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, 2008
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.
83 * Target Architecture Definition::
84 * Target Descriptions::
85 * Target Vector Definition::
90 * Versions and Branches::
91 * Start of New Year Procedure::
96 * GDB Observers:: @value{GDBN} Currently available observers
97 * GNU Free Documentation License:: The license for this documentation
103 @chapter Requirements
104 @cindex requirements for @value{GDBN}
106 Before diving into the internals, you should understand the formal
107 requirements and other expectations for @value{GDBN}. Although some
108 of these may seem obvious, there have been proposals for @value{GDBN}
109 that have run counter to these requirements.
111 First of all, @value{GDBN} is a debugger. It's not designed to be a
112 front panel for embedded systems. It's not a text editor. It's not a
113 shell. It's not a programming environment.
115 @value{GDBN} is an interactive tool. Although a batch mode is
116 available, @value{GDBN}'s primary role is to interact with a human
119 @value{GDBN} should be responsive to the user. A programmer hot on
120 the trail of a nasty bug, and operating under a looming deadline, is
121 going to be very impatient of everything, including the response time
122 to debugger commands.
124 @value{GDBN} should be relatively permissive, such as for expressions.
125 While the compiler should be picky (or have the option to be made
126 picky), since source code lives for a long time usually, the
127 programmer doing debugging shouldn't be spending time figuring out to
128 mollify the debugger.
130 @value{GDBN} will be called upon to deal with really large programs.
131 Executable sizes of 50 to 100 megabytes occur regularly, and we've
132 heard reports of programs approaching 1 gigabyte in size.
134 @value{GDBN} should be able to run everywhere. No other debugger is
135 available for even half as many configurations as @value{GDBN}
139 @node Overall Structure
141 @chapter Overall Structure
143 @value{GDBN} consists of three major subsystems: user interface,
144 symbol handling (the @dfn{symbol side}), and target system handling (the
147 The user interface consists of several actual interfaces, plus
150 The symbol side consists of object file readers, debugging info
151 interpreters, symbol table management, source language expression
152 parsing, type and value printing.
154 The target side consists of execution control, stack frame analysis, and
155 physical target manipulation.
157 The target side/symbol side division is not formal, and there are a
158 number of exceptions. For instance, core file support involves symbolic
159 elements (the basic core file reader is in BFD) and target elements (it
160 supplies the contents of memory and the values of registers). Instead,
161 this division is useful for understanding how the minor subsystems
164 @section The Symbol Side
166 The symbolic side of @value{GDBN} can be thought of as ``everything
167 you can do in @value{GDBN} without having a live program running''.
168 For instance, you can look at the types of variables, and evaluate
169 many kinds of expressions.
171 @section The Target Side
173 The target side of @value{GDBN} is the ``bits and bytes manipulator''.
174 Although it may make reference to symbolic info here and there, most
175 of the target side will run with only a stripped executable
176 available---or even no executable at all, in remote debugging cases.
178 Operations such as disassembly, stack frame crawls, and register
179 display, are able to work with no symbolic info at all. In some cases,
180 such as disassembly, @value{GDBN} will use symbolic info to present addresses
181 relative to symbols rather than as raw numbers, but it will work either
184 @section Configurations
188 @dfn{Host} refers to attributes of the system where @value{GDBN} runs.
189 @dfn{Target} refers to the system where the program being debugged
190 executes. In most cases they are the same machine, in which case a
191 third type of @dfn{Native} attributes come into play.
193 Defines and include files needed to build on the host are host support.
194 Examples are tty support, system defined types, host byte order, host
197 Defines and information needed to handle the target format are target
198 dependent. Examples are the stack frame format, instruction set,
199 breakpoint instruction, registers, and how to set up and tear down the stack
202 Information that is only needed when the host and target are the same,
203 is native dependent. One example is Unix child process support; if the
204 host and target are not the same, doing a fork to start the target
205 process is a bad idea. The various macros needed for finding the
206 registers in the @code{upage}, running @code{ptrace}, and such are all
207 in the native-dependent files.
209 Another example of native-dependent code is support for features that
210 are really part of the target environment, but which require
211 @code{#include} files that are only available on the host system. Core
212 file handling and @code{setjmp} handling are two common cases.
214 When you want to make @value{GDBN} work ``native'' on a particular machine, you
215 have to include all three kinds of information.
217 @section Source Tree Structure
218 @cindex @value{GDBN} source tree structure
220 The @value{GDBN} source directory has a mostly flat structure---there
221 are only a few subdirectories. A file's name usually gives a hint as
222 to what it does; for example, @file{stabsread.c} reads stabs,
223 @file{dwarf2read.c} reads @sc{DWARF 2}, etc.
225 Files that are related to some common task have names that share
226 common substrings. For example, @file{*-thread.c} files deal with
227 debugging threads on various platforms; @file{*read.c} files deal with
228 reading various kinds of symbol and object files; @file{inf*.c} files
229 deal with direct control of the @dfn{inferior program} (@value{GDBN}
230 parlance for the program being debugged).
232 There are several dozens of files in the @file{*-tdep.c} family.
233 @samp{tdep} stands for @dfn{target-dependent code}---each of these
234 files implements debug support for a specific target architecture
235 (sparc, mips, etc). Usually, only one of these will be used in a
236 specific @value{GDBN} configuration (sometimes two, closely related).
238 Similarly, there are many @file{*-nat.c} files, each one for native
239 debugging on a specific system (e.g., @file{sparc-linux-nat.c} is for
240 native debugging of Sparc machines running the Linux kernel).
242 The few subdirectories of the source tree are:
246 Code that implements @dfn{CLI}, the @value{GDBN} Command-Line
247 Interpreter. @xref{User Interface, Command Interpreter}.
250 Code for the @value{GDBN} remote server.
253 Code for Insight, the @value{GDBN} TK-based GUI front-end.
256 The @dfn{GDB/MI}, the @value{GDBN} Machine Interface interpreter.
259 Target signal translation code.
262 Code for @dfn{TUI}, the @value{GDBN} Text-mode full-screen User
263 Interface. @xref{User Interface, TUI}.
271 @value{GDBN} uses a number of debugging-specific algorithms. They are
272 often not very complicated, but get lost in the thicket of special
273 cases and real-world issues. This chapter describes the basic
274 algorithms and mentions some of the specific target definitions that
277 @section Prologue Analysis
279 @cindex prologue analysis
280 @cindex call frame information
281 @cindex CFI (call frame information)
282 To produce a backtrace and allow the user to manipulate older frames'
283 variables and arguments, @value{GDBN} needs to find the base addresses
284 of older frames, and discover where those frames' registers have been
285 saved. Since a frame's ``callee-saves'' registers get saved by
286 younger frames if and when they're reused, a frame's registers may be
287 scattered unpredictably across younger frames. This means that
288 changing the value of a register-allocated variable in an older frame
289 may actually entail writing to a save slot in some younger frame.
291 Modern versions of GCC emit Dwarf call frame information (``CFI''),
292 which describes how to find frame base addresses and saved registers.
293 But CFI is not always available, so as a fallback @value{GDBN} uses a
294 technique called @dfn{prologue analysis} to find frame sizes and saved
295 registers. A prologue analyzer disassembles the function's machine
296 code starting from its entry point, and looks for instructions that
297 allocate frame space, save the stack pointer in a frame pointer
298 register, save registers, and so on. Obviously, this can't be done
299 accurately in general, but it's tractable to do well enough to be very
300 helpful. Prologue analysis predates the GNU toolchain's support for
301 CFI; at one time, prologue analysis was the only mechanism
302 @value{GDBN} used for stack unwinding at all, when the function
303 calling conventions didn't specify a fixed frame layout.
305 In the olden days, function prologues were generated by hand-written,
306 target-specific code in GCC, and treated as opaque and untouchable by
307 optimizers. Looking at this code, it was usually straightforward to
308 write a prologue analyzer for @value{GDBN} that would accurately
309 understand all the prologues GCC would generate. However, over time
310 GCC became more aggressive about instruction scheduling, and began to
311 understand more about the semantics of the prologue instructions
312 themselves; in response, @value{GDBN}'s analyzers became more complex
313 and fragile. Keeping the prologue analyzers working as GCC (and the
314 instruction sets themselves) evolved became a substantial task.
316 @cindex @file{prologue-value.c}
317 @cindex abstract interpretation of function prologues
318 @cindex pseudo-evaluation of function prologues
319 To try to address this problem, the code in @file{prologue-value.h}
320 and @file{prologue-value.c} provides a general framework for writing
321 prologue analyzers that are simpler and more robust than ad-hoc
322 analyzers. When we analyze a prologue using the prologue-value
323 framework, we're really doing ``abstract interpretation'' or
324 ``pseudo-evaluation'': running the function's code in simulation, but
325 using conservative approximations of the values registers and memory
326 would hold when the code actually runs. For example, if our function
327 starts with the instruction:
330 addi r1, 42 # add 42 to r1
333 we don't know exactly what value will be in @code{r1} after executing
334 this instruction, but we do know it'll be 42 greater than its original
337 If we then see an instruction like:
340 addi r1, 22 # add 22 to r1
343 we still don't know what @code{r1's} value is, but again, we can say
344 it is now 64 greater than its original value.
346 If the next instruction were:
349 mov r2, r1 # set r2 to r1's value
352 then we can say that @code{r2's} value is now the original value of
355 It's common for prologues to save registers on the stack, so we'll
356 need to track the values of stack frame slots, as well as the
357 registers. So after an instruction like this:
363 then we'd know that the stack slot four bytes above the frame pointer
364 holds the original value of @code{r1} plus 64.
368 Of course, this can only go so far before it gets unreasonable. If we
369 wanted to be able to say anything about the value of @code{r1} after
373 xor r1, r3 # exclusive-or r1 and r3, place result in r1
376 then things would get pretty complex. But remember, we're just doing
377 a conservative approximation; if exclusive-or instructions aren't
378 relevant to prologues, we can just say @code{r1}'s value is now
379 ``unknown''. We can ignore things that are too complex, if that loss of
380 information is acceptable for our application.
382 So when we say ``conservative approximation'' here, what we mean is an
383 approximation that is either accurate, or marked ``unknown'', but
386 Using this framework, a prologue analyzer is simply an interpreter for
387 machine code, but one that uses conservative approximations for the
388 contents of registers and memory instead of actual values. Starting
389 from the function's entry point, you simulate instructions up to the
390 current PC, or an instruction that you don't know how to simulate.
391 Now you can examine the state of the registers and stack slots you've
397 To see how large your stack frame is, just check the value of the
398 stack pointer register; if it's the original value of the SP
399 minus a constant, then that constant is the stack frame's size.
400 If the SP's value has been marked as ``unknown'', then that means
401 the prologue has done something too complex for us to track, and
402 we don't know the frame size.
405 To see where we've saved the previous frame's registers, we just
406 search the values we've tracked --- stack slots, usually, but
407 registers, too, if you want --- for something equal to the register's
408 original value. If the calling conventions suggest a standard place
409 to save a given register, then we can check there first, but really,
410 anything that will get us back the original value will probably work.
413 This does take some work. But prologue analyzers aren't
414 quick-and-simple pattern patching to recognize a few fixed prologue
415 forms any more; they're big, hairy functions. Along with inferior
416 function calls, prologue analysis accounts for a substantial portion
417 of the time needed to stabilize a @value{GDBN} port. So it's
418 worthwhile to look for an approach that will be easier to understand
419 and maintain. In the approach described above:
424 It's easier to see that the analyzer is correct: you just see
425 whether the analyzer properly (albeit conservatively) simulates
426 the effect of each instruction.
429 It's easier to extend the analyzer: you can add support for new
430 instructions, and know that you haven't broken anything that
431 wasn't already broken before.
434 It's orthogonal: to gather new information, you don't need to
435 complicate the code for each instruction. As long as your domain
436 of conservative values is already detailed enough to tell you
437 what you need, then all the existing instruction simulations are
438 already gathering the right data for you.
442 The file @file{prologue-value.h} contains detailed comments explaining
443 the framework and how to use it.
446 @section Breakpoint Handling
449 In general, a breakpoint is a user-designated location in the program
450 where the user wants to regain control if program execution ever reaches
453 There are two main ways to implement breakpoints; either as ``hardware''
454 breakpoints or as ``software'' breakpoints.
456 @cindex hardware breakpoints
457 @cindex program counter
458 Hardware breakpoints are sometimes available as a builtin debugging
459 features with some chips. Typically these work by having dedicated
460 register into which the breakpoint address may be stored. If the PC
461 (shorthand for @dfn{program counter})
462 ever matches a value in a breakpoint registers, the CPU raises an
463 exception and reports it to @value{GDBN}.
465 Another possibility is when an emulator is in use; many emulators
466 include circuitry that watches the address lines coming out from the
467 processor, and force it to stop if the address matches a breakpoint's
470 A third possibility is that the target already has the ability to do
471 breakpoints somehow; for instance, a ROM monitor may do its own
472 software breakpoints. So although these are not literally ``hardware
473 breakpoints'', from @value{GDBN}'s point of view they work the same;
474 @value{GDBN} need not do anything more than set the breakpoint and wait
475 for something to happen.
477 Since they depend on hardware resources, hardware breakpoints may be
478 limited in number; when the user asks for more, @value{GDBN} will
479 start trying to set software breakpoints. (On some architectures,
480 notably the 32-bit x86 platforms, @value{GDBN} cannot always know
481 whether there's enough hardware resources to insert all the hardware
482 breakpoints and watchpoints. On those platforms, @value{GDBN} prints
483 an error message only when the program being debugged is continued.)
485 @cindex software breakpoints
486 Software breakpoints require @value{GDBN} to do somewhat more work.
487 The basic theory is that @value{GDBN} will replace a program
488 instruction with a trap, illegal divide, or some other instruction
489 that will cause an exception, and then when it's encountered,
490 @value{GDBN} will take the exception and stop the program. When the
491 user says to continue, @value{GDBN} will restore the original
492 instruction, single-step, re-insert the trap, and continue on.
494 Since it literally overwrites the program being tested, the program area
495 must be writable, so this technique won't work on programs in ROM. It
496 can also distort the behavior of programs that examine themselves,
497 although such a situation would be highly unusual.
499 Also, the software breakpoint instruction should be the smallest size of
500 instruction, so it doesn't overwrite an instruction that might be a jump
501 target, and cause disaster when the program jumps into the middle of the
502 breakpoint instruction. (Strictly speaking, the breakpoint must be no
503 larger than the smallest interval between instructions that may be jump
504 targets; perhaps there is an architecture where only even-numbered
505 instructions may jumped to.) Note that it's possible for an instruction
506 set not to have any instructions usable for a software breakpoint,
507 although in practice only the ARC has failed to define such an
511 The basic definition of the software breakpoint is the macro
514 Basic breakpoint object handling is in @file{breakpoint.c}. However,
515 much of the interesting breakpoint action is in @file{infrun.c}.
518 @cindex insert or remove software breakpoint
519 @findex target_remove_breakpoint
520 @findex target_insert_breakpoint
521 @item target_remove_breakpoint (@var{bp_tgt})
522 @itemx target_insert_breakpoint (@var{bp_tgt})
523 Insert or remove a software breakpoint at address
524 @code{@var{bp_tgt}->placed_address}. Returns zero for success,
525 non-zero for failure. On input, @var{bp_tgt} contains the address of the
526 breakpoint, and is otherwise initialized to zero. The fields of the
527 @code{struct bp_target_info} pointed to by @var{bp_tgt} are updated
528 to contain other information about the breakpoint on output. The field
529 @code{placed_address} may be updated if the breakpoint was placed at a
530 related address; the field @code{shadow_contents} contains the real
531 contents of the bytes where the breakpoint has been inserted,
532 if reading memory would return the breakpoint instead of the
533 underlying memory; the field @code{shadow_len} is the length of
534 memory cached in @code{shadow_contents}, if any; and the field
535 @code{placed_size} is optionally set and used by the target, if
536 it could differ from @code{shadow_len}.
538 For example, the remote target @samp{Z0} packet does not require
539 shadowing memory, so @code{shadow_len} is left at zero. However,
540 the length reported by @code{gdbarch_breakpoint_from_pc} is cached in
541 @code{placed_size}, so that a matching @samp{z0} packet can be
542 used to remove the breakpoint.
544 @cindex insert or remove hardware breakpoint
545 @findex target_remove_hw_breakpoint
546 @findex target_insert_hw_breakpoint
547 @item target_remove_hw_breakpoint (@var{bp_tgt})
548 @itemx target_insert_hw_breakpoint (@var{bp_tgt})
549 Insert or remove a hardware-assisted breakpoint at address
550 @code{@var{bp_tgt}->placed_address}. Returns zero for success,
551 non-zero for failure. See @code{target_insert_breakpoint} for
552 a description of the @code{struct bp_target_info} pointed to by
553 @var{bp_tgt}; the @code{shadow_contents} and
554 @code{shadow_len} members are not used for hardware breakpoints,
555 but @code{placed_size} may be.
558 @section Single Stepping
560 @section Signal Handling
562 @section Thread Handling
564 @section Inferior Function Calls
566 @section Longjmp Support
568 @cindex @code{longjmp} debugging
569 @value{GDBN} has support for figuring out that the target is doing a
570 @code{longjmp} and for stopping at the target of the jump, if we are
571 stepping. This is done with a few specialized internal breakpoints,
572 which are visible in the output of the @samp{maint info breakpoint}
575 @findex gdbarch_get_longjmp_target
576 To make this work, you need to define a function called
577 @code{gdbarch_get_longjmp_target}, which will examine the @code{jmp_buf}
578 structure and extract the longjmp target address. Since @code{jmp_buf}
579 is target specific, you will need to define it in the appropriate
580 @file{tm-@var{target}.h} file. Look in @file{tm-sun4os4.h} and
581 @file{sparc-tdep.c} for examples of how to do this.
586 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
587 breakpoints}) which break when data is accessed rather than when some
588 instruction is executed. When you have data which changes without
589 your knowing what code does that, watchpoints are the silver bullet to
590 hunt down and kill such bugs.
592 @cindex hardware watchpoints
593 @cindex software watchpoints
594 Watchpoints can be either hardware-assisted or not; the latter type is
595 known as ``software watchpoints.'' @value{GDBN} always uses
596 hardware-assisted watchpoints if they are available, and falls back on
597 software watchpoints otherwise. Typical situations where @value{GDBN}
598 will use software watchpoints are:
602 The watched memory region is too large for the underlying hardware
603 watchpoint support. For example, each x86 debug register can watch up
604 to 4 bytes of memory, so trying to watch data structures whose size is
605 more than 16 bytes will cause @value{GDBN} to use software
609 The value of the expression to be watched depends on data held in
610 registers (as opposed to memory).
613 Too many different watchpoints requested. (On some architectures,
614 this situation is impossible to detect until the debugged program is
615 resumed.) Note that x86 debug registers are used both for hardware
616 breakpoints and for watchpoints, so setting too many hardware
617 breakpoints might cause watchpoint insertion to fail.
620 No hardware-assisted watchpoints provided by the target
624 Software watchpoints are very slow, since @value{GDBN} needs to
625 single-step the program being debugged and test the value of the
626 watched expression(s) after each instruction. The rest of this
627 section is mostly irrelevant for software watchpoints.
629 When the inferior stops, @value{GDBN} tries to establish, among other
630 possible reasons, whether it stopped due to a watchpoint being hit.
631 It first uses @code{STOPPED_BY_WATCHPOINT} to see if any watchpoint
632 was hit. If not, all watchpoint checking is skipped.
634 Then @value{GDBN} calls @code{target_stopped_data_address} exactly
635 once. This method returns the address of the watchpoint which
636 triggered, if the target can determine it. If the triggered address
637 is available, @value{GDBN} compares the address returned by this
638 method with each watched memory address in each active watchpoint.
639 For data-read and data-access watchpoints, @value{GDBN} announces
640 every watchpoint that watches the triggered address as being hit.
641 For this reason, data-read and data-access watchpoints
642 @emph{require} that the triggered address be available; if not, read
643 and access watchpoints will never be considered hit. For data-write
644 watchpoints, if the triggered address is available, @value{GDBN}
645 considers only those watchpoints which match that address;
646 otherwise, @value{GDBN} considers all data-write watchpoints. For
647 each data-write watchpoint that @value{GDBN} considers, it evaluates
648 the expression whose value is being watched, and tests whether the
649 watched value has changed. Watchpoints whose watched values have
650 changed are announced as hit.
652 @value{GDBN} uses several macros and primitives to support hardware
656 @findex TARGET_HAS_HARDWARE_WATCHPOINTS
657 @item TARGET_HAS_HARDWARE_WATCHPOINTS
658 If defined, the target supports hardware watchpoints.
660 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
661 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
662 Return the number of hardware watchpoints of type @var{type} that are
663 possible to be set. The value is positive if @var{count} watchpoints
664 of this type can be set, zero if setting watchpoints of this type is
665 not supported, and negative if @var{count} is more than the maximum
666 number of watchpoints of type @var{type} that can be set. @var{other}
667 is non-zero if other types of watchpoints are currently enabled (there
668 are architectures which cannot set watchpoints of different types at
671 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
672 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
673 Return non-zero if hardware watchpoints can be used to watch a region
674 whose address is @var{addr} and whose length in bytes is @var{len}.
676 @cindex insert or remove hardware watchpoint
677 @findex target_insert_watchpoint
678 @findex target_remove_watchpoint
679 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
680 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
681 Insert or remove a hardware watchpoint starting at @var{addr}, for
682 @var{len} bytes. @var{type} is the watchpoint type, one of the
683 possible values of the enumerated data type @code{target_hw_bp_type},
684 defined by @file{breakpoint.h} as follows:
687 enum target_hw_bp_type
689 hw_write = 0, /* Common (write) HW watchpoint */
690 hw_read = 1, /* Read HW watchpoint */
691 hw_access = 2, /* Access (read or write) HW watchpoint */
692 hw_execute = 3 /* Execute HW breakpoint */
697 These two macros should return 0 for success, non-zero for failure.
699 @findex target_stopped_data_address
700 @item target_stopped_data_address (@var{addr_p})
701 If the inferior has some watchpoint that triggered, place the address
702 associated with the watchpoint at the location pointed to by
703 @var{addr_p} and return non-zero. Otherwise, return zero. This
704 is required for data-read and data-access watchpoints. It is
705 not required for data-write watchpoints, but @value{GDBN} uses
706 it to improve handling of those also.
708 @value{GDBN} will only call this method once per watchpoint stop,
709 immediately after calling @code{STOPPED_BY_WATCHPOINT}. If the
710 target's watchpoint indication is sticky, i.e., stays set after
711 resuming, this method should clear it. For instance, the x86 debug
712 control register has sticky triggered flags.
714 @findex target_watchpoint_addr_within_range
715 @item target_watchpoint_addr_within_range (@var{target}, @var{addr}, @var{start}, @var{length})
716 Check whether @var{addr} (as returned by @code{target_stopped_data_address})
717 lies within the hardware-defined watchpoint region described by
718 @var{start} and @var{length}. This only needs to be provided if the
719 granularity of a watchpoint is greater than one byte, i.e., if the
720 watchpoint can also trigger on nearby addresses outside of the watched
723 @findex HAVE_STEPPABLE_WATCHPOINT
724 @item HAVE_STEPPABLE_WATCHPOINT
725 If defined to a non-zero value, it is not necessary to disable a
726 watchpoint to step over it. Like @code{gdbarch_have_nonsteppable_watchpoint},
727 this is usually set when watchpoints trigger at the instruction
728 which will perform an interesting read or write. It should be
729 set if there is a temporary disable bit which allows the processor
730 to step over the interesting instruction without raising the
731 watchpoint exception again.
733 @findex gdbarch_have_nonsteppable_watchpoint
734 @item int gdbarch_have_nonsteppable_watchpoint (@var{gdbarch})
735 If it returns a non-zero value, @value{GDBN} should disable a
736 watchpoint to step the inferior over it. This is usually set when
737 watchpoints trigger at the instruction which will perform an
738 interesting read or write.
740 @findex HAVE_CONTINUABLE_WATCHPOINT
741 @item HAVE_CONTINUABLE_WATCHPOINT
742 If defined to a non-zero value, it is possible to continue the
743 inferior after a watchpoint has been hit. This is usually set
744 when watchpoints trigger at the instruction following an interesting
747 @findex CANNOT_STEP_HW_WATCHPOINTS
748 @item CANNOT_STEP_HW_WATCHPOINTS
749 If this is defined to a non-zero value, @value{GDBN} will remove all
750 watchpoints before stepping the inferior.
752 @findex STOPPED_BY_WATCHPOINT
753 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
754 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
755 the type @code{struct target_waitstatus}, defined by @file{target.h}.
756 Normally, this macro is defined to invoke the function pointed to by
757 the @code{to_stopped_by_watchpoint} member of the structure (of the
758 type @code{target_ops}, defined on @file{target.h}) that describes the
759 target-specific operations; @code{to_stopped_by_watchpoint} ignores
760 the @var{wait_status} argument.
762 @value{GDBN} does not require the non-zero value returned by
763 @code{STOPPED_BY_WATCHPOINT} to be 100% correct, so if a target cannot
764 determine for sure whether the inferior stopped due to a watchpoint,
765 it could return non-zero ``just in case''.
768 @subsection Watchpoints and Threads
769 @cindex watchpoints, with threads
771 @value{GDBN} only supports process-wide watchpoints, which trigger
772 in all threads. @value{GDBN} uses the thread ID to make watchpoints
773 act as if they were thread-specific, but it cannot set hardware
774 watchpoints that only trigger in a specific thread. Therefore, even
775 if the target supports threads, per-thread debug registers, and
776 watchpoints which only affect a single thread, it should set the
777 per-thread debug registers for all threads to the same value. On
778 @sc{gnu}/Linux native targets, this is accomplished by using
779 @code{ALL_LWPS} in @code{target_insert_watchpoint} and
780 @code{target_remove_watchpoint} and by using
781 @code{linux_set_new_thread} to register a handler for newly created
784 @value{GDBN}'s @sc{gnu}/Linux support only reports a single event
785 at a time, although multiple events can trigger simultaneously for
786 multi-threaded programs. When multiple events occur, @file{linux-nat.c}
787 queues subsequent events and returns them the next time the program
788 is resumed. This means that @code{STOPPED_BY_WATCHPOINT} and
789 @code{target_stopped_data_address} only need to consult the current
790 thread's state---the thread indicated by @code{inferior_ptid}. If
791 two threads have hit watchpoints simultaneously, those routines
792 will be called a second time for the second thread.
794 @subsection x86 Watchpoints
795 @cindex x86 debug registers
796 @cindex watchpoints, on x86
798 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
799 registers designed to facilitate debugging. @value{GDBN} provides a
800 generic library of functions that x86-based ports can use to implement
801 support for watchpoints and hardware-assisted breakpoints. This
802 subsection documents the x86 watchpoint facilities in @value{GDBN}.
804 To use the generic x86 watchpoint support, a port should do the
808 @findex I386_USE_GENERIC_WATCHPOINTS
810 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
811 target-dependent headers.
814 Include the @file{config/i386/nm-i386.h} header file @emph{after}
815 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
818 Add @file{i386-nat.o} to the value of the Make variable
819 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}) or
820 @code{TDEPFILES} (@pxref{Target Architecture Definition, TDEPFILES}).
823 Provide implementations for the @code{I386_DR_LOW_*} macros described
824 below. Typically, each macro should call a target-specific function
825 which does the real work.
828 The x86 watchpoint support works by maintaining mirror images of the
829 debug registers. Values are copied between the mirror images and the
830 real debug registers via a set of macros which each target needs to
834 @findex I386_DR_LOW_SET_CONTROL
835 @item I386_DR_LOW_SET_CONTROL (@var{val})
836 Set the Debug Control (DR7) register to the value @var{val}.
838 @findex I386_DR_LOW_SET_ADDR
839 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
840 Put the address @var{addr} into the debug register number @var{idx}.
842 @findex I386_DR_LOW_RESET_ADDR
843 @item I386_DR_LOW_RESET_ADDR (@var{idx})
844 Reset (i.e.@: zero out) the address stored in the debug register
847 @findex I386_DR_LOW_GET_STATUS
848 @item I386_DR_LOW_GET_STATUS
849 Return the value of the Debug Status (DR6) register. This value is
850 used immediately after it is returned by
851 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
855 For each one of the 4 debug registers (whose indices are from 0 to 3)
856 that store addresses, a reference count is maintained by @value{GDBN},
857 to allow sharing of debug registers by several watchpoints. This
858 allows users to define several watchpoints that watch the same
859 expression, but with different conditions and/or commands, without
860 wasting debug registers which are in short supply. @value{GDBN}
861 maintains the reference counts internally, targets don't have to do
862 anything to use this feature.
864 The x86 debug registers can each watch a region that is 1, 2, or 4
865 bytes long. The ia32 architecture requires that each watched region
866 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
867 region on 4-byte boundary. However, the x86 watchpoint support in
868 @value{GDBN} can watch unaligned regions and regions larger than 4
869 bytes (up to 16 bytes) by allocating several debug registers to watch
870 a single region. This allocation of several registers per a watched
871 region is also done automatically without target code intervention.
873 The generic x86 watchpoint support provides the following API for the
874 @value{GDBN}'s application code:
877 @findex i386_region_ok_for_watchpoint
878 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
879 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
880 this function. It counts the number of debug registers required to
881 watch a given region, and returns a non-zero value if that number is
882 less than 4, the number of debug registers available to x86
885 @findex i386_stopped_data_address
886 @item i386_stopped_data_address (@var{addr_p})
888 @code{target_stopped_data_address} is set to call this function.
890 function examines the breakpoint condition bits in the DR6 Debug
891 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
892 macro, and returns the address associated with the first bit that is
895 @findex i386_stopped_by_watchpoint
896 @item i386_stopped_by_watchpoint (void)
897 The macro @code{STOPPED_BY_WATCHPOINT}
898 is set to call this function. The
899 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
900 function examines the breakpoint condition bits in the DR6 Debug
901 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
902 macro, and returns true if any bit is set. Otherwise, false is
905 @findex i386_insert_watchpoint
906 @findex i386_remove_watchpoint
907 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
908 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
909 Insert or remove a watchpoint. The macros
910 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
911 are set to call these functions. @code{i386_insert_watchpoint} first
912 looks for a debug register which is already set to watch the same
913 region for the same access types; if found, it just increments the
914 reference count of that debug register, thus implementing debug
915 register sharing between watchpoints. If no such register is found,
916 the function looks for a vacant debug register, sets its mirrored
917 value to @var{addr}, sets the mirrored value of DR7 Debug Control
918 register as appropriate for the @var{len} and @var{type} parameters,
919 and then passes the new values of the debug register and DR7 to the
920 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
921 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
922 required to cover the given region, the above process is repeated for
925 @code{i386_remove_watchpoint} does the opposite: it resets the address
926 in the mirrored value of the debug register and its read/write and
927 length bits in the mirrored value of DR7, then passes these new
928 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
929 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
930 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
931 decrements the reference count, and only calls
932 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
933 the count goes to zero.
935 @findex i386_insert_hw_breakpoint
936 @findex i386_remove_hw_breakpoint
937 @item i386_insert_hw_breakpoint (@var{bp_tgt})
938 @itemx i386_remove_hw_breakpoint (@var{bp_tgt})
939 These functions insert and remove hardware-assisted breakpoints. The
940 macros @code{target_insert_hw_breakpoint} and
941 @code{target_remove_hw_breakpoint} are set to call these functions.
942 The argument is a @code{struct bp_target_info *}, as described in
943 the documentation for @code{target_insert_breakpoint}.
944 These functions work like @code{i386_insert_watchpoint} and
945 @code{i386_remove_watchpoint}, respectively, except that they set up
946 the debug registers to watch instruction execution, and each
947 hardware-assisted breakpoint always requires exactly one debug
950 @findex i386_stopped_by_hwbp
951 @item i386_stopped_by_hwbp (void)
952 This function returns non-zero if the inferior has some watchpoint or
953 hardware breakpoint that triggered. It works like
954 @code{i386_stopped_data_address}, except that it doesn't record the
955 address whose watchpoint triggered.
957 @findex i386_cleanup_dregs
958 @item i386_cleanup_dregs (void)
959 This function clears all the reference counts, addresses, and control
960 bits in the mirror images of the debug registers. It doesn't affect
961 the actual debug registers in the inferior process.
968 x86 processors support setting watchpoints on I/O reads or writes.
969 However, since no target supports this (as of March 2001), and since
970 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
971 watchpoints, this feature is not yet available to @value{GDBN} running
975 x86 processors can enable watchpoints locally, for the current task
976 only, or globally, for all the tasks. For each debug register,
977 there's a bit in the DR7 Debug Control register that determines
978 whether the associated address is watched locally or globally. The
979 current implementation of x86 watchpoint support in @value{GDBN}
980 always sets watchpoints to be locally enabled, since global
981 watchpoints might interfere with the underlying OS and are probably
982 unavailable in many platforms.
988 In the abstract, a checkpoint is a point in the execution history of
989 the program, which the user may wish to return to at some later time.
991 Internally, a checkpoint is a saved copy of the program state, including
992 whatever information is required in order to restore the program to that
993 state at a later time. This can be expected to include the state of
994 registers and memory, and may include external state such as the state
995 of open files and devices.
997 There are a number of ways in which checkpoints may be implemented
998 in gdb, e.g.@: as corefiles, as forked processes, and as some opaque
999 method implemented on the target side.
1001 A corefile can be used to save an image of target memory and register
1002 state, which can in principle be restored later --- but corefiles do
1003 not typically include information about external entities such as
1004 open files. Currently this method is not implemented in gdb.
1006 A forked process can save the state of user memory and registers,
1007 as well as some subset of external (kernel) state. This method
1008 is used to implement checkpoints on Linux, and in principle might
1009 be used on other systems.
1011 Some targets, e.g.@: simulators, might have their own built-in
1012 method for saving checkpoints, and gdb might be able to take
1013 advantage of that capability without necessarily knowing any
1014 details of how it is done.
1017 @section Observing changes in @value{GDBN} internals
1018 @cindex observer pattern interface
1019 @cindex notifications about changes in internals
1021 In order to function properly, several modules need to be notified when
1022 some changes occur in the @value{GDBN} internals. Traditionally, these
1023 modules have relied on several paradigms, the most common ones being
1024 hooks and gdb-events. Unfortunately, none of these paradigms was
1025 versatile enough to become the standard notification mechanism in
1026 @value{GDBN}. The fact that they only supported one ``client'' was also
1027 a strong limitation.
1029 A new paradigm, based on the Observer pattern of the @cite{Design
1030 Patterns} book, has therefore been implemented. The goal was to provide
1031 a new interface overcoming the issues with the notification mechanisms
1032 previously available. This new interface needed to be strongly typed,
1033 easy to extend, and versatile enough to be used as the standard
1034 interface when adding new notifications.
1036 See @ref{GDB Observers} for a brief description of the observers
1037 currently implemented in GDB. The rationale for the current
1038 implementation is also briefly discussed.
1040 @node User Interface
1042 @chapter User Interface
1044 @value{GDBN} has several user interfaces. Although the command-line interface
1045 is the most common and most familiar, there are others.
1047 @section Command Interpreter
1049 @cindex command interpreter
1051 The command interpreter in @value{GDBN} is fairly simple. It is designed to
1052 allow for the set of commands to be augmented dynamically, and also
1053 has a recursive subcommand capability, where the first argument to
1054 a command may itself direct a lookup on a different command list.
1056 For instance, the @samp{set} command just starts a lookup on the
1057 @code{setlist} command list, while @samp{set thread} recurses
1058 to the @code{set_thread_cmd_list}.
1062 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
1063 the main command list, and should be used for those commands. The usual
1064 place to add commands is in the @code{_initialize_@var{xyz}} routines at
1065 the ends of most source files.
1067 @findex add_setshow_cmd
1068 @findex add_setshow_cmd_full
1069 To add paired @samp{set} and @samp{show} commands, use
1070 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
1071 a slightly simpler interface which is useful when you don't need to
1072 further modify the new command structures, while the latter returns
1073 the new command structures for manipulation.
1075 @cindex deprecating commands
1076 @findex deprecate_cmd
1077 Before removing commands from the command set it is a good idea to
1078 deprecate them for some time. Use @code{deprecate_cmd} on commands or
1079 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
1080 @code{struct cmd_list_element} as it's first argument. You can use the
1081 return value from @code{add_com} or @code{add_cmd} to deprecate the
1082 command immediately after it is created.
1084 The first time a command is used the user will be warned and offered a
1085 replacement (if one exists). Note that the replacement string passed to
1086 @code{deprecate_cmd} should be the full name of the command, i.e., the
1087 entire string the user should type at the command line.
1089 @section UI-Independent Output---the @code{ui_out} Functions
1090 @c This section is based on the documentation written by Fernando
1091 @c Nasser <fnasser@redhat.com>.
1093 @cindex @code{ui_out} functions
1094 The @code{ui_out} functions present an abstraction level for the
1095 @value{GDBN} output code. They hide the specifics of different user
1096 interfaces supported by @value{GDBN}, and thus free the programmer
1097 from the need to write several versions of the same code, one each for
1098 every UI, to produce output.
1100 @subsection Overview and Terminology
1102 In general, execution of each @value{GDBN} command produces some sort
1103 of output, and can even generate an input request.
1105 Output can be generated for the following purposes:
1109 to display a @emph{result} of an operation;
1112 to convey @emph{info} or produce side-effects of a requested
1116 to provide a @emph{notification} of an asynchronous event (including
1117 progress indication of a prolonged asynchronous operation);
1120 to display @emph{error messages} (including warnings);
1123 to show @emph{debug data};
1126 to @emph{query} or prompt a user for input (a special case).
1130 This section mainly concentrates on how to build result output,
1131 although some of it also applies to other kinds of output.
1133 Generation of output that displays the results of an operation
1134 involves one or more of the following:
1138 output of the actual data
1141 formatting the output as appropriate for console output, to make it
1142 easily readable by humans
1145 machine oriented formatting--a more terse formatting to allow for easy
1146 parsing by programs which read @value{GDBN}'s output
1149 annotation, whose purpose is to help legacy GUIs to identify interesting
1153 The @code{ui_out} routines take care of the first three aspects.
1154 Annotations are provided by separate annotation routines. Note that use
1155 of annotations for an interface between a GUI and @value{GDBN} is
1158 Output can be in the form of a single item, which we call a @dfn{field};
1159 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
1160 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
1161 header and a body. In a BNF-like form:
1164 @item <table> @expansion{}
1165 @code{<header> <body>}
1166 @item <header> @expansion{}
1167 @code{@{ <column> @}}
1168 @item <column> @expansion{}
1169 @code{<width> <alignment> <title>}
1170 @item <body> @expansion{}
1175 @subsection General Conventions
1177 Most @code{ui_out} routines are of type @code{void}, the exceptions are
1178 @code{ui_out_stream_new} (which returns a pointer to the newly created
1179 object) and the @code{make_cleanup} routines.
1181 The first parameter is always the @code{ui_out} vector object, a pointer
1182 to a @code{struct ui_out}.
1184 The @var{format} parameter is like in @code{printf} family of functions.
1185 When it is present, there must also be a variable list of arguments
1186 sufficient used to satisfy the @code{%} specifiers in the supplied
1189 When a character string argument is not used in a @code{ui_out} function
1190 call, a @code{NULL} pointer has to be supplied instead.
1193 @subsection Table, Tuple and List Functions
1195 @cindex list output functions
1196 @cindex table output functions
1197 @cindex tuple output functions
1198 This section introduces @code{ui_out} routines for building lists,
1199 tuples and tables. The routines to output the actual data items
1200 (fields) are presented in the next section.
1202 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
1203 containing information about an object; a @dfn{list} is a sequence of
1204 fields where each field describes an identical object.
1206 Use the @dfn{table} functions when your output consists of a list of
1207 rows (tuples) and the console output should include a heading. Use this
1208 even when you are listing just one object but you still want the header.
1210 @cindex nesting level in @code{ui_out} functions
1211 Tables can not be nested. Tuples and lists can be nested up to a
1212 maximum of five levels.
1214 The overall structure of the table output code is something like this:
1229 Here is the description of table-, tuple- and list-related @code{ui_out}
1232 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
1233 The function @code{ui_out_table_begin} marks the beginning of the output
1234 of a table. It should always be called before any other @code{ui_out}
1235 function for a given table. @var{nbrofcols} is the number of columns in
1236 the table. @var{nr_rows} is the number of rows in the table.
1237 @var{tblid} is an optional string identifying the table. The string
1238 pointed to by @var{tblid} is copied by the implementation of
1239 @code{ui_out_table_begin}, so the application can free the string if it
1240 was @code{malloc}ed.
1242 The companion function @code{ui_out_table_end}, described below, marks
1243 the end of the table's output.
1246 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
1247 @code{ui_out_table_header} provides the header information for a single
1248 table column. You call this function several times, one each for every
1249 column of the table, after @code{ui_out_table_begin}, but before
1250 @code{ui_out_table_body}.
1252 The value of @var{width} gives the column width in characters. The
1253 value of @var{alignment} is one of @code{left}, @code{center}, and
1254 @code{right}, and it specifies how to align the header: left-justify,
1255 center, or right-justify it. @var{colhdr} points to a string that
1256 specifies the column header; the implementation copies that string, so
1257 column header strings in @code{malloc}ed storage can be freed after the
1261 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
1262 This function delimits the table header from the table body.
1265 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
1266 This function signals the end of a table's output. It should be called
1267 after the table body has been produced by the list and field output
1270 There should be exactly one call to @code{ui_out_table_end} for each
1271 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
1272 will signal an internal error.
1275 The output of the tuples that represent the table rows must follow the
1276 call to @code{ui_out_table_body} and precede the call to
1277 @code{ui_out_table_end}. You build a tuple by calling
1278 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
1279 calls to functions which actually output fields between them.
1281 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
1282 This function marks the beginning of a tuple output. @var{id} points
1283 to an optional string that identifies the tuple; it is copied by the
1284 implementation, and so strings in @code{malloc}ed storage can be freed
1288 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
1289 This function signals an end of a tuple output. There should be exactly
1290 one call to @code{ui_out_tuple_end} for each call to
1291 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
1295 @deftypefun struct cleanup *make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1296 This function first opens the tuple and then establishes a cleanup
1297 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
1298 and correct implementation of the non-portable@footnote{The function
1299 cast is not portable ISO C.} code sequence:
1301 struct cleanup *old_cleanup;
1302 ui_out_tuple_begin (uiout, "...");
1303 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
1308 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
1309 This function marks the beginning of a list output. @var{id} points to
1310 an optional string that identifies the list; it is copied by the
1311 implementation, and so strings in @code{malloc}ed storage can be freed
1315 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
1316 This function signals an end of a list output. There should be exactly
1317 one call to @code{ui_out_list_end} for each call to
1318 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1322 @deftypefun struct cleanup *make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1323 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1324 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
1325 that will close the list.
1328 @subsection Item Output Functions
1330 @cindex item output functions
1331 @cindex field output functions
1333 The functions described below produce output for the actual data
1334 items, or fields, which contain information about the object.
1336 Choose the appropriate function accordingly to your particular needs.
1338 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1339 This is the most general output function. It produces the
1340 representation of the data in the variable-length argument list
1341 according to formatting specifications in @var{format}, a
1342 @code{printf}-like format string. The optional argument @var{fldname}
1343 supplies the name of the field. The data items themselves are
1344 supplied as additional arguments after @var{format}.
1346 This generic function should be used only when it is not possible to
1347 use one of the specialized versions (see below).
1350 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1351 This function outputs a value of an @code{int} variable. It uses the
1352 @code{"%d"} output conversion specification. @var{fldname} specifies
1353 the name of the field.
1356 @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})
1357 This function outputs a value of an @code{int} variable. It differs from
1358 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1359 @var{fldname} specifies
1360 the name of the field.
1363 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1364 This function outputs an address.
1367 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1368 This function outputs a string using the @code{"%s"} conversion
1372 Sometimes, there's a need to compose your output piece by piece using
1373 functions that operate on a stream, such as @code{value_print} or
1374 @code{fprintf_symbol_filtered}. These functions accept an argument of
1375 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1376 used to store the data stream used for the output. When you use one
1377 of these functions, you need a way to pass their results stored in a
1378 @code{ui_file} object to the @code{ui_out} functions. To this end,
1379 you first create a @code{ui_stream} object by calling
1380 @code{ui_out_stream_new}, pass the @code{stream} member of that
1381 @code{ui_stream} object to @code{value_print} and similar functions,
1382 and finally call @code{ui_out_field_stream} to output the field you
1383 constructed. When the @code{ui_stream} object is no longer needed,
1384 you should destroy it and free its memory by calling
1385 @code{ui_out_stream_delete}.
1387 @deftypefun struct ui_stream *ui_out_stream_new (struct ui_out *@var{uiout})
1388 This function creates a new @code{ui_stream} object which uses the
1389 same output methods as the @code{ui_out} object whose pointer is
1390 passed in @var{uiout}. It returns a pointer to the newly created
1391 @code{ui_stream} object.
1394 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1395 This functions destroys a @code{ui_stream} object specified by
1399 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1400 This function consumes all the data accumulated in
1401 @code{streambuf->stream} and outputs it like
1402 @code{ui_out_field_string} does. After a call to
1403 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1404 the stream is still valid and may be used for producing more fields.
1407 @strong{Important:} If there is any chance that your code could bail
1408 out before completing output generation and reaching the point where
1409 @code{ui_out_stream_delete} is called, it is necessary to set up a
1410 cleanup, to avoid leaking memory and other resources. Here's a
1411 skeleton code to do that:
1414 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1415 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1420 If the function already has the old cleanup chain set (for other kinds
1421 of cleanups), you just have to add your cleanup to it:
1424 mybuf = ui_out_stream_new (uiout);
1425 make_cleanup (ui_out_stream_delete, mybuf);
1428 Note that with cleanups in place, you should not call
1429 @code{ui_out_stream_delete} directly, or you would attempt to free the
1432 @subsection Utility Output Functions
1434 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1435 This function skips a field in a table. Use it if you have to leave
1436 an empty field without disrupting the table alignment. The argument
1437 @var{fldname} specifies a name for the (missing) filed.
1440 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1441 This function outputs the text in @var{string} in a way that makes it
1442 easy to be read by humans. For example, the console implementation of
1443 this method filters the text through a built-in pager, to prevent it
1444 from scrolling off the visible portion of the screen.
1446 Use this function for printing relatively long chunks of text around
1447 the actual field data: the text it produces is not aligned according
1448 to the table's format. Use @code{ui_out_field_string} to output a
1449 string field, and use @code{ui_out_message}, described below, to
1450 output short messages.
1453 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1454 This function outputs @var{nspaces} spaces. It is handy to align the
1455 text produced by @code{ui_out_text} with the rest of the table or
1459 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1460 This function produces a formatted message, provided that the current
1461 verbosity level is at least as large as given by @var{verbosity}. The
1462 current verbosity level is specified by the user with the @samp{set
1463 verbositylevel} command.@footnote{As of this writing (April 2001),
1464 setting verbosity level is not yet implemented, and is always returned
1465 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1466 argument more than zero will cause the message to never be printed.}
1469 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1470 This function gives the console output filter (a paging filter) a hint
1471 of where to break lines which are too long. Ignored for all other
1472 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1473 be printed to indent the wrapped text on the next line; it must remain
1474 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1475 explicit newline is produced by one of the other functions. If
1476 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1479 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1480 This function flushes whatever output has been accumulated so far, if
1481 the UI buffers output.
1485 @subsection Examples of Use of @code{ui_out} functions
1487 @cindex using @code{ui_out} functions
1488 @cindex @code{ui_out} functions, usage examples
1489 This section gives some practical examples of using the @code{ui_out}
1490 functions to generalize the old console-oriented code in
1491 @value{GDBN}. The examples all come from functions defined on the
1492 @file{breakpoints.c} file.
1494 This example, from the @code{breakpoint_1} function, shows how to
1497 The original code was:
1500 if (!found_a_breakpoint++)
1502 annotate_breakpoints_headers ();
1505 printf_filtered ("Num ");
1507 printf_filtered ("Type ");
1509 printf_filtered ("Disp ");
1511 printf_filtered ("Enb ");
1515 printf_filtered ("Address ");
1518 printf_filtered ("What\n");
1520 annotate_breakpoints_table ();
1524 Here's the new version:
1527 nr_printable_breakpoints = @dots{};
1530 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1532 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1534 if (nr_printable_breakpoints > 0)
1535 annotate_breakpoints_headers ();
1536 if (nr_printable_breakpoints > 0)
1538 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1539 if (nr_printable_breakpoints > 0)
1541 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1542 if (nr_printable_breakpoints > 0)
1544 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1545 if (nr_printable_breakpoints > 0)
1547 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1550 if (nr_printable_breakpoints > 0)
1552 if (gdbarch_addr_bit (current_gdbarch) <= 32)
1553 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1555 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1557 if (nr_printable_breakpoints > 0)
1559 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1560 ui_out_table_body (uiout);
1561 if (nr_printable_breakpoints > 0)
1562 annotate_breakpoints_table ();
1565 This example, from the @code{print_one_breakpoint} function, shows how
1566 to produce the actual data for the table whose structure was defined
1567 in the above example. The original code was:
1572 printf_filtered ("%-3d ", b->number);
1574 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1575 || ((int) b->type != bptypes[(int) b->type].type))
1576 internal_error ("bptypes table does not describe type #%d.",
1578 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1580 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1582 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1586 This is the new version:
1590 ui_out_tuple_begin (uiout, "bkpt");
1592 ui_out_field_int (uiout, "number", b->number);
1594 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1595 || ((int) b->type != bptypes[(int) b->type].type))
1596 internal_error ("bptypes table does not describe type #%d.",
1598 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1600 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1602 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1606 This example, also from @code{print_one_breakpoint}, shows how to
1607 produce a complicated output field using the @code{print_expression}
1608 functions which requires a stream to be passed. It also shows how to
1609 automate stream destruction with cleanups. The original code was:
1613 print_expression (b->exp, gdb_stdout);
1619 struct ui_stream *stb = ui_out_stream_new (uiout);
1620 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1623 print_expression (b->exp, stb->stream);
1624 ui_out_field_stream (uiout, "what", local_stream);
1627 This example, also from @code{print_one_breakpoint}, shows how to use
1628 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1633 if (b->dll_pathname == NULL)
1634 printf_filtered ("<any library> ");
1636 printf_filtered ("library \"%s\" ", b->dll_pathname);
1643 if (b->dll_pathname == NULL)
1645 ui_out_field_string (uiout, "what", "<any library>");
1646 ui_out_spaces (uiout, 1);
1650 ui_out_text (uiout, "library \"");
1651 ui_out_field_string (uiout, "what", b->dll_pathname);
1652 ui_out_text (uiout, "\" ");
1656 The following example from @code{print_one_breakpoint} shows how to
1657 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1662 if (b->forked_inferior_pid != 0)
1663 printf_filtered ("process %d ", b->forked_inferior_pid);
1670 if (b->forked_inferior_pid != 0)
1672 ui_out_text (uiout, "process ");
1673 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1674 ui_out_spaces (uiout, 1);
1678 Here's an example of using @code{ui_out_field_string}. The original
1683 if (b->exec_pathname != NULL)
1684 printf_filtered ("program \"%s\" ", b->exec_pathname);
1691 if (b->exec_pathname != NULL)
1693 ui_out_text (uiout, "program \"");
1694 ui_out_field_string (uiout, "what", b->exec_pathname);
1695 ui_out_text (uiout, "\" ");
1699 Finally, here's an example of printing an address. The original code:
1703 printf_filtered ("%s ",
1704 hex_string_custom ((unsigned long) b->address, 8));
1711 ui_out_field_core_addr (uiout, "Address", b->address);
1715 @section Console Printing
1724 @cindex @code{libgdb}
1725 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1726 to provide an API to @value{GDBN}'s functionality.
1729 @cindex @code{libgdb}
1730 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1731 better able to support graphical and other environments.
1733 Since @code{libgdb} development is on-going, its architecture is still
1734 evolving. The following components have so far been identified:
1738 Observer - @file{gdb-events.h}.
1740 Builder - @file{ui-out.h}
1742 Event Loop - @file{event-loop.h}
1744 Library - @file{gdb.h}
1747 The model that ties these components together is described below.
1749 @section The @code{libgdb} Model
1751 A client of @code{libgdb} interacts with the library in two ways.
1755 As an observer (using @file{gdb-events}) receiving notifications from
1756 @code{libgdb} of any internal state changes (break point changes, run
1759 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1760 obtain various status values from @value{GDBN}.
1763 Since @code{libgdb} could have multiple clients (e.g., a GUI supporting
1764 the existing @value{GDBN} CLI), those clients must co-operate when
1765 controlling @code{libgdb}. In particular, a client must ensure that
1766 @code{libgdb} is idle (i.e. no other client is using @code{libgdb})
1767 before responding to a @file{gdb-event} by making a query.
1769 @section CLI support
1771 At present @value{GDBN}'s CLI is very much entangled in with the core of
1772 @code{libgdb}. Consequently, a client wishing to include the CLI in
1773 their interface needs to carefully co-ordinate its own and the CLI's
1776 It is suggested that the client set @code{libgdb} up to be bi-modal
1777 (alternate between CLI and client query modes). The notes below sketch
1782 The client registers itself as an observer of @code{libgdb}.
1784 The client create and install @code{cli-out} builder using its own
1785 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1786 @code{gdb_stdout} streams.
1788 The client creates a separate custom @code{ui-out} builder that is only
1789 used while making direct queries to @code{libgdb}.
1792 When the client receives input intended for the CLI, it simply passes it
1793 along. Since the @code{cli-out} builder is installed by default, all
1794 the CLI output in response to that command is routed (pronounced rooted)
1795 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1796 At the same time, the client is kept abreast of internal changes by
1797 virtue of being a @code{libgdb} observer.
1799 The only restriction on the client is that it must wait until
1800 @code{libgdb} becomes idle before initiating any queries (using the
1801 client's custom builder).
1803 @section @code{libgdb} components
1805 @subheading Observer - @file{gdb-events.h}
1806 @file{gdb-events} provides the client with a very raw mechanism that can
1807 be used to implement an observer. At present it only allows for one
1808 observer and that observer must, internally, handle the need to delay
1809 the processing of any event notifications until after @code{libgdb} has
1810 finished the current command.
1812 @subheading Builder - @file{ui-out.h}
1813 @file{ui-out} provides the infrastructure necessary for a client to
1814 create a builder. That builder is then passed down to @code{libgdb}
1815 when doing any queries.
1817 @subheading Event Loop - @file{event-loop.h}
1818 @c There could be an entire section on the event-loop
1819 @file{event-loop}, currently non-re-entrant, provides a simple event
1820 loop. A client would need to either plug its self into this loop or,
1821 implement a new event-loop that GDB would use.
1823 The event-loop will eventually be made re-entrant. This is so that
1824 @value{GDBN} can better handle the problem of some commands blocking
1825 instead of returning.
1827 @subheading Library - @file{gdb.h}
1828 @file{libgdb} is the most obvious component of this system. It provides
1829 the query interface. Each function is parameterized by a @code{ui-out}
1830 builder. The result of the query is constructed using that builder
1831 before the query function returns.
1834 @chapter Stack Frames
1837 @cindex call stack frame
1838 A frame is a construct that @value{GDBN} uses to keep track of calling
1839 and called functions.
1841 @cindex unwind frame
1842 @value{GDBN}'s frame model, a fresh design, was implemented with the
1843 need to support @sc{dwarf}'s Call Frame Information in mind. In fact,
1844 the term ``unwind'' is taken directly from that specification.
1845 Developers wishing to learn more about unwinders, are encouraged to
1846 read the @sc{dwarf} specification, available from
1847 @url{http://www.dwarfstd.org}.
1849 @findex frame_register_unwind
1850 @findex get_frame_register
1851 @value{GDBN}'s model is that you find a frame's registers by
1852 ``unwinding'' them from the next younger frame. That is,
1853 @samp{get_frame_register} which returns the value of a register in
1854 frame #1 (the next-to-youngest frame), is implemented by calling frame
1855 #0's @code{frame_register_unwind} (the youngest frame). But then the
1856 obvious question is: how do you access the registers of the youngest
1859 @cindex sentinel frame
1860 @findex get_frame_type
1861 @vindex SENTINEL_FRAME
1862 To answer this question, GDB has the @dfn{sentinel} frame, the
1863 ``-1st'' frame. Unwinding registers from the sentinel frame gives you
1864 the current values of the youngest real frame's registers. If @var{f}
1865 is a sentinel frame, then @code{get_frame_type (@var{f}) @equiv{}
1868 @section Selecting an Unwinder
1870 @findex frame_unwind_prepend_unwinder
1871 @findex frame_unwind_append_unwinder
1872 The architecture registers a list of frame unwinders (@code{struct
1873 frame_unwind}), using the functions
1874 @code{frame_unwind_prepend_unwinder} and
1875 @code{frame_unwind_append_unwinder}. Each unwinder includes a
1876 sniffer. Whenever @value{GDBN} needs to unwind a frame (to fetch the
1877 previous frame's registers or the current frame's ID), it calls
1878 registered sniffers in order to find one which recognizes the frame.
1879 The first time a sniffer returns non-zero, the corresponding unwinder
1880 is assigned to the frame.
1882 @section Unwinding the Frame ID
1885 Every frame has an associated ID, of type @code{struct frame_id}.
1886 The ID includes the stack base and function start address for
1887 the frame. The ID persists through the entire life of the frame,
1888 including while other called frames are running; it is used to
1889 locate an appropriate @code{struct frame_info} from the cache.
1891 Every time the inferior stops, and at various other times, the frame
1892 cache is flushed. Because of this, parts of @value{GDBN} which need
1893 to keep track of individual frames cannot use pointers to @code{struct
1894 frame_info}. A frame ID provides a stable reference to a frame, even
1895 when the unwinder must be run again to generate a new @code{struct
1896 frame_info} for the same frame.
1898 The frame's unwinder's @code{this_id} method is called to find the ID.
1899 Note that this is different from register unwinding, where the next
1900 frame's @code{prev_register} is called to unwind this frame's
1903 Both stack base and function address are required to identify the
1904 frame, because a recursive function has the same function address for
1905 two consecutive frames and a leaf function may have the same stack
1906 address as its caller. On some platforms, a third address is part of
1907 the ID to further disambiguate frames---for instance, on IA-64
1908 the separate register stack address is included in the ID.
1910 An invalid frame ID (@code{null_frame_id}) returned from the
1911 @code{this_id} method means to stop unwinding after this frame.
1913 @section Unwinding Registers
1915 Each unwinder includes a @code{prev_register} method. This method
1916 takes a frame, an associated cache pointer, and a register number.
1917 It returns a @code{struct value *} describing the requested register,
1918 as saved by this frame. This is the value of the register that is
1919 current in this frame's caller.
1921 The returned value must have the same type as the register. It may
1922 have any lvalue type. In most circumstances one of these routines
1923 will generate the appropriate value:
1926 @item frame_unwind_got_optimized
1927 @findex frame_unwind_got_optimized
1928 This register was not saved.
1930 @item frame_unwind_got_register
1931 @findex frame_unwind_got_register
1932 This register was copied into another register in this frame. This
1933 is also used for unchanged registers; they are ``copied'' into the
1936 @item frame_unwind_got_memory
1937 @findex frame_unwind_got_memory
1938 This register was saved in memory.
1940 @item frame_unwind_got_constant
1941 @findex frame_unwind_got_constant
1942 This register was not saved, but the unwinder can compute the previous
1943 value some other way.
1945 @item frame_unwind_got_address
1946 @findex frame_unwind_got_address
1947 Same as @code{frame_unwind_got_constant}, except that the value is a target
1948 address. This is frequently used for the stack pointer, which is not
1949 explicitly saved but has a known offset from this frame's stack
1950 pointer. For architectures with a flat unified address space, this is
1951 generally the same as @code{frame_unwind_got_constant}.
1954 @node Symbol Handling
1956 @chapter Symbol Handling
1958 Symbols are a key part of @value{GDBN}'s operation. Symbols include variables,
1959 functions, and types.
1961 @section Symbol Reading
1963 @cindex symbol reading
1964 @cindex reading of symbols
1965 @cindex symbol files
1966 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
1967 file is the file containing the program which @value{GDBN} is
1968 debugging. @value{GDBN} can be directed to use a different file for
1969 symbols (with the @samp{symbol-file} command), and it can also read
1970 more symbols via the @samp{add-file} and @samp{load} commands, or while
1971 reading symbols from shared libraries.
1973 @findex find_sym_fns
1974 Symbol files are initially opened by code in @file{symfile.c} using
1975 the BFD library (@pxref{Support Libraries}). BFD identifies the type
1976 of the file by examining its header. @code{find_sym_fns} then uses
1977 this identification to locate a set of symbol-reading functions.
1979 @findex add_symtab_fns
1980 @cindex @code{sym_fns} structure
1981 @cindex adding a symbol-reading module
1982 Symbol-reading modules identify themselves to @value{GDBN} by calling
1983 @code{add_symtab_fns} during their module initialization. The argument
1984 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
1985 name (or name prefix) of the symbol format, the length of the prefix,
1986 and pointers to four functions. These functions are called at various
1987 times to process symbol files whose identification matches the specified
1990 The functions supplied by each module are:
1993 @item @var{xyz}_symfile_init(struct sym_fns *sf)
1995 @cindex secondary symbol file
1996 Called from @code{symbol_file_add} when we are about to read a new
1997 symbol file. This function should clean up any internal state (possibly
1998 resulting from half-read previous files, for example) and prepare to
1999 read a new symbol file. Note that the symbol file which we are reading
2000 might be a new ``main'' symbol file, or might be a secondary symbol file
2001 whose symbols are being added to the existing symbol table.
2003 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
2004 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
2005 new symbol file being read. Its @code{private} field has been zeroed,
2006 and can be modified as desired. Typically, a struct of private
2007 information will be @code{malloc}'d, and a pointer to it will be placed
2008 in the @code{private} field.
2010 There is no result from @code{@var{xyz}_symfile_init}, but it can call
2011 @code{error} if it detects an unavoidable problem.
2013 @item @var{xyz}_new_init()
2015 Called from @code{symbol_file_add} when discarding existing symbols.
2016 This function needs only handle the symbol-reading module's internal
2017 state; the symbol table data structures visible to the rest of
2018 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
2019 arguments and no result. It may be called after
2020 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
2021 may be called alone if all symbols are simply being discarded.
2023 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
2025 Called from @code{symbol_file_add} to actually read the symbols from a
2026 symbol-file into a set of psymtabs or symtabs.
2028 @code{sf} points to the @code{struct sym_fns} originally passed to
2029 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
2030 the offset between the file's specified start address and its true
2031 address in memory. @code{mainline} is 1 if this is the main symbol
2032 table being read, and 0 if a secondary symbol file (e.g., shared library
2033 or dynamically loaded file) is being read.@refill
2036 In addition, if a symbol-reading module creates psymtabs when
2037 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
2038 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
2039 from any point in the @value{GDBN} symbol-handling code.
2042 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
2044 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
2045 the psymtab has not already been read in and had its @code{pst->symtab}
2046 pointer set. The argument is the psymtab to be fleshed-out into a
2047 symtab. Upon return, @code{pst->readin} should have been set to 1, and
2048 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
2049 zero if there were no symbols in that part of the symbol file.
2052 @section Partial Symbol Tables
2054 @value{GDBN} has three types of symbol tables:
2057 @cindex full symbol table
2060 Full symbol tables (@dfn{symtabs}). These contain the main
2061 information about symbols and addresses.
2065 Partial symbol tables (@dfn{psymtabs}). These contain enough
2066 information to know when to read the corresponding part of the full
2069 @cindex minimal symbol table
2072 Minimal symbol tables (@dfn{msymtabs}). These contain information
2073 gleaned from non-debugging symbols.
2076 @cindex partial symbol table
2077 This section describes partial symbol tables.
2079 A psymtab is constructed by doing a very quick pass over an executable
2080 file's debugging information. Small amounts of information are
2081 extracted---enough to identify which parts of the symbol table will
2082 need to be re-read and fully digested later, when the user needs the
2083 information. The speed of this pass causes @value{GDBN} to start up very
2084 quickly. Later, as the detailed rereading occurs, it occurs in small
2085 pieces, at various times, and the delay therefrom is mostly invisible to
2087 @c (@xref{Symbol Reading}.)
2089 The symbols that show up in a file's psymtab should be, roughly, those
2090 visible to the debugger's user when the program is not running code from
2091 that file. These include external symbols and types, static symbols and
2092 types, and @code{enum} values declared at file scope.
2094 The psymtab also contains the range of instruction addresses that the
2095 full symbol table would represent.
2097 @cindex finding a symbol
2098 @cindex symbol lookup
2099 The idea is that there are only two ways for the user (or much of the
2100 code in the debugger) to reference a symbol:
2103 @findex find_pc_function
2104 @findex find_pc_line
2106 By its address (e.g., execution stops at some address which is inside a
2107 function in this file). The address will be noticed to be in the
2108 range of this psymtab, and the full symtab will be read in.
2109 @code{find_pc_function}, @code{find_pc_line}, and other
2110 @code{find_pc_@dots{}} functions handle this.
2112 @cindex lookup_symbol
2115 (e.g., the user asks to print a variable, or set a breakpoint on a
2116 function). Global names and file-scope names will be found in the
2117 psymtab, which will cause the symtab to be pulled in. Local names will
2118 have to be qualified by a global name, or a file-scope name, in which
2119 case we will have already read in the symtab as we evaluated the
2120 qualifier. Or, a local symbol can be referenced when we are ``in'' a
2121 local scope, in which case the first case applies. @code{lookup_symbol}
2122 does most of the work here.
2125 The only reason that psymtabs exist is to cause a symtab to be read in
2126 at the right moment. Any symbol that can be elided from a psymtab,
2127 while still causing that to happen, should not appear in it. Since
2128 psymtabs don't have the idea of scope, you can't put local symbols in
2129 them anyway. Psymtabs don't have the idea of the type of a symbol,
2130 either, so types need not appear, unless they will be referenced by
2133 It is a bug for @value{GDBN} to behave one way when only a psymtab has
2134 been read, and another way if the corresponding symtab has been read
2135 in. Such bugs are typically caused by a psymtab that does not contain
2136 all the visible symbols, or which has the wrong instruction address
2139 The psymtab for a particular section of a symbol file (objfile) could be
2140 thrown away after the symtab has been read in. The symtab should always
2141 be searched before the psymtab, so the psymtab will never be used (in a
2142 bug-free environment). Currently, psymtabs are allocated on an obstack,
2143 and all the psymbols themselves are allocated in a pair of large arrays
2144 on an obstack, so there is little to be gained by trying to free them
2145 unless you want to do a lot more work.
2149 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
2151 @cindex fundamental types
2152 These are the fundamental types that @value{GDBN} uses internally. Fundamental
2153 types from the various debugging formats (stabs, ELF, etc) are mapped
2154 into one of these. They are basically a union of all fundamental types
2155 that @value{GDBN} knows about for all the languages that @value{GDBN}
2158 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
2161 Each time @value{GDBN} builds an internal type, it marks it with one
2162 of these types. The type may be a fundamental type, such as
2163 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
2164 which is a pointer to another type. Typically, several @code{FT_*}
2165 types map to one @code{TYPE_CODE_*} type, and are distinguished by
2166 other members of the type struct, such as whether the type is signed
2167 or unsigned, and how many bits it uses.
2169 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
2171 These are instances of type structs that roughly correspond to
2172 fundamental types and are created as global types for @value{GDBN} to
2173 use for various ugly historical reasons. We eventually want to
2174 eliminate these. Note for example that @code{builtin_type_int}
2175 initialized in @file{gdbtypes.c} is basically the same as a
2176 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
2177 an @code{FT_INTEGER} fundamental type. The difference is that the
2178 @code{builtin_type} is not associated with any particular objfile, and
2179 only one instance exists, while @file{c-lang.c} builds as many
2180 @code{TYPE_CODE_INT} types as needed, with each one associated with
2181 some particular objfile.
2183 @section Object File Formats
2184 @cindex object file formats
2188 @cindex @code{a.out} format
2189 The @code{a.out} format is the original file format for Unix. It
2190 consists of three sections: @code{text}, @code{data}, and @code{bss},
2191 which are for program code, initialized data, and uninitialized data,
2194 The @code{a.out} format is so simple that it doesn't have any reserved
2195 place for debugging information. (Hey, the original Unix hackers used
2196 @samp{adb}, which is a machine-language debugger!) The only debugging
2197 format for @code{a.out} is stabs, which is encoded as a set of normal
2198 symbols with distinctive attributes.
2200 The basic @code{a.out} reader is in @file{dbxread.c}.
2205 The COFF format was introduced with System V Release 3 (SVR3) Unix.
2206 COFF files may have multiple sections, each prefixed by a header. The
2207 number of sections is limited.
2209 The COFF specification includes support for debugging. Although this
2210 was a step forward, the debugging information was woefully limited. For
2211 instance, it was not possible to represent code that came from an
2214 The COFF reader is in @file{coffread.c}.
2218 @cindex ECOFF format
2219 ECOFF is an extended COFF originally introduced for Mips and Alpha
2222 The basic ECOFF reader is in @file{mipsread.c}.
2226 @cindex XCOFF format
2227 The IBM RS/6000 running AIX uses an object file format called XCOFF.
2228 The COFF sections, symbols, and line numbers are used, but debugging
2229 symbols are @code{dbx}-style stabs whose strings are located in the
2230 @code{.debug} section (rather than the string table). For more
2231 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
2233 The shared library scheme has a clean interface for figuring out what
2234 shared libraries are in use, but the catch is that everything which
2235 refers to addresses (symbol tables and breakpoints at least) needs to be
2236 relocated for both shared libraries and the main executable. At least
2237 using the standard mechanism this can only be done once the program has
2238 been run (or the core file has been read).
2242 @cindex PE-COFF format
2243 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
2244 executables. PE is basically COFF with additional headers.
2246 While BFD includes special PE support, @value{GDBN} needs only the basic
2252 The ELF format came with System V Release 4 (SVR4) Unix. ELF is similar
2253 to COFF in being organized into a number of sections, but it removes
2254 many of COFF's limitations.
2256 The basic ELF reader is in @file{elfread.c}.
2261 SOM is HP's object file and debug format (not to be confused with IBM's
2262 SOM, which is a cross-language ABI).
2264 The SOM reader is in @file{somread.c}.
2266 @section Debugging File Formats
2268 This section describes characteristics of debugging information that
2269 are independent of the object file format.
2273 @cindex stabs debugging info
2274 @code{stabs} started out as special symbols within the @code{a.out}
2275 format. Since then, it has been encapsulated into other file
2276 formats, such as COFF and ELF.
2278 While @file{dbxread.c} does some of the basic stab processing,
2279 including for encapsulated versions, @file{stabsread.c} does
2284 @cindex COFF debugging info
2285 The basic COFF definition includes debugging information. The level
2286 of support is minimal and non-extensible, and is not often used.
2288 @subsection Mips debug (Third Eye)
2290 @cindex ECOFF debugging info
2291 ECOFF includes a definition of a special debug format.
2293 The file @file{mdebugread.c} implements reading for this format.
2297 @cindex DWARF 2 debugging info
2298 DWARF 2 is an improved but incompatible version of DWARF 1.
2300 The DWARF 2 reader is in @file{dwarf2read.c}.
2302 @subsection Compressed DWARF 2
2304 @cindex Compressed DWARF 2 debugging info
2305 Compressed DWARF 2 is not technically a separate debugging format, but
2306 merely DWARF 2 debug information that has been compressed. In this
2307 format, every object-file section holding DWARF 2 debugging
2308 information is compressed and prepended with a header. (The section
2309 is also typically renamed, so a section called @code{.debug_info} in a
2310 DWARF 2 binary would be called @code{.zdebug_info} in a compressed
2311 DWARF 2 binary.) The header is 12 bytes long:
2315 4 bytes: the literal string ``ZLIB''
2317 8 bytes: the uncompressed size of the section, in big-endian byte
2321 The same reader is used for both compressed an normal DWARF 2 info.
2322 Section decompression is done in @code{zlib_decompress_section} in
2323 @file{dwarf2read.c}.
2327 @cindex SOM debugging info
2328 Like COFF, the SOM definition includes debugging information.
2330 @section Adding a New Symbol Reader to @value{GDBN}
2332 @cindex adding debugging info reader
2333 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
2334 there is probably little to be done.
2336 If you need to add a new object file format, you must first add it to
2337 BFD. This is beyond the scope of this document.
2339 You must then arrange for the BFD code to provide access to the
2340 debugging symbols. Generally @value{GDBN} will have to call swapping routines
2341 from BFD and a few other BFD internal routines to locate the debugging
2342 information. As much as possible, @value{GDBN} should not depend on the BFD
2343 internal data structures.
2345 For some targets (e.g., COFF), there is a special transfer vector used
2346 to call swapping routines, since the external data structures on various
2347 platforms have different sizes and layouts. Specialized routines that
2348 will only ever be implemented by one object file format may be called
2349 directly. This interface should be described in a file
2350 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
2352 @section Memory Management for Symbol Files
2354 Most memory associated with a loaded symbol file is stored on
2355 its @code{objfile_obstack}. This includes symbols, types,
2356 namespace data, and other information produced by the symbol readers.
2358 Because this data lives on the objfile's obstack, it is automatically
2359 released when the objfile is unloaded or reloaded. Therefore one
2360 objfile must not reference symbol or type data from another objfile;
2361 they could be unloaded at different times.
2363 User convenience variables, et cetera, have associated types. Normally
2364 these types live in the associated objfile. However, when the objfile
2365 is unloaded, those types are deep copied to global memory, so that
2366 the values of the user variables and history items are not lost.
2369 @node Language Support
2371 @chapter Language Support
2373 @cindex language support
2374 @value{GDBN}'s language support is mainly driven by the symbol reader,
2375 although it is possible for the user to set the source language
2378 @value{GDBN} chooses the source language by looking at the extension
2379 of the file recorded in the debug info; @file{.c} means C, @file{.f}
2380 means Fortran, etc. It may also use a special-purpose language
2381 identifier if the debug format supports it, like with DWARF.
2383 @section Adding a Source Language to @value{GDBN}
2385 @cindex adding source language
2386 To add other languages to @value{GDBN}'s expression parser, follow the
2390 @item Create the expression parser.
2392 @cindex expression parser
2393 This should reside in a file @file{@var{lang}-exp.y}. Routines for
2394 building parsed expressions into a @code{union exp_element} list are in
2397 @cindex language parser
2398 Since we can't depend upon everyone having Bison, and YACC produces
2399 parsers that define a bunch of global names, the following lines
2400 @strong{must} be included at the top of the YACC parser, to prevent the
2401 various parsers from defining the same global names:
2404 #define yyparse @var{lang}_parse
2405 #define yylex @var{lang}_lex
2406 #define yyerror @var{lang}_error
2407 #define yylval @var{lang}_lval
2408 #define yychar @var{lang}_char
2409 #define yydebug @var{lang}_debug
2410 #define yypact @var{lang}_pact
2411 #define yyr1 @var{lang}_r1
2412 #define yyr2 @var{lang}_r2
2413 #define yydef @var{lang}_def
2414 #define yychk @var{lang}_chk
2415 #define yypgo @var{lang}_pgo
2416 #define yyact @var{lang}_act
2417 #define yyexca @var{lang}_exca
2418 #define yyerrflag @var{lang}_errflag
2419 #define yynerrs @var{lang}_nerrs
2422 At the bottom of your parser, define a @code{struct language_defn} and
2423 initialize it with the right values for your language. Define an
2424 @code{initialize_@var{lang}} routine and have it call
2425 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
2426 that your language exists. You'll need some other supporting variables
2427 and functions, which will be used via pointers from your
2428 @code{@var{lang}_language_defn}. See the declaration of @code{struct
2429 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
2430 for more information.
2432 @item Add any evaluation routines, if necessary
2434 @cindex expression evaluation routines
2435 @findex evaluate_subexp
2436 @findex prefixify_subexp
2437 @findex length_of_subexp
2438 If you need new opcodes (that represent the operations of the language),
2439 add them to the enumerated type in @file{expression.h}. Add support
2440 code for these operations in the @code{evaluate_subexp} function
2441 defined in the file @file{eval.c}. Add cases
2442 for new opcodes in two functions from @file{parse.c}:
2443 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
2444 the number of @code{exp_element}s that a given operation takes up.
2446 @item Update some existing code
2448 Add an enumerated identifier for your language to the enumerated type
2449 @code{enum language} in @file{defs.h}.
2451 Update the routines in @file{language.c} so your language is included.
2452 These routines include type predicates and such, which (in some cases)
2453 are language dependent. If your language does not appear in the switch
2454 statement, an error is reported.
2456 @vindex current_language
2457 Also included in @file{language.c} is the code that updates the variable
2458 @code{current_language}, and the routines that translate the
2459 @code{language_@var{lang}} enumerated identifier into a printable
2462 @findex _initialize_language
2463 Update the function @code{_initialize_language} to include your
2464 language. This function picks the default language upon startup, so is
2465 dependent upon which languages that @value{GDBN} is built for.
2467 @findex allocate_symtab
2468 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2469 code so that the language of each symtab (source file) is set properly.
2470 This is used to determine the language to use at each stack frame level.
2471 Currently, the language is set based upon the extension of the source
2472 file. If the language can be better inferred from the symbol
2473 information, please set the language of the symtab in the symbol-reading
2476 @findex print_subexp
2477 @findex op_print_tab
2478 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2479 expression opcodes you have added to @file{expression.h}. Also, add the
2480 printed representations of your operators to @code{op_print_tab}.
2482 @item Add a place of call
2485 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2486 @code{parse_exp_1} (defined in @file{parse.c}).
2488 @item Use macros to trim code
2490 @cindex trimming language-dependent code
2491 The user has the option of building @value{GDBN} for some or all of the
2492 languages. If the user decides to build @value{GDBN} for the language
2493 @var{lang}, then every file dependent on @file{language.h} will have the
2494 macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to
2495 leave out large routines that the user won't need if he or she is not
2496 using your language.
2498 Note that you do not need to do this in your YACC parser, since if @value{GDBN}
2499 is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the
2500 compiled form of your parser) is not linked into @value{GDBN} at all.
2502 See the file @file{configure.in} for how @value{GDBN} is configured
2503 for different languages.
2505 @item Edit @file{Makefile.in}
2507 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2508 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2509 not get linked in, or, worse yet, it may not get @code{tar}red into the
2514 @node Host Definition
2516 @chapter Host Definition
2518 With the advent of Autoconf, it's rarely necessary to have host
2519 definition machinery anymore. The following information is provided,
2520 mainly, as an historical reference.
2522 @section Adding a New Host
2524 @cindex adding a new host
2525 @cindex host, adding
2526 @value{GDBN}'s host configuration support normally happens via Autoconf.
2527 New host-specific definitions should not be needed. Older hosts
2528 @value{GDBN} still use the host-specific definitions and files listed
2529 below, but these mostly exist for historical reasons, and will
2530 eventually disappear.
2533 @item gdb/config/@var{arch}/@var{xyz}.mh
2534 This file once contained both host and native configuration information
2535 (@pxref{Native Debugging}) for the machine @var{xyz}. The host
2536 configuration information is now handed by Autoconf.
2538 Host configuration information included a definition of
2539 @code{XM_FILE=xm-@var{xyz}.h} and possibly definitions for @code{CC},
2540 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2541 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2543 New host only configurations do not need this file.
2545 @item gdb/config/@var{arch}/xm-@var{xyz}.h
2546 This file once contained definitions and includes required when hosting
2547 gdb on machine @var{xyz}. Those definitions and includes are now
2548 handled by Autoconf.
2550 New host and native configurations do not need this file.
2552 @emph{Maintainer's note: Some hosts continue to use the @file{xm-xyz.h}
2553 file to define the macros @var{HOST_FLOAT_FORMAT},
2554 @var{HOST_DOUBLE_FORMAT} and @var{HOST_LONG_DOUBLE_FORMAT}. That code
2555 also needs to be replaced with either an Autoconf or run-time test.}
2559 @subheading Generic Host Support Files
2561 @cindex generic host support
2562 There are some ``generic'' versions of routines that can be used by
2563 various systems. These can be customized in various ways by macros
2564 defined in your @file{xm-@var{xyz}.h} file. If these routines work for
2565 the @var{xyz} host, you can just include the generic file's name (with
2566 @samp{.o}, not @samp{.c}) in @code{XDEPFILES}.
2568 Otherwise, if your machine needs custom support routines, you will need
2569 to write routines that perform the same functions as the generic file.
2570 Put them into @code{@var{xyz}-xdep.c}, and put @code{@var{xyz}-xdep.o}
2571 into @code{XDEPFILES}.
2574 @cindex remote debugging support
2575 @cindex serial line support
2577 This contains serial line support for Unix systems. This is always
2578 included, via the makefile variable @code{SER_HARDWIRE}; override this
2579 variable in the @file{.mh} file to avoid it.
2582 This contains serial line support for 32-bit programs running under DOS,
2583 using the DJGPP (a.k.a.@: GO32) execution environment.
2585 @cindex TCP remote support
2587 This contains generic TCP support using sockets.
2590 @section Host Conditionals
2592 When @value{GDBN} is configured and compiled, various macros are
2593 defined or left undefined, to control compilation based on the
2594 attributes of the host system. These macros and their meanings (or if
2595 the meaning is not documented here, then one of the source files where
2596 they are used is indicated) are:
2599 @item @value{GDBN}INIT_FILENAME
2600 The default name of @value{GDBN}'s initialization file (normally
2604 This macro is deprecated.
2606 @item SIGWINCH_HANDLER
2607 If your host defines @code{SIGWINCH}, you can define this to be the name
2608 of a function to be called if @code{SIGWINCH} is received.
2610 @item SIGWINCH_HANDLER_BODY
2611 Define this to expand into code that will define the function named by
2612 the expansion of @code{SIGWINCH_HANDLER}.
2614 @item CRLF_SOURCE_FILES
2615 @cindex DOS text files
2616 Define this if host files use @code{\r\n} rather than @code{\n} as a
2617 line terminator. This will cause source file listings to omit @code{\r}
2618 characters when printing and it will allow @code{\r\n} line endings of files
2619 which are ``sourced'' by gdb. It must be possible to open files in binary
2620 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2622 @item DEFAULT_PROMPT
2624 The default value of the prompt string (normally @code{"(gdb) "}).
2627 @cindex terminal device
2628 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2631 Define this if binary files are opened the same way as text files.
2635 In some cases, use the system call @code{mmap} for reading symbol
2636 tables. For some machines this allows for sharing and quick updates.
2639 Define this if the host system has @code{termio.h}.
2646 Values for host-side constants.
2649 Substitute for isatty, if not available.
2652 This is the longest integer type available on the host. If not defined,
2653 it will default to @code{long long} or @code{long}, depending on
2654 @code{CC_HAS_LONG_LONG}.
2656 @item CC_HAS_LONG_LONG
2657 @cindex @code{long long} data type
2658 Define this if the host C compiler supports @code{long long}. This is set
2659 by the @code{configure} script.
2661 @item PRINTF_HAS_LONG_LONG
2662 Define this if the host can handle printing of long long integers via
2663 the printf format conversion specifier @code{ll}. This is set by the
2664 @code{configure} script.
2666 @item HAVE_LONG_DOUBLE
2667 Define this if the host C compiler supports @code{long double}. This is
2668 set by the @code{configure} script.
2670 @item PRINTF_HAS_LONG_DOUBLE
2671 Define this if the host can handle printing of long double float-point
2672 numbers via the printf format conversion specifier @code{Lg}. This is
2673 set by the @code{configure} script.
2675 @item SCANF_HAS_LONG_DOUBLE
2676 Define this if the host can handle the parsing of long double
2677 float-point numbers via the scanf format conversion specifier
2678 @code{Lg}. This is set by the @code{configure} script.
2680 @item LSEEK_NOT_LINEAR
2681 Define this if @code{lseek (n)} does not necessarily move to byte number
2682 @code{n} in the file. This is only used when reading source files. It
2683 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2686 This macro is used as the argument to @code{lseek} (or, most commonly,
2687 @code{bfd_seek}). FIXME, should be replaced by SEEK_SET instead,
2688 which is the POSIX equivalent.
2691 If defined, this should be one or more tokens, such as @code{volatile},
2692 that can be used in both the declaration and definition of functions to
2693 indicate that they never return. The default is already set correctly
2694 if compiling with GCC. This will almost never need to be defined.
2697 If defined, this should be one or more tokens, such as
2698 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2699 of functions to indicate that they never return. The default is already
2700 set correctly if compiling with GCC. This will almost never need to be
2705 Define these to appropriate value for the system @code{lseek}, if not already
2709 This is the signal for stopping @value{GDBN}. Defaults to
2710 @code{SIGTSTP}. (Only redefined for the Convex.)
2713 Means that System V (prior to SVR4) include files are in use. (FIXME:
2714 This symbol is abused in @file{infrun.c}, @file{regex.c}, and
2715 @file{utils.c} for other things, at the moment.)
2718 Define this to help placate @code{lint} in some situations.
2721 Define this to override the defaults of @code{__volatile__} or
2726 @node Target Architecture Definition
2728 @chapter Target Architecture Definition
2730 @cindex target architecture definition
2731 @value{GDBN}'s target architecture defines what sort of
2732 machine-language programs @value{GDBN} can work with, and how it works
2735 The target architecture object is implemented as the C structure
2736 @code{struct gdbarch *}. The structure, and its methods, are generated
2737 using the Bourne shell script @file{gdbarch.sh}.
2740 * OS ABI Variant Handling::
2741 * Initialize New Architecture::
2742 * Registers and Memory::
2743 * Pointers and Addresses::
2745 * Raw and Virtual Registers::
2746 * Register and Memory Data::
2747 * Frame Interpretation::
2748 * Inferior Call Setup::
2749 * Compiler Characteristics::
2750 * Target Conditionals::
2751 * Adding a New Target::
2754 @node OS ABI Variant Handling
2755 @section Operating System ABI Variant Handling
2756 @cindex OS ABI variants
2758 @value{GDBN} provides a mechanism for handling variations in OS
2759 ABIs. An OS ABI variant may have influence over any number of
2760 variables in the target architecture definition. There are two major
2761 components in the OS ABI mechanism: sniffers and handlers.
2763 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2764 (the architecture may be wildcarded) in an attempt to determine the
2765 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2766 to be @dfn{generic}, while sniffers for a specific architecture are
2767 considered to be @dfn{specific}. A match from a specific sniffer
2768 overrides a match from a generic sniffer. Multiple sniffers for an
2769 architecture/flavour may exist, in order to differentiate between two
2770 different operating systems which use the same basic file format. The
2771 OS ABI framework provides a generic sniffer for ELF-format files which
2772 examines the @code{EI_OSABI} field of the ELF header, as well as note
2773 sections known to be used by several operating systems.
2775 @cindex fine-tuning @code{gdbarch} structure
2776 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2777 selected OS ABI. There may be only one handler for a given OS ABI
2778 for each BFD architecture.
2780 The following OS ABI variants are defined in @file{defs.h}:
2784 @findex GDB_OSABI_UNINITIALIZED
2785 @item GDB_OSABI_UNINITIALIZED
2786 Used for struct gdbarch_info if ABI is still uninitialized.
2788 @findex GDB_OSABI_UNKNOWN
2789 @item GDB_OSABI_UNKNOWN
2790 The ABI of the inferior is unknown. The default @code{gdbarch}
2791 settings for the architecture will be used.
2793 @findex GDB_OSABI_SVR4
2794 @item GDB_OSABI_SVR4
2795 UNIX System V Release 4.
2797 @findex GDB_OSABI_HURD
2798 @item GDB_OSABI_HURD
2799 GNU using the Hurd kernel.
2801 @findex GDB_OSABI_SOLARIS
2802 @item GDB_OSABI_SOLARIS
2805 @findex GDB_OSABI_OSF1
2806 @item GDB_OSABI_OSF1
2807 OSF/1, including Digital UNIX and Compaq Tru64 UNIX.
2809 @findex GDB_OSABI_LINUX
2810 @item GDB_OSABI_LINUX
2811 GNU using the Linux kernel.
2813 @findex GDB_OSABI_FREEBSD_AOUT
2814 @item GDB_OSABI_FREEBSD_AOUT
2815 FreeBSD using the @code{a.out} executable format.
2817 @findex GDB_OSABI_FREEBSD_ELF
2818 @item GDB_OSABI_FREEBSD_ELF
2819 FreeBSD using the ELF executable format.
2821 @findex GDB_OSABI_NETBSD_AOUT
2822 @item GDB_OSABI_NETBSD_AOUT
2823 NetBSD using the @code{a.out} executable format.
2825 @findex GDB_OSABI_NETBSD_ELF
2826 @item GDB_OSABI_NETBSD_ELF
2827 NetBSD using the ELF executable format.
2829 @findex GDB_OSABI_OPENBSD_ELF
2830 @item GDB_OSABI_OPENBSD_ELF
2831 OpenBSD using the ELF executable format.
2833 @findex GDB_OSABI_WINCE
2834 @item GDB_OSABI_WINCE
2837 @findex GDB_OSABI_GO32
2838 @item GDB_OSABI_GO32
2841 @findex GDB_OSABI_IRIX
2842 @item GDB_OSABI_IRIX
2845 @findex GDB_OSABI_INTERIX
2846 @item GDB_OSABI_INTERIX
2847 Interix (Posix layer for MS-Windows systems).
2849 @findex GDB_OSABI_HPUX_ELF
2850 @item GDB_OSABI_HPUX_ELF
2851 HP/UX using the ELF executable format.
2853 @findex GDB_OSABI_HPUX_SOM
2854 @item GDB_OSABI_HPUX_SOM
2855 HP/UX using the SOM executable format.
2857 @findex GDB_OSABI_QNXNTO
2858 @item GDB_OSABI_QNXNTO
2861 @findex GDB_OSABI_CYGWIN
2862 @item GDB_OSABI_CYGWIN
2865 @findex GDB_OSABI_AIX
2871 Here are the functions that make up the OS ABI framework:
2873 @deftypefun const char *gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2874 Return the name of the OS ABI corresponding to @var{osabi}.
2877 @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}))
2878 Register the OS ABI handler specified by @var{init_osabi} for the
2879 architecture, machine type and OS ABI specified by @var{arch},
2880 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2881 machine type, which implies the architecture's default machine type,
2885 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2886 Register the OS ABI file sniffer specified by @var{sniffer} for the
2887 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2888 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2889 be generic, and is allowed to examine @var{flavour}-flavoured files for
2893 @deftypefun enum gdb_osabi gdbarch_lookup_osabi (bfd *@var{abfd})
2894 Examine the file described by @var{abfd} to determine its OS ABI.
2895 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2899 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2900 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2901 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2902 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2903 architecture, a warning will be issued and the debugging session will continue
2904 with the defaults already established for @var{gdbarch}.
2907 @deftypefun void generic_elf_osabi_sniff_abi_tag_sections (bfd *@var{abfd}, asection *@var{sect}, void *@var{obj})
2908 Helper routine for ELF file sniffers. Examine the file described by
2909 @var{abfd} and look at ABI tag note sections to determine the OS ABI
2910 from the note. This function should be called via
2911 @code{bfd_map_over_sections}.
2914 @node Initialize New Architecture
2915 @section Initializing a New Architecture
2917 Each @code{gdbarch} is associated with a single @sc{bfd} architecture,
2918 via a @code{bfd_arch_@var{arch}} constant. The @code{gdbarch} is
2919 registered by a call to @code{register_gdbarch_init}, usually from
2920 the file's @code{_initialize_@var{filename}} routine, which will
2921 be automatically called during @value{GDBN} startup. The arguments
2922 are a @sc{bfd} architecture constant and an initialization function.
2924 The initialization function has this type:
2927 static struct gdbarch *
2928 @var{arch}_gdbarch_init (struct gdbarch_info @var{info},
2929 struct gdbarch_list *@var{arches})
2932 The @var{info} argument contains parameters used to select the correct
2933 architecture, and @var{arches} is a list of architectures which
2934 have already been created with the same @code{bfd_arch_@var{arch}}
2937 The initialization function should first make sure that @var{info}
2938 is acceptable, and return @code{NULL} if it is not. Then, it should
2939 search through @var{arches} for an exact match to @var{info}, and
2940 return one if found. Lastly, if no exact match was found, it should
2941 create a new architecture based on @var{info} and return it.
2943 Only information in @var{info} should be used to choose the new
2944 architecture. Historically, @var{info} could be sparse, and
2945 defaults would be collected from the first element on @var{arches}.
2946 However, @value{GDBN} now fills in @var{info} more thoroughly,
2947 so new @code{gdbarch} initialization functions should not take
2948 defaults from @var{arches}.
2950 @node Registers and Memory
2951 @section Registers and Memory
2953 @value{GDBN}'s model of the target machine is rather simple.
2954 @value{GDBN} assumes the machine includes a bank of registers and a
2955 block of memory. Each register may have a different size.
2957 @value{GDBN} does not have a magical way to match up with the
2958 compiler's idea of which registers are which; however, it is critical
2959 that they do match up accurately. The only way to make this work is
2960 to get accurate information about the order that the compiler uses,
2961 and to reflect that in the @code{gdbarch_register_name} and related functions.
2963 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
2965 @node Pointers and Addresses
2966 @section Pointers Are Not Always Addresses
2967 @cindex pointer representation
2968 @cindex address representation
2969 @cindex word-addressed machines
2970 @cindex separate data and code address spaces
2971 @cindex spaces, separate data and code address
2972 @cindex address spaces, separate data and code
2973 @cindex code pointers, word-addressed
2974 @cindex converting between pointers and addresses
2975 @cindex D10V addresses
2977 On almost all 32-bit architectures, the representation of a pointer is
2978 indistinguishable from the representation of some fixed-length number
2979 whose value is the byte address of the object pointed to. On such
2980 machines, the words ``pointer'' and ``address'' can be used interchangeably.
2981 However, architectures with smaller word sizes are often cramped for
2982 address space, so they may choose a pointer representation that breaks this
2983 identity, and allows a larger code address space.
2985 For example, the Renesas D10V is a 16-bit VLIW processor whose
2986 instructions are 32 bits long@footnote{Some D10V instructions are
2987 actually pairs of 16-bit sub-instructions. However, since you can't
2988 jump into the middle of such a pair, code addresses can only refer to
2989 full 32 bit instructions, which is what matters in this explanation.}.
2990 If the D10V used ordinary byte addresses to refer to code locations,
2991 then the processor would only be able to address 64kb of instructions.
2992 However, since instructions must be aligned on four-byte boundaries, the
2993 low two bits of any valid instruction's byte address are always
2994 zero---byte addresses waste two bits. So instead of byte addresses,
2995 the D10V uses word addresses---byte addresses shifted right two bits---to
2996 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
2999 However, this means that code pointers and data pointers have different
3000 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
3001 @code{0xC020} when used as a data address, but refers to byte address
3002 @code{0x30080} when used as a code address.
3004 (The D10V also uses separate code and data address spaces, which also
3005 affects the correspondence between pointers and addresses, but we're
3006 going to ignore that here; this example is already too long.)
3008 To cope with architectures like this---the D10V is not the only
3009 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
3010 byte numbers, and @dfn{pointers}, which are the target's representation
3011 of an address of a particular type of data. In the example above,
3012 @code{0xC020} is the pointer, which refers to one of the addresses
3013 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
3014 @value{GDBN} provides functions for turning a pointer into an address
3015 and vice versa, in the appropriate way for the current architecture.
3017 Unfortunately, since addresses and pointers are identical on almost all
3018 processors, this distinction tends to bit-rot pretty quickly. Thus,
3019 each time you port @value{GDBN} to an architecture which does
3020 distinguish between pointers and addresses, you'll probably need to
3021 clean up some architecture-independent code.
3023 Here are functions which convert between pointers and addresses:
3025 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
3026 Treat the bytes at @var{buf} as a pointer or reference of type
3027 @var{type}, and return the address it represents, in a manner
3028 appropriate for the current architecture. This yields an address
3029 @value{GDBN} can use to read target memory, disassemble, etc. Note that
3030 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
3033 For example, if the current architecture is the Intel x86, this function
3034 extracts a little-endian integer of the appropriate length from
3035 @var{buf} and returns it. However, if the current architecture is the
3036 D10V, this function will return a 16-bit integer extracted from
3037 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
3039 If @var{type} is not a pointer or reference type, then this function
3040 will signal an internal error.
3043 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
3044 Store the address @var{addr} in @var{buf}, in the proper format for a
3045 pointer of type @var{type} in the current architecture. Note that
3046 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
3049 For example, if the current architecture is the Intel x86, this function
3050 stores @var{addr} unmodified as a little-endian integer of the
3051 appropriate length in @var{buf}. However, if the current architecture
3052 is the D10V, this function divides @var{addr} by four if @var{type} is
3053 a pointer to a function, and then stores it in @var{buf}.
3055 If @var{type} is not a pointer or reference type, then this function
3056 will signal an internal error.
3059 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
3060 Assuming that @var{val} is a pointer, return the address it represents,
3061 as appropriate for the current architecture.
3063 This function actually works on integral values, as well as pointers.
3064 For pointers, it performs architecture-specific conversions as
3065 described above for @code{extract_typed_address}.
3068 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
3069 Create and return a value representing a pointer of type @var{type} to
3070 the address @var{addr}, as appropriate for the current architecture.
3071 This function performs architecture-specific conversions as described
3072 above for @code{store_typed_address}.
3075 Here are two functions which architectures can define to indicate the
3076 relationship between pointers and addresses. These have default
3077 definitions, appropriate for architectures on which all pointers are
3078 simple unsigned byte addresses.
3080 @deftypefun CORE_ADDR gdbarch_pointer_to_address (struct gdbarch *@var{current_gdbarch}, struct type *@var{type}, char *@var{buf})
3081 Assume that @var{buf} holds a pointer of type @var{type}, in the
3082 appropriate format for the current architecture. Return the byte
3083 address the pointer refers to.
3085 This function may safely assume that @var{type} is either a pointer or a
3086 C@t{++} reference type.
3089 @deftypefun void gdbarch_address_to_pointer (struct gdbarch *@var{current_gdbarch}, struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
3090 Store in @var{buf} a pointer of type @var{type} representing the address
3091 @var{addr}, in the appropriate format for the current architecture.
3093 This function may safely assume that @var{type} is either a pointer or a
3094 C@t{++} reference type.
3097 @node Address Classes
3098 @section Address Classes
3099 @cindex address classes
3100 @cindex DW_AT_byte_size
3101 @cindex DW_AT_address_class
3103 Sometimes information about different kinds of addresses is available
3104 via the debug information. For example, some programming environments
3105 define addresses of several different sizes. If the debug information
3106 distinguishes these kinds of address classes through either the size
3107 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
3108 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
3109 following macros should be defined in order to disambiguate these
3110 types within @value{GDBN} as well as provide the added information to
3111 a @value{GDBN} user when printing type expressions.
3113 @deftypefun int gdbarch_address_class_type_flags (struct gdbarch *@var{current_gdbarch}, int @var{byte_size}, int @var{dwarf2_addr_class})
3114 Returns the type flags needed to construct a pointer type whose size
3115 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
3116 This function is normally called from within a symbol reader. See
3117 @file{dwarf2read.c}.
3120 @deftypefun char *gdbarch_address_class_type_flags_to_name (struct gdbarch *@var{current_gdbarch}, int @var{type_flags})
3121 Given the type flags representing an address class qualifier, return
3124 @deftypefun int gdbarch_address_class_name_to_type_flags (struct gdbarch *@var{current_gdbarch}, int @var{name}, int *var{type_flags_ptr})
3125 Given an address qualifier name, set the @code{int} referenced by @var{type_flags_ptr} to the type flags
3126 for that address class qualifier.
3129 Since the need for address classes is rather rare, none of
3130 the address class functions are defined by default. Predicate
3131 functions are provided to detect when they are defined.
3133 Consider a hypothetical architecture in which addresses are normally
3134 32-bits wide, but 16-bit addresses are also supported. Furthermore,
3135 suppose that the @w{DWARF 2} information for this architecture simply
3136 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
3137 of these "short" pointers. The following functions could be defined
3138 to implement the address class functions:
3141 somearch_address_class_type_flags (int byte_size,
3142 int dwarf2_addr_class)
3145 return TYPE_FLAG_ADDRESS_CLASS_1;
3151 somearch_address_class_type_flags_to_name (int type_flags)
3153 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
3160 somearch_address_class_name_to_type_flags (char *name,
3161 int *type_flags_ptr)
3163 if (strcmp (name, "short") == 0)
3165 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
3173 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
3174 to indicate the presence of one of these "short" pointers. E.g, if
3175 the debug information indicates that @code{short_ptr_var} is one of these
3176 short pointers, @value{GDBN} might show the following behavior:
3179 (gdb) ptype short_ptr_var
3180 type = int * @@short
3184 @node Raw and Virtual Registers
3185 @section Raw and Virtual Register Representations
3186 @cindex raw register representation
3187 @cindex virtual register representation
3188 @cindex representations, raw and virtual registers
3190 @emph{Maintainer note: This section is pretty much obsolete. The
3191 functionality described here has largely been replaced by
3192 pseudo-registers and the mechanisms described in @ref{Target
3193 Architecture Definition, , Using Different Register and Memory Data
3194 Representations}. See also @uref{http://www.gnu.org/software/gdb/bugs/,
3195 Bug Tracking Database} and
3196 @uref{http://sources.redhat.com/gdb/current/ari/, ARI Index} for more
3197 up-to-date information.}
3199 Some architectures use one representation for a value when it lives in a
3200 register, but use a different representation when it lives in memory.
3201 In @value{GDBN}'s terminology, the @dfn{raw} representation is the one used in
3202 the target registers, and the @dfn{virtual} representation is the one
3203 used in memory, and within @value{GDBN} @code{struct value} objects.
3205 @emph{Maintainer note: Notice that the same mechanism is being used to
3206 both convert a register to a @code{struct value} and alternative
3209 For almost all data types on almost all architectures, the virtual and
3210 raw representations are identical, and no special handling is needed.
3211 However, they do occasionally differ. For example:
3215 The x86 architecture supports an 80-bit @code{long double} type. However, when
3216 we store those values in memory, they occupy twelve bytes: the
3217 floating-point number occupies the first ten, and the final two bytes
3218 are unused. This keeps the values aligned on four-byte boundaries,
3219 allowing more efficient access. Thus, the x86 80-bit floating-point
3220 type is the raw representation, and the twelve-byte loosely-packed
3221 arrangement is the virtual representation.
3224 Some 64-bit MIPS targets present 32-bit registers to @value{GDBN} as 64-bit
3225 registers, with garbage in their upper bits. @value{GDBN} ignores the top 32
3226 bits. Thus, the 64-bit form, with garbage in the upper 32 bits, is the
3227 raw representation, and the trimmed 32-bit representation is the
3228 virtual representation.
3231 In general, the raw representation is determined by the architecture, or
3232 @value{GDBN}'s interface to the architecture, while the virtual representation
3233 can be chosen for @value{GDBN}'s convenience. @value{GDBN}'s register file,
3234 @code{registers}, holds the register contents in raw format, and the
3235 @value{GDBN} remote protocol transmits register values in raw format.
3237 Your architecture may define the following macros to request
3238 conversions between the raw and virtual format:
3240 @deftypefn {Target Macro} int REGISTER_CONVERTIBLE (int @var{reg})
3241 Return non-zero if register number @var{reg}'s value needs different raw
3242 and virtual formats.
3244 You should not use @code{REGISTER_CONVERT_TO_VIRTUAL} for a register
3245 unless this macro returns a non-zero value for that register.
3248 @deftypefn {Target Macro} int DEPRECATED_REGISTER_RAW_SIZE (int @var{reg})
3249 The size of register number @var{reg}'s raw value. This is the number
3250 of bytes the register will occupy in @code{registers}, or in a @value{GDBN}
3251 remote protocol packet.
3254 @deftypefn {Target Macro} int DEPRECATED_REGISTER_VIRTUAL_SIZE (int @var{reg})
3255 The size of register number @var{reg}'s value, in its virtual format.
3256 This is the size a @code{struct value}'s buffer will have, holding that
3260 @deftypefn {Target Macro} struct type *DEPRECATED_REGISTER_VIRTUAL_TYPE (int @var{reg})
3261 This is the type of the virtual representation of register number
3262 @var{reg}. Note that there is no need for a macro giving a type for the
3263 register's raw form; once the register's value has been obtained, @value{GDBN}
3264 always uses the virtual form.
3267 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_VIRTUAL (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
3268 Convert the value of register number @var{reg} to @var{type}, which
3269 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
3270 at @var{from} holds the register's value in raw format; the macro should
3271 convert the value to virtual format, and place it at @var{to}.
3273 Note that @code{REGISTER_CONVERT_TO_VIRTUAL} and
3274 @code{REGISTER_CONVERT_TO_RAW} take their @var{reg} and @var{type}
3275 arguments in different orders.
3277 You should only use @code{REGISTER_CONVERT_TO_VIRTUAL} with registers
3278 for which the @code{REGISTER_CONVERTIBLE} macro returns a non-zero
3282 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_RAW (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3283 Convert the value of register number @var{reg} to @var{type}, which
3284 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
3285 at @var{from} holds the register's value in raw format; the macro should
3286 convert the value to virtual format, and place it at @var{to}.
3288 Note that REGISTER_CONVERT_TO_VIRTUAL and REGISTER_CONVERT_TO_RAW take
3289 their @var{reg} and @var{type} arguments in different orders.
3293 @node Register and Memory Data
3294 @section Using Different Register and Memory Data Representations
3295 @cindex register representation
3296 @cindex memory representation
3297 @cindex representations, register and memory
3298 @cindex register data formats, converting
3299 @cindex @code{struct value}, converting register contents to
3301 @emph{Maintainer's note: The way GDB manipulates registers is undergoing
3302 significant change. Many of the macros and functions referred to in this
3303 section are likely to be subject to further revision. See
3304 @uref{http://sources.redhat.com/gdb/current/ari/, A.R. Index} and
3305 @uref{http://www.gnu.org/software/gdb/bugs, Bug Tracking Database} for
3306 further information. cagney/2002-05-06.}
3308 Some architectures can represent a data object in a register using a
3309 form that is different to the objects more normal memory representation.
3315 The Alpha architecture can represent 32 bit integer values in
3316 floating-point registers.
3319 The x86 architecture supports 80-bit floating-point registers. The
3320 @code{long double} data type occupies 96 bits in memory but only 80 bits
3321 when stored in a register.
3325 In general, the register representation of a data type is determined by
3326 the architecture, or @value{GDBN}'s interface to the architecture, while
3327 the memory representation is determined by the Application Binary
3330 For almost all data types on almost all architectures, the two
3331 representations are identical, and no special handling is needed.
3332 However, they do occasionally differ. Your architecture may define the
3333 following macros to request conversions between the register and memory
3334 representations of a data type:
3336 @deftypefun int gdbarch_convert_register_p (struct gdbarch *@var{gdbarch}, int @var{reg})
3337 Return non-zero if the representation of a data value stored in this
3338 register may be different to the representation of that same data value
3339 when stored in memory.
3341 When non-zero, the macros @code{gdbarch_register_to_value} and
3342 @code{value_to_register} are used to perform any necessary conversion.
3344 This function should return zero for the register's native type, when
3345 no conversion is necessary.
3348 @deftypefun void gdbarch_register_to_value (struct gdbarch *@var{gdbarch}, int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
3349 Convert the value of register number @var{reg} to a data object of type
3350 @var{type}. The buffer at @var{from} holds the register's value in raw
3351 format; the converted value should be placed in the buffer at @var{to}.
3353 Note that @code{gdbarch_register_to_value} and @code{gdbarch_value_to_register}
3354 take their @var{reg} and @var{type} arguments in different orders.
3356 You should only use @code{gdbarch_register_to_value} with registers for which
3357 the @code{gdbarch_convert_register_p} function returns a non-zero value.
3360 @deftypefun void gdbarch_value_to_register (struct gdbarch *@var{gdbarch}, struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3361 Convert a data value of type @var{type} to register number @var{reg}'
3364 Note that @code{gdbarch_register_to_value} and @code{gdbarch_value_to_register}
3365 take their @var{reg} and @var{type} arguments in different orders.
3367 You should only use @code{gdbarch_value_to_register} with registers for which
3368 the @code{gdbarch_convert_register_p} function returns a non-zero value.
3371 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_TYPE (int @var{regnum}, struct type *@var{type}, char *@var{buf})
3372 See @file{mips-tdep.c}. It does not do what you want.
3375 @node Frame Interpretation
3376 @section Frame Interpretation
3378 @node Inferior Call Setup
3379 @section Inferior Call Setup
3381 @node Compiler Characteristics
3382 @section Compiler Characteristics
3384 @node Target Conditionals
3385 @section Target Conditionals
3387 This section describes the macros and functions that you can use to define the
3392 @item CORE_ADDR gdbarch_addr_bits_remove (@var{gdbarch}, @var{addr})
3393 @findex gdbarch_addr_bits_remove
3394 If a raw machine instruction address includes any bits that are not
3395 really part of the address, then this function is used to zero those bits in
3396 @var{addr}. This is only used for addresses of instructions, and even then not
3399 For example, the two low-order bits of the PC on the Hewlett-Packard PA
3400 2.0 architecture contain the privilege level of the corresponding
3401 instruction. Since instructions must always be aligned on four-byte
3402 boundaries, the processor masks out these bits to generate the actual
3403 address of the instruction. @code{gdbarch_addr_bits_remove} would then for
3404 example look like that:
3406 arch_addr_bits_remove (CORE_ADDR addr)
3408 return (addr &= ~0x3);
3412 @item int address_class_name_to_type_flags (@var{gdbarch}, @var{name}, @var{type_flags_ptr})
3413 @findex address_class_name_to_type_flags
3414 If @var{name} is a valid address class qualifier name, set the @code{int}
3415 referenced by @var{type_flags_ptr} to the mask representing the qualifier
3416 and return 1. If @var{name} is not a valid address class qualifier name,
3419 The value for @var{type_flags_ptr} should be one of
3420 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
3421 possibly some combination of these values or'd together.
3422 @xref{Target Architecture Definition, , Address Classes}.
3424 @item int address_class_name_to_type_flags_p (@var{gdbarch})
3425 @findex address_class_name_to_type_flags_p
3426 Predicate which indicates whether @code{address_class_name_to_type_flags}
3429 @item int gdbarch_address_class_type_flags (@var{gdbarch}, @var{byte_size}, @var{dwarf2_addr_class})
3430 @findex gdbarch_address_class_type_flags
3431 Given a pointers byte size (as described by the debug information) and
3432 the possible @code{DW_AT_address_class} value, return the type flags
3433 used by @value{GDBN} to represent this address class. The value
3434 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
3435 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
3436 values or'd together.
3437 @xref{Target Architecture Definition, , Address Classes}.
3439 @item int gdbarch_address_class_type_flags_p (@var{gdbarch})
3440 @findex gdbarch_address_class_type_flags_p
3441 Predicate which indicates whether @code{gdbarch_address_class_type_flags_p} has
3444 @item const char *gdbarch_address_class_type_flags_to_name (@var{gdbarch}, @var{type_flags})
3445 @findex gdbarch_address_class_type_flags_to_name
3446 Return the name of the address class qualifier associated with the type
3447 flags given by @var{type_flags}.
3449 @item int gdbarch_address_class_type_flags_to_name_p (@var{gdbarch})
3450 @findex gdbarch_address_class_type_flags_to_name_p
3451 Predicate which indicates whether @code{gdbarch_address_class_type_flags_to_name} has been defined.
3452 @xref{Target Architecture Definition, , Address Classes}.
3454 @item void gdbarch_address_to_pointer (@var{gdbarch}, @var{type}, @var{buf}, @var{addr})
3455 @findex gdbarch_address_to_pointer
3456 Store in @var{buf} a pointer of type @var{type} representing the address
3457 @var{addr}, in the appropriate format for the current architecture.
3458 This function may safely assume that @var{type} is either a pointer or a
3459 C@t{++} reference type.
3460 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3462 @item int gdbarch_believe_pcc_promotion (@var{gdbarch})
3463 @findex gdbarch_believe_pcc_promotion
3464 Used to notify if the compiler promotes a @code{short} or @code{char}
3465 parameter to an @code{int}, but still reports the parameter as its
3466 original type, rather than the promoted type.
3468 @item gdbarch_bits_big_endian (@var{gdbarch})
3469 @findex gdbarch_bits_big_endian
3470 This is used if the numbering of bits in the targets does @strong{not} match
3471 the endianness of the target byte order. A value of 1 means that the bits
3472 are numbered in a big-endian bit order, 0 means little-endian.
3474 @item set_gdbarch_bits_big_endian (@var{gdbarch}, @var{bits_big_endian})
3475 @findex set_gdbarch_bits_big_endian
3476 Calling set_gdbarch_bits_big_endian with a value of 1 indicates that the
3477 bits in the target are numbered in a big-endian bit order, 0 indicates
3482 This is the character array initializer for the bit pattern to put into
3483 memory where a breakpoint is set. Although it's common to use a trap
3484 instruction for a breakpoint, it's not required; for instance, the bit
3485 pattern could be an invalid instruction. The breakpoint must be no
3486 longer than the shortest instruction of the architecture.
3488 @code{BREAKPOINT} has been deprecated in favor of
3489 @code{gdbarch_breakpoint_from_pc}.
3491 @item BIG_BREAKPOINT
3492 @itemx LITTLE_BREAKPOINT
3493 @findex LITTLE_BREAKPOINT
3494 @findex BIG_BREAKPOINT
3495 Similar to BREAKPOINT, but used for bi-endian targets.
3497 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
3498 favor of @code{gdbarch_breakpoint_from_pc}.
3500 @item const gdb_byte *gdbarch_breakpoint_from_pc (@var{gdbarch}, @var{pcptr}, @var{lenptr})
3501 @findex gdbarch_breakpoint_from_pc
3502 @anchor{gdbarch_breakpoint_from_pc} Use the program counter to determine the
3503 contents and size of a breakpoint instruction. It returns a pointer to
3504 a string of bytes that encode a breakpoint instruction, stores the
3505 length of the string to @code{*@var{lenptr}}, and adjusts the program
3506 counter (if necessary) to point to the actual memory location where the
3507 breakpoint should be inserted.
3509 Although it is common to use a trap instruction for a breakpoint, it's
3510 not required; for instance, the bit pattern could be an invalid
3511 instruction. The breakpoint must be no longer than the shortest
3512 instruction of the architecture.
3514 Replaces all the other @var{BREAKPOINT} macros.
3516 @item int gdbarch_memory_insert_breakpoint (@var{gdbarch}, @var{bp_tgt})
3517 @itemx gdbarch_memory_remove_breakpoint (@var{gdbarch}, @var{bp_tgt})
3518 @findex gdbarch_memory_remove_breakpoint
3519 @findex gdbarch_memory_insert_breakpoint
3520 Insert or remove memory based breakpoints. Reasonable defaults
3521 (@code{default_memory_insert_breakpoint} and
3522 @code{default_memory_remove_breakpoint} respectively) have been
3523 provided so that it is not necessary to set these for most
3524 architectures. Architectures which may want to set
3525 @code{gdbarch_memory_insert_breakpoint} and @code{gdbarch_memory_remove_breakpoint} will likely have instructions that are oddly sized or are not stored in a
3526 conventional manner.
3528 It may also be desirable (from an efficiency standpoint) to define
3529 custom breakpoint insertion and removal routines if
3530 @code{gdbarch_breakpoint_from_pc} needs to read the target's memory for some
3533 @item CORE_ADDR gdbarch_adjust_breakpoint_address (@var{gdbarch}, @var{bpaddr})
3534 @findex gdbarch_adjust_breakpoint_address
3535 @cindex breakpoint address adjusted
3536 Given an address at which a breakpoint is desired, return a breakpoint
3537 address adjusted to account for architectural constraints on
3538 breakpoint placement. This method is not needed by most targets.
3540 The FR-V target (see @file{frv-tdep.c}) requires this method.
3541 The FR-V is a VLIW architecture in which a number of RISC-like
3542 instructions are grouped (packed) together into an aggregate
3543 instruction or instruction bundle. When the processor executes
3544 one of these bundles, the component instructions are executed
3547 In the course of optimization, the compiler may group instructions
3548 from distinct source statements into the same bundle. The line number
3549 information associated with one of the latter statements will likely
3550 refer to some instruction other than the first one in the bundle. So,
3551 if the user attempts to place a breakpoint on one of these latter
3552 statements, @value{GDBN} must be careful to @emph{not} place the break
3553 instruction on any instruction other than the first one in the bundle.
3554 (Remember though that the instructions within a bundle execute
3555 in parallel, so the @emph{first} instruction is the instruction
3556 at the lowest address and has nothing to do with execution order.)
3558 The FR-V's @code{gdbarch_adjust_breakpoint_address} method will adjust a
3559 breakpoint's address by scanning backwards for the beginning of
3560 the bundle, returning the address of the bundle.
3562 Since the adjustment of a breakpoint may significantly alter a user's
3563 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
3564 is initially set and each time that that breakpoint is hit.
3566 @item int gdbarch_call_dummy_location (@var{gdbarch})
3567 @findex gdbarch_call_dummy_location
3568 See the file @file{inferior.h}.
3570 This method has been replaced by @code{gdbarch_push_dummy_code}
3571 (@pxref{gdbarch_push_dummy_code}).
3573 @item int gdbarch_cannot_fetch_register (@var{gdbarch}, @var{regum})
3574 @findex gdbarch_cannot_fetch_register
3575 This function should return nonzero if @var{regno} cannot be fetched
3576 from an inferior process. This is only relevant if
3577 @code{FETCH_INFERIOR_REGISTERS} is not defined.
3579 @item int gdbarch_cannot_store_register (@var{gdbarch}, @var{regnum})
3580 @findex gdbarch_cannot_store_register
3581 This function should return nonzero if @var{regno} should not be
3582 written to the target. This is often the case for program counters,
3583 status words, and other special registers. This function returns 0 as
3584 default so that @value{GDBN} will assume that all registers may be written.
3586 @item int gdbarch_convert_register_p (@var{gdbarch}, @var{regnum}, struct type *@var{type})
3587 @findex gdbarch_convert_register_p
3588 Return non-zero if register @var{regnum} represents data values of type
3589 @var{type} in a non-standard form.
3590 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3592 @item CORE_ADDR gdbarch_decr_pc_after_break (@var{gdbarch})
3593 @findex gdbarch_decr_pc_after_break
3594 This function shall return the amount by which to decrement the PC after the
3595 program encounters a breakpoint. This is often the number of bytes in
3596 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
3598 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
3599 @findex DISABLE_UNSETTABLE_BREAK
3600 If defined, this should evaluate to 1 if @var{addr} is in a shared
3601 library in which breakpoints cannot be set and so should be disabled.
3603 @item void gdbarch_print_float_info (@var{gdbarch}, @var{file}, @var{frame}, @var{args})
3604 @findex gdbarch_print_float_info
3605 If defined, then the @samp{info float} command will print information about
3606 the processor's floating point unit.
3608 @item void gdbarch_print_registers_info (@var{gdbarch}, @var{frame}, @var{regnum}, @var{all})
3609 @findex gdbarch_print_registers_info
3610 If defined, pretty print the value of the register @var{regnum} for the
3611 specified @var{frame}. If the value of @var{regnum} is -1, pretty print
3612 either all registers (@var{all} is non zero) or a select subset of
3613 registers (@var{all} is zero).
3615 The default method prints one register per line, and if @var{all} is
3616 zero omits floating-point registers.
3618 @item int gdbarch_print_vector_info (@var{gdbarch}, @var{file}, @var{frame}, @var{args})
3619 @findex gdbarch_print_vector_info
3620 If defined, then the @samp{info vector} command will call this function
3621 to print information about the processor's vector unit.
3623 By default, the @samp{info vector} command will print all vector
3624 registers (the register's type having the vector attribute).
3626 @item int gdbarch_dwarf2_reg_to_regnum (@var{gdbarch}, @var{dwarf2_regnr})
3627 @findex gdbarch_dwarf2_reg_to_regnum
3628 Convert DWARF2 register number @var{dwarf2_regnr} into @value{GDBN} regnum.
3629 If not defined, no conversion will be performed.
3631 @item int gdbarch_ecoff_reg_to_regnum (@var{gdbarch}, @var{ecoff_regnr})
3632 @findex gdbarch_ecoff_reg_to_regnum
3633 Convert ECOFF register number @var{ecoff_regnr} into @value{GDBN} regnum. If
3634 not defined, no conversion will be performed.
3636 @item DEPRECATED_FP_REGNUM
3637 @findex DEPRECATED_FP_REGNUM
3638 If the virtual frame pointer is kept in a register, then define this
3639 macro to be the number (greater than or equal to zero) of that register.
3641 This should only need to be defined if @code{DEPRECATED_TARGET_READ_FP}
3644 @item DEPRECATED_FRAMELESS_FUNCTION_INVOCATION(@var{fi})
3645 @findex DEPRECATED_FRAMELESS_FUNCTION_INVOCATION
3646 Define this to an expression that returns 1 if the function invocation
3647 represented by @var{fi} does not have a stack frame associated with it.
3650 @item CORE_ADDR frame_align (@var{gdbarch}, @var{address})
3651 @anchor{frame_align}
3653 Define this to adjust @var{address} so that it meets the alignment
3654 requirements for the start of a new stack frame. A stack frame's
3655 alignment requirements are typically stronger than a target processors
3656 stack alignment requirements.
3658 This function is used to ensure that, when creating a dummy frame, both
3659 the initial stack pointer and (if needed) the address of the return
3660 value are correctly aligned.
3662 This function always adjusts the address in the direction of stack
3665 By default, no frame based stack alignment is performed.
3667 @item int gdbarch_frame_red_zone_size (@var{gdbarch})
3668 @findex gdbarch_frame_red_zone_size
3669 The number of bytes, beyond the innermost-stack-address, reserved by the
3670 @sc{abi}. A function is permitted to use this scratch area (instead of
3671 allocating extra stack space).
3673 When performing an inferior function call, to ensure that it does not
3674 modify this area, @value{GDBN} adjusts the innermost-stack-address by
3675 @var{gdbarch_frame_red_zone_size} bytes before pushing parameters onto the
3678 By default, zero bytes are allocated. The value must be aligned
3679 (@pxref{frame_align}).
3681 The @sc{amd64} (nee x86-64) @sc{abi} documentation refers to the
3682 @emph{red zone} when describing this scratch area.
3685 @item DEPRECATED_FRAME_CHAIN(@var{frame})
3686 @findex DEPRECATED_FRAME_CHAIN
3687 Given @var{frame}, return a pointer to the calling frame.
3689 @item DEPRECATED_FRAME_CHAIN_VALID(@var{chain}, @var{thisframe})
3690 @findex DEPRECATED_FRAME_CHAIN_VALID
3691 Define this to be an expression that returns zero if the given frame is an
3692 outermost frame, with no caller, and nonzero otherwise. Most normal
3693 situations can be handled without defining this macro, including @code{NULL}
3694 chain pointers, dummy frames, and frames whose PC values are inside the
3695 startup file (e.g.@: @file{crt0.o}), inside @code{main}, or inside
3698 @item DEPRECATED_FRAME_INIT_SAVED_REGS(@var{frame})
3699 @findex DEPRECATED_FRAME_INIT_SAVED_REGS
3700 See @file{frame.h}. Determines the address of all registers in the
3701 current stack frame storing each in @code{frame->saved_regs}. Space for
3702 @code{frame->saved_regs} shall be allocated by
3703 @code{DEPRECATED_FRAME_INIT_SAVED_REGS} using
3704 @code{frame_saved_regs_zalloc}.
3706 @code{FRAME_FIND_SAVED_REGS} is deprecated.
3708 @item int gdbarch_frame_num_args (@var{gdbarch}, @var{frame})
3709 @findex gdbarch_frame_num_args
3710 For the frame described by @var{frame} return the number of arguments that
3711 are being passed. If the number of arguments is not known, return
3714 @item DEPRECATED_FRAME_SAVED_PC(@var{frame})
3715 @findex DEPRECATED_FRAME_SAVED_PC
3716 @anchor{DEPRECATED_FRAME_SAVED_PC} Given @var{frame}, return the pc
3717 saved there. This is the return address.
3719 This method is deprecated. @xref{gdbarch_unwind_pc}.
3721 @item CORE_ADDR gdbarch_unwind_pc (@var{next_frame})
3722 @findex gdbarch_unwind_pc
3723 @anchor{gdbarch_unwind_pc} Return the instruction address, in
3724 @var{next_frame}'s caller, at which execution will resume after
3725 @var{next_frame} returns. This is commonly referred to as the return address.
3727 The implementation, which must be frame agnostic (work with any frame),
3728 is typically no more than:
3732 pc = frame_unwind_register_unsigned (next_frame, S390_PC_REGNUM);
3733 return gdbarch_addr_bits_remove (gdbarch, pc);
3737 @xref{DEPRECATED_FRAME_SAVED_PC}, which this method replaces.
3739 @item CORE_ADDR gdbarch_unwind_sp (@var{gdbarch}, @var{next_frame})
3740 @findex gdbarch_unwind_sp
3741 @anchor{gdbarch_unwind_sp} Return the frame's inner most stack address. This is
3742 commonly referred to as the frame's @dfn{stack pointer}.
3744 The implementation, which must be frame agnostic (work with any frame),
3745 is typically no more than:
3749 sp = frame_unwind_register_unsigned (next_frame, S390_SP_REGNUM);
3750 return gdbarch_addr_bits_remove (gdbarch, sp);
3754 @xref{TARGET_READ_SP}, which this method replaces.
3756 @item FUNCTION_EPILOGUE_SIZE
3757 @findex FUNCTION_EPILOGUE_SIZE
3758 For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the
3759 function end symbol is 0. For such targets, you must define
3760 @code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size of a
3761 function's epilogue.
3763 @item DEPRECATED_FUNCTION_START_OFFSET
3764 @findex DEPRECATED_FUNCTION_START_OFFSET
3765 An integer, giving the offset in bytes from a function's address (as
3766 used in the values of symbols, function pointers, etc.), and the
3767 function's first genuine instruction.
3769 This is zero on almost all machines: the function's address is usually
3770 the address of its first instruction. However, on the VAX, for
3771 example, each function starts with two bytes containing a bitmask
3772 indicating which registers to save upon entry to the function. The
3773 VAX @code{call} instructions check this value, and save the
3774 appropriate registers automatically. Thus, since the offset from the
3775 function's address to its first instruction is two bytes,
3776 @code{DEPRECATED_FUNCTION_START_OFFSET} would be 2 on the VAX.
3778 @item GCC_COMPILED_FLAG_SYMBOL
3779 @itemx GCC2_COMPILED_FLAG_SYMBOL
3780 @findex GCC2_COMPILED_FLAG_SYMBOL
3781 @findex GCC_COMPILED_FLAG_SYMBOL
3782 If defined, these are the names of the symbols that @value{GDBN} will
3783 look for to detect that GCC compiled the file. The default symbols
3784 are @code{gcc_compiled.} and @code{gcc2_compiled.},
3785 respectively. (Currently only defined for the Delta 68.)
3787 @item gdbarch_get_longjmp_target
3788 @findex gdbarch_get_longjmp_target
3789 For most machines, this is a target-dependent parameter. On the
3790 DECstation and the Iris, this is a native-dependent parameter, since
3791 the header file @file{setjmp.h} is needed to define it.
3793 This macro determines the target PC address that @code{longjmp} will jump to,
3794 assuming that we have just stopped at a @code{longjmp} breakpoint. It takes a
3795 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3796 pointer. It examines the current state of the machine as needed.
3798 @item DEPRECATED_IBM6000_TARGET
3799 @findex DEPRECATED_IBM6000_TARGET
3800 Shows that we are configured for an IBM RS/6000 system. This
3801 conditional should be eliminated (FIXME) and replaced by
3802 feature-specific macros. It was introduced in a haste and we are
3803 repenting at leisure.
3805 @item I386_USE_GENERIC_WATCHPOINTS
3806 An x86-based target can define this to use the generic x86 watchpoint
3807 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3809 @item int gdbarch_inner_than (@var{gdbarch}, @var{lhs}, @var{rhs})
3810 @findex gdbarch_inner_than
3811 Returns non-zero if stack address @var{lhs} is inner than (nearer to the
3812 stack top) stack address @var{rhs}. Let the function return
3813 @w{@code{lhs < rhs}} if the target's stack grows downward in memory, or
3814 @w{@code{lhs > rsh}} if the stack grows upward.
3816 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{addr})
3817 @findex gdbarch_in_function_epilogue_p
3818 Returns non-zero if the given @var{addr} is in the epilogue of a function.
3819 The epilogue of a function is defined as the part of a function where
3820 the stack frame of the function already has been destroyed up to the
3821 final `return from function call' instruction.
3823 @item int gdbarch_in_solib_return_trampoline (@var{gdbarch}, @var{pc}, @var{name})
3824 @findex gdbarch_in_solib_return_trampoline
3825 Define this function to return nonzero if the program is stopped in the
3826 trampoline that returns from a shared library.
3828 @item IN_SOLIB_DYNSYM_RESOLVE_CODE (@var{pc})
3829 @findex IN_SOLIB_DYNSYM_RESOLVE_CODE
3830 Define this to return nonzero if the program is stopped in the
3833 @item SKIP_SOLIB_RESOLVER (@var{pc})
3834 @findex SKIP_SOLIB_RESOLVER
3835 Define this to evaluate to the (nonzero) address at which execution
3836 should continue to get past the dynamic linker's symbol resolution
3837 function. A zero value indicates that it is not important or necessary
3838 to set a breakpoint to get through the dynamic linker and that single
3839 stepping will suffice.
3841 @item CORE_ADDR gdbarch_integer_to_address (@var{gdbarch}, @var{type}, @var{buf})
3842 @findex gdbarch_integer_to_address
3843 @cindex converting integers to addresses
3844 Define this when the architecture needs to handle non-pointer to address
3845 conversions specially. Converts that value to an address according to
3846 the current architectures conventions.
3848 @emph{Pragmatics: When the user copies a well defined expression from
3849 their source code and passes it, as a parameter, to @value{GDBN}'s
3850 @code{print} command, they should get the same value as would have been
3851 computed by the target program. Any deviation from this rule can cause
3852 major confusion and annoyance, and needs to be justified carefully. In
3853 other words, @value{GDBN} doesn't really have the freedom to do these
3854 conversions in clever and useful ways. It has, however, been pointed
3855 out that users aren't complaining about how @value{GDBN} casts integers
3856 to pointers; they are complaining that they can't take an address from a
3857 disassembly listing and give it to @code{x/i}. Adding an architecture
3858 method like @code{gdbarch_integer_to_address} certainly makes it possible for
3859 @value{GDBN} to ``get it right'' in all circumstances.}
3861 @xref{Target Architecture Definition, , Pointers Are Not Always
3864 @item CORE_ADDR gdbarch_pointer_to_address (@var{gdbarch}, @var{type}, @var{buf})
3865 @findex gdbarch_pointer_to_address
3866 Assume that @var{buf} holds a pointer of type @var{type}, in the
3867 appropriate format for the current architecture. Return the byte
3868 address the pointer refers to.
3869 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3871 @item void gdbarch_register_to_value(@var{gdbarch}, @var{frame}, @var{regnum}, @var{type}, @var{fur})
3872 @findex gdbarch_register_to_value
3873 Convert the raw contents of register @var{regnum} into a value of type
3875 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3877 @item register_reggroup_p (@var{gdbarch}, @var{regnum}, @var{reggroup})
3878 @findex register_reggroup_p
3879 @cindex register groups
3880 Return non-zero if register @var{regnum} is a member of the register
3881 group @var{reggroup}.
3883 By default, registers are grouped as follows:
3886 @item float_reggroup
3887 Any register with a valid name and a floating-point type.
3888 @item vector_reggroup
3889 Any register with a valid name and a vector type.
3890 @item general_reggroup
3891 Any register with a valid name and a type other than vector or
3892 floating-point. @samp{float_reggroup}.
3894 @itemx restore_reggroup
3896 Any register with a valid name.
3899 @item DEPRECATED_REGISTER_VIRTUAL_SIZE (@var{reg})
3900 @findex DEPRECATED_REGISTER_VIRTUAL_SIZE
3901 Return the virtual size of @var{reg}; defaults to the size of the
3902 register's virtual type.
3903 Return the virtual size of @var{reg}.
3904 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3906 @item DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})
3907 @findex REGISTER_VIRTUAL_TYPE
3908 Return the virtual type of @var{reg}.
3909 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3911 @item struct type *register_type (@var{gdbarch}, @var{reg})
3912 @findex register_type
3913 If defined, return the type of register @var{reg}. This function
3914 supersedes @code{DEPRECATED_REGISTER_VIRTUAL_TYPE}. @xref{Target Architecture
3915 Definition, , Raw and Virtual Register Representations}.
3917 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
3918 @findex REGISTER_CONVERT_TO_VIRTUAL
3919 Convert the value of register @var{reg} from its raw form to its virtual
3921 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3923 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
3924 @findex REGISTER_CONVERT_TO_RAW
3925 Convert the value of register @var{reg} from its virtual form to its raw
3927 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3929 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
3930 @findex regset_from_core_section
3931 Return the appropriate register set for a core file section with name
3932 @var{sect_name} and size @var{sect_size}.
3934 @item SOFTWARE_SINGLE_STEP_P()
3935 @findex SOFTWARE_SINGLE_STEP_P
3936 Define this as 1 if the target does not have a hardware single-step
3937 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
3939 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breakpoints_p})
3940 @findex SOFTWARE_SINGLE_STEP
3941 A function that inserts or removes (depending on
3942 @var{insert_breakpoints_p}) breakpoints at each possible destinations of
3943 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
3946 @item set_gdbarch_sofun_address_maybe_missing (@var{gdbarch}, @var{set})
3947 @findex set_gdbarch_sofun_address_maybe_missing
3948 Somebody clever observed that, the more actual addresses you have in the
3949 debug information, the more time the linker has to spend relocating
3950 them. So whenever there's some other way the debugger could find the
3951 address it needs, you should omit it from the debug info, to make
3954 Calling @code{set_gdbarch_sofun_address_maybe_missing} with a non-zero
3955 argument @var{set} indicates that a particular set of hacks of this sort
3956 are in use, affecting @code{N_SO} and @code{N_FUN} entries in stabs-format
3957 debugging information. @code{N_SO} stabs mark the beginning and ending
3958 addresses of compilation units in the text segment. @code{N_FUN} stabs
3959 mark the starts and ends of functions.
3961 In this case, @value{GDBN} assumes two things:
3965 @code{N_FUN} stabs have an address of zero. Instead of using those
3966 addresses, you should find the address where the function starts by
3967 taking the function name from the stab, and then looking that up in the
3968 minsyms (the linker/assembler symbol table). In other words, the stab
3969 has the name, and the linker/assembler symbol table is the only place
3970 that carries the address.
3973 @code{N_SO} stabs have an address of zero, too. You just look at the
3974 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab, and
3975 guess the starting and ending addresses of the compilation unit from them.
3978 @item int gdbarch_pc_regnum (@var{gdbarch})
3979 @findex gdbarch_pc_regnum
3980 If the program counter is kept in a register, then let this function return
3981 the number (greater than or equal to zero) of that register.
3983 This should only need to be defined if @code{gdbarch_read_pc} and
3984 @code{gdbarch_write_pc} are not defined.
3986 @item int gdbarch_stabs_argument_has_addr (@var{gdbarch}, @var{type})
3987 @findex gdbarch_stabs_argument_has_addr
3988 @anchor{gdbarch_stabs_argument_has_addr} Define this function to return
3989 nonzero if a function argument of type @var{type} is passed by reference
3992 @item PROCESS_LINENUMBER_HOOK
3993 @findex PROCESS_LINENUMBER_HOOK
3994 A hook defined for XCOFF reading.
3996 @item gdbarch_ps_regnum (@var{gdbarch}
3997 @findex gdbarch_ps_regnum
3998 If defined, this function returns the number of the processor status
4000 (This definition is only used in generic code when parsing "$ps".)
4002 @item CORE_ADDR gdbarch_push_dummy_call (@var{gdbarch}, @var{function}, @var{regcache}, @var{bp_addr}, @var{nargs}, @var{args}, @var{sp}, @var{struct_return}, @var{struct_addr})
4003 @findex gdbarch_push_dummy_call
4004 @findex DEPRECATED_PUSH_ARGUMENTS.
4005 @anchor{gdbarch_push_dummy_call} Define this to push the dummy frame's call to
4006 the inferior function onto the stack. In addition to pushing @var{nargs}, the
4007 code should push @var{struct_addr} (when @var{struct_return} is non-zero), and
4008 the return address (@var{bp_addr}).
4010 @var{function} is a pointer to a @code{struct value}; on architectures that use
4011 function descriptors, this contains the function descriptor value.
4013 Returns the updated top-of-stack pointer.
4015 This method replaces @code{DEPRECATED_PUSH_ARGUMENTS}.
4017 @item CORE_ADDR gdbarch_push_dummy_code (@var{gdbarch}, @var{sp}, @var{funaddr}, @var{using_gcc}, @var{args}, @var{nargs}, @var{value_type}, @var{real_pc}, @var{bp_addr}, @var{regcache})
4018 @findex gdbarch_push_dummy_code
4019 @anchor{gdbarch_push_dummy_code} Given a stack based call dummy, push the
4020 instruction sequence (including space for a breakpoint) to which the
4021 called function should return.
4023 Set @var{bp_addr} to the address at which the breakpoint instruction
4024 should be inserted, @var{real_pc} to the resume address when starting
4025 the call sequence, and return the updated inner-most stack address.
4027 By default, the stack is grown sufficient to hold a frame-aligned
4028 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
4029 reserved for that breakpoint, and @var{real_pc} set to @var{funaddr}.
4031 This method replaces @w{@code{gdbarch_call_dummy_location (@var{gdbarch})}} and
4032 @code{DEPRECATED_REGISTER_SIZE}.
4034 @item const char *gdbarch_register_name (@var{gdbarch}, @var{regnr})
4035 @findex gdbarch_register_name
4036 Return the name of register @var{regnr} as a string. May return @code{NULL}
4037 to indicate that @var{regnr} is not a valid register.
4039 @item int gdbarch_sdb_reg_to_regnum (@var{gdbarch}, @var{sdb_regnr})
4040 @findex gdbarch_sdb_reg_to_regnum
4041 Use this function to convert sdb register @var{sdb_regnr} into @value{GDBN}
4042 regnum. If not defined, no conversion will be done.
4044 @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})
4045 @findex gdbarch_return_value
4046 @anchor{gdbarch_return_value} Given a function with a return-value of
4047 type @var{rettype}, return which return-value convention that function
4050 @value{GDBN} currently recognizes two function return-value conventions:
4051 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
4052 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
4053 value is found in memory and the address of that memory location is
4054 passed in as the function's first parameter.
4056 If the register convention is being used, and @var{writebuf} is
4057 non-@code{NULL}, also copy the return-value in @var{writebuf} into
4060 If the register convention is being used, and @var{readbuf} is
4061 non-@code{NULL}, also copy the return value from @var{regcache} into
4062 @var{readbuf} (@var{regcache} contains a copy of the registers from the
4063 just returned function).
4065 @emph{Maintainer note: This method replaces separate predicate, extract,
4066 store methods. By having only one method, the logic needed to determine
4067 the return-value convention need only be implemented in one place. If
4068 @value{GDBN} were written in an @sc{oo} language, this method would
4069 instead return an object that knew how to perform the register
4070 return-value extract and store.}
4072 @emph{Maintainer note: This method does not take a @var{gcc_p}
4073 parameter, and such a parameter should not be added. If an architecture
4074 that requires per-compiler or per-function information be identified,
4075 then the replacement of @var{rettype} with @code{struct value}
4076 @var{function} should be pursued.}
4078 @emph{Maintainer note: The @var{regcache} parameter limits this methods
4079 to the inner most frame. While replacing @var{regcache} with a
4080 @code{struct frame_info} @var{frame} parameter would remove that
4081 limitation there has yet to be a demonstrated need for such a change.}
4083 @item void gdbarch_skip_permanent_breakpoint (@var{gdbarch}, @var{regcache})
4084 @findex gdbarch_skip_permanent_breakpoint
4085 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
4086 steps over a breakpoint by removing it, stepping one instruction, and
4087 re-inserting the breakpoint. However, permanent breakpoints are
4088 hardwired into the inferior, and can't be removed, so this strategy
4089 doesn't work. Calling @code{gdbarch_skip_permanent_breakpoint} adjusts the
4090 processor's state so that execution will resume just after the breakpoint.
4091 This function does the right thing even when the breakpoint is in the delay slot
4092 of a branch or jump.
4094 @item CORE_ADDR gdbarch_skip_prologue (@var{gdbarch}, @var{ip})
4095 @findex gdbarch_skip_prologue
4096 A function that returns the address of the ``real'' code beyond the
4097 function entry prologue found at @var{ip}.
4099 @item CORE_ADDR gdbarch_skip_trampoline_code (@var{gdbarch}, @var{frame}, @var{pc})
4100 @findex gdbarch_skip_trampoline_code
4101 If the target machine has trampoline code that sits between callers and
4102 the functions being called, then define this function to return a new PC
4103 that is at the start of the real function.
4105 @item int gdbarch_sp_regnum (@var{gdbarch})
4106 @findex gdbarch_sp_regnum
4107 If the stack-pointer is kept in a register, then use this function to return
4108 the number (greater than or equal to zero) of that register, or -1 if
4109 there is no such register.
4111 @item int gdbarch_stab_reg_to_regnum (@var{gdbarch}, @var{stab_regnr})
4112 @findex gdbarch_stab_reg_to_regnum
4113 Use this function to convert stab register @var{stab_regnr} into @value{GDBN}
4114 regnum. If not defined, no conversion will be done.
4116 @item SYMBOL_RELOADING_DEFAULT
4117 @findex SYMBOL_RELOADING_DEFAULT
4118 The default value of the ``symbol-reloading'' variable. (Never defined in
4121 @item TARGET_CHAR_BIT
4122 @findex TARGET_CHAR_BIT
4123 Number of bits in a char; defaults to 8.
4125 @item int gdbarch_char_signed (@var{gdbarch})
4126 @findex gdbarch_char_signed
4127 Non-zero if @code{char} is normally signed on this architecture; zero if
4128 it should be unsigned.
4130 The ISO C standard requires the compiler to treat @code{char} as
4131 equivalent to either @code{signed char} or @code{unsigned char}; any
4132 character in the standard execution set is supposed to be positive.
4133 Most compilers treat @code{char} as signed, but @code{char} is unsigned
4134 on the IBM S/390, RS6000, and PowerPC targets.
4136 @item int gdbarch_double_bit (@var{gdbarch})
4137 @findex gdbarch_double_bit
4138 Number of bits in a double float; defaults to @w{@code{8 * TARGET_CHAR_BIT}}.
4140 @item int gdbarch_float_bit (@var{gdbarch})
4141 @findex gdbarch_float_bit
4142 Number of bits in a float; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4144 @item int gdbarch_int_bit (@var{gdbarch})
4145 @findex gdbarch_int_bit
4146 Number of bits in an integer; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4148 @item int gdbarch_long_bit (@var{gdbarch})
4149 @findex gdbarch_long_bit
4150 Number of bits in a long integer; defaults to @w{@code{4 * TARGET_CHAR_BIT}}.
4152 @item int gdbarch_long_double_bit (@var{gdbarch})
4153 @findex gdbarch_long_double_bit
4154 Number of bits in a long double float;
4155 defaults to @w{@code{2 * gdbarch_double_bit (@var{gdbarch})}}.
4157 @item int gdbarch_long_long_bit (@var{gdbarch})
4158 @findex gdbarch_long_long_bit
4159 Number of bits in a long long integer; defaults to
4160 @w{@code{2 * gdbarch_long_bit (@var{gdbarch})}}.
4162 @item int gdbarch_ptr_bit (@var{gdbarch})
4163 @findex gdbarch_ptr_bit
4164 Number of bits in a pointer; defaults to
4165 @w{@code{gdbarch_int_bit (@var{gdbarch})}}.
4167 @item int gdbarch_short_bit (@var{gdbarch})
4168 @findex gdbarch_short_bit
4169 Number of bits in a short integer; defaults to @w{@code{2 * TARGET_CHAR_BIT}}.
4171 @item CORE_ADDR gdbarch_read_pc (@var{gdbarch}, @var{regcache})
4172 @findex gdbarch_read_pc
4173 @itemx gdbarch_write_pc (@var{gdbarch}, @var{regcache}, @var{val})
4174 @findex gdbarch_write_pc
4175 @anchor{gdbarch_write_pc}
4176 @itemx TARGET_READ_SP
4177 @findex TARGET_READ_SP
4178 @itemx TARGET_READ_FP
4179 @findex TARGET_READ_FP
4180 @findex gdbarch_read_pc
4181 @findex gdbarch_write_pc
4184 @anchor{TARGET_READ_SP} These change the behavior of @code{gdbarch_read_pc},
4185 @code{gdbarch_write_pc}, and @code{read_sp}. For most targets, these may be
4186 left undefined. @value{GDBN} will call the read and write register
4187 functions with the relevant @code{_REGNUM} argument.
4189 These macros and functions are useful when a target keeps one of these
4190 registers in a hard to get at place; for example, part in a segment register
4191 and part in an ordinary register.
4193 @xref{gdbarch_unwind_sp}, which replaces @code{TARGET_READ_SP}.
4195 @item void gdbarch_virtual_frame_pointer (@var{gdbarch}, @var{pc}, @var{frame_regnum}, @var{frame_offset})
4196 @findex gdbarch_virtual_frame_pointer
4197 Returns a @code{(register, offset)} pair representing the virtual frame
4198 pointer in use at the code address @var{pc}. If virtual frame pointers
4199 are not used, a default definition simply returns
4200 @code{DEPRECATED_FP_REGNUM}, with an offset of zero.
4202 @item TARGET_HAS_HARDWARE_WATCHPOINTS
4203 If non-zero, the target has support for hardware-assisted
4204 watchpoints. @xref{Algorithms, watchpoints}, for more details and
4205 other related macros.
4207 @item int gdbarch_print_insn (@var{gdbarch}, @var{vma}, @var{info})
4208 @findex gdbarch_print_insn
4209 This is the function used by @value{GDBN} to print an assembly
4210 instruction. It prints the instruction at address @var{vma} in
4211 debugged memory and returns the length of the instruction, in bytes. If
4212 a target doesn't define its own printing routine, it defaults to an
4213 accessor function for the global pointer
4214 @code{deprecated_tm_print_insn}. This usually points to a function in
4215 the @code{opcodes} library (@pxref{Support Libraries, ,Opcodes}).
4216 @var{info} is a structure (of type @code{disassemble_info}) defined in
4217 @file{include/dis-asm.h} used to pass information to the instruction
4220 @item frame_id gdbarch_dummy_id (@var{gdbarch}, @var{frame})
4221 @findex gdbarch_dummy_id
4222 @anchor{gdbarch_dummy_id} Given @var{frame} return a @w{@code{struct
4223 frame_id}} that uniquely identifies an inferior function call's dummy
4224 frame. The value returned must match the dummy frame stack value
4225 previously saved by @code{call_function_by_hand}.
4227 @item DEPRECATED_USE_STRUCT_CONVENTION (@var{gcc_p}, @var{type})
4228 @findex DEPRECATED_USE_STRUCT_CONVENTION
4229 If defined, this must be an expression that is nonzero if a value of the
4230 given @var{type} being returned from a function must have space
4231 allocated for it on the stack. @var{gcc_p} is true if the function
4232 being considered is known to have been compiled by GCC; this is helpful
4233 for systems where GCC is known to use different calling convention than
4236 This method has been deprecated in favour of @code{gdbarch_return_value}
4237 (@pxref{gdbarch_return_value}).
4239 @item void gdbarch_value_to_register (@var{gdbarch}, @var{frame}, @var{type}, @var{buf})
4240 @findex gdbarch_value_to_register
4241 Convert a value of type @var{type} into the raw contents of a register.
4242 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
4246 Motorola M68K target conditionals.
4250 Define this to be the 4-bit location of the breakpoint trap vector. If
4251 not defined, it will default to @code{0xf}.
4253 @item REMOTE_BPT_VECTOR
4254 Defaults to @code{1}.
4256 @item const char *gdbarch_name_of_malloc (@var{gdbarch})
4257 @findex gdbarch_name_of_malloc
4258 A string containing the name of the function to call in order to
4259 allocate some memory in the inferior. The default value is "malloc".
4263 @node Adding a New Target
4264 @section Adding a New Target
4266 @cindex adding a target
4267 The following files add a target to @value{GDBN}:
4271 @item gdb/config/@var{arch}/@var{ttt}.mt
4272 Contains a Makefile fragment specific to this target. Specifies what
4273 object files are needed for target @var{ttt}, by defining
4274 @samp{TDEPFILES=@dots{}} and @samp{TDEPLIBS=@dots{}}. Also specifies
4275 the header file which describes @var{ttt}, by defining @samp{TM_FILE=
4278 You can also define @samp{TM_CFLAGS}, @samp{TM_CLIBS}, @samp{TM_CDEPS},
4279 but these are now deprecated, replaced by autoconf, and may go away in
4280 future versions of @value{GDBN}.
4282 @item gdb/@var{ttt}-tdep.c
4283 Contains any miscellaneous code required for this target machine. On
4284 some machines it doesn't exist at all. Sometimes the macros in
4285 @file{tm-@var{ttt}.h} become very complicated, so they are implemented
4286 as functions here instead, and the macro is simply defined to call the
4287 function. This is vastly preferable, since it is easier to understand
4290 @item gdb/@var{arch}-tdep.c
4291 @itemx gdb/@var{arch}-tdep.h
4292 This often exists to describe the basic layout of the target machine's
4293 processor chip (registers, stack, etc.). If used, it is included by
4294 @file{@var{ttt}-tdep.h}. It can be shared among many targets that use
4297 @item gdb/config/@var{arch}/tm-@var{ttt}.h
4298 (@file{tm.h} is a link to this file, created by @code{configure}). Contains
4299 macro definitions about the target machine's registers, stack frame
4300 format and instructions.
4302 New targets do not need this file and should not create it.
4304 @item gdb/config/@var{arch}/tm-@var{arch}.h
4305 This often exists to describe the basic layout of the target machine's
4306 processor chip (registers, stack, etc.). If used, it is included by
4307 @file{tm-@var{ttt}.h}. It can be shared among many targets that use the
4310 New targets do not need this file and should not create it.
4314 If you are adding a new operating system for an existing CPU chip, add a
4315 @file{config/tm-@var{os}.h} file that describes the operating system
4316 facilities that are unusual (extra symbol table info; the breakpoint
4317 instruction needed; etc.). Then write a @file{@var{arch}/tm-@var{os}.h}
4318 that just @code{#include}s @file{tm-@var{arch}.h} and
4319 @file{config/tm-@var{os}.h}.
4321 @node Target Descriptions
4322 @chapter Target Descriptions
4323 @cindex target descriptions
4325 The target architecture definition (@pxref{Target Architecture Definition})
4326 contains @value{GDBN}'s hard-coded knowledge about an architecture. For
4327 some platforms, it is handy to have more flexible knowledge about a specific
4328 instance of the architecture---for instance, a processor or development board.
4329 @dfn{Target descriptions} provide a mechanism for the user to tell @value{GDBN}
4330 more about what their target supports, or for the target to tell @value{GDBN}
4333 For details on writing, automatically supplying, and manually selecting
4334 target descriptions, see @ref{Target Descriptions, , , gdb,
4335 Debugging with @value{GDBN}}. This section will cover some related
4336 topics about the @value{GDBN} internals.
4339 * Target Descriptions Implementation::
4340 * Adding Target Described Register Support::
4343 @node Target Descriptions Implementation
4344 @section Target Descriptions Implementation
4345 @cindex target descriptions, implementation
4347 Before @value{GDBN} connects to a new target, or runs a new program on
4348 an existing target, it discards any existing target description and
4349 reverts to a default gdbarch. Then, after connecting, it looks for a
4350 new target description by calling @code{target_find_description}.
4352 A description may come from a user specified file (XML), the remote
4353 @samp{qXfer:features:read} packet (also XML), or from any custom
4354 @code{to_read_description} routine in the target vector. For instance,
4355 the remote target supports guessing whether a MIPS target is 32-bit or
4356 64-bit based on the size of the @samp{g} packet.
4358 If any target description is found, @value{GDBN} creates a new gdbarch
4359 incorporating the description by calling @code{gdbarch_update_p}. Any
4360 @samp{<architecture>} element is handled first, to determine which
4361 architecture's gdbarch initialization routine is called to create the
4362 new architecture. Then the initialization routine is called, and has
4363 a chance to adjust the constructed architecture based on the contents
4364 of the target description. For instance, it can recognize any
4365 properties set by a @code{to_read_description} routine. Also
4366 see @ref{Adding Target Described Register Support}.
4368 @node Adding Target Described Register Support
4369 @section Adding Target Described Register Support
4370 @cindex target descriptions, adding register support
4372 Target descriptions can report additional registers specific to an
4373 instance of the target. But it takes a little work in the architecture
4374 specific routines to support this.
4376 A target description must either have no registers or a complete
4377 set---this avoids complexity in trying to merge standard registers
4378 with the target defined registers. It is the architecture's
4379 responsibility to validate that a description with registers has
4380 everything it needs. To keep architecture code simple, the same
4381 mechanism is used to assign fixed internal register numbers to
4384 If @code{tdesc_has_registers} returns 1, the description contains
4385 registers. The architecture's @code{gdbarch_init} routine should:
4390 Call @code{tdesc_data_alloc} to allocate storage, early, before
4391 searching for a matching gdbarch or allocating a new one.
4394 Use @code{tdesc_find_feature} to locate standard features by name.
4397 Use @code{tdesc_numbered_register} and @code{tdesc_numbered_register_choices}
4398 to locate the expected registers in the standard features.
4401 Return @code{NULL} if a required feature is missing, or if any standard
4402 feature is missing expected registers. This will produce a warning that
4403 the description was incomplete.
4406 Free the allocated data before returning, unless @code{tdesc_use_registers}
4410 Call @code{set_gdbarch_num_regs} as usual, with a number higher than any
4411 fixed number passed to @code{tdesc_numbered_register}.
4414 Call @code{tdesc_use_registers} after creating a new gdbarch, before
4419 After @code{tdesc_use_registers} has been called, the architecture's
4420 @code{register_name}, @code{register_type}, and @code{register_reggroup_p}
4421 routines will not be called; that information will be taken from
4422 the target description. @code{num_regs} may be increased to account
4423 for any additional registers in the description.
4425 Pseudo-registers require some extra care:
4430 Using @code{tdesc_numbered_register} allows the architecture to give
4431 constant register numbers to standard architectural registers, e.g.@:
4432 as an @code{enum} in @file{@var{arch}-tdep.h}. But because
4433 pseudo-registers are always numbered above @code{num_regs},
4434 which may be increased by the description, constant numbers
4435 can not be used for pseudos. They must be numbered relative to
4436 @code{num_regs} instead.
4439 The description will not describe pseudo-registers, so the
4440 architecture must call @code{set_tdesc_pseudo_register_name},
4441 @code{set_tdesc_pseudo_register_type}, and
4442 @code{set_tdesc_pseudo_register_reggroup_p} to supply routines
4443 describing pseudo registers. These routines will be passed
4444 internal register numbers, so the same routines used for the
4445 gdbarch equivalents are usually suitable.
4450 @node Target Vector Definition
4452 @chapter Target Vector Definition
4453 @cindex target vector
4455 The target vector defines the interface between @value{GDBN}'s
4456 abstract handling of target systems, and the nitty-gritty code that
4457 actually exercises control over a process or a serial port.
4458 @value{GDBN} includes some 30-40 different target vectors; however,
4459 each configuration of @value{GDBN} includes only a few of them.
4462 * Managing Execution State::
4463 * Existing Targets::
4466 @node Managing Execution State
4467 @section Managing Execution State
4468 @cindex execution state
4470 A target vector can be completely inactive (not pushed on the target
4471 stack), active but not running (pushed, but not connected to a fully
4472 manifested inferior), or completely active (pushed, with an accessible
4473 inferior). Most targets are only completely inactive or completely
4474 active, but some support persistent connections to a target even
4475 when the target has exited or not yet started.
4477 For example, connecting to the simulator using @code{target sim} does
4478 not create a running program. Neither registers nor memory are
4479 accessible until @code{run}. Similarly, after @code{kill}, the
4480 program can not continue executing. But in both cases @value{GDBN}
4481 remains connected to the simulator, and target-specific commands
4482 are directed to the simulator.
4484 A target which only supports complete activation should push itself
4485 onto the stack in its @code{to_open} routine (by calling
4486 @code{push_target}), and unpush itself from the stack in its
4487 @code{to_mourn_inferior} routine (by calling @code{unpush_target}).
4489 A target which supports both partial and complete activation should
4490 still call @code{push_target} in @code{to_open}, but not call
4491 @code{unpush_target} in @code{to_mourn_inferior}. Instead, it should
4492 call either @code{target_mark_running} or @code{target_mark_exited}
4493 in its @code{to_open}, depending on whether the target is fully active
4494 after connection. It should also call @code{target_mark_running} any
4495 time the inferior becomes fully active (e.g.@: in
4496 @code{to_create_inferior} and @code{to_attach}), and
4497 @code{target_mark_exited} when the inferior becomes inactive (in
4498 @code{to_mourn_inferior}). The target should also make sure to call
4499 @code{target_mourn_inferior} from its @code{to_kill}, to return the
4500 target to inactive state.
4502 @node Existing Targets
4503 @section Existing Targets
4506 @subsection File Targets
4508 Both executables and core files have target vectors.
4510 @subsection Standard Protocol and Remote Stubs
4512 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code
4513 that runs in the target system. @value{GDBN} provides several sample
4514 @dfn{stubs} that can be integrated into target programs or operating
4515 systems for this purpose; they are named @file{*-stub.c}.
4517 The @value{GDBN} user's manual describes how to put such a stub into
4518 your target code. What follows is a discussion of integrating the
4519 SPARC stub into a complicated operating system (rather than a simple
4520 program), by Stu Grossman, the author of this stub.
4522 The trap handling code in the stub assumes the following upon entry to
4527 %l1 and %l2 contain pc and npc respectively at the time of the trap;
4533 you are in the correct trap window.
4536 As long as your trap handler can guarantee those conditions, then there
4537 is no reason why you shouldn't be able to ``share'' traps with the stub.
4538 The stub has no requirement that it be jumped to directly from the
4539 hardware trap vector. That is why it calls @code{exceptionHandler()},
4540 which is provided by the external environment. For instance, this could
4541 set up the hardware traps to actually execute code which calls the stub
4542 first, and then transfers to its own trap handler.
4544 For the most point, there probably won't be much of an issue with
4545 ``sharing'' traps, as the traps we use are usually not used by the kernel,
4546 and often indicate unrecoverable error conditions. Anyway, this is all
4547 controlled by a table, and is trivial to modify. The most important
4548 trap for us is for @code{ta 1}. Without that, we can't single step or
4549 do breakpoints. Everything else is unnecessary for the proper operation
4550 of the debugger/stub.
4552 From reading the stub, it's probably not obvious how breakpoints work.
4553 They are simply done by deposit/examine operations from @value{GDBN}.
4555 @subsection ROM Monitor Interface
4557 @subsection Custom Protocols
4559 @subsection Transport Layer
4561 @subsection Builtin Simulator
4564 @node Native Debugging
4566 @chapter Native Debugging
4567 @cindex native debugging
4569 Several files control @value{GDBN}'s configuration for native support:
4573 @item gdb/config/@var{arch}/@var{xyz}.mh
4574 Specifies Makefile fragments needed by a @emph{native} configuration on
4575 machine @var{xyz}. In particular, this lists the required
4576 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
4577 Also specifies the header file which describes native support on
4578 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
4579 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
4580 @samp{NAT_CDEPS}, etc.; see @file{Makefile.in}.
4582 @emph{Maintainer's note: The @file{.mh} suffix is because this file
4583 originally contained @file{Makefile} fragments for hosting @value{GDBN}
4584 on machine @var{xyz}. While the file is no longer used for this
4585 purpose, the @file{.mh} suffix remains. Perhaps someone will
4586 eventually rename these fragments so that they have a @file{.mn}
4589 @item gdb/config/@var{arch}/nm-@var{xyz}.h
4590 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
4591 macro definitions describing the native system environment, such as
4592 child process control and core file support.
4594 @item gdb/@var{xyz}-nat.c
4595 Contains any miscellaneous C code required for this native support of
4596 this machine. On some machines it doesn't exist at all.
4599 There are some ``generic'' versions of routines that can be used by
4600 various systems. These can be customized in various ways by macros
4601 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
4602 the @var{xyz} host, you can just include the generic file's name (with
4603 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
4605 Otherwise, if your machine needs custom support routines, you will need
4606 to write routines that perform the same functions as the generic file.
4607 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
4608 into @code{NATDEPFILES}.
4612 This contains the @emph{target_ops vector} that supports Unix child
4613 processes on systems which use ptrace and wait to control the child.
4616 This contains the @emph{target_ops vector} that supports Unix child
4617 processes on systems which use /proc to control the child.
4620 This does the low-level grunge that uses Unix system calls to do a ``fork
4621 and exec'' to start up a child process.
4624 This is the low level interface to inferior processes for systems using
4625 the Unix @code{ptrace} call in a vanilla way.
4628 @section Native core file Support
4629 @cindex native core files
4632 @findex fetch_core_registers
4633 @item core-aout.c::fetch_core_registers()
4634 Support for reading registers out of a core file. This routine calls
4635 @code{register_addr()}, see below. Now that BFD is used to read core
4636 files, virtually all machines should use @code{core-aout.c}, and should
4637 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
4638 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
4640 @item core-aout.c::register_addr()
4641 If your @code{nm-@var{xyz}.h} file defines the macro
4642 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
4643 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
4644 register number @code{regno}. @code{blockend} is the offset within the
4645 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
4646 @file{core-aout.c} will define the @code{register_addr()} function and
4647 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
4648 you are using the standard @code{fetch_core_registers()}, you will need
4649 to define your own version of @code{register_addr()}, put it into your
4650 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
4651 the @code{NATDEPFILES} list. If you have your own
4652 @code{fetch_core_registers()}, you may not need a separate
4653 @code{register_addr()}. Many custom @code{fetch_core_registers()}
4654 implementations simply locate the registers themselves.@refill
4657 When making @value{GDBN} run native on a new operating system, to make it
4658 possible to debug core files, you will need to either write specific
4659 code for parsing your OS's core files, or customize
4660 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
4661 machine uses to define the struct of registers that is accessible
4662 (possibly in the u-area) in a core file (rather than
4663 @file{machine/reg.h}), and an include file that defines whatever header
4664 exists on a core file (e.g., the u-area or a @code{struct core}). Then
4665 modify @code{trad_unix_core_file_p} to use these values to set up the
4666 section information for the data segment, stack segment, any other
4667 segments in the core file (perhaps shared library contents or control
4668 information), ``registers'' segment, and if there are two discontiguous
4669 sets of registers (e.g., integer and float), the ``reg2'' segment. This
4670 section information basically delimits areas in the core file in a
4671 standard way, which the section-reading routines in BFD know how to seek
4674 Then back in @value{GDBN}, you need a matching routine called
4675 @code{fetch_core_registers}. If you can use the generic one, it's in
4676 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
4677 It will be passed a char pointer to the entire ``registers'' segment,
4678 its length, and a zero; or a char pointer to the entire ``regs2''
4679 segment, its length, and a 2. The routine should suck out the supplied
4680 register values and install them into @value{GDBN}'s ``registers'' array.
4682 If your system uses @file{/proc} to control processes, and uses ELF
4683 format core files, then you may be able to use the same routines for
4684 reading the registers out of processes and out of core files.
4692 @section shared libraries
4694 @section Native Conditionals
4695 @cindex native conditionals
4697 When @value{GDBN} is configured and compiled, various macros are
4698 defined or left undefined, to control compilation when the host and
4699 target systems are the same. These macros should be defined (or left
4700 undefined) in @file{nm-@var{system}.h}.
4704 @item CHILD_PREPARE_TO_STORE
4705 @findex CHILD_PREPARE_TO_STORE
4706 If the machine stores all registers at once in the child process, then
4707 define this to ensure that all values are correct. This usually entails
4708 a read from the child.
4710 [Note that this is incorrectly defined in @file{xm-@var{system}.h} files
4713 @item FETCH_INFERIOR_REGISTERS
4714 @findex FETCH_INFERIOR_REGISTERS
4715 Define this if the native-dependent code will provide its own routines
4716 @code{fetch_inferior_registers} and @code{store_inferior_registers} in
4717 @file{@var{host}-nat.c}. If this symbol is @emph{not} defined, and
4718 @file{infptrace.c} is included in this configuration, the default
4719 routines in @file{infptrace.c} are used for these functions.
4721 @item int gdbarch_fp0_regnum (@var{gdbarch})
4722 @findex gdbarch_fp0_regnum
4723 This functions normally returns the number of the first floating
4724 point register, if the machine has such registers. As such, it would
4725 appear only in target-specific code. However, @file{/proc} support uses this
4726 to decide whether floats are in use on this target.
4728 @item int gdbarch_get_longjmp_target (@var{gdbarch})
4729 @findex gdbarch_get_longjmp_target
4730 For most machines, this is a target-dependent parameter. On the
4731 DECstation and the Iris, this is a native-dependent parameter, since
4732 @file{setjmp.h} is needed to define it.
4734 This function determines the target PC address that @code{longjmp} will jump to,
4735 assuming that we have just stopped at a longjmp breakpoint. It takes a
4736 @code{CORE_ADDR *} as argument, and stores the target PC value through this
4737 pointer. It examines the current state of the machine as needed.
4739 @item I386_USE_GENERIC_WATCHPOINTS
4740 An x86-based machine can define this to use the generic x86 watchpoint
4741 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4743 @item ONE_PROCESS_WRITETEXT
4744 @findex ONE_PROCESS_WRITETEXT
4745 Define this to be able to, when a breakpoint insertion fails, warn the
4746 user that another process may be running with the same executable.
4749 @findex PROC_NAME_FMT
4750 Defines the format for the name of a @file{/proc} device. Should be
4751 defined in @file{nm.h} @emph{only} in order to override the default
4752 definition in @file{procfs.c}.
4754 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
4756 Define this to expand into an expression that will cause the symbols in
4757 @var{filename} to be added to @value{GDBN}'s symbol table. If
4758 @var{readsyms} is zero symbols are not read but any necessary low level
4759 processing for @var{filename} is still done.
4761 @item SOLIB_CREATE_INFERIOR_HOOK
4762 @findex SOLIB_CREATE_INFERIOR_HOOK
4763 Define this to expand into any shared-library-relocation code that you
4764 want to be run just after the child process has been forked.
4766 @item START_INFERIOR_TRAPS_EXPECTED
4767 @findex START_INFERIOR_TRAPS_EXPECTED
4768 When starting an inferior, @value{GDBN} normally expects to trap
4770 the shell execs, and once when the program itself execs. If the actual
4771 number of traps is something other than 2, then define this macro to
4772 expand into the number expected.
4776 @node Support Libraries
4778 @chapter Support Libraries
4783 BFD provides support for @value{GDBN} in several ways:
4786 @item identifying executable and core files
4787 BFD will identify a variety of file types, including a.out, coff, and
4788 several variants thereof, as well as several kinds of core files.
4790 @item access to sections of files
4791 BFD parses the file headers to determine the names, virtual addresses,
4792 sizes, and file locations of all the various named sections in files
4793 (such as the text section or the data section). @value{GDBN} simply
4794 calls BFD to read or write section @var{x} at byte offset @var{y} for
4797 @item specialized core file support
4798 BFD provides routines to determine the failing command name stored in a
4799 core file, the signal with which the program failed, and whether a core
4800 file matches (i.e.@: could be a core dump of) a particular executable
4803 @item locating the symbol information
4804 @value{GDBN} uses an internal interface of BFD to determine where to find the
4805 symbol information in an executable file or symbol-file. @value{GDBN} itself
4806 handles the reading of symbols, since BFD does not ``understand'' debug
4807 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
4812 @cindex opcodes library
4814 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
4815 library because it's also used in binutils, for @file{objdump}).
4818 @cindex readline library
4819 The @code{readline} library provides a set of functions for use by applications
4820 that allow users to edit command lines as they are typed in.
4823 @cindex @code{libiberty} library
4825 The @code{libiberty} library provides a set of functions and features
4826 that integrate and improve on functionality found in modern operating
4827 systems. Broadly speaking, such features can be divided into three
4828 groups: supplemental functions (functions that may be missing in some
4829 environments and operating systems), replacement functions (providing
4830 a uniform and easier to use interface for commonly used standard
4831 functions), and extensions (which provide additional functionality
4832 beyond standard functions).
4834 @value{GDBN} uses various features provided by the @code{libiberty}
4835 library, for instance the C@t{++} demangler, the @acronym{IEEE}
4836 floating format support functions, the input options parser
4837 @samp{getopt}, the @samp{obstack} extension, and other functions.
4839 @subsection @code{obstacks} in @value{GDBN}
4840 @cindex @code{obstacks}
4842 The obstack mechanism provides a convenient way to allocate and free
4843 chunks of memory. Each obstack is a pool of memory that is managed
4844 like a stack. Objects (of any nature, size and alignment) are
4845 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
4846 @code{libiberty}'s documentation for a more detailed explanation of
4849 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
4850 object files. There is an obstack associated with each internal
4851 representation of an object file. Lots of things get allocated on
4852 these @code{obstacks}: dictionary entries, blocks, blockvectors,
4853 symbols, minimal symbols, types, vectors of fundamental types, class
4854 fields of types, object files section lists, object files section
4855 offset lists, line tables, symbol tables, partial symbol tables,
4856 string tables, symbol table private data, macros tables, debug
4857 information sections and entries, import and export lists (som),
4858 unwind information (hppa), dwarf2 location expressions data. Plus
4859 various strings such as directory names strings, debug format strings,
4862 An essential and convenient property of all data on @code{obstacks} is
4863 that memory for it gets allocated (with @code{obstack_alloc}) at
4864 various times during a debugging session, but it is released all at
4865 once using the @code{obstack_free} function. The @code{obstack_free}
4866 function takes a pointer to where in the stack it must start the
4867 deletion from (much like the cleanup chains have a pointer to where to
4868 start the cleanups). Because of the stack like structure of the
4869 @code{obstacks}, this allows to free only a top portion of the
4870 obstack. There are a few instances in @value{GDBN} where such thing
4871 happens. Calls to @code{obstack_free} are done after some local data
4872 is allocated to the obstack. Only the local data is deleted from the
4873 obstack. Of course this assumes that nothing between the
4874 @code{obstack_alloc} and the @code{obstack_free} allocates anything
4875 else on the same obstack. For this reason it is best and safest to
4876 use temporary @code{obstacks}.
4878 Releasing the whole obstack is also not safe per se. It is safe only
4879 under the condition that we know the @code{obstacks} memory is no
4880 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
4881 when we get rid of the whole objfile(s), for instance upon reading a
4885 @cindex regular expressions library
4896 @item SIGN_EXTEND_CHAR
4898 @item SWITCH_ENUM_BUG
4907 @section Array Containers
4908 @cindex Array Containers
4911 Often it is necessary to manipulate a dynamic array of a set of
4912 objects. C forces some bookkeeping on this, which can get cumbersome
4913 and repetitive. The @file{vec.h} file contains macros for defining
4914 and using a typesafe vector type. The functions defined will be
4915 inlined when compiling, and so the abstraction cost should be zero.
4916 Domain checks are added to detect programming errors.
4918 An example use would be an array of symbols or section information.
4919 The array can be grown as symbols are read in (or preallocated), and
4920 the accessor macros provided keep care of all the necessary
4921 bookkeeping. Because the arrays are type safe, there is no danger of
4922 accidentally mixing up the contents. Think of these as C++ templates,
4923 but implemented in C.
4925 Because of the different behavior of structure objects, scalar objects
4926 and of pointers, there are three flavors of vector, one for each of
4927 these variants. Both the structure object and pointer variants pass
4928 pointers to objects around --- in the former case the pointers are
4929 stored into the vector and in the latter case the pointers are
4930 dereferenced and the objects copied into the vector. The scalar
4931 object variant is suitable for @code{int}-like objects, and the vector
4932 elements are returned by value.
4934 There are both @code{index} and @code{iterate} accessors. The iterator
4935 returns a boolean iteration condition and updates the iteration
4936 variable passed by reference. Because the iterator will be inlined,
4937 the address-of can be optimized away.
4939 The vectors are implemented using the trailing array idiom, thus they
4940 are not resizeable without changing the address of the vector object
4941 itself. This means you cannot have variables or fields of vector type
4942 --- always use a pointer to a vector. The one exception is the final
4943 field of a structure, which could be a vector type. You will have to
4944 use the @code{embedded_size} & @code{embedded_init} calls to create
4945 such objects, and they will probably not be resizeable (so don't use
4946 the @dfn{safe} allocation variants). The trailing array idiom is used
4947 (rather than a pointer to an array of data), because, if we allow
4948 @code{NULL} to also represent an empty vector, empty vectors occupy
4949 minimal space in the structure containing them.
4951 Each operation that increases the number of active elements is
4952 available in @dfn{quick} and @dfn{safe} variants. The former presumes
4953 that there is sufficient allocated space for the operation to succeed
4954 (it dies if there is not). The latter will reallocate the vector, if
4955 needed. Reallocation causes an exponential increase in vector size.
4956 If you know you will be adding N elements, it would be more efficient
4957 to use the reserve operation before adding the elements with the
4958 @dfn{quick} operation. This will ensure there are at least as many
4959 elements as you ask for, it will exponentially increase if there are
4960 too few spare slots. If you want reserve a specific number of slots,
4961 but do not want the exponential increase (for instance, you know this
4962 is the last allocation), use a negative number for reservation. You
4963 can also create a vector of a specific size from the get go.
4965 You should prefer the push and pop operations, as they append and
4966 remove from the end of the vector. If you need to remove several items
4967 in one go, use the truncate operation. The insert and remove
4968 operations allow you to change elements in the middle of the vector.
4969 There are two remove operations, one which preserves the element
4970 ordering @code{ordered_remove}, and one which does not
4971 @code{unordered_remove}. The latter function copies the end element
4972 into the removed slot, rather than invoke a memmove operation. The
4973 @code{lower_bound} function will determine where to place an item in
4974 the array using insert that will maintain sorted order.
4976 If you need to directly manipulate a vector, then the @code{address}
4977 accessor will return the address of the start of the vector. Also the
4978 @code{space} predicate will tell you whether there is spare capacity in the
4979 vector. You will not normally need to use these two functions.
4981 Vector types are defined using a
4982 @code{DEF_VEC_@{O,P,I@}(@var{typename})} macro. Variables of vector
4983 type are declared using a @code{VEC(@var{typename})} macro. The
4984 characters @code{O}, @code{P} and @code{I} indicate whether
4985 @var{typename} is an object (@code{O}), pointer (@code{P}) or integral
4986 (@code{I}) type. Be careful to pick the correct one, as you'll get an
4987 awkward and inefficient API if you use the wrong one. There is a
4988 check, which results in a compile-time warning, for the @code{P} and
4989 @code{I} versions, but there is no check for the @code{O} versions, as
4990 that is not possible in plain C.
4992 An example of their use would be,
4995 DEF_VEC_P(tree); // non-managed tree vector.
4998 VEC(tree) *v; // A (pointer to) a vector of tree pointers.
5001 struct my_struct *s;
5003 if (VEC_length(tree, s->v)) @{ we have some contents @}
5004 VEC_safe_push(tree, s->v, decl); // append some decl onto the end
5005 for (ix = 0; VEC_iterate(tree, s->v, ix, elt); ix++)
5006 @{ do something with elt @}
5010 The @file{vec.h} file provides details on how to invoke the various
5011 accessors provided. They are enumerated here:
5015 Return the number of items in the array,
5018 Return true if the array has no elements.
5022 Return the last or arbitrary item in the array.
5025 Access an array element and indicate whether the array has been
5030 Create and destroy an array.
5032 @item VEC_embedded_size
5033 @itemx VEC_embedded_init
5034 Helpers for embedding an array as the final element of another struct.
5040 Return the amount of free space in an array.
5043 Ensure a certain amount of free space.
5045 @item VEC_quick_push
5046 @itemx VEC_safe_push
5047 Append to an array, either assuming the space is available, or making
5051 Remove the last item from an array.
5054 Remove several items from the end of an array.
5057 Add several items to the end of an array.
5060 Overwrite an item in the array.
5062 @item VEC_quick_insert
5063 @itemx VEC_safe_insert
5064 Insert an item into the middle of the array. Either the space must
5065 already exist, or the space is created.
5067 @item VEC_ordered_remove
5068 @itemx VEC_unordered_remove
5069 Remove an item from the array, preserving order or not.
5071 @item VEC_block_remove
5072 Remove a set of items from the array.
5075 Provide the address of the first element.
5077 @item VEC_lower_bound
5078 Binary search the array.
5088 This chapter covers topics that are lower-level than the major
5089 algorithms of @value{GDBN}.
5094 Cleanups are a structured way to deal with things that need to be done
5097 When your code does something (e.g., @code{xmalloc} some memory, or
5098 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
5099 the memory or @code{close} the file), it can make a cleanup. The
5100 cleanup will be done at some future point: when the command is finished
5101 and control returns to the top level; when an error occurs and the stack
5102 is unwound; or when your code decides it's time to explicitly perform
5103 cleanups. Alternatively you can elect to discard the cleanups you
5109 @item struct cleanup *@var{old_chain};
5110 Declare a variable which will hold a cleanup chain handle.
5112 @findex make_cleanup
5113 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
5114 Make a cleanup which will cause @var{function} to be called with
5115 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
5116 handle that can later be passed to @code{do_cleanups} or
5117 @code{discard_cleanups}. Unless you are going to call
5118 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
5119 from @code{make_cleanup}.
5122 @item do_cleanups (@var{old_chain});
5123 Do all cleanups added to the chain since the corresponding
5124 @code{make_cleanup} call was made.
5126 @findex discard_cleanups
5127 @item discard_cleanups (@var{old_chain});
5128 Same as @code{do_cleanups} except that it just removes the cleanups from
5129 the chain and does not call the specified functions.
5132 Cleanups are implemented as a chain. The handle returned by
5133 @code{make_cleanups} includes the cleanup passed to the call and any
5134 later cleanups appended to the chain (but not yet discarded or
5138 make_cleanup (a, 0);
5140 struct cleanup *old = make_cleanup (b, 0);
5148 will call @code{c()} and @code{b()} but will not call @code{a()}. The
5149 cleanup that calls @code{a()} will remain in the cleanup chain, and will
5150 be done later unless otherwise discarded.@refill
5152 Your function should explicitly do or discard the cleanups it creates.
5153 Failing to do this leads to non-deterministic behavior since the caller
5154 will arbitrarily do or discard your functions cleanups. This need leads
5155 to two common cleanup styles.
5157 The first style is try/finally. Before it exits, your code-block calls
5158 @code{do_cleanups} with the old cleanup chain and thus ensures that your
5159 code-block's cleanups are always performed. For instance, the following
5160 code-segment avoids a memory leak problem (even when @code{error} is
5161 called and a forced stack unwind occurs) by ensuring that the
5162 @code{xfree} will always be called:
5165 struct cleanup *old = make_cleanup (null_cleanup, 0);
5166 data = xmalloc (sizeof blah);
5167 make_cleanup (xfree, data);
5172 The second style is try/except. Before it exits, your code-block calls
5173 @code{discard_cleanups} with the old cleanup chain and thus ensures that
5174 any created cleanups are not performed. For instance, the following
5175 code segment, ensures that the file will be closed but only if there is
5179 FILE *file = fopen ("afile", "r");
5180 struct cleanup *old = make_cleanup (close_file, file);
5182 discard_cleanups (old);
5186 Some functions, e.g., @code{fputs_filtered()} or @code{error()}, specify
5187 that they ``should not be called when cleanups are not in place''. This
5188 means that any actions you need to reverse in the case of an error or
5189 interruption must be on the cleanup chain before you call these
5190 functions, since they might never return to your code (they
5191 @samp{longjmp} instead).
5193 @section Per-architecture module data
5194 @cindex per-architecture module data
5195 @cindex multi-arch data
5196 @cindex data-pointer, per-architecture/per-module
5198 The multi-arch framework includes a mechanism for adding module
5199 specific per-architecture data-pointers to the @code{struct gdbarch}
5200 architecture object.
5202 A module registers one or more per-architecture data-pointers using:
5204 @deftypefun struct gdbarch_data *gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
5205 @var{pre_init} is used to, on-demand, allocate an initial value for a
5206 per-architecture data-pointer using the architecture's obstack (passed
5207 in as a parameter). Since @var{pre_init} can be called during
5208 architecture creation, it is not parameterized with the architecture.
5209 and must not call modules that use per-architecture data.
5212 @deftypefun struct gdbarch_data *gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
5213 @var{post_init} is used to obtain an initial value for a
5214 per-architecture data-pointer @emph{after}. Since @var{post_init} is
5215 always called after architecture creation, it both receives the fully
5216 initialized architecture and is free to call modules that use
5217 per-architecture data (care needs to be taken to ensure that those
5218 other modules do not try to call back to this module as that will
5219 create in cycles in the initialization call graph).
5222 These functions return a @code{struct gdbarch_data} that is used to
5223 identify the per-architecture data-pointer added for that module.
5225 The per-architecture data-pointer is accessed using the function:
5227 @deftypefun void *gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
5228 Given the architecture @var{arch} and module data handle
5229 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
5230 or @code{gdbarch_data_register_post_init}), this function returns the
5231 current value of the per-architecture data-pointer. If the data
5232 pointer is @code{NULL}, it is first initialized by calling the
5233 corresponding @var{pre_init} or @var{post_init} method.
5236 The examples below assume the following definitions:
5239 struct nozel @{ int total; @};
5240 static struct gdbarch_data *nozel_handle;
5243 A module can extend the architecture vector, adding additional
5244 per-architecture data, using the @var{pre_init} method. The module's
5245 per-architecture data is then initialized during architecture
5248 In the below, the module's per-architecture @emph{nozel} is added. An
5249 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
5250 from @code{gdbarch_init}.
5254 nozel_pre_init (struct obstack *obstack)
5256 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
5263 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
5265 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5266 data->total = nozel;
5270 A module can on-demand create architecture dependant data structures
5271 using @code{post_init}.
5273 In the below, the nozel's total is computed on-demand by
5274 @code{nozel_post_init} using information obtained from the
5279 nozel_post_init (struct gdbarch *gdbarch)
5281 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
5282 nozel->total = gdbarch@dots{} (gdbarch);
5289 nozel_total (struct gdbarch *gdbarch)
5291 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5296 @section Wrapping Output Lines
5297 @cindex line wrap in output
5300 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
5301 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
5302 added in places that would be good breaking points. The utility
5303 routines will take care of actually wrapping if the line width is
5306 The argument to @code{wrap_here} is an indentation string which is
5307 printed @emph{only} if the line breaks there. This argument is saved
5308 away and used later. It must remain valid until the next call to
5309 @code{wrap_here} or until a newline has been printed through the
5310 @code{*_filtered} functions. Don't pass in a local variable and then
5313 It is usually best to call @code{wrap_here} after printing a comma or
5314 space. If you call it before printing a space, make sure that your
5315 indentation properly accounts for the leading space that will print if
5316 the line wraps there.
5318 Any function or set of functions that produce filtered output must
5319 finish by printing a newline, to flush the wrap buffer, before switching
5320 to unfiltered (@code{printf}) output. Symbol reading routines that
5321 print warnings are a good example.
5323 @section @value{GDBN} Coding Standards
5324 @cindex coding standards
5326 @value{GDBN} follows the GNU coding standards, as described in
5327 @file{etc/standards.texi}. This file is also available for anonymous
5328 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
5329 of the standard; in general, when the GNU standard recommends a practice
5330 but does not require it, @value{GDBN} requires it.
5332 @value{GDBN} follows an additional set of coding standards specific to
5333 @value{GDBN}, as described in the following sections.
5338 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
5341 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
5344 @subsection Memory Management
5346 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
5347 @code{calloc}, @code{free} and @code{asprintf}.
5349 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
5350 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
5351 these functions do not return when the memory pool is empty. Instead,
5352 they unwind the stack using cleanups. These functions return
5353 @code{NULL} when requested to allocate a chunk of memory of size zero.
5355 @emph{Pragmatics: By using these functions, the need to check every
5356 memory allocation is removed. These functions provide portable
5359 @value{GDBN} does not use the function @code{free}.
5361 @value{GDBN} uses the function @code{xfree} to return memory to the
5362 memory pool. Consistent with ISO-C, this function ignores a request to
5363 free a @code{NULL} pointer.
5365 @emph{Pragmatics: On some systems @code{free} fails when passed a
5366 @code{NULL} pointer.}
5368 @value{GDBN} can use the non-portable function @code{alloca} for the
5369 allocation of small temporary values (such as strings).
5371 @emph{Pragmatics: This function is very non-portable. Some systems
5372 restrict the memory being allocated to no more than a few kilobytes.}
5374 @value{GDBN} uses the string function @code{xstrdup} and the print
5375 function @code{xstrprintf}.
5377 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
5378 functions such as @code{sprintf} are very prone to buffer overflow
5382 @subsection Compiler Warnings
5383 @cindex compiler warnings
5385 With few exceptions, developers should avoid the configuration option
5386 @samp{--disable-werror} when building @value{GDBN}. The exceptions
5387 are listed in the file @file{gdb/MAINTAINERS}. The default, when
5388 building with @sc{gcc}, is @samp{--enable-werror}.
5390 This option causes @value{GDBN} (when built using GCC) to be compiled
5391 with a carefully selected list of compiler warning flags. Any warnings
5392 from those flags are treated as errors.
5394 The current list of warning flags includes:
5398 Recommended @sc{gcc} warnings.
5400 @item -Wdeclaration-after-statement
5402 @sc{gcc} 3.x (and later) and @sc{c99} allow declarations mixed with
5403 code, but @sc{gcc} 2.x and @sc{c89} do not.
5405 @item -Wpointer-arith
5407 @item -Wformat-nonliteral
5408 Non-literal format strings, with a few exceptions, are bugs - they
5409 might contain unintended user-supplied format specifiers.
5410 Since @value{GDBN} uses the @code{format printf} attribute on all
5411 @code{printf} like functions this checks not just @code{printf} calls
5412 but also calls to functions such as @code{fprintf_unfiltered}.
5414 @item -Wno-pointer-sign
5415 In version 4.0, GCC began warning about pointer argument passing or
5416 assignment even when the source and destination differed only in
5417 signedness. However, most @value{GDBN} code doesn't distinguish
5418 carefully between @code{char} and @code{unsigned char}. In early 2006
5419 the @value{GDBN} developers decided correcting these warnings wasn't
5420 worth the time it would take.
5422 @item -Wno-unused-parameter
5423 Due to the way that @value{GDBN} is implemented many functions have
5424 unused parameters. Consequently this warning is avoided. The macro
5425 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
5426 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
5431 @itemx -Wno-char-subscripts
5432 These are warnings which might be useful for @value{GDBN}, but are
5433 currently too noisy to enable with @samp{-Werror}.
5437 @subsection Formatting
5439 @cindex source code formatting
5440 The standard GNU recommendations for formatting must be followed
5443 A function declaration should not have its name in column zero. A
5444 function definition should have its name in column zero.
5448 static void foo (void);
5456 @emph{Pragmatics: This simplifies scripting. Function definitions can
5457 be found using @samp{^function-name}.}
5459 There must be a space between a function or macro name and the opening
5460 parenthesis of its argument list (except for macro definitions, as
5461 required by C). There must not be a space after an open paren/bracket
5462 or before a close paren/bracket.
5464 While additional whitespace is generally helpful for reading, do not use
5465 more than one blank line to separate blocks, and avoid adding whitespace
5466 after the end of a program line (as of 1/99, some 600 lines had
5467 whitespace after the semicolon). Excess whitespace causes difficulties
5468 for @code{diff} and @code{patch} utilities.
5470 Pointers are declared using the traditional K&R C style:
5484 @subsection Comments
5486 @cindex comment formatting
5487 The standard GNU requirements on comments must be followed strictly.
5489 Block comments must appear in the following form, with no @code{/*}- or
5490 @code{*/}-only lines, and no leading @code{*}:
5493 /* Wait for control to return from inferior to debugger. If inferior
5494 gets a signal, we may decide to start it up again instead of
5495 returning. That is why there is a loop in this function. When
5496 this function actually returns it means the inferior should be left
5497 stopped and @value{GDBN} should read more commands. */
5500 (Note that this format is encouraged by Emacs; tabbing for a multi-line
5501 comment works correctly, and @kbd{M-q} fills the block consistently.)
5503 Put a blank line between the block comments preceding function or
5504 variable definitions, and the definition itself.
5506 In general, put function-body comments on lines by themselves, rather
5507 than trying to fit them into the 20 characters left at the end of a
5508 line, since either the comment or the code will inevitably get longer
5509 than will fit, and then somebody will have to move it anyhow.
5513 @cindex C data types
5514 Code must not depend on the sizes of C data types, the format of the
5515 host's floating point numbers, the alignment of anything, or the order
5516 of evaluation of expressions.
5518 @cindex function usage
5519 Use functions freely. There are only a handful of compute-bound areas
5520 in @value{GDBN} that might be affected by the overhead of a function
5521 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
5522 limited by the target interface (whether serial line or system call).
5524 However, use functions with moderation. A thousand one-line functions
5525 are just as hard to understand as a single thousand-line function.
5527 @emph{Macros are bad, M'kay.}
5528 (But if you have to use a macro, make sure that the macro arguments are
5529 protected with parentheses.)
5533 Declarations like @samp{struct foo *} should be used in preference to
5534 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
5537 @subsection Function Prototypes
5538 @cindex function prototypes
5540 Prototypes must be used when both @emph{declaring} and @emph{defining}
5541 a function. Prototypes for @value{GDBN} functions must include both the
5542 argument type and name, with the name matching that used in the actual
5543 function definition.
5545 All external functions should have a declaration in a header file that
5546 callers include, except for @code{_initialize_*} functions, which must
5547 be external so that @file{init.c} construction works, but shouldn't be
5548 visible to random source files.
5550 Where a source file needs a forward declaration of a static function,
5551 that declaration must appear in a block near the top of the source file.
5554 @subsection Internal Error Recovery
5556 During its execution, @value{GDBN} can encounter two types of errors.
5557 User errors and internal errors. User errors include not only a user
5558 entering an incorrect command but also problems arising from corrupt
5559 object files and system errors when interacting with the target.
5560 Internal errors include situations where @value{GDBN} has detected, at
5561 run time, a corrupt or erroneous situation.
5563 When reporting an internal error, @value{GDBN} uses
5564 @code{internal_error} and @code{gdb_assert}.
5566 @value{GDBN} must not call @code{abort} or @code{assert}.
5568 @emph{Pragmatics: There is no @code{internal_warning} function. Either
5569 the code detected a user error, recovered from it and issued a
5570 @code{warning} or the code failed to correctly recover from the user
5571 error and issued an @code{internal_error}.}
5573 @subsection File Names
5575 Any file used when building the core of @value{GDBN} must be in lower
5576 case. Any file used when building the core of @value{GDBN} must be 8.3
5577 unique. These requirements apply to both source and generated files.
5579 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
5580 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
5581 is introduced to the build process both @file{Makefile.in} and
5582 @file{configure.in} need to be modified accordingly. Compare the
5583 convoluted conversion process needed to transform @file{COPYING} into
5584 @file{copying.c} with the conversion needed to transform
5585 @file{version.in} into @file{version.c}.}
5587 Any file non 8.3 compliant file (that is not used when building the core
5588 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
5590 @emph{Pragmatics: This is clearly a compromise.}
5592 When @value{GDBN} has a local version of a system header file (ex
5593 @file{string.h}) the file name based on the POSIX header prefixed with
5594 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
5595 independent: they should use only macros defined by @file{configure},
5596 the compiler, or the host; they should include only system headers; they
5597 should refer only to system types. They may be shared between multiple
5598 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
5600 For other files @samp{-} is used as the separator.
5603 @subsection Include Files
5605 A @file{.c} file should include @file{defs.h} first.
5607 A @file{.c} file should directly include the @code{.h} file of every
5608 declaration and/or definition it directly refers to. It cannot rely on
5611 A @file{.h} file should directly include the @code{.h} file of every
5612 declaration and/or definition it directly refers to. It cannot rely on
5613 indirect inclusion. Exception: The file @file{defs.h} does not need to
5614 be directly included.
5616 An external declaration should only appear in one include file.
5618 An external declaration should never appear in a @code{.c} file.
5619 Exception: a declaration for the @code{_initialize} function that
5620 pacifies @option{-Wmissing-declaration}.
5622 A @code{typedef} definition should only appear in one include file.
5624 An opaque @code{struct} declaration can appear in multiple @file{.h}
5625 files. Where possible, a @file{.h} file should use an opaque
5626 @code{struct} declaration instead of an include.
5628 All @file{.h} files should be wrapped in:
5631 #ifndef INCLUDE_FILE_NAME_H
5632 #define INCLUDE_FILE_NAME_H
5638 @subsection Clean Design and Portable Implementation
5641 In addition to getting the syntax right, there's the little question of
5642 semantics. Some things are done in certain ways in @value{GDBN} because long
5643 experience has shown that the more obvious ways caused various kinds of
5646 @cindex assumptions about targets
5647 You can't assume the byte order of anything that comes from a target
5648 (including @var{value}s, object files, and instructions). Such things
5649 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
5650 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
5651 such as @code{bfd_get_32}.
5653 You can't assume that you know what interface is being used to talk to
5654 the target system. All references to the target must go through the
5655 current @code{target_ops} vector.
5657 You can't assume that the host and target machines are the same machine
5658 (except in the ``native'' support modules). In particular, you can't
5659 assume that the target machine's header files will be available on the
5660 host machine. Target code must bring along its own header files --
5661 written from scratch or explicitly donated by their owner, to avoid
5665 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
5666 to write the code portably than to conditionalize it for various
5669 @cindex system dependencies
5670 New @code{#ifdef}'s which test for specific compilers or manufacturers
5671 or operating systems are unacceptable. All @code{#ifdef}'s should test
5672 for features. The information about which configurations contain which
5673 features should be segregated into the configuration files. Experience
5674 has proven far too often that a feature unique to one particular system
5675 often creeps into other systems; and that a conditional based on some
5676 predefined macro for your current system will become worthless over
5677 time, as new versions of your system come out that behave differently
5678 with regard to this feature.
5680 Adding code that handles specific architectures, operating systems,
5681 target interfaces, or hosts, is not acceptable in generic code.
5683 @cindex portable file name handling
5684 @cindex file names, portability
5685 One particularly notorious area where system dependencies tend to
5686 creep in is handling of file names. The mainline @value{GDBN} code
5687 assumes Posix semantics of file names: absolute file names begin with
5688 a forward slash @file{/}, slashes are used to separate leading
5689 directories, case-sensitive file names. These assumptions are not
5690 necessarily true on non-Posix systems such as MS-Windows. To avoid
5691 system-dependent code where you need to take apart or construct a file
5692 name, use the following portable macros:
5695 @findex HAVE_DOS_BASED_FILE_SYSTEM
5696 @item HAVE_DOS_BASED_FILE_SYSTEM
5697 This preprocessing symbol is defined to a non-zero value on hosts
5698 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
5699 symbol to write conditional code which should only be compiled for
5702 @findex IS_DIR_SEPARATOR
5703 @item IS_DIR_SEPARATOR (@var{c})
5704 Evaluates to a non-zero value if @var{c} is a directory separator
5705 character. On Unix and GNU/Linux systems, only a slash @file{/} is
5706 such a character, but on Windows, both @file{/} and @file{\} will
5709 @findex IS_ABSOLUTE_PATH
5710 @item IS_ABSOLUTE_PATH (@var{file})
5711 Evaluates to a non-zero value if @var{file} is an absolute file name.
5712 For Unix and GNU/Linux hosts, a name which begins with a slash
5713 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
5714 @file{x:\bar} are also absolute file names.
5716 @findex FILENAME_CMP
5717 @item FILENAME_CMP (@var{f1}, @var{f2})
5718 Calls a function which compares file names @var{f1} and @var{f2} as
5719 appropriate for the underlying host filesystem. For Posix systems,
5720 this simply calls @code{strcmp}; on case-insensitive filesystems it
5721 will call @code{strcasecmp} instead.
5723 @findex DIRNAME_SEPARATOR
5724 @item DIRNAME_SEPARATOR
5725 Evaluates to a character which separates directories in
5726 @code{PATH}-style lists, typically held in environment variables.
5727 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
5729 @findex SLASH_STRING
5731 This evaluates to a constant string you should use to produce an
5732 absolute filename from leading directories and the file's basename.
5733 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
5734 @code{"\\"} for some Windows-based ports.
5737 In addition to using these macros, be sure to use portable library
5738 functions whenever possible. For example, to extract a directory or a
5739 basename part from a file name, use the @code{dirname} and
5740 @code{basename} library functions (available in @code{libiberty} for
5741 platforms which don't provide them), instead of searching for a slash
5742 with @code{strrchr}.
5744 Another way to generalize @value{GDBN} along a particular interface is with an
5745 attribute struct. For example, @value{GDBN} has been generalized to handle
5746 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
5747 by defining the @code{target_ops} structure and having a current target (as
5748 well as a stack of targets below it, for memory references). Whenever
5749 something needs to be done that depends on which remote interface we are
5750 using, a flag in the current target_ops structure is tested (e.g.,
5751 @code{target_has_stack}), or a function is called through a pointer in the
5752 current target_ops structure. In this way, when a new remote interface
5753 is added, only one module needs to be touched---the one that actually
5754 implements the new remote interface. Other examples of
5755 attribute-structs are BFD access to multiple kinds of object file
5756 formats, or @value{GDBN}'s access to multiple source languages.
5758 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
5759 the code interfacing between @code{ptrace} and the rest of
5760 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
5761 something was very painful. In @value{GDBN} 4.x, these have all been
5762 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
5763 with variations between systems the same way any system-independent
5764 file would (hooks, @code{#if defined}, etc.), and machines which are
5765 radically different don't need to use @file{infptrace.c} at all.
5767 All debugging code must be controllable using the @samp{set debug
5768 @var{module}} command. Do not use @code{printf} to print trace
5769 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
5770 @code{#ifdef DEBUG}.
5775 @chapter Porting @value{GDBN}
5776 @cindex porting to new machines
5778 Most of the work in making @value{GDBN} compile on a new machine is in
5779 specifying the configuration of the machine. This is done in a
5780 dizzying variety of header files and configuration scripts, which we
5781 hope to make more sensible soon. Let's say your new host is called an
5782 @var{xyz} (e.g., @samp{sun4}), and its full three-part configuration
5783 name is @code{@var{arch}-@var{xvend}-@var{xos}} (e.g.,
5784 @samp{sparc-sun-sunos4}). In particular:
5788 In the top level directory, edit @file{config.sub} and add @var{arch},
5789 @var{xvend}, and @var{xos} to the lists of supported architectures,
5790 vendors, and operating systems near the bottom of the file. Also, add
5791 @var{xyz} as an alias that maps to
5792 @code{@var{arch}-@var{xvend}-@var{xos}}. You can test your changes by
5796 ./config.sub @var{xyz}
5803 ./config.sub @code{@var{arch}-@var{xvend}-@var{xos}}
5807 which should both respond with @code{@var{arch}-@var{xvend}-@var{xos}}
5808 and no error messages.
5811 You need to port BFD, if that hasn't been done already. Porting BFD is
5812 beyond the scope of this manual.
5815 To configure @value{GDBN} itself, edit @file{gdb/configure.host} to recognize
5816 your system and set @code{gdb_host} to @var{xyz}, and (unless your
5817 desired target is already available) also edit @file{gdb/configure.tgt},
5818 setting @code{gdb_target} to something appropriate (for instance,
5821 @emph{Maintainer's note: Work in progress. The file
5822 @file{gdb/configure.host} originally needed to be modified when either a
5823 new native target or a new host machine was being added to @value{GDBN}.
5824 Recent changes have removed this requirement. The file now only needs
5825 to be modified when adding a new native configuration. This will likely
5826 changed again in the future.}
5829 Finally, you'll need to specify and define @value{GDBN}'s host-, native-, and
5830 target-dependent @file{.h} and @file{.c} files used for your
5834 @node Versions and Branches
5835 @chapter Versions and Branches
5839 @value{GDBN}'s version is determined by the file
5840 @file{gdb/version.in} and takes one of the following forms:
5843 @item @var{major}.@var{minor}
5844 @itemx @var{major}.@var{minor}.@var{patchlevel}
5845 an official release (e.g., 6.2 or 6.2.1)
5846 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}
5847 a snapshot taken at @var{YYYY}-@var{MM}-@var{DD}-gmt (e.g.,
5848 6.1.50.20020302, 6.1.90.20020304, or 6.1.0.20020308)
5849 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}-cvs
5850 a @sc{cvs} check out drawn on @var{YYYY}-@var{MM}-@var{DD} (e.g.,
5851 6.1.50.20020302-cvs, 6.1.90.20020304-cvs, or 6.1.0.20020308-cvs)
5852 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD} (@var{vendor})
5853 a vendor specific release of @value{GDBN}, that while based on@*
5854 @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD},
5855 may include additional changes
5858 @value{GDBN}'s mainline uses the @var{major} and @var{minor} version
5859 numbers from the most recent release branch, with a @var{patchlevel}
5860 of 50. At the time each new release branch is created, the mainline's
5861 @var{major} and @var{minor} version numbers are updated.
5863 @value{GDBN}'s release branch is similar. When the branch is cut, the
5864 @var{patchlevel} is changed from 50 to 90. As draft releases are
5865 drawn from the branch, the @var{patchlevel} is incremented. Once the
5866 first release (@var{major}.@var{minor}) has been made, the
5867 @var{patchlevel} is set to 0 and updates have an incremented
5870 For snapshots, and @sc{cvs} check outs, it is also possible to
5871 identify the @sc{cvs} origin:
5874 @item @var{major}.@var{minor}.50.@var{YYYY}@var{MM}@var{DD}
5875 drawn from the @sc{head} of mainline @sc{cvs} (e.g., 6.1.50.20020302)
5876 @item @var{major}.@var{minor}.90.@var{YYYY}@var{MM}@var{DD}
5877 @itemx @var{major}.@var{minor}.91.@var{YYYY}@var{MM}@var{DD} @dots{}
5878 drawn from a release branch prior to the release (e.g.,
5880 @item @var{major}.@var{minor}.0.@var{YYYY}@var{MM}@var{DD}
5881 @itemx @var{major}.@var{minor}.1.@var{YYYY}@var{MM}@var{DD} @dots{}
5882 drawn from a release branch after the release (e.g., 6.2.0.20020308)
5885 If the previous @value{GDBN} version is 6.1 and the current version is
5886 6.2, then, substituting 6 for @var{major} and 1 or 2 for @var{minor},
5887 here's an illustration of a typical sequence:
5894 +--------------------------.
5897 6.2.50.20020303-cvs 6.1.90 (draft #1)
5899 6.2.50.20020304-cvs 6.1.90.20020304-cvs
5901 6.2.50.20020305-cvs 6.1.91 (draft #2)
5903 6.2.50.20020306-cvs 6.1.91.20020306-cvs
5905 6.2.50.20020307-cvs 6.2 (release)
5907 6.2.50.20020308-cvs 6.2.0.20020308-cvs
5909 6.2.50.20020309-cvs 6.2.1 (update)
5911 6.2.50.20020310-cvs <branch closed>
5915 +--------------------------.
5918 6.3.50.20020312-cvs 6.2.90 (draft #1)
5922 @section Release Branches
5923 @cindex Release Branches
5925 @value{GDBN} draws a release series (6.2, 6.2.1, @dots{}) from a
5926 single release branch, and identifies that branch using the @sc{cvs}
5930 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-branchpoint
5931 gdb_@var{major}_@var{minor}-branch
5932 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-release
5935 @emph{Pragmatics: To help identify the date at which a branch or
5936 release is made, both the branchpoint and release tags include the
5937 date that they are cut (@var{YYYY}@var{MM}@var{DD}) in the tag. The
5938 branch tag, denoting the head of the branch, does not need this.}
5940 @section Vendor Branches
5941 @cindex vendor branches
5943 To avoid version conflicts, vendors are expected to modify the file
5944 @file{gdb/version.in} to include a vendor unique alphabetic identifier
5945 (an official @value{GDBN} release never uses alphabetic characters in
5946 its version identifier). E.g., @samp{6.2widgit2}, or @samp{6.2 (Widgit
5949 @section Experimental Branches
5950 @cindex experimental branches
5952 @subsection Guidelines
5954 @value{GDBN} permits the creation of branches, cut from the @sc{cvs}
5955 repository, for experimental development. Branches make it possible
5956 for developers to share preliminary work, and maintainers to examine
5957 significant new developments.
5959 The following are a set of guidelines for creating such branches:
5963 @item a branch has an owner
5964 The owner can set further policy for a branch, but may not change the
5965 ground rules. In particular, they can set a policy for commits (be it
5966 adding more reviewers or deciding who can commit).
5968 @item all commits are posted
5969 All changes committed to a branch shall also be posted to
5970 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} patches
5971 mailing list}. While commentary on such changes are encouraged, people
5972 should remember that the changes only apply to a branch.
5974 @item all commits are covered by an assignment
5975 This ensures that all changes belong to the Free Software Foundation,
5976 and avoids the possibility that the branch may become contaminated.
5978 @item a branch is focused
5979 A focused branch has a single objective or goal, and does not contain
5980 unnecessary or irrelevant changes. Cleanups, where identified, being
5981 be pushed into the mainline as soon as possible.
5983 @item a branch tracks mainline
5984 This keeps the level of divergence under control. It also keeps the
5985 pressure on developers to push cleanups and other stuff into the
5988 @item a branch shall contain the entire @value{GDBN} module
5989 The @value{GDBN} module @code{gdb} should be specified when creating a
5990 branch (branches of individual files should be avoided). @xref{Tags}.
5992 @item a branch shall be branded using @file{version.in}
5993 The file @file{gdb/version.in} shall be modified so that it identifies
5994 the branch @var{owner} and branch @var{name}, e.g.,
5995 @samp{6.2.50.20030303_owner_name} or @samp{6.2 (Owner Name)}.
6002 To simplify the identification of @value{GDBN} branches, the following
6003 branch tagging convention is strongly recommended:
6007 @item @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
6008 @itemx @var{owner}_@var{name}-@var{YYYYMMDD}-branch
6009 The branch point and corresponding branch tag. @var{YYYYMMDD} is the
6010 date that the branch was created. A branch is created using the
6011 sequence: @anchor{experimental branch tags}
6013 cvs rtag @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint gdb
6014 cvs rtag -b -r @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint \
6015 @var{owner}_@var{name}-@var{YYYYMMDD}-branch gdb
6018 @item @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
6019 The tagged point, on the mainline, that was used when merging the branch
6020 on @var{yyyymmdd}. To merge in all changes since the branch was cut,
6021 use a command sequence like:
6023 cvs rtag @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint gdb
6025 -j@var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
6026 -j@var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
6029 Similar sequences can be used to just merge in changes since the last
6035 For further information on @sc{cvs}, see
6036 @uref{http://www.gnu.org/software/cvs/, Concurrent Versions System}.
6038 @node Start of New Year Procedure
6039 @chapter Start of New Year Procedure
6040 @cindex new year procedure
6042 At the start of each new year, the following actions should be performed:
6046 Rotate the ChangeLog file
6048 The current @file{ChangeLog} file should be renamed into
6049 @file{ChangeLog-YYYY} where YYYY is the year that has just passed.
6050 A new @file{ChangeLog} file should be created, and its contents should
6051 contain a reference to the previous ChangeLog. The following should
6052 also be preserved at the end of the new ChangeLog, in order to provide
6053 the appropriate settings when editing this file with Emacs:
6059 version-control: never
6064 Add an entry for the newly created ChangeLog file (@file{ChangeLog-YYYY})
6065 in @file{gdb/config/djgpp/fnchange.lst}.
6068 Update the copyright year in the startup message
6070 Update the copyright year in file @file{top.c}, function
6071 @code{print_gdb_version}.
6074 Add the new year in the copyright notices of all source and documentation
6075 files. This can be done semi-automatically by running the @code{copyright.sh}
6076 script. This script requires Emacs 22 or later to be installed.
6082 @chapter Releasing @value{GDBN}
6083 @cindex making a new release of gdb
6085 @section Branch Commit Policy
6087 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
6088 5.1 and 5.2 all used the below:
6092 The @file{gdb/MAINTAINERS} file still holds.
6094 Don't fix something on the branch unless/until it is also fixed in the
6095 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
6096 file is better than committing a hack.
6098 When considering a patch for the branch, suggested criteria include:
6099 Does it fix a build? Does it fix the sequence @kbd{break main; run}
6100 when debugging a static binary?
6102 The further a change is from the core of @value{GDBN}, the less likely
6103 the change will worry anyone (e.g., target specific code).
6105 Only post a proposal to change the core of @value{GDBN} after you've
6106 sent individual bribes to all the people listed in the
6107 @file{MAINTAINERS} file @t{;-)}
6110 @emph{Pragmatics: Provided updates are restricted to non-core
6111 functionality there is little chance that a broken change will be fatal.
6112 This means that changes such as adding a new architectures or (within
6113 reason) support for a new host are considered acceptable.}
6116 @section Obsoleting code
6118 Before anything else, poke the other developers (and around the source
6119 code) to see if there is anything that can be removed from @value{GDBN}
6120 (an old target, an unused file).
6122 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
6123 line. Doing this means that it is easy to identify something that has
6124 been obsoleted when greping through the sources.
6126 The process is done in stages --- this is mainly to ensure that the
6127 wider @value{GDBN} community has a reasonable opportunity to respond.
6128 Remember, everything on the Internet takes a week.
6132 Post the proposal on @email{gdb@@sources.redhat.com, the GDB mailing
6133 list} Creating a bug report to track the task's state, is also highly
6138 Post the proposal on @email{gdb-announce@@sources.redhat.com, the GDB
6139 Announcement mailing list}.
6143 Go through and edit all relevant files and lines so that they are
6144 prefixed with the word @code{OBSOLETE}.
6146 Wait until the next GDB version, containing this obsolete code, has been
6149 Remove the obsolete code.
6153 @emph{Maintainer note: While removing old code is regrettable it is
6154 hopefully better for @value{GDBN}'s long term development. Firstly it
6155 helps the developers by removing code that is either no longer relevant
6156 or simply wrong. Secondly since it removes any history associated with
6157 the file (effectively clearing the slate) the developer has a much freer
6158 hand when it comes to fixing broken files.}
6162 @section Before the Branch
6164 The most important objective at this stage is to find and fix simple
6165 changes that become a pain to track once the branch is created. For
6166 instance, configuration problems that stop @value{GDBN} from even
6167 building. If you can't get the problem fixed, document it in the
6168 @file{gdb/PROBLEMS} file.
6170 @subheading Prompt for @file{gdb/NEWS}
6172 People always forget. Send a post reminding them but also if you know
6173 something interesting happened add it yourself. The @code{schedule}
6174 script will mention this in its e-mail.
6176 @subheading Review @file{gdb/README}
6178 Grab one of the nightly snapshots and then walk through the
6179 @file{gdb/README} looking for anything that can be improved. The
6180 @code{schedule} script will mention this in its e-mail.
6182 @subheading Refresh any imported files.
6184 A number of files are taken from external repositories. They include:
6188 @file{texinfo/texinfo.tex}
6190 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
6193 @file{etc/standards.texi}, @file{etc/make-stds.texi}
6196 @subheading Check the ARI
6198 @uref{http://sources.redhat.com/gdb/ari,,A.R.I.} is an @code{awk} script
6199 (Awk Regression Index ;-) that checks for a number of errors and coding
6200 conventions. The checks include things like using @code{malloc} instead
6201 of @code{xmalloc} and file naming problems. There shouldn't be any
6204 @subsection Review the bug data base
6206 Close anything obviously fixed.
6208 @subsection Check all cross targets build
6210 The targets are listed in @file{gdb/MAINTAINERS}.
6213 @section Cut the Branch
6215 @subheading Create the branch
6220 $ V=`echo $v | sed 's/\./_/g'`
6221 $ D=`date -u +%Y-%m-%d`
6224 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6225 -D $D-gmt gdb_$V-$D-branchpoint insight
6226 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag
6227 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight
6230 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6231 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight
6232 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6233 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight
6241 By using @kbd{-D YYYY-MM-DD-gmt}, the branch is forced to an exact
6244 The trunk is first tagged so that the branch point can easily be found.
6246 Insight, which includes @value{GDBN}, is tagged at the same time.
6248 @file{version.in} gets bumped to avoid version number conflicts.
6250 The reading of @file{.cvsrc} is disabled using @file{-f}.
6253 @subheading Update @file{version.in}
6258 $ V=`echo $v | sed 's/\./_/g'`
6262 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src co \
6263 -r gdb_$V-branch src/gdb/version.in
6264 cvs -f -d :ext:sources.redhat.com:/cvs/src co
6265 -r gdb_5_2-branch src/gdb/version.in
6267 U src/gdb/version.in
6269 $ echo $u.90-0000-00-00-cvs > version.in
6271 5.1.90-0000-00-00-cvs
6272 $ cvs -f commit version.in
6277 @file{0000-00-00} is used as a date to pump prime the version.in update
6280 @file{.90} and the previous branch version are used as fairly arbitrary
6281 initial branch version number.
6285 @subheading Update the web and news pages
6289 @subheading Tweak cron to track the new branch
6291 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
6292 This file needs to be updated so that:
6296 A daily timestamp is added to the file @file{version.in}.
6298 The new branch is included in the snapshot process.
6302 See the file @file{gdbadmin/cron/README} for how to install the updated
6305 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
6306 any changes. That file is copied to both the branch/ and current/
6307 snapshot directories.
6310 @subheading Update the NEWS and README files
6312 The @file{NEWS} file needs to be updated so that on the branch it refers
6313 to @emph{changes in the current release} while on the trunk it also
6314 refers to @emph{changes since the current release}.
6316 The @file{README} file needs to be updated so that it refers to the
6319 @subheading Post the branch info
6321 Send an announcement to the mailing lists:
6325 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6327 @email{gdb@@sources.redhat.com, GDB Discussion mailing list} and
6328 @email{gdb-testers@@sources.redhat.com, GDB Testers mailing list}
6331 @emph{Pragmatics: The branch creation is sent to the announce list to
6332 ensure that people people not subscribed to the higher volume discussion
6335 The announcement should include:
6341 How to check out the branch using CVS.
6343 The date/number of weeks until the release.
6345 The branch commit policy still holds.
6348 @section Stabilize the branch
6350 Something goes here.
6352 @section Create a Release
6354 The process of creating and then making available a release is broken
6355 down into a number of stages. The first part addresses the technical
6356 process of creating a releasable tar ball. The later stages address the
6357 process of releasing that tar ball.
6359 When making a release candidate just the first section is needed.
6361 @subsection Create a release candidate
6363 The objective at this stage is to create a set of tar balls that can be
6364 made available as a formal release (or as a less formal release
6367 @subsubheading Freeze the branch
6369 Send out an e-mail notifying everyone that the branch is frozen to
6370 @email{gdb-patches@@sources.redhat.com}.
6372 @subsubheading Establish a few defaults.
6377 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
6379 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
6383 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
6385 /home/gdbadmin/bin/autoconf
6394 Check the @code{autoconf} version carefully. You want to be using the
6395 version taken from the @file{binutils} snapshot directory, which can be
6396 found at @uref{ftp://sources.redhat.com/pub/binutils/}. It is very
6397 unlikely that a system installed version of @code{autoconf} (e.g.,
6398 @file{/usr/bin/autoconf}) is correct.
6401 @subsubheading Check out the relevant modules:
6404 $ for m in gdb insight
6406 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
6416 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
6417 any confusion between what is written here and what your local
6418 @code{cvs} really does.
6421 @subsubheading Update relevant files.
6427 Major releases get their comments added as part of the mainline. Minor
6428 releases should probably mention any significant bugs that were fixed.
6430 Don't forget to include the @file{ChangeLog} entry.
6433 $ emacs gdb/src/gdb/NEWS
6438 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
6439 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6444 You'll need to update:
6456 $ emacs gdb/src/gdb/README
6461 $ cp gdb/src/gdb/README insight/src/gdb/README
6462 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6465 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
6466 before the initial branch was cut so just a simple substitute is needed
6469 @emph{Maintainer note: Other projects generate @file{README} and
6470 @file{INSTALL} from the core documentation. This might be worth
6473 @item gdb/version.in
6476 $ echo $v > gdb/src/gdb/version.in
6477 $ cat gdb/src/gdb/version.in
6479 $ emacs gdb/src/gdb/version.in
6482 ... Bump to version ...
6484 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
6485 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6490 @subsubheading Do the dirty work
6492 This is identical to the process used to create the daily snapshot.
6495 $ for m in gdb insight
6497 ( cd $m/src && gmake -f src-release $m.tar )
6501 If the top level source directory does not have @file{src-release}
6502 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
6505 $ for m in gdb insight
6507 ( cd $m/src && gmake -f Makefile.in $m.tar )
6511 @subsubheading Check the source files
6513 You're looking for files that have mysteriously disappeared.
6514 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
6515 for the @file{version.in} update @kbd{cronjob}.
6518 $ ( cd gdb/src && cvs -f -q -n update )
6522 @dots{} lots of generated files @dots{}
6527 @dots{} lots of generated files @dots{}
6532 @emph{Don't worry about the @file{gdb.info-??} or
6533 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
6534 was also generated only something strange with CVS means that they
6535 didn't get suppressed). Fixing it would be nice though.}
6537 @subsubheading Create compressed versions of the release
6543 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
6544 $ for m in gdb insight
6546 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
6547 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
6557 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
6558 in that mode, @code{gzip} does not know the name of the file and, hence,
6559 can not include it in the compressed file. This is also why the release
6560 process runs @code{tar} and @code{bzip2} as separate passes.
6563 @subsection Sanity check the tar ball
6565 Pick a popular machine (Solaris/PPC?) and try the build on that.
6568 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
6573 $ ./gdb/gdb ./gdb/gdb
6577 Breakpoint 1 at 0x80732bc: file main.c, line 734.
6579 Starting program: /tmp/gdb-5.2/gdb/gdb
6581 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
6582 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
6584 $1 = @{argc = 136426532, argv = 0x821b7f0@}
6588 @subsection Make a release candidate available
6590 If this is a release candidate then the only remaining steps are:
6594 Commit @file{version.in} and @file{ChangeLog}
6596 Tweak @file{version.in} (and @file{ChangeLog} to read
6597 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
6598 process can restart.
6600 Make the release candidate available in
6601 @uref{ftp://sources.redhat.com/pub/gdb/snapshots/branch}
6603 Notify the relevant mailing lists ( @email{gdb@@sources.redhat.com} and
6604 @email{gdb-testers@@sources.redhat.com} that the candidate is available.
6607 @subsection Make a formal release available
6609 (And you thought all that was required was to post an e-mail.)
6611 @subsubheading Install on sware
6613 Copy the new files to both the release and the old release directory:
6616 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
6617 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
6621 Clean up the releases directory so that only the most recent releases
6622 are available (e.g. keep 5.2 and 5.2.1 but remove 5.1):
6625 $ cd ~ftp/pub/gdb/releases
6630 Update the file @file{README} and @file{.message} in the releases
6637 $ ln README .message
6640 @subsubheading Update the web pages.
6644 @item htdocs/download/ANNOUNCEMENT
6645 This file, which is posted as the official announcement, includes:
6648 General announcement.
6650 News. If making an @var{M}.@var{N}.1 release, retain the news from
6651 earlier @var{M}.@var{N} release.
6656 @item htdocs/index.html
6657 @itemx htdocs/news/index.html
6658 @itemx htdocs/download/index.html
6659 These files include:
6662 Announcement of the most recent release.
6664 News entry (remember to update both the top level and the news directory).
6666 These pages also need to be regenerate using @code{index.sh}.
6668 @item download/onlinedocs/
6669 You need to find the magic command that is used to generate the online
6670 docs from the @file{.tar.bz2}. The best way is to look in the output
6671 from one of the nightly @code{cron} jobs and then just edit accordingly.
6675 $ ~/ss/update-web-docs \
6676 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6678 /www/sourceware/htdocs/gdb/download/onlinedocs \
6683 Just like the online documentation. Something like:
6686 $ /bin/sh ~/ss/update-web-ari \
6687 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6689 /www/sourceware/htdocs/gdb/download/ari \
6695 @subsubheading Shadow the pages onto gnu
6697 Something goes here.
6700 @subsubheading Install the @value{GDBN} tar ball on GNU
6702 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
6703 @file{~ftp/gnu/gdb}.
6705 @subsubheading Make the @file{ANNOUNCEMENT}
6707 Post the @file{ANNOUNCEMENT} file you created above to:
6711 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6713 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
6714 day or so to let things get out)
6716 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
6721 The release is out but you're still not finished.
6723 @subsubheading Commit outstanding changes
6725 In particular you'll need to commit any changes to:
6729 @file{gdb/ChangeLog}
6731 @file{gdb/version.in}
6738 @subsubheading Tag the release
6743 $ d=`date -u +%Y-%m-%d`
6746 $ ( cd insight/src/gdb && cvs -f -q update )
6747 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
6750 Insight is used since that contains more of the release than
6753 @subsubheading Mention the release on the trunk
6755 Just put something in the @file{ChangeLog} so that the trunk also
6756 indicates when the release was made.
6758 @subsubheading Restart @file{gdb/version.in}
6760 If @file{gdb/version.in} does not contain an ISO date such as
6761 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
6762 committed all the release changes it can be set to
6763 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
6764 is important - it affects the snapshot process).
6766 Don't forget the @file{ChangeLog}.
6768 @subsubheading Merge into trunk
6770 The files committed to the branch may also need changes merged into the
6773 @subsubheading Revise the release schedule
6775 Post a revised release schedule to @email{gdb@@sources.redhat.com, GDB
6776 Discussion List} with an updated announcement. The schedule can be
6777 generated by running:
6780 $ ~/ss/schedule `date +%s` schedule
6784 The first parameter is approximate date/time in seconds (from the epoch)
6785 of the most recent release.
6787 Also update the schedule @code{cronjob}.
6789 @section Post release
6791 Remove any @code{OBSOLETE} code.
6798 The testsuite is an important component of the @value{GDBN} package.
6799 While it is always worthwhile to encourage user testing, in practice
6800 this is rarely sufficient; users typically use only a small subset of
6801 the available commands, and it has proven all too common for a change
6802 to cause a significant regression that went unnoticed for some time.
6804 The @value{GDBN} testsuite uses the DejaGNU testing framework. The
6805 tests themselves are calls to various @code{Tcl} procs; the framework
6806 runs all the procs and summarizes the passes and fails.
6808 @section Using the Testsuite
6810 @cindex running the test suite
6811 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
6812 testsuite's objdir) and type @code{make check}. This just sets up some
6813 environment variables and invokes DejaGNU's @code{runtest} script. While
6814 the testsuite is running, you'll get mentions of which test file is in use,
6815 and a mention of any unexpected passes or fails. When the testsuite is
6816 finished, you'll get a summary that looks like this:
6821 # of expected passes 6016
6822 # of unexpected failures 58
6823 # of unexpected successes 5
6824 # of expected failures 183
6825 # of unresolved testcases 3
6826 # of untested testcases 5
6829 To run a specific test script, type:
6831 make check RUNTESTFLAGS='@var{tests}'
6833 where @var{tests} is a list of test script file names, separated by
6836 The ideal test run consists of expected passes only; however, reality
6837 conspires to keep us from this ideal. Unexpected failures indicate
6838 real problems, whether in @value{GDBN} or in the testsuite. Expected
6839 failures are still failures, but ones which have been decided are too
6840 hard to deal with at the time; for instance, a test case might work
6841 everywhere except on AIX, and there is no prospect of the AIX case
6842 being fixed in the near future. Expected failures should not be added
6843 lightly, since you may be masking serious bugs in @value{GDBN}.
6844 Unexpected successes are expected fails that are passing for some
6845 reason, while unresolved and untested cases often indicate some minor
6846 catastrophe, such as the compiler being unable to deal with a test
6849 When making any significant change to @value{GDBN}, you should run the
6850 testsuite before and after the change, to confirm that there are no
6851 regressions. Note that truly complete testing would require that you
6852 run the testsuite with all supported configurations and a variety of
6853 compilers; however this is more than really necessary. In many cases
6854 testing with a single configuration is sufficient. Other useful
6855 options are to test one big-endian (Sparc) and one little-endian (x86)
6856 host, a cross config with a builtin simulator (powerpc-eabi,
6857 mips-elf), or a 64-bit host (Alpha).
6859 If you add new functionality to @value{GDBN}, please consider adding
6860 tests for it as well; this way future @value{GDBN} hackers can detect
6861 and fix their changes that break the functionality you added.
6862 Similarly, if you fix a bug that was not previously reported as a test
6863 failure, please add a test case for it. Some cases are extremely
6864 difficult to test, such as code that handles host OS failures or bugs
6865 in particular versions of compilers, and it's OK not to try to write
6866 tests for all of those.
6868 DejaGNU supports separate build, host, and target machines. However,
6869 some @value{GDBN} test scripts do not work if the build machine and
6870 the host machine are not the same. In such an environment, these scripts
6871 will give a result of ``UNRESOLVED'', like this:
6874 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
6877 @section Testsuite Organization
6879 @cindex test suite organization
6880 The testsuite is entirely contained in @file{gdb/testsuite}. While the
6881 testsuite includes some makefiles and configury, these are very minimal,
6882 and used for little besides cleaning up, since the tests themselves
6883 handle the compilation of the programs that @value{GDBN} will run. The file
6884 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
6885 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
6886 configuration-specific files, typically used for special-purpose
6887 definitions of procs like @code{gdb_load} and @code{gdb_start}.
6889 The tests themselves are to be found in @file{testsuite/gdb.*} and
6890 subdirectories of those. The names of the test files must always end
6891 with @file{.exp}. DejaGNU collects the test files by wildcarding
6892 in the test directories, so both subdirectories and individual files
6893 get chosen and run in alphabetical order.
6895 The following table lists the main types of subdirectories and what they
6896 are for. Since DejaGNU finds test files no matter where they are
6897 located, and since each test file sets up its own compilation and
6898 execution environment, this organization is simply for convenience and
6903 This is the base testsuite. The tests in it should apply to all
6904 configurations of @value{GDBN} (but generic native-only tests may live here).
6905 The test programs should be in the subset of C that is valid K&R,
6906 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
6909 @item gdb.@var{lang}
6910 Language-specific tests for any language @var{lang} besides C. Examples are
6911 @file{gdb.cp} and @file{gdb.java}.
6913 @item gdb.@var{platform}
6914 Non-portable tests. The tests are specific to a specific configuration
6915 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
6918 @item gdb.@var{compiler}
6919 Tests specific to a particular compiler. As of this writing (June
6920 1999), there aren't currently any groups of tests in this category that
6921 couldn't just as sensibly be made platform-specific, but one could
6922 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
6925 @item gdb.@var{subsystem}
6926 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
6927 instance, @file{gdb.disasm} exercises various disassemblers, while
6928 @file{gdb.stabs} tests pathways through the stabs symbol reader.
6931 @section Writing Tests
6932 @cindex writing tests
6934 In many areas, the @value{GDBN} tests are already quite comprehensive; you
6935 should be able to copy existing tests to handle new cases.
6937 You should try to use @code{gdb_test} whenever possible, since it
6938 includes cases to handle all the unexpected errors that might happen.
6939 However, it doesn't cost anything to add new test procedures; for
6940 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
6941 calls @code{gdb_test} multiple times.
6943 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
6944 necessary. Even if @value{GDBN} has several valid responses to
6945 a command, you can use @code{gdb_test_multiple}. Like @code{gdb_test},
6946 @code{gdb_test_multiple} recognizes internal errors and unexpected
6949 Do not write tests which expect a literal tab character from @value{GDBN}.
6950 On some operating systems (e.g.@: OpenBSD) the TTY layer expands tabs to
6951 spaces, so by the time @value{GDBN}'s output reaches expect the tab is gone.
6953 The source language programs do @emph{not} need to be in a consistent
6954 style. Since @value{GDBN} is used to debug programs written in many different
6955 styles, it's worth having a mix of styles in the testsuite; for
6956 instance, some @value{GDBN} bugs involving the display of source lines would
6957 never manifest themselves if the programs used GNU coding style
6964 Check the @file{README} file, it often has useful information that does not
6965 appear anywhere else in the directory.
6968 * Getting Started:: Getting started working on @value{GDBN}
6969 * Debugging GDB:: Debugging @value{GDBN} with itself
6972 @node Getting Started,,, Hints
6974 @section Getting Started
6976 @value{GDBN} is a large and complicated program, and if you first starting to
6977 work on it, it can be hard to know where to start. Fortunately, if you
6978 know how to go about it, there are ways to figure out what is going on.
6980 This manual, the @value{GDBN} Internals manual, has information which applies
6981 generally to many parts of @value{GDBN}.
6983 Information about particular functions or data structures are located in
6984 comments with those functions or data structures. If you run across a
6985 function or a global variable which does not have a comment correctly
6986 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
6987 free to submit a bug report, with a suggested comment if you can figure
6988 out what the comment should say. If you find a comment which is
6989 actually wrong, be especially sure to report that.
6991 Comments explaining the function of macros defined in host, target, or
6992 native dependent files can be in several places. Sometimes they are
6993 repeated every place the macro is defined. Sometimes they are where the
6994 macro is used. Sometimes there is a header file which supplies a
6995 default definition of the macro, and the comment is there. This manual
6996 also documents all the available macros.
6997 @c (@pxref{Host Conditionals}, @pxref{Target
6998 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
7001 Start with the header files. Once you have some idea of how
7002 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
7003 @file{gdbtypes.h}), you will find it much easier to understand the
7004 code which uses and creates those symbol tables.
7006 You may wish to process the information you are getting somehow, to
7007 enhance your understanding of it. Summarize it, translate it to another
7008 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
7009 the code to predict what a test case would do and write the test case
7010 and verify your prediction, etc. If you are reading code and your eyes
7011 are starting to glaze over, this is a sign you need to use a more active
7014 Once you have a part of @value{GDBN} to start with, you can find more
7015 specifically the part you are looking for by stepping through each
7016 function with the @code{next} command. Do not use @code{step} or you
7017 will quickly get distracted; when the function you are stepping through
7018 calls another function try only to get a big-picture understanding
7019 (perhaps using the comment at the beginning of the function being
7020 called) of what it does. This way you can identify which of the
7021 functions being called by the function you are stepping through is the
7022 one which you are interested in. You may need to examine the data
7023 structures generated at each stage, with reference to the comments in
7024 the header files explaining what the data structures are supposed to
7027 Of course, this same technique can be used if you are just reading the
7028 code, rather than actually stepping through it. The same general
7029 principle applies---when the code you are looking at calls something
7030 else, just try to understand generally what the code being called does,
7031 rather than worrying about all its details.
7033 @cindex command implementation
7034 A good place to start when tracking down some particular area is with
7035 a command which invokes that feature. Suppose you want to know how
7036 single-stepping works. As a @value{GDBN} user, you know that the
7037 @code{step} command invokes single-stepping. The command is invoked
7038 via command tables (see @file{command.h}); by convention the function
7039 which actually performs the command is formed by taking the name of
7040 the command and adding @samp{_command}, or in the case of an
7041 @code{info} subcommand, @samp{_info}. For example, the @code{step}
7042 command invokes the @code{step_command} function and the @code{info
7043 display} command invokes @code{display_info}. When this convention is
7044 not followed, you might have to use @code{grep} or @kbd{M-x
7045 tags-search} in emacs, or run @value{GDBN} on itself and set a
7046 breakpoint in @code{execute_command}.
7048 @cindex @code{bug-gdb} mailing list
7049 If all of the above fail, it may be appropriate to ask for information
7050 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
7051 wondering if anyone could give me some tips about understanding
7052 @value{GDBN}''---if we had some magic secret we would put it in this manual.
7053 Suggestions for improving the manual are always welcome, of course.
7055 @node Debugging GDB,,,Hints
7057 @section Debugging @value{GDBN} with itself
7058 @cindex debugging @value{GDBN}
7060 If @value{GDBN} is limping on your machine, this is the preferred way to get it
7061 fully functional. Be warned that in some ancient Unix systems, like
7062 Ultrix 4.2, a program can't be running in one process while it is being
7063 debugged in another. Rather than typing the command @kbd{@w{./gdb
7064 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
7065 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
7067 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
7068 @file{.gdbinit} file that sets up some simple things to make debugging
7069 gdb easier. The @code{info} command, when executed without a subcommand
7070 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
7071 gdb. See @file{.gdbinit} for details.
7073 If you use emacs, you will probably want to do a @code{make TAGS} after
7074 you configure your distribution; this will put the machine dependent
7075 routines for your local machine where they will be accessed first by
7078 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
7079 have run @code{fixincludes} if you are compiling with gcc.
7081 @section Submitting Patches
7083 @cindex submitting patches
7084 Thanks for thinking of offering your changes back to the community of
7085 @value{GDBN} users. In general we like to get well designed enhancements.
7086 Thanks also for checking in advance about the best way to transfer the
7089 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
7090 This manual summarizes what we believe to be clean design for @value{GDBN}.
7092 If the maintainers don't have time to put the patch in when it arrives,
7093 or if there is any question about a patch, it goes into a large queue
7094 with everyone else's patches and bug reports.
7096 @cindex legal papers for code contributions
7097 The legal issue is that to incorporate substantial changes requires a
7098 copyright assignment from you and/or your employer, granting ownership
7099 of the changes to the Free Software Foundation. You can get the
7100 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
7101 and asking for it. We recommend that people write in "All programs
7102 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
7103 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
7105 contributed with only one piece of legalese pushed through the
7106 bureaucracy and filed with the FSF. We can't start merging changes until
7107 this paperwork is received by the FSF (their rules, which we follow
7108 since we maintain it for them).
7110 Technically, the easiest way to receive changes is to receive each
7111 feature as a small context diff or unidiff, suitable for @code{patch}.
7112 Each message sent to me should include the changes to C code and
7113 header files for a single feature, plus @file{ChangeLog} entries for
7114 each directory where files were modified, and diffs for any changes
7115 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
7116 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
7117 single feature, they can be split down into multiple messages.
7119 In this way, if we read and like the feature, we can add it to the
7120 sources with a single patch command, do some testing, and check it in.
7121 If you leave out the @file{ChangeLog}, we have to write one. If you leave
7122 out the doc, we have to puzzle out what needs documenting. Etc., etc.
7124 The reason to send each change in a separate message is that we will not
7125 install some of the changes. They'll be returned to you with questions
7126 or comments. If we're doing our job correctly, the message back to you
7127 will say what you have to fix in order to make the change acceptable.
7128 The reason to have separate messages for separate features is so that
7129 the acceptable changes can be installed while one or more changes are
7130 being reworked. If multiple features are sent in a single message, we
7131 tend to not put in the effort to sort out the acceptable changes from
7132 the unacceptable, so none of the features get installed until all are
7135 If this sounds painful or authoritarian, well, it is. But we get a lot
7136 of bug reports and a lot of patches, and many of them don't get
7137 installed because we don't have the time to finish the job that the bug
7138 reporter or the contributor could have done. Patches that arrive
7139 complete, working, and well designed, tend to get installed on the day
7140 they arrive. The others go into a queue and get installed as time
7141 permits, which, since the maintainers have many demands to meet, may not
7142 be for quite some time.
7144 Please send patches directly to
7145 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} maintainers}.
7147 @section Obsolete Conditionals
7148 @cindex obsolete code
7150 Fragments of old code in @value{GDBN} sometimes reference or set the following
7151 configuration macros. They should not be used by new code, and old uses
7152 should be removed as those parts of the debugger are otherwise touched.
7155 @item STACK_END_ADDR
7156 This macro used to define where the end of the stack appeared, for use
7157 in interpreting core file formats that don't record this address in the
7158 core file itself. This information is now configured in BFD, and @value{GDBN}
7159 gets the info portably from there. The values in @value{GDBN}'s configuration
7160 files should be moved into BFD configuration files (if needed there),
7161 and deleted from all of @value{GDBN}'s config files.
7163 Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR
7164 is so old that it has never been converted to use BFD. Now that's old!
7168 @section Build Script
7170 @cindex build script
7172 The script @file{gdb_buildall.sh} builds @value{GDBN} with flag
7173 @option{--enable-targets=all} set. This builds @value{GDBN} with all supported
7174 targets activated. This helps testing @value{GDBN} when doing changes that
7175 affect more than one architecture and is much faster than using
7176 @file{gdb_mbuild.sh}.
7178 After building @value{GDBN} the script checks which architectures are
7179 supported and then switches the current architecture to each of those to get
7180 information about the architecture. The test results are stored in log files
7181 in the directory the script was called from.
7183 @include observer.texi