when editing this file with Emacs:
6147 version-control: never
6152 Add an entry for the newly created ChangeLog file (@file{ChangeLog-YYYY})
6153 in @file{gdb/config/djgpp/fnchange.lst}.
6156 Update the copyright year in the startup message
6158 Update the copyright year in file @file{top.c}, function
6159 @code{print_gdb_version}.
6164 @chapter Releasing @value{GDBN}
6165 @cindex making a new release of gdb
6167 @section Branch Commit Policy
6169 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
6170 5.1 and 5.2 all used the below:
6174 The @file{gdb/MAINTAINERS} file still holds.
6176 Don't fix something on the branch unless/until it is also fixed in the
6177 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
6178 file is better than committing a hack.
6180 When considering a patch for the branch, suggested criteria include:
6181 Does it fix a build? Does it fix the sequence @kbd{break main; run}
6182 when debugging a static binary?
6184 The further a change is from the core of @value{GDBN}, the less likely
6185 the change will worry anyone (e.g., target specific code).
6187 Only post a proposal to change the core of @value{GDBN} after you've
6188 sent individual bribes to all the people listed in the
6189 @file{MAINTAINERS} file @t{;-)}
6192 @emph{Pragmatics: Provided updates are restricted to non-core
6193 functionality there is little chance that a broken change will be fatal.
6194 This means that changes such as adding a new architectures or (within
6195 reason) support for a new host are considered acceptable.}
6198 @section Obsoleting code
6200 Before anything else, poke the other developers (and around the source
6201 code) to see if there is anything that can be removed from @value{GDBN}
6202 (an old target, an unused file).
6204 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
6205 line. Doing this means that it is easy to identify something that has
6206 been obsoleted when greping through the sources.
6208 The process is done in stages --- this is mainly to ensure that the
6209 wider @value{GDBN} community has a reasonable opportunity to respond.
6210 Remember, everything on the Internet takes a week.
6214 Post the proposal on @email{gdb@@sources.redhat.com, the GDB mailing
6215 list} Creating a bug report to track the task's state, is also highly
6220 Post the proposal on @email{gdb-announce@@sources.redhat.com, the GDB
6221 Announcement mailing list}.
6225 Go through and edit all relevant files and lines so that they are
6226 prefixed with the word @code{OBSOLETE}.
6228 Wait until the next GDB version, containing this obsolete code, has been
6231 Remove the obsolete code.
6235 @emph{Maintainer note: While removing old code is regrettable it is
6236 hopefully better for @value{GDBN}'s long term development. Firstly it
6237 helps the developers by removing code that is either no longer relevant
6238 or simply wrong. Secondly since it removes any history associated with
6239 the file (effectively clearing the slate) the developer has a much freer
6240 hand when it comes to fixing broken files.}
6244 @section Before the Branch
6246 The most important objective at this stage is to find and fix simple
6247 changes that become a pain to track once the branch is created. For
6248 instance, configuration problems that stop @value{GDBN} from even
6249 building. If you can't get the problem fixed, document it in the
6250 @file{gdb/PROBLEMS} file.
6252 @subheading Prompt for @file{gdb/NEWS}
6254 People always forget. Send a post reminding them but also if you know
6255 something interesting happened add it yourself. The @code{schedule}
6256 script will mention this in its e-mail.
6258 @subheading Review @file{gdb/README}
6260 Grab one of the nightly snapshots and then walk through the
6261 @file{gdb/README} looking for anything that can be improved. The
6262 @code{schedule} script will mention this in its e-mail.
6264 @subheading Refresh any imported files.
6266 A number of files are taken from external repositories. They include:
6270 @file{texinfo/texinfo.tex}
6272 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
6275 @file{etc/standards.texi}, @file{etc/make-stds.texi}
6278 @subheading Check the ARI
6280 @uref{http://sources.redhat.com/gdb/ari,,A.R.I.} is an @code{awk} script
6281 (Awk Regression Index ;-) that checks for a number of errors and coding
6282 conventions. The checks include things like using @code{malloc} instead
6283 of @code{xmalloc} and file naming problems. There shouldn't be any
6286 @subsection Review the bug data base
6288 Close anything obviously fixed.
6290 @subsection Check all cross targets build
6292 The targets are listed in @file{gdb/MAINTAINERS}.
6295 @section Cut the Branch
6297 @subheading Create the branch
6302 $ V=`echo $v | sed 's/\./_/g'`
6303 $ D=`date -u +%Y-%m-%d`
6306 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6307 -D $D-gmt gdb_$V-$D-branchpoint insight
6308 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag
6309 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight
6312 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6313 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight
6314 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6315 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight
6323 By using @kbd{-D YYYY-MM-DD-gmt}, the branch is forced to an exact
6326 The trunk is first tagged so that the branch point can easily be found.
6328 Insight, which includes @value{GDBN}, is tagged at the same time.
6330 @file{version.in} gets bumped to avoid version number conflicts.
6332 The reading of @file{.cvsrc} is disabled using @file{-f}.
6335 @subheading Update @file{version.in}
6340 $ V=`echo $v | sed 's/\./_/g'`
6344 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src co \
6345 -r gdb_$V-branch src/gdb/version.in
6346 cvs -f -d :ext:sources.redhat.com:/cvs/src co
6347 -r gdb_5_2-branch src/gdb/version.in
6349 U src/gdb/version.in
6351 $ echo $u.90-0000-00-00-cvs > version.in
6353 5.1.90-0000-00-00-cvs
6354 $ cvs -f commit version.in
6359 @file{0000-00-00} is used as a date to pump prime the version.in update
6362 @file{.90} and the previous branch version are used as fairly arbitrary
6363 initial branch version number.
6367 @subheading Update the web and news pages
6371 @subheading Tweak cron to track the new branch
6373 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
6374 This file needs to be updated so that:
6378 A daily timestamp is added to the file @file{version.in}.
6380 The new branch is included in the snapshot process.
6384 See the file @file{gdbadmin/cron/README} for how to install the updated
6387 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
6388 any changes. That file is copied to both the branch/ and current/
6389 snapshot directories.
6392 @subheading Update the NEWS and README files
6394 The @file{NEWS} file needs to be updated so that on the branch it refers
6395 to @emph{changes in the current release} while on the trunk it also
6396 refers to @emph{changes since the current release}.
6398 The @file{README} file needs to be updated so that it refers to the
6401 @subheading Post the branch info
6403 Send an announcement to the mailing lists:
6407 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6409 @email{gdb@@sources.redhat.com, GDB Discsussion mailing list} and
6410 @email{gdb-testers@@sources.redhat.com, GDB Discsussion mailing list}
6413 @emph{Pragmatics: The branch creation is sent to the announce list to
6414 ensure that people people not subscribed to the higher volume discussion
6417 The announcement should include:
6423 How to check out the branch using CVS.
6425 The date/number of weeks until the release.
6427 The branch commit policy still holds.
6430 @section Stabilize the branch
6432 Something goes here.
6434 @section Create a Release
6436 The process of creating and then making available a release is broken
6437 down into a number of stages. The first part addresses the technical
6438 process of creating a releasable tar ball. The later stages address the
6439 process of releasing that tar ball.
6441 When making a release candidate just the first section is needed.
6443 @subsection Create a release candidate
6445 The objective at this stage is to create a set of tar balls that can be
6446 made available as a formal release (or as a less formal release
6449 @subsubheading Freeze the branch
6451 Send out an e-mail notifying everyone that the branch is frozen to
6452 @email{gdb-patches@@sources.redhat.com}.
6454 @subsubheading Establish a few defaults.
6459 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
6461 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
6465 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
6467 /home/gdbadmin/bin/autoconf
6476 Check the @code{autoconf} version carefully. You want to be using the
6477 version taken from the @file{binutils} snapshot directory, which can be
6478 found at @uref{ftp://sources.redhat.com/pub/binutils/}. It is very
6479 unlikely that a system installed version of @code{autoconf} (e.g.,
6480 @file{/usr/bin/autoconf}) is correct.
6483 @subsubheading Check out the relevant modules:
6486 $ for m in gdb insight
6488 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
6498 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
6499 any confusion between what is written here and what your local
6500 @code{cvs} really does.
6503 @subsubheading Update relevant files.
6509 Major releases get their comments added as part of the mainline. Minor
6510 releases should probably mention any significant bugs that were fixed.
6512 Don't forget to include the @file{ChangeLog} entry.
6515 $ emacs gdb/src/gdb/NEWS
6520 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
6521 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6526 You'll need to update:
6538 $ emacs gdb/src/gdb/README
6543 $ cp gdb/src/gdb/README insight/src/gdb/README
6544 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6547 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
6548 before the initial branch was cut so just a simple substitute is needed
6551 @emph{Maintainer note: Other projects generate @file{README} and
6552 @file{INSTALL} from the core documentation. This might be worth
6555 @item gdb/version.in
6558 $ echo $v > gdb/src/gdb/version.in
6559 $ cat gdb/src/gdb/version.in
6561 $ emacs gdb/src/gdb/version.in
6564 ... Bump to version ...
6566 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
6567 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6572 @subsubheading Do the dirty work
6574 This is identical to the process used to create the daily snapshot.
6577 $ for m in gdb insight
6579 ( cd $m/src && gmake -f src-release $m.tar )
6583 If the top level source directory does not have @file{src-release}
6584 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
6587 $ for m in gdb insight
6589 ( cd $m/src && gmake -f Makefile.in $m.tar )
6593 @subsubheading Check the source files
6595 You're looking for files that have mysteriously disappeared.
6596 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
6597 for the @file{version.in} update @kbd{cronjob}.
6600 $ ( cd gdb/src && cvs -f -q -n update )
6604 @dots{} lots of generated files @dots{}
6609 @dots{} lots of generated files @dots{}
6614 @emph{Don't worry about the @file{gdb.info-??} or
6615 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
6616 was also generated only something strange with CVS means that they
6617 didn't get supressed). Fixing it would be nice though.}
6619 @subsubheading Create compressed versions of the release
6625 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
6626 $ for m in gdb insight
6628 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
6629 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
6639 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
6640 in that mode, @code{gzip} does not know the name of the file and, hence,
6641 can not include it in the compressed file. This is also why the release
6642 process runs @code{tar} and @code{bzip2} as separate passes.
6645 @subsection Sanity check the tar ball
6647 Pick a popular machine (Solaris/PPC?) and try the build on that.
6650 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
6655 $ ./gdb/gdb ./gdb/gdb
6659 Breakpoint 1 at 0x80732bc: file main.c, line 734.
6661 Starting program: /tmp/gdb-5.2/gdb/gdb
6663 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
6664 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
6666 $1 = @{argc = 136426532, argv = 0x821b7f0@}
6670 @subsection Make a release candidate available
6672 If this is a release candidate then the only remaining steps are:
6676 Commit @file{version.in} and @file{ChangeLog}
6678 Tweak @file{version.in} (and @file{ChangeLog} to read
6679 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
6680 process can restart.
6682 Make the release candidate available in
6683 @uref{ftp://sources.redhat.com/pub/gdb/snapshots/branch}
6685 Notify the relevant mailing lists ( @email{gdb@@sources.redhat.com} and
6686 @email{gdb-testers@@sources.redhat.com} that the candidate is available.
6689 @subsection Make a formal release available
6691 (And you thought all that was required was to post an e-mail.)
6693 @subsubheading Install on sware
6695 Copy the new files to both the release and the old release directory:
6698 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
6699 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
6703 Clean up the releases directory so that only the most recent releases
6704 are available (e.g. keep 5.2 and 5.2.1 but remove 5.1):
6707 $ cd ~ftp/pub/gdb/releases
6712 Update the file @file{README} and @file{.message} in the releases
6719 $ ln README .message
6722 @subsubheading Update the web pages.
6726 @item htdocs/download/ANNOUNCEMENT
6727 This file, which is posted as the official announcement, includes:
6730 General announcement.
6732 News. If making an @var{M}.@var{N}.1 release, retain the news from
6733 earlier @var{M}.@var{N} release.
6738 @item htdocs/index.html
6739 @itemx htdocs/news/index.html
6740 @itemx htdocs/download/index.html
6741 These files include:
6744 Announcement of the most recent release.
6746 News entry (remember to update both the top level and the news directory).
6748 These pages also need to be regenerate using @code{index.sh}.
6750 @item download/onlinedocs/
6751 You need to find the magic command that is used to generate the online
6752 docs from the @file{.tar.bz2}. The best way is to look in the output
6753 from one of the nightly @code{cron} jobs and then just edit accordingly.
6757 $ ~/ss/update-web-docs \
6758 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6760 /www/sourceware/htdocs/gdb/download/onlinedocs \
6765 Just like the online documentation. Something like:
6768 $ /bin/sh ~/ss/update-web-ari \
6769 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6771 /www/sourceware/htdocs/gdb/download/ari \
6777 @subsubheading Shadow the pages onto gnu
6779 Something goes here.
6782 @subsubheading Install the @value{GDBN} tar ball on GNU
6784 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
6785 @file{~ftp/gnu/gdb}.
6787 @subsubheading Make the @file{ANNOUNCEMENT}
6789 Post the @file{ANNOUNCEMENT} file you created above to:
6793 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6795 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
6796 day or so to let things get out)
6798 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
6803 The release is out but you're still not finished.
6805 @subsubheading Commit outstanding changes
6807 In particular you'll need to commit any changes to:
6811 @file{gdb/ChangeLog}
6813 @file{gdb/version.in}
6820 @subsubheading Tag the release
6825 $ d=`date -u +%Y-%m-%d`
6828 $ ( cd insight/src/gdb && cvs -f -q update )
6829 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
6832 Insight is used since that contains more of the release than
6835 @subsubheading Mention the release on the trunk
6837 Just put something in the @file{ChangeLog} so that the trunk also
6838 indicates when the release was made.
6840 @subsubheading Restart @file{gdb/version.in}
6842 If @file{gdb/version.in} does not contain an ISO date such as
6843 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
6844 committed all the release changes it can be set to
6845 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
6846 is important - it affects the snapshot process).
6848 Don't forget the @file{ChangeLog}.
6850 @subsubheading Merge into trunk
6852 The files committed to the branch may also need changes merged into the
6855 @subsubheading Revise the release schedule
6857 Post a revised release schedule to @email{gdb@@sources.redhat.com, GDB
6858 Discussion List} with an updated announcement. The schedule can be
6859 generated by running:
6862 $ ~/ss/schedule `date +%s` schedule
6866 The first parameter is approximate date/time in seconds (from the epoch)
6867 of the most recent release.
6869 Also update the schedule @code{cronjob}.
6871 @section Post release
6873 Remove any @code{OBSOLETE} code.
6880 The testsuite is an important component of the @value{GDBN} package.
6881 While it is always worthwhile to encourage user testing, in practice
6882 this is rarely sufficient; users typically use only a small subset of
6883 the available commands, and it has proven all too common for a change
6884 to cause a significant regression that went unnoticed for some time.
6886 The @value{GDBN} testsuite uses the DejaGNU testing framework. The
6887 tests themselves are calls to various @code{Tcl} procs; the framework
6888 runs all the procs and summarizes the passes and fails.
6890 @section Using the Testsuite
6892 @cindex running the test suite
6893 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
6894 testsuite's objdir) and type @code{make check}. This just sets up some
6895 environment variables and invokes DejaGNU's @code{runtest} script. While
6896 the testsuite is running, you'll get mentions of which test file is in use,
6897 and a mention of any unexpected passes or fails. When the testsuite is
6898 finished, you'll get a summary that looks like this:
6903 # of expected passes 6016
6904 # of unexpected failures 58
6905 # of unexpected successes 5
6906 # of expected failures 183
6907 # of unresolved testcases 3
6908 # of untested testcases 5
6911 To run a specific test script, type:
6913 make check RUNTESTFLAGS='@var{tests}'
6915 where @var{tests} is a list of test script file names, separated by
6918 The ideal test run consists of expected passes only; however, reality
6919 conspires to keep us from this ideal. Unexpected failures indicate
6920 real problems, whether in @value{GDBN} or in the testsuite. Expected
6921 failures are still failures, but ones which have been decided are too
6922 hard to deal with at the time; for instance, a test case might work
6923 everywhere except on AIX, and there is no prospect of the AIX case
6924 being fixed in the near future. Expected failures should not be added
6925 lightly, since you may be masking serious bugs in @value{GDBN}.
6926 Unexpected successes are expected fails that are passing for some
6927 reason, while unresolved and untested cases often indicate some minor
6928 catastrophe, such as the compiler being unable to deal with a test
6931 When making any significant change to @value{GDBN}, you should run the
6932 testsuite before and after the change, to confirm that there are no
6933 regressions. Note that truly complete testing would require that you
6934 run the testsuite with all supported configurations and a variety of
6935 compilers; however this is more than really necessary. In many cases
6936 testing with a single configuration is sufficient. Other useful
6937 options are to test one big-endian (Sparc) and one little-endian (x86)
6938 host, a cross config with a builtin simulator (powerpc-eabi,
6939 mips-elf), or a 64-bit host (Alpha).
6941 If you add new functionality to @value{GDBN}, please consider adding
6942 tests for it as well; this way future @value{GDBN} hackers can detect
6943 and fix their changes that break the functionality you added.
6944 Similarly, if you fix a bug that was not previously reported as a test
6945 failure, please add a test case for it. Some cases are extremely
6946 difficult to test, such as code that handles host OS failures or bugs
6947 in particular versions of compilers, and it's OK not to try to write
6948 tests for all of those.
6950 DejaGNU supports separate build, host, and target machines. However,
6951 some @value{GDBN} test scripts do not work if the build machine and
6952 the host machine are not the same. In such an environment, these scripts
6953 will give a result of ``UNRESOLVED'', like this:
6956 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
6959 @section Testsuite Organization
6961 @cindex test suite organization
6962 The testsuite is entirely contained in @file{gdb/testsuite}. While the
6963 testsuite includes some makefiles and configury, these are very minimal,
6964 and used for little besides cleaning up, since the tests themselves
6965 handle the compilation of the programs that @value{GDBN} will run. The file
6966 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
6967 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
6968 configuration-specific files, typically used for special-purpose
6969 definitions of procs like @code{gdb_load} and @code{gdb_start}.
6971 The tests themselves are to be found in @file{testsuite/gdb.*} and
6972 subdirectories of those. The names of the test files must always end
6973 with @file{.exp}. DejaGNU collects the test files by wildcarding
6974 in the test directories, so both subdirectories and individual files
6975 get chosen and run in alphabetical order.
6977 The following table lists the main types of subdirectories and what they
6978 are for. Since DejaGNU finds test files no matter where they are
6979 located, and since each test file sets up its own compilation and
6980 execution environment, this organization is simply for convenience and
6985 This is the base testsuite. The tests in it should apply to all
6986 configurations of @value{GDBN} (but generic native-only tests may live here).
6987 The test programs should be in the subset of C that is valid K&R,
6988 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
6991 @item gdb.@var{lang}
6992 Language-specific tests for any language @var{lang} besides C. Examples are
6993 @file{gdb.cp} and @file{gdb.java}.
6995 @item gdb.@var{platform}
6996 Non-portable tests. The tests are specific to a specific configuration
6997 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
7000 @item gdb.@var{compiler}
7001 Tests specific to a particular compiler. As of this writing (June
7002 1999), there aren't currently any groups of tests in this category that
7003 couldn't just as sensibly be made platform-specific, but one could
7004 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
7007 @item gdb.@var{subsystem}
7008 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
7009 instance, @file{gdb.disasm} exercises various disassemblers, while
7010 @file{gdb.stabs} tests pathways through the stabs symbol reader.
7013 @section Writing Tests
7014 @cindex writing tests
7016 In many areas, the @value{GDBN} tests are already quite comprehensive; you
7017 should be able to copy existing tests to handle new cases.
7019 You should try to use @code{gdb_test} whenever possible, since it
7020 includes cases to handle all the unexpected errors that might happen.
7021 However, it doesn't cost anything to add new test procedures; for
7022 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
7023 calls @code{gdb_test} multiple times.
7025 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
7026 necessary, such as when @value{GDBN} has several valid responses to a command.
7028 The source language programs do @emph{not} need to be in a consistent
7029 style. Since @value{GDBN} is used to debug programs written in many different
7030 styles, it's worth having a mix of styles in the testsuite; for
7031 instance, some @value{GDBN} bugs involving the display of source lines would
7032 never manifest themselves if the programs used GNU coding style
7039 Check the @file{README} file, it often has useful information that does not
7040 appear anywhere else in the directory.
7043 * Getting Started:: Getting started working on @value{GDBN}
7044 * Debugging GDB:: Debugging @value{GDBN} with itself
7047 @node Getting Started,,, Hints
7049 @section Getting Started
7051 @value{GDBN} is a large and complicated program, and if you first starting to
7052 work on it, it can be hard to know where to start. Fortunately, if you
7053 know how to go about it, there are ways to figure out what is going on.
7055 This manual, the @value{GDBN} Internals manual, has information which applies
7056 generally to many parts of @value{GDBN}.
7058 Information about particular functions or data structures are located in
7059 comments with those functions or data structures. If you run across a
7060 function or a global variable which does not have a comment correctly
7061 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
7062 free to submit a bug report, with a suggested comment if you can figure
7063 out what the comment should say. If you find a comment which is
7064 actually wrong, be especially sure to report that.
7066 Comments explaining the function of macros defined in host, target, or
7067 native dependent files can be in several places. Sometimes they are
7068 repeated every place the macro is defined. Sometimes they are where the
7069 macro is used. Sometimes there is a header file which supplies a
7070 default definition of the macro, and the comment is there. This manual
7071 also documents all the available macros.
7072 @c (@pxref{Host Conditionals}, @pxref{Target
7073 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
7076 Start with the header files. Once you have some idea of how
7077 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
7078 @file{gdbtypes.h}), you will find it much easier to understand the
7079 code which uses and creates those symbol tables.
7081 You may wish to process the information you are getting somehow, to
7082 enhance your understanding of it. Summarize it, translate it to another
7083 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
7084 the code to predict what a test case would do and write the test case
7085 and verify your prediction, etc. If you are reading code and your eyes
7086 are starting to glaze over, this is a sign you need to use a more active
7089 Once you have a part of @value{GDBN} to start with, you can find more
7090 specifically the part you are looking for by stepping through each
7091 function with the @code{next} command. Do not use @code{step} or you
7092 will quickly get distracted; when the function you are stepping through
7093 calls another function try only to get a big-picture understanding
7094 (perhaps using the comment at the beginning of the function being
7095 called) of what it does. This way you can identify which of the
7096 functions being called by the function you are stepping through is the
7097 one which you are interested in. You may need to examine the data
7098 structures generated at each stage, with reference to the comments in
7099 the header files explaining what the data structures are supposed to
7102 Of course, this same technique can be used if you are just reading the
7103 code, rather than actually stepping through it. The same general
7104 principle applies---when the code you are looking at calls something
7105 else, just try to understand generally what the code being called does,
7106 rather than worrying about all its details.
7108 @cindex command implementation
7109 A good place to start when tracking down some particular area is with
7110 a command which invokes that feature. Suppose you want to know how
7111 single-stepping works. As a @value{GDBN} user, you know that the
7112 @code{step} command invokes single-stepping. The command is invoked
7113 via command tables (see @file{command.h}); by convention the function
7114 which actually performs the command is formed by taking the name of
7115 the command and adding @samp{_command}, or in the case of an
7116 @code{info} subcommand, @samp{_info}. For example, the @code{step}
7117 command invokes the @code{step_command} function and the @code{info
7118 display} command invokes @code{display_info}. When this convention is
7119 not followed, you might have to use @code{grep} or @kbd{M-x
7120 tags-search} in emacs, or run @value{GDBN} on itself and set a
7121 breakpoint in @code{execute_command}.
7123 @cindex @code{bug-gdb} mailing list
7124 If all of the above fail, it may be appropriate to ask for information
7125 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
7126 wondering if anyone could give me some tips about understanding
7127 @value{GDBN}''---if we had some magic secret we would put it in this manual.
7128 Suggestions for improving the manual are always welcome, of course.
7130 @node Debugging GDB,,,Hints
7132 @section Debugging @value{GDBN} with itself
7133 @cindex debugging @value{GDBN}
7135 If @value{GDBN} is limping on your machine, this is the preferred way to get it
7136 fully functional. Be warned that in some ancient Unix systems, like
7137 Ultrix 4.2, a program can't be running in one process while it is being
7138 debugged in another. Rather than typing the command @kbd{@w{./gdb
7139 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
7140 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
7142 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
7143 @file{.gdbinit} file that sets up some simple things to make debugging
7144 gdb easier. The @code{info} command, when executed without a subcommand
7145 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
7146 gdb. See @file{.gdbinit} for details.
7148 If you use emacs, you will probably want to do a @code{make TAGS} after
7149 you configure your distribution; this will put the machine dependent
7150 routines for your local machine where they will be accessed first by
7153 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
7154 have run @code{fixincludes} if you are compiling with gcc.
7156 @section Submitting Patches
7158 @cindex submitting patches
7159 Thanks for thinking of offering your changes back to the community of
7160 @value{GDBN} users. In general we like to get well designed enhancements.
7161 Thanks also for checking in advance about the best way to transfer the
7164 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
7165 This manual summarizes what we believe to be clean design for @value{GDBN}.
7167 If the maintainers don't have time to put the patch in when it arrives,
7168 or if there is any question about a patch, it goes into a large queue
7169 with everyone else's patches and bug reports.
7171 @cindex legal papers for code contributions
7172 The legal issue is that to incorporate substantial changes requires a
7173 copyright assignment from you and/or your employer, granting ownership
7174 of the changes to the Free Software Foundation. You can get the
7175 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
7176 and asking for it. We recommend that people write in "All programs
7177 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
7178 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
7180 contributed with only one piece of legalese pushed through the
7181 bureaucracy and filed with the FSF. We can't start merging changes until
7182 this paperwork is received by the FSF (their rules, which we follow
7183 since we maintain it for them).
7185 Technically, the easiest way to receive changes is to receive each
7186 feature as a small context diff or unidiff, suitable for @code{patch}.
7187 Each message sent to me should include the changes to C code and
7188 header files for a single feature, plus @file{ChangeLog} entries for
7189 each directory where files were modified, and diffs for any changes
7190 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
7191 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
7192 single feature, they can be split down into multiple messages.
7194 In this way, if we read and like the feature, we can add it to the
7195 sources with a single patch command, do some testing, and check it in.
7196 If you leave out the @file{ChangeLog}, we have to write one. If you leave
7197 out the doc, we have to puzzle out what needs documenting. Etc., etc.
7199 The reason to send each change in a separate message is that we will not
7200 install some of the changes. They'll be returned to you with questions
7201 or comments. If we're doing our job correctly, the message back to you
7202 will say what you have to fix in order to make the change acceptable.
7203 The reason to have separate messages for separate features is so that
7204 the acceptable changes can be installed while one or more changes are
7205 being reworked. If multiple features are sent in a single message, we
7206 tend to not put in the effort to sort out the acceptable changes from
7207 the unacceptable, so none of the features get installed until all are
7210 If this sounds painful or authoritarian, well, it is. But we get a lot
7211 of bug reports and a lot of patches, and many of them don't get
7212 installed because we don't have the time to finish the job that the bug
7213 reporter or the contributor could have done. Patches that arrive
7214 complete, working, and well designed, tend to get installed on the day
7215 they arrive. The others go into a queue and get installed as time
7216 permits, which, since the maintainers have many demands to meet, may not
7217 be for quite some time.
7219 Please send patches directly to
7220 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} maintainers}.
7222 @section Obsolete Conditionals
7223 @cindex obsolete code
7225 Fragments of old code in @value{GDBN} sometimes reference or set the following
7226 configuration macros. They should not be used by new code, and old uses
7227 should be removed as those parts of the debugger are otherwise touched.
7230 @item STACK_END_ADDR
7231 This macro used to define where the end of the stack appeared, for use
7232 in interpreting core file formats that don't record this address in the
7233 core file itself. This information is now configured in BFD, and @value{GDBN}
7234 gets the info portably from there. The values in @value{GDBN}'s configuration
7235 files should be moved into BFD configuration files (if needed there),
7236 and deleted from all of @value{GDBN}'s config files.
7238 Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR
7239 is so old that it has never been converted to use BFD. Now that's old!
7243 @include observer.texi