1 \input texinfo @c -*-texinfo-*-
2 @c Copyright (C) 1988-2019 Free Software Foundation, Inc.
5 @c makeinfo ignores cmds prev to setfilename, so its arg cannot make use
6 @c of @set vars. However, you can override filename with makeinfo -o.
13 @settitle Debugging with @value{GDBN}
14 @setchapternewpage odd
23 @c To avoid file-name clashes between index.html and Index.html, when
24 @c the manual is produced on a Posix host and then moved to a
25 @c case-insensitive filesystem (e.g., MS-Windows), we separate the
26 @c indices into two: Concept Index and all the rest.
30 @c readline appendices use @vindex, @findex and @ftable,
31 @c annotate.texi and gdbmi use @findex.
34 @c !!set GDB manual's edition---not the same as GDB version!
35 @c This is updated by GNU Press.
38 @c !!set GDB edit command default editor
41 @c THIS MANUAL REQUIRES TEXINFO 4.0 OR LATER.
43 @c This is a dir.info fragment to support semi-automated addition of
44 @c manuals to an info tree.
45 @dircategory Software development
47 * Gdb: (gdb). The GNU debugger.
48 * gdbserver: (gdb) Server. The GNU debugging server.
52 @c man begin COPYRIGHT
53 Copyright @copyright{} 1988-2019 Free Software Foundation, Inc.
55 Permission is granted to copy, distribute and/or modify this document
56 under the terms of the GNU Free Documentation License, Version 1.3 or
57 any later version published by the Free Software Foundation; with the
58 Invariant Sections being ``Free Software'' and ``Free Software Needs
59 Free Documentation'', with the Front-Cover Texts being ``A GNU Manual,''
60 and with the Back-Cover Texts as in (a) below.
62 (a) The FSF's Back-Cover Text is: ``You are free to copy and modify
63 this GNU Manual. Buying copies from GNU Press supports the FSF in
64 developing GNU and promoting software freedom.''
69 This file documents the @sc{gnu} debugger @value{GDBN}.
71 This is the @value{EDITION} Edition, of @cite{Debugging with
72 @value{GDBN}: the @sc{gnu} Source-Level Debugger} for @value{GDBN}
73 @ifset VERSION_PACKAGE
74 @value{VERSION_PACKAGE}
76 Version @value{GDBVN}.
82 @title Debugging with @value{GDBN}
83 @subtitle The @sc{gnu} Source-Level Debugger
85 @subtitle @value{EDITION} Edition, for @value{GDBN} version @value{GDBVN}
86 @ifset VERSION_PACKAGE
88 @subtitle @value{VERSION_PACKAGE}
90 @author Richard Stallman, Roland Pesch, Stan Shebs, et al.
94 \hfill (Send bugs and comments on @value{GDBN} to @value{BUGURL}.)\par
95 \hfill {\it Debugging with @value{GDBN}}\par
96 \hfill \TeX{}info \texinfoversion\par
100 @vskip 0pt plus 1filll
101 Published by the Free Software Foundation @*
102 51 Franklin Street, Fifth Floor,
103 Boston, MA 02110-1301, USA@*
104 ISBN 978-0-9831592-3-0 @*
111 @node Top, Summary, (dir), (dir)
113 @top Debugging with @value{GDBN}
115 This file describes @value{GDBN}, the @sc{gnu} symbolic debugger.
117 This is the @value{EDITION} Edition, for @value{GDBN}
118 @ifset VERSION_PACKAGE
119 @value{VERSION_PACKAGE}
121 Version @value{GDBVN}.
123 Copyright (C) 1988-2019 Free Software Foundation, Inc.
125 This edition of the GDB manual is dedicated to the memory of Fred
126 Fish. Fred was a long-standing contributor to GDB and to Free
127 software in general. We will miss him.
130 * Summary:: Summary of @value{GDBN}
131 * Sample Session:: A sample @value{GDBN} session
133 * Invocation:: Getting in and out of @value{GDBN}
134 * Commands:: @value{GDBN} commands
135 * Running:: Running programs under @value{GDBN}
136 * Stopping:: Stopping and continuing
137 * Reverse Execution:: Running programs backward
138 * Process Record and Replay:: Recording inferior's execution and replaying it
139 * Stack:: Examining the stack
140 * Source:: Examining source files
141 * Data:: Examining data
142 * Optimized Code:: Debugging optimized code
143 * Macros:: Preprocessor Macros
144 * Tracepoints:: Debugging remote targets non-intrusively
145 * Overlays:: Debugging programs that use overlays
147 * Languages:: Using @value{GDBN} with different languages
149 * Symbols:: Examining the symbol table
150 * Altering:: Altering execution
151 * GDB Files:: @value{GDBN} files
152 * Targets:: Specifying a debugging target
153 * Remote Debugging:: Debugging remote programs
154 * Configurations:: Configuration-specific information
155 * Controlling GDB:: Controlling @value{GDBN}
156 * Extending GDB:: Extending @value{GDBN}
157 * Interpreters:: Command Interpreters
158 * TUI:: @value{GDBN} Text User Interface
159 * Emacs:: Using @value{GDBN} under @sc{gnu} Emacs
160 * GDB/MI:: @value{GDBN}'s Machine Interface.
161 * Annotations:: @value{GDBN}'s annotation interface.
162 * JIT Interface:: Using the JIT debugging interface.
163 * In-Process Agent:: In-Process Agent
165 * GDB Bugs:: Reporting bugs in @value{GDBN}
167 @ifset SYSTEM_READLINE
168 * Command Line Editing: (rluserman). Command Line Editing
169 * Using History Interactively: (history). Using History Interactively
171 @ifclear SYSTEM_READLINE
172 * Command Line Editing:: Command Line Editing
173 * Using History Interactively:: Using History Interactively
175 * In Memoriam:: In Memoriam
176 * Formatting Documentation:: How to format and print @value{GDBN} documentation
177 * Installing GDB:: Installing GDB
178 * Maintenance Commands:: Maintenance Commands
179 * Remote Protocol:: GDB Remote Serial Protocol
180 * Agent Expressions:: The GDB Agent Expression Mechanism
181 * Target Descriptions:: How targets can describe themselves to
183 * Operating System Information:: Getting additional information from
185 * Trace File Format:: GDB trace file format
186 * Index Section Format:: .gdb_index section format
187 * Man Pages:: Manual pages
188 * Copying:: GNU General Public License says
189 how you can copy and share GDB
190 * GNU Free Documentation License:: The license for this documentation
191 * Concept Index:: Index of @value{GDBN} concepts
192 * Command and Variable Index:: Index of @value{GDBN} commands, variables,
193 functions, and Python data types
201 @unnumbered Summary of @value{GDBN}
203 The purpose of a debugger such as @value{GDBN} is to allow you to see what is
204 going on ``inside'' another program while it executes---or what another
205 program was doing at the moment it crashed.
207 @value{GDBN} can do four main kinds of things (plus other things in support of
208 these) to help you catch bugs in the act:
212 Start your program, specifying anything that might affect its behavior.
215 Make your program stop on specified conditions.
218 Examine what has happened, when your program has stopped.
221 Change things in your program, so you can experiment with correcting the
222 effects of one bug and go on to learn about another.
225 You can use @value{GDBN} to debug programs written in C and C@t{++}.
226 For more information, see @ref{Supported Languages,,Supported Languages}.
227 For more information, see @ref{C,,C and C++}.
229 Support for D is partial. For information on D, see
233 Support for Modula-2 is partial. For information on Modula-2, see
234 @ref{Modula-2,,Modula-2}.
236 Support for OpenCL C is partial. For information on OpenCL C, see
237 @ref{OpenCL C,,OpenCL C}.
240 Debugging Pascal programs which use sets, subranges, file variables, or
241 nested functions does not currently work. @value{GDBN} does not support
242 entering expressions, printing values, or similar features using Pascal
246 @value{GDBN} can be used to debug programs written in Fortran, although
247 it may be necessary to refer to some variables with a trailing
250 @value{GDBN} can be used to debug programs written in Objective-C,
251 using either the Apple/NeXT or the GNU Objective-C runtime.
254 * Free Software:: Freely redistributable software
255 * Free Documentation:: Free Software Needs Free Documentation
256 * Contributors:: Contributors to GDB
260 @unnumberedsec Free Software
262 @value{GDBN} is @dfn{free software}, protected by the @sc{gnu}
263 General Public License
264 (GPL). The GPL gives you the freedom to copy or adapt a licensed
265 program---but every person getting a copy also gets with it the
266 freedom to modify that copy (which means that they must get access to
267 the source code), and the freedom to distribute further copies.
268 Typical software companies use copyrights to limit your freedoms; the
269 Free Software Foundation uses the GPL to preserve these freedoms.
271 Fundamentally, the General Public License is a license which says that
272 you have these freedoms and that you cannot take these freedoms away
275 @node Free Documentation
276 @unnumberedsec Free Software Needs Free Documentation
278 The biggest deficiency in the free software community today is not in
279 the software---it is the lack of good free documentation that we can
280 include with the free software. Many of our most important
281 programs do not come with free reference manuals and free introductory
282 texts. Documentation is an essential part of any software package;
283 when an important free software package does not come with a free
284 manual and a free tutorial, that is a major gap. We have many such
287 Consider Perl, for instance. The tutorial manuals that people
288 normally use are non-free. How did this come about? Because the
289 authors of those manuals published them with restrictive terms---no
290 copying, no modification, source files not available---which exclude
291 them from the free software world.
293 That wasn't the first time this sort of thing happened, and it was far
294 from the last. Many times we have heard a GNU user eagerly describe a
295 manual that he is writing, his intended contribution to the community,
296 only to learn that he had ruined everything by signing a publication
297 contract to make it non-free.
299 Free documentation, like free software, is a matter of freedom, not
300 price. The problem with the non-free manual is not that publishers
301 charge a price for printed copies---that in itself is fine. (The Free
302 Software Foundation sells printed copies of manuals, too.) The
303 problem is the restrictions on the use of the manual. Free manuals
304 are available in source code form, and give you permission to copy and
305 modify. Non-free manuals do not allow this.
307 The criteria of freedom for a free manual are roughly the same as for
308 free software. Redistribution (including the normal kinds of
309 commercial redistribution) must be permitted, so that the manual can
310 accompany every copy of the program, both on-line and on paper.
312 Permission for modification of the technical content is crucial too.
313 When people modify the software, adding or changing features, if they
314 are conscientious they will change the manual too---so they can
315 provide accurate and clear documentation for the modified program. A
316 manual that leaves you no choice but to write a new manual to document
317 a changed version of the program is not really available to our
320 Some kinds of limits on the way modification is handled are
321 acceptable. For example, requirements to preserve the original
322 author's copyright notice, the distribution terms, or the list of
323 authors, are ok. It is also no problem to require modified versions
324 to include notice that they were modified. Even entire sections that
325 may not be deleted or changed are acceptable, as long as they deal
326 with nontechnical topics (like this one). These kinds of restrictions
327 are acceptable because they don't obstruct the community's normal use
330 However, it must be possible to modify all the @emph{technical}
331 content of the manual, and then distribute the result in all the usual
332 media, through all the usual channels. Otherwise, the restrictions
333 obstruct the use of the manual, it is not free, and we need another
334 manual to replace it.
336 Please spread the word about this issue. Our community continues to
337 lose manuals to proprietary publishing. If we spread the word that
338 free software needs free reference manuals and free tutorials, perhaps
339 the next person who wants to contribute by writing documentation will
340 realize, before it is too late, that only free manuals contribute to
341 the free software community.
343 If you are writing documentation, please insist on publishing it under
344 the GNU Free Documentation License or another free documentation
345 license. Remember that this decision requires your approval---you
346 don't have to let the publisher decide. Some commercial publishers
347 will use a free license if you insist, but they will not propose the
348 option; it is up to you to raise the issue and say firmly that this is
349 what you want. If the publisher you are dealing with refuses, please
350 try other publishers. If you're not sure whether a proposed license
351 is free, write to @email{licensing@@gnu.org}.
353 You can encourage commercial publishers to sell more free, copylefted
354 manuals and tutorials by buying them, and particularly by buying
355 copies from the publishers that paid for their writing or for major
356 improvements. Meanwhile, try to avoid buying non-free documentation
357 at all. Check the distribution terms of a manual before you buy it,
358 and insist that whoever seeks your business must respect your freedom.
359 Check the history of the book, and try to reward the publishers that
360 have paid or pay the authors to work on it.
362 The Free Software Foundation maintains a list of free documentation
363 published by other publishers, at
364 @url{http://www.fsf.org/doc/other-free-books.html}.
367 @unnumberedsec Contributors to @value{GDBN}
369 Richard Stallman was the original author of @value{GDBN}, and of many
370 other @sc{gnu} programs. Many others have contributed to its
371 development. This section attempts to credit major contributors. One
372 of the virtues of free software is that everyone is free to contribute
373 to it; with regret, we cannot actually acknowledge everyone here. The
374 file @file{ChangeLog} in the @value{GDBN} distribution approximates a
375 blow-by-blow account.
377 Changes much prior to version 2.0 are lost in the mists of time.
380 @emph{Plea:} Additions to this section are particularly welcome. If you
381 or your friends (or enemies, to be evenhanded) have been unfairly
382 omitted from this list, we would like to add your names!
385 So that they may not regard their many labors as thankless, we
386 particularly thank those who shepherded @value{GDBN} through major
388 Andrew Cagney (releases 6.3, 6.2, 6.1, 6.0, 5.3, 5.2, 5.1 and 5.0);
389 Jim Blandy (release 4.18);
390 Jason Molenda (release 4.17);
391 Stan Shebs (release 4.14);
392 Fred Fish (releases 4.16, 4.15, 4.13, 4.12, 4.11, 4.10, and 4.9);
393 Stu Grossman and John Gilmore (releases 4.8, 4.7, 4.6, 4.5, and 4.4);
394 John Gilmore (releases 4.3, 4.2, 4.1, 4.0, and 3.9);
395 Jim Kingdon (releases 3.5, 3.4, and 3.3);
396 and Randy Smith (releases 3.2, 3.1, and 3.0).
398 Richard Stallman, assisted at various times by Peter TerMaat, Chris
399 Hanson, and Richard Mlynarik, handled releases through 2.8.
401 Michael Tiemann is the author of most of the @sc{gnu} C@t{++} support
402 in @value{GDBN}, with significant additional contributions from Per
403 Bothner and Daniel Berlin. James Clark wrote the @sc{gnu} C@t{++}
404 demangler. Early work on C@t{++} was by Peter TerMaat (who also did
405 much general update work leading to release 3.0).
407 @value{GDBN} uses the BFD subroutine library to examine multiple
408 object-file formats; BFD was a joint project of David V.
409 Henkel-Wallace, Rich Pixley, Steve Chamberlain, and John Gilmore.
411 David Johnson wrote the original COFF support; Pace Willison did
412 the original support for encapsulated COFF.
414 Brent Benson of Harris Computer Systems contributed DWARF 2 support.
416 Adam de Boor and Bradley Davis contributed the ISI Optimum V support.
417 Per Bothner, Noboyuki Hikichi, and Alessandro Forin contributed MIPS
419 Jean-Daniel Fekete contributed Sun 386i support.
420 Chris Hanson improved the HP9000 support.
421 Noboyuki Hikichi and Tomoyuki Hasei contributed Sony/News OS 3 support.
422 David Johnson contributed Encore Umax support.
423 Jyrki Kuoppala contributed Altos 3068 support.
424 Jeff Law contributed HP PA and SOM support.
425 Keith Packard contributed NS32K support.
426 Doug Rabson contributed Acorn Risc Machine support.
427 Bob Rusk contributed Harris Nighthawk CX-UX support.
428 Chris Smith contributed Convex support (and Fortran debugging).
429 Jonathan Stone contributed Pyramid support.
430 Michael Tiemann contributed SPARC support.
431 Tim Tucker contributed support for the Gould NP1 and Gould Powernode.
432 Pace Willison contributed Intel 386 support.
433 Jay Vosburgh contributed Symmetry support.
434 Marko Mlinar contributed OpenRISC 1000 support.
436 Andreas Schwab contributed M68K @sc{gnu}/Linux support.
438 Rich Schaefer and Peter Schauer helped with support of SunOS shared
441 Jay Fenlason and Roland McGrath ensured that @value{GDBN} and GAS agree
442 about several machine instruction sets.
444 Patrick Duval, Ted Goldstein, Vikram Koka and Glenn Engel helped develop
445 remote debugging. Intel Corporation, Wind River Systems, AMD, and ARM
446 contributed remote debugging modules for the i960, VxWorks, A29K UDI,
447 and RDI targets, respectively.
449 Brian Fox is the author of the readline libraries providing
450 command-line editing and command history.
452 Andrew Beers of SUNY Buffalo wrote the language-switching code, the
453 Modula-2 support, and contributed the Languages chapter of this manual.
455 Fred Fish wrote most of the support for Unix System Vr4.
456 He also enhanced the command-completion support to cover C@t{++} overloaded
459 Hitachi America (now Renesas America), Ltd. sponsored the support for
460 H8/300, H8/500, and Super-H processors.
462 NEC sponsored the support for the v850, Vr4xxx, and Vr5xxx processors.
464 Mitsubishi (now Renesas) sponsored the support for D10V, D30V, and M32R/D
467 Toshiba sponsored the support for the TX39 Mips processor.
469 Matsushita sponsored the support for the MN10200 and MN10300 processors.
471 Fujitsu sponsored the support for SPARClite and FR30 processors.
473 Kung Hsu, Jeff Law, and Rick Sladkey added support for hardware
476 Michael Snyder added support for tracepoints.
478 Stu Grossman wrote gdbserver.
480 Jim Kingdon, Peter Schauer, Ian Taylor, and Stu Grossman made
481 nearly innumerable bug fixes and cleanups throughout @value{GDBN}.
483 The following people at the Hewlett-Packard Company contributed
484 support for the PA-RISC 2.0 architecture, HP-UX 10.20, 10.30, and 11.0
485 (narrow mode), HP's implementation of kernel threads, HP's aC@t{++}
486 compiler, and the Text User Interface (nee Terminal User Interface):
487 Ben Krepp, Richard Title, John Bishop, Susan Macchia, Kathy Mann,
488 Satish Pai, India Paul, Steve Rehrauer, and Elena Zannoni. Kim Haase
489 provided HP-specific information in this manual.
491 DJ Delorie ported @value{GDBN} to MS-DOS, for the DJGPP project.
492 Robert Hoehne made significant contributions to the DJGPP port.
494 Cygnus Solutions has sponsored @value{GDBN} maintenance and much of its
495 development since 1991. Cygnus engineers who have worked on @value{GDBN}
496 fulltime include Mark Alexander, Jim Blandy, Per Bothner, Kevin
497 Buettner, Edith Epstein, Chris Faylor, Fred Fish, Martin Hunt, Jim
498 Ingham, John Gilmore, Stu Grossman, Kung Hsu, Jim Kingdon, John Metzler,
499 Fernando Nasser, Geoffrey Noer, Dawn Perchik, Rich Pixley, Zdenek
500 Radouch, Keith Seitz, Stan Shebs, David Taylor, and Elena Zannoni. In
501 addition, Dave Brolley, Ian Carmichael, Steve Chamberlain, Nick Clifton,
502 JT Conklin, Stan Cox, DJ Delorie, Ulrich Drepper, Frank Eigler, Doug
503 Evans, Sean Fagan, David Henkel-Wallace, Richard Henderson, Jeff
504 Holcomb, Jeff Law, Jim Lemke, Tom Lord, Bob Manson, Michael Meissner,
505 Jason Merrill, Catherine Moore, Drew Moseley, Ken Raeburn, Gavin
506 Romig-Koch, Rob Savoye, Jamie Smith, Mike Stump, Ian Taylor, Angela
507 Thomas, Michael Tiemann, Tom Tromey, Ron Unrau, Jim Wilson, and David
508 Zuhn have made contributions both large and small.
510 Andrew Cagney, Fernando Nasser, and Elena Zannoni, while working for
511 Cygnus Solutions, implemented the original @sc{gdb/mi} interface.
513 Jim Blandy added support for preprocessor macros, while working for Red
516 Andrew Cagney designed @value{GDBN}'s architecture vector. Many
517 people including Andrew Cagney, Stephane Carrez, Randolph Chung, Nick
518 Duffek, Richard Henderson, Mark Kettenis, Grace Sainsbury, Kei
519 Sakamoto, Yoshinori Sato, Michael Snyder, Andreas Schwab, Jason
520 Thorpe, Corinna Vinschen, Ulrich Weigand, and Elena Zannoni, helped
521 with the migration of old architectures to this new framework.
523 Andrew Cagney completely re-designed and re-implemented @value{GDBN}'s
524 unwinder framework, this consisting of a fresh new design featuring
525 frame IDs, independent frame sniffers, and the sentinel frame. Mark
526 Kettenis implemented the @sc{dwarf 2} unwinder, Jeff Johnston the
527 libunwind unwinder, and Andrew Cagney the dummy, sentinel, tramp, and
528 trad unwinders. The architecture-specific changes, each involving a
529 complete rewrite of the architecture's frame code, were carried out by
530 Jim Blandy, Joel Brobecker, Kevin Buettner, Andrew Cagney, Stephane
531 Carrez, Randolph Chung, Orjan Friberg, Richard Henderson, Daniel
532 Jacobowitz, Jeff Johnston, Mark Kettenis, Theodore A. Roth, Kei
533 Sakamoto, Yoshinori Sato, Michael Snyder, Corinna Vinschen, and Ulrich
536 Christian Zankel, Ross Morley, Bob Wilson, and Maxim Grigoriev from
537 Tensilica, Inc.@: contributed support for Xtensa processors. Others
538 who have worked on the Xtensa port of @value{GDBN} in the past include
539 Steve Tjiang, John Newlin, and Scott Foehner.
541 Michael Eager and staff of Xilinx, Inc., contributed support for the
542 Xilinx MicroBlaze architecture.
544 Initial support for the FreeBSD/mips target and native configuration
545 was developed by SRI International and the University of Cambridge
546 Computer Laboratory under DARPA/AFRL contract FA8750-10-C-0237
547 ("CTSRD"), as part of the DARPA CRASH research programme.
549 Initial support for the FreeBSD/riscv target and native configuration
550 was developed by SRI International and the University of Cambridge
551 Computer Laboratory (Department of Computer Science and Technology)
552 under DARPA contract HR0011-18-C-0016 ("ECATS"), as part of the DARPA
553 SSITH research programme.
555 The original port to the OpenRISC 1000 is believed to be due to
556 Alessandro Forin and Per Bothner. More recent ports have been the work
557 of Jeremy Bennett, Franck Jullien, Stefan Wallentowitz and
561 @chapter A Sample @value{GDBN} Session
563 You can use this manual at your leisure to read all about @value{GDBN}.
564 However, a handful of commands are enough to get started using the
565 debugger. This chapter illustrates those commands.
568 In this sample session, we emphasize user input like this: @b{input},
569 to make it easier to pick out from the surrounding output.
572 @c FIXME: this example may not be appropriate for some configs, where
573 @c FIXME...primary interest is in remote use.
575 One of the preliminary versions of @sc{gnu} @code{m4} (a generic macro
576 processor) exhibits the following bug: sometimes, when we change its
577 quote strings from the default, the commands used to capture one macro
578 definition within another stop working. In the following short @code{m4}
579 session, we define a macro @code{foo} which expands to @code{0000}; we
580 then use the @code{m4} built-in @code{defn} to define @code{bar} as the
581 same thing. However, when we change the open quote string to
582 @code{<QUOTE>} and the close quote string to @code{<UNQUOTE>}, the same
583 procedure fails to define a new synonym @code{baz}:
592 @b{define(bar,defn(`foo'))}
596 @b{changequote(<QUOTE>,<UNQUOTE>)}
598 @b{define(baz,defn(<QUOTE>foo<UNQUOTE>))}
601 m4: End of input: 0: fatal error: EOF in string
605 Let us use @value{GDBN} to try to see what is going on.
608 $ @b{@value{GDBP} m4}
609 @c FIXME: this falsifies the exact text played out, to permit smallbook
610 @c FIXME... format to come out better.
611 @value{GDBN} is free software and you are welcome to distribute copies
612 of it under certain conditions; type "show copying" to see
614 There is absolutely no warranty for @value{GDBN}; type "show warranty"
617 @value{GDBN} @value{GDBVN}, Copyright 1999 Free Software Foundation, Inc...
622 @value{GDBN} reads only enough symbol data to know where to find the
623 rest when needed; as a result, the first prompt comes up very quickly.
624 We now tell @value{GDBN} to use a narrower display width than usual, so
625 that examples fit in this manual.
628 (@value{GDBP}) @b{set width 70}
632 We need to see how the @code{m4} built-in @code{changequote} works.
633 Having looked at the source, we know the relevant subroutine is
634 @code{m4_changequote}, so we set a breakpoint there with the @value{GDBN}
635 @code{break} command.
638 (@value{GDBP}) @b{break m4_changequote}
639 Breakpoint 1 at 0x62f4: file builtin.c, line 879.
643 Using the @code{run} command, we start @code{m4} running under @value{GDBN}
644 control; as long as control does not reach the @code{m4_changequote}
645 subroutine, the program runs as usual:
648 (@value{GDBP}) @b{run}
649 Starting program: /work/Editorial/gdb/gnu/m4/m4
657 To trigger the breakpoint, we call @code{changequote}. @value{GDBN}
658 suspends execution of @code{m4}, displaying information about the
659 context where it stops.
662 @b{changequote(<QUOTE>,<UNQUOTE>)}
664 Breakpoint 1, m4_changequote (argc=3, argv=0x33c70)
666 879 if (bad_argc(TOKEN_DATA_TEXT(argv[0]),argc,1,3))
670 Now we use the command @code{n} (@code{next}) to advance execution to
671 the next line of the current function.
675 882 set_quotes((argc >= 2) ? TOKEN_DATA_TEXT(argv[1])\
680 @code{set_quotes} looks like a promising subroutine. We can go into it
681 by using the command @code{s} (@code{step}) instead of @code{next}.
682 @code{step} goes to the next line to be executed in @emph{any}
683 subroutine, so it steps into @code{set_quotes}.
687 set_quotes (lq=0x34c78 "<QUOTE>", rq=0x34c88 "<UNQUOTE>")
689 530 if (lquote != def_lquote)
693 The display that shows the subroutine where @code{m4} is now
694 suspended (and its arguments) is called a stack frame display. It
695 shows a summary of the stack. We can use the @code{backtrace}
696 command (which can also be spelled @code{bt}), to see where we are
697 in the stack as a whole: the @code{backtrace} command displays a
698 stack frame for each active subroutine.
701 (@value{GDBP}) @b{bt}
702 #0 set_quotes (lq=0x34c78 "<QUOTE>", rq=0x34c88 "<UNQUOTE>")
704 #1 0x6344 in m4_changequote (argc=3, argv=0x33c70)
706 #2 0x8174 in expand_macro (sym=0x33320) at macro.c:242
707 #3 0x7a88 in expand_token (obs=0x0, t=209696, td=0xf7fffa30)
709 #4 0x79dc in expand_input () at macro.c:40
710 #5 0x2930 in main (argc=0, argv=0xf7fffb20) at m4.c:195
714 We step through a few more lines to see what happens. The first two
715 times, we can use @samp{s}; the next two times we use @code{n} to avoid
716 falling into the @code{xstrdup} subroutine.
720 0x3b5c 532 if (rquote != def_rquote)
722 0x3b80 535 lquote = (lq == nil || *lq == '\0') ? \
723 def_lquote : xstrdup(lq);
725 536 rquote = (rq == nil || *rq == '\0') ? def_rquote\
728 538 len_lquote = strlen(rquote);
732 The last line displayed looks a little odd; we can examine the variables
733 @code{lquote} and @code{rquote} to see if they are in fact the new left
734 and right quotes we specified. We use the command @code{p}
735 (@code{print}) to see their values.
738 (@value{GDBP}) @b{p lquote}
739 $1 = 0x35d40 "<QUOTE>"
740 (@value{GDBP}) @b{p rquote}
741 $2 = 0x35d50 "<UNQUOTE>"
745 @code{lquote} and @code{rquote} are indeed the new left and right quotes.
746 To look at some context, we can display ten lines of source
747 surrounding the current line with the @code{l} (@code{list}) command.
753 535 lquote = (lq == nil || *lq == '\0') ? def_lquote\
755 536 rquote = (rq == nil || *rq == '\0') ? def_rquote\
758 538 len_lquote = strlen(rquote);
759 539 len_rquote = strlen(lquote);
766 Let us step past the two lines that set @code{len_lquote} and
767 @code{len_rquote}, and then examine the values of those variables.
771 539 len_rquote = strlen(lquote);
774 (@value{GDBP}) @b{p len_lquote}
776 (@value{GDBP}) @b{p len_rquote}
781 That certainly looks wrong, assuming @code{len_lquote} and
782 @code{len_rquote} are meant to be the lengths of @code{lquote} and
783 @code{rquote} respectively. We can set them to better values using
784 the @code{p} command, since it can print the value of
785 any expression---and that expression can include subroutine calls and
789 (@value{GDBP}) @b{p len_lquote=strlen(lquote)}
791 (@value{GDBP}) @b{p len_rquote=strlen(rquote)}
796 Is that enough to fix the problem of using the new quotes with the
797 @code{m4} built-in @code{defn}? We can allow @code{m4} to continue
798 executing with the @code{c} (@code{continue}) command, and then try the
799 example that caused trouble initially:
805 @b{define(baz,defn(<QUOTE>foo<UNQUOTE>))}
812 Success! The new quotes now work just as well as the default ones. The
813 problem seems to have been just the two typos defining the wrong
814 lengths. We allow @code{m4} exit by giving it an EOF as input:
818 Program exited normally.
822 The message @samp{Program exited normally.} is from @value{GDBN}; it
823 indicates @code{m4} has finished executing. We can end our @value{GDBN}
824 session with the @value{GDBN} @code{quit} command.
827 (@value{GDBP}) @b{quit}
831 @chapter Getting In and Out of @value{GDBN}
833 This chapter discusses how to start @value{GDBN}, and how to get out of it.
837 type @samp{@value{GDBP}} to start @value{GDBN}.
839 type @kbd{quit} or @kbd{Ctrl-d} to exit.
843 * Invoking GDB:: How to start @value{GDBN}
844 * Quitting GDB:: How to quit @value{GDBN}
845 * Shell Commands:: How to use shell commands inside @value{GDBN}
846 * Logging Output:: How to log @value{GDBN}'s output to a file
850 @section Invoking @value{GDBN}
852 Invoke @value{GDBN} by running the program @code{@value{GDBP}}. Once started,
853 @value{GDBN} reads commands from the terminal until you tell it to exit.
855 You can also run @code{@value{GDBP}} with a variety of arguments and options,
856 to specify more of your debugging environment at the outset.
858 The command-line options described here are designed
859 to cover a variety of situations; in some environments, some of these
860 options may effectively be unavailable.
862 The most usual way to start @value{GDBN} is with one argument,
863 specifying an executable program:
866 @value{GDBP} @var{program}
870 You can also start with both an executable program and a core file
874 @value{GDBP} @var{program} @var{core}
877 You can, instead, specify a process ID as a second argument, if you want
878 to debug a running process:
881 @value{GDBP} @var{program} 1234
885 would attach @value{GDBN} to process @code{1234} (unless you also have a file
886 named @file{1234}; @value{GDBN} does check for a core file first).
888 Taking advantage of the second command-line argument requires a fairly
889 complete operating system; when you use @value{GDBN} as a remote
890 debugger attached to a bare board, there may not be any notion of
891 ``process'', and there is often no way to get a core dump. @value{GDBN}
892 will warn you if it is unable to attach or to read core dumps.
894 You can optionally have @code{@value{GDBP}} pass any arguments after the
895 executable file to the inferior using @code{--args}. This option stops
898 @value{GDBP} --args gcc -O2 -c foo.c
900 This will cause @code{@value{GDBP}} to debug @code{gcc}, and to set
901 @code{gcc}'s command-line arguments (@pxref{Arguments}) to @samp{-O2 -c foo.c}.
903 You can run @code{@value{GDBP}} without printing the front material, which describes
904 @value{GDBN}'s non-warranty, by specifying @code{--silent}
905 (or @code{-q}/@code{--quiet}):
908 @value{GDBP} --silent
912 You can further control how @value{GDBN} starts up by using command-line
913 options. @value{GDBN} itself can remind you of the options available.
923 to display all available options and briefly describe their use
924 (@samp{@value{GDBP} -h} is a shorter equivalent).
926 All options and command line arguments you give are processed
927 in sequential order. The order makes a difference when the
928 @samp{-x} option is used.
932 * File Options:: Choosing files
933 * Mode Options:: Choosing modes
934 * Startup:: What @value{GDBN} does during startup
938 @subsection Choosing Files
940 When @value{GDBN} starts, it reads any arguments other than options as
941 specifying an executable file and core file (or process ID). This is
942 the same as if the arguments were specified by the @samp{-se} and
943 @samp{-c} (or @samp{-p}) options respectively. (@value{GDBN} reads the
944 first argument that does not have an associated option flag as
945 equivalent to the @samp{-se} option followed by that argument; and the
946 second argument that does not have an associated option flag, if any, as
947 equivalent to the @samp{-c}/@samp{-p} option followed by that argument.)
948 If the second argument begins with a decimal digit, @value{GDBN} will
949 first attempt to attach to it as a process, and if that fails, attempt
950 to open it as a corefile. If you have a corefile whose name begins with
951 a digit, you can prevent @value{GDBN} from treating it as a pid by
952 prefixing it with @file{./}, e.g.@: @file{./12345}.
954 If @value{GDBN} has not been configured to included core file support,
955 such as for most embedded targets, then it will complain about a second
956 argument and ignore it.
958 Many options have both long and short forms; both are shown in the
959 following list. @value{GDBN} also recognizes the long forms if you truncate
960 them, so long as enough of the option is present to be unambiguous.
961 (If you prefer, you can flag option arguments with @samp{--} rather
962 than @samp{-}, though we illustrate the more usual convention.)
964 @c NOTE: the @cindex entries here use double dashes ON PURPOSE. This
965 @c way, both those who look for -foo and --foo in the index, will find
969 @item -symbols @var{file}
971 @cindex @code{--symbols}
973 Read symbol table from file @var{file}.
975 @item -exec @var{file}
977 @cindex @code{--exec}
979 Use file @var{file} as the executable file to execute when appropriate,
980 and for examining pure data in conjunction with a core dump.
984 Read symbol table from file @var{file} and use it as the executable
987 @item -core @var{file}
989 @cindex @code{--core}
991 Use file @var{file} as a core dump to examine.
993 @item -pid @var{number}
994 @itemx -p @var{number}
997 Connect to process ID @var{number}, as with the @code{attach} command.
999 @item -command @var{file}
1000 @itemx -x @var{file}
1001 @cindex @code{--command}
1003 Execute commands from file @var{file}. The contents of this file is
1004 evaluated exactly as the @code{source} command would.
1005 @xref{Command Files,, Command files}.
1007 @item -eval-command @var{command}
1008 @itemx -ex @var{command}
1009 @cindex @code{--eval-command}
1011 Execute a single @value{GDBN} command.
1013 This option may be used multiple times to call multiple commands. It may
1014 also be interleaved with @samp{-command} as required.
1017 @value{GDBP} -ex 'target sim' -ex 'load' \
1018 -x setbreakpoints -ex 'run' a.out
1021 @item -init-command @var{file}
1022 @itemx -ix @var{file}
1023 @cindex @code{--init-command}
1025 Execute commands from file @var{file} before loading the inferior (but
1026 after loading gdbinit files).
1029 @item -init-eval-command @var{command}
1030 @itemx -iex @var{command}
1031 @cindex @code{--init-eval-command}
1033 Execute a single @value{GDBN} command before loading the inferior (but
1034 after loading gdbinit files).
1037 @item -directory @var{directory}
1038 @itemx -d @var{directory}
1039 @cindex @code{--directory}
1041 Add @var{directory} to the path to search for source and script files.
1045 @cindex @code{--readnow}
1047 Read each symbol file's entire symbol table immediately, rather than
1048 the default, which is to read it incrementally as it is needed.
1049 This makes startup slower, but makes future operations faster.
1052 @anchor{--readnever}
1053 @cindex @code{--readnever}, command-line option
1054 Do not read each symbol file's symbolic debug information. This makes
1055 startup faster but at the expense of not being able to perform
1056 symbolic debugging. DWARF unwind information is also not read,
1057 meaning backtraces may become incomplete or inaccurate. One use of
1058 this is when a user simply wants to do the following sequence: attach,
1059 dump core, detach. Loading the debugging information in this case is
1060 an unnecessary cause of delay.
1064 @subsection Choosing Modes
1066 You can run @value{GDBN} in various alternative modes---for example, in
1067 batch mode or quiet mode.
1075 Do not execute commands found in any initialization file.
1076 There are three init files, loaded in the following order:
1079 @item @file{system.gdbinit}
1080 This is the system-wide init file.
1081 Its location is specified with the @code{--with-system-gdbinit}
1082 configure option (@pxref{System-wide configuration}).
1083 It is loaded first when @value{GDBN} starts, before command line options
1084 have been processed.
1085 @item @file{~/.gdbinit}
1086 This is the init file in your home directory.
1087 It is loaded next, after @file{system.gdbinit}, and before
1088 command options have been processed.
1089 @item @file{./.gdbinit}
1090 This is the init file in the current directory.
1091 It is loaded last, after command line options other than @code{-x} and
1092 @code{-ex} have been processed. Command line options @code{-x} and
1093 @code{-ex} are processed last, after @file{./.gdbinit} has been loaded.
1096 For further documentation on startup processing, @xref{Startup}.
1097 For documentation on how to write command files,
1098 @xref{Command Files,,Command Files}.
1103 Do not execute commands found in @file{~/.gdbinit}, the init file
1104 in your home directory.
1110 @cindex @code{--quiet}
1111 @cindex @code{--silent}
1113 ``Quiet''. Do not print the introductory and copyright messages. These
1114 messages are also suppressed in batch mode.
1117 @cindex @code{--batch}
1118 Run in batch mode. Exit with status @code{0} after processing all the
1119 command files specified with @samp{-x} (and all commands from
1120 initialization files, if not inhibited with @samp{-n}). Exit with
1121 nonzero status if an error occurs in executing the @value{GDBN} commands
1122 in the command files. Batch mode also disables pagination, sets unlimited
1123 terminal width and height @pxref{Screen Size}, and acts as if @kbd{set confirm
1124 off} were in effect (@pxref{Messages/Warnings}).
1126 Batch mode may be useful for running @value{GDBN} as a filter, for
1127 example to download and run a program on another computer; in order to
1128 make this more useful, the message
1131 Program exited normally.
1135 (which is ordinarily issued whenever a program running under
1136 @value{GDBN} control terminates) is not issued when running in batch
1140 @cindex @code{--batch-silent}
1141 Run in batch mode exactly like @samp{-batch}, but totally silently. All
1142 @value{GDBN} output to @code{stdout} is prevented (@code{stderr} is
1143 unaffected). This is much quieter than @samp{-silent} and would be useless
1144 for an interactive session.
1146 This is particularly useful when using targets that give @samp{Loading section}
1147 messages, for example.
1149 Note that targets that give their output via @value{GDBN}, as opposed to
1150 writing directly to @code{stdout}, will also be made silent.
1152 @item -return-child-result
1153 @cindex @code{--return-child-result}
1154 The return code from @value{GDBN} will be the return code from the child
1155 process (the process being debugged), with the following exceptions:
1159 @value{GDBN} exits abnormally. E.g., due to an incorrect argument or an
1160 internal error. In this case the exit code is the same as it would have been
1161 without @samp{-return-child-result}.
1163 The user quits with an explicit value. E.g., @samp{quit 1}.
1165 The child process never runs, or is not allowed to terminate, in which case
1166 the exit code will be -1.
1169 This option is useful in conjunction with @samp{-batch} or @samp{-batch-silent},
1170 when @value{GDBN} is being used as a remote program loader or simulator
1175 @cindex @code{--nowindows}
1177 ``No windows''. If @value{GDBN} comes with a graphical user interface
1178 (GUI) built in, then this option tells @value{GDBN} to only use the command-line
1179 interface. If no GUI is available, this option has no effect.
1183 @cindex @code{--windows}
1185 If @value{GDBN} includes a GUI, then this option requires it to be
1188 @item -cd @var{directory}
1190 Run @value{GDBN} using @var{directory} as its working directory,
1191 instead of the current directory.
1193 @item -data-directory @var{directory}
1194 @itemx -D @var{directory}
1195 @cindex @code{--data-directory}
1197 Run @value{GDBN} using @var{directory} as its data directory.
1198 The data directory is where @value{GDBN} searches for its
1199 auxiliary files. @xref{Data Files}.
1203 @cindex @code{--fullname}
1205 @sc{gnu} Emacs sets this option when it runs @value{GDBN} as a
1206 subprocess. It tells @value{GDBN} to output the full file name and line
1207 number in a standard, recognizable fashion each time a stack frame is
1208 displayed (which includes each time your program stops). This
1209 recognizable format looks like two @samp{\032} characters, followed by
1210 the file name, line number and character position separated by colons,
1211 and a newline. The Emacs-to-@value{GDBN} interface program uses the two
1212 @samp{\032} characters as a signal to display the source code for the
1215 @item -annotate @var{level}
1216 @cindex @code{--annotate}
1217 This option sets the @dfn{annotation level} inside @value{GDBN}. Its
1218 effect is identical to using @samp{set annotate @var{level}}
1219 (@pxref{Annotations}). The annotation @var{level} controls how much
1220 information @value{GDBN} prints together with its prompt, values of
1221 expressions, source lines, and other types of output. Level 0 is the
1222 normal, level 1 is for use when @value{GDBN} is run as a subprocess of
1223 @sc{gnu} Emacs, level 3 is the maximum annotation suitable for programs
1224 that control @value{GDBN}, and level 2 has been deprecated.
1226 The annotation mechanism has largely been superseded by @sc{gdb/mi}
1230 @cindex @code{--args}
1231 Change interpretation of command line so that arguments following the
1232 executable file are passed as command line arguments to the inferior.
1233 This option stops option processing.
1235 @item -baud @var{bps}
1237 @cindex @code{--baud}
1239 Set the line speed (baud rate or bits per second) of any serial
1240 interface used by @value{GDBN} for remote debugging.
1242 @item -l @var{timeout}
1244 Set the timeout (in seconds) of any communication used by @value{GDBN}
1245 for remote debugging.
1247 @item -tty @var{device}
1248 @itemx -t @var{device}
1249 @cindex @code{--tty}
1251 Run using @var{device} for your program's standard input and output.
1252 @c FIXME: kingdon thinks there is more to -tty. Investigate.
1254 @c resolve the situation of these eventually
1256 @cindex @code{--tui}
1257 Activate the @dfn{Text User Interface} when starting. The Text User
1258 Interface manages several text windows on the terminal, showing
1259 source, assembly, registers and @value{GDBN} command outputs
1260 (@pxref{TUI, ,@value{GDBN} Text User Interface}). Do not use this
1261 option if you run @value{GDBN} from Emacs (@pxref{Emacs, ,
1262 Using @value{GDBN} under @sc{gnu} Emacs}).
1264 @item -interpreter @var{interp}
1265 @cindex @code{--interpreter}
1266 Use the interpreter @var{interp} for interface with the controlling
1267 program or device. This option is meant to be set by programs which
1268 communicate with @value{GDBN} using it as a back end.
1269 @xref{Interpreters, , Command Interpreters}.
1271 @samp{--interpreter=mi} (or @samp{--interpreter=mi3}) causes
1272 @value{GDBN} to use the @dfn{@sc{gdb/mi} interface} version 3 (@pxref{GDB/MI, ,
1273 The @sc{gdb/mi} Interface}) included since @value{GDBN} version 9.1. @sc{gdb/mi}
1274 version 2 (@code{mi2}), included in @value{GDBN} 6.0 and version 1 (@code{mi1}),
1275 included in @value{GDBN} 5.3, are also available. Earlier @sc{gdb/mi}
1276 interfaces are no longer supported.
1279 @cindex @code{--write}
1280 Open the executable and core files for both reading and writing. This
1281 is equivalent to the @samp{set write on} command inside @value{GDBN}
1285 @cindex @code{--statistics}
1286 This option causes @value{GDBN} to print statistics about time and
1287 memory usage after it completes each command and returns to the prompt.
1290 @cindex @code{--version}
1291 This option causes @value{GDBN} to print its version number and
1292 no-warranty blurb, and exit.
1294 @item -configuration
1295 @cindex @code{--configuration}
1296 This option causes @value{GDBN} to print details about its build-time
1297 configuration parameters, and then exit. These details can be
1298 important when reporting @value{GDBN} bugs (@pxref{GDB Bugs}).
1303 @subsection What @value{GDBN} Does During Startup
1304 @cindex @value{GDBN} startup
1306 Here's the description of what @value{GDBN} does during session startup:
1310 Sets up the command interpreter as specified by the command line
1311 (@pxref{Mode Options, interpreter}).
1315 Reads the system-wide @dfn{init file} (if @option{--with-system-gdbinit} was
1316 used when building @value{GDBN}; @pxref{System-wide configuration,
1317 ,System-wide configuration and settings}) and executes all the commands in
1320 @anchor{Home Directory Init File}
1322 Reads the init file (if any) in your home directory@footnote{On
1323 DOS/Windows systems, the home directory is the one pointed to by the
1324 @code{HOME} environment variable.} and executes all the commands in
1327 @anchor{Option -init-eval-command}
1329 Executes commands and command files specified by the @samp{-iex} and
1330 @samp{-ix} options in their specified order. Usually you should use the
1331 @samp{-ex} and @samp{-x} options instead, but this way you can apply
1332 settings before @value{GDBN} init files get executed and before inferior
1336 Processes command line options and operands.
1338 @anchor{Init File in the Current Directory during Startup}
1340 Reads and executes the commands from init file (if any) in the current
1341 working directory as long as @samp{set auto-load local-gdbinit} is set to
1342 @samp{on} (@pxref{Init File in the Current Directory}).
1343 This is only done if the current directory is
1344 different from your home directory. Thus, you can have more than one
1345 init file, one generic in your home directory, and another, specific
1346 to the program you are debugging, in the directory where you invoke
1350 If the command line specified a program to debug, or a process to
1351 attach to, or a core file, @value{GDBN} loads any auto-loaded
1352 scripts provided for the program or for its loaded shared libraries.
1353 @xref{Auto-loading}.
1355 If you wish to disable the auto-loading during startup,
1356 you must do something like the following:
1359 $ gdb -iex "set auto-load python-scripts off" myprogram
1362 Option @samp{-ex} does not work because the auto-loading is then turned
1366 Executes commands and command files specified by the @samp{-ex} and
1367 @samp{-x} options in their specified order. @xref{Command Files}, for
1368 more details about @value{GDBN} command files.
1371 Reads the command history recorded in the @dfn{history file}.
1372 @xref{Command History}, for more details about the command history and the
1373 files where @value{GDBN} records it.
1376 Init files use the same syntax as @dfn{command files} (@pxref{Command
1377 Files}) and are processed by @value{GDBN} in the same way. The init
1378 file in your home directory can set options (such as @samp{set
1379 complaints}) that affect subsequent processing of command line options
1380 and operands. Init files are not executed if you use the @samp{-nx}
1381 option (@pxref{Mode Options, ,Choosing Modes}).
1383 To display the list of init files loaded by gdb at startup, you
1384 can use @kbd{gdb --help}.
1386 @cindex init file name
1387 @cindex @file{.gdbinit}
1388 @cindex @file{gdb.ini}
1389 The @value{GDBN} init files are normally called @file{.gdbinit}.
1390 The DJGPP port of @value{GDBN} uses the name @file{gdb.ini}, due to
1391 the limitations of file names imposed by DOS filesystems. The Windows
1392 port of @value{GDBN} uses the standard name, but if it finds a
1393 @file{gdb.ini} file in your home directory, it warns you about that
1394 and suggests to rename the file to the standard name.
1398 @section Quitting @value{GDBN}
1399 @cindex exiting @value{GDBN}
1400 @cindex leaving @value{GDBN}
1403 @kindex quit @r{[}@var{expression}@r{]}
1404 @kindex q @r{(@code{quit})}
1405 @item quit @r{[}@var{expression}@r{]}
1407 To exit @value{GDBN}, use the @code{quit} command (abbreviated
1408 @code{q}), or type an end-of-file character (usually @kbd{Ctrl-d}). If you
1409 do not supply @var{expression}, @value{GDBN} will terminate normally;
1410 otherwise it will terminate using the result of @var{expression} as the
1415 An interrupt (often @kbd{Ctrl-c}) does not exit from @value{GDBN}, but rather
1416 terminates the action of any @value{GDBN} command that is in progress and
1417 returns to @value{GDBN} command level. It is safe to type the interrupt
1418 character at any time because @value{GDBN} does not allow it to take effect
1419 until a time when it is safe.
1421 If you have been using @value{GDBN} to control an attached process or
1422 device, you can release it with the @code{detach} command
1423 (@pxref{Attach, ,Debugging an Already-running Process}).
1425 @node Shell Commands
1426 @section Shell Commands
1428 If you need to execute occasional shell commands during your
1429 debugging session, there is no need to leave or suspend @value{GDBN}; you can
1430 just use the @code{shell} command.
1435 @cindex shell escape
1436 @item shell @var{command-string}
1437 @itemx !@var{command-string}
1438 Invoke a standard shell to execute @var{command-string}.
1439 Note that no space is needed between @code{!} and @var{command-string}.
1440 If it exists, the environment variable @code{SHELL} determines which
1441 shell to run. Otherwise @value{GDBN} uses the default shell
1442 (@file{/bin/sh} on Unix systems, @file{COMMAND.COM} on MS-DOS, etc.).
1445 The utility @code{make} is often needed in development environments.
1446 You do not have to use the @code{shell} command for this purpose in
1451 @cindex calling make
1452 @item make @var{make-args}
1453 Execute the @code{make} program with the specified
1454 arguments. This is equivalent to @samp{shell make @var{make-args}}.
1457 @node Logging Output
1458 @section Logging Output
1459 @cindex logging @value{GDBN} output
1460 @cindex save @value{GDBN} output to a file
1462 You may want to save the output of @value{GDBN} commands to a file.
1463 There are several commands to control @value{GDBN}'s logging.
1467 @item set logging on
1469 @item set logging off
1471 @cindex logging file name
1472 @item set logging file @var{file}
1473 Change the name of the current logfile. The default logfile is @file{gdb.txt}.
1474 @item set logging overwrite [on|off]
1475 By default, @value{GDBN} will append to the logfile. Set @code{overwrite} if
1476 you want @code{set logging on} to overwrite the logfile instead.
1477 @item set logging redirect [on|off]
1478 By default, @value{GDBN} output will go to both the terminal and the logfile.
1479 Set @code{redirect} if you want output to go only to the log file.
1480 @item set logging debugredirect [on|off]
1481 By default, @value{GDBN} debug output will go to both the terminal and the logfile.
1482 Set @code{debugredirect} if you want debug output to go only to the log file.
1483 @kindex show logging
1485 Show the current values of the logging settings.
1489 @chapter @value{GDBN} Commands
1491 You can abbreviate a @value{GDBN} command to the first few letters of the command
1492 name, if that abbreviation is unambiguous; and you can repeat certain
1493 @value{GDBN} commands by typing just @key{RET}. You can also use the @key{TAB}
1494 key to get @value{GDBN} to fill out the rest of a word in a command (or to
1495 show you the alternatives available, if there is more than one possibility).
1498 * Command Syntax:: How to give commands to @value{GDBN}
1499 * Completion:: Command completion
1500 * Help:: How to ask @value{GDBN} for help
1503 @node Command Syntax
1504 @section Command Syntax
1506 A @value{GDBN} command is a single line of input. There is no limit on
1507 how long it can be. It starts with a command name, which is followed by
1508 arguments whose meaning depends on the command name. For example, the
1509 command @code{step} accepts an argument which is the number of times to
1510 step, as in @samp{step 5}. You can also use the @code{step} command
1511 with no arguments. Some commands do not allow any arguments.
1513 @cindex abbreviation
1514 @value{GDBN} command names may always be truncated if that abbreviation is
1515 unambiguous. Other possible command abbreviations are listed in the
1516 documentation for individual commands. In some cases, even ambiguous
1517 abbreviations are allowed; for example, @code{s} is specially defined as
1518 equivalent to @code{step} even though there are other commands whose
1519 names start with @code{s}. You can test abbreviations by using them as
1520 arguments to the @code{help} command.
1522 @cindex repeating commands
1523 @kindex RET @r{(repeat last command)}
1524 A blank line as input to @value{GDBN} (typing just @key{RET}) means to
1525 repeat the previous command. Certain commands (for example, @code{run})
1526 will not repeat this way; these are commands whose unintentional
1527 repetition might cause trouble and which you are unlikely to want to
1528 repeat. User-defined commands can disable this feature; see
1529 @ref{Define, dont-repeat}.
1531 The @code{list} and @code{x} commands, when you repeat them with
1532 @key{RET}, construct new arguments rather than repeating
1533 exactly as typed. This permits easy scanning of source or memory.
1535 @value{GDBN} can also use @key{RET} in another way: to partition lengthy
1536 output, in a way similar to the common utility @code{more}
1537 (@pxref{Screen Size,,Screen Size}). Since it is easy to press one
1538 @key{RET} too many in this situation, @value{GDBN} disables command
1539 repetition after any command that generates this sort of display.
1541 @kindex # @r{(a comment)}
1543 Any text from a @kbd{#} to the end of the line is a comment; it does
1544 nothing. This is useful mainly in command files (@pxref{Command
1545 Files,,Command Files}).
1547 @cindex repeating command sequences
1548 @kindex Ctrl-o @r{(operate-and-get-next)}
1549 The @kbd{Ctrl-o} binding is useful for repeating a complex sequence of
1550 commands. This command accepts the current line, like @key{RET}, and
1551 then fetches the next line relative to the current line from the history
1555 @section Command Completion
1558 @cindex word completion
1559 @value{GDBN} can fill in the rest of a word in a command for you, if there is
1560 only one possibility; it can also show you what the valid possibilities
1561 are for the next word in a command, at any time. This works for @value{GDBN}
1562 commands, @value{GDBN} subcommands, and the names of symbols in your program.
1564 Press the @key{TAB} key whenever you want @value{GDBN} to fill out the rest
1565 of a word. If there is only one possibility, @value{GDBN} fills in the
1566 word, and waits for you to finish the command (or press @key{RET} to
1567 enter it). For example, if you type
1569 @c FIXME "@key" does not distinguish its argument sufficiently to permit
1570 @c complete accuracy in these examples; space introduced for clarity.
1571 @c If texinfo enhancements make it unnecessary, it would be nice to
1572 @c replace " @key" by "@key" in the following...
1574 (@value{GDBP}) info bre @key{TAB}
1578 @value{GDBN} fills in the rest of the word @samp{breakpoints}, since that is
1579 the only @code{info} subcommand beginning with @samp{bre}:
1582 (@value{GDBP}) info breakpoints
1586 You can either press @key{RET} at this point, to run the @code{info
1587 breakpoints} command, or backspace and enter something else, if
1588 @samp{breakpoints} does not look like the command you expected. (If you
1589 were sure you wanted @code{info breakpoints} in the first place, you
1590 might as well just type @key{RET} immediately after @samp{info bre},
1591 to exploit command abbreviations rather than command completion).
1593 If there is more than one possibility for the next word when you press
1594 @key{TAB}, @value{GDBN} sounds a bell. You can either supply more
1595 characters and try again, or just press @key{TAB} a second time;
1596 @value{GDBN} displays all the possible completions for that word. For
1597 example, you might want to set a breakpoint on a subroutine whose name
1598 begins with @samp{make_}, but when you type @kbd{b make_@key{TAB}} @value{GDBN}
1599 just sounds the bell. Typing @key{TAB} again displays all the
1600 function names in your program that begin with those characters, for
1604 (@value{GDBP}) b make_ @key{TAB}
1605 @exdent @value{GDBN} sounds bell; press @key{TAB} again, to see:
1606 make_a_section_from_file make_environ
1607 make_abs_section make_function_type
1608 make_blockvector make_pointer_type
1609 make_cleanup make_reference_type
1610 make_command make_symbol_completion_list
1611 (@value{GDBP}) b make_
1615 After displaying the available possibilities, @value{GDBN} copies your
1616 partial input (@samp{b make_} in the example) so you can finish the
1619 If you just want to see the list of alternatives in the first place, you
1620 can press @kbd{M-?} rather than pressing @key{TAB} twice. @kbd{M-?}
1621 means @kbd{@key{META} ?}. You can type this either by holding down a
1622 key designated as the @key{META} shift on your keyboard (if there is
1623 one) while typing @kbd{?}, or as @key{ESC} followed by @kbd{?}.
1625 If the number of possible completions is large, @value{GDBN} will
1626 print as much of the list as it has collected, as well as a message
1627 indicating that the list may be truncated.
1630 (@value{GDBP}) b m@key{TAB}@key{TAB}
1632 <... the rest of the possible completions ...>
1633 *** List may be truncated, max-completions reached. ***
1638 This behavior can be controlled with the following commands:
1641 @kindex set max-completions
1642 @item set max-completions @var{limit}
1643 @itemx set max-completions unlimited
1644 Set the maximum number of completion candidates. @value{GDBN} will
1645 stop looking for more completions once it collects this many candidates.
1646 This is useful when completing on things like function names as collecting
1647 all the possible candidates can be time consuming.
1648 The default value is 200. A value of zero disables tab-completion.
1649 Note that setting either no limit or a very large limit can make
1651 @kindex show max-completions
1652 @item show max-completions
1653 Show the maximum number of candidates that @value{GDBN} will collect and show
1657 @cindex quotes in commands
1658 @cindex completion of quoted strings
1659 Sometimes the string you need, while logically a ``word'', may contain
1660 parentheses or other characters that @value{GDBN} normally excludes from
1661 its notion of a word. To permit word completion to work in this
1662 situation, you may enclose words in @code{'} (single quote marks) in
1663 @value{GDBN} commands.
1665 A likely situation where you might need this is in typing an
1666 expression that involves a C@t{++} symbol name with template
1667 parameters. This is because when completing expressions, GDB treats
1668 the @samp{<} character as word delimiter, assuming that it's the
1669 less-than comparison operator (@pxref{C Operators, , C and C@t{++}
1672 For example, when you want to call a C@t{++} template function
1673 interactively using the @code{print} or @code{call} commands, you may
1674 need to distinguish whether you mean the version of @code{name} that
1675 was specialized for @code{int}, @code{name<int>()}, or the version
1676 that was specialized for @code{float}, @code{name<float>()}. To use
1677 the word-completion facilities in this situation, type a single quote
1678 @code{'} at the beginning of the function name. This alerts
1679 @value{GDBN} that it may need to consider more information than usual
1680 when you press @key{TAB} or @kbd{M-?} to request word completion:
1683 (@value{GDBP}) p 'func< @kbd{M-?}
1684 func<int>() func<float>()
1685 (@value{GDBP}) p 'func<
1688 When setting breakpoints however (@pxref{Specify Location}), you don't
1689 usually need to type a quote before the function name, because
1690 @value{GDBN} understands that you want to set a breakpoint on a
1694 (@value{GDBP}) b func< @kbd{M-?}
1695 func<int>() func<float>()
1696 (@value{GDBP}) b func<
1699 This is true even in the case of typing the name of C@t{++} overloaded
1700 functions (multiple definitions of the same function, distinguished by
1701 argument type). For example, when you want to set a breakpoint you
1702 don't need to distinguish whether you mean the version of @code{name}
1703 that takes an @code{int} parameter, @code{name(int)}, or the version
1704 that takes a @code{float} parameter, @code{name(float)}.
1707 (@value{GDBP}) b bubble( @kbd{M-?}
1708 bubble(int) bubble(double)
1709 (@value{GDBP}) b bubble(dou @kbd{M-?}
1713 See @ref{quoting names} for a description of other scenarios that
1716 For more information about overloaded functions, see @ref{C Plus Plus
1717 Expressions, ,C@t{++} Expressions}. You can use the command @code{set
1718 overload-resolution off} to disable overload resolution;
1719 see @ref{Debugging C Plus Plus, ,@value{GDBN} Features for C@t{++}}.
1721 @cindex completion of structure field names
1722 @cindex structure field name completion
1723 @cindex completion of union field names
1724 @cindex union field name completion
1725 When completing in an expression which looks up a field in a
1726 structure, @value{GDBN} also tries@footnote{The completer can be
1727 confused by certain kinds of invalid expressions. Also, it only
1728 examines the static type of the expression, not the dynamic type.} to
1729 limit completions to the field names available in the type of the
1733 (@value{GDBP}) p gdb_stdout.@kbd{M-?}
1734 magic to_fputs to_rewind
1735 to_data to_isatty to_write
1736 to_delete to_put to_write_async_safe
1741 This is because the @code{gdb_stdout} is a variable of the type
1742 @code{struct ui_file} that is defined in @value{GDBN} sources as
1749 ui_file_flush_ftype *to_flush;
1750 ui_file_write_ftype *to_write;
1751 ui_file_write_async_safe_ftype *to_write_async_safe;
1752 ui_file_fputs_ftype *to_fputs;
1753 ui_file_read_ftype *to_read;
1754 ui_file_delete_ftype *to_delete;
1755 ui_file_isatty_ftype *to_isatty;
1756 ui_file_rewind_ftype *to_rewind;
1757 ui_file_put_ftype *to_put;
1764 @section Getting Help
1765 @cindex online documentation
1768 You can always ask @value{GDBN} itself for information on its commands,
1769 using the command @code{help}.
1772 @kindex h @r{(@code{help})}
1775 You can use @code{help} (abbreviated @code{h}) with no arguments to
1776 display a short list of named classes of commands:
1780 List of classes of commands:
1782 aliases -- Aliases of other commands
1783 breakpoints -- Making program stop at certain points
1784 data -- Examining data
1785 files -- Specifying and examining files
1786 internals -- Maintenance commands
1787 obscure -- Obscure features
1788 running -- Running the program
1789 stack -- Examining the stack
1790 status -- Status inquiries
1791 support -- Support facilities
1792 tracepoints -- Tracing of program execution without
1793 stopping the program
1794 user-defined -- User-defined commands
1796 Type "help" followed by a class name for a list of
1797 commands in that class.
1798 Type "help" followed by command name for full
1800 Command name abbreviations are allowed if unambiguous.
1803 @c the above line break eliminates huge line overfull...
1805 @item help @var{class}
1806 Using one of the general help classes as an argument, you can get a
1807 list of the individual commands in that class. For example, here is the
1808 help display for the class @code{status}:
1811 (@value{GDBP}) help status
1816 @c Line break in "show" line falsifies real output, but needed
1817 @c to fit in smallbook page size.
1818 info -- Generic command for showing things
1819 about the program being debugged
1820 show -- Generic command for showing things
1823 Type "help" followed by command name for full
1825 Command name abbreviations are allowed if unambiguous.
1829 @item help @var{command}
1830 With a command name as @code{help} argument, @value{GDBN} displays a
1831 short paragraph on how to use that command.
1834 @item apropos @var{args}
1835 The @code{apropos} command searches through all of the @value{GDBN}
1836 commands, and their documentation, for the regular expression specified in
1837 @var{args}. It prints out all matches found. For example:
1848 alias -- Define a new command that is an alias of an existing command
1849 aliases -- Aliases of other commands
1850 d -- Delete some breakpoints or auto-display expressions
1851 del -- Delete some breakpoints or auto-display expressions
1852 delete -- Delete some breakpoints or auto-display expressions
1857 @item complete @var{args}
1858 The @code{complete @var{args}} command lists all the possible completions
1859 for the beginning of a command. Use @var{args} to specify the beginning of the
1860 command you want completed. For example:
1866 @noindent results in:
1877 @noindent This is intended for use by @sc{gnu} Emacs.
1880 In addition to @code{help}, you can use the @value{GDBN} commands @code{info}
1881 and @code{show} to inquire about the state of your program, or the state
1882 of @value{GDBN} itself. Each command supports many topics of inquiry; this
1883 manual introduces each of them in the appropriate context. The listings
1884 under @code{info} and under @code{show} in the Command, Variable, and
1885 Function Index point to all the sub-commands. @xref{Command and Variable
1891 @kindex i @r{(@code{info})}
1893 This command (abbreviated @code{i}) is for describing the state of your
1894 program. For example, you can show the arguments passed to a function
1895 with @code{info args}, list the registers currently in use with @code{info
1896 registers}, or list the breakpoints you have set with @code{info breakpoints}.
1897 You can get a complete list of the @code{info} sub-commands with
1898 @w{@code{help info}}.
1902 You can assign the result of an expression to an environment variable with
1903 @code{set}. For example, you can set the @value{GDBN} prompt to a $-sign with
1904 @code{set prompt $}.
1908 In contrast to @code{info}, @code{show} is for describing the state of
1909 @value{GDBN} itself.
1910 You can change most of the things you can @code{show}, by using the
1911 related command @code{set}; for example, you can control what number
1912 system is used for displays with @code{set radix}, or simply inquire
1913 which is currently in use with @code{show radix}.
1916 To display all the settable parameters and their current
1917 values, you can use @code{show} with no arguments; you may also use
1918 @code{info set}. Both commands produce the same display.
1919 @c FIXME: "info set" violates the rule that "info" is for state of
1920 @c FIXME...program. Ck w/ GNU: "info set" to be called something else,
1921 @c FIXME...or change desc of rule---eg "state of prog and debugging session"?
1925 Here are several miscellaneous @code{show} subcommands, all of which are
1926 exceptional in lacking corresponding @code{set} commands:
1929 @kindex show version
1930 @cindex @value{GDBN} version number
1932 Show what version of @value{GDBN} is running. You should include this
1933 information in @value{GDBN} bug-reports. If multiple versions of
1934 @value{GDBN} are in use at your site, you may need to determine which
1935 version of @value{GDBN} you are running; as @value{GDBN} evolves, new
1936 commands are introduced, and old ones may wither away. Also, many
1937 system vendors ship variant versions of @value{GDBN}, and there are
1938 variant versions of @value{GDBN} in @sc{gnu}/Linux distributions as well.
1939 The version number is the same as the one announced when you start
1942 @kindex show copying
1943 @kindex info copying
1944 @cindex display @value{GDBN} copyright
1947 Display information about permission for copying @value{GDBN}.
1949 @kindex show warranty
1950 @kindex info warranty
1952 @itemx info warranty
1953 Display the @sc{gnu} ``NO WARRANTY'' statement, or a warranty,
1954 if your version of @value{GDBN} comes with one.
1956 @kindex show configuration
1957 @item show configuration
1958 Display detailed information about the way @value{GDBN} was configured
1959 when it was built. This displays the optional arguments passed to the
1960 @file{configure} script and also configuration parameters detected
1961 automatically by @command{configure}. When reporting a @value{GDBN}
1962 bug (@pxref{GDB Bugs}), it is important to include this information in
1968 @chapter Running Programs Under @value{GDBN}
1970 When you run a program under @value{GDBN}, you must first generate
1971 debugging information when you compile it.
1973 You may start @value{GDBN} with its arguments, if any, in an environment
1974 of your choice. If you are doing native debugging, you may redirect
1975 your program's input and output, debug an already running process, or
1976 kill a child process.
1979 * Compilation:: Compiling for debugging
1980 * Starting:: Starting your program
1981 * Arguments:: Your program's arguments
1982 * Environment:: Your program's environment
1984 * Working Directory:: Your program's working directory
1985 * Input/Output:: Your program's input and output
1986 * Attach:: Debugging an already-running process
1987 * Kill Process:: Killing the child process
1989 * Inferiors and Programs:: Debugging multiple inferiors and programs
1990 * Threads:: Debugging programs with multiple threads
1991 * Forks:: Debugging forks
1992 * Checkpoint/Restart:: Setting a @emph{bookmark} to return to later
1996 @section Compiling for Debugging
1998 In order to debug a program effectively, you need to generate
1999 debugging information when you compile it. This debugging information
2000 is stored in the object file; it describes the data type of each
2001 variable or function and the correspondence between source line numbers
2002 and addresses in the executable code.
2004 To request debugging information, specify the @samp{-g} option when you run
2007 Programs that are to be shipped to your customers are compiled with
2008 optimizations, using the @samp{-O} compiler option. However, some
2009 compilers are unable to handle the @samp{-g} and @samp{-O} options
2010 together. Using those compilers, you cannot generate optimized
2011 executables containing debugging information.
2013 @value{NGCC}, the @sc{gnu} C/C@t{++} compiler, supports @samp{-g} with or
2014 without @samp{-O}, making it possible to debug optimized code. We
2015 recommend that you @emph{always} use @samp{-g} whenever you compile a
2016 program. You may think your program is correct, but there is no sense
2017 in pushing your luck. For more information, see @ref{Optimized Code}.
2019 Older versions of the @sc{gnu} C compiler permitted a variant option
2020 @w{@samp{-gg}} for debugging information. @value{GDBN} no longer supports this
2021 format; if your @sc{gnu} C compiler has this option, do not use it.
2023 @value{GDBN} knows about preprocessor macros and can show you their
2024 expansion (@pxref{Macros}). Most compilers do not include information
2025 about preprocessor macros in the debugging information if you specify
2026 the @option{-g} flag alone. Version 3.1 and later of @value{NGCC},
2027 the @sc{gnu} C compiler, provides macro information if you are using
2028 the DWARF debugging format, and specify the option @option{-g3}.
2030 @xref{Debugging Options,,Options for Debugging Your Program or GCC,
2031 gcc, Using the @sc{gnu} Compiler Collection (GCC)}, for more
2032 information on @value{NGCC} options affecting debug information.
2034 You will have the best debugging experience if you use the latest
2035 version of the DWARF debugging format that your compiler supports.
2036 DWARF is currently the most expressive and best supported debugging
2037 format in @value{GDBN}.
2041 @section Starting your Program
2047 @kindex r @r{(@code{run})}
2050 Use the @code{run} command to start your program under @value{GDBN}.
2051 You must first specify the program name with an argument to
2052 @value{GDBN} (@pxref{Invocation, ,Getting In and Out of
2053 @value{GDBN}}), or by using the @code{file} or @code{exec-file}
2054 command (@pxref{Files, ,Commands to Specify Files}).
2058 If you are running your program in an execution environment that
2059 supports processes, @code{run} creates an inferior process and makes
2060 that process run your program. In some environments without processes,
2061 @code{run} jumps to the start of your program. Other targets,
2062 like @samp{remote}, are always running. If you get an error
2063 message like this one:
2066 The "remote" target does not support "run".
2067 Try "help target" or "continue".
2071 then use @code{continue} to run your program. You may need @code{load}
2072 first (@pxref{load}).
2074 The execution of a program is affected by certain information it
2075 receives from its superior. @value{GDBN} provides ways to specify this
2076 information, which you must do @emph{before} starting your program. (You
2077 can change it after starting your program, but such changes only affect
2078 your program the next time you start it.) This information may be
2079 divided into four categories:
2082 @item The @emph{arguments.}
2083 Specify the arguments to give your program as the arguments of the
2084 @code{run} command. If a shell is available on your target, the shell
2085 is used to pass the arguments, so that you may use normal conventions
2086 (such as wildcard expansion or variable substitution) in describing
2088 In Unix systems, you can control which shell is used with the
2089 @code{SHELL} environment variable. If you do not define @code{SHELL},
2090 @value{GDBN} uses the default shell (@file{/bin/sh}). You can disable
2091 use of any shell with the @code{set startup-with-shell} command (see
2094 @item The @emph{environment.}
2095 Your program normally inherits its environment from @value{GDBN}, but you can
2096 use the @value{GDBN} commands @code{set environment} and @code{unset
2097 environment} to change parts of the environment that affect
2098 your program. @xref{Environment, ,Your Program's Environment}.
2100 @item The @emph{working directory.}
2101 You can set your program's working directory with the command
2102 @kbd{set cwd}. If you do not set any working directory with this
2103 command, your program will inherit @value{GDBN}'s working directory if
2104 native debugging, or the remote server's working directory if remote
2105 debugging. @xref{Working Directory, ,Your Program's Working
2108 @item The @emph{standard input and output.}
2109 Your program normally uses the same device for standard input and
2110 standard output as @value{GDBN} is using. You can redirect input and output
2111 in the @code{run} command line, or you can use the @code{tty} command to
2112 set a different device for your program.
2113 @xref{Input/Output, ,Your Program's Input and Output}.
2116 @emph{Warning:} While input and output redirection work, you cannot use
2117 pipes to pass the output of the program you are debugging to another
2118 program; if you attempt this, @value{GDBN} is likely to wind up debugging the
2122 When you issue the @code{run} command, your program begins to execute
2123 immediately. @xref{Stopping, ,Stopping and Continuing}, for discussion
2124 of how to arrange for your program to stop. Once your program has
2125 stopped, you may call functions in your program, using the @code{print}
2126 or @code{call} commands. @xref{Data, ,Examining Data}.
2128 If the modification time of your symbol file has changed since the last
2129 time @value{GDBN} read its symbols, @value{GDBN} discards its symbol
2130 table, and reads it again. When it does this, @value{GDBN} tries to retain
2131 your current breakpoints.
2136 @cindex run to main procedure
2137 The name of the main procedure can vary from language to language.
2138 With C or C@t{++}, the main procedure name is always @code{main}, but
2139 other languages such as Ada do not require a specific name for their
2140 main procedure. The debugger provides a convenient way to start the
2141 execution of the program and to stop at the beginning of the main
2142 procedure, depending on the language used.
2144 The @samp{start} command does the equivalent of setting a temporary
2145 breakpoint at the beginning of the main procedure and then invoking
2146 the @samp{run} command.
2148 @cindex elaboration phase
2149 Some programs contain an @dfn{elaboration} phase where some startup code is
2150 executed before the main procedure is called. This depends on the
2151 languages used to write your program. In C@t{++}, for instance,
2152 constructors for static and global objects are executed before
2153 @code{main} is called. It is therefore possible that the debugger stops
2154 before reaching the main procedure. However, the temporary breakpoint
2155 will remain to halt execution.
2157 Specify the arguments to give to your program as arguments to the
2158 @samp{start} command. These arguments will be given verbatim to the
2159 underlying @samp{run} command. Note that the same arguments will be
2160 reused if no argument is provided during subsequent calls to
2161 @samp{start} or @samp{run}.
2163 It is sometimes necessary to debug the program during elaboration. In
2164 these cases, using the @code{start} command would stop the execution
2165 of your program too late, as the program would have already completed
2166 the elaboration phase. Under these circumstances, either insert
2167 breakpoints in your elaboration code before running your program or
2168 use the @code{starti} command.
2172 @cindex run to first instruction
2173 The @samp{starti} command does the equivalent of setting a temporary
2174 breakpoint at the first instruction of a program's execution and then
2175 invoking the @samp{run} command. For programs containing an
2176 elaboration phase, the @code{starti} command will stop execution at
2177 the start of the elaboration phase.
2179 @anchor{set exec-wrapper}
2180 @kindex set exec-wrapper
2181 @item set exec-wrapper @var{wrapper}
2182 @itemx show exec-wrapper
2183 @itemx unset exec-wrapper
2184 When @samp{exec-wrapper} is set, the specified wrapper is used to
2185 launch programs for debugging. @value{GDBN} starts your program
2186 with a shell command of the form @kbd{exec @var{wrapper}
2187 @var{program}}. Quoting is added to @var{program} and its
2188 arguments, but not to @var{wrapper}, so you should add quotes if
2189 appropriate for your shell. The wrapper runs until it executes
2190 your program, and then @value{GDBN} takes control.
2192 You can use any program that eventually calls @code{execve} with
2193 its arguments as a wrapper. Several standard Unix utilities do
2194 this, e.g.@: @code{env} and @code{nohup}. Any Unix shell script ending
2195 with @code{exec "$@@"} will also work.
2197 For example, you can use @code{env} to pass an environment variable to
2198 the debugged program, without setting the variable in your shell's
2202 (@value{GDBP}) set exec-wrapper env 'LD_PRELOAD=libtest.so'
2206 This command is available when debugging locally on most targets, excluding
2207 @sc{djgpp}, Cygwin, MS Windows, and QNX Neutrino.
2209 @kindex set startup-with-shell
2210 @anchor{set startup-with-shell}
2211 @item set startup-with-shell
2212 @itemx set startup-with-shell on
2213 @itemx set startup-with-shell off
2214 @itemx show startup-with-shell
2215 On Unix systems, by default, if a shell is available on your target,
2216 @value{GDBN}) uses it to start your program. Arguments of the
2217 @code{run} command are passed to the shell, which does variable
2218 substitution, expands wildcard characters and performs redirection of
2219 I/O. In some circumstances, it may be useful to disable such use of a
2220 shell, for example, when debugging the shell itself or diagnosing
2221 startup failures such as:
2225 Starting program: ./a.out
2226 During startup program terminated with signal SIGSEGV, Segmentation fault.
2230 which indicates the shell or the wrapper specified with
2231 @samp{exec-wrapper} crashed, not your program. Most often, this is
2232 caused by something odd in your shell's non-interactive mode
2233 initialization file---such as @file{.cshrc} for C-shell,
2234 $@file{.zshenv} for the Z shell, or the file specified in the
2235 @samp{BASH_ENV} environment variable for BASH.
2237 @anchor{set auto-connect-native-target}
2238 @kindex set auto-connect-native-target
2239 @item set auto-connect-native-target
2240 @itemx set auto-connect-native-target on
2241 @itemx set auto-connect-native-target off
2242 @itemx show auto-connect-native-target
2244 By default, if not connected to any target yet (e.g., with
2245 @code{target remote}), the @code{run} command starts your program as a
2246 native process under @value{GDBN}, on your local machine. If you're
2247 sure you don't want to debug programs on your local machine, you can
2248 tell @value{GDBN} to not connect to the native target automatically
2249 with the @code{set auto-connect-native-target off} command.
2251 If @code{on}, which is the default, and if @value{GDBN} is not
2252 connected to a target already, the @code{run} command automaticaly
2253 connects to the native target, if one is available.
2255 If @code{off}, and if @value{GDBN} is not connected to a target
2256 already, the @code{run} command fails with an error:
2260 Don't know how to run. Try "help target".
2263 If @value{GDBN} is already connected to a target, @value{GDBN} always
2264 uses it with the @code{run} command.
2266 In any case, you can explicitly connect to the native target with the
2267 @code{target native} command. For example,
2270 (@value{GDBP}) set auto-connect-native-target off
2272 Don't know how to run. Try "help target".
2273 (@value{GDBP}) target native
2275 Starting program: ./a.out
2276 [Inferior 1 (process 10421) exited normally]
2279 In case you connected explicitly to the @code{native} target,
2280 @value{GDBN} remains connected even if all inferiors exit, ready for
2281 the next @code{run} command. Use the @code{disconnect} command to
2284 Examples of other commands that likewise respect the
2285 @code{auto-connect-native-target} setting: @code{attach}, @code{info
2286 proc}, @code{info os}.
2288 @kindex set disable-randomization
2289 @item set disable-randomization
2290 @itemx set disable-randomization on
2291 This option (enabled by default in @value{GDBN}) will turn off the native
2292 randomization of the virtual address space of the started program. This option
2293 is useful for multiple debugging sessions to make the execution better
2294 reproducible and memory addresses reusable across debugging sessions.
2296 This feature is implemented only on certain targets, including @sc{gnu}/Linux.
2297 On @sc{gnu}/Linux you can get the same behavior using
2300 (@value{GDBP}) set exec-wrapper setarch `uname -m` -R
2303 @item set disable-randomization off
2304 Leave the behavior of the started executable unchanged. Some bugs rear their
2305 ugly heads only when the program is loaded at certain addresses. If your bug
2306 disappears when you run the program under @value{GDBN}, that might be because
2307 @value{GDBN} by default disables the address randomization on platforms, such
2308 as @sc{gnu}/Linux, which do that for stand-alone programs. Use @kbd{set
2309 disable-randomization off} to try to reproduce such elusive bugs.
2311 On targets where it is available, virtual address space randomization
2312 protects the programs against certain kinds of security attacks. In these
2313 cases the attacker needs to know the exact location of a concrete executable
2314 code. Randomizing its location makes it impossible to inject jumps misusing
2315 a code at its expected addresses.
2317 Prelinking shared libraries provides a startup performance advantage but it
2318 makes addresses in these libraries predictable for privileged processes by
2319 having just unprivileged access at the target system. Reading the shared
2320 library binary gives enough information for assembling the malicious code
2321 misusing it. Still even a prelinked shared library can get loaded at a new
2322 random address just requiring the regular relocation process during the
2323 startup. Shared libraries not already prelinked are always loaded at
2324 a randomly chosen address.
2326 Position independent executables (PIE) contain position independent code
2327 similar to the shared libraries and therefore such executables get loaded at
2328 a randomly chosen address upon startup. PIE executables always load even
2329 already prelinked shared libraries at a random address. You can build such
2330 executable using @command{gcc -fPIE -pie}.
2332 Heap (malloc storage), stack and custom mmap areas are always placed randomly
2333 (as long as the randomization is enabled).
2335 @item show disable-randomization
2336 Show the current setting of the explicit disable of the native randomization of
2337 the virtual address space of the started program.
2342 @section Your Program's Arguments
2344 @cindex arguments (to your program)
2345 The arguments to your program can be specified by the arguments of the
2347 They are passed to a shell, which expands wildcard characters and
2348 performs redirection of I/O, and thence to your program. Your
2349 @code{SHELL} environment variable (if it exists) specifies what shell
2350 @value{GDBN} uses. If you do not define @code{SHELL}, @value{GDBN} uses
2351 the default shell (@file{/bin/sh} on Unix).
2353 On non-Unix systems, the program is usually invoked directly by
2354 @value{GDBN}, which emulates I/O redirection via the appropriate system
2355 calls, and the wildcard characters are expanded by the startup code of
2356 the program, not by the shell.
2358 @code{run} with no arguments uses the same arguments used by the previous
2359 @code{run}, or those set by the @code{set args} command.
2364 Specify the arguments to be used the next time your program is run. If
2365 @code{set args} has no arguments, @code{run} executes your program
2366 with no arguments. Once you have run your program with arguments,
2367 using @code{set args} before the next @code{run} is the only way to run
2368 it again without arguments.
2372 Show the arguments to give your program when it is started.
2376 @section Your Program's Environment
2378 @cindex environment (of your program)
2379 The @dfn{environment} consists of a set of environment variables and
2380 their values. Environment variables conventionally record such things as
2381 your user name, your home directory, your terminal type, and your search
2382 path for programs to run. Usually you set up environment variables with
2383 the shell and they are inherited by all the other programs you run. When
2384 debugging, it can be useful to try running your program with a modified
2385 environment without having to start @value{GDBN} over again.
2389 @item path @var{directory}
2390 Add @var{directory} to the front of the @code{PATH} environment variable
2391 (the search path for executables) that will be passed to your program.
2392 The value of @code{PATH} used by @value{GDBN} does not change.
2393 You may specify several directory names, separated by whitespace or by a
2394 system-dependent separator character (@samp{:} on Unix, @samp{;} on
2395 MS-DOS and MS-Windows). If @var{directory} is already in the path, it
2396 is moved to the front, so it is searched sooner.
2398 You can use the string @samp{$cwd} to refer to whatever is the current
2399 working directory at the time @value{GDBN} searches the path. If you
2400 use @samp{.} instead, it refers to the directory where you executed the
2401 @code{path} command. @value{GDBN} replaces @samp{.} in the
2402 @var{directory} argument (with the current path) before adding
2403 @var{directory} to the search path.
2404 @c 'path' is explicitly nonrepeatable, but RMS points out it is silly to
2405 @c document that, since repeating it would be a no-op.
2409 Display the list of search paths for executables (the @code{PATH}
2410 environment variable).
2412 @kindex show environment
2413 @item show environment @r{[}@var{varname}@r{]}
2414 Print the value of environment variable @var{varname} to be given to
2415 your program when it starts. If you do not supply @var{varname},
2416 print the names and values of all environment variables to be given to
2417 your program. You can abbreviate @code{environment} as @code{env}.
2419 @kindex set environment
2420 @anchor{set environment}
2421 @item set environment @var{varname} @r{[}=@var{value}@r{]}
2422 Set environment variable @var{varname} to @var{value}. The value
2423 changes for your program (and the shell @value{GDBN} uses to launch
2424 it), not for @value{GDBN} itself. The @var{value} may be any string; the
2425 values of environment variables are just strings, and any
2426 interpretation is supplied by your program itself. The @var{value}
2427 parameter is optional; if it is eliminated, the variable is set to a
2429 @c "any string" here does not include leading, trailing
2430 @c blanks. Gnu asks: does anyone care?
2432 For example, this command:
2439 tells the debugged program, when subsequently run, that its user is named
2440 @samp{foo}. (The spaces around @samp{=} are used for clarity here; they
2441 are not actually required.)
2443 Note that on Unix systems, @value{GDBN} runs your program via a shell,
2444 which also inherits the environment set with @code{set environment}.
2445 If necessary, you can avoid that by using the @samp{env} program as a
2446 wrapper instead of using @code{set environment}. @xref{set
2447 exec-wrapper}, for an example doing just that.
2449 Environment variables that are set by the user are also transmitted to
2450 @command{gdbserver} to be used when starting the remote inferior.
2451 @pxref{QEnvironmentHexEncoded}.
2453 @kindex unset environment
2454 @anchor{unset environment}
2455 @item unset environment @var{varname}
2456 Remove variable @var{varname} from the environment to be passed to your
2457 program. This is different from @samp{set env @var{varname} =};
2458 @code{unset environment} removes the variable from the environment,
2459 rather than assigning it an empty value.
2461 Environment variables that are unset by the user are also unset on
2462 @command{gdbserver} when starting the remote inferior.
2463 @pxref{QEnvironmentUnset}.
2466 @emph{Warning:} On Unix systems, @value{GDBN} runs your program using
2467 the shell indicated by your @code{SHELL} environment variable if it
2468 exists (or @code{/bin/sh} if not). If your @code{SHELL} variable
2469 names a shell that runs an initialization file when started
2470 non-interactively---such as @file{.cshrc} for C-shell, $@file{.zshenv}
2471 for the Z shell, or the file specified in the @samp{BASH_ENV}
2472 environment variable for BASH---any variables you set in that file
2473 affect your program. You may wish to move setting of environment
2474 variables to files that are only run when you sign on, such as
2475 @file{.login} or @file{.profile}.
2477 @node Working Directory
2478 @section Your Program's Working Directory
2480 @cindex working directory (of your program)
2481 Each time you start your program with @code{run}, the inferior will be
2482 initialized with the current working directory specified by the
2483 @kbd{set cwd} command. If no directory has been specified by this
2484 command, then the inferior will inherit @value{GDBN}'s current working
2485 directory as its working directory if native debugging, or it will
2486 inherit the remote server's current working directory if remote
2491 @cindex change inferior's working directory
2492 @anchor{set cwd command}
2493 @item set cwd @r{[}@var{directory}@r{]}
2494 Set the inferior's working directory to @var{directory}, which will be
2495 @code{glob}-expanded in order to resolve tildes (@file{~}). If no
2496 argument has been specified, the command clears the setting and resets
2497 it to an empty state. This setting has no effect on @value{GDBN}'s
2498 working directory, and it only takes effect the next time you start
2499 the inferior. The @file{~} in @var{directory} is a short for the
2500 @dfn{home directory}, usually pointed to by the @env{HOME} environment
2501 variable. On MS-Windows, if @env{HOME} is not defined, @value{GDBN}
2502 uses the concatenation of @env{HOMEDRIVE} and @env{HOMEPATH} as
2505 You can also change @value{GDBN}'s current working directory by using
2506 the @code{cd} command.
2510 @cindex show inferior's working directory
2512 Show the inferior's working directory. If no directory has been
2513 specified by @kbd{set cwd}, then the default inferior's working
2514 directory is the same as @value{GDBN}'s working directory.
2517 @cindex change @value{GDBN}'s working directory
2519 @item cd @r{[}@var{directory}@r{]}
2520 Set the @value{GDBN} working directory to @var{directory}. If not
2521 given, @var{directory} uses @file{'~'}.
2523 The @value{GDBN} working directory serves as a default for the
2524 commands that specify files for @value{GDBN} to operate on.
2525 @xref{Files, ,Commands to Specify Files}.
2526 @xref{set cwd command}.
2530 Print the @value{GDBN} working directory.
2533 It is generally impossible to find the current working directory of
2534 the process being debugged (since a program can change its directory
2535 during its run). If you work on a system where @value{GDBN} supports
2536 the @code{info proc} command (@pxref{Process Information}), you can
2537 use the @code{info proc} command to find out the
2538 current working directory of the debuggee.
2541 @section Your Program's Input and Output
2546 By default, the program you run under @value{GDBN} does input and output to
2547 the same terminal that @value{GDBN} uses. @value{GDBN} switches the terminal
2548 to its own terminal modes to interact with you, but it records the terminal
2549 modes your program was using and switches back to them when you continue
2550 running your program.
2553 @kindex info terminal
2555 Displays information recorded by @value{GDBN} about the terminal modes your
2559 You can redirect your program's input and/or output using shell
2560 redirection with the @code{run} command. For example,
2567 starts your program, diverting its output to the file @file{outfile}.
2570 @cindex controlling terminal
2571 Another way to specify where your program should do input and output is
2572 with the @code{tty} command. This command accepts a file name as
2573 argument, and causes this file to be the default for future @code{run}
2574 commands. It also resets the controlling terminal for the child
2575 process, for future @code{run} commands. For example,
2582 directs that processes started with subsequent @code{run} commands
2583 default to do input and output on the terminal @file{/dev/ttyb} and have
2584 that as their controlling terminal.
2586 An explicit redirection in @code{run} overrides the @code{tty} command's
2587 effect on the input/output device, but not its effect on the controlling
2590 When you use the @code{tty} command or redirect input in the @code{run}
2591 command, only the input @emph{for your program} is affected. The input
2592 for @value{GDBN} still comes from your terminal. @code{tty} is an alias
2593 for @code{set inferior-tty}.
2595 @cindex inferior tty
2596 @cindex set inferior controlling terminal
2597 You can use the @code{show inferior-tty} command to tell @value{GDBN} to
2598 display the name of the terminal that will be used for future runs of your
2602 @item set inferior-tty [ @var{tty} ]
2603 @kindex set inferior-tty
2604 Set the tty for the program being debugged to @var{tty}. Omitting @var{tty}
2605 restores the default behavior, which is to use the same terminal as
2608 @item show inferior-tty
2609 @kindex show inferior-tty
2610 Show the current tty for the program being debugged.
2614 @section Debugging an Already-running Process
2619 @item attach @var{process-id}
2620 This command attaches to a running process---one that was started
2621 outside @value{GDBN}. (@code{info files} shows your active
2622 targets.) The command takes as argument a process ID. The usual way to
2623 find out the @var{process-id} of a Unix process is with the @code{ps} utility,
2624 or with the @samp{jobs -l} shell command.
2626 @code{attach} does not repeat if you press @key{RET} a second time after
2627 executing the command.
2630 To use @code{attach}, your program must be running in an environment
2631 which supports processes; for example, @code{attach} does not work for
2632 programs on bare-board targets that lack an operating system. You must
2633 also have permission to send the process a signal.
2635 When you use @code{attach}, the debugger finds the program running in
2636 the process first by looking in the current working directory, then (if
2637 the program is not found) by using the source file search path
2638 (@pxref{Source Path, ,Specifying Source Directories}). You can also use
2639 the @code{file} command to load the program. @xref{Files, ,Commands to
2642 The first thing @value{GDBN} does after arranging to debug the specified
2643 process is to stop it. You can examine and modify an attached process
2644 with all the @value{GDBN} commands that are ordinarily available when
2645 you start processes with @code{run}. You can insert breakpoints; you
2646 can step and continue; you can modify storage. If you would rather the
2647 process continue running, you may use the @code{continue} command after
2648 attaching @value{GDBN} to the process.
2653 When you have finished debugging the attached process, you can use the
2654 @code{detach} command to release it from @value{GDBN} control. Detaching
2655 the process continues its execution. After the @code{detach} command,
2656 that process and @value{GDBN} become completely independent once more, and you
2657 are ready to @code{attach} another process or start one with @code{run}.
2658 @code{detach} does not repeat if you press @key{RET} again after
2659 executing the command.
2662 If you exit @value{GDBN} while you have an attached process, you detach
2663 that process. If you use the @code{run} command, you kill that process.
2664 By default, @value{GDBN} asks for confirmation if you try to do either of these
2665 things; you can control whether or not you need to confirm by using the
2666 @code{set confirm} command (@pxref{Messages/Warnings, ,Optional Warnings and
2670 @section Killing the Child Process
2675 Kill the child process in which your program is running under @value{GDBN}.
2678 This command is useful if you wish to debug a core dump instead of a
2679 running process. @value{GDBN} ignores any core dump file while your program
2682 On some operating systems, a program cannot be executed outside @value{GDBN}
2683 while you have breakpoints set on it inside @value{GDBN}. You can use the
2684 @code{kill} command in this situation to permit running your program
2685 outside the debugger.
2687 The @code{kill} command is also useful if you wish to recompile and
2688 relink your program, since on many systems it is impossible to modify an
2689 executable file while it is running in a process. In this case, when you
2690 next type @code{run}, @value{GDBN} notices that the file has changed, and
2691 reads the symbol table again (while trying to preserve your current
2692 breakpoint settings).
2694 @node Inferiors and Programs
2695 @section Debugging Multiple Inferiors and Programs
2697 @value{GDBN} lets you run and debug multiple programs in a single
2698 session. In addition, @value{GDBN} on some systems may let you run
2699 several programs simultaneously (otherwise you have to exit from one
2700 before starting another). In the most general case, you can have
2701 multiple threads of execution in each of multiple processes, launched
2702 from multiple executables.
2705 @value{GDBN} represents the state of each program execution with an
2706 object called an @dfn{inferior}. An inferior typically corresponds to
2707 a process, but is more general and applies also to targets that do not
2708 have processes. Inferiors may be created before a process runs, and
2709 may be retained after a process exits. Inferiors have unique
2710 identifiers that are different from process ids. Usually each
2711 inferior will also have its own distinct address space, although some
2712 embedded targets may have several inferiors running in different parts
2713 of a single address space. Each inferior may in turn have multiple
2714 threads running in it.
2716 To find out what inferiors exist at any moment, use @w{@code{info
2720 @kindex info inferiors [ @var{id}@dots{} ]
2721 @item info inferiors
2722 Print a list of all inferiors currently being managed by @value{GDBN}.
2723 By default all inferiors are printed, but the argument @var{id}@dots{}
2724 -- a space separated list of inferior numbers -- can be used to limit
2725 the display to just the requested inferiors.
2727 @value{GDBN} displays for each inferior (in this order):
2731 the inferior number assigned by @value{GDBN}
2734 the target system's inferior identifier
2737 the name of the executable the inferior is running.
2742 An asterisk @samp{*} preceding the @value{GDBN} inferior number
2743 indicates the current inferior.
2747 @c end table here to get a little more width for example
2750 (@value{GDBP}) info inferiors
2751 Num Description Executable
2752 2 process 2307 hello
2753 * 1 process 3401 goodbye
2756 To switch focus between inferiors, use the @code{inferior} command:
2759 @kindex inferior @var{infno}
2760 @item inferior @var{infno}
2761 Make inferior number @var{infno} the current inferior. The argument
2762 @var{infno} is the inferior number assigned by @value{GDBN}, as shown
2763 in the first field of the @samp{info inferiors} display.
2766 @vindex $_inferior@r{, convenience variable}
2767 The debugger convenience variable @samp{$_inferior} contains the
2768 number of the current inferior. You may find this useful in writing
2769 breakpoint conditional expressions, command scripts, and so forth.
2770 @xref{Convenience Vars,, Convenience Variables}, for general
2771 information on convenience variables.
2773 You can get multiple executables into a debugging session via the
2774 @code{add-inferior} and @w{@code{clone-inferior}} commands. On some
2775 systems @value{GDBN} can add inferiors to the debug session
2776 automatically by following calls to @code{fork} and @code{exec}. To
2777 remove inferiors from the debugging session use the
2778 @w{@code{remove-inferiors}} command.
2781 @kindex add-inferior
2782 @item add-inferior [ -copies @var{n} ] [ -exec @var{executable} ]
2783 Adds @var{n} inferiors to be run using @var{executable} as the
2784 executable; @var{n} defaults to 1. If no executable is specified,
2785 the inferiors begins empty, with no program. You can still assign or
2786 change the program assigned to the inferior at any time by using the
2787 @code{file} command with the executable name as its argument.
2789 @kindex clone-inferior
2790 @item clone-inferior [ -copies @var{n} ] [ @var{infno} ]
2791 Adds @var{n} inferiors ready to execute the same program as inferior
2792 @var{infno}; @var{n} defaults to 1, and @var{infno} defaults to the
2793 number of the current inferior. This is a convenient command when you
2794 want to run another instance of the inferior you are debugging.
2797 (@value{GDBP}) info inferiors
2798 Num Description Executable
2799 * 1 process 29964 helloworld
2800 (@value{GDBP}) clone-inferior
2803 (@value{GDBP}) info inferiors
2804 Num Description Executable
2806 * 1 process 29964 helloworld
2809 You can now simply switch focus to inferior 2 and run it.
2811 @kindex remove-inferiors
2812 @item remove-inferiors @var{infno}@dots{}
2813 Removes the inferior or inferiors @var{infno}@dots{}. It is not
2814 possible to remove an inferior that is running with this command. For
2815 those, use the @code{kill} or @code{detach} command first.
2819 To quit debugging one of the running inferiors that is not the current
2820 inferior, you can either detach from it by using the @w{@code{detach
2821 inferior}} command (allowing it to run independently), or kill it
2822 using the @w{@code{kill inferiors}} command:
2825 @kindex detach inferiors @var{infno}@dots{}
2826 @item detach inferior @var{infno}@dots{}
2827 Detach from the inferior or inferiors identified by @value{GDBN}
2828 inferior number(s) @var{infno}@dots{}. Note that the inferior's entry
2829 still stays on the list of inferiors shown by @code{info inferiors},
2830 but its Description will show @samp{<null>}.
2832 @kindex kill inferiors @var{infno}@dots{}
2833 @item kill inferiors @var{infno}@dots{}
2834 Kill the inferior or inferiors identified by @value{GDBN} inferior
2835 number(s) @var{infno}@dots{}. Note that the inferior's entry still
2836 stays on the list of inferiors shown by @code{info inferiors}, but its
2837 Description will show @samp{<null>}.
2840 After the successful completion of a command such as @code{detach},
2841 @code{detach inferiors}, @code{kill} or @code{kill inferiors}, or after
2842 a normal process exit, the inferior is still valid and listed with
2843 @code{info inferiors}, ready to be restarted.
2846 To be notified when inferiors are started or exit under @value{GDBN}'s
2847 control use @w{@code{set print inferior-events}}:
2850 @kindex set print inferior-events
2851 @cindex print messages on inferior start and exit
2852 @item set print inferior-events
2853 @itemx set print inferior-events on
2854 @itemx set print inferior-events off
2855 The @code{set print inferior-events} command allows you to enable or
2856 disable printing of messages when @value{GDBN} notices that new
2857 inferiors have started or that inferiors have exited or have been
2858 detached. By default, these messages will not be printed.
2860 @kindex show print inferior-events
2861 @item show print inferior-events
2862 Show whether messages will be printed when @value{GDBN} detects that
2863 inferiors have started, exited or have been detached.
2866 Many commands will work the same with multiple programs as with a
2867 single program: e.g., @code{print myglobal} will simply display the
2868 value of @code{myglobal} in the current inferior.
2871 Occasionaly, when debugging @value{GDBN} itself, it may be useful to
2872 get more info about the relationship of inferiors, programs, address
2873 spaces in a debug session. You can do that with the @w{@code{maint
2874 info program-spaces}} command.
2877 @kindex maint info program-spaces
2878 @item maint info program-spaces
2879 Print a list of all program spaces currently being managed by
2882 @value{GDBN} displays for each program space (in this order):
2886 the program space number assigned by @value{GDBN}
2889 the name of the executable loaded into the program space, with e.g.,
2890 the @code{file} command.
2895 An asterisk @samp{*} preceding the @value{GDBN} program space number
2896 indicates the current program space.
2898 In addition, below each program space line, @value{GDBN} prints extra
2899 information that isn't suitable to display in tabular form. For
2900 example, the list of inferiors bound to the program space.
2903 (@value{GDBP}) maint info program-spaces
2907 Bound inferiors: ID 1 (process 21561)
2910 Here we can see that no inferior is running the program @code{hello},
2911 while @code{process 21561} is running the program @code{goodbye}. On
2912 some targets, it is possible that multiple inferiors are bound to the
2913 same program space. The most common example is that of debugging both
2914 the parent and child processes of a @code{vfork} call. For example,
2917 (@value{GDBP}) maint info program-spaces
2920 Bound inferiors: ID 2 (process 18050), ID 1 (process 18045)
2923 Here, both inferior 2 and inferior 1 are running in the same program
2924 space as a result of inferior 1 having executed a @code{vfork} call.
2928 @section Debugging Programs with Multiple Threads
2930 @cindex threads of execution
2931 @cindex multiple threads
2932 @cindex switching threads
2933 In some operating systems, such as GNU/Linux and Solaris, a single program
2934 may have more than one @dfn{thread} of execution. The precise semantics
2935 of threads differ from one operating system to another, but in general
2936 the threads of a single program are akin to multiple processes---except
2937 that they share one address space (that is, they can all examine and
2938 modify the same variables). On the other hand, each thread has its own
2939 registers and execution stack, and perhaps private memory.
2941 @value{GDBN} provides these facilities for debugging multi-thread
2945 @item automatic notification of new threads
2946 @item @samp{thread @var{thread-id}}, a command to switch among threads
2947 @item @samp{info threads}, a command to inquire about existing threads
2948 @item @samp{thread apply [@var{thread-id-list} | all] @var{args}},
2949 a command to apply a command to a list of threads
2950 @item thread-specific breakpoints
2951 @item @samp{set print thread-events}, which controls printing of
2952 messages on thread start and exit.
2953 @item @samp{set libthread-db-search-path @var{path}}, which lets
2954 the user specify which @code{libthread_db} to use if the default choice
2955 isn't compatible with the program.
2958 @cindex focus of debugging
2959 @cindex current thread
2960 The @value{GDBN} thread debugging facility allows you to observe all
2961 threads while your program runs---but whenever @value{GDBN} takes
2962 control, one thread in particular is always the focus of debugging.
2963 This thread is called the @dfn{current thread}. Debugging commands show
2964 program information from the perspective of the current thread.
2966 @cindex @code{New} @var{systag} message
2967 @cindex thread identifier (system)
2968 @c FIXME-implementors!! It would be more helpful if the [New...] message
2969 @c included GDB's numeric thread handle, so you could just go to that
2970 @c thread without first checking `info threads'.
2971 Whenever @value{GDBN} detects a new thread in your program, it displays
2972 the target system's identification for the thread with a message in the
2973 form @samp{[New @var{systag}]}, where @var{systag} is a thread identifier
2974 whose form varies depending on the particular system. For example, on
2975 @sc{gnu}/Linux, you might see
2978 [New Thread 0x41e02940 (LWP 25582)]
2982 when @value{GDBN} notices a new thread. In contrast, on other systems,
2983 the @var{systag} is simply something like @samp{process 368}, with no
2986 @c FIXME!! (1) Does the [New...] message appear even for the very first
2987 @c thread of a program, or does it only appear for the
2988 @c second---i.e.@: when it becomes obvious we have a multithread
2990 @c (2) *Is* there necessarily a first thread always? Or do some
2991 @c multithread systems permit starting a program with multiple
2992 @c threads ab initio?
2994 @anchor{thread numbers}
2995 @cindex thread number, per inferior
2996 @cindex thread identifier (GDB)
2997 For debugging purposes, @value{GDBN} associates its own thread number
2998 ---always a single integer---with each thread of an inferior. This
2999 number is unique between all threads of an inferior, but not unique
3000 between threads of different inferiors.
3002 @cindex qualified thread ID
3003 You can refer to a given thread in an inferior using the qualified
3004 @var{inferior-num}.@var{thread-num} syntax, also known as
3005 @dfn{qualified thread ID}, with @var{inferior-num} being the inferior
3006 number and @var{thread-num} being the thread number of the given
3007 inferior. For example, thread @code{2.3} refers to thread number 3 of
3008 inferior 2. If you omit @var{inferior-num} (e.g., @code{thread 3}),
3009 then @value{GDBN} infers you're referring to a thread of the current
3012 Until you create a second inferior, @value{GDBN} does not show the
3013 @var{inferior-num} part of thread IDs, even though you can always use
3014 the full @var{inferior-num}.@var{thread-num} form to refer to threads
3015 of inferior 1, the initial inferior.
3017 @anchor{thread ID lists}
3018 @cindex thread ID lists
3019 Some commands accept a space-separated @dfn{thread ID list} as
3020 argument. A list element can be:
3024 A thread ID as shown in the first field of the @samp{info threads}
3025 display, with or without an inferior qualifier. E.g., @samp{2.1} or
3029 A range of thread numbers, again with or without an inferior
3030 qualifier, as in @var{inf}.@var{thr1}-@var{thr2} or
3031 @var{thr1}-@var{thr2}. E.g., @samp{1.2-4} or @samp{2-4}.
3034 All threads of an inferior, specified with a star wildcard, with or
3035 without an inferior qualifier, as in @var{inf}.@code{*} (e.g.,
3036 @samp{1.*}) or @code{*}. The former refers to all threads of the
3037 given inferior, and the latter form without an inferior qualifier
3038 refers to all threads of the current inferior.
3042 For example, if the current inferior is 1, and inferior 7 has one
3043 thread with ID 7.1, the thread list @samp{1 2-3 4.5 6.7-9 7.*}
3044 includes threads 1 to 3 of inferior 1, thread 5 of inferior 4, threads
3045 7 to 9 of inferior 6 and all threads of inferior 7. That is, in
3046 expanded qualified form, the same as @samp{1.1 1.2 1.3 4.5 6.7 6.8 6.9
3050 @anchor{global thread numbers}
3051 @cindex global thread number
3052 @cindex global thread identifier (GDB)
3053 In addition to a @emph{per-inferior} number, each thread is also
3054 assigned a unique @emph{global} number, also known as @dfn{global
3055 thread ID}, a single integer. Unlike the thread number component of
3056 the thread ID, no two threads have the same global ID, even when
3057 you're debugging multiple inferiors.
3059 From @value{GDBN}'s perspective, a process always has at least one
3060 thread. In other words, @value{GDBN} assigns a thread number to the
3061 program's ``main thread'' even if the program is not multi-threaded.
3063 @vindex $_thread@r{, convenience variable}
3064 @vindex $_gthread@r{, convenience variable}
3065 The debugger convenience variables @samp{$_thread} and
3066 @samp{$_gthread} contain, respectively, the per-inferior thread number
3067 and the global thread number of the current thread. You may find this
3068 useful in writing breakpoint conditional expressions, command scripts,
3069 and so forth. @xref{Convenience Vars,, Convenience Variables}, for
3070 general information on convenience variables.
3072 If @value{GDBN} detects the program is multi-threaded, it augments the
3073 usual message about stopping at a breakpoint with the ID and name of
3074 the thread that hit the breakpoint.
3077 Thread 2 "client" hit Breakpoint 1, send_message () at client.c:68
3080 Likewise when the program receives a signal:
3083 Thread 1 "main" received signal SIGINT, Interrupt.
3087 @kindex info threads
3088 @item info threads @r{[}@var{thread-id-list}@r{]}
3090 Display information about one or more threads. With no arguments
3091 displays information about all threads. You can specify the list of
3092 threads that you want to display using the thread ID list syntax
3093 (@pxref{thread ID lists}).
3095 @value{GDBN} displays for each thread (in this order):
3099 the per-inferior thread number assigned by @value{GDBN}
3102 the global thread number assigned by @value{GDBN}, if the @samp{-gid}
3103 option was specified
3106 the target system's thread identifier (@var{systag})
3109 the thread's name, if one is known. A thread can either be named by
3110 the user (see @code{thread name}, below), or, in some cases, by the
3114 the current stack frame summary for that thread
3118 An asterisk @samp{*} to the left of the @value{GDBN} thread number
3119 indicates the current thread.
3123 @c end table here to get a little more width for example
3126 (@value{GDBP}) info threads
3128 * 1 process 35 thread 13 main (argc=1, argv=0x7ffffff8)
3129 2 process 35 thread 23 0x34e5 in sigpause ()
3130 3 process 35 thread 27 0x34e5 in sigpause ()
3134 If you're debugging multiple inferiors, @value{GDBN} displays thread
3135 IDs using the qualified @var{inferior-num}.@var{thread-num} format.
3136 Otherwise, only @var{thread-num} is shown.
3138 If you specify the @samp{-gid} option, @value{GDBN} displays a column
3139 indicating each thread's global thread ID:
3142 (@value{GDBP}) info threads
3143 Id GId Target Id Frame
3144 1.1 1 process 35 thread 13 main (argc=1, argv=0x7ffffff8)
3145 1.2 3 process 35 thread 23 0x34e5 in sigpause ()
3146 1.3 4 process 35 thread 27 0x34e5 in sigpause ()
3147 * 2.1 2 process 65 thread 1 main (argc=1, argv=0x7ffffff8)
3150 On Solaris, you can display more information about user threads with a
3151 Solaris-specific command:
3154 @item maint info sol-threads
3155 @kindex maint info sol-threads
3156 @cindex thread info (Solaris)
3157 Display info on Solaris user threads.
3161 @kindex thread @var{thread-id}
3162 @item thread @var{thread-id}
3163 Make thread ID @var{thread-id} the current thread. The command
3164 argument @var{thread-id} is the @value{GDBN} thread ID, as shown in
3165 the first field of the @samp{info threads} display, with or without an
3166 inferior qualifier (e.g., @samp{2.1} or @samp{1}).
3168 @value{GDBN} responds by displaying the system identifier of the
3169 thread you selected, and its current stack frame summary:
3172 (@value{GDBP}) thread 2
3173 [Switching to thread 2 (Thread 0xb7fdab70 (LWP 12747))]
3174 #0 some_function (ignore=0x0) at example.c:8
3175 8 printf ("hello\n");
3179 As with the @samp{[New @dots{}]} message, the form of the text after
3180 @samp{Switching to} depends on your system's conventions for identifying
3183 @kindex thread apply
3184 @cindex apply command to several threads
3185 @item thread apply [@var{thread-id-list} | all [-ascending]] [@var{flag}]@dots{} @var{command}
3186 The @code{thread apply} command allows you to apply the named
3187 @var{command} to one or more threads. Specify the threads that you
3188 want affected using the thread ID list syntax (@pxref{thread ID
3189 lists}), or specify @code{all} to apply to all threads. To apply a
3190 command to all threads in descending order, type @kbd{thread apply all
3191 @var{command}}. To apply a command to all threads in ascending order,
3192 type @kbd{thread apply all -ascending @var{command}}.
3194 The @var{flag} arguments control what output to produce and how to handle
3195 errors raised when applying @var{command} to a thread. @var{flag}
3196 must start with a @code{-} directly followed by one letter in
3197 @code{qcs}. If several flags are provided, they must be given
3198 individually, such as @code{-c -q}.
3200 By default, @value{GDBN} displays some thread information before the
3201 output produced by @var{command}, and an error raised during the
3202 execution of a @var{command} will abort @code{thread apply}. The
3203 following flags can be used to fine-tune this behavior:
3207 The flag @code{-c}, which stands for @samp{continue}, causes any
3208 errors in @var{command} to be displayed, and the execution of
3209 @code{thread apply} then continues.
3211 The flag @code{-s}, which stands for @samp{silent}, causes any errors
3212 or empty output produced by a @var{command} to be silently ignored.
3213 That is, the execution continues, but the thread information and errors
3216 The flag @code{-q} (@samp{quiet}) disables printing the thread
3220 Flags @code{-c} and @code{-s} cannot be used together.
3223 @cindex apply command to all threads (ignoring errors and empty output)
3224 @item taas @var{command}
3225 Shortcut for @code{thread apply all -s @var{command}}.
3226 Applies @var{command} on all threads, ignoring errors and empty output.
3229 @cindex apply a command to all frames of all threads (ignoring errors and empty output)
3230 @item tfaas @var{command}
3231 Shortcut for @code{thread apply all -s frame apply all -s @var{command}}.
3232 Applies @var{command} on all frames of all threads, ignoring errors
3233 and empty output. Note that the flag @code{-s} is specified twice:
3234 The first @code{-s} ensures that @code{thread apply} only shows the thread
3235 information of the threads for which @code{frame apply} produces
3236 some output. The second @code{-s} is needed to ensure that @code{frame
3237 apply} shows the frame information of a frame only if the
3238 @var{command} successfully produced some output.
3240 It can for example be used to print a local variable or a function
3241 argument without knowing the thread or frame where this variable or argument
3244 (@value{GDBP}) tfaas p some_local_var_i_do_not_remember_where_it_is
3249 @cindex name a thread
3250 @item thread name [@var{name}]
3251 This command assigns a name to the current thread. If no argument is
3252 given, any existing user-specified name is removed. The thread name
3253 appears in the @samp{info threads} display.
3255 On some systems, such as @sc{gnu}/Linux, @value{GDBN} is able to
3256 determine the name of the thread as given by the OS. On these
3257 systems, a name specified with @samp{thread name} will override the
3258 system-give name, and removing the user-specified name will cause
3259 @value{GDBN} to once again display the system-specified name.
3262 @cindex search for a thread
3263 @item thread find [@var{regexp}]
3264 Search for and display thread ids whose name or @var{systag}
3265 matches the supplied regular expression.
3267 As well as being the complement to the @samp{thread name} command,
3268 this command also allows you to identify a thread by its target
3269 @var{systag}. For instance, on @sc{gnu}/Linux, the target @var{systag}
3273 (@value{GDBN}) thread find 26688
3274 Thread 4 has target id 'Thread 0x41e02940 (LWP 26688)'
3275 (@value{GDBN}) info thread 4
3277 4 Thread 0x41e02940 (LWP 26688) 0x00000031ca6cd372 in select ()
3280 @kindex set print thread-events
3281 @cindex print messages on thread start and exit
3282 @item set print thread-events
3283 @itemx set print thread-events on
3284 @itemx set print thread-events off
3285 The @code{set print thread-events} command allows you to enable or
3286 disable printing of messages when @value{GDBN} notices that new threads have
3287 started or that threads have exited. By default, these messages will
3288 be printed if detection of these events is supported by the target.
3289 Note that these messages cannot be disabled on all targets.
3291 @kindex show print thread-events
3292 @item show print thread-events
3293 Show whether messages will be printed when @value{GDBN} detects that threads
3294 have started and exited.
3297 @xref{Thread Stops,,Stopping and Starting Multi-thread Programs}, for
3298 more information about how @value{GDBN} behaves when you stop and start
3299 programs with multiple threads.
3301 @xref{Set Watchpoints,,Setting Watchpoints}, for information about
3302 watchpoints in programs with multiple threads.
3304 @anchor{set libthread-db-search-path}
3306 @kindex set libthread-db-search-path
3307 @cindex search path for @code{libthread_db}
3308 @item set libthread-db-search-path @r{[}@var{path}@r{]}
3309 If this variable is set, @var{path} is a colon-separated list of
3310 directories @value{GDBN} will use to search for @code{libthread_db}.
3311 If you omit @var{path}, @samp{libthread-db-search-path} will be reset to
3312 its default value (@code{$sdir:$pdir} on @sc{gnu}/Linux and Solaris systems).
3313 Internally, the default value comes from the @code{LIBTHREAD_DB_SEARCH_PATH}
3316 On @sc{gnu}/Linux and Solaris systems, @value{GDBN} uses a ``helper''
3317 @code{libthread_db} library to obtain information about threads in the
3318 inferior process. @value{GDBN} will use @samp{libthread-db-search-path}
3319 to find @code{libthread_db}. @value{GDBN} also consults first if inferior
3320 specific thread debugging library loading is enabled
3321 by @samp{set auto-load libthread-db} (@pxref{libthread_db.so.1 file}).
3323 A special entry @samp{$sdir} for @samp{libthread-db-search-path}
3324 refers to the default system directories that are
3325 normally searched for loading shared libraries. The @samp{$sdir} entry
3326 is the only kind not needing to be enabled by @samp{set auto-load libthread-db}
3327 (@pxref{libthread_db.so.1 file}).
3329 A special entry @samp{$pdir} for @samp{libthread-db-search-path}
3330 refers to the directory from which @code{libpthread}
3331 was loaded in the inferior process.
3333 For any @code{libthread_db} library @value{GDBN} finds in above directories,
3334 @value{GDBN} attempts to initialize it with the current inferior process.
3335 If this initialization fails (which could happen because of a version
3336 mismatch between @code{libthread_db} and @code{libpthread}), @value{GDBN}
3337 will unload @code{libthread_db}, and continue with the next directory.
3338 If none of @code{libthread_db} libraries initialize successfully,
3339 @value{GDBN} will issue a warning and thread debugging will be disabled.
3341 Setting @code{libthread-db-search-path} is currently implemented
3342 only on some platforms.
3344 @kindex show libthread-db-search-path
3345 @item show libthread-db-search-path
3346 Display current libthread_db search path.
3348 @kindex set debug libthread-db
3349 @kindex show debug libthread-db
3350 @cindex debugging @code{libthread_db}
3351 @item set debug libthread-db
3352 @itemx show debug libthread-db
3353 Turns on or off display of @code{libthread_db}-related events.
3354 Use @code{1} to enable, @code{0} to disable.
3358 @section Debugging Forks
3360 @cindex fork, debugging programs which call
3361 @cindex multiple processes
3362 @cindex processes, multiple
3363 On most systems, @value{GDBN} has no special support for debugging
3364 programs which create additional processes using the @code{fork}
3365 function. When a program forks, @value{GDBN} will continue to debug the
3366 parent process and the child process will run unimpeded. If you have
3367 set a breakpoint in any code which the child then executes, the child
3368 will get a @code{SIGTRAP} signal which (unless it catches the signal)
3369 will cause it to terminate.
3371 However, if you want to debug the child process there is a workaround
3372 which isn't too painful. Put a call to @code{sleep} in the code which
3373 the child process executes after the fork. It may be useful to sleep
3374 only if a certain environment variable is set, or a certain file exists,
3375 so that the delay need not occur when you don't want to run @value{GDBN}
3376 on the child. While the child is sleeping, use the @code{ps} program to
3377 get its process ID. Then tell @value{GDBN} (a new invocation of
3378 @value{GDBN} if you are also debugging the parent process) to attach to
3379 the child process (@pxref{Attach}). From that point on you can debug
3380 the child process just like any other process which you attached to.
3382 On some systems, @value{GDBN} provides support for debugging programs
3383 that create additional processes using the @code{fork} or @code{vfork}
3384 functions. On @sc{gnu}/Linux platforms, this feature is supported
3385 with kernel version 2.5.46 and later.
3387 The fork debugging commands are supported in native mode and when
3388 connected to @code{gdbserver} in either @code{target remote} mode or
3389 @code{target extended-remote} mode.
3391 By default, when a program forks, @value{GDBN} will continue to debug
3392 the parent process and the child process will run unimpeded.
3394 If you want to follow the child process instead of the parent process,
3395 use the command @w{@code{set follow-fork-mode}}.
3398 @kindex set follow-fork-mode
3399 @item set follow-fork-mode @var{mode}
3400 Set the debugger response to a program call of @code{fork} or
3401 @code{vfork}. A call to @code{fork} or @code{vfork} creates a new
3402 process. The @var{mode} argument can be:
3406 The original process is debugged after a fork. The child process runs
3407 unimpeded. This is the default.
3410 The new process is debugged after a fork. The parent process runs
3415 @kindex show follow-fork-mode
3416 @item show follow-fork-mode
3417 Display the current debugger response to a @code{fork} or @code{vfork} call.
3420 @cindex debugging multiple processes
3421 On Linux, if you want to debug both the parent and child processes, use the
3422 command @w{@code{set detach-on-fork}}.
3425 @kindex set detach-on-fork
3426 @item set detach-on-fork @var{mode}
3427 Tells gdb whether to detach one of the processes after a fork, or
3428 retain debugger control over them both.
3432 The child process (or parent process, depending on the value of
3433 @code{follow-fork-mode}) will be detached and allowed to run
3434 independently. This is the default.
3437 Both processes will be held under the control of @value{GDBN}.
3438 One process (child or parent, depending on the value of
3439 @code{follow-fork-mode}) is debugged as usual, while the other
3444 @kindex show detach-on-fork
3445 @item show detach-on-fork
3446 Show whether detach-on-fork mode is on/off.
3449 If you choose to set @samp{detach-on-fork} mode off, then @value{GDBN}
3450 will retain control of all forked processes (including nested forks).
3451 You can list the forked processes under the control of @value{GDBN} by
3452 using the @w{@code{info inferiors}} command, and switch from one fork
3453 to another by using the @code{inferior} command (@pxref{Inferiors and
3454 Programs, ,Debugging Multiple Inferiors and Programs}).
3456 To quit debugging one of the forked processes, you can either detach
3457 from it by using the @w{@code{detach inferiors}} command (allowing it
3458 to run independently), or kill it using the @w{@code{kill inferiors}}
3459 command. @xref{Inferiors and Programs, ,Debugging Multiple Inferiors
3462 If you ask to debug a child process and a @code{vfork} is followed by an
3463 @code{exec}, @value{GDBN} executes the new target up to the first
3464 breakpoint in the new target. If you have a breakpoint set on
3465 @code{main} in your original program, the breakpoint will also be set on
3466 the child process's @code{main}.
3468 On some systems, when a child process is spawned by @code{vfork}, you
3469 cannot debug the child or parent until an @code{exec} call completes.
3471 If you issue a @code{run} command to @value{GDBN} after an @code{exec}
3472 call executes, the new target restarts. To restart the parent
3473 process, use the @code{file} command with the parent executable name
3474 as its argument. By default, after an @code{exec} call executes,
3475 @value{GDBN} discards the symbols of the previous executable image.
3476 You can change this behaviour with the @w{@code{set follow-exec-mode}}
3480 @kindex set follow-exec-mode
3481 @item set follow-exec-mode @var{mode}
3483 Set debugger response to a program call of @code{exec}. An
3484 @code{exec} call replaces the program image of a process.
3486 @code{follow-exec-mode} can be:
3490 @value{GDBN} creates a new inferior and rebinds the process to this
3491 new inferior. The program the process was running before the
3492 @code{exec} call can be restarted afterwards by restarting the
3498 (@value{GDBP}) info inferiors
3500 Id Description Executable
3503 process 12020 is executing new program: prog2
3504 Program exited normally.
3505 (@value{GDBP}) info inferiors
3506 Id Description Executable
3512 @value{GDBN} keeps the process bound to the same inferior. The new
3513 executable image replaces the previous executable loaded in the
3514 inferior. Restarting the inferior after the @code{exec} call, with
3515 e.g., the @code{run} command, restarts the executable the process was
3516 running after the @code{exec} call. This is the default mode.
3521 (@value{GDBP}) info inferiors
3522 Id Description Executable
3525 process 12020 is executing new program: prog2
3526 Program exited normally.
3527 (@value{GDBP}) info inferiors
3528 Id Description Executable
3535 @code{follow-exec-mode} is supported in native mode and
3536 @code{target extended-remote} mode.
3538 You can use the @code{catch} command to make @value{GDBN} stop whenever
3539 a @code{fork}, @code{vfork}, or @code{exec} call is made. @xref{Set
3540 Catchpoints, ,Setting Catchpoints}.
3542 @node Checkpoint/Restart
3543 @section Setting a @emph{Bookmark} to Return to Later
3548 @cindex snapshot of a process
3549 @cindex rewind program state
3551 On certain operating systems@footnote{Currently, only
3552 @sc{gnu}/Linux.}, @value{GDBN} is able to save a @dfn{snapshot} of a
3553 program's state, called a @dfn{checkpoint}, and come back to it
3556 Returning to a checkpoint effectively undoes everything that has
3557 happened in the program since the @code{checkpoint} was saved. This
3558 includes changes in memory, registers, and even (within some limits)
3559 system state. Effectively, it is like going back in time to the
3560 moment when the checkpoint was saved.
3562 Thus, if you're stepping thru a program and you think you're
3563 getting close to the point where things go wrong, you can save
3564 a checkpoint. Then, if you accidentally go too far and miss
3565 the critical statement, instead of having to restart your program
3566 from the beginning, you can just go back to the checkpoint and
3567 start again from there.
3569 This can be especially useful if it takes a lot of time or
3570 steps to reach the point where you think the bug occurs.
3572 To use the @code{checkpoint}/@code{restart} method of debugging:
3577 Save a snapshot of the debugged program's current execution state.
3578 The @code{checkpoint} command takes no arguments, but each checkpoint
3579 is assigned a small integer id, similar to a breakpoint id.
3581 @kindex info checkpoints
3582 @item info checkpoints
3583 List the checkpoints that have been saved in the current debugging
3584 session. For each checkpoint, the following information will be
3591 @item Source line, or label
3594 @kindex restart @var{checkpoint-id}
3595 @item restart @var{checkpoint-id}
3596 Restore the program state that was saved as checkpoint number
3597 @var{checkpoint-id}. All program variables, registers, stack frames
3598 etc.@: will be returned to the values that they had when the checkpoint
3599 was saved. In essence, gdb will ``wind back the clock'' to the point
3600 in time when the checkpoint was saved.
3602 Note that breakpoints, @value{GDBN} variables, command history etc.
3603 are not affected by restoring a checkpoint. In general, a checkpoint
3604 only restores things that reside in the program being debugged, not in
3607 @kindex delete checkpoint @var{checkpoint-id}
3608 @item delete checkpoint @var{checkpoint-id}
3609 Delete the previously-saved checkpoint identified by @var{checkpoint-id}.
3613 Returning to a previously saved checkpoint will restore the user state
3614 of the program being debugged, plus a significant subset of the system
3615 (OS) state, including file pointers. It won't ``un-write'' data from
3616 a file, but it will rewind the file pointer to the previous location,
3617 so that the previously written data can be overwritten. For files
3618 opened in read mode, the pointer will also be restored so that the
3619 previously read data can be read again.
3621 Of course, characters that have been sent to a printer (or other
3622 external device) cannot be ``snatched back'', and characters received
3623 from eg.@: a serial device can be removed from internal program buffers,
3624 but they cannot be ``pushed back'' into the serial pipeline, ready to
3625 be received again. Similarly, the actual contents of files that have
3626 been changed cannot be restored (at this time).
3628 However, within those constraints, you actually can ``rewind'' your
3629 program to a previously saved point in time, and begin debugging it
3630 again --- and you can change the course of events so as to debug a
3631 different execution path this time.
3633 @cindex checkpoints and process id
3634 Finally, there is one bit of internal program state that will be
3635 different when you return to a checkpoint --- the program's process
3636 id. Each checkpoint will have a unique process id (or @var{pid}),
3637 and each will be different from the program's original @var{pid}.
3638 If your program has saved a local copy of its process id, this could
3639 potentially pose a problem.
3641 @subsection A Non-obvious Benefit of Using Checkpoints
3643 On some systems such as @sc{gnu}/Linux, address space randomization
3644 is performed on new processes for security reasons. This makes it
3645 difficult or impossible to set a breakpoint, or watchpoint, on an
3646 absolute address if you have to restart the program, since the
3647 absolute location of a symbol will change from one execution to the
3650 A checkpoint, however, is an @emph{identical} copy of a process.
3651 Therefore if you create a checkpoint at (eg.@:) the start of main,
3652 and simply return to that checkpoint instead of restarting the
3653 process, you can avoid the effects of address randomization and
3654 your symbols will all stay in the same place.
3657 @chapter Stopping and Continuing
3659 The principal purposes of using a debugger are so that you can stop your
3660 program before it terminates; or so that, if your program runs into
3661 trouble, you can investigate and find out why.
3663 Inside @value{GDBN}, your program may stop for any of several reasons,
3664 such as a signal, a breakpoint, or reaching a new line after a
3665 @value{GDBN} command such as @code{step}. You may then examine and
3666 change variables, set new breakpoints or remove old ones, and then
3667 continue execution. Usually, the messages shown by @value{GDBN} provide
3668 ample explanation of the status of your program---but you can also
3669 explicitly request this information at any time.
3672 @kindex info program
3674 Display information about the status of your program: whether it is
3675 running or not, what process it is, and why it stopped.
3679 * Breakpoints:: Breakpoints, watchpoints, and catchpoints
3680 * Continuing and Stepping:: Resuming execution
3681 * Skipping Over Functions and Files::
3682 Skipping over functions and files
3684 * Thread Stops:: Stopping and starting multi-thread programs
3688 @section Breakpoints, Watchpoints, and Catchpoints
3691 A @dfn{breakpoint} makes your program stop whenever a certain point in
3692 the program is reached. For each breakpoint, you can add conditions to
3693 control in finer detail whether your program stops. You can set
3694 breakpoints with the @code{break} command and its variants (@pxref{Set
3695 Breaks, ,Setting Breakpoints}), to specify the place where your program
3696 should stop by line number, function name or exact address in the
3699 On some systems, you can set breakpoints in shared libraries before
3700 the executable is run.
3703 @cindex data breakpoints
3704 @cindex memory tracing
3705 @cindex breakpoint on memory address
3706 @cindex breakpoint on variable modification
3707 A @dfn{watchpoint} is a special breakpoint that stops your program
3708 when the value of an expression changes. The expression may be a value
3709 of a variable, or it could involve values of one or more variables
3710 combined by operators, such as @samp{a + b}. This is sometimes called
3711 @dfn{data breakpoints}. You must use a different command to set
3712 watchpoints (@pxref{Set Watchpoints, ,Setting Watchpoints}), but aside
3713 from that, you can manage a watchpoint like any other breakpoint: you
3714 enable, disable, and delete both breakpoints and watchpoints using the
3717 You can arrange to have values from your program displayed automatically
3718 whenever @value{GDBN} stops at a breakpoint. @xref{Auto Display,,
3722 @cindex breakpoint on events
3723 A @dfn{catchpoint} is another special breakpoint that stops your program
3724 when a certain kind of event occurs, such as the throwing of a C@t{++}
3725 exception or the loading of a library. As with watchpoints, you use a
3726 different command to set a catchpoint (@pxref{Set Catchpoints, ,Setting
3727 Catchpoints}), but aside from that, you can manage a catchpoint like any
3728 other breakpoint. (To stop when your program receives a signal, use the
3729 @code{handle} command; see @ref{Signals, ,Signals}.)
3731 @cindex breakpoint numbers
3732 @cindex numbers for breakpoints
3733 @value{GDBN} assigns a number to each breakpoint, watchpoint, or
3734 catchpoint when you create it; these numbers are successive integers
3735 starting with one. In many of the commands for controlling various
3736 features of breakpoints you use the breakpoint number to say which
3737 breakpoint you want to change. Each breakpoint may be @dfn{enabled} or
3738 @dfn{disabled}; if disabled, it has no effect on your program until you
3741 @cindex breakpoint ranges
3742 @cindex breakpoint lists
3743 @cindex ranges of breakpoints
3744 @cindex lists of breakpoints
3745 Some @value{GDBN} commands accept a space-separated list of breakpoints
3746 on which to operate. A list element can be either a single breakpoint number,
3747 like @samp{5}, or a range of such numbers, like @samp{5-7}.
3748 When a breakpoint list is given to a command, all breakpoints in that list
3752 * Set Breaks:: Setting breakpoints
3753 * Set Watchpoints:: Setting watchpoints
3754 * Set Catchpoints:: Setting catchpoints
3755 * Delete Breaks:: Deleting breakpoints
3756 * Disabling:: Disabling breakpoints
3757 * Conditions:: Break conditions
3758 * Break Commands:: Breakpoint command lists
3759 * Dynamic Printf:: Dynamic printf
3760 * Save Breakpoints:: How to save breakpoints in a file
3761 * Static Probe Points:: Listing static probe points
3762 * Error in Breakpoints:: ``Cannot insert breakpoints''
3763 * Breakpoint-related Warnings:: ``Breakpoint address adjusted...''
3767 @subsection Setting Breakpoints
3769 @c FIXME LMB what does GDB do if no code on line of breakpt?
3770 @c consider in particular declaration with/without initialization.
3772 @c FIXME 2 is there stuff on this already? break at fun start, already init?
3775 @kindex b @r{(@code{break})}
3776 @vindex $bpnum@r{, convenience variable}
3777 @cindex latest breakpoint
3778 Breakpoints are set with the @code{break} command (abbreviated
3779 @code{b}). The debugger convenience variable @samp{$bpnum} records the
3780 number of the breakpoint you've set most recently; see @ref{Convenience
3781 Vars,, Convenience Variables}, for a discussion of what you can do with
3782 convenience variables.
3785 @item break @var{location}
3786 Set a breakpoint at the given @var{location}, which can specify a
3787 function name, a line number, or an address of an instruction.
3788 (@xref{Specify Location}, for a list of all the possible ways to
3789 specify a @var{location}.) The breakpoint will stop your program just
3790 before it executes any of the code in the specified @var{location}.
3792 When using source languages that permit overloading of symbols, such as
3793 C@t{++}, a function name may refer to more than one possible place to break.
3794 @xref{Ambiguous Expressions,,Ambiguous Expressions}, for a discussion of
3797 It is also possible to insert a breakpoint that will stop the program
3798 only if a specific thread (@pxref{Thread-Specific Breakpoints})
3799 or a specific task (@pxref{Ada Tasks}) hits that breakpoint.
3802 When called without any arguments, @code{break} sets a breakpoint at
3803 the next instruction to be executed in the selected stack frame
3804 (@pxref{Stack, ,Examining the Stack}). In any selected frame but the
3805 innermost, this makes your program stop as soon as control
3806 returns to that frame. This is similar to the effect of a
3807 @code{finish} command in the frame inside the selected frame---except
3808 that @code{finish} does not leave an active breakpoint. If you use
3809 @code{break} without an argument in the innermost frame, @value{GDBN} stops
3810 the next time it reaches the current location; this may be useful
3813 @value{GDBN} normally ignores breakpoints when it resumes execution, until at
3814 least one instruction has been executed. If it did not do this, you
3815 would be unable to proceed past a breakpoint without first disabling the
3816 breakpoint. This rule applies whether or not the breakpoint already
3817 existed when your program stopped.
3819 @item break @dots{} if @var{cond}
3820 Set a breakpoint with condition @var{cond}; evaluate the expression
3821 @var{cond} each time the breakpoint is reached, and stop only if the
3822 value is nonzero---that is, if @var{cond} evaluates as true.
3823 @samp{@dots{}} stands for one of the possible arguments described
3824 above (or no argument) specifying where to break. @xref{Conditions,
3825 ,Break Conditions}, for more information on breakpoint conditions.
3828 @item tbreak @var{args}
3829 Set a breakpoint enabled only for one stop. The @var{args} are the
3830 same as for the @code{break} command, and the breakpoint is set in the same
3831 way, but the breakpoint is automatically deleted after the first time your
3832 program stops there. @xref{Disabling, ,Disabling Breakpoints}.
3835 @cindex hardware breakpoints
3836 @item hbreak @var{args}
3837 Set a hardware-assisted breakpoint. The @var{args} are the same as for the
3838 @code{break} command and the breakpoint is set in the same way, but the
3839 breakpoint requires hardware support and some target hardware may not
3840 have this support. The main purpose of this is EPROM/ROM code
3841 debugging, so you can set a breakpoint at an instruction without
3842 changing the instruction. This can be used with the new trap-generation
3843 provided by SPARClite DSU and most x86-based targets. These targets
3844 will generate traps when a program accesses some data or instruction
3845 address that is assigned to the debug registers. However the hardware
3846 breakpoint registers can take a limited number of breakpoints. For
3847 example, on the DSU, only two data breakpoints can be set at a time, and
3848 @value{GDBN} will reject this command if more than two are used. Delete
3849 or disable unused hardware breakpoints before setting new ones
3850 (@pxref{Disabling, ,Disabling Breakpoints}).
3851 @xref{Conditions, ,Break Conditions}.
3852 For remote targets, you can restrict the number of hardware
3853 breakpoints @value{GDBN} will use, see @ref{set remote
3854 hardware-breakpoint-limit}.
3857 @item thbreak @var{args}
3858 Set a hardware-assisted breakpoint enabled only for one stop. The @var{args}
3859 are the same as for the @code{hbreak} command and the breakpoint is set in
3860 the same way. However, like the @code{tbreak} command,
3861 the breakpoint is automatically deleted after the
3862 first time your program stops there. Also, like the @code{hbreak}
3863 command, the breakpoint requires hardware support and some target hardware
3864 may not have this support. @xref{Disabling, ,Disabling Breakpoints}.
3865 See also @ref{Conditions, ,Break Conditions}.
3868 @cindex regular expression
3869 @cindex breakpoints at functions matching a regexp
3870 @cindex set breakpoints in many functions
3871 @item rbreak @var{regex}
3872 Set breakpoints on all functions matching the regular expression
3873 @var{regex}. This command sets an unconditional breakpoint on all
3874 matches, printing a list of all breakpoints it set. Once these
3875 breakpoints are set, they are treated just like the breakpoints set with
3876 the @code{break} command. You can delete them, disable them, or make
3877 them conditional the same way as any other breakpoint.
3879 In programs using different languages, @value{GDBN} chooses the syntax
3880 to print the list of all breakpoints it sets according to the
3881 @samp{set language} value: using @samp{set language auto}
3882 (see @ref{Automatically, ,Set Language Automatically}) means to use the
3883 language of the breakpoint's function, other values mean to use
3884 the manually specified language (see @ref{Manually, ,Set Language Manually}).
3886 The syntax of the regular expression is the standard one used with tools
3887 like @file{grep}. Note that this is different from the syntax used by
3888 shells, so for instance @code{foo*} matches all functions that include
3889 an @code{fo} followed by zero or more @code{o}s. There is an implicit
3890 @code{.*} leading and trailing the regular expression you supply, so to
3891 match only functions that begin with @code{foo}, use @code{^foo}.
3893 @cindex non-member C@t{++} functions, set breakpoint in
3894 When debugging C@t{++} programs, @code{rbreak} is useful for setting
3895 breakpoints on overloaded functions that are not members of any special
3898 @cindex set breakpoints on all functions
3899 The @code{rbreak} command can be used to set breakpoints in
3900 @strong{all} the functions in a program, like this:
3903 (@value{GDBP}) rbreak .
3906 @item rbreak @var{file}:@var{regex}
3907 If @code{rbreak} is called with a filename qualification, it limits
3908 the search for functions matching the given regular expression to the
3909 specified @var{file}. This can be used, for example, to set breakpoints on
3910 every function in a given file:
3913 (@value{GDBP}) rbreak file.c:.
3916 The colon separating the filename qualifier from the regex may
3917 optionally be surrounded by spaces.
3919 @kindex info breakpoints
3920 @cindex @code{$_} and @code{info breakpoints}
3921 @item info breakpoints @r{[}@var{list}@dots{}@r{]}
3922 @itemx info break @r{[}@var{list}@dots{}@r{]}
3923 Print a table of all breakpoints, watchpoints, and catchpoints set and
3924 not deleted. Optional argument @var{n} means print information only
3925 about the specified breakpoint(s) (or watchpoint(s) or catchpoint(s)).
3926 For each breakpoint, following columns are printed:
3929 @item Breakpoint Numbers
3931 Breakpoint, watchpoint, or catchpoint.
3933 Whether the breakpoint is marked to be disabled or deleted when hit.
3934 @item Enabled or Disabled
3935 Enabled breakpoints are marked with @samp{y}. @samp{n} marks breakpoints
3936 that are not enabled.
3938 Where the breakpoint is in your program, as a memory address. For a
3939 pending breakpoint whose address is not yet known, this field will
3940 contain @samp{<PENDING>}. Such breakpoint won't fire until a shared
3941 library that has the symbol or line referred by breakpoint is loaded.
3942 See below for details. A breakpoint with several locations will
3943 have @samp{<MULTIPLE>} in this field---see below for details.
3945 Where the breakpoint is in the source for your program, as a file and
3946 line number. For a pending breakpoint, the original string passed to
3947 the breakpoint command will be listed as it cannot be resolved until
3948 the appropriate shared library is loaded in the future.
3952 If a breakpoint is conditional, there are two evaluation modes: ``host'' and
3953 ``target''. If mode is ``host'', breakpoint condition evaluation is done by
3954 @value{GDBN} on the host's side. If it is ``target'', then the condition
3955 is evaluated by the target. The @code{info break} command shows
3956 the condition on the line following the affected breakpoint, together with
3957 its condition evaluation mode in between parentheses.
3959 Breakpoint commands, if any, are listed after that. A pending breakpoint is
3960 allowed to have a condition specified for it. The condition is not parsed for
3961 validity until a shared library is loaded that allows the pending
3962 breakpoint to resolve to a valid location.
3965 @code{info break} with a breakpoint
3966 number @var{n} as argument lists only that breakpoint. The
3967 convenience variable @code{$_} and the default examining-address for
3968 the @code{x} command are set to the address of the last breakpoint
3969 listed (@pxref{Memory, ,Examining Memory}).
3972 @code{info break} displays a count of the number of times the breakpoint
3973 has been hit. This is especially useful in conjunction with the
3974 @code{ignore} command. You can ignore a large number of breakpoint
3975 hits, look at the breakpoint info to see how many times the breakpoint
3976 was hit, and then run again, ignoring one less than that number. This
3977 will get you quickly to the last hit of that breakpoint.
3980 For a breakpoints with an enable count (xref) greater than 1,
3981 @code{info break} also displays that count.
3985 @value{GDBN} allows you to set any number of breakpoints at the same place in
3986 your program. There is nothing silly or meaningless about this. When
3987 the breakpoints are conditional, this is even useful
3988 (@pxref{Conditions, ,Break Conditions}).
3990 @cindex multiple locations, breakpoints
3991 @cindex breakpoints, multiple locations
3992 It is possible that a breakpoint corresponds to several locations
3993 in your program. Examples of this situation are:
3997 Multiple functions in the program may have the same name.
4000 For a C@t{++} constructor, the @value{NGCC} compiler generates several
4001 instances of the function body, used in different cases.
4004 For a C@t{++} template function, a given line in the function can
4005 correspond to any number of instantiations.
4008 For an inlined function, a given source line can correspond to
4009 several places where that function is inlined.
4012 In all those cases, @value{GDBN} will insert a breakpoint at all
4013 the relevant locations.
4015 A breakpoint with multiple locations is displayed in the breakpoint
4016 table using several rows---one header row, followed by one row for
4017 each breakpoint location. The header row has @samp{<MULTIPLE>} in the
4018 address column. The rows for individual locations contain the actual
4019 addresses for locations, and show the functions to which those
4020 locations belong. The number column for a location is of the form
4021 @var{breakpoint-number}.@var{location-number}.
4026 Num Type Disp Enb Address What
4027 1 breakpoint keep y <MULTIPLE>
4029 breakpoint already hit 1 time
4030 1.1 y 0x080486a2 in void foo<int>() at t.cc:8
4031 1.2 y 0x080486ca in void foo<double>() at t.cc:8
4034 You cannot delete the individual locations from a breakpoint. However,
4035 each location can be individually enabled or disabled by passing
4036 @var{breakpoint-number}.@var{location-number} as argument to the
4037 @code{enable} and @code{disable} commands. It's also possible to
4038 @code{enable} and @code{disable} a range of @var{location-number}
4039 locations using a @var{breakpoint-number} and two @var{location-number}s,
4040 in increasing order, separated by a hyphen, like
4041 @kbd{@var{breakpoint-number}.@var{location-number1}-@var{location-number2}},
4042 in which case @value{GDBN} acts on all the locations in the range (inclusive).
4043 Disabling or enabling the parent breakpoint (@pxref{Disabling}) affects
4044 all of the locations that belong to that breakpoint.
4046 @cindex pending breakpoints
4047 It's quite common to have a breakpoint inside a shared library.
4048 Shared libraries can be loaded and unloaded explicitly,
4049 and possibly repeatedly, as the program is executed. To support
4050 this use case, @value{GDBN} updates breakpoint locations whenever
4051 any shared library is loaded or unloaded. Typically, you would
4052 set a breakpoint in a shared library at the beginning of your
4053 debugging session, when the library is not loaded, and when the
4054 symbols from the library are not available. When you try to set
4055 breakpoint, @value{GDBN} will ask you if you want to set
4056 a so called @dfn{pending breakpoint}---breakpoint whose address
4057 is not yet resolved.
4059 After the program is run, whenever a new shared library is loaded,
4060 @value{GDBN} reevaluates all the breakpoints. When a newly loaded
4061 shared library contains the symbol or line referred to by some
4062 pending breakpoint, that breakpoint is resolved and becomes an
4063 ordinary breakpoint. When a library is unloaded, all breakpoints
4064 that refer to its symbols or source lines become pending again.
4066 This logic works for breakpoints with multiple locations, too. For
4067 example, if you have a breakpoint in a C@t{++} template function, and
4068 a newly loaded shared library has an instantiation of that template,
4069 a new location is added to the list of locations for the breakpoint.
4071 Except for having unresolved address, pending breakpoints do not
4072 differ from regular breakpoints. You can set conditions or commands,
4073 enable and disable them and perform other breakpoint operations.
4075 @value{GDBN} provides some additional commands for controlling what
4076 happens when the @samp{break} command cannot resolve breakpoint
4077 address specification to an address:
4079 @kindex set breakpoint pending
4080 @kindex show breakpoint pending
4082 @item set breakpoint pending auto
4083 This is the default behavior. When @value{GDBN} cannot find the breakpoint
4084 location, it queries you whether a pending breakpoint should be created.
4086 @item set breakpoint pending on
4087 This indicates that an unrecognized breakpoint location should automatically
4088 result in a pending breakpoint being created.
4090 @item set breakpoint pending off
4091 This indicates that pending breakpoints are not to be created. Any
4092 unrecognized breakpoint location results in an error. This setting does
4093 not affect any pending breakpoints previously created.
4095 @item show breakpoint pending
4096 Show the current behavior setting for creating pending breakpoints.
4099 The settings above only affect the @code{break} command and its
4100 variants. Once breakpoint is set, it will be automatically updated
4101 as shared libraries are loaded and unloaded.
4103 @cindex automatic hardware breakpoints
4104 For some targets, @value{GDBN} can automatically decide if hardware or
4105 software breakpoints should be used, depending on whether the
4106 breakpoint address is read-only or read-write. This applies to
4107 breakpoints set with the @code{break} command as well as to internal
4108 breakpoints set by commands like @code{next} and @code{finish}. For
4109 breakpoints set with @code{hbreak}, @value{GDBN} will always use hardware
4112 You can control this automatic behaviour with the following commands:
4114 @kindex set breakpoint auto-hw
4115 @kindex show breakpoint auto-hw
4117 @item set breakpoint auto-hw on
4118 This is the default behavior. When @value{GDBN} sets a breakpoint, it
4119 will try to use the target memory map to decide if software or hardware
4120 breakpoint must be used.
4122 @item set breakpoint auto-hw off
4123 This indicates @value{GDBN} should not automatically select breakpoint
4124 type. If the target provides a memory map, @value{GDBN} will warn when
4125 trying to set software breakpoint at a read-only address.
4128 @value{GDBN} normally implements breakpoints by replacing the program code
4129 at the breakpoint address with a special instruction, which, when
4130 executed, given control to the debugger. By default, the program
4131 code is so modified only when the program is resumed. As soon as
4132 the program stops, @value{GDBN} restores the original instructions. This
4133 behaviour guards against leaving breakpoints inserted in the
4134 target should gdb abrubptly disconnect. However, with slow remote
4135 targets, inserting and removing breakpoint can reduce the performance.
4136 This behavior can be controlled with the following commands::
4138 @kindex set breakpoint always-inserted
4139 @kindex show breakpoint always-inserted
4141 @item set breakpoint always-inserted off
4142 All breakpoints, including newly added by the user, are inserted in
4143 the target only when the target is resumed. All breakpoints are
4144 removed from the target when it stops. This is the default mode.
4146 @item set breakpoint always-inserted on
4147 Causes all breakpoints to be inserted in the target at all times. If
4148 the user adds a new breakpoint, or changes an existing breakpoint, the
4149 breakpoints in the target are updated immediately. A breakpoint is
4150 removed from the target only when breakpoint itself is deleted.
4153 @value{GDBN} handles conditional breakpoints by evaluating these conditions
4154 when a breakpoint breaks. If the condition is true, then the process being
4155 debugged stops, otherwise the process is resumed.
4157 If the target supports evaluating conditions on its end, @value{GDBN} may
4158 download the breakpoint, together with its conditions, to it.
4160 This feature can be controlled via the following commands:
4162 @kindex set breakpoint condition-evaluation
4163 @kindex show breakpoint condition-evaluation
4165 @item set breakpoint condition-evaluation host
4166 This option commands @value{GDBN} to evaluate the breakpoint
4167 conditions on the host's side. Unconditional breakpoints are sent to
4168 the target which in turn receives the triggers and reports them back to GDB
4169 for condition evaluation. This is the standard evaluation mode.
4171 @item set breakpoint condition-evaluation target
4172 This option commands @value{GDBN} to download breakpoint conditions
4173 to the target at the moment of their insertion. The target
4174 is responsible for evaluating the conditional expression and reporting
4175 breakpoint stop events back to @value{GDBN} whenever the condition
4176 is true. Due to limitations of target-side evaluation, some conditions
4177 cannot be evaluated there, e.g., conditions that depend on local data
4178 that is only known to the host. Examples include
4179 conditional expressions involving convenience variables, complex types
4180 that cannot be handled by the agent expression parser and expressions
4181 that are too long to be sent over to the target, specially when the
4182 target is a remote system. In these cases, the conditions will be
4183 evaluated by @value{GDBN}.
4185 @item set breakpoint condition-evaluation auto
4186 This is the default mode. If the target supports evaluating breakpoint
4187 conditions on its end, @value{GDBN} will download breakpoint conditions to
4188 the target (limitations mentioned previously apply). If the target does
4189 not support breakpoint condition evaluation, then @value{GDBN} will fallback
4190 to evaluating all these conditions on the host's side.
4194 @cindex negative breakpoint numbers
4195 @cindex internal @value{GDBN} breakpoints
4196 @value{GDBN} itself sometimes sets breakpoints in your program for
4197 special purposes, such as proper handling of @code{longjmp} (in C
4198 programs). These internal breakpoints are assigned negative numbers,
4199 starting with @code{-1}; @samp{info breakpoints} does not display them.
4200 You can see these breakpoints with the @value{GDBN} maintenance command
4201 @samp{maint info breakpoints} (@pxref{maint info breakpoints}).
4204 @node Set Watchpoints
4205 @subsection Setting Watchpoints
4207 @cindex setting watchpoints
4208 You can use a watchpoint to stop execution whenever the value of an
4209 expression changes, without having to predict a particular place where
4210 this may happen. (This is sometimes called a @dfn{data breakpoint}.)
4211 The expression may be as simple as the value of a single variable, or
4212 as complex as many variables combined by operators. Examples include:
4216 A reference to the value of a single variable.
4219 An address cast to an appropriate data type. For example,
4220 @samp{*(int *)0x12345678} will watch a 4-byte region at the specified
4221 address (assuming an @code{int} occupies 4 bytes).
4224 An arbitrarily complex expression, such as @samp{a*b + c/d}. The
4225 expression can use any operators valid in the program's native
4226 language (@pxref{Languages}).
4229 You can set a watchpoint on an expression even if the expression can
4230 not be evaluated yet. For instance, you can set a watchpoint on
4231 @samp{*global_ptr} before @samp{global_ptr} is initialized.
4232 @value{GDBN} will stop when your program sets @samp{global_ptr} and
4233 the expression produces a valid value. If the expression becomes
4234 valid in some other way than changing a variable (e.g.@: if the memory
4235 pointed to by @samp{*global_ptr} becomes readable as the result of a
4236 @code{malloc} call), @value{GDBN} may not stop until the next time
4237 the expression changes.
4239 @cindex software watchpoints
4240 @cindex hardware watchpoints
4241 Depending on your system, watchpoints may be implemented in software or
4242 hardware. @value{GDBN} does software watchpointing by single-stepping your
4243 program and testing the variable's value each time, which is hundreds of
4244 times slower than normal execution. (But this may still be worth it, to
4245 catch errors where you have no clue what part of your program is the
4248 On some systems, such as most PowerPC or x86-based targets,
4249 @value{GDBN} includes support for hardware watchpoints, which do not
4250 slow down the running of your program.
4254 @item watch @r{[}-l@r{|}-location@r{]} @var{expr} @r{[}thread @var{thread-id}@r{]} @r{[}mask @var{maskvalue}@r{]}
4255 Set a watchpoint for an expression. @value{GDBN} will break when the
4256 expression @var{expr} is written into by the program and its value
4257 changes. The simplest (and the most popular) use of this command is
4258 to watch the value of a single variable:
4261 (@value{GDBP}) watch foo
4264 If the command includes a @code{@r{[}thread @var{thread-id}@r{]}}
4265 argument, @value{GDBN} breaks only when the thread identified by
4266 @var{thread-id} changes the value of @var{expr}. If any other threads
4267 change the value of @var{expr}, @value{GDBN} will not break. Note
4268 that watchpoints restricted to a single thread in this way only work
4269 with Hardware Watchpoints.
4271 Ordinarily a watchpoint respects the scope of variables in @var{expr}
4272 (see below). The @code{-location} argument tells @value{GDBN} to
4273 instead watch the memory referred to by @var{expr}. In this case,
4274 @value{GDBN} will evaluate @var{expr}, take the address of the result,
4275 and watch the memory at that address. The type of the result is used
4276 to determine the size of the watched memory. If the expression's
4277 result does not have an address, then @value{GDBN} will print an
4280 The @code{@r{[}mask @var{maskvalue}@r{]}} argument allows creation
4281 of masked watchpoints, if the current architecture supports this
4282 feature (e.g., PowerPC Embedded architecture, see @ref{PowerPC
4283 Embedded}.) A @dfn{masked watchpoint} specifies a mask in addition
4284 to an address to watch. The mask specifies that some bits of an address
4285 (the bits which are reset in the mask) should be ignored when matching
4286 the address accessed by the inferior against the watchpoint address.
4287 Thus, a masked watchpoint watches many addresses simultaneously---those
4288 addresses whose unmasked bits are identical to the unmasked bits in the
4289 watchpoint address. The @code{mask} argument implies @code{-location}.
4293 (@value{GDBP}) watch foo mask 0xffff00ff
4294 (@value{GDBP}) watch *0xdeadbeef mask 0xffffff00
4298 @item rwatch @r{[}-l@r{|}-location@r{]} @var{expr} @r{[}thread @var{thread-id}@r{]} @r{[}mask @var{maskvalue}@r{]}
4299 Set a watchpoint that will break when the value of @var{expr} is read
4303 @item awatch @r{[}-l@r{|}-location@r{]} @var{expr} @r{[}thread @var{thread-id}@r{]} @r{[}mask @var{maskvalue}@r{]}
4304 Set a watchpoint that will break when @var{expr} is either read from
4305 or written into by the program.
4307 @kindex info watchpoints @r{[}@var{list}@dots{}@r{]}
4308 @item info watchpoints @r{[}@var{list}@dots{}@r{]}
4309 This command prints a list of watchpoints, using the same format as
4310 @code{info break} (@pxref{Set Breaks}).
4313 If you watch for a change in a numerically entered address you need to
4314 dereference it, as the address itself is just a constant number which will
4315 never change. @value{GDBN} refuses to create a watchpoint that watches
4316 a never-changing value:
4319 (@value{GDBP}) watch 0x600850
4320 Cannot watch constant value 0x600850.
4321 (@value{GDBP}) watch *(int *) 0x600850
4322 Watchpoint 1: *(int *) 6293584
4325 @value{GDBN} sets a @dfn{hardware watchpoint} if possible. Hardware
4326 watchpoints execute very quickly, and the debugger reports a change in
4327 value at the exact instruction where the change occurs. If @value{GDBN}
4328 cannot set a hardware watchpoint, it sets a software watchpoint, which
4329 executes more slowly and reports the change in value at the next
4330 @emph{statement}, not the instruction, after the change occurs.
4332 @cindex use only software watchpoints
4333 You can force @value{GDBN} to use only software watchpoints with the
4334 @kbd{set can-use-hw-watchpoints 0} command. With this variable set to
4335 zero, @value{GDBN} will never try to use hardware watchpoints, even if
4336 the underlying system supports them. (Note that hardware-assisted
4337 watchpoints that were set @emph{before} setting
4338 @code{can-use-hw-watchpoints} to zero will still use the hardware
4339 mechanism of watching expression values.)
4342 @item set can-use-hw-watchpoints
4343 @kindex set can-use-hw-watchpoints
4344 Set whether or not to use hardware watchpoints.
4346 @item show can-use-hw-watchpoints
4347 @kindex show can-use-hw-watchpoints
4348 Show the current mode of using hardware watchpoints.
4351 For remote targets, you can restrict the number of hardware
4352 watchpoints @value{GDBN} will use, see @ref{set remote
4353 hardware-breakpoint-limit}.
4355 When you issue the @code{watch} command, @value{GDBN} reports
4358 Hardware watchpoint @var{num}: @var{expr}
4362 if it was able to set a hardware watchpoint.
4364 Currently, the @code{awatch} and @code{rwatch} commands can only set
4365 hardware watchpoints, because accesses to data that don't change the
4366 value of the watched expression cannot be detected without examining
4367 every instruction as it is being executed, and @value{GDBN} does not do
4368 that currently. If @value{GDBN} finds that it is unable to set a
4369 hardware breakpoint with the @code{awatch} or @code{rwatch} command, it
4370 will print a message like this:
4373 Expression cannot be implemented with read/access watchpoint.
4376 Sometimes, @value{GDBN} cannot set a hardware watchpoint because the
4377 data type of the watched expression is wider than what a hardware
4378 watchpoint on the target machine can handle. For example, some systems
4379 can only watch regions that are up to 4 bytes wide; on such systems you
4380 cannot set hardware watchpoints for an expression that yields a
4381 double-precision floating-point number (which is typically 8 bytes
4382 wide). As a work-around, it might be possible to break the large region
4383 into a series of smaller ones and watch them with separate watchpoints.
4385 If you set too many hardware watchpoints, @value{GDBN} might be unable
4386 to insert all of them when you resume the execution of your program.
4387 Since the precise number of active watchpoints is unknown until such
4388 time as the program is about to be resumed, @value{GDBN} might not be
4389 able to warn you about this when you set the watchpoints, and the
4390 warning will be printed only when the program is resumed:
4393 Hardware watchpoint @var{num}: Could not insert watchpoint
4397 If this happens, delete or disable some of the watchpoints.
4399 Watching complex expressions that reference many variables can also
4400 exhaust the resources available for hardware-assisted watchpoints.
4401 That's because @value{GDBN} needs to watch every variable in the
4402 expression with separately allocated resources.
4404 If you call a function interactively using @code{print} or @code{call},
4405 any watchpoints you have set will be inactive until @value{GDBN} reaches another
4406 kind of breakpoint or the call completes.
4408 @value{GDBN} automatically deletes watchpoints that watch local
4409 (automatic) variables, or expressions that involve such variables, when
4410 they go out of scope, that is, when the execution leaves the block in
4411 which these variables were defined. In particular, when the program
4412 being debugged terminates, @emph{all} local variables go out of scope,
4413 and so only watchpoints that watch global variables remain set. If you
4414 rerun the program, you will need to set all such watchpoints again. One
4415 way of doing that would be to set a code breakpoint at the entry to the
4416 @code{main} function and when it breaks, set all the watchpoints.
4418 @cindex watchpoints and threads
4419 @cindex threads and watchpoints
4420 In multi-threaded programs, watchpoints will detect changes to the
4421 watched expression from every thread.
4424 @emph{Warning:} In multi-threaded programs, software watchpoints
4425 have only limited usefulness. If @value{GDBN} creates a software
4426 watchpoint, it can only watch the value of an expression @emph{in a
4427 single thread}. If you are confident that the expression can only
4428 change due to the current thread's activity (and if you are also
4429 confident that no other thread can become current), then you can use
4430 software watchpoints as usual. However, @value{GDBN} may not notice
4431 when a non-current thread's activity changes the expression. (Hardware
4432 watchpoints, in contrast, watch an expression in all threads.)
4435 @xref{set remote hardware-watchpoint-limit}.
4437 @node Set Catchpoints
4438 @subsection Setting Catchpoints
4439 @cindex catchpoints, setting
4440 @cindex exception handlers
4441 @cindex event handling
4443 You can use @dfn{catchpoints} to cause the debugger to stop for certain
4444 kinds of program events, such as C@t{++} exceptions or the loading of a
4445 shared library. Use the @code{catch} command to set a catchpoint.
4449 @item catch @var{event}
4450 Stop when @var{event} occurs. The @var{event} can be any of the following:
4453 @item throw @r{[}@var{regexp}@r{]}
4454 @itemx rethrow @r{[}@var{regexp}@r{]}
4455 @itemx catch @r{[}@var{regexp}@r{]}
4457 @kindex catch rethrow
4459 @cindex stop on C@t{++} exceptions
4460 The throwing, re-throwing, or catching of a C@t{++} exception.
4462 If @var{regexp} is given, then only exceptions whose type matches the
4463 regular expression will be caught.
4465 @vindex $_exception@r{, convenience variable}
4466 The convenience variable @code{$_exception} is available at an
4467 exception-related catchpoint, on some systems. This holds the
4468 exception being thrown.
4470 There are currently some limitations to C@t{++} exception handling in
4475 The support for these commands is system-dependent. Currently, only
4476 systems using the @samp{gnu-v3} C@t{++} ABI (@pxref{ABI}) are
4480 The regular expression feature and the @code{$_exception} convenience
4481 variable rely on the presence of some SDT probes in @code{libstdc++}.
4482 If these probes are not present, then these features cannot be used.
4483 These probes were first available in the GCC 4.8 release, but whether
4484 or not they are available in your GCC also depends on how it was
4488 The @code{$_exception} convenience variable is only valid at the
4489 instruction at which an exception-related catchpoint is set.
4492 When an exception-related catchpoint is hit, @value{GDBN} stops at a
4493 location in the system library which implements runtime exception
4494 support for C@t{++}, usually @code{libstdc++}. You can use @code{up}
4495 (@pxref{Selection}) to get to your code.
4498 If you call a function interactively, @value{GDBN} normally returns
4499 control to you when the function has finished executing. If the call
4500 raises an exception, however, the call may bypass the mechanism that
4501 returns control to you and cause your program either to abort or to
4502 simply continue running until it hits a breakpoint, catches a signal
4503 that @value{GDBN} is listening for, or exits. This is the case even if
4504 you set a catchpoint for the exception; catchpoints on exceptions are
4505 disabled within interactive calls. @xref{Calling}, for information on
4506 controlling this with @code{set unwind-on-terminating-exception}.
4509 You cannot raise an exception interactively.
4512 You cannot install an exception handler interactively.
4515 @item exception @r{[}@var{name}@r{]}
4516 @kindex catch exception
4517 @cindex Ada exception catching
4518 @cindex catch Ada exceptions
4519 An Ada exception being raised. If an exception name is specified
4520 at the end of the command (eg @code{catch exception Program_Error}),
4521 the debugger will stop only when this specific exception is raised.
4522 Otherwise, the debugger stops execution when any Ada exception is raised.
4524 When inserting an exception catchpoint on a user-defined exception whose
4525 name is identical to one of the exceptions defined by the language, the
4526 fully qualified name must be used as the exception name. Otherwise,
4527 @value{GDBN} will assume that it should stop on the pre-defined exception
4528 rather than the user-defined one. For instance, assuming an exception
4529 called @code{Constraint_Error} is defined in package @code{Pck}, then
4530 the command to use to catch such exceptions is @kbd{catch exception
4531 Pck.Constraint_Error}.
4533 @item exception unhandled
4534 @kindex catch exception unhandled
4535 An exception that was raised but is not handled by the program.
4537 @item handlers @r{[}@var{name}@r{]}
4538 @kindex catch handlers
4539 @cindex Ada exception handlers catching
4540 @cindex catch Ada exceptions when handled
4541 An Ada exception being handled. If an exception name is
4542 specified at the end of the command
4543 (eg @kbd{catch handlers Program_Error}), the debugger will stop
4544 only when this specific exception is handled.
4545 Otherwise, the debugger stops execution when any Ada exception is handled.
4547 When inserting a handlers catchpoint on a user-defined
4548 exception whose name is identical to one of the exceptions
4549 defined by the language, the fully qualified name must be used
4550 as the exception name. Otherwise, @value{GDBN} will assume that it
4551 should stop on the pre-defined exception rather than the
4552 user-defined one. For instance, assuming an exception called
4553 @code{Constraint_Error} is defined in package @code{Pck}, then the
4554 command to use to catch such exceptions handling is
4555 @kbd{catch handlers Pck.Constraint_Error}.
4558 @kindex catch assert
4559 A failed Ada assertion.
4563 @cindex break on fork/exec
4564 A call to @code{exec}.
4566 @anchor{catch syscall}
4568 @itemx syscall @r{[}@var{name} @r{|} @var{number} @r{|} @r{group:}@var{groupname} @r{|} @r{g:}@var{groupname}@r{]} @dots{}
4569 @kindex catch syscall
4570 @cindex break on a system call.
4571 A call to or return from a system call, a.k.a.@: @dfn{syscall}. A
4572 syscall is a mechanism for application programs to request a service
4573 from the operating system (OS) or one of the OS system services.
4574 @value{GDBN} can catch some or all of the syscalls issued by the
4575 debuggee, and show the related information for each syscall. If no
4576 argument is specified, calls to and returns from all system calls
4579 @var{name} can be any system call name that is valid for the
4580 underlying OS. Just what syscalls are valid depends on the OS. On
4581 GNU and Unix systems, you can find the full list of valid syscall
4582 names on @file{/usr/include/asm/unistd.h}.
4584 @c For MS-Windows, the syscall names and the corresponding numbers
4585 @c can be found, e.g., on this URL:
4586 @c http://www.metasploit.com/users/opcode/syscalls.html
4587 @c but we don't support Windows syscalls yet.
4589 Normally, @value{GDBN} knows in advance which syscalls are valid for
4590 each OS, so you can use the @value{GDBN} command-line completion
4591 facilities (@pxref{Completion,, command completion}) to list the
4594 You may also specify the system call numerically. A syscall's
4595 number is the value passed to the OS's syscall dispatcher to
4596 identify the requested service. When you specify the syscall by its
4597 name, @value{GDBN} uses its database of syscalls to convert the name
4598 into the corresponding numeric code, but using the number directly
4599 may be useful if @value{GDBN}'s database does not have the complete
4600 list of syscalls on your system (e.g., because @value{GDBN} lags
4601 behind the OS upgrades).
4603 You may specify a group of related syscalls to be caught at once using
4604 the @code{group:} syntax (@code{g:} is a shorter equivalent). For
4605 instance, on some platforms @value{GDBN} allows you to catch all
4606 network related syscalls, by passing the argument @code{group:network}
4607 to @code{catch syscall}. Note that not all syscall groups are
4608 available in every system. You can use the command completion
4609 facilities (@pxref{Completion,, command completion}) to list the
4610 syscall groups available on your environment.
4612 The example below illustrates how this command works if you don't provide
4616 (@value{GDBP}) catch syscall
4617 Catchpoint 1 (syscall)
4619 Starting program: /tmp/catch-syscall
4621 Catchpoint 1 (call to syscall 'close'), \
4622 0xffffe424 in __kernel_vsyscall ()
4626 Catchpoint 1 (returned from syscall 'close'), \
4627 0xffffe424 in __kernel_vsyscall ()
4631 Here is an example of catching a system call by name:
4634 (@value{GDBP}) catch syscall chroot
4635 Catchpoint 1 (syscall 'chroot' [61])
4637 Starting program: /tmp/catch-syscall
4639 Catchpoint 1 (call to syscall 'chroot'), \
4640 0xffffe424 in __kernel_vsyscall ()
4644 Catchpoint 1 (returned from syscall 'chroot'), \
4645 0xffffe424 in __kernel_vsyscall ()
4649 An example of specifying a system call numerically. In the case
4650 below, the syscall number has a corresponding entry in the XML
4651 file, so @value{GDBN} finds its name and prints it:
4654 (@value{GDBP}) catch syscall 252
4655 Catchpoint 1 (syscall(s) 'exit_group')
4657 Starting program: /tmp/catch-syscall
4659 Catchpoint 1 (call to syscall 'exit_group'), \
4660 0xffffe424 in __kernel_vsyscall ()
4664 Program exited normally.
4668 Here is an example of catching a syscall group:
4671 (@value{GDBP}) catch syscall group:process
4672 Catchpoint 1 (syscalls 'exit' [1] 'fork' [2] 'waitpid' [7]
4673 'execve' [11] 'wait4' [114] 'clone' [120] 'vfork' [190]
4674 'exit_group' [252] 'waitid' [284] 'unshare' [310])
4676 Starting program: /tmp/catch-syscall
4678 Catchpoint 1 (call to syscall fork), 0x00007ffff7df4e27 in open64 ()
4679 from /lib64/ld-linux-x86-64.so.2
4685 However, there can be situations when there is no corresponding name
4686 in XML file for that syscall number. In this case, @value{GDBN} prints
4687 a warning message saying that it was not able to find the syscall name,
4688 but the catchpoint will be set anyway. See the example below:
4691 (@value{GDBP}) catch syscall 764
4692 warning: The number '764' does not represent a known syscall.
4693 Catchpoint 2 (syscall 764)
4697 If you configure @value{GDBN} using the @samp{--without-expat} option,
4698 it will not be able to display syscall names. Also, if your
4699 architecture does not have an XML file describing its system calls,
4700 you will not be able to see the syscall names. It is important to
4701 notice that these two features are used for accessing the syscall
4702 name database. In either case, you will see a warning like this:
4705 (@value{GDBP}) catch syscall
4706 warning: Could not open "syscalls/i386-linux.xml"
4707 warning: Could not load the syscall XML file 'syscalls/i386-linux.xml'.
4708 GDB will not be able to display syscall names.
4709 Catchpoint 1 (syscall)
4713 Of course, the file name will change depending on your architecture and system.
4715 Still using the example above, you can also try to catch a syscall by its
4716 number. In this case, you would see something like:
4719 (@value{GDBP}) catch syscall 252
4720 Catchpoint 1 (syscall(s) 252)
4723 Again, in this case @value{GDBN} would not be able to display syscall's names.
4727 A call to @code{fork}.
4731 A call to @code{vfork}.
4733 @item load @r{[}@var{regexp}@r{]}
4734 @itemx unload @r{[}@var{regexp}@r{]}
4736 @kindex catch unload
4737 The loading or unloading of a shared library. If @var{regexp} is
4738 given, then the catchpoint will stop only if the regular expression
4739 matches one of the affected libraries.
4741 @item signal @r{[}@var{signal}@dots{} @r{|} @samp{all}@r{]}
4742 @kindex catch signal
4743 The delivery of a signal.
4745 With no arguments, this catchpoint will catch any signal that is not
4746 used internally by @value{GDBN}, specifically, all signals except
4747 @samp{SIGTRAP} and @samp{SIGINT}.
4749 With the argument @samp{all}, all signals, including those used by
4750 @value{GDBN}, will be caught. This argument cannot be used with other
4753 Otherwise, the arguments are a list of signal names as given to
4754 @code{handle} (@pxref{Signals}). Only signals specified in this list
4757 One reason that @code{catch signal} can be more useful than
4758 @code{handle} is that you can attach commands and conditions to the
4761 When a signal is caught by a catchpoint, the signal's @code{stop} and
4762 @code{print} settings, as specified by @code{handle}, are ignored.
4763 However, whether the signal is still delivered to the inferior depends
4764 on the @code{pass} setting; this can be changed in the catchpoint's
4769 @item tcatch @var{event}
4771 Set a catchpoint that is enabled only for one stop. The catchpoint is
4772 automatically deleted after the first time the event is caught.
4776 Use the @code{info break} command to list the current catchpoints.
4780 @subsection Deleting Breakpoints
4782 @cindex clearing breakpoints, watchpoints, catchpoints
4783 @cindex deleting breakpoints, watchpoints, catchpoints
4784 It is often necessary to eliminate a breakpoint, watchpoint, or
4785 catchpoint once it has done its job and you no longer want your program
4786 to stop there. This is called @dfn{deleting} the breakpoint. A
4787 breakpoint that has been deleted no longer exists; it is forgotten.
4789 With the @code{clear} command you can delete breakpoints according to
4790 where they are in your program. With the @code{delete} command you can
4791 delete individual breakpoints, watchpoints, or catchpoints by specifying
4792 their breakpoint numbers.
4794 It is not necessary to delete a breakpoint to proceed past it. @value{GDBN}
4795 automatically ignores breakpoints on the first instruction to be executed
4796 when you continue execution without changing the execution address.
4801 Delete any breakpoints at the next instruction to be executed in the
4802 selected stack frame (@pxref{Selection, ,Selecting a Frame}). When
4803 the innermost frame is selected, this is a good way to delete a
4804 breakpoint where your program just stopped.
4806 @item clear @var{location}
4807 Delete any breakpoints set at the specified @var{location}.
4808 @xref{Specify Location}, for the various forms of @var{location}; the
4809 most useful ones are listed below:
4812 @item clear @var{function}
4813 @itemx clear @var{filename}:@var{function}
4814 Delete any breakpoints set at entry to the named @var{function}.
4816 @item clear @var{linenum}
4817 @itemx clear @var{filename}:@var{linenum}
4818 Delete any breakpoints set at or within the code of the specified
4819 @var{linenum} of the specified @var{filename}.
4822 @cindex delete breakpoints
4824 @kindex d @r{(@code{delete})}
4825 @item delete @r{[}breakpoints@r{]} @r{[}@var{list}@dots{}@r{]}
4826 Delete the breakpoints, watchpoints, or catchpoints of the breakpoint
4827 list specified as argument. If no argument is specified, delete all
4828 breakpoints (@value{GDBN} asks confirmation, unless you have @code{set
4829 confirm off}). You can abbreviate this command as @code{d}.
4833 @subsection Disabling Breakpoints
4835 @cindex enable/disable a breakpoint
4836 Rather than deleting a breakpoint, watchpoint, or catchpoint, you might
4837 prefer to @dfn{disable} it. This makes the breakpoint inoperative as if
4838 it had been deleted, but remembers the information on the breakpoint so
4839 that you can @dfn{enable} it again later.
4841 You disable and enable breakpoints, watchpoints, and catchpoints with
4842 the @code{enable} and @code{disable} commands, optionally specifying
4843 one or more breakpoint numbers as arguments. Use @code{info break} to
4844 print a list of all breakpoints, watchpoints, and catchpoints if you
4845 do not know which numbers to use.
4847 Disabling and enabling a breakpoint that has multiple locations
4848 affects all of its locations.
4850 A breakpoint, watchpoint, or catchpoint can have any of several
4851 different states of enablement:
4855 Enabled. The breakpoint stops your program. A breakpoint set
4856 with the @code{break} command starts out in this state.
4858 Disabled. The breakpoint has no effect on your program.
4860 Enabled once. The breakpoint stops your program, but then becomes
4863 Enabled for a count. The breakpoint stops your program for the next
4864 N times, then becomes disabled.
4866 Enabled for deletion. The breakpoint stops your program, but
4867 immediately after it does so it is deleted permanently. A breakpoint
4868 set with the @code{tbreak} command starts out in this state.
4871 You can use the following commands to enable or disable breakpoints,
4872 watchpoints, and catchpoints:
4876 @kindex dis @r{(@code{disable})}
4877 @item disable @r{[}breakpoints@r{]} @r{[}@var{list}@dots{}@r{]}
4878 Disable the specified breakpoints---or all breakpoints, if none are
4879 listed. A disabled breakpoint has no effect but is not forgotten. All
4880 options such as ignore-counts, conditions and commands are remembered in
4881 case the breakpoint is enabled again later. You may abbreviate
4882 @code{disable} as @code{dis}.
4885 @item enable @r{[}breakpoints@r{]} @r{[}@var{list}@dots{}@r{]}
4886 Enable the specified breakpoints (or all defined breakpoints). They
4887 become effective once again in stopping your program.
4889 @item enable @r{[}breakpoints@r{]} once @var{list}@dots{}
4890 Enable the specified breakpoints temporarily. @value{GDBN} disables any
4891 of these breakpoints immediately after stopping your program.
4893 @item enable @r{[}breakpoints@r{]} count @var{count} @var{list}@dots{}
4894 Enable the specified breakpoints temporarily. @value{GDBN} records
4895 @var{count} with each of the specified breakpoints, and decrements a
4896 breakpoint's count when it is hit. When any count reaches 0,
4897 @value{GDBN} disables that breakpoint. If a breakpoint has an ignore
4898 count (@pxref{Conditions, ,Break Conditions}), that will be
4899 decremented to 0 before @var{count} is affected.
4901 @item enable @r{[}breakpoints@r{]} delete @var{list}@dots{}
4902 Enable the specified breakpoints to work once, then die. @value{GDBN}
4903 deletes any of these breakpoints as soon as your program stops there.
4904 Breakpoints set by the @code{tbreak} command start out in this state.
4907 @c FIXME: I think the following ``Except for [...] @code{tbreak}'' is
4908 @c confusing: tbreak is also initially enabled.
4909 Except for a breakpoint set with @code{tbreak} (@pxref{Set Breaks,
4910 ,Setting Breakpoints}), breakpoints that you set are initially enabled;
4911 subsequently, they become disabled or enabled only when you use one of
4912 the commands above. (The command @code{until} can set and delete a
4913 breakpoint of its own, but it does not change the state of your other
4914 breakpoints; see @ref{Continuing and Stepping, ,Continuing and
4918 @subsection Break Conditions
4919 @cindex conditional breakpoints
4920 @cindex breakpoint conditions
4922 @c FIXME what is scope of break condition expr? Context where wanted?
4923 @c in particular for a watchpoint?
4924 The simplest sort of breakpoint breaks every time your program reaches a
4925 specified place. You can also specify a @dfn{condition} for a
4926 breakpoint. A condition is just a Boolean expression in your
4927 programming language (@pxref{Expressions, ,Expressions}). A breakpoint with
4928 a condition evaluates the expression each time your program reaches it,
4929 and your program stops only if the condition is @emph{true}.
4931 This is the converse of using assertions for program validation; in that
4932 situation, you want to stop when the assertion is violated---that is,
4933 when the condition is false. In C, if you want to test an assertion expressed
4934 by the condition @var{assert}, you should set the condition
4935 @samp{! @var{assert}} on the appropriate breakpoint.
4937 Conditions are also accepted for watchpoints; you may not need them,
4938 since a watchpoint is inspecting the value of an expression anyhow---but
4939 it might be simpler, say, to just set a watchpoint on a variable name,
4940 and specify a condition that tests whether the new value is an interesting
4943 Break conditions can have side effects, and may even call functions in
4944 your program. This can be useful, for example, to activate functions
4945 that log program progress, or to use your own print functions to
4946 format special data structures. The effects are completely predictable
4947 unless there is another enabled breakpoint at the same address. (In
4948 that case, @value{GDBN} might see the other breakpoint first and stop your
4949 program without checking the condition of this one.) Note that
4950 breakpoint commands are usually more convenient and flexible than break
4952 purpose of performing side effects when a breakpoint is reached
4953 (@pxref{Break Commands, ,Breakpoint Command Lists}).
4955 Breakpoint conditions can also be evaluated on the target's side if
4956 the target supports it. Instead of evaluating the conditions locally,
4957 @value{GDBN} encodes the expression into an agent expression
4958 (@pxref{Agent Expressions}) suitable for execution on the target,
4959 independently of @value{GDBN}. Global variables become raw memory
4960 locations, locals become stack accesses, and so forth.
4962 In this case, @value{GDBN} will only be notified of a breakpoint trigger
4963 when its condition evaluates to true. This mechanism may provide faster
4964 response times depending on the performance characteristics of the target
4965 since it does not need to keep @value{GDBN} informed about
4966 every breakpoint trigger, even those with false conditions.
4968 Break conditions can be specified when a breakpoint is set, by using
4969 @samp{if} in the arguments to the @code{break} command. @xref{Set
4970 Breaks, ,Setting Breakpoints}. They can also be changed at any time
4971 with the @code{condition} command.
4973 You can also use the @code{if} keyword with the @code{watch} command.
4974 The @code{catch} command does not recognize the @code{if} keyword;
4975 @code{condition} is the only way to impose a further condition on a
4980 @item condition @var{bnum} @var{expression}
4981 Specify @var{expression} as the break condition for breakpoint,
4982 watchpoint, or catchpoint number @var{bnum}. After you set a condition,
4983 breakpoint @var{bnum} stops your program only if the value of
4984 @var{expression} is true (nonzero, in C). When you use
4985 @code{condition}, @value{GDBN} checks @var{expression} immediately for
4986 syntactic correctness, and to determine whether symbols in it have
4987 referents in the context of your breakpoint. If @var{expression} uses
4988 symbols not referenced in the context of the breakpoint, @value{GDBN}
4989 prints an error message:
4992 No symbol "foo" in current context.
4997 not actually evaluate @var{expression} at the time the @code{condition}
4998 command (or a command that sets a breakpoint with a condition, like
4999 @code{break if @dots{}}) is given, however. @xref{Expressions, ,Expressions}.
5001 @item condition @var{bnum}
5002 Remove the condition from breakpoint number @var{bnum}. It becomes
5003 an ordinary unconditional breakpoint.
5006 @cindex ignore count (of breakpoint)
5007 A special case of a breakpoint condition is to stop only when the
5008 breakpoint has been reached a certain number of times. This is so
5009 useful that there is a special way to do it, using the @dfn{ignore
5010 count} of the breakpoint. Every breakpoint has an ignore count, which
5011 is an integer. Most of the time, the ignore count is zero, and
5012 therefore has no effect. But if your program reaches a breakpoint whose
5013 ignore count is positive, then instead of stopping, it just decrements
5014 the ignore count by one and continues. As a result, if the ignore count
5015 value is @var{n}, the breakpoint does not stop the next @var{n} times
5016 your program reaches it.
5020 @item ignore @var{bnum} @var{count}
5021 Set the ignore count of breakpoint number @var{bnum} to @var{count}.
5022 The next @var{count} times the breakpoint is reached, your program's
5023 execution does not stop; other than to decrement the ignore count, @value{GDBN}
5026 To make the breakpoint stop the next time it is reached, specify
5029 When you use @code{continue} to resume execution of your program from a
5030 breakpoint, you can specify an ignore count directly as an argument to
5031 @code{continue}, rather than using @code{ignore}. @xref{Continuing and
5032 Stepping,,Continuing and Stepping}.
5034 If a breakpoint has a positive ignore count and a condition, the
5035 condition is not checked. Once the ignore count reaches zero,
5036 @value{GDBN} resumes checking the condition.
5038 You could achieve the effect of the ignore count with a condition such
5039 as @w{@samp{$foo-- <= 0}} using a debugger convenience variable that
5040 is decremented each time. @xref{Convenience Vars, ,Convenience
5044 Ignore counts apply to breakpoints, watchpoints, and catchpoints.
5047 @node Break Commands
5048 @subsection Breakpoint Command Lists
5050 @cindex breakpoint commands
5051 You can give any breakpoint (or watchpoint or catchpoint) a series of
5052 commands to execute when your program stops due to that breakpoint. For
5053 example, you might want to print the values of certain expressions, or
5054 enable other breakpoints.
5058 @kindex end@r{ (breakpoint commands)}
5059 @item commands @r{[}@var{list}@dots{}@r{]}
5060 @itemx @dots{} @var{command-list} @dots{}
5062 Specify a list of commands for the given breakpoints. The commands
5063 themselves appear on the following lines. Type a line containing just
5064 @code{end} to terminate the commands.
5066 To remove all commands from a breakpoint, type @code{commands} and
5067 follow it immediately with @code{end}; that is, give no commands.
5069 With no argument, @code{commands} refers to the last breakpoint,
5070 watchpoint, or catchpoint set (not to the breakpoint most recently
5071 encountered). If the most recent breakpoints were set with a single
5072 command, then the @code{commands} will apply to all the breakpoints
5073 set by that command. This applies to breakpoints set by
5074 @code{rbreak}, and also applies when a single @code{break} command
5075 creates multiple breakpoints (@pxref{Ambiguous Expressions,,Ambiguous
5079 Pressing @key{RET} as a means of repeating the last @value{GDBN} command is
5080 disabled within a @var{command-list}.
5082 You can use breakpoint commands to start your program up again. Simply
5083 use the @code{continue} command, or @code{step}, or any other command
5084 that resumes execution.
5086 Any other commands in the command list, after a command that resumes
5087 execution, are ignored. This is because any time you resume execution
5088 (even with a simple @code{next} or @code{step}), you may encounter
5089 another breakpoint---which could have its own command list, leading to
5090 ambiguities about which list to execute.
5093 If the first command you specify in a command list is @code{silent}, the
5094 usual message about stopping at a breakpoint is not printed. This may
5095 be desirable for breakpoints that are to print a specific message and
5096 then continue. If none of the remaining commands print anything, you
5097 see no sign that the breakpoint was reached. @code{silent} is
5098 meaningful only at the beginning of a breakpoint command list.
5100 The commands @code{echo}, @code{output}, and @code{printf} allow you to
5101 print precisely controlled output, and are often useful in silent
5102 breakpoints. @xref{Output, ,Commands for Controlled Output}.
5104 For example, here is how you could use breakpoint commands to print the
5105 value of @code{x} at entry to @code{foo} whenever @code{x} is positive.
5111 printf "x is %d\n",x
5116 One application for breakpoint commands is to compensate for one bug so
5117 you can test for another. Put a breakpoint just after the erroneous line
5118 of code, give it a condition to detect the case in which something
5119 erroneous has been done, and give it commands to assign correct values
5120 to any variables that need them. End with the @code{continue} command
5121 so that your program does not stop, and start with the @code{silent}
5122 command so that no output is produced. Here is an example:
5133 @node Dynamic Printf
5134 @subsection Dynamic Printf
5136 @cindex dynamic printf
5138 The dynamic printf command @code{dprintf} combines a breakpoint with
5139 formatted printing of your program's data to give you the effect of
5140 inserting @code{printf} calls into your program on-the-fly, without
5141 having to recompile it.
5143 In its most basic form, the output goes to the GDB console. However,
5144 you can set the variable @code{dprintf-style} for alternate handling.
5145 For instance, you can ask to format the output by calling your
5146 program's @code{printf} function. This has the advantage that the
5147 characters go to the program's output device, so they can recorded in
5148 redirects to files and so forth.
5150 If you are doing remote debugging with a stub or agent, you can also
5151 ask to have the printf handled by the remote agent. In addition to
5152 ensuring that the output goes to the remote program's device along
5153 with any other output the program might produce, you can also ask that
5154 the dprintf remain active even after disconnecting from the remote
5155 target. Using the stub/agent is also more efficient, as it can do
5156 everything without needing to communicate with @value{GDBN}.
5160 @item dprintf @var{location},@var{template},@var{expression}[,@var{expression}@dots{}]
5161 Whenever execution reaches @var{location}, print the values of one or
5162 more @var{expressions} under the control of the string @var{template}.
5163 To print several values, separate them with commas.
5165 @item set dprintf-style @var{style}
5166 Set the dprintf output to be handled in one of several different
5167 styles enumerated below. A change of style affects all existing
5168 dynamic printfs immediately. (If you need individual control over the
5169 print commands, simply define normal breakpoints with
5170 explicitly-supplied command lists.)
5174 @kindex dprintf-style gdb
5175 Handle the output using the @value{GDBN} @code{printf} command.
5178 @kindex dprintf-style call
5179 Handle the output by calling a function in your program (normally
5183 @kindex dprintf-style agent
5184 Have the remote debugging agent (such as @code{gdbserver}) handle
5185 the output itself. This style is only available for agents that
5186 support running commands on the target.
5189 @item set dprintf-function @var{function}
5190 Set the function to call if the dprintf style is @code{call}. By
5191 default its value is @code{printf}. You may set it to any expression.
5192 that @value{GDBN} can evaluate to a function, as per the @code{call}
5195 @item set dprintf-channel @var{channel}
5196 Set a ``channel'' for dprintf. If set to a non-empty value,
5197 @value{GDBN} will evaluate it as an expression and pass the result as
5198 a first argument to the @code{dprintf-function}, in the manner of
5199 @code{fprintf} and similar functions. Otherwise, the dprintf format
5200 string will be the first argument, in the manner of @code{printf}.
5202 As an example, if you wanted @code{dprintf} output to go to a logfile
5203 that is a standard I/O stream assigned to the variable @code{mylog},
5204 you could do the following:
5207 (gdb) set dprintf-style call
5208 (gdb) set dprintf-function fprintf
5209 (gdb) set dprintf-channel mylog
5210 (gdb) dprintf 25,"at line 25, glob=%d\n",glob
5211 Dprintf 1 at 0x123456: file main.c, line 25.
5213 1 dprintf keep y 0x00123456 in main at main.c:25
5214 call (void) fprintf (mylog,"at line 25, glob=%d\n",glob)
5219 Note that the @code{info break} displays the dynamic printf commands
5220 as normal breakpoint commands; you can thus easily see the effect of
5221 the variable settings.
5223 @item set disconnected-dprintf on
5224 @itemx set disconnected-dprintf off
5225 @kindex set disconnected-dprintf
5226 Choose whether @code{dprintf} commands should continue to run if
5227 @value{GDBN} has disconnected from the target. This only applies
5228 if the @code{dprintf-style} is @code{agent}.
5230 @item show disconnected-dprintf off
5231 @kindex show disconnected-dprintf
5232 Show the current choice for disconnected @code{dprintf}.
5236 @value{GDBN} does not check the validity of function and channel,
5237 relying on you to supply values that are meaningful for the contexts
5238 in which they are being used. For instance, the function and channel
5239 may be the values of local variables, but if that is the case, then
5240 all enabled dynamic prints must be at locations within the scope of
5241 those locals. If evaluation fails, @value{GDBN} will report an error.
5243 @node Save Breakpoints
5244 @subsection How to save breakpoints to a file
5246 To save breakpoint definitions to a file use the @w{@code{save
5247 breakpoints}} command.
5250 @kindex save breakpoints
5251 @cindex save breakpoints to a file for future sessions
5252 @item save breakpoints [@var{filename}]
5253 This command saves all current breakpoint definitions together with
5254 their commands and ignore counts, into a file @file{@var{filename}}
5255 suitable for use in a later debugging session. This includes all
5256 types of breakpoints (breakpoints, watchpoints, catchpoints,
5257 tracepoints). To read the saved breakpoint definitions, use the
5258 @code{source} command (@pxref{Command Files}). Note that watchpoints
5259 with expressions involving local variables may fail to be recreated
5260 because it may not be possible to access the context where the
5261 watchpoint is valid anymore. Because the saved breakpoint definitions
5262 are simply a sequence of @value{GDBN} commands that recreate the
5263 breakpoints, you can edit the file in your favorite editing program,
5264 and remove the breakpoint definitions you're not interested in, or
5265 that can no longer be recreated.
5268 @node Static Probe Points
5269 @subsection Static Probe Points
5271 @cindex static probe point, SystemTap
5272 @cindex static probe point, DTrace
5273 @value{GDBN} supports @dfn{SDT} probes in the code. @acronym{SDT} stands
5274 for Statically Defined Tracing, and the probes are designed to have a tiny
5275 runtime code and data footprint, and no dynamic relocations.
5277 Currently, the following types of probes are supported on
5278 ELF-compatible systems:
5282 @item @code{SystemTap} (@uref{http://sourceware.org/systemtap/})
5283 @acronym{SDT} probes@footnote{See
5284 @uref{http://sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps}
5285 for more information on how to add @code{SystemTap} @acronym{SDT}
5286 probes in your applications.}. @code{SystemTap} probes are usable
5287 from assembly, C and C@t{++} languages@footnote{See
5288 @uref{http://sourceware.org/systemtap/wiki/UserSpaceProbeImplementation}
5289 for a good reference on how the @acronym{SDT} probes are implemented.}.
5291 @item @code{DTrace} (@uref{http://oss.oracle.com/projects/DTrace})
5292 @acronym{USDT} probes. @code{DTrace} probes are usable from C and
5296 @cindex semaphores on static probe points
5297 Some @code{SystemTap} probes have an associated semaphore variable;
5298 for instance, this happens automatically if you defined your probe
5299 using a DTrace-style @file{.d} file. If your probe has a semaphore,
5300 @value{GDBN} will automatically enable it when you specify a
5301 breakpoint using the @samp{-probe-stap} notation. But, if you put a
5302 breakpoint at a probe's location by some other method (e.g.,
5303 @code{break file:line}), then @value{GDBN} will not automatically set
5304 the semaphore. @code{DTrace} probes do not support semaphores.
5306 You can examine the available static static probes using @code{info
5307 probes}, with optional arguments:
5311 @item info probes @r{[}@var{type}@r{]} @r{[}@var{provider} @r{[}@var{name} @r{[}@var{objfile}@r{]}@r{]}@r{]}
5312 If given, @var{type} is either @code{stap} for listing
5313 @code{SystemTap} probes or @code{dtrace} for listing @code{DTrace}
5314 probes. If omitted all probes are listed regardless of their types.
5316 If given, @var{provider} is a regular expression used to match against provider
5317 names when selecting which probes to list. If omitted, probes by all
5318 probes from all providers are listed.
5320 If given, @var{name} is a regular expression to match against probe names
5321 when selecting which probes to list. If omitted, probe names are not
5322 considered when deciding whether to display them.
5324 If given, @var{objfile} is a regular expression used to select which
5325 object files (executable or shared libraries) to examine. If not
5326 given, all object files are considered.
5328 @item info probes all
5329 List the available static probes, from all types.
5332 @cindex enabling and disabling probes
5333 Some probe points can be enabled and/or disabled. The effect of
5334 enabling or disabling a probe depends on the type of probe being
5335 handled. Some @code{DTrace} probes can be enabled or
5336 disabled, but @code{SystemTap} probes cannot be disabled.
5338 You can enable (or disable) one or more probes using the following
5339 commands, with optional arguments:
5342 @kindex enable probes
5343 @item enable probes @r{[}@var{provider} @r{[}@var{name} @r{[}@var{objfile}@r{]}@r{]}@r{]}
5344 If given, @var{provider} is a regular expression used to match against
5345 provider names when selecting which probes to enable. If omitted,
5346 all probes from all providers are enabled.
5348 If given, @var{name} is a regular expression to match against probe
5349 names when selecting which probes to enable. If omitted, probe names
5350 are not considered when deciding whether to enable them.
5352 If given, @var{objfile} is a regular expression used to select which
5353 object files (executable or shared libraries) to examine. If not
5354 given, all object files are considered.
5356 @kindex disable probes
5357 @item disable probes @r{[}@var{provider} @r{[}@var{name} @r{[}@var{objfile}@r{]}@r{]}@r{]}
5358 See the @code{enable probes} command above for a description of the
5359 optional arguments accepted by this command.
5362 @vindex $_probe_arg@r{, convenience variable}
5363 A probe may specify up to twelve arguments. These are available at the
5364 point at which the probe is defined---that is, when the current PC is
5365 at the probe's location. The arguments are available using the
5366 convenience variables (@pxref{Convenience Vars})
5367 @code{$_probe_arg0}@dots{}@code{$_probe_arg11}. In @code{SystemTap}
5368 probes each probe argument is an integer of the appropriate size;
5369 types are not preserved. In @code{DTrace} probes types are preserved
5370 provided that they are recognized as such by @value{GDBN}; otherwise
5371 the value of the probe argument will be a long integer. The
5372 convenience variable @code{$_probe_argc} holds the number of arguments
5373 at the current probe point.
5375 These variables are always available, but attempts to access them at
5376 any location other than a probe point will cause @value{GDBN} to give
5380 @c @ifclear BARETARGET
5381 @node Error in Breakpoints
5382 @subsection ``Cannot insert breakpoints''
5384 If you request too many active hardware-assisted breakpoints and
5385 watchpoints, you will see this error message:
5387 @c FIXME: the precise wording of this message may change; the relevant
5388 @c source change is not committed yet (Sep 3, 1999).
5390 Stopped; cannot insert breakpoints.
5391 You may have requested too many hardware breakpoints and watchpoints.
5395 This message is printed when you attempt to resume the program, since
5396 only then @value{GDBN} knows exactly how many hardware breakpoints and
5397 watchpoints it needs to insert.
5399 When this message is printed, you need to disable or remove some of the
5400 hardware-assisted breakpoints and watchpoints, and then continue.
5402 @node Breakpoint-related Warnings
5403 @subsection ``Breakpoint address adjusted...''
5404 @cindex breakpoint address adjusted
5406 Some processor architectures place constraints on the addresses at
5407 which breakpoints may be placed. For architectures thus constrained,
5408 @value{GDBN} will attempt to adjust the breakpoint's address to comply
5409 with the constraints dictated by the architecture.
5411 One example of such an architecture is the Fujitsu FR-V. The FR-V is
5412 a VLIW architecture in which a number of RISC-like instructions may be
5413 bundled together for parallel execution. The FR-V architecture
5414 constrains the location of a breakpoint instruction within such a
5415 bundle to the instruction with the lowest address. @value{GDBN}
5416 honors this constraint by adjusting a breakpoint's address to the
5417 first in the bundle.
5419 It is not uncommon for optimized code to have bundles which contain
5420 instructions from different source statements, thus it may happen that
5421 a breakpoint's address will be adjusted from one source statement to
5422 another. Since this adjustment may significantly alter @value{GDBN}'s
5423 breakpoint related behavior from what the user expects, a warning is
5424 printed when the breakpoint is first set and also when the breakpoint
5427 A warning like the one below is printed when setting a breakpoint
5428 that's been subject to address adjustment:
5431 warning: Breakpoint address adjusted from 0x00010414 to 0x00010410.
5434 Such warnings are printed both for user settable and @value{GDBN}'s
5435 internal breakpoints. If you see one of these warnings, you should
5436 verify that a breakpoint set at the adjusted address will have the
5437 desired affect. If not, the breakpoint in question may be removed and
5438 other breakpoints may be set which will have the desired behavior.
5439 E.g., it may be sufficient to place the breakpoint at a later
5440 instruction. A conditional breakpoint may also be useful in some
5441 cases to prevent the breakpoint from triggering too often.
5443 @value{GDBN} will also issue a warning when stopping at one of these
5444 adjusted breakpoints:
5447 warning: Breakpoint 1 address previously adjusted from 0x00010414
5451 When this warning is encountered, it may be too late to take remedial
5452 action except in cases where the breakpoint is hit earlier or more
5453 frequently than expected.
5455 @node Continuing and Stepping
5456 @section Continuing and Stepping
5460 @cindex resuming execution
5461 @dfn{Continuing} means resuming program execution until your program
5462 completes normally. In contrast, @dfn{stepping} means executing just
5463 one more ``step'' of your program, where ``step'' may mean either one
5464 line of source code, or one machine instruction (depending on what
5465 particular command you use). Either when continuing or when stepping,
5466 your program may stop even sooner, due to a breakpoint or a signal. (If
5467 it stops due to a signal, you may want to use @code{handle}, or use
5468 @samp{signal 0} to resume execution (@pxref{Signals, ,Signals}),
5469 or you may step into the signal's handler (@pxref{stepping and signal
5474 @kindex c @r{(@code{continue})}
5475 @kindex fg @r{(resume foreground execution)}
5476 @item continue @r{[}@var{ignore-count}@r{]}
5477 @itemx c @r{[}@var{ignore-count}@r{]}
5478 @itemx fg @r{[}@var{ignore-count}@r{]}
5479 Resume program execution, at the address where your program last stopped;
5480 any breakpoints set at that address are bypassed. The optional argument
5481 @var{ignore-count} allows you to specify a further number of times to
5482 ignore a breakpoint at this location; its effect is like that of
5483 @code{ignore} (@pxref{Conditions, ,Break Conditions}).
5485 The argument @var{ignore-count} is meaningful only when your program
5486 stopped due to a breakpoint. At other times, the argument to
5487 @code{continue} is ignored.
5489 The synonyms @code{c} and @code{fg} (for @dfn{foreground}, as the
5490 debugged program is deemed to be the foreground program) are provided
5491 purely for convenience, and have exactly the same behavior as
5495 To resume execution at a different place, you can use @code{return}
5496 (@pxref{Returning, ,Returning from a Function}) to go back to the
5497 calling function; or @code{jump} (@pxref{Jumping, ,Continuing at a
5498 Different Address}) to go to an arbitrary location in your program.
5500 A typical technique for using stepping is to set a breakpoint
5501 (@pxref{Breakpoints, ,Breakpoints; Watchpoints; and Catchpoints}) at the
5502 beginning of the function or the section of your program where a problem
5503 is believed to lie, run your program until it stops at that breakpoint,
5504 and then step through the suspect area, examining the variables that are
5505 interesting, until you see the problem happen.
5509 @kindex s @r{(@code{step})}
5511 Continue running your program until control reaches a different source
5512 line, then stop it and return control to @value{GDBN}. This command is
5513 abbreviated @code{s}.
5516 @c "without debugging information" is imprecise; actually "without line
5517 @c numbers in the debugging information". (gcc -g1 has debugging info but
5518 @c not line numbers). But it seems complex to try to make that
5519 @c distinction here.
5520 @emph{Warning:} If you use the @code{step} command while control is
5521 within a function that was compiled without debugging information,
5522 execution proceeds until control reaches a function that does have
5523 debugging information. Likewise, it will not step into a function which
5524 is compiled without debugging information. To step through functions
5525 without debugging information, use the @code{stepi} command, described
5529 The @code{step} command only stops at the first instruction of a source
5530 line. This prevents the multiple stops that could otherwise occur in
5531 @code{switch} statements, @code{for} loops, etc. @code{step} continues
5532 to stop if a function that has debugging information is called within
5533 the line. In other words, @code{step} @emph{steps inside} any functions
5534 called within the line.
5536 Also, the @code{step} command only enters a function if there is line
5537 number information for the function. Otherwise it acts like the
5538 @code{next} command. This avoids problems when using @code{cc -gl}
5539 on @acronym{MIPS} machines. Previously, @code{step} entered subroutines if there
5540 was any debugging information about the routine.
5542 @item step @var{count}
5543 Continue running as in @code{step}, but do so @var{count} times. If a
5544 breakpoint is reached, or a signal not related to stepping occurs before
5545 @var{count} steps, stepping stops right away.
5548 @kindex n @r{(@code{next})}
5549 @item next @r{[}@var{count}@r{]}
5550 Continue to the next source line in the current (innermost) stack frame.
5551 This is similar to @code{step}, but function calls that appear within
5552 the line of code are executed without stopping. Execution stops when
5553 control reaches a different line of code at the original stack level
5554 that was executing when you gave the @code{next} command. This command
5555 is abbreviated @code{n}.
5557 An argument @var{count} is a repeat count, as for @code{step}.
5560 @c FIX ME!! Do we delete this, or is there a way it fits in with
5561 @c the following paragraph? --- Vctoria
5563 @c @code{next} within a function that lacks debugging information acts like
5564 @c @code{step}, but any function calls appearing within the code of the
5565 @c function are executed without stopping.
5567 The @code{next} command only stops at the first instruction of a
5568 source line. This prevents multiple stops that could otherwise occur in
5569 @code{switch} statements, @code{for} loops, etc.
5571 @kindex set step-mode
5573 @cindex functions without line info, and stepping
5574 @cindex stepping into functions with no line info
5575 @itemx set step-mode on
5576 The @code{set step-mode on} command causes the @code{step} command to
5577 stop at the first instruction of a function which contains no debug line
5578 information rather than stepping over it.
5580 This is useful in cases where you may be interested in inspecting the
5581 machine instructions of a function which has no symbolic info and do not
5582 want @value{GDBN} to automatically skip over this function.
5584 @item set step-mode off
5585 Causes the @code{step} command to step over any functions which contains no
5586 debug information. This is the default.
5588 @item show step-mode
5589 Show whether @value{GDBN} will stop in or step over functions without
5590 source line debug information.
5593 @kindex fin @r{(@code{finish})}
5595 Continue running until just after function in the selected stack frame
5596 returns. Print the returned value (if any). This command can be
5597 abbreviated as @code{fin}.
5599 Contrast this with the @code{return} command (@pxref{Returning,
5600 ,Returning from a Function}).
5603 @kindex u @r{(@code{until})}
5604 @cindex run until specified location
5607 Continue running until a source line past the current line, in the
5608 current stack frame, is reached. This command is used to avoid single
5609 stepping through a loop more than once. It is like the @code{next}
5610 command, except that when @code{until} encounters a jump, it
5611 automatically continues execution until the program counter is greater
5612 than the address of the jump.
5614 This means that when you reach the end of a loop after single stepping
5615 though it, @code{until} makes your program continue execution until it
5616 exits the loop. In contrast, a @code{next} command at the end of a loop
5617 simply steps back to the beginning of the loop, which forces you to step
5618 through the next iteration.
5620 @code{until} always stops your program if it attempts to exit the current
5623 @code{until} may produce somewhat counterintuitive results if the order
5624 of machine code does not match the order of the source lines. For
5625 example, in the following excerpt from a debugging session, the @code{f}
5626 (@code{frame}) command shows that execution is stopped at line
5627 @code{206}; yet when we use @code{until}, we get to line @code{195}:
5631 #0 main (argc=4, argv=0xf7fffae8) at m4.c:206
5633 (@value{GDBP}) until
5634 195 for ( ; argc > 0; NEXTARG) @{
5637 This happened because, for execution efficiency, the compiler had
5638 generated code for the loop closure test at the end, rather than the
5639 start, of the loop---even though the test in a C @code{for}-loop is
5640 written before the body of the loop. The @code{until} command appeared
5641 to step back to the beginning of the loop when it advanced to this
5642 expression; however, it has not really gone to an earlier
5643 statement---not in terms of the actual machine code.
5645 @code{until} with no argument works by means of single
5646 instruction stepping, and hence is slower than @code{until} with an
5649 @item until @var{location}
5650 @itemx u @var{location}
5651 Continue running your program until either the specified @var{location} is
5652 reached, or the current stack frame returns. The location is any of
5653 the forms described in @ref{Specify Location}.
5654 This form of the command uses temporary breakpoints, and
5655 hence is quicker than @code{until} without an argument. The specified
5656 location is actually reached only if it is in the current frame. This
5657 implies that @code{until} can be used to skip over recursive function
5658 invocations. For instance in the code below, if the current location is
5659 line @code{96}, issuing @code{until 99} will execute the program up to
5660 line @code{99} in the same invocation of factorial, i.e., after the inner
5661 invocations have returned.
5664 94 int factorial (int value)
5666 96 if (value > 1) @{
5667 97 value *= factorial (value - 1);
5674 @kindex advance @var{location}
5675 @item advance @var{location}
5676 Continue running the program up to the given @var{location}. An argument is
5677 required, which should be of one of the forms described in
5678 @ref{Specify Location}.
5679 Execution will also stop upon exit from the current stack
5680 frame. This command is similar to @code{until}, but @code{advance} will
5681 not skip over recursive function calls, and the target location doesn't
5682 have to be in the same frame as the current one.
5686 @kindex si @r{(@code{stepi})}
5688 @itemx stepi @var{arg}
5690 Execute one machine instruction, then stop and return to the debugger.
5692 It is often useful to do @samp{display/i $pc} when stepping by machine
5693 instructions. This makes @value{GDBN} automatically display the next
5694 instruction to be executed, each time your program stops. @xref{Auto
5695 Display,, Automatic Display}.
5697 An argument is a repeat count, as in @code{step}.
5701 @kindex ni @r{(@code{nexti})}
5703 @itemx nexti @var{arg}
5705 Execute one machine instruction, but if it is a function call,
5706 proceed until the function returns.
5708 An argument is a repeat count, as in @code{next}.
5712 @anchor{range stepping}
5713 @cindex range stepping
5714 @cindex target-assisted range stepping
5715 By default, and if available, @value{GDBN} makes use of
5716 target-assisted @dfn{range stepping}. In other words, whenever you
5717 use a stepping command (e.g., @code{step}, @code{next}), @value{GDBN}
5718 tells the target to step the corresponding range of instruction
5719 addresses instead of issuing multiple single-steps. This speeds up
5720 line stepping, particularly for remote targets. Ideally, there should
5721 be no reason you would want to turn range stepping off. However, it's
5722 possible that a bug in the debug info, a bug in the remote stub (for
5723 remote targets), or even a bug in @value{GDBN} could make line
5724 stepping behave incorrectly when target-assisted range stepping is
5725 enabled. You can use the following command to turn off range stepping
5729 @kindex set range-stepping
5730 @kindex show range-stepping
5731 @item set range-stepping
5732 @itemx show range-stepping
5733 Control whether range stepping is enabled.
5735 If @code{on}, and the target supports it, @value{GDBN} tells the
5736 target to step a range of addresses itself, instead of issuing
5737 multiple single-steps. If @code{off}, @value{GDBN} always issues
5738 single-steps, even if range stepping is supported by the target. The
5739 default is @code{on}.
5743 @node Skipping Over Functions and Files
5744 @section Skipping Over Functions and Files
5745 @cindex skipping over functions and files
5747 The program you are debugging may contain some functions which are
5748 uninteresting to debug. The @code{skip} command lets you tell @value{GDBN} to
5749 skip a function, all functions in a file or a particular function in
5750 a particular file when stepping.
5752 For example, consider the following C function:
5763 Suppose you wish to step into the functions @code{foo} and @code{bar}, but you
5764 are not interested in stepping through @code{boring}. If you run @code{step}
5765 at line 103, you'll enter @code{boring()}, but if you run @code{next}, you'll
5766 step over both @code{foo} and @code{boring}!
5768 One solution is to @code{step} into @code{boring} and use the @code{finish}
5769 command to immediately exit it. But this can become tedious if @code{boring}
5770 is called from many places.
5772 A more flexible solution is to execute @kbd{skip boring}. This instructs
5773 @value{GDBN} never to step into @code{boring}. Now when you execute
5774 @code{step} at line 103, you'll step over @code{boring} and directly into
5777 Functions may be skipped by providing either a function name, linespec
5778 (@pxref{Specify Location}), regular expression that matches the function's
5779 name, file name or a @code{glob}-style pattern that matches the file name.
5781 On Posix systems the form of the regular expression is
5782 ``Extended Regular Expressions''. See for example @samp{man 7 regex}
5783 on @sc{gnu}/Linux systems. On non-Posix systems the form of the regular
5784 expression is whatever is provided by the @code{regcomp} function of
5785 the underlying system.
5786 See for example @samp{man 7 glob} on @sc{gnu}/Linux systems for a
5787 description of @code{glob}-style patterns.
5791 @item skip @r{[}@var{options}@r{]}
5792 The basic form of the @code{skip} command takes zero or more options
5793 that specify what to skip.
5794 The @var{options} argument is any useful combination of the following:
5797 @item -file @var{file}
5798 @itemx -fi @var{file}
5799 Functions in @var{file} will be skipped over when stepping.
5801 @item -gfile @var{file-glob-pattern}
5802 @itemx -gfi @var{file-glob-pattern}
5803 @cindex skipping over files via glob-style patterns
5804 Functions in files matching @var{file-glob-pattern} will be skipped
5808 (gdb) skip -gfi utils/*.c
5811 @item -function @var{linespec}
5812 @itemx -fu @var{linespec}
5813 Functions named by @var{linespec} or the function containing the line
5814 named by @var{linespec} will be skipped over when stepping.
5815 @xref{Specify Location}.
5817 @item -rfunction @var{regexp}
5818 @itemx -rfu @var{regexp}
5819 @cindex skipping over functions via regular expressions
5820 Functions whose name matches @var{regexp} will be skipped over when stepping.
5822 This form is useful for complex function names.
5823 For example, there is generally no need to step into C@t{++} @code{std::string}
5824 constructors or destructors. Plus with C@t{++} templates it can be hard to
5825 write out the full name of the function, and often it doesn't matter what
5826 the template arguments are. Specifying the function to be skipped as a
5827 regular expression makes this easier.
5830 (gdb) skip -rfu ^std::(allocator|basic_string)<.*>::~?\1 *\(
5833 If you want to skip every templated C@t{++} constructor and destructor
5834 in the @code{std} namespace you can do:
5837 (gdb) skip -rfu ^std::([a-zA-z0-9_]+)<.*>::~?\1 *\(
5841 If no options are specified, the function you're currently debugging
5844 @kindex skip function
5845 @item skip function @r{[}@var{linespec}@r{]}
5846 After running this command, the function named by @var{linespec} or the
5847 function containing the line named by @var{linespec} will be skipped over when
5848 stepping. @xref{Specify Location}.
5850 If you do not specify @var{linespec}, the function you're currently debugging
5853 (If you have a function called @code{file} that you want to skip, use
5854 @kbd{skip function file}.)
5857 @item skip file @r{[}@var{filename}@r{]}
5858 After running this command, any function whose source lives in @var{filename}
5859 will be skipped over when stepping.
5862 (gdb) skip file boring.c
5863 File boring.c will be skipped when stepping.
5866 If you do not specify @var{filename}, functions whose source lives in the file
5867 you're currently debugging will be skipped.
5870 Skips can be listed, deleted, disabled, and enabled, much like breakpoints.
5871 These are the commands for managing your list of skips:
5875 @item info skip @r{[}@var{range}@r{]}
5876 Print details about the specified skip(s). If @var{range} is not specified,
5877 print a table with details about all functions and files marked for skipping.
5878 @code{info skip} prints the following information about each skip:
5882 A number identifying this skip.
5883 @item Enabled or Disabled
5884 Enabled skips are marked with @samp{y}.
5885 Disabled skips are marked with @samp{n}.
5887 If the file name is a @samp{glob} pattern this is @samp{y}.
5888 Otherwise it is @samp{n}.
5890 The name or @samp{glob} pattern of the file to be skipped.
5891 If no file is specified this is @samp{<none>}.
5893 If the function name is a @samp{regular expression} this is @samp{y}.
5894 Otherwise it is @samp{n}.
5896 The name or regular expression of the function to skip.
5897 If no function is specified this is @samp{<none>}.
5901 @item skip delete @r{[}@var{range}@r{]}
5902 Delete the specified skip(s). If @var{range} is not specified, delete all
5906 @item skip enable @r{[}@var{range}@r{]}
5907 Enable the specified skip(s). If @var{range} is not specified, enable all
5910 @kindex skip disable
5911 @item skip disable @r{[}@var{range}@r{]}
5912 Disable the specified skip(s). If @var{range} is not specified, disable all
5915 @kindex set debug skip
5916 @item set debug skip @r{[}on|off@r{]}
5917 Set whether to print the debug output about skipping files and functions.
5919 @kindex show debug skip
5920 @item show debug skip
5921 Show whether the debug output about skipping files and functions is printed.
5929 A signal is an asynchronous event that can happen in a program. The
5930 operating system defines the possible kinds of signals, and gives each
5931 kind a name and a number. For example, in Unix @code{SIGINT} is the
5932 signal a program gets when you type an interrupt character (often @kbd{Ctrl-c});
5933 @code{SIGSEGV} is the signal a program gets from referencing a place in
5934 memory far away from all the areas in use; @code{SIGALRM} occurs when
5935 the alarm clock timer goes off (which happens only if your program has
5936 requested an alarm).
5938 @cindex fatal signals
5939 Some signals, including @code{SIGALRM}, are a normal part of the
5940 functioning of your program. Others, such as @code{SIGSEGV}, indicate
5941 errors; these signals are @dfn{fatal} (they kill your program immediately) if the
5942 program has not specified in advance some other way to handle the signal.
5943 @code{SIGINT} does not indicate an error in your program, but it is normally
5944 fatal so it can carry out the purpose of the interrupt: to kill the program.
5946 @value{GDBN} has the ability to detect any occurrence of a signal in your
5947 program. You can tell @value{GDBN} in advance what to do for each kind of
5950 @cindex handling signals
5951 Normally, @value{GDBN} is set up to let the non-erroneous signals like
5952 @code{SIGALRM} be silently passed to your program
5953 (so as not to interfere with their role in the program's functioning)
5954 but to stop your program immediately whenever an error signal happens.
5955 You can change these settings with the @code{handle} command.
5958 @kindex info signals
5962 Print a table of all the kinds of signals and how @value{GDBN} has been told to
5963 handle each one. You can use this to see the signal numbers of all
5964 the defined types of signals.
5966 @item info signals @var{sig}
5967 Similar, but print information only about the specified signal number.
5969 @code{info handle} is an alias for @code{info signals}.
5971 @item catch signal @r{[}@var{signal}@dots{} @r{|} @samp{all}@r{]}
5972 Set a catchpoint for the indicated signals. @xref{Set Catchpoints},
5973 for details about this command.
5976 @item handle @var{signal} @r{[}@var{keywords}@dots{}@r{]}
5977 Change the way @value{GDBN} handles signal @var{signal}. The @var{signal}
5978 can be the number of a signal or its name (with or without the
5979 @samp{SIG} at the beginning); a list of signal numbers of the form
5980 @samp{@var{low}-@var{high}}; or the word @samp{all}, meaning all the
5981 known signals. Optional arguments @var{keywords}, described below,
5982 say what change to make.
5986 The keywords allowed by the @code{handle} command can be abbreviated.
5987 Their full names are:
5991 @value{GDBN} should not stop your program when this signal happens. It may
5992 still print a message telling you that the signal has come in.
5995 @value{GDBN} should stop your program when this signal happens. This implies
5996 the @code{print} keyword as well.
5999 @value{GDBN} should print a message when this signal happens.
6002 @value{GDBN} should not mention the occurrence of the signal at all. This
6003 implies the @code{nostop} keyword as well.
6007 @value{GDBN} should allow your program to see this signal; your program
6008 can handle the signal, or else it may terminate if the signal is fatal
6009 and not handled. @code{pass} and @code{noignore} are synonyms.
6013 @value{GDBN} should not allow your program to see this signal.
6014 @code{nopass} and @code{ignore} are synonyms.
6018 When a signal stops your program, the signal is not visible to the
6020 continue. Your program sees the signal then, if @code{pass} is in
6021 effect for the signal in question @emph{at that time}. In other words,
6022 after @value{GDBN} reports a signal, you can use the @code{handle}
6023 command with @code{pass} or @code{nopass} to control whether your
6024 program sees that signal when you continue.
6026 The default is set to @code{nostop}, @code{noprint}, @code{pass} for
6027 non-erroneous signals such as @code{SIGALRM}, @code{SIGWINCH} and
6028 @code{SIGCHLD}, and to @code{stop}, @code{print}, @code{pass} for the
6031 You can also use the @code{signal} command to prevent your program from
6032 seeing a signal, or cause it to see a signal it normally would not see,
6033 or to give it any signal at any time. For example, if your program stopped
6034 due to some sort of memory reference error, you might store correct
6035 values into the erroneous variables and continue, hoping to see more
6036 execution; but your program would probably terminate immediately as
6037 a result of the fatal signal once it saw the signal. To prevent this,
6038 you can continue with @samp{signal 0}. @xref{Signaling, ,Giving your
6041 @cindex stepping and signal handlers
6042 @anchor{stepping and signal handlers}
6044 @value{GDBN} optimizes for stepping the mainline code. If a signal
6045 that has @code{handle nostop} and @code{handle pass} set arrives while
6046 a stepping command (e.g., @code{stepi}, @code{step}, @code{next}) is
6047 in progress, @value{GDBN} lets the signal handler run and then resumes
6048 stepping the mainline code once the signal handler returns. In other
6049 words, @value{GDBN} steps over the signal handler. This prevents
6050 signals that you've specified as not interesting (with @code{handle
6051 nostop}) from changing the focus of debugging unexpectedly. Note that
6052 the signal handler itself may still hit a breakpoint, stop for another
6053 signal that has @code{handle stop} in effect, or for any other event
6054 that normally results in stopping the stepping command sooner. Also
6055 note that @value{GDBN} still informs you that the program received a
6056 signal if @code{handle print} is set.
6058 @anchor{stepping into signal handlers}
6060 If you set @code{handle pass} for a signal, and your program sets up a
6061 handler for it, then issuing a stepping command, such as @code{step}
6062 or @code{stepi}, when your program is stopped due to the signal will
6063 step @emph{into} the signal handler (if the target supports that).
6065 Likewise, if you use the @code{queue-signal} command to queue a signal
6066 to be delivered to the current thread when execution of the thread
6067 resumes (@pxref{Signaling, ,Giving your Program a Signal}), then a
6068 stepping command will step into the signal handler.
6070 Here's an example, using @code{stepi} to step to the first instruction
6071 of @code{SIGUSR1}'s handler:
6074 (@value{GDBP}) handle SIGUSR1
6075 Signal Stop Print Pass to program Description
6076 SIGUSR1 Yes Yes Yes User defined signal 1
6080 Program received signal SIGUSR1, User defined signal 1.
6081 main () sigusr1.c:28
6084 sigusr1_handler () at sigusr1.c:9
6088 The same, but using @code{queue-signal} instead of waiting for the
6089 program to receive the signal first:
6094 (@value{GDBP}) queue-signal SIGUSR1
6096 sigusr1_handler () at sigusr1.c:9
6101 @cindex extra signal information
6102 @anchor{extra signal information}
6104 On some targets, @value{GDBN} can inspect extra signal information
6105 associated with the intercepted signal, before it is actually
6106 delivered to the program being debugged. This information is exported
6107 by the convenience variable @code{$_siginfo}, and consists of data
6108 that is passed by the kernel to the signal handler at the time of the
6109 receipt of a signal. The data type of the information itself is
6110 target dependent. You can see the data type using the @code{ptype
6111 $_siginfo} command. On Unix systems, it typically corresponds to the
6112 standard @code{siginfo_t} type, as defined in the @file{signal.h}
6115 Here's an example, on a @sc{gnu}/Linux system, printing the stray
6116 referenced address that raised a segmentation fault.
6120 (@value{GDBP}) continue
6121 Program received signal SIGSEGV, Segmentation fault.
6122 0x0000000000400766 in main ()
6124 (@value{GDBP}) ptype $_siginfo
6131 struct @{...@} _kill;
6132 struct @{...@} _timer;
6134 struct @{...@} _sigchld;
6135 struct @{...@} _sigfault;
6136 struct @{...@} _sigpoll;
6139 (@value{GDBP}) ptype $_siginfo._sifields._sigfault
6143 (@value{GDBP}) p $_siginfo._sifields._sigfault.si_addr
6144 $1 = (void *) 0x7ffff7ff7000
6148 Depending on target support, @code{$_siginfo} may also be writable.
6150 @cindex Intel MPX boundary violations
6151 @cindex boundary violations, Intel MPX
6152 On some targets, a @code{SIGSEGV} can be caused by a boundary
6153 violation, i.e., accessing an address outside of the allowed range.
6154 In those cases @value{GDBN} may displays additional information,
6155 depending on how @value{GDBN} has been told to handle the signal.
6156 With @code{handle stop SIGSEGV}, @value{GDBN} displays the violation
6157 kind: "Upper" or "Lower", the memory address accessed and the
6158 bounds, while with @code{handle nostop SIGSEGV} no additional
6159 information is displayed.
6161 The usual output of a segfault is:
6163 Program received signal SIGSEGV, Segmentation fault
6164 0x0000000000400d7c in upper () at i386-mpx-sigsegv.c:68
6165 68 value = *(p + len);
6168 While a bound violation is presented as:
6170 Program received signal SIGSEGV, Segmentation fault
6171 Upper bound violation while accessing address 0x7fffffffc3b3
6172 Bounds: [lower = 0x7fffffffc390, upper = 0x7fffffffc3a3]
6173 0x0000000000400d7c in upper () at i386-mpx-sigsegv.c:68
6174 68 value = *(p + len);
6178 @section Stopping and Starting Multi-thread Programs
6180 @cindex stopped threads
6181 @cindex threads, stopped
6183 @cindex continuing threads
6184 @cindex threads, continuing
6186 @value{GDBN} supports debugging programs with multiple threads
6187 (@pxref{Threads,, Debugging Programs with Multiple Threads}). There
6188 are two modes of controlling execution of your program within the
6189 debugger. In the default mode, referred to as @dfn{all-stop mode},
6190 when any thread in your program stops (for example, at a breakpoint
6191 or while being stepped), all other threads in the program are also stopped by
6192 @value{GDBN}. On some targets, @value{GDBN} also supports
6193 @dfn{non-stop mode}, in which other threads can continue to run freely while
6194 you examine the stopped thread in the debugger.
6197 * All-Stop Mode:: All threads stop when GDB takes control
6198 * Non-Stop Mode:: Other threads continue to execute
6199 * Background Execution:: Running your program asynchronously
6200 * Thread-Specific Breakpoints:: Controlling breakpoints
6201 * Interrupted System Calls:: GDB may interfere with system calls
6202 * Observer Mode:: GDB does not alter program behavior
6206 @subsection All-Stop Mode
6208 @cindex all-stop mode
6210 In all-stop mode, whenever your program stops under @value{GDBN} for any reason,
6211 @emph{all} threads of execution stop, not just the current thread. This
6212 allows you to examine the overall state of the program, including
6213 switching between threads, without worrying that things may change
6216 Conversely, whenever you restart the program, @emph{all} threads start
6217 executing. @emph{This is true even when single-stepping} with commands
6218 like @code{step} or @code{next}.
6220 In particular, @value{GDBN} cannot single-step all threads in lockstep.
6221 Since thread scheduling is up to your debugging target's operating
6222 system (not controlled by @value{GDBN}), other threads may
6223 execute more than one statement while the current thread completes a
6224 single step. Moreover, in general other threads stop in the middle of a
6225 statement, rather than at a clean statement boundary, when the program
6228 You might even find your program stopped in another thread after
6229 continuing or even single-stepping. This happens whenever some other
6230 thread runs into a breakpoint, a signal, or an exception before the
6231 first thread completes whatever you requested.
6233 @cindex automatic thread selection
6234 @cindex switching threads automatically
6235 @cindex threads, automatic switching
6236 Whenever @value{GDBN} stops your program, due to a breakpoint or a
6237 signal, it automatically selects the thread where that breakpoint or
6238 signal happened. @value{GDBN} alerts you to the context switch with a
6239 message such as @samp{[Switching to Thread @var{n}]} to identify the
6242 On some OSes, you can modify @value{GDBN}'s default behavior by
6243 locking the OS scheduler to allow only a single thread to run.
6246 @item set scheduler-locking @var{mode}
6247 @cindex scheduler locking mode
6248 @cindex lock scheduler
6249 Set the scheduler locking mode. It applies to normal execution,
6250 record mode, and replay mode. If it is @code{off}, then there is no
6251 locking and any thread may run at any time. If @code{on}, then only
6252 the current thread may run when the inferior is resumed. The
6253 @code{step} mode optimizes for single-stepping; it prevents other
6254 threads from preempting the current thread while you are stepping, so
6255 that the focus of debugging does not change unexpectedly. Other
6256 threads never get a chance to run when you step, and they are
6257 completely free to run when you use commands like @samp{continue},
6258 @samp{until}, or @samp{finish}. However, unless another thread hits a
6259 breakpoint during its timeslice, @value{GDBN} does not change the
6260 current thread away from the thread that you are debugging. The
6261 @code{replay} mode behaves like @code{off} in record mode and like
6262 @code{on} in replay mode.
6264 @item show scheduler-locking
6265 Display the current scheduler locking mode.
6268 @cindex resume threads of multiple processes simultaneously
6269 By default, when you issue one of the execution commands such as
6270 @code{continue}, @code{next} or @code{step}, @value{GDBN} allows only
6271 threads of the current inferior to run. For example, if @value{GDBN}
6272 is attached to two inferiors, each with two threads, the
6273 @code{continue} command resumes only the two threads of the current
6274 inferior. This is useful, for example, when you debug a program that
6275 forks and you want to hold the parent stopped (so that, for instance,
6276 it doesn't run to exit), while you debug the child. In other
6277 situations, you may not be interested in inspecting the current state
6278 of any of the processes @value{GDBN} is attached to, and you may want
6279 to resume them all until some breakpoint is hit. In the latter case,
6280 you can instruct @value{GDBN} to allow all threads of all the
6281 inferiors to run with the @w{@code{set schedule-multiple}} command.
6284 @kindex set schedule-multiple
6285 @item set schedule-multiple
6286 Set the mode for allowing threads of multiple processes to be resumed
6287 when an execution command is issued. When @code{on}, all threads of
6288 all processes are allowed to run. When @code{off}, only the threads
6289 of the current process are resumed. The default is @code{off}. The
6290 @code{scheduler-locking} mode takes precedence when set to @code{on},
6291 or while you are stepping and set to @code{step}.
6293 @item show schedule-multiple
6294 Display the current mode for resuming the execution of threads of
6299 @subsection Non-Stop Mode
6301 @cindex non-stop mode
6303 @c This section is really only a place-holder, and needs to be expanded
6304 @c with more details.
6306 For some multi-threaded targets, @value{GDBN} supports an optional
6307 mode of operation in which you can examine stopped program threads in
6308 the debugger while other threads continue to execute freely. This
6309 minimizes intrusion when debugging live systems, such as programs
6310 where some threads have real-time constraints or must continue to
6311 respond to external events. This is referred to as @dfn{non-stop} mode.
6313 In non-stop mode, when a thread stops to report a debugging event,
6314 @emph{only} that thread is stopped; @value{GDBN} does not stop other
6315 threads as well, in contrast to the all-stop mode behavior. Additionally,
6316 execution commands such as @code{continue} and @code{step} apply by default
6317 only to the current thread in non-stop mode, rather than all threads as
6318 in all-stop mode. This allows you to control threads explicitly in
6319 ways that are not possible in all-stop mode --- for example, stepping
6320 one thread while allowing others to run freely, stepping
6321 one thread while holding all others stopped, or stepping several threads
6322 independently and simultaneously.
6324 To enter non-stop mode, use this sequence of commands before you run
6325 or attach to your program:
6328 # If using the CLI, pagination breaks non-stop.
6331 # Finally, turn it on!
6335 You can use these commands to manipulate the non-stop mode setting:
6338 @kindex set non-stop
6339 @item set non-stop on
6340 Enable selection of non-stop mode.
6341 @item set non-stop off
6342 Disable selection of non-stop mode.
6343 @kindex show non-stop
6345 Show the current non-stop enablement setting.
6348 Note these commands only reflect whether non-stop mode is enabled,
6349 not whether the currently-executing program is being run in non-stop mode.
6350 In particular, the @code{set non-stop} preference is only consulted when
6351 @value{GDBN} starts or connects to the target program, and it is generally
6352 not possible to switch modes once debugging has started. Furthermore,
6353 since not all targets support non-stop mode, even when you have enabled
6354 non-stop mode, @value{GDBN} may still fall back to all-stop operation by
6357 In non-stop mode, all execution commands apply only to the current thread
6358 by default. That is, @code{continue} only continues one thread.
6359 To continue all threads, issue @code{continue -a} or @code{c -a}.
6361 You can use @value{GDBN}'s background execution commands
6362 (@pxref{Background Execution}) to run some threads in the background
6363 while you continue to examine or step others from @value{GDBN}.
6364 The MI execution commands (@pxref{GDB/MI Program Execution}) are
6365 always executed asynchronously in non-stop mode.
6367 Suspending execution is done with the @code{interrupt} command when
6368 running in the background, or @kbd{Ctrl-c} during foreground execution.
6369 In all-stop mode, this stops the whole process;
6370 but in non-stop mode the interrupt applies only to the current thread.
6371 To stop the whole program, use @code{interrupt -a}.
6373 Other execution commands do not currently support the @code{-a} option.
6375 In non-stop mode, when a thread stops, @value{GDBN} doesn't automatically make
6376 that thread current, as it does in all-stop mode. This is because the
6377 thread stop notifications are asynchronous with respect to @value{GDBN}'s
6378 command interpreter, and it would be confusing if @value{GDBN} unexpectedly
6379 changed to a different thread just as you entered a command to operate on the
6380 previously current thread.
6382 @node Background Execution
6383 @subsection Background Execution
6385 @cindex foreground execution
6386 @cindex background execution
6387 @cindex asynchronous execution
6388 @cindex execution, foreground, background and asynchronous
6390 @value{GDBN}'s execution commands have two variants: the normal
6391 foreground (synchronous) behavior, and a background
6392 (asynchronous) behavior. In foreground execution, @value{GDBN} waits for
6393 the program to report that some thread has stopped before prompting for
6394 another command. In background execution, @value{GDBN} immediately gives
6395 a command prompt so that you can issue other commands while your program runs.
6397 If the target doesn't support async mode, @value{GDBN} issues an error
6398 message if you attempt to use the background execution commands.
6400 @cindex @code{&}, background execution of commands
6401 To specify background execution, add a @code{&} to the command. For example,
6402 the background form of the @code{continue} command is @code{continue&}, or
6403 just @code{c&}. The execution commands that accept background execution
6409 @xref{Starting, , Starting your Program}.
6413 @xref{Attach, , Debugging an Already-running Process}.
6417 @xref{Continuing and Stepping, step}.
6421 @xref{Continuing and Stepping, stepi}.
6425 @xref{Continuing and Stepping, next}.
6429 @xref{Continuing and Stepping, nexti}.
6433 @xref{Continuing and Stepping, continue}.
6437 @xref{Continuing and Stepping, finish}.
6441 @xref{Continuing and Stepping, until}.
6445 Background execution is especially useful in conjunction with non-stop
6446 mode for debugging programs with multiple threads; see @ref{Non-Stop Mode}.
6447 However, you can also use these commands in the normal all-stop mode with
6448 the restriction that you cannot issue another execution command until the
6449 previous one finishes. Examples of commands that are valid in all-stop
6450 mode while the program is running include @code{help} and @code{info break}.
6452 You can interrupt your program while it is running in the background by
6453 using the @code{interrupt} command.
6460 Suspend execution of the running program. In all-stop mode,
6461 @code{interrupt} stops the whole process, but in non-stop mode, it stops
6462 only the current thread. To stop the whole program in non-stop mode,
6463 use @code{interrupt -a}.
6466 @node Thread-Specific Breakpoints
6467 @subsection Thread-Specific Breakpoints
6469 When your program has multiple threads (@pxref{Threads,, Debugging
6470 Programs with Multiple Threads}), you can choose whether to set
6471 breakpoints on all threads, or on a particular thread.
6474 @cindex breakpoints and threads
6475 @cindex thread breakpoints
6476 @kindex break @dots{} thread @var{thread-id}
6477 @item break @var{location} thread @var{thread-id}
6478 @itemx break @var{location} thread @var{thread-id} if @dots{}
6479 @var{location} specifies source lines; there are several ways of
6480 writing them (@pxref{Specify Location}), but the effect is always to
6481 specify some source line.
6483 Use the qualifier @samp{thread @var{thread-id}} with a breakpoint command
6484 to specify that you only want @value{GDBN} to stop the program when a
6485 particular thread reaches this breakpoint. The @var{thread-id} specifier
6486 is one of the thread identifiers assigned by @value{GDBN}, shown
6487 in the first column of the @samp{info threads} display.
6489 If you do not specify @samp{thread @var{thread-id}} when you set a
6490 breakpoint, the breakpoint applies to @emph{all} threads of your
6493 You can use the @code{thread} qualifier on conditional breakpoints as
6494 well; in this case, place @samp{thread @var{thread-id}} before or
6495 after the breakpoint condition, like this:
6498 (@value{GDBP}) break frik.c:13 thread 28 if bartab > lim
6503 Thread-specific breakpoints are automatically deleted when
6504 @value{GDBN} detects the corresponding thread is no longer in the
6505 thread list. For example:
6509 Thread-specific breakpoint 3 deleted - thread 28 no longer in the thread list.
6512 There are several ways for a thread to disappear, such as a regular
6513 thread exit, but also when you detach from the process with the
6514 @code{detach} command (@pxref{Attach, ,Debugging an Already-running
6515 Process}), or if @value{GDBN} loses the remote connection
6516 (@pxref{Remote Debugging}), etc. Note that with some targets,
6517 @value{GDBN} is only able to detect a thread has exited when the user
6518 explictly asks for the thread list with the @code{info threads}
6521 @node Interrupted System Calls
6522 @subsection Interrupted System Calls
6524 @cindex thread breakpoints and system calls
6525 @cindex system calls and thread breakpoints
6526 @cindex premature return from system calls
6527 There is an unfortunate side effect when using @value{GDBN} to debug
6528 multi-threaded programs. If one thread stops for a
6529 breakpoint, or for some other reason, and another thread is blocked in a
6530 system call, then the system call may return prematurely. This is a
6531 consequence of the interaction between multiple threads and the signals
6532 that @value{GDBN} uses to implement breakpoints and other events that
6535 To handle this problem, your program should check the return value of
6536 each system call and react appropriately. This is good programming
6539 For example, do not write code like this:
6545 The call to @code{sleep} will return early if a different thread stops
6546 at a breakpoint or for some other reason.
6548 Instead, write this:
6553 unslept = sleep (unslept);
6556 A system call is allowed to return early, so the system is still
6557 conforming to its specification. But @value{GDBN} does cause your
6558 multi-threaded program to behave differently than it would without
6561 Also, @value{GDBN} uses internal breakpoints in the thread library to
6562 monitor certain events such as thread creation and thread destruction.
6563 When such an event happens, a system call in another thread may return
6564 prematurely, even though your program does not appear to stop.
6567 @subsection Observer Mode
6569 If you want to build on non-stop mode and observe program behavior
6570 without any chance of disruption by @value{GDBN}, you can set
6571 variables to disable all of the debugger's attempts to modify state,
6572 whether by writing memory, inserting breakpoints, etc. These operate
6573 at a low level, intercepting operations from all commands.
6575 When all of these are set to @code{off}, then @value{GDBN} is said to
6576 be @dfn{observer mode}. As a convenience, the variable
6577 @code{observer} can be set to disable these, plus enable non-stop
6580 Note that @value{GDBN} will not prevent you from making nonsensical
6581 combinations of these settings. For instance, if you have enabled
6582 @code{may-insert-breakpoints} but disabled @code{may-write-memory},
6583 then breakpoints that work by writing trap instructions into the code
6584 stream will still not be able to be placed.
6589 @item set observer on
6590 @itemx set observer off
6591 When set to @code{on}, this disables all the permission variables
6592 below (except for @code{insert-fast-tracepoints}), plus enables
6593 non-stop debugging. Setting this to @code{off} switches back to
6594 normal debugging, though remaining in non-stop mode.
6597 Show whether observer mode is on or off.
6599 @kindex may-write-registers
6600 @item set may-write-registers on
6601 @itemx set may-write-registers off
6602 This controls whether @value{GDBN} will attempt to alter the values of
6603 registers, such as with assignment expressions in @code{print}, or the
6604 @code{jump} command. It defaults to @code{on}.
6606 @item show may-write-registers
6607 Show the current permission to write registers.
6609 @kindex may-write-memory
6610 @item set may-write-memory on
6611 @itemx set may-write-memory off
6612 This controls whether @value{GDBN} will attempt to alter the contents
6613 of memory, such as with assignment expressions in @code{print}. It
6614 defaults to @code{on}.
6616 @item show may-write-memory
6617 Show the current permission to write memory.
6619 @kindex may-insert-breakpoints
6620 @item set may-insert-breakpoints on
6621 @itemx set may-insert-breakpoints off
6622 This controls whether @value{GDBN} will attempt to insert breakpoints.
6623 This affects all breakpoints, including internal breakpoints defined
6624 by @value{GDBN}. It defaults to @code{on}.
6626 @item show may-insert-breakpoints
6627 Show the current permission to insert breakpoints.
6629 @kindex may-insert-tracepoints
6630 @item set may-insert-tracepoints on
6631 @itemx set may-insert-tracepoints off
6632 This controls whether @value{GDBN} will attempt to insert (regular)
6633 tracepoints at the beginning of a tracing experiment. It affects only
6634 non-fast tracepoints, fast tracepoints being under the control of
6635 @code{may-insert-fast-tracepoints}. It defaults to @code{on}.
6637 @item show may-insert-tracepoints
6638 Show the current permission to insert tracepoints.
6640 @kindex may-insert-fast-tracepoints
6641 @item set may-insert-fast-tracepoints on
6642 @itemx set may-insert-fast-tracepoints off
6643 This controls whether @value{GDBN} will attempt to insert fast
6644 tracepoints at the beginning of a tracing experiment. It affects only
6645 fast tracepoints, regular (non-fast) tracepoints being under the
6646 control of @code{may-insert-tracepoints}. It defaults to @code{on}.
6648 @item show may-insert-fast-tracepoints
6649 Show the current permission to insert fast tracepoints.
6651 @kindex may-interrupt
6652 @item set may-interrupt on
6653 @itemx set may-interrupt off
6654 This controls whether @value{GDBN} will attempt to interrupt or stop
6655 program execution. When this variable is @code{off}, the
6656 @code{interrupt} command will have no effect, nor will
6657 @kbd{Ctrl-c}. It defaults to @code{on}.
6659 @item show may-interrupt
6660 Show the current permission to interrupt or stop the program.
6664 @node Reverse Execution
6665 @chapter Running programs backward
6666 @cindex reverse execution
6667 @cindex running programs backward
6669 When you are debugging a program, it is not unusual to realize that
6670 you have gone too far, and some event of interest has already happened.
6671 If the target environment supports it, @value{GDBN} can allow you to
6672 ``rewind'' the program by running it backward.
6674 A target environment that supports reverse execution should be able
6675 to ``undo'' the changes in machine state that have taken place as the
6676 program was executing normally. Variables, registers etc.@: should
6677 revert to their previous values. Obviously this requires a great
6678 deal of sophistication on the part of the target environment; not
6679 all target environments can support reverse execution.
6681 When a program is executed in reverse, the instructions that
6682 have most recently been executed are ``un-executed'', in reverse
6683 order. The program counter runs backward, following the previous
6684 thread of execution in reverse. As each instruction is ``un-executed'',
6685 the values of memory and/or registers that were changed by that
6686 instruction are reverted to their previous states. After executing
6687 a piece of source code in reverse, all side effects of that code
6688 should be ``undone'', and all variables should be returned to their
6689 prior values@footnote{
6690 Note that some side effects are easier to undo than others. For instance,
6691 memory and registers are relatively easy, but device I/O is hard. Some
6692 targets may be able undo things like device I/O, and some may not.
6694 The contract between @value{GDBN} and the reverse executing target
6695 requires only that the target do something reasonable when
6696 @value{GDBN} tells it to execute backwards, and then report the
6697 results back to @value{GDBN}. Whatever the target reports back to
6698 @value{GDBN}, @value{GDBN} will report back to the user. @value{GDBN}
6699 assumes that the memory and registers that the target reports are in a
6700 consistant state, but @value{GDBN} accepts whatever it is given.
6703 On some platforms, @value{GDBN} has built-in support for reverse
6704 execution, activated with the @code{record} or @code{record btrace}
6705 commands. @xref{Process Record and Replay}. Some remote targets,
6706 typically full system emulators, support reverse execution directly
6707 without requiring any special command.
6709 If you are debugging in a target environment that supports
6710 reverse execution, @value{GDBN} provides the following commands.
6713 @kindex reverse-continue
6714 @kindex rc @r{(@code{reverse-continue})}
6715 @item reverse-continue @r{[}@var{ignore-count}@r{]}
6716 @itemx rc @r{[}@var{ignore-count}@r{]}
6717 Beginning at the point where your program last stopped, start executing
6718 in reverse. Reverse execution will stop for breakpoints and synchronous
6719 exceptions (signals), just like normal execution. Behavior of
6720 asynchronous signals depends on the target environment.
6722 @kindex reverse-step
6723 @kindex rs @r{(@code{step})}
6724 @item reverse-step @r{[}@var{count}@r{]}
6725 Run the program backward until control reaches the start of a
6726 different source line; then stop it, and return control to @value{GDBN}.
6728 Like the @code{step} command, @code{reverse-step} will only stop
6729 at the beginning of a source line. It ``un-executes'' the previously
6730 executed source line. If the previous source line included calls to
6731 debuggable functions, @code{reverse-step} will step (backward) into
6732 the called function, stopping at the beginning of the @emph{last}
6733 statement in the called function (typically a return statement).
6735 Also, as with the @code{step} command, if non-debuggable functions are
6736 called, @code{reverse-step} will run thru them backward without stopping.
6738 @kindex reverse-stepi
6739 @kindex rsi @r{(@code{reverse-stepi})}
6740 @item reverse-stepi @r{[}@var{count}@r{]}
6741 Reverse-execute one machine instruction. Note that the instruction
6742 to be reverse-executed is @emph{not} the one pointed to by the program
6743 counter, but the instruction executed prior to that one. For instance,
6744 if the last instruction was a jump, @code{reverse-stepi} will take you
6745 back from the destination of the jump to the jump instruction itself.
6747 @kindex reverse-next
6748 @kindex rn @r{(@code{reverse-next})}
6749 @item reverse-next @r{[}@var{count}@r{]}
6750 Run backward to the beginning of the previous line executed in
6751 the current (innermost) stack frame. If the line contains function
6752 calls, they will be ``un-executed'' without stopping. Starting from
6753 the first line of a function, @code{reverse-next} will take you back
6754 to the caller of that function, @emph{before} the function was called,
6755 just as the normal @code{next} command would take you from the last
6756 line of a function back to its return to its caller
6757 @footnote{Unless the code is too heavily optimized.}.
6759 @kindex reverse-nexti
6760 @kindex rni @r{(@code{reverse-nexti})}
6761 @item reverse-nexti @r{[}@var{count}@r{]}
6762 Like @code{nexti}, @code{reverse-nexti} executes a single instruction
6763 in reverse, except that called functions are ``un-executed'' atomically.
6764 That is, if the previously executed instruction was a return from
6765 another function, @code{reverse-nexti} will continue to execute
6766 in reverse until the call to that function (from the current stack
6769 @kindex reverse-finish
6770 @item reverse-finish
6771 Just as the @code{finish} command takes you to the point where the
6772 current function returns, @code{reverse-finish} takes you to the point
6773 where it was called. Instead of ending up at the end of the current
6774 function invocation, you end up at the beginning.
6776 @kindex set exec-direction
6777 @item set exec-direction
6778 Set the direction of target execution.
6779 @item set exec-direction reverse
6780 @cindex execute forward or backward in time
6781 @value{GDBN} will perform all execution commands in reverse, until the
6782 exec-direction mode is changed to ``forward''. Affected commands include
6783 @code{step, stepi, next, nexti, continue, and finish}. The @code{return}
6784 command cannot be used in reverse mode.
6785 @item set exec-direction forward
6786 @value{GDBN} will perform all execution commands in the normal fashion.
6787 This is the default.
6791 @node Process Record and Replay
6792 @chapter Recording Inferior's Execution and Replaying It
6793 @cindex process record and replay
6794 @cindex recording inferior's execution and replaying it
6796 On some platforms, @value{GDBN} provides a special @dfn{process record
6797 and replay} target that can record a log of the process execution, and
6798 replay it later with both forward and reverse execution commands.
6801 When this target is in use, if the execution log includes the record
6802 for the next instruction, @value{GDBN} will debug in @dfn{replay
6803 mode}. In the replay mode, the inferior does not really execute code
6804 instructions. Instead, all the events that normally happen during
6805 code execution are taken from the execution log. While code is not
6806 really executed in replay mode, the values of registers (including the
6807 program counter register) and the memory of the inferior are still
6808 changed as they normally would. Their contents are taken from the
6812 If the record for the next instruction is not in the execution log,
6813 @value{GDBN} will debug in @dfn{record mode}. In this mode, the
6814 inferior executes normally, and @value{GDBN} records the execution log
6817 The process record and replay target supports reverse execution
6818 (@pxref{Reverse Execution}), even if the platform on which the
6819 inferior runs does not. However, the reverse execution is limited in
6820 this case by the range of the instructions recorded in the execution
6821 log. In other words, reverse execution on platforms that don't
6822 support it directly can only be done in the replay mode.
6824 When debugging in the reverse direction, @value{GDBN} will work in
6825 replay mode as long as the execution log includes the record for the
6826 previous instruction; otherwise, it will work in record mode, if the
6827 platform supports reverse execution, or stop if not.
6829 Currently, process record and replay is supported on ARM, Aarch64,
6830 Moxie, PowerPC, PowerPC64, S/390, and x86 (i386/amd64) running
6831 GNU/Linux. Process record and replay can be used both when native
6832 debugging, and when remote debugging via @code{gdbserver}.
6834 For architecture environments that support process record and replay,
6835 @value{GDBN} provides the following commands:
6838 @kindex target record
6839 @kindex target record-full
6840 @kindex target record-btrace
6843 @kindex record btrace
6844 @kindex record btrace bts
6845 @kindex record btrace pt
6851 @kindex rec btrace bts
6852 @kindex rec btrace pt
6855 @item record @var{method}
6856 This command starts the process record and replay target. The
6857 recording method can be specified as parameter. Without a parameter
6858 the command uses the @code{full} recording method. The following
6859 recording methods are available:
6863 Full record/replay recording using @value{GDBN}'s software record and
6864 replay implementation. This method allows replaying and reverse
6867 @item btrace @var{format}
6868 Hardware-supported instruction recording, supported on Intel
6869 processors. This method does not record data. Further, the data is
6870 collected in a ring buffer so old data will be overwritten when the
6871 buffer is full. It allows limited reverse execution. Variables and
6872 registers are not available during reverse execution. In remote
6873 debugging, recording continues on disconnect. Recorded data can be
6874 inspected after reconnecting. The recording may be stopped using
6877 The recording format can be specified as parameter. Without a parameter
6878 the command chooses the recording format. The following recording
6879 formats are available:
6883 @cindex branch trace store
6884 Use the @dfn{Branch Trace Store} (@acronym{BTS}) recording format. In
6885 this format, the processor stores a from/to record for each executed
6886 branch in the btrace ring buffer.
6889 @cindex Intel Processor Trace
6890 Use the @dfn{Intel Processor Trace} recording format. In this
6891 format, the processor stores the execution trace in a compressed form
6892 that is afterwards decoded by @value{GDBN}.
6894 The trace can be recorded with very low overhead. The compressed
6895 trace format also allows small trace buffers to already contain a big
6896 number of instructions compared to @acronym{BTS}.
6898 Decoding the recorded execution trace, on the other hand, is more
6899 expensive than decoding @acronym{BTS} trace. This is mostly due to the
6900 increased number of instructions to process. You should increase the
6901 buffer-size with care.
6904 Not all recording formats may be available on all processors.
6907 The process record and replay target can only debug a process that is
6908 already running. Therefore, you need first to start the process with
6909 the @kbd{run} or @kbd{start} commands, and then start the recording
6910 with the @kbd{record @var{method}} command.
6912 @cindex displaced stepping, and process record and replay
6913 Displaced stepping (@pxref{Maintenance Commands,, displaced stepping})
6914 will be automatically disabled when process record and replay target
6915 is started. That's because the process record and replay target
6916 doesn't support displaced stepping.
6918 @cindex non-stop mode, and process record and replay
6919 @cindex asynchronous execution, and process record and replay
6920 If the inferior is in the non-stop mode (@pxref{Non-Stop Mode}) or in
6921 the asynchronous execution mode (@pxref{Background Execution}), not
6922 all recording methods are available. The @code{full} recording method
6923 does not support these two modes.
6928 Stop the process record and replay target. When process record and
6929 replay target stops, the entire execution log will be deleted and the
6930 inferior will either be terminated, or will remain in its final state.
6932 When you stop the process record and replay target in record mode (at
6933 the end of the execution log), the inferior will be stopped at the
6934 next instruction that would have been recorded. In other words, if
6935 you record for a while and then stop recording, the inferior process
6936 will be left in the same state as if the recording never happened.
6938 On the other hand, if the process record and replay target is stopped
6939 while in replay mode (that is, not at the end of the execution log,
6940 but at some earlier point), the inferior process will become ``live''
6941 at that earlier state, and it will then be possible to continue the
6942 usual ``live'' debugging of the process from that state.
6944 When the inferior process exits, or @value{GDBN} detaches from it,
6945 process record and replay target will automatically stop itself.
6949 Go to a specific location in the execution log. There are several
6950 ways to specify the location to go to:
6953 @item record goto begin
6954 @itemx record goto start
6955 Go to the beginning of the execution log.
6957 @item record goto end
6958 Go to the end of the execution log.
6960 @item record goto @var{n}
6961 Go to instruction number @var{n} in the execution log.
6965 @item record save @var{filename}
6966 Save the execution log to a file @file{@var{filename}}.
6967 Default filename is @file{gdb_record.@var{process_id}}, where
6968 @var{process_id} is the process ID of the inferior.
6970 This command may not be available for all recording methods.
6972 @kindex record restore
6973 @item record restore @var{filename}
6974 Restore the execution log from a file @file{@var{filename}}.
6975 File must have been created with @code{record save}.
6977 @kindex set record full
6978 @item set record full insn-number-max @var{limit}
6979 @itemx set record full insn-number-max unlimited
6980 Set the limit of instructions to be recorded for the @code{full}
6981 recording method. Default value is 200000.
6983 If @var{limit} is a positive number, then @value{GDBN} will start
6984 deleting instructions from the log once the number of the record
6985 instructions becomes greater than @var{limit}. For every new recorded
6986 instruction, @value{GDBN} will delete the earliest recorded
6987 instruction to keep the number of recorded instructions at the limit.
6988 (Since deleting recorded instructions loses information, @value{GDBN}
6989 lets you control what happens when the limit is reached, by means of
6990 the @code{stop-at-limit} option, described below.)
6992 If @var{limit} is @code{unlimited} or zero, @value{GDBN} will never
6993 delete recorded instructions from the execution log. The number of
6994 recorded instructions is limited only by the available memory.
6996 @kindex show record full
6997 @item show record full insn-number-max
6998 Show the limit of instructions to be recorded with the @code{full}
7001 @item set record full stop-at-limit
7002 Control the behavior of the @code{full} recording method when the
7003 number of recorded instructions reaches the limit. If ON (the
7004 default), @value{GDBN} will stop when the limit is reached for the
7005 first time and ask you whether you want to stop the inferior or
7006 continue running it and recording the execution log. If you decide
7007 to continue recording, each new recorded instruction will cause the
7008 oldest one to be deleted.
7010 If this option is OFF, @value{GDBN} will automatically delete the
7011 oldest record to make room for each new one, without asking.
7013 @item show record full stop-at-limit
7014 Show the current setting of @code{stop-at-limit}.
7016 @item set record full memory-query
7017 Control the behavior when @value{GDBN} is unable to record memory
7018 changes caused by an instruction for the @code{full} recording method.
7019 If ON, @value{GDBN} will query whether to stop the inferior in that
7022 If this option is OFF (the default), @value{GDBN} will automatically
7023 ignore the effect of such instructions on memory. Later, when
7024 @value{GDBN} replays this execution log, it will mark the log of this
7025 instruction as not accessible, and it will not affect the replay
7028 @item show record full memory-query
7029 Show the current setting of @code{memory-query}.
7031 @kindex set record btrace
7032 The @code{btrace} record target does not trace data. As a
7033 convenience, when replaying, @value{GDBN} reads read-only memory off
7034 the live program directly, assuming that the addresses of the
7035 read-only areas don't change. This for example makes it possible to
7036 disassemble code while replaying, but not to print variables.
7037 In some cases, being able to inspect variables might be useful.
7038 You can use the following command for that:
7040 @item set record btrace replay-memory-access
7041 Control the behavior of the @code{btrace} recording method when
7042 accessing memory during replay. If @code{read-only} (the default),
7043 @value{GDBN} will only allow accesses to read-only memory.
7044 If @code{read-write}, @value{GDBN} will allow accesses to read-only
7045 and to read-write memory. Beware that the accessed memory corresponds
7046 to the live target and not necessarily to the current replay
7049 @item set record btrace cpu @var{identifier}
7050 Set the processor to be used for enabling workarounds for processor
7051 errata when decoding the trace.
7053 Processor errata are defects in processor operation, caused by its
7054 design or manufacture. They can cause a trace not to match the
7055 specification. This, in turn, may cause trace decode to fail.
7056 @value{GDBN} can detect erroneous trace packets and correct them, thus
7057 avoiding the decoding failures. These corrections are known as
7058 @dfn{errata workarounds}, and are enabled based on the processor on
7059 which the trace was recorded.
7061 By default, @value{GDBN} attempts to detect the processor
7062 automatically, and apply the necessary workarounds for it. However,
7063 you may need to specify the processor if @value{GDBN} does not yet
7064 support it. This command allows you to do that, and also allows to
7065 disable the workarounds.
7067 The argument @var{identifier} identifies the @sc{cpu} and is of the
7068 form: @code{@var{vendor}:@var{procesor identifier}}. In addition,
7069 there are two special identifiers, @code{none} and @code{auto}
7072 The following vendor identifiers and corresponding processor
7073 identifiers are currently supported:
7075 @multitable @columnfractions .1 .9
7078 @tab @var{family}/@var{model}[/@var{stepping}]
7082 On GNU/Linux systems, the processor @var{family}, @var{model}, and
7083 @var{stepping} can be obtained from @code{/proc/cpuinfo}.
7085 If @var{identifier} is @code{auto}, enable errata workarounds for the
7086 processor on which the trace was recorded. If @var{identifier} is
7087 @code{none}, errata workarounds are disabled.
7089 For example, when using an old @value{GDBN} on a new system, decode
7090 may fail because @value{GDBN} does not support the new processor. It
7091 often suffices to specify an older processor that @value{GDBN}
7096 Active record target: record-btrace
7097 Recording format: Intel Processor Trace.
7099 Failed to configure the Intel Processor Trace decoder: unknown cpu.
7100 (gdb) set record btrace cpu intel:6/158
7102 Active record target: record-btrace
7103 Recording format: Intel Processor Trace.
7105 Recorded 84872 instructions in 3189 functions (0 gaps) for thread 1 (...).
7108 @kindex show record btrace
7109 @item show record btrace replay-memory-access
7110 Show the current setting of @code{replay-memory-access}.
7112 @item show record btrace cpu
7113 Show the processor to be used for enabling trace decode errata
7116 @kindex set record btrace bts
7117 @item set record btrace bts buffer-size @var{size}
7118 @itemx set record btrace bts buffer-size unlimited
7119 Set the requested ring buffer size for branch tracing in @acronym{BTS}
7120 format. Default is 64KB.
7122 If @var{size} is a positive number, then @value{GDBN} will try to
7123 allocate a buffer of at least @var{size} bytes for each new thread
7124 that uses the btrace recording method and the @acronym{BTS} format.
7125 The actually obtained buffer size may differ from the requested
7126 @var{size}. Use the @code{info record} command to see the actual
7127 buffer size for each thread that uses the btrace recording method and
7128 the @acronym{BTS} format.
7130 If @var{limit} is @code{unlimited} or zero, @value{GDBN} will try to
7131 allocate a buffer of 4MB.
7133 Bigger buffers mean longer traces. On the other hand, @value{GDBN} will
7134 also need longer to process the branch trace data before it can be used.
7136 @item show record btrace bts buffer-size @var{size}
7137 Show the current setting of the requested ring buffer size for branch
7138 tracing in @acronym{BTS} format.
7140 @kindex set record btrace pt
7141 @item set record btrace pt buffer-size @var{size}
7142 @itemx set record btrace pt buffer-size unlimited
7143 Set the requested ring buffer size for branch tracing in Intel
7144 Processor Trace format. Default is 16KB.
7146 If @var{size} is a positive number, then @value{GDBN} will try to
7147 allocate a buffer of at least @var{size} bytes for each new thread
7148 that uses the btrace recording method and the Intel Processor Trace
7149 format. The actually obtained buffer size may differ from the
7150 requested @var{size}. Use the @code{info record} command to see the
7151 actual buffer size for each thread.
7153 If @var{limit} is @code{unlimited} or zero, @value{GDBN} will try to
7154 allocate a buffer of 4MB.
7156 Bigger buffers mean longer traces. On the other hand, @value{GDBN} will
7157 also need longer to process the branch trace data before it can be used.
7159 @item show record btrace pt buffer-size @var{size}
7160 Show the current setting of the requested ring buffer size for branch
7161 tracing in Intel Processor Trace format.
7165 Show various statistics about the recording depending on the recording
7170 For the @code{full} recording method, it shows the state of process
7171 record and its in-memory execution log buffer, including:
7175 Whether in record mode or replay mode.
7177 Lowest recorded instruction number (counting from when the current execution log started recording instructions).
7179 Highest recorded instruction number.
7181 Current instruction about to be replayed (if in replay mode).
7183 Number of instructions contained in the execution log.
7185 Maximum number of instructions that may be contained in the execution log.
7189 For the @code{btrace} recording method, it shows:
7195 Number of instructions that have been recorded.
7197 Number of blocks of sequential control-flow formed by the recorded
7200 Whether in record mode or replay mode.
7203 For the @code{bts} recording format, it also shows:
7206 Size of the perf ring buffer.
7209 For the @code{pt} recording format, it also shows:
7212 Size of the perf ring buffer.
7216 @kindex record delete
7219 When record target runs in replay mode (``in the past''), delete the
7220 subsequent execution log and begin to record a new execution log starting
7221 from the current address. This means you will abandon the previously
7222 recorded ``future'' and begin recording a new ``future''.
7224 @kindex record instruction-history
7225 @kindex rec instruction-history
7226 @item record instruction-history
7227 Disassembles instructions from the recorded execution log. By
7228 default, ten instructions are disassembled. This can be changed using
7229 the @code{set record instruction-history-size} command. Instructions
7230 are printed in execution order.
7232 It can also print mixed source+disassembly if you specify the the
7233 @code{/m} or @code{/s} modifier, and print the raw instructions in hex
7234 as well as in symbolic form by specifying the @code{/r} modifier.
7236 The current position marker is printed for the instruction at the
7237 current program counter value. This instruction can appear multiple
7238 times in the trace and the current position marker will be printed
7239 every time. To omit the current position marker, specify the
7242 To better align the printed instructions when the trace contains
7243 instructions from more than one function, the function name may be
7244 omitted by specifying the @code{/f} modifier.
7246 Speculatively executed instructions are prefixed with @samp{?}. This
7247 feature is not available for all recording formats.
7249 There are several ways to specify what part of the execution log to
7253 @item record instruction-history @var{insn}
7254 Disassembles ten instructions starting from instruction number
7257 @item record instruction-history @var{insn}, +/-@var{n}
7258 Disassembles @var{n} instructions around instruction number
7259 @var{insn}. If @var{n} is preceded with @code{+}, disassembles
7260 @var{n} instructions after instruction number @var{insn}. If
7261 @var{n} is preceded with @code{-}, disassembles @var{n}
7262 instructions before instruction number @var{insn}.
7264 @item record instruction-history
7265 Disassembles ten more instructions after the last disassembly.
7267 @item record instruction-history -
7268 Disassembles ten more instructions before the last disassembly.
7270 @item record instruction-history @var{begin}, @var{end}
7271 Disassembles instructions beginning with instruction number
7272 @var{begin} until instruction number @var{end}. The instruction
7273 number @var{end} is included.
7276 This command may not be available for all recording methods.
7279 @item set record instruction-history-size @var{size}
7280 @itemx set record instruction-history-size unlimited
7281 Define how many instructions to disassemble in the @code{record
7282 instruction-history} command. The default value is 10.
7283 A @var{size} of @code{unlimited} means unlimited instructions.
7286 @item show record instruction-history-size
7287 Show how many instructions to disassemble in the @code{record
7288 instruction-history} command.
7290 @kindex record function-call-history
7291 @kindex rec function-call-history
7292 @item record function-call-history
7293 Prints the execution history at function granularity. It prints one
7294 line for each sequence of instructions that belong to the same
7295 function giving the name of that function, the source lines
7296 for this instruction sequence (if the @code{/l} modifier is
7297 specified), and the instructions numbers that form the sequence (if
7298 the @code{/i} modifier is specified). The function names are indented
7299 to reflect the call stack depth if the @code{/c} modifier is
7300 specified. The @code{/l}, @code{/i}, and @code{/c} modifiers can be
7304 (@value{GDBP}) @b{list 1, 10}
7315 (@value{GDBP}) @b{record function-call-history /ilc}
7316 1 bar inst 1,4 at foo.c:6,8
7317 2 foo inst 5,10 at foo.c:2,3
7318 3 bar inst 11,13 at foo.c:9,10
7321 By default, ten lines are printed. This can be changed using the
7322 @code{set record function-call-history-size} command. Functions are
7323 printed in execution order. There are several ways to specify what
7327 @item record function-call-history @var{func}
7328 Prints ten functions starting from function number @var{func}.
7330 @item record function-call-history @var{func}, +/-@var{n}
7331 Prints @var{n} functions around function number @var{func}. If
7332 @var{n} is preceded with @code{+}, prints @var{n} functions after
7333 function number @var{func}. If @var{n} is preceded with @code{-},
7334 prints @var{n} functions before function number @var{func}.
7336 @item record function-call-history
7337 Prints ten more functions after the last ten-line print.
7339 @item record function-call-history -
7340 Prints ten more functions before the last ten-line print.
7342 @item record function-call-history @var{begin}, @var{end}
7343 Prints functions beginning with function number @var{begin} until
7344 function number @var{end}. The function number @var{end} is included.
7347 This command may not be available for all recording methods.
7349 @item set record function-call-history-size @var{size}
7350 @itemx set record function-call-history-size unlimited
7351 Define how many lines to print in the
7352 @code{record function-call-history} command. The default value is 10.
7353 A size of @code{unlimited} means unlimited lines.
7355 @item show record function-call-history-size
7356 Show how many lines to print in the
7357 @code{record function-call-history} command.
7362 @chapter Examining the Stack
7364 When your program has stopped, the first thing you need to know is where it
7365 stopped and how it got there.
7368 Each time your program performs a function call, information about the call
7370 That information includes the location of the call in your program,
7371 the arguments of the call,
7372 and the local variables of the function being called.
7373 The information is saved in a block of data called a @dfn{stack frame}.
7374 The stack frames are allocated in a region of memory called the @dfn{call
7377 When your program stops, the @value{GDBN} commands for examining the
7378 stack allow you to see all of this information.
7380 @cindex selected frame
7381 One of the stack frames is @dfn{selected} by @value{GDBN} and many
7382 @value{GDBN} commands refer implicitly to the selected frame. In
7383 particular, whenever you ask @value{GDBN} for the value of a variable in
7384 your program, the value is found in the selected frame. There are
7385 special @value{GDBN} commands to select whichever frame you are
7386 interested in. @xref{Selection, ,Selecting a Frame}.
7388 When your program stops, @value{GDBN} automatically selects the
7389 currently executing frame and describes it briefly, similar to the
7390 @code{frame} command (@pxref{Frame Info, ,Information about a Frame}).
7393 * Frames:: Stack frames
7394 * Backtrace:: Backtraces
7395 * Selection:: Selecting a frame
7396 * Frame Info:: Information on a frame
7397 * Frame Apply:: Applying a command to several frames
7398 * Frame Filter Management:: Managing frame filters
7403 @section Stack Frames
7405 @cindex frame, definition
7407 The call stack is divided up into contiguous pieces called @dfn{stack
7408 frames}, or @dfn{frames} for short; each frame is the data associated
7409 with one call to one function. The frame contains the arguments given
7410 to the function, the function's local variables, and the address at
7411 which the function is executing.
7413 @cindex initial frame
7414 @cindex outermost frame
7415 @cindex innermost frame
7416 When your program is started, the stack has only one frame, that of the
7417 function @code{main}. This is called the @dfn{initial} frame or the
7418 @dfn{outermost} frame. Each time a function is called, a new frame is
7419 made. Each time a function returns, the frame for that function invocation
7420 is eliminated. If a function is recursive, there can be many frames for
7421 the same function. The frame for the function in which execution is
7422 actually occurring is called the @dfn{innermost} frame. This is the most
7423 recently created of all the stack frames that still exist.
7425 @cindex frame pointer
7426 Inside your program, stack frames are identified by their addresses. A
7427 stack frame consists of many bytes, each of which has its own address; each
7428 kind of computer has a convention for choosing one byte whose
7429 address serves as the address of the frame. Usually this address is kept
7430 in a register called the @dfn{frame pointer register}
7431 (@pxref{Registers, $fp}) while execution is going on in that frame.
7434 @cindex frame number
7435 @value{GDBN} labels each existing stack frame with a @dfn{level}, a
7436 number that is zero for the innermost frame, one for the frame that
7437 called it, and so on upward. These level numbers give you a way of
7438 designating stack frames in @value{GDBN} commands. The terms
7439 @dfn{frame number} and @dfn{frame level} can be used interchangeably to
7440 describe this number.
7442 @c The -fomit-frame-pointer below perennially causes hbox overflow
7443 @c underflow problems.
7444 @cindex frameless execution
7445 Some compilers provide a way to compile functions so that they operate
7446 without stack frames. (For example, the @value{NGCC} option
7448 @samp{-fomit-frame-pointer}
7450 generates functions without a frame.)
7451 This is occasionally done with heavily used library functions to save
7452 the frame setup time. @value{GDBN} has limited facilities for dealing
7453 with these function invocations. If the innermost function invocation
7454 has no stack frame, @value{GDBN} nevertheless regards it as though
7455 it had a separate frame, which is numbered zero as usual, allowing
7456 correct tracing of the function call chain. However, @value{GDBN} has
7457 no provision for frameless functions elsewhere in the stack.
7463 @cindex call stack traces
7464 A backtrace is a summary of how your program got where it is. It shows one
7465 line per frame, for many frames, starting with the currently executing
7466 frame (frame zero), followed by its caller (frame one), and on up the
7469 @anchor{backtrace-command}
7471 @kindex bt @r{(@code{backtrace})}
7472 To print a backtrace of the entire stack, use the @code{backtrace}
7473 command, or its alias @code{bt}. This command will print one line per
7474 frame for frames in the stack. By default, all stack frames are
7475 printed. You can stop the backtrace at any time by typing the system
7476 interrupt character, normally @kbd{Ctrl-c}.
7479 @item backtrace [@var{args}@dots{}]
7480 @itemx bt [@var{args}@dots{}]
7481 Print the backtrace of the entire stack. The optional @var{args} can
7482 be one of the following:
7487 Print only the innermost @var{n} frames, where @var{n} is a positive
7492 Print only the outermost @var{n} frames, where @var{n} is a positive
7496 Print the values of the local variables also. This can be combined
7497 with a number to limit the number of frames shown.
7500 Do not run Python frame filters on this backtrace. @xref{Frame
7501 Filter API}, for more information. Additionally use @ref{disable
7502 frame-filter all} to turn off all frame filters. This is only
7503 relevant when @value{GDBN} has been configured with @code{Python}
7507 A Python frame filter might decide to ``elide'' some frames. Normally
7508 such elided frames are still printed, but they are indented relative
7509 to the filtered frames that cause them to be elided. The @code{hide}
7510 option causes elided frames to not be printed at all.
7516 The names @code{where} and @code{info stack} (abbreviated @code{info s})
7517 are additional aliases for @code{backtrace}.
7519 @cindex multiple threads, backtrace
7520 In a multi-threaded program, @value{GDBN} by default shows the
7521 backtrace only for the current thread. To display the backtrace for
7522 several or all of the threads, use the command @code{thread apply}
7523 (@pxref{Threads, thread apply}). For example, if you type @kbd{thread
7524 apply all backtrace}, @value{GDBN} will display the backtrace for all
7525 the threads; this is handy when you debug a core dump of a
7526 multi-threaded program.
7528 Each line in the backtrace shows the frame number and the function name.
7529 The program counter value is also shown---unless you use @code{set
7530 print address off}. The backtrace also shows the source file name and
7531 line number, as well as the arguments to the function. The program
7532 counter value is omitted if it is at the beginning of the code for that
7535 Here is an example of a backtrace. It was made with the command
7536 @samp{bt 3}, so it shows the innermost three frames.
7540 #0 m4_traceon (obs=0x24eb0, argc=1, argv=0x2b8c8)
7542 #1 0x6e38 in expand_macro (sym=0x2b600, data=...) at macro.c:242
7543 #2 0x6840 in expand_token (obs=0x0, t=177664, td=0xf7fffb08)
7545 (More stack frames follow...)
7550 The display for frame zero does not begin with a program counter
7551 value, indicating that your program has stopped at the beginning of the
7552 code for line @code{993} of @code{builtin.c}.
7555 The value of parameter @code{data} in frame 1 has been replaced by
7556 @code{@dots{}}. By default, @value{GDBN} prints the value of a parameter
7557 only if it is a scalar (integer, pointer, enumeration, etc). See command
7558 @kbd{set print frame-arguments} in @ref{Print Settings} for more details
7559 on how to configure the way function parameter values are printed.
7561 @cindex optimized out, in backtrace
7562 @cindex function call arguments, optimized out
7563 If your program was compiled with optimizations, some compilers will
7564 optimize away arguments passed to functions if those arguments are
7565 never used after the call. Such optimizations generate code that
7566 passes arguments through registers, but doesn't store those arguments
7567 in the stack frame. @value{GDBN} has no way of displaying such
7568 arguments in stack frames other than the innermost one. Here's what
7569 such a backtrace might look like:
7573 #0 m4_traceon (obs=0x24eb0, argc=1, argv=0x2b8c8)
7575 #1 0x6e38 in expand_macro (sym=<optimized out>) at macro.c:242
7576 #2 0x6840 in expand_token (obs=0x0, t=<optimized out>, td=0xf7fffb08)
7578 (More stack frames follow...)
7583 The values of arguments that were not saved in their stack frames are
7584 shown as @samp{<optimized out>}.
7586 If you need to display the values of such optimized-out arguments,
7587 either deduce that from other variables whose values depend on the one
7588 you are interested in, or recompile without optimizations.
7590 @cindex backtrace beyond @code{main} function
7591 @cindex program entry point
7592 @cindex startup code, and backtrace
7593 Most programs have a standard user entry point---a place where system
7594 libraries and startup code transition into user code. For C this is
7595 @code{main}@footnote{
7596 Note that embedded programs (the so-called ``free-standing''
7597 environment) are not required to have a @code{main} function as the
7598 entry point. They could even have multiple entry points.}.
7599 When @value{GDBN} finds the entry function in a backtrace
7600 it will terminate the backtrace, to avoid tracing into highly
7601 system-specific (and generally uninteresting) code.
7603 If you need to examine the startup code, or limit the number of levels
7604 in a backtrace, you can change this behavior:
7607 @item set backtrace past-main
7608 @itemx set backtrace past-main on
7609 @kindex set backtrace
7610 Backtraces will continue past the user entry point.
7612 @item set backtrace past-main off
7613 Backtraces will stop when they encounter the user entry point. This is the
7616 @item show backtrace past-main
7617 @kindex show backtrace
7618 Display the current user entry point backtrace policy.
7620 @item set backtrace past-entry
7621 @itemx set backtrace past-entry on
7622 Backtraces will continue past the internal entry point of an application.
7623 This entry point is encoded by the linker when the application is built,
7624 and is likely before the user entry point @code{main} (or equivalent) is called.
7626 @item set backtrace past-entry off
7627 Backtraces will stop when they encounter the internal entry point of an
7628 application. This is the default.
7630 @item show backtrace past-entry
7631 Display the current internal entry point backtrace policy.
7633 @item set backtrace limit @var{n}
7634 @itemx set backtrace limit 0
7635 @itemx set backtrace limit unlimited
7636 @cindex backtrace limit
7637 Limit the backtrace to @var{n} levels. A value of @code{unlimited}
7638 or zero means unlimited levels.
7640 @item show backtrace limit
7641 Display the current limit on backtrace levels.
7644 You can control how file names are displayed.
7647 @item set filename-display
7648 @itemx set filename-display relative
7649 @cindex filename-display
7650 Display file names relative to the compilation directory. This is the default.
7652 @item set filename-display basename
7653 Display only basename of a filename.
7655 @item set filename-display absolute
7656 Display an absolute filename.
7658 @item show filename-display
7659 Show the current way to display filenames.
7663 @section Selecting a Frame
7665 Most commands for examining the stack and other data in your program work on
7666 whichever stack frame is selected at the moment. Here are the commands for
7667 selecting a stack frame; all of them finish by printing a brief description
7668 of the stack frame just selected.
7671 @kindex frame@r{, selecting}
7672 @kindex f @r{(@code{frame})}
7673 @item frame @r{[} @var{frame-selection-spec} @r{]}
7674 @item f @r{[} @var{frame-selection-spec} @r{]}
7675 The @command{frame} command allows different stack frames to be
7676 selected. The @var{frame-selection-spec} can be any of the following:
7681 @item level @var{num}
7682 Select frame level @var{num}. Recall that frame zero is the innermost
7683 (currently executing) frame, frame one is the frame that called the
7684 innermost one, and so on. The highest level frame is usually the one
7687 As this is the most common method of navigating the frame stack, the
7688 string @command{level} can be omitted. For example, the following two
7689 commands are equivalent:
7692 (@value{GDBP}) frame 3
7693 (@value{GDBP}) frame level 3
7696 @kindex frame address
7697 @item address @var{stack-address}
7698 Select the frame with stack address @var{stack-address}. The
7699 @var{stack-address} for a frame can be seen in the output of
7700 @command{info frame}, for example:
7704 Stack level 1, frame at 0x7fffffffda30:
7705 rip = 0x40066d in b (amd64-entry-value.cc:59); saved rip 0x4004c5
7706 tail call frame, caller of frame at 0x7fffffffda30
7707 source language c++.
7708 Arglist at unknown address.
7709 Locals at unknown address, Previous frame's sp is 0x7fffffffda30
7712 The @var{stack-address} for this frame is @code{0x7fffffffda30} as
7713 indicated by the line:
7716 Stack level 1, frame at 0x7fffffffda30:
7719 @kindex frame function
7720 @item function @var{function-name}
7721 Select the stack frame for function @var{function-name}. If there are
7722 multiple stack frames for function @var{function-name} then the inner
7723 most stack frame is selected.
7726 @item view @var{stack-address} @r{[} @var{pc-addr} @r{]}
7727 View a frame that is not part of @value{GDBN}'s backtrace. The frame
7728 viewed has stack address @var{stack-addr}, and optionally, a program
7729 counter address of @var{pc-addr}.
7731 This is useful mainly if the chaining of stack frames has been
7732 damaged by a bug, making it impossible for @value{GDBN} to assign
7733 numbers properly to all frames. In addition, this can be useful
7734 when your program has multiple stacks and switches between them.
7736 When viewing a frame outside the current backtrace using
7737 @command{frame view} then you can always return to the original
7738 stack using one of the previous stack frame selection instructions,
7739 for example @command{frame level 0}.
7745 Move @var{n} frames up the stack; @var{n} defaults to 1. For positive
7746 numbers @var{n}, this advances toward the outermost frame, to higher
7747 frame numbers, to frames that have existed longer.
7750 @kindex do @r{(@code{down})}
7752 Move @var{n} frames down the stack; @var{n} defaults to 1. For
7753 positive numbers @var{n}, this advances toward the innermost frame, to
7754 lower frame numbers, to frames that were created more recently.
7755 You may abbreviate @code{down} as @code{do}.
7758 All of these commands end by printing two lines of output describing the
7759 frame. The first line shows the frame number, the function name, the
7760 arguments, and the source file and line number of execution in that
7761 frame. The second line shows the text of that source line.
7769 #1 0x22f0 in main (argc=1, argv=0xf7fffbf4, env=0xf7fffbfc)
7771 10 read_input_file (argv[i]);
7775 After such a printout, the @code{list} command with no arguments
7776 prints ten lines centered on the point of execution in the frame.
7777 You can also edit the program at the point of execution with your favorite
7778 editing program by typing @code{edit}.
7779 @xref{List, ,Printing Source Lines},
7783 @kindex select-frame
7784 @item select-frame @r{[} @var{frame-selection-spec} @r{]}
7785 The @code{select-frame} command is a variant of @code{frame} that does
7786 not display the new frame after selecting it. This command is
7787 intended primarily for use in @value{GDBN} command scripts, where the
7788 output might be unnecessary and distracting. The
7789 @var{frame-selection-spec} is as for the @command{frame} command
7790 described in @ref{Selection, ,Selecting a Frame}.
7792 @kindex down-silently
7794 @item up-silently @var{n}
7795 @itemx down-silently @var{n}
7796 These two commands are variants of @code{up} and @code{down},
7797 respectively; they differ in that they do their work silently, without
7798 causing display of the new frame. They are intended primarily for use
7799 in @value{GDBN} command scripts, where the output might be unnecessary and
7804 @section Information About a Frame
7806 There are several other commands to print information about the selected
7812 When used without any argument, this command does not change which
7813 frame is selected, but prints a brief description of the currently
7814 selected stack frame. It can be abbreviated @code{f}. With an
7815 argument, this command is used to select a stack frame.
7816 @xref{Selection, ,Selecting a Frame}.
7819 @kindex info f @r{(@code{info frame})}
7822 This command prints a verbose description of the selected stack frame,
7827 the address of the frame
7829 the address of the next frame down (called by this frame)
7831 the address of the next frame up (caller of this frame)
7833 the language in which the source code corresponding to this frame is written
7835 the address of the frame's arguments
7837 the address of the frame's local variables
7839 the program counter saved in it (the address of execution in the caller frame)
7841 which registers were saved in the frame
7844 @noindent The verbose description is useful when
7845 something has gone wrong that has made the stack format fail to fit
7846 the usual conventions.
7848 @item info frame @r{[} @var{frame-selection-spec} @r{]}
7849 @itemx info f @r{[} @var{frame-selection-spec} @r{]}
7850 Print a verbose description of the frame selected by
7851 @var{frame-selection-spec}. The @var{frame-selection-spec} is the
7852 same as for the @command{frame} command (@pxref{Selection, ,Selecting
7853 a Frame}). The selected frame remains unchanged by this command.
7856 @item info args [-q]
7857 Print the arguments of the selected frame, each on a separate line.
7859 The optional flag @samp{-q}, which stands for @samp{quiet}, disables
7860 printing header information and messages explaining why no argument
7863 @item info args [-q] [-t @var{type_regexp}] [@var{regexp}]
7864 Like @kbd{info args}, but only print the arguments selected
7865 with the provided regexp(s).
7867 If @var{regexp} is provided, print only the arguments whose names
7868 match the regular expression @var{regexp}.
7870 If @var{type_regexp} is provided, print only the arguments whose
7871 types, as printed by the @code{whatis} command, match
7872 the regular expression @var{type_regexp}.
7873 If @var{type_regexp} contains space(s), it should be enclosed in
7874 quote characters. If needed, use backslash to escape the meaning
7875 of special characters or quotes.
7877 If both @var{regexp} and @var{type_regexp} are provided, an argument
7878 is printed only if its name matches @var{regexp} and its type matches
7881 @item info locals [-q]
7883 Print the local variables of the selected frame, each on a separate
7884 line. These are all variables (declared either static or automatic)
7885 accessible at the point of execution of the selected frame.
7887 The optional flag @samp{-q}, which stands for @samp{quiet}, disables
7888 printing header information and messages explaining why no local variables
7891 @item info locals [-q] [-t @var{type_regexp}] [@var{regexp}]
7892 Like @kbd{info locals}, but only print the local variables selected
7893 with the provided regexp(s).
7895 If @var{regexp} is provided, print only the local variables whose names
7896 match the regular expression @var{regexp}.
7898 If @var{type_regexp} is provided, print only the local variables whose
7899 types, as printed by the @code{whatis} command, match
7900 the regular expression @var{type_regexp}.
7901 If @var{type_regexp} contains space(s), it should be enclosed in
7902 quote characters. If needed, use backslash to escape the meaning
7903 of special characters or quotes.
7905 If both @var{regexp} and @var{type_regexp} are provided, a local variable
7906 is printed only if its name matches @var{regexp} and its type matches
7909 The command @kbd{info locals -q -t @var{type_regexp}} can usefully be
7910 combined with the commands @kbd{frame apply} and @kbd{thread apply}.
7911 For example, your program might use Resource Acquisition Is
7912 Initialization types (RAII) such as @code{lock_something_t}: each
7913 local variable of type @code{lock_something_t} automatically places a
7914 lock that is destroyed when the variable goes out of scope. You can
7915 then list all acquired locks in your program by doing
7917 thread apply all -s frame apply all -s info locals -q -t lock_something_t
7920 or the equivalent shorter form
7922 tfaas i lo -q -t lock_something_t
7928 @section Applying a Command to Several Frames.
7930 @cindex apply command to several frames
7932 @item frame apply [all | @var{count} | @var{-count} | level @var{level}@dots{}] [@var{flag}]@dots{} @var{command}
7933 The @code{frame apply} command allows you to apply the named
7934 @var{command} to one or more frames.
7938 Specify @code{all} to apply @var{command} to all frames.
7941 Use @var{count} to apply @var{command} to the innermost @var{count}
7942 frames, where @var{count} is a positive number.
7945 Use @var{-count} to apply @var{command} to the outermost @var{count}
7946 frames, where @var{count} is a positive number.
7949 Use @code{level} to apply @var{command} to the set of frames identified
7950 by the @var{level} list. @var{level} is a frame level or a range of frame
7951 levels as @var{level1}-@var{level2}. The frame level is the number shown
7952 in the first field of the @samp{backtrace} command output.
7953 E.g., @samp{2-4 6-8 3} indicates to apply @var{command} for the frames
7954 at levels 2, 3, 4, 6, 7, 8, and then again on frame at level 3.
7960 Note that the frames on which @code{frame apply} applies a command are
7961 also influenced by the @code{set backtrace} settings such as @code{set
7962 backtrace past-main} and @code{set backtrace limit N}. See
7963 @xref{Backtrace,,Backtraces}.
7965 The @var{flag} arguments control what output to produce and how to handle
7966 errors raised when applying @var{command} to a frame. @var{flag}
7967 must start with a @code{-} directly followed by one letter in
7968 @code{qcs}. If several flags are provided, they must be given
7969 individually, such as @code{-c -q}.
7971 By default, @value{GDBN} displays some frame information before the
7972 output produced by @var{command}, and an error raised during the
7973 execution of a @var{command} will abort @code{frame apply}. The
7974 following flags can be used to fine-tune this behavior:
7978 The flag @code{-c}, which stands for @samp{continue}, causes any
7979 errors in @var{command} to be displayed, and the execution of
7980 @code{frame apply} then continues.
7982 The flag @code{-s}, which stands for @samp{silent}, causes any errors
7983 or empty output produced by a @var{command} to be silently ignored.
7984 That is, the execution continues, but the frame information and errors
7987 The flag @code{-q} (@samp{quiet}) disables printing the frame
7991 The following example shows how the flags @code{-c} and @code{-s} are
7992 working when applying the command @code{p j} to all frames, where
7993 variable @code{j} can only be successfully printed in the outermost
7994 @code{#1 main} frame.
7998 (gdb) frame apply all p j
7999 #0 some_function (i=5) at fun.c:4
8000 No symbol "j" in current context.
8001 (gdb) frame apply all -c p j
8002 #0 some_function (i=5) at fun.c:4
8003 No symbol "j" in current context.
8004 #1 0x565555fb in main (argc=1, argv=0xffffd2c4) at fun.c:11
8006 (gdb) frame apply all -s p j
8007 #1 0x565555fb in main (argc=1, argv=0xffffd2c4) at fun.c:11
8013 By default, @samp{frame apply}, prints the frame location
8014 information before the command output:
8018 (gdb) frame apply all p $sp
8019 #0 some_function (i=5) at fun.c:4
8020 $4 = (void *) 0xffffd1e0
8021 #1 0x565555fb in main (argc=1, argv=0xffffd2c4) at fun.c:11
8022 $5 = (void *) 0xffffd1f0
8027 If flag @code{-q} is given, no frame information is printed:
8030 (gdb) frame apply all -q p $sp
8031 $12 = (void *) 0xffffd1e0
8032 $13 = (void *) 0xffffd1f0
8040 @cindex apply a command to all frames (ignoring errors and empty output)
8041 @item faas @var{command}
8042 Shortcut for @code{frame apply all -s @var{command}}.
8043 Applies @var{command} on all frames, ignoring errors and empty output.
8045 It can for example be used to print a local variable or a function
8046 argument without knowing the frame where this variable or argument
8049 (@value{GDBP}) faas p some_local_var_i_do_not_remember_where_it_is
8052 Note that the command @code{tfaas @var{command}} applies @var{command}
8053 on all frames of all threads. See @xref{Threads,,Threads}.
8057 @node Frame Filter Management
8058 @section Management of Frame Filters.
8059 @cindex managing frame filters
8061 Frame filters are Python based utilities to manage and decorate the
8062 output of frames. @xref{Frame Filter API}, for further information.
8064 Managing frame filters is performed by several commands available
8065 within @value{GDBN}, detailed here.
8068 @kindex info frame-filter
8069 @item info frame-filter
8070 Print a list of installed frame filters from all dictionaries, showing
8071 their name, priority and enabled status.
8073 @kindex disable frame-filter
8074 @anchor{disable frame-filter all}
8075 @item disable frame-filter @var{filter-dictionary} @var{filter-name}
8076 Disable a frame filter in the dictionary matching
8077 @var{filter-dictionary} and @var{filter-name}. The
8078 @var{filter-dictionary} may be @code{all}, @code{global},
8079 @code{progspace}, or the name of the object file where the frame filter
8080 dictionary resides. When @code{all} is specified, all frame filters
8081 across all dictionaries are disabled. The @var{filter-name} is the name
8082 of the frame filter and is used when @code{all} is not the option for
8083 @var{filter-dictionary}. A disabled frame-filter is not deleted, it
8084 may be enabled again later.
8086 @kindex enable frame-filter
8087 @item enable frame-filter @var{filter-dictionary} @var{filter-name}
8088 Enable a frame filter in the dictionary matching
8089 @var{filter-dictionary} and @var{filter-name}. The
8090 @var{filter-dictionary} may be @code{all}, @code{global},
8091 @code{progspace} or the name of the object file where the frame filter
8092 dictionary resides. When @code{all} is specified, all frame filters across
8093 all dictionaries are enabled. The @var{filter-name} is the name of the frame
8094 filter and is used when @code{all} is not the option for
8095 @var{filter-dictionary}.
8100 (gdb) info frame-filter
8102 global frame-filters:
8103 Priority Enabled Name
8104 1000 No PrimaryFunctionFilter
8107 progspace /build/test frame-filters:
8108 Priority Enabled Name
8109 100 Yes ProgspaceFilter
8111 objfile /build/test frame-filters:
8112 Priority Enabled Name
8113 999 Yes BuildProgra Filter
8115 (gdb) disable frame-filter /build/test BuildProgramFilter
8116 (gdb) info frame-filter
8118 global frame-filters:
8119 Priority Enabled Name
8120 1000 No PrimaryFunctionFilter
8123 progspace /build/test frame-filters:
8124 Priority Enabled Name
8125 100 Yes ProgspaceFilter
8127 objfile /build/test frame-filters:
8128 Priority Enabled Name
8129 999 No BuildProgramFilter
8131 (gdb) enable frame-filter global PrimaryFunctionFilter
8132 (gdb) info frame-filter
8134 global frame-filters:
8135 Priority Enabled Name
8136 1000 Yes PrimaryFunctionFilter
8139 progspace /build/test frame-filters:
8140 Priority Enabled Name
8141 100 Yes ProgspaceFilter
8143 objfile /build/test frame-filters:
8144 Priority Enabled Name
8145 999 No BuildProgramFilter
8148 @kindex set frame-filter priority
8149 @item set frame-filter priority @var{filter-dictionary} @var{filter-name} @var{priority}
8150 Set the @var{priority} of a frame filter in the dictionary matching
8151 @var{filter-dictionary}, and the frame filter name matching
8152 @var{filter-name}. The @var{filter-dictionary} may be @code{global},
8153 @code{progspace} or the name of the object file where the frame filter
8154 dictionary resides. The @var{priority} is an integer.
8156 @kindex show frame-filter priority
8157 @item show frame-filter priority @var{filter-dictionary} @var{filter-name}
8158 Show the @var{priority} of a frame filter in the dictionary matching
8159 @var{filter-dictionary}, and the frame filter name matching
8160 @var{filter-name}. The @var{filter-dictionary} may be @code{global},
8161 @code{progspace} or the name of the object file where the frame filter
8167 (gdb) info frame-filter
8169 global frame-filters:
8170 Priority Enabled Name
8171 1000 Yes PrimaryFunctionFilter
8174 progspace /build/test frame-filters:
8175 Priority Enabled Name
8176 100 Yes ProgspaceFilter
8178 objfile /build/test frame-filters:
8179 Priority Enabled Name
8180 999 No BuildProgramFilter
8182 (gdb) set frame-filter priority global Reverse 50
8183 (gdb) info frame-filter
8185 global frame-filters:
8186 Priority Enabled Name
8187 1000 Yes PrimaryFunctionFilter
8190 progspace /build/test frame-filters:
8191 Priority Enabled Name
8192 100 Yes ProgspaceFilter
8194 objfile /build/test frame-filters:
8195 Priority Enabled Name
8196 999 No BuildProgramFilter
8201 @chapter Examining Source Files
8203 @value{GDBN} can print parts of your program's source, since the debugging
8204 information recorded in the program tells @value{GDBN} what source files were
8205 used to build it. When your program stops, @value{GDBN} spontaneously prints
8206 the line where it stopped. Likewise, when you select a stack frame
8207 (@pxref{Selection, ,Selecting a Frame}), @value{GDBN} prints the line where
8208 execution in that frame has stopped. You can print other portions of
8209 source files by explicit command.
8211 If you use @value{GDBN} through its @sc{gnu} Emacs interface, you may
8212 prefer to use Emacs facilities to view source; see @ref{Emacs, ,Using
8213 @value{GDBN} under @sc{gnu} Emacs}.
8216 * List:: Printing source lines
8217 * Specify Location:: How to specify code locations
8218 * Edit:: Editing source files
8219 * Search:: Searching source files
8220 * Source Path:: Specifying source directories
8221 * Machine Code:: Source and machine code
8225 @section Printing Source Lines
8228 @kindex l @r{(@code{list})}
8229 To print lines from a source file, use the @code{list} command
8230 (abbreviated @code{l}). By default, ten lines are printed.
8231 There are several ways to specify what part of the file you want to
8232 print; see @ref{Specify Location}, for the full list.
8234 Here are the forms of the @code{list} command most commonly used:
8237 @item list @var{linenum}
8238 Print lines centered around line number @var{linenum} in the
8239 current source file.
8241 @item list @var{function}
8242 Print lines centered around the beginning of function
8246 Print more lines. If the last lines printed were printed with a
8247 @code{list} command, this prints lines following the last lines
8248 printed; however, if the last line printed was a solitary line printed
8249 as part of displaying a stack frame (@pxref{Stack, ,Examining the
8250 Stack}), this prints lines centered around that line.
8253 Print lines just before the lines last printed.
8256 @cindex @code{list}, how many lines to display
8257 By default, @value{GDBN} prints ten source lines with any of these forms of
8258 the @code{list} command. You can change this using @code{set listsize}:
8261 @kindex set listsize
8262 @item set listsize @var{count}
8263 @itemx set listsize unlimited
8264 Make the @code{list} command display @var{count} source lines (unless
8265 the @code{list} argument explicitly specifies some other number).
8266 Setting @var{count} to @code{unlimited} or 0 means there's no limit.
8268 @kindex show listsize
8270 Display the number of lines that @code{list} prints.
8273 Repeating a @code{list} command with @key{RET} discards the argument,
8274 so it is equivalent to typing just @code{list}. This is more useful
8275 than listing the same lines again. An exception is made for an
8276 argument of @samp{-}; that argument is preserved in repetition so that
8277 each repetition moves up in the source file.
8279 In general, the @code{list} command expects you to supply zero, one or two
8280 @dfn{locations}. Locations specify source lines; there are several ways
8281 of writing them (@pxref{Specify Location}), but the effect is always
8282 to specify some source line.
8284 Here is a complete description of the possible arguments for @code{list}:
8287 @item list @var{location}
8288 Print lines centered around the line specified by @var{location}.
8290 @item list @var{first},@var{last}
8291 Print lines from @var{first} to @var{last}. Both arguments are
8292 locations. When a @code{list} command has two locations, and the
8293 source file of the second location is omitted, this refers to
8294 the same source file as the first location.
8296 @item list ,@var{last}
8297 Print lines ending with @var{last}.
8299 @item list @var{first},
8300 Print lines starting with @var{first}.
8303 Print lines just after the lines last printed.
8306 Print lines just before the lines last printed.
8309 As described in the preceding table.
8312 @node Specify Location
8313 @section Specifying a Location
8314 @cindex specifying location
8316 @cindex source location
8319 * Linespec Locations:: Linespec locations
8320 * Explicit Locations:: Explicit locations
8321 * Address Locations:: Address locations
8324 Several @value{GDBN} commands accept arguments that specify a location
8325 of your program's code. Since @value{GDBN} is a source-level
8326 debugger, a location usually specifies some line in the source code.
8327 Locations may be specified using three different formats:
8328 linespec locations, explicit locations, or address locations.
8330 @node Linespec Locations
8331 @subsection Linespec Locations
8332 @cindex linespec locations
8334 A @dfn{linespec} is a colon-separated list of source location parameters such
8335 as file name, function name, etc. Here are all the different ways of
8336 specifying a linespec:
8340 Specifies the line number @var{linenum} of the current source file.
8343 @itemx +@var{offset}
8344 Specifies the line @var{offset} lines before or after the @dfn{current
8345 line}. For the @code{list} command, the current line is the last one
8346 printed; for the breakpoint commands, this is the line at which
8347 execution stopped in the currently selected @dfn{stack frame}
8348 (@pxref{Frames, ,Frames}, for a description of stack frames.) When
8349 used as the second of the two linespecs in a @code{list} command,
8350 this specifies the line @var{offset} lines up or down from the first
8353 @item @var{filename}:@var{linenum}
8354 Specifies the line @var{linenum} in the source file @var{filename}.
8355 If @var{filename} is a relative file name, then it will match any
8356 source file name with the same trailing components. For example, if
8357 @var{filename} is @samp{gcc/expr.c}, then it will match source file
8358 name of @file{/build/trunk/gcc/expr.c}, but not
8359 @file{/build/trunk/libcpp/expr.c} or @file{/build/trunk/gcc/x-expr.c}.
8361 @item @var{function}
8362 Specifies the line that begins the body of the function @var{function}.
8363 For example, in C, this is the line with the open brace.
8365 By default, in C@t{++} and Ada, @var{function} is interpreted as
8366 specifying all functions named @var{function} in all scopes. For
8367 C@t{++}, this means in all namespaces and classes. For Ada, this
8368 means in all packages.
8370 For example, assuming a program with C@t{++} symbols named
8371 @code{A::B::func} and @code{B::func}, both commands @w{@kbd{break
8372 func}} and @w{@kbd{break B::func}} set a breakpoint on both symbols.
8374 Commands that accept a linespec let you override this with the
8375 @code{-qualified} option. For example, @w{@kbd{break -qualified
8376 func}} sets a breakpoint on a free-function named @code{func} ignoring
8377 any C@t{++} class methods and namespace functions called @code{func}.
8379 @xref{Explicit Locations}.
8381 @item @var{function}:@var{label}
8382 Specifies the line where @var{label} appears in @var{function}.
8384 @item @var{filename}:@var{function}
8385 Specifies the line that begins the body of the function @var{function}
8386 in the file @var{filename}. You only need the file name with a
8387 function name to avoid ambiguity when there are identically named
8388 functions in different source files.
8391 Specifies the line at which the label named @var{label} appears
8392 in the function corresponding to the currently selected stack frame.
8393 If there is no current selected stack frame (for instance, if the inferior
8394 is not running), then @value{GDBN} will not search for a label.
8396 @cindex breakpoint at static probe point
8397 @item -pstap|-probe-stap @r{[}@var{objfile}:@r{[}@var{provider}:@r{]}@r{]}@var{name}
8398 The @sc{gnu}/Linux tool @code{SystemTap} provides a way for
8399 applications to embed static probes. @xref{Static Probe Points}, for more
8400 information on finding and using static probes. This form of linespec
8401 specifies the location of such a static probe.
8403 If @var{objfile} is given, only probes coming from that shared library
8404 or executable matching @var{objfile} as a regular expression are considered.
8405 If @var{provider} is given, then only probes from that provider are considered.
8406 If several probes match the spec, @value{GDBN} will insert a breakpoint at
8407 each one of those probes.
8410 @node Explicit Locations
8411 @subsection Explicit Locations
8412 @cindex explicit locations
8414 @dfn{Explicit locations} allow the user to directly specify the source
8415 location's parameters using option-value pairs.
8417 Explicit locations are useful when several functions, labels, or
8418 file names have the same name (base name for files) in the program's
8419 sources. In these cases, explicit locations point to the source
8420 line you meant more accurately and unambiguously. Also, using
8421 explicit locations might be faster in large programs.
8423 For example, the linespec @samp{foo:bar} may refer to a function @code{bar}
8424 defined in the file named @file{foo} or the label @code{bar} in a function
8425 named @code{foo}. @value{GDBN} must search either the file system or
8426 the symbol table to know.
8428 The list of valid explicit location options is summarized in the
8432 @item -source @var{filename}
8433 The value specifies the source file name. To differentiate between
8434 files with the same base name, prepend as many directories as is necessary
8435 to uniquely identify the desired file, e.g., @file{foo/bar/baz.c}. Otherwise
8436 @value{GDBN} will use the first file it finds with the given base
8437 name. This option requires the use of either @code{-function} or @code{-line}.
8439 @item -function @var{function}
8440 The value specifies the name of a function. Operations
8441 on function locations unmodified by other options (such as @code{-label}
8442 or @code{-line}) refer to the line that begins the body of the function.
8443 In C, for example, this is the line with the open brace.
8445 By default, in C@t{++} and Ada, @var{function} is interpreted as
8446 specifying all functions named @var{function} in all scopes. For
8447 C@t{++}, this means in all namespaces and classes. For Ada, this
8448 means in all packages.
8450 For example, assuming a program with C@t{++} symbols named
8451 @code{A::B::func} and @code{B::func}, both commands @w{@kbd{break
8452 -function func}} and @w{@kbd{break -function B::func}} set a
8453 breakpoint on both symbols.
8455 You can use the @kbd{-qualified} flag to override this (see below).
8459 This flag makes @value{GDBN} interpret a function name specified with
8460 @kbd{-function} as a complete fully-qualified name.
8462 For example, assuming a C@t{++} program with symbols named
8463 @code{A::B::func} and @code{B::func}, the @w{@kbd{break -qualified
8464 -function B::func}} command sets a breakpoint on @code{B::func}, only.
8466 (Note: the @kbd{-qualified} option can precede a linespec as well
8467 (@pxref{Linespec Locations}), so the particular example above could be
8468 simplified as @w{@kbd{break -qualified B::func}}.)
8470 @item -label @var{label}
8471 The value specifies the name of a label. When the function
8472 name is not specified, the label is searched in the function of the currently
8473 selected stack frame.
8475 @item -line @var{number}
8476 The value specifies a line offset for the location. The offset may either
8477 be absolute (@code{-line 3}) or relative (@code{-line +3}), depending on
8478 the command. When specified without any other options, the line offset is
8479 relative to the current line.
8482 Explicit location options may be abbreviated by omitting any non-unique
8483 trailing characters from the option name, e.g., @w{@kbd{break -s main.c -li 3}}.
8485 @node Address Locations
8486 @subsection Address Locations
8487 @cindex address locations
8489 @dfn{Address locations} indicate a specific program address. They have
8490 the generalized form *@var{address}.
8492 For line-oriented commands, such as @code{list} and @code{edit}, this
8493 specifies a source line that contains @var{address}. For @code{break} and
8494 other breakpoint-oriented commands, this can be used to set breakpoints in
8495 parts of your program which do not have debugging information or
8498 Here @var{address} may be any expression valid in the current working
8499 language (@pxref{Languages, working language}) that specifies a code
8500 address. In addition, as a convenience, @value{GDBN} extends the
8501 semantics of expressions used in locations to cover several situations
8502 that frequently occur during debugging. Here are the various forms
8506 @item @var{expression}
8507 Any expression valid in the current working language.
8509 @item @var{funcaddr}
8510 An address of a function or procedure derived from its name. In C,
8511 C@t{++}, Objective-C, Fortran, minimal, and assembly, this is
8512 simply the function's name @var{function} (and actually a special case
8513 of a valid expression). In Pascal and Modula-2, this is
8514 @code{&@var{function}}. In Ada, this is @code{@var{function}'Address}
8515 (although the Pascal form also works).
8517 This form specifies the address of the function's first instruction,
8518 before the stack frame and arguments have been set up.
8520 @item '@var{filename}':@var{funcaddr}
8521 Like @var{funcaddr} above, but also specifies the name of the source
8522 file explicitly. This is useful if the name of the function does not
8523 specify the function unambiguously, e.g., if there are several
8524 functions with identical names in different source files.
8528 @section Editing Source Files
8529 @cindex editing source files
8532 @kindex e @r{(@code{edit})}
8533 To edit the lines in a source file, use the @code{edit} command.
8534 The editing program of your choice
8535 is invoked with the current line set to
8536 the active line in the program.
8537 Alternatively, there are several ways to specify what part of the file you
8538 want to print if you want to see other parts of the program:
8541 @item edit @var{location}
8542 Edit the source file specified by @code{location}. Editing starts at
8543 that @var{location}, e.g., at the specified source line of the
8544 specified file. @xref{Specify Location}, for all the possible forms
8545 of the @var{location} argument; here are the forms of the @code{edit}
8546 command most commonly used:
8549 @item edit @var{number}
8550 Edit the current source file with @var{number} as the active line number.
8552 @item edit @var{function}
8553 Edit the file containing @var{function} at the beginning of its definition.
8558 @subsection Choosing your Editor
8559 You can customize @value{GDBN} to use any editor you want
8561 The only restriction is that your editor (say @code{ex}), recognizes the
8562 following command-line syntax:
8564 ex +@var{number} file
8566 The optional numeric value +@var{number} specifies the number of the line in
8567 the file where to start editing.}.
8568 By default, it is @file{@value{EDITOR}}, but you can change this
8569 by setting the environment variable @code{EDITOR} before using
8570 @value{GDBN}. For example, to configure @value{GDBN} to use the
8571 @code{vi} editor, you could use these commands with the @code{sh} shell:
8577 or in the @code{csh} shell,
8579 setenv EDITOR /usr/bin/vi
8584 @section Searching Source Files
8585 @cindex searching source files
8587 There are two commands for searching through the current source file for a
8592 @kindex forward-search
8593 @kindex fo @r{(@code{forward-search})}
8594 @item forward-search @var{regexp}
8595 @itemx search @var{regexp}
8596 The command @samp{forward-search @var{regexp}} checks each line,
8597 starting with the one following the last line listed, for a match for
8598 @var{regexp}. It lists the line that is found. You can use the
8599 synonym @samp{search @var{regexp}} or abbreviate the command name as
8602 @kindex reverse-search
8603 @item reverse-search @var{regexp}
8604 The command @samp{reverse-search @var{regexp}} checks each line, starting
8605 with the one before the last line listed and going backward, for a match
8606 for @var{regexp}. It lists the line that is found. You can abbreviate
8607 this command as @code{rev}.
8611 @section Specifying Source Directories
8614 @cindex directories for source files
8615 Executable programs sometimes do not record the directories of the source
8616 files from which they were compiled, just the names. Even when they do,
8617 the directories could be moved between the compilation and your debugging
8618 session. @value{GDBN} has a list of directories to search for source files;
8619 this is called the @dfn{source path}. Each time @value{GDBN} wants a source file,
8620 it tries all the directories in the list, in the order they are present
8621 in the list, until it finds a file with the desired name.
8623 For example, suppose an executable references the file
8624 @file{/usr/src/foo-1.0/lib/foo.c}, and our source path is
8625 @file{/mnt/cross}. The file is first looked up literally; if this
8626 fails, @file{/mnt/cross/usr/src/foo-1.0/lib/foo.c} is tried; if this
8627 fails, @file{/mnt/cross/foo.c} is opened; if this fails, an error
8628 message is printed. @value{GDBN} does not look up the parts of the
8629 source file name, such as @file{/mnt/cross/src/foo-1.0/lib/foo.c}.
8630 Likewise, the subdirectories of the source path are not searched: if
8631 the source path is @file{/mnt/cross}, and the binary refers to
8632 @file{foo.c}, @value{GDBN} would not find it under
8633 @file{/mnt/cross/usr/src/foo-1.0/lib}.
8635 Plain file names, relative file names with leading directories, file
8636 names containing dots, etc.@: are all treated as described above; for
8637 instance, if the source path is @file{/mnt/cross}, and the source file
8638 is recorded as @file{../lib/foo.c}, @value{GDBN} would first try
8639 @file{../lib/foo.c}, then @file{/mnt/cross/../lib/foo.c}, and after
8640 that---@file{/mnt/cross/foo.c}.
8642 Note that the executable search path is @emph{not} used to locate the
8645 Whenever you reset or rearrange the source path, @value{GDBN} clears out
8646 any information it has cached about where source files are found and where
8647 each line is in the file.
8651 When you start @value{GDBN}, its source path includes only @samp{cdir}
8652 and @samp{cwd}, in that order.
8653 To add other directories, use the @code{directory} command.
8655 The search path is used to find both program source files and @value{GDBN}
8656 script files (read using the @samp{-command} option and @samp{source} command).
8658 In addition to the source path, @value{GDBN} provides a set of commands
8659 that manage a list of source path substitution rules. A @dfn{substitution
8660 rule} specifies how to rewrite source directories stored in the program's
8661 debug information in case the sources were moved to a different
8662 directory between compilation and debugging. A rule is made of
8663 two strings, the first specifying what needs to be rewritten in
8664 the path, and the second specifying how it should be rewritten.
8665 In @ref{set substitute-path}, we name these two parts @var{from} and
8666 @var{to} respectively. @value{GDBN} does a simple string replacement
8667 of @var{from} with @var{to} at the start of the directory part of the
8668 source file name, and uses that result instead of the original file
8669 name to look up the sources.
8671 Using the previous example, suppose the @file{foo-1.0} tree has been
8672 moved from @file{/usr/src} to @file{/mnt/cross}, then you can tell
8673 @value{GDBN} to replace @file{/usr/src} in all source path names with
8674 @file{/mnt/cross}. The first lookup will then be
8675 @file{/mnt/cross/foo-1.0/lib/foo.c} in place of the original location
8676 of @file{/usr/src/foo-1.0/lib/foo.c}. To define a source path
8677 substitution rule, use the @code{set substitute-path} command
8678 (@pxref{set substitute-path}).
8680 To avoid unexpected substitution results, a rule is applied only if the
8681 @var{from} part of the directory name ends at a directory separator.
8682 For instance, a rule substituting @file{/usr/source} into
8683 @file{/mnt/cross} will be applied to @file{/usr/source/foo-1.0} but
8684 not to @file{/usr/sourceware/foo-2.0}. And because the substitution
8685 is applied only at the beginning of the directory name, this rule will
8686 not be applied to @file{/root/usr/source/baz.c} either.
8688 In many cases, you can achieve the same result using the @code{directory}
8689 command. However, @code{set substitute-path} can be more efficient in
8690 the case where the sources are organized in a complex tree with multiple
8691 subdirectories. With the @code{directory} command, you need to add each
8692 subdirectory of your project. If you moved the entire tree while
8693 preserving its internal organization, then @code{set substitute-path}
8694 allows you to direct the debugger to all the sources with one single
8697 @code{set substitute-path} is also more than just a shortcut command.
8698 The source path is only used if the file at the original location no
8699 longer exists. On the other hand, @code{set substitute-path} modifies
8700 the debugger behavior to look at the rewritten location instead. So, if
8701 for any reason a source file that is not relevant to your executable is
8702 located at the original location, a substitution rule is the only
8703 method available to point @value{GDBN} at the new location.
8705 @cindex @samp{--with-relocated-sources}
8706 @cindex default source path substitution
8707 You can configure a default source path substitution rule by
8708 configuring @value{GDBN} with the
8709 @samp{--with-relocated-sources=@var{dir}} option. The @var{dir}
8710 should be the name of a directory under @value{GDBN}'s configured
8711 prefix (set with @samp{--prefix} or @samp{--exec-prefix}), and
8712 directory names in debug information under @var{dir} will be adjusted
8713 automatically if the installed @value{GDBN} is moved to a new
8714 location. This is useful if @value{GDBN}, libraries or executables
8715 with debug information and corresponding source code are being moved
8719 @item directory @var{dirname} @dots{}
8720 @item dir @var{dirname} @dots{}
8721 Add directory @var{dirname} to the front of the source path. Several
8722 directory names may be given to this command, separated by @samp{:}
8723 (@samp{;} on MS-DOS and MS-Windows, where @samp{:} usually appears as
8724 part of absolute file names) or
8725 whitespace. You may specify a directory that is already in the source
8726 path; this moves it forward, so @value{GDBN} searches it sooner.
8730 @vindex $cdir@r{, convenience variable}
8731 @vindex $cwd@r{, convenience variable}
8732 @cindex compilation directory
8733 @cindex current directory
8734 @cindex working directory
8735 @cindex directory, current
8736 @cindex directory, compilation
8737 You can use the string @samp{$cdir} to refer to the compilation
8738 directory (if one is recorded), and @samp{$cwd} to refer to the current
8739 working directory. @samp{$cwd} is not the same as @samp{.}---the former
8740 tracks the current working directory as it changes during your @value{GDBN}
8741 session, while the latter is immediately expanded to the current
8742 directory at the time you add an entry to the source path.
8745 Reset the source path to its default value (@samp{$cdir:$cwd} on Unix systems). This requires confirmation.
8747 @c RET-repeat for @code{directory} is explicitly disabled, but since
8748 @c repeating it would be a no-op we do not say that. (thanks to RMS)
8750 @item set directories @var{path-list}
8751 @kindex set directories
8752 Set the source path to @var{path-list}.
8753 @samp{$cdir:$cwd} are added if missing.
8755 @item show directories
8756 @kindex show directories
8757 Print the source path: show which directories it contains.
8759 @anchor{set substitute-path}
8760 @item set substitute-path @var{from} @var{to}
8761 @kindex set substitute-path
8762 Define a source path substitution rule, and add it at the end of the
8763 current list of existing substitution rules. If a rule with the same
8764 @var{from} was already defined, then the old rule is also deleted.
8766 For example, if the file @file{/foo/bar/baz.c} was moved to
8767 @file{/mnt/cross/baz.c}, then the command
8770 (@value{GDBP}) set substitute-path /foo/bar /mnt/cross
8774 will tell @value{GDBN} to replace @samp{/foo/bar} with
8775 @samp{/mnt/cross}, which will allow @value{GDBN} to find the file
8776 @file{baz.c} even though it was moved.
8778 In the case when more than one substitution rule have been defined,
8779 the rules are evaluated one by one in the order where they have been
8780 defined. The first one matching, if any, is selected to perform
8783 For instance, if we had entered the following commands:
8786 (@value{GDBP}) set substitute-path /usr/src/include /mnt/include
8787 (@value{GDBP}) set substitute-path /usr/src /mnt/src
8791 @value{GDBN} would then rewrite @file{/usr/src/include/defs.h} into
8792 @file{/mnt/include/defs.h} by using the first rule. However, it would
8793 use the second rule to rewrite @file{/usr/src/lib/foo.c} into
8794 @file{/mnt/src/lib/foo.c}.
8797 @item unset substitute-path [path]
8798 @kindex unset substitute-path
8799 If a path is specified, search the current list of substitution rules
8800 for a rule that would rewrite that path. Delete that rule if found.
8801 A warning is emitted by the debugger if no rule could be found.
8803 If no path is specified, then all substitution rules are deleted.
8805 @item show substitute-path [path]
8806 @kindex show substitute-path
8807 If a path is specified, then print the source path substitution rule
8808 which would rewrite that path, if any.
8810 If no path is specified, then print all existing source path substitution
8815 If your source path is cluttered with directories that are no longer of
8816 interest, @value{GDBN} may sometimes cause confusion by finding the wrong
8817 versions of source. You can correct the situation as follows:
8821 Use @code{directory} with no argument to reset the source path to its default value.
8824 Use @code{directory} with suitable arguments to reinstall the
8825 directories you want in the source path. You can add all the
8826 directories in one command.
8830 @section Source and Machine Code
8831 @cindex source line and its code address
8833 You can use the command @code{info line} to map source lines to program
8834 addresses (and vice versa), and the command @code{disassemble} to display
8835 a range of addresses as machine instructions. You can use the command
8836 @code{set disassemble-next-line} to set whether to disassemble next
8837 source line when execution stops. When run under @sc{gnu} Emacs
8838 mode, the @code{info line} command causes the arrow to point to the
8839 line specified. Also, @code{info line} prints addresses in symbolic form as
8845 @itemx info line @var{location}
8846 Print the starting and ending addresses of the compiled code for
8847 source line @var{location}. You can specify source lines in any of
8848 the ways documented in @ref{Specify Location}. With no @var{location}
8849 information about the current source line is printed.
8852 For example, we can use @code{info line} to discover the location of
8853 the object code for the first line of function
8854 @code{m4_changequote}:
8857 (@value{GDBP}) info line m4_changequote
8858 Line 895 of "builtin.c" starts at pc 0x634c <m4_changequote> and \
8859 ends at 0x6350 <m4_changequote+4>.
8863 @cindex code address and its source line
8864 We can also inquire (using @code{*@var{addr}} as the form for
8865 @var{location}) what source line covers a particular address:
8867 (@value{GDBP}) info line *0x63ff
8868 Line 926 of "builtin.c" starts at pc 0x63e4 <m4_changequote+152> and \
8869 ends at 0x6404 <m4_changequote+184>.
8872 @cindex @code{$_} and @code{info line}
8873 @cindex @code{x} command, default address
8874 @kindex x@r{(examine), and} info line
8875 After @code{info line}, the default address for the @code{x} command
8876 is changed to the starting address of the line, so that @samp{x/i} is
8877 sufficient to begin examining the machine code (@pxref{Memory,
8878 ,Examining Memory}). Also, this address is saved as the value of the
8879 convenience variable @code{$_} (@pxref{Convenience Vars, ,Convenience
8882 @cindex info line, repeated calls
8883 After @code{info line}, using @code{info line} again without
8884 specifying a location will display information about the next source
8889 @cindex assembly instructions
8890 @cindex instructions, assembly
8891 @cindex machine instructions
8892 @cindex listing machine instructions
8894 @itemx disassemble /m
8895 @itemx disassemble /s
8896 @itemx disassemble /r
8897 This specialized command dumps a range of memory as machine
8898 instructions. It can also print mixed source+disassembly by specifying
8899 the @code{/m} or @code{/s} modifier and print the raw instructions in hex
8900 as well as in symbolic form by specifying the @code{/r} modifier.
8901 The default memory range is the function surrounding the
8902 program counter of the selected frame. A single argument to this
8903 command is a program counter value; @value{GDBN} dumps the function
8904 surrounding this value. When two arguments are given, they should
8905 be separated by a comma, possibly surrounded by whitespace. The
8906 arguments specify a range of addresses to dump, in one of two forms:
8909 @item @var{start},@var{end}
8910 the addresses from @var{start} (inclusive) to @var{end} (exclusive)
8911 @item @var{start},+@var{length}
8912 the addresses from @var{start} (inclusive) to
8913 @code{@var{start}+@var{length}} (exclusive).
8917 When 2 arguments are specified, the name of the function is also
8918 printed (since there could be several functions in the given range).
8920 The argument(s) can be any expression yielding a numeric value, such as
8921 @samp{0x32c4}, @samp{&main+10} or @samp{$pc - 8}.
8923 If the range of memory being disassembled contains current program counter,
8924 the instruction at that location is shown with a @code{=>} marker.
8927 The following example shows the disassembly of a range of addresses of
8928 HP PA-RISC 2.0 code:
8931 (@value{GDBP}) disas 0x32c4, 0x32e4
8932 Dump of assembler code from 0x32c4 to 0x32e4:
8933 0x32c4 <main+204>: addil 0,dp
8934 0x32c8 <main+208>: ldw 0x22c(sr0,r1),r26
8935 0x32cc <main+212>: ldil 0x3000,r31
8936 0x32d0 <main+216>: ble 0x3f8(sr4,r31)
8937 0x32d4 <main+220>: ldo 0(r31),rp
8938 0x32d8 <main+224>: addil -0x800,dp
8939 0x32dc <main+228>: ldo 0x588(r1),r26
8940 0x32e0 <main+232>: ldil 0x3000,r31
8941 End of assembler dump.
8944 Here is an example showing mixed source+assembly for Intel x86
8945 with @code{/m} or @code{/s}, when the program is stopped just after
8946 function prologue in a non-optimized function with no inline code.
8949 (@value{GDBP}) disas /m main
8950 Dump of assembler code for function main:
8952 0x08048330 <+0>: push %ebp
8953 0x08048331 <+1>: mov %esp,%ebp
8954 0x08048333 <+3>: sub $0x8,%esp
8955 0x08048336 <+6>: and $0xfffffff0,%esp
8956 0x08048339 <+9>: sub $0x10,%esp
8958 6 printf ("Hello.\n");
8959 => 0x0804833c <+12>: movl $0x8048440,(%esp)
8960 0x08048343 <+19>: call 0x8048284 <puts@@plt>
8964 0x08048348 <+24>: mov $0x0,%eax
8965 0x0804834d <+29>: leave
8966 0x0804834e <+30>: ret
8968 End of assembler dump.
8971 The @code{/m} option is deprecated as its output is not useful when
8972 there is either inlined code or re-ordered code.
8973 The @code{/s} option is the preferred choice.
8974 Here is an example for AMD x86-64 showing the difference between
8975 @code{/m} output and @code{/s} output.
8976 This example has one inline function defined in a header file,
8977 and the code is compiled with @samp{-O2} optimization.
8978 Note how the @code{/m} output is missing the disassembly of
8979 several instructions that are present in the @code{/s} output.
9009 (@value{GDBP}) disas /m main
9010 Dump of assembler code for function main:
9014 0x0000000000400400 <+0>: mov 0x200c2e(%rip),%eax # 0x601034 <y>
9015 0x0000000000400417 <+23>: mov %eax,0x200c13(%rip) # 0x601030 <x>
9019 0x000000000040041d <+29>: xor %eax,%eax
9020 0x000000000040041f <+31>: retq
9021 0x0000000000400420 <+32>: add %eax,%eax
9022 0x0000000000400422 <+34>: jmp 0x400417 <main+23>
9024 End of assembler dump.
9025 (@value{GDBP}) disas /s main
9026 Dump of assembler code for function main:
9030 0x0000000000400400 <+0>: mov 0x200c2e(%rip),%eax # 0x601034 <y>
9034 0x0000000000400406 <+6>: test %eax,%eax
9035 0x0000000000400408 <+8>: js 0x400420 <main+32>
9040 0x000000000040040a <+10>: lea 0xa(%rax),%edx
9041 0x000000000040040d <+13>: test %eax,%eax
9042 0x000000000040040f <+15>: mov $0x1,%eax
9043 0x0000000000400414 <+20>: cmovne %edx,%eax
9047 0x0000000000400417 <+23>: mov %eax,0x200c13(%rip) # 0x601030 <x>
9051 0x000000000040041d <+29>: xor %eax,%eax
9052 0x000000000040041f <+31>: retq
9056 0x0000000000400420 <+32>: add %eax,%eax
9057 0x0000000000400422 <+34>: jmp 0x400417 <main+23>
9058 End of assembler dump.
9061 Here is another example showing raw instructions in hex for AMD x86-64,
9064 (gdb) disas /r 0x400281,+10
9065 Dump of assembler code from 0x400281 to 0x40028b:
9066 0x0000000000400281: 38 36 cmp %dh,(%rsi)
9067 0x0000000000400283: 2d 36 34 2e 73 sub $0x732e3436,%eax
9068 0x0000000000400288: 6f outsl %ds:(%rsi),(%dx)
9069 0x0000000000400289: 2e 32 00 xor %cs:(%rax),%al
9070 End of assembler dump.
9073 Addresses cannot be specified as a location (@pxref{Specify Location}).
9074 So, for example, if you want to disassemble function @code{bar}
9075 in file @file{foo.c}, you must type @samp{disassemble 'foo.c'::bar}
9076 and not @samp{disassemble foo.c:bar}.
9078 Some architectures have more than one commonly-used set of instruction
9079 mnemonics or other syntax.
9081 For programs that were dynamically linked and use shared libraries,
9082 instructions that call functions or branch to locations in the shared
9083 libraries might show a seemingly bogus location---it's actually a
9084 location of the relocation table. On some architectures, @value{GDBN}
9085 might be able to resolve these to actual function names.
9088 @kindex set disassembler-options
9089 @cindex disassembler options
9090 @item set disassembler-options @var{option1}[,@var{option2}@dots{}]
9091 This command controls the passing of target specific information to
9092 the disassembler. For a list of valid options, please refer to the
9093 @code{-M}/@code{--disassembler-options} section of the @samp{objdump}
9094 manual and/or the output of @kbd{objdump --help}
9095 (@pxref{objdump,,objdump,binutils,The GNU Binary Utilities}).
9096 The default value is the empty string.
9098 If it is necessary to specify more than one disassembler option, then
9099 multiple options can be placed together into a comma separated list.
9100 Currently this command is only supported on targets ARM, MIPS, PowerPC
9103 @kindex show disassembler-options
9104 @item show disassembler-options
9105 Show the current setting of the disassembler options.
9109 @kindex set disassembly-flavor
9110 @cindex Intel disassembly flavor
9111 @cindex AT&T disassembly flavor
9112 @item set disassembly-flavor @var{instruction-set}
9113 Select the instruction set to use when disassembling the
9114 program via the @code{disassemble} or @code{x/i} commands.
9116 Currently this command is only defined for the Intel x86 family. You
9117 can set @var{instruction-set} to either @code{intel} or @code{att}.
9118 The default is @code{att}, the AT&T flavor used by default by Unix
9119 assemblers for x86-based targets.
9121 @kindex show disassembly-flavor
9122 @item show disassembly-flavor
9123 Show the current setting of the disassembly flavor.
9127 @kindex set disassemble-next-line
9128 @kindex show disassemble-next-line
9129 @item set disassemble-next-line
9130 @itemx show disassemble-next-line
9131 Control whether or not @value{GDBN} will disassemble the next source
9132 line or instruction when execution stops. If ON, @value{GDBN} will
9133 display disassembly of the next source line when execution of the
9134 program being debugged stops. This is @emph{in addition} to
9135 displaying the source line itself, which @value{GDBN} always does if
9136 possible. If the next source line cannot be displayed for some reason
9137 (e.g., if @value{GDBN} cannot find the source file, or there's no line
9138 info in the debug info), @value{GDBN} will display disassembly of the
9139 next @emph{instruction} instead of showing the next source line. If
9140 AUTO, @value{GDBN} will display disassembly of next instruction only
9141 if the source line cannot be displayed. This setting causes
9142 @value{GDBN} to display some feedback when you step through a function
9143 with no line info or whose source file is unavailable. The default is
9144 OFF, which means never display the disassembly of the next line or
9150 @chapter Examining Data
9152 @cindex printing data
9153 @cindex examining data
9156 The usual way to examine data in your program is with the @code{print}
9157 command (abbreviated @code{p}), or its synonym @code{inspect}. It
9158 evaluates and prints the value of an expression of the language your
9159 program is written in (@pxref{Languages, ,Using @value{GDBN} with
9160 Different Languages}). It may also print the expression using a
9161 Python-based pretty-printer (@pxref{Pretty Printing}).
9164 @item print @var{expr}
9165 @itemx print /@var{f} @var{expr}
9166 @var{expr} is an expression (in the source language). By default the
9167 value of @var{expr} is printed in a format appropriate to its data type;
9168 you can choose a different format by specifying @samp{/@var{f}}, where
9169 @var{f} is a letter specifying the format; see @ref{Output Formats,,Output
9173 @itemx print /@var{f}
9174 @cindex reprint the last value
9175 If you omit @var{expr}, @value{GDBN} displays the last value again (from the
9176 @dfn{value history}; @pxref{Value History, ,Value History}). This allows you to
9177 conveniently inspect the same value in an alternative format.
9180 A more low-level way of examining data is with the @code{x} command.
9181 It examines data in memory at a specified address and prints it in a
9182 specified format. @xref{Memory, ,Examining Memory}.
9184 If you are interested in information about types, or about how the
9185 fields of a struct or a class are declared, use the @code{ptype @var{exp}}
9186 command rather than @code{print}. @xref{Symbols, ,Examining the Symbol
9189 @cindex exploring hierarchical data structures
9191 Another way of examining values of expressions and type information is
9192 through the Python extension command @code{explore} (available only if
9193 the @value{GDBN} build is configured with @code{--with-python}). It
9194 offers an interactive way to start at the highest level (or, the most
9195 abstract level) of the data type of an expression (or, the data type
9196 itself) and explore all the way down to leaf scalar values/fields
9197 embedded in the higher level data types.
9200 @item explore @var{arg}
9201 @var{arg} is either an expression (in the source language), or a type
9202 visible in the current context of the program being debugged.
9205 The working of the @code{explore} command can be illustrated with an
9206 example. If a data type @code{struct ComplexStruct} is defined in your
9216 struct ComplexStruct
9218 struct SimpleStruct *ss_p;
9224 followed by variable declarations as
9227 struct SimpleStruct ss = @{ 10, 1.11 @};
9228 struct ComplexStruct cs = @{ &ss, @{ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 @} @};
9232 then, the value of the variable @code{cs} can be explored using the
9233 @code{explore} command as follows.
9237 The value of `cs' is a struct/class of type `struct ComplexStruct' with
9238 the following fields:
9240 ss_p = <Enter 0 to explore this field of type `struct SimpleStruct *'>
9241 arr = <Enter 1 to explore this field of type `int [10]'>
9243 Enter the field number of choice:
9247 Since the fields of @code{cs} are not scalar values, you are being
9248 prompted to chose the field you want to explore. Let's say you choose
9249 the field @code{ss_p} by entering @code{0}. Then, since this field is a
9250 pointer, you will be asked if it is pointing to a single value. From
9251 the declaration of @code{cs} above, it is indeed pointing to a single
9252 value, hence you enter @code{y}. If you enter @code{n}, then you will
9253 be asked if it were pointing to an array of values, in which case this
9254 field will be explored as if it were an array.
9257 `cs.ss_p' is a pointer to a value of type `struct SimpleStruct'
9258 Continue exploring it as a pointer to a single value [y/n]: y
9259 The value of `*(cs.ss_p)' is a struct/class of type `struct
9260 SimpleStruct' with the following fields:
9262 i = 10 .. (Value of type `int')
9263 d = 1.1100000000000001 .. (Value of type `double')
9265 Press enter to return to parent value:
9269 If the field @code{arr} of @code{cs} was chosen for exploration by
9270 entering @code{1} earlier, then since it is as array, you will be
9271 prompted to enter the index of the element in the array that you want
9275 `cs.arr' is an array of `int'.
9276 Enter the index of the element you want to explore in `cs.arr': 5
9278 `(cs.arr)[5]' is a scalar value of type `int'.
9282 Press enter to return to parent value:
9285 In general, at any stage of exploration, you can go deeper towards the
9286 leaf values by responding to the prompts appropriately, or hit the
9287 return key to return to the enclosing data structure (the @i{higher}
9288 level data structure).
9290 Similar to exploring values, you can use the @code{explore} command to
9291 explore types. Instead of specifying a value (which is typically a
9292 variable name or an expression valid in the current context of the
9293 program being debugged), you specify a type name. If you consider the
9294 same example as above, your can explore the type
9295 @code{struct ComplexStruct} by passing the argument
9296 @code{struct ComplexStruct} to the @code{explore} command.
9299 (gdb) explore struct ComplexStruct
9303 By responding to the prompts appropriately in the subsequent interactive
9304 session, you can explore the type @code{struct ComplexStruct} in a
9305 manner similar to how the value @code{cs} was explored in the above
9308 The @code{explore} command also has two sub-commands,
9309 @code{explore value} and @code{explore type}. The former sub-command is
9310 a way to explicitly specify that value exploration of the argument is
9311 being invoked, while the latter is a way to explicitly specify that type
9312 exploration of the argument is being invoked.
9315 @item explore value @var{expr}
9316 @cindex explore value
9317 This sub-command of @code{explore} explores the value of the
9318 expression @var{expr} (if @var{expr} is an expression valid in the
9319 current context of the program being debugged). The behavior of this
9320 command is identical to that of the behavior of the @code{explore}
9321 command being passed the argument @var{expr}.
9323 @item explore type @var{arg}
9324 @cindex explore type
9325 This sub-command of @code{explore} explores the type of @var{arg} (if
9326 @var{arg} is a type visible in the current context of program being
9327 debugged), or the type of the value/expression @var{arg} (if @var{arg}
9328 is an expression valid in the current context of the program being
9329 debugged). If @var{arg} is a type, then the behavior of this command is
9330 identical to that of the @code{explore} command being passed the
9331 argument @var{arg}. If @var{arg} is an expression, then the behavior of
9332 this command will be identical to that of the @code{explore} command
9333 being passed the type of @var{arg} as the argument.
9337 * Expressions:: Expressions
9338 * Ambiguous Expressions:: Ambiguous Expressions
9339 * Variables:: Program variables
9340 * Arrays:: Artificial arrays
9341 * Output Formats:: Output formats
9342 * Memory:: Examining memory
9343 * Auto Display:: Automatic display
9344 * Print Settings:: Print settings
9345 * Pretty Printing:: Python pretty printing
9346 * Value History:: Value history
9347 * Convenience Vars:: Convenience variables
9348 * Convenience Funs:: Convenience functions
9349 * Registers:: Registers
9350 * Floating Point Hardware:: Floating point hardware
9351 * Vector Unit:: Vector Unit
9352 * OS Information:: Auxiliary data provided by operating system
9353 * Memory Region Attributes:: Memory region attributes
9354 * Dump/Restore Files:: Copy between memory and a file
9355 * Core File Generation:: Cause a program dump its core
9356 * Character Sets:: Debugging programs that use a different
9357 character set than GDB does
9358 * Caching Target Data:: Data caching for targets
9359 * Searching Memory:: Searching memory for a sequence of bytes
9360 * Value Sizes:: Managing memory allocated for values
9364 @section Expressions
9367 @code{print} and many other @value{GDBN} commands accept an expression and
9368 compute its value. Any kind of constant, variable or operator defined
9369 by the programming language you are using is valid in an expression in
9370 @value{GDBN}. This includes conditional expressions, function calls,
9371 casts, and string constants. It also includes preprocessor macros, if
9372 you compiled your program to include this information; see
9375 @cindex arrays in expressions
9376 @value{GDBN} supports array constants in expressions input by
9377 the user. The syntax is @{@var{element}, @var{element}@dots{}@}. For example,
9378 you can use the command @code{print @{1, 2, 3@}} to create an array
9379 of three integers. If you pass an array to a function or assign it
9380 to a program variable, @value{GDBN} copies the array to memory that
9381 is @code{malloc}ed in the target program.
9383 Because C is so widespread, most of the expressions shown in examples in
9384 this manual are in C. @xref{Languages, , Using @value{GDBN} with Different
9385 Languages}, for information on how to use expressions in other
9388 In this section, we discuss operators that you can use in @value{GDBN}
9389 expressions regardless of your programming language.
9391 @cindex casts, in expressions
9392 Casts are supported in all languages, not just in C, because it is so
9393 useful to cast a number into a pointer in order to examine a structure
9394 at that address in memory.
9395 @c FIXME: casts supported---Mod2 true?
9397 @value{GDBN} supports these operators, in addition to those common
9398 to programming languages:
9402 @samp{@@} is a binary operator for treating parts of memory as arrays.
9403 @xref{Arrays, ,Artificial Arrays}, for more information.
9406 @samp{::} allows you to specify a variable in terms of the file or
9407 function where it is defined. @xref{Variables, ,Program Variables}.
9409 @cindex @{@var{type}@}
9410 @cindex type casting memory
9411 @cindex memory, viewing as typed object
9412 @cindex casts, to view memory
9413 @item @{@var{type}@} @var{addr}
9414 Refers to an object of type @var{type} stored at address @var{addr} in
9415 memory. The address @var{addr} may be any expression whose value is
9416 an integer or pointer (but parentheses are required around binary
9417 operators, just as in a cast). This construct is allowed regardless
9418 of what kind of data is normally supposed to reside at @var{addr}.
9421 @node Ambiguous Expressions
9422 @section Ambiguous Expressions
9423 @cindex ambiguous expressions
9425 Expressions can sometimes contain some ambiguous elements. For instance,
9426 some programming languages (notably Ada, C@t{++} and Objective-C) permit
9427 a single function name to be defined several times, for application in
9428 different contexts. This is called @dfn{overloading}. Another example
9429 involving Ada is generics. A @dfn{generic package} is similar to C@t{++}
9430 templates and is typically instantiated several times, resulting in
9431 the same function name being defined in different contexts.
9433 In some cases and depending on the language, it is possible to adjust
9434 the expression to remove the ambiguity. For instance in C@t{++}, you
9435 can specify the signature of the function you want to break on, as in
9436 @kbd{break @var{function}(@var{types})}. In Ada, using the fully
9437 qualified name of your function often makes the expression unambiguous
9440 When an ambiguity that needs to be resolved is detected, the debugger
9441 has the capability to display a menu of numbered choices for each
9442 possibility, and then waits for the selection with the prompt @samp{>}.
9443 The first option is always @samp{[0] cancel}, and typing @kbd{0 @key{RET}}
9444 aborts the current command. If the command in which the expression was
9445 used allows more than one choice to be selected, the next option in the
9446 menu is @samp{[1] all}, and typing @kbd{1 @key{RET}} selects all possible
9449 For example, the following session excerpt shows an attempt to set a
9450 breakpoint at the overloaded symbol @code{String::after}.
9451 We choose three particular definitions of that function name:
9453 @c FIXME! This is likely to change to show arg type lists, at least
9456 (@value{GDBP}) b String::after
9459 [2] file:String.cc; line number:867
9460 [3] file:String.cc; line number:860
9461 [4] file:String.cc; line number:875
9462 [5] file:String.cc; line number:853
9463 [6] file:String.cc; line number:846
9464 [7] file:String.cc; line number:735
9466 Breakpoint 1 at 0xb26c: file String.cc, line 867.
9467 Breakpoint 2 at 0xb344: file String.cc, line 875.
9468 Breakpoint 3 at 0xafcc: file String.cc, line 846.
9469 Multiple breakpoints were set.
9470 Use the "delete" command to delete unwanted
9477 @kindex set multiple-symbols
9478 @item set multiple-symbols @var{mode}
9479 @cindex multiple-symbols menu
9481 This option allows you to adjust the debugger behavior when an expression
9484 By default, @var{mode} is set to @code{all}. If the command with which
9485 the expression is used allows more than one choice, then @value{GDBN}
9486 automatically selects all possible choices. For instance, inserting
9487 a breakpoint on a function using an ambiguous name results in a breakpoint
9488 inserted on each possible match. However, if a unique choice must be made,
9489 then @value{GDBN} uses the menu to help you disambiguate the expression.
9490 For instance, printing the address of an overloaded function will result
9491 in the use of the menu.
9493 When @var{mode} is set to @code{ask}, the debugger always uses the menu
9494 when an ambiguity is detected.
9496 Finally, when @var{mode} is set to @code{cancel}, the debugger reports
9497 an error due to the ambiguity and the command is aborted.
9499 @kindex show multiple-symbols
9500 @item show multiple-symbols
9501 Show the current value of the @code{multiple-symbols} setting.
9505 @section Program Variables
9507 The most common kind of expression to use is the name of a variable
9510 Variables in expressions are understood in the selected stack frame
9511 (@pxref{Selection, ,Selecting a Frame}); they must be either:
9515 global (or file-static)
9522 visible according to the scope rules of the
9523 programming language from the point of execution in that frame
9526 @noindent This means that in the function
9541 you can examine and use the variable @code{a} whenever your program is
9542 executing within the function @code{foo}, but you can only use or
9543 examine the variable @code{b} while your program is executing inside
9544 the block where @code{b} is declared.
9546 @cindex variable name conflict
9547 There is an exception: you can refer to a variable or function whose
9548 scope is a single source file even if the current execution point is not
9549 in this file. But it is possible to have more than one such variable or
9550 function with the same name (in different source files). If that
9551 happens, referring to that name has unpredictable effects. If you wish,
9552 you can specify a static variable in a particular function or file by
9553 using the colon-colon (@code{::}) notation:
9555 @cindex colon-colon, context for variables/functions
9557 @c info cannot cope with a :: index entry, but why deprive hard copy readers?
9558 @cindex @code{::}, context for variables/functions
9561 @var{file}::@var{variable}
9562 @var{function}::@var{variable}
9566 Here @var{file} or @var{function} is the name of the context for the
9567 static @var{variable}. In the case of file names, you can use quotes to
9568 make sure @value{GDBN} parses the file name as a single word---for example,
9569 to print a global value of @code{x} defined in @file{f2.c}:
9572 (@value{GDBP}) p 'f2.c'::x
9575 The @code{::} notation is normally used for referring to
9576 static variables, since you typically disambiguate uses of local variables
9577 in functions by selecting the appropriate frame and using the
9578 simple name of the variable. However, you may also use this notation
9579 to refer to local variables in frames enclosing the selected frame:
9588 process (a); /* Stop here */
9599 For example, if there is a breakpoint at the commented line,
9600 here is what you might see
9601 when the program stops after executing the call @code{bar(0)}:
9606 (@value{GDBP}) p bar::a
9609 #2 0x080483d0 in foo (a=5) at foobar.c:12
9612 (@value{GDBP}) p bar::a
9616 @cindex C@t{++} scope resolution
9617 These uses of @samp{::} are very rarely in conflict with the very
9618 similar use of the same notation in C@t{++}. When they are in
9619 conflict, the C@t{++} meaning takes precedence; however, this can be
9620 overridden by quoting the file or function name with single quotes.
9622 For example, suppose the program is stopped in a method of a class
9623 that has a field named @code{includefile}, and there is also an
9624 include file named @file{includefile} that defines a variable,
9628 (@value{GDBP}) p includefile
9630 (@value{GDBP}) p includefile::some_global
9631 A syntax error in expression, near `'.
9632 (@value{GDBP}) p 'includefile'::some_global
9636 @cindex wrong values
9637 @cindex variable values, wrong
9638 @cindex function entry/exit, wrong values of variables
9639 @cindex optimized code, wrong values of variables
9641 @emph{Warning:} Occasionally, a local variable may appear to have the
9642 wrong value at certain points in a function---just after entry to a new
9643 scope, and just before exit.
9645 You may see this problem when you are stepping by machine instructions.
9646 This is because, on most machines, it takes more than one instruction to
9647 set up a stack frame (including local variable definitions); if you are
9648 stepping by machine instructions, variables may appear to have the wrong
9649 values until the stack frame is completely built. On exit, it usually
9650 also takes more than one machine instruction to destroy a stack frame;
9651 after you begin stepping through that group of instructions, local
9652 variable definitions may be gone.
9654 This may also happen when the compiler does significant optimizations.
9655 To be sure of always seeing accurate values, turn off all optimization
9658 @cindex ``No symbol "foo" in current context''
9659 Another possible effect of compiler optimizations is to optimize
9660 unused variables out of existence, or assign variables to registers (as
9661 opposed to memory addresses). Depending on the support for such cases
9662 offered by the debug info format used by the compiler, @value{GDBN}
9663 might not be able to display values for such local variables. If that
9664 happens, @value{GDBN} will print a message like this:
9667 No symbol "foo" in current context.
9670 To solve such problems, either recompile without optimizations, or use a
9671 different debug info format, if the compiler supports several such
9672 formats. @xref{Compilation}, for more information on choosing compiler
9673 options. @xref{C, ,C and C@t{++}}, for more information about debug
9674 info formats that are best suited to C@t{++} programs.
9676 If you ask to print an object whose contents are unknown to
9677 @value{GDBN}, e.g., because its data type is not completely specified
9678 by the debug information, @value{GDBN} will say @samp{<incomplete
9679 type>}. @xref{Symbols, incomplete type}, for more about this.
9681 @cindex no debug info variables
9682 If you try to examine or use the value of a (global) variable for
9683 which @value{GDBN} has no type information, e.g., because the program
9684 includes no debug information, @value{GDBN} displays an error message.
9685 @xref{Symbols, unknown type}, for more about unknown types. If you
9686 cast the variable to its declared type, @value{GDBN} gets the
9687 variable's value using the cast-to type as the variable's type. For
9688 example, in a C program:
9691 (@value{GDBP}) p var
9692 'var' has unknown type; cast it to its declared type
9693 (@value{GDBP}) p (float) var
9697 If you append @kbd{@@entry} string to a function parameter name you get its
9698 value at the time the function got called. If the value is not available an
9699 error message is printed. Entry values are available only with some compilers.
9700 Entry values are normally also printed at the function parameter list according
9701 to @ref{set print entry-values}.
9704 Breakpoint 1, d (i=30) at gdb.base/entry-value.c:29
9710 (gdb) print i@@entry
9714 Strings are identified as arrays of @code{char} values without specified
9715 signedness. Arrays of either @code{signed char} or @code{unsigned char} get
9716 printed as arrays of 1 byte sized integers. @code{-fsigned-char} or
9717 @code{-funsigned-char} @value{NGCC} options have no effect as @value{GDBN}
9718 defines literal string type @code{"char"} as @code{char} without a sign.
9723 signed char var1[] = "A";
9726 You get during debugging
9731 $2 = @{65 'A', 0 '\0'@}
9735 @section Artificial Arrays
9737 @cindex artificial array
9739 @kindex @@@r{, referencing memory as an array}
9740 It is often useful to print out several successive objects of the
9741 same type in memory; a section of an array, or an array of
9742 dynamically determined size for which only a pointer exists in the
9745 You can do this by referring to a contiguous span of memory as an
9746 @dfn{artificial array}, using the binary operator @samp{@@}. The left
9747 operand of @samp{@@} should be the first element of the desired array
9748 and be an individual object. The right operand should be the desired length
9749 of the array. The result is an array value whose elements are all of
9750 the type of the left argument. The first element is actually the left
9751 argument; the second element comes from bytes of memory immediately
9752 following those that hold the first element, and so on. Here is an
9753 example. If a program says
9756 int *array = (int *) malloc (len * sizeof (int));
9760 you can print the contents of @code{array} with
9766 The left operand of @samp{@@} must reside in memory. Array values made
9767 with @samp{@@} in this way behave just like other arrays in terms of
9768 subscripting, and are coerced to pointers when used in expressions.
9769 Artificial arrays most often appear in expressions via the value history
9770 (@pxref{Value History, ,Value History}), after printing one out.
9772 Another way to create an artificial array is to use a cast.
9773 This re-interprets a value as if it were an array.
9774 The value need not be in memory:
9776 (@value{GDBP}) p/x (short[2])0x12345678
9777 $1 = @{0x1234, 0x5678@}
9780 As a convenience, if you leave the array length out (as in
9781 @samp{(@var{type}[])@var{value}}) @value{GDBN} calculates the size to fill
9782 the value (as @samp{sizeof(@var{value})/sizeof(@var{type})}:
9784 (@value{GDBP}) p/x (short[])0x12345678
9785 $2 = @{0x1234, 0x5678@}
9788 Sometimes the artificial array mechanism is not quite enough; in
9789 moderately complex data structures, the elements of interest may not
9790 actually be adjacent---for example, if you are interested in the values
9791 of pointers in an array. One useful work-around in this situation is
9792 to use a convenience variable (@pxref{Convenience Vars, ,Convenience
9793 Variables}) as a counter in an expression that prints the first
9794 interesting value, and then repeat that expression via @key{RET}. For
9795 instance, suppose you have an array @code{dtab} of pointers to
9796 structures, and you are interested in the values of a field @code{fv}
9797 in each structure. Here is an example of what you might type:
9807 @node Output Formats
9808 @section Output Formats
9810 @cindex formatted output
9811 @cindex output formats
9812 By default, @value{GDBN} prints a value according to its data type. Sometimes
9813 this is not what you want. For example, you might want to print a number
9814 in hex, or a pointer in decimal. Or you might want to view data in memory
9815 at a certain address as a character string or as an instruction. To do
9816 these things, specify an @dfn{output format} when you print a value.
9818 The simplest use of output formats is to say how to print a value
9819 already computed. This is done by starting the arguments of the
9820 @code{print} command with a slash and a format letter. The format
9821 letters supported are:
9825 Regard the bits of the value as an integer, and print the integer in
9829 Print as integer in signed decimal.
9832 Print as integer in unsigned decimal.
9835 Print as integer in octal.
9838 Print as integer in binary. The letter @samp{t} stands for ``two''.
9839 @footnote{@samp{b} cannot be used because these format letters are also
9840 used with the @code{x} command, where @samp{b} stands for ``byte'';
9841 see @ref{Memory,,Examining Memory}.}
9844 @cindex unknown address, locating
9845 @cindex locate address
9846 Print as an address, both absolute in hexadecimal and as an offset from
9847 the nearest preceding symbol. You can use this format used to discover
9848 where (in what function) an unknown address is located:
9851 (@value{GDBP}) p/a 0x54320
9852 $3 = 0x54320 <_initialize_vx+396>
9856 The command @code{info symbol 0x54320} yields similar results.
9857 @xref{Symbols, info symbol}.
9860 Regard as an integer and print it as a character constant. This
9861 prints both the numerical value and its character representation. The
9862 character representation is replaced with the octal escape @samp{\nnn}
9863 for characters outside the 7-bit @sc{ascii} range.
9865 Without this format, @value{GDBN} displays @code{char},
9866 @w{@code{unsigned char}}, and @w{@code{signed char}} data as character
9867 constants. Single-byte members of vectors are displayed as integer
9871 Regard the bits of the value as a floating point number and print
9872 using typical floating point syntax.
9875 @cindex printing strings
9876 @cindex printing byte arrays
9877 Regard as a string, if possible. With this format, pointers to single-byte
9878 data are displayed as null-terminated strings and arrays of single-byte data
9879 are displayed as fixed-length strings. Other values are displayed in their
9882 Without this format, @value{GDBN} displays pointers to and arrays of
9883 @code{char}, @w{@code{unsigned char}}, and @w{@code{signed char}} as
9884 strings. Single-byte members of a vector are displayed as an integer
9888 Like @samp{x} formatting, the value is treated as an integer and
9889 printed as hexadecimal, but leading zeros are printed to pad the value
9890 to the size of the integer type.
9893 @cindex raw printing
9894 Print using the @samp{raw} formatting. By default, @value{GDBN} will
9895 use a Python-based pretty-printer, if one is available (@pxref{Pretty
9896 Printing}). This typically results in a higher-level display of the
9897 value's contents. The @samp{r} format bypasses any Python
9898 pretty-printer which might exist.
9901 For example, to print the program counter in hex (@pxref{Registers}), type
9908 Note that no space is required before the slash; this is because command
9909 names in @value{GDBN} cannot contain a slash.
9911 To reprint the last value in the value history with a different format,
9912 you can use the @code{print} command with just a format and no
9913 expression. For example, @samp{p/x} reprints the last value in hex.
9916 @section Examining Memory
9918 You can use the command @code{x} (for ``examine'') to examine memory in
9919 any of several formats, independently of your program's data types.
9921 @cindex examining memory
9923 @kindex x @r{(examine memory)}
9924 @item x/@var{nfu} @var{addr}
9927 Use the @code{x} command to examine memory.
9930 @var{n}, @var{f}, and @var{u} are all optional parameters that specify how
9931 much memory to display and how to format it; @var{addr} is an
9932 expression giving the address where you want to start displaying memory.
9933 If you use defaults for @var{nfu}, you need not type the slash @samp{/}.
9934 Several commands set convenient defaults for @var{addr}.
9937 @item @var{n}, the repeat count
9938 The repeat count is a decimal integer; the default is 1. It specifies
9939 how much memory (counting by units @var{u}) to display. If a negative
9940 number is specified, memory is examined backward from @var{addr}.
9941 @c This really is **decimal**; unaffected by 'set radix' as of GDB
9944 @item @var{f}, the display format
9945 The display format is one of the formats used by @code{print}
9946 (@samp{x}, @samp{d}, @samp{u}, @samp{o}, @samp{t}, @samp{a}, @samp{c},
9947 @samp{f}, @samp{s}), and in addition @samp{i} (for machine instructions).
9948 The default is @samp{x} (hexadecimal) initially. The default changes
9949 each time you use either @code{x} or @code{print}.
9951 @item @var{u}, the unit size
9952 The unit size is any of
9958 Halfwords (two bytes).
9960 Words (four bytes). This is the initial default.
9962 Giant words (eight bytes).
9965 Each time you specify a unit size with @code{x}, that size becomes the
9966 default unit the next time you use @code{x}. For the @samp{i} format,
9967 the unit size is ignored and is normally not written. For the @samp{s} format,
9968 the unit size defaults to @samp{b}, unless it is explicitly given.
9969 Use @kbd{x /hs} to display 16-bit char strings and @kbd{x /ws} to display
9970 32-bit strings. The next use of @kbd{x /s} will again display 8-bit strings.
9971 Note that the results depend on the programming language of the
9972 current compilation unit. If the language is C, the @samp{s}
9973 modifier will use the UTF-16 encoding while @samp{w} will use
9974 UTF-32. The encoding is set by the programming language and cannot
9977 @item @var{addr}, starting display address
9978 @var{addr} is the address where you want @value{GDBN} to begin displaying
9979 memory. The expression need not have a pointer value (though it may);
9980 it is always interpreted as an integer address of a byte of memory.
9981 @xref{Expressions, ,Expressions}, for more information on expressions. The default for
9982 @var{addr} is usually just after the last address examined---but several
9983 other commands also set the default address: @code{info breakpoints} (to
9984 the address of the last breakpoint listed), @code{info line} (to the
9985 starting address of a line), and @code{print} (if you use it to display
9986 a value from memory).
9989 For example, @samp{x/3uh 0x54320} is a request to display three halfwords
9990 (@code{h}) of memory, formatted as unsigned decimal integers (@samp{u}),
9991 starting at address @code{0x54320}. @samp{x/4xw $sp} prints the four
9992 words (@samp{w}) of memory above the stack pointer (here, @samp{$sp};
9993 @pxref{Registers, ,Registers}) in hexadecimal (@samp{x}).
9995 You can also specify a negative repeat count to examine memory backward
9996 from the given address. For example, @samp{x/-3uh 0x54320} prints three
9997 halfwords (@code{h}) at @code{0x54314}, @code{0x54328}, and @code{0x5431c}.
9999 Since the letters indicating unit sizes are all distinct from the
10000 letters specifying output formats, you do not have to remember whether
10001 unit size or format comes first; either order works. The output
10002 specifications @samp{4xw} and @samp{4wx} mean exactly the same thing.
10003 (However, the count @var{n} must come first; @samp{wx4} does not work.)
10005 Even though the unit size @var{u} is ignored for the formats @samp{s}
10006 and @samp{i}, you might still want to use a count @var{n}; for example,
10007 @samp{3i} specifies that you want to see three machine instructions,
10008 including any operands. For convenience, especially when used with
10009 the @code{display} command, the @samp{i} format also prints branch delay
10010 slot instructions, if any, beyond the count specified, which immediately
10011 follow the last instruction that is within the count. The command
10012 @code{disassemble} gives an alternative way of inspecting machine
10013 instructions; see @ref{Machine Code,,Source and Machine Code}.
10015 If a negative repeat count is specified for the formats @samp{s} or @samp{i},
10016 the command displays null-terminated strings or instructions before the given
10017 address as many as the absolute value of the given number. For the @samp{i}
10018 format, we use line number information in the debug info to accurately locate
10019 instruction boundaries while disassembling backward. If line info is not
10020 available, the command stops examining memory with an error message.
10022 All the defaults for the arguments to @code{x} are designed to make it
10023 easy to continue scanning memory with minimal specifications each time
10024 you use @code{x}. For example, after you have inspected three machine
10025 instructions with @samp{x/3i @var{addr}}, you can inspect the next seven
10026 with just @samp{x/7}. If you use @key{RET} to repeat the @code{x} command,
10027 the repeat count @var{n} is used again; the other arguments default as
10028 for successive uses of @code{x}.
10030 When examining machine instructions, the instruction at current program
10031 counter is shown with a @code{=>} marker. For example:
10034 (@value{GDBP}) x/5i $pc-6
10035 0x804837f <main+11>: mov %esp,%ebp
10036 0x8048381 <main+13>: push %ecx
10037 0x8048382 <main+14>: sub $0x4,%esp
10038 => 0x8048385 <main+17>: movl $0x8048460,(%esp)
10039 0x804838c <main+24>: call 0x80482d4 <puts@@plt>
10042 @cindex @code{$_}, @code{$__}, and value history
10043 The addresses and contents printed by the @code{x} command are not saved
10044 in the value history because there is often too much of them and they
10045 would get in the way. Instead, @value{GDBN} makes these values available for
10046 subsequent use in expressions as values of the convenience variables
10047 @code{$_} and @code{$__}. After an @code{x} command, the last address
10048 examined is available for use in expressions in the convenience variable
10049 @code{$_}. The contents of that address, as examined, are available in
10050 the convenience variable @code{$__}.
10052 If the @code{x} command has a repeat count, the address and contents saved
10053 are from the last memory unit printed; this is not the same as the last
10054 address printed if several units were printed on the last line of output.
10056 @anchor{addressable memory unit}
10057 @cindex addressable memory unit
10058 Most targets have an addressable memory unit size of 8 bits. This means
10059 that to each memory address are associated 8 bits of data. Some
10060 targets, however, have other addressable memory unit sizes.
10061 Within @value{GDBN} and this document, the term
10062 @dfn{addressable memory unit} (or @dfn{memory unit} for short) is used
10063 when explicitly referring to a chunk of data of that size. The word
10064 @dfn{byte} is used to refer to a chunk of data of 8 bits, regardless of
10065 the addressable memory unit size of the target. For most systems,
10066 addressable memory unit is a synonym of byte.
10068 @cindex remote memory comparison
10069 @cindex target memory comparison
10070 @cindex verify remote memory image
10071 @cindex verify target memory image
10072 When you are debugging a program running on a remote target machine
10073 (@pxref{Remote Debugging}), you may wish to verify the program's image
10074 in the remote machine's memory against the executable file you
10075 downloaded to the target. Or, on any target, you may want to check
10076 whether the program has corrupted its own read-only sections. The
10077 @code{compare-sections} command is provided for such situations.
10080 @kindex compare-sections
10081 @item compare-sections @r{[}@var{section-name}@r{|}@code{-r}@r{]}
10082 Compare the data of a loadable section @var{section-name} in the
10083 executable file of the program being debugged with the same section in
10084 the target machine's memory, and report any mismatches. With no
10085 arguments, compares all loadable sections. With an argument of
10086 @code{-r}, compares all loadable read-only sections.
10088 Note: for remote targets, this command can be accelerated if the
10089 target supports computing the CRC checksum of a block of memory
10090 (@pxref{qCRC packet}).
10094 @section Automatic Display
10095 @cindex automatic display
10096 @cindex display of expressions
10098 If you find that you want to print the value of an expression frequently
10099 (to see how it changes), you might want to add it to the @dfn{automatic
10100 display list} so that @value{GDBN} prints its value each time your program stops.
10101 Each expression added to the list is given a number to identify it;
10102 to remove an expression from the list, you specify that number.
10103 The automatic display looks like this:
10107 3: bar[5] = (struct hack *) 0x3804
10111 This display shows item numbers, expressions and their current values. As with
10112 displays you request manually using @code{x} or @code{print}, you can
10113 specify the output format you prefer; in fact, @code{display} decides
10114 whether to use @code{print} or @code{x} depending your format
10115 specification---it uses @code{x} if you specify either the @samp{i}
10116 or @samp{s} format, or a unit size; otherwise it uses @code{print}.
10120 @item display @var{expr}
10121 Add the expression @var{expr} to the list of expressions to display
10122 each time your program stops. @xref{Expressions, ,Expressions}.
10124 @code{display} does not repeat if you press @key{RET} again after using it.
10126 @item display/@var{fmt} @var{expr}
10127 For @var{fmt} specifying only a display format and not a size or
10128 count, add the expression @var{expr} to the auto-display list but
10129 arrange to display it each time in the specified format @var{fmt}.
10130 @xref{Output Formats,,Output Formats}.
10132 @item display/@var{fmt} @var{addr}
10133 For @var{fmt} @samp{i} or @samp{s}, or including a unit-size or a
10134 number of units, add the expression @var{addr} as a memory address to
10135 be examined each time your program stops. Examining means in effect
10136 doing @samp{x/@var{fmt} @var{addr}}. @xref{Memory, ,Examining Memory}.
10139 For example, @samp{display/i $pc} can be helpful, to see the machine
10140 instruction about to be executed each time execution stops (@samp{$pc}
10141 is a common name for the program counter; @pxref{Registers, ,Registers}).
10144 @kindex delete display
10146 @item undisplay @var{dnums}@dots{}
10147 @itemx delete display @var{dnums}@dots{}
10148 Remove items from the list of expressions to display. Specify the
10149 numbers of the displays that you want affected with the command
10150 argument @var{dnums}. It can be a single display number, one of the
10151 numbers shown in the first field of the @samp{info display} display;
10152 or it could be a range of display numbers, as in @code{2-4}.
10154 @code{undisplay} does not repeat if you press @key{RET} after using it.
10155 (Otherwise you would just get the error @samp{No display number @dots{}}.)
10157 @kindex disable display
10158 @item disable display @var{dnums}@dots{}
10159 Disable the display of item numbers @var{dnums}. A disabled display
10160 item is not printed automatically, but is not forgotten. It may be
10161 enabled again later. Specify the numbers of the displays that you
10162 want affected with the command argument @var{dnums}. It can be a
10163 single display number, one of the numbers shown in the first field of
10164 the @samp{info display} display; or it could be a range of display
10165 numbers, as in @code{2-4}.
10167 @kindex enable display
10168 @item enable display @var{dnums}@dots{}
10169 Enable display of item numbers @var{dnums}. It becomes effective once
10170 again in auto display of its expression, until you specify otherwise.
10171 Specify the numbers of the displays that you want affected with the
10172 command argument @var{dnums}. It can be a single display number, one
10173 of the numbers shown in the first field of the @samp{info display}
10174 display; or it could be a range of display numbers, as in @code{2-4}.
10177 Display the current values of the expressions on the list, just as is
10178 done when your program stops.
10180 @kindex info display
10182 Print the list of expressions previously set up to display
10183 automatically, each one with its item number, but without showing the
10184 values. This includes disabled expressions, which are marked as such.
10185 It also includes expressions which would not be displayed right now
10186 because they refer to automatic variables not currently available.
10189 @cindex display disabled out of scope
10190 If a display expression refers to local variables, then it does not make
10191 sense outside the lexical context for which it was set up. Such an
10192 expression is disabled when execution enters a context where one of its
10193 variables is not defined. For example, if you give the command
10194 @code{display last_char} while inside a function with an argument
10195 @code{last_char}, @value{GDBN} displays this argument while your program
10196 continues to stop inside that function. When it stops elsewhere---where
10197 there is no variable @code{last_char}---the display is disabled
10198 automatically. The next time your program stops where @code{last_char}
10199 is meaningful, you can enable the display expression once again.
10201 @node Print Settings
10202 @section Print Settings
10204 @cindex format options
10205 @cindex print settings
10206 @value{GDBN} provides the following ways to control how arrays, structures,
10207 and symbols are printed.
10210 These settings are useful for debugging programs in any language:
10214 @item set print address
10215 @itemx set print address on
10216 @cindex print/don't print memory addresses
10217 @value{GDBN} prints memory addresses showing the location of stack
10218 traces, structure values, pointer values, breakpoints, and so forth,
10219 even when it also displays the contents of those addresses. The default
10220 is @code{on}. For example, this is what a stack frame display looks like with
10221 @code{set print address on}:
10226 #0 set_quotes (lq=0x34c78 "<<", rq=0x34c88 ">>")
10228 530 if (lquote != def_lquote)
10232 @item set print address off
10233 Do not print addresses when displaying their contents. For example,
10234 this is the same stack frame displayed with @code{set print address off}:
10238 (@value{GDBP}) set print addr off
10240 #0 set_quotes (lq="<<", rq=">>") at input.c:530
10241 530 if (lquote != def_lquote)
10245 You can use @samp{set print address off} to eliminate all machine
10246 dependent displays from the @value{GDBN} interface. For example, with
10247 @code{print address off}, you should get the same text for backtraces on
10248 all machines---whether or not they involve pointer arguments.
10251 @item show print address
10252 Show whether or not addresses are to be printed.
10255 When @value{GDBN} prints a symbolic address, it normally prints the
10256 closest earlier symbol plus an offset. If that symbol does not uniquely
10257 identify the address (for example, it is a name whose scope is a single
10258 source file), you may need to clarify. One way to do this is with
10259 @code{info line}, for example @samp{info line *0x4537}. Alternately,
10260 you can set @value{GDBN} to print the source file and line number when
10261 it prints a symbolic address:
10264 @item set print symbol-filename on
10265 @cindex source file and line of a symbol
10266 @cindex symbol, source file and line
10267 Tell @value{GDBN} to print the source file name and line number of a
10268 symbol in the symbolic form of an address.
10270 @item set print symbol-filename off
10271 Do not print source file name and line number of a symbol. This is the
10274 @item show print symbol-filename
10275 Show whether or not @value{GDBN} will print the source file name and
10276 line number of a symbol in the symbolic form of an address.
10279 Another situation where it is helpful to show symbol filenames and line
10280 numbers is when disassembling code; @value{GDBN} shows you the line
10281 number and source file that corresponds to each instruction.
10283 Also, you may wish to see the symbolic form only if the address being
10284 printed is reasonably close to the closest earlier symbol:
10287 @item set print max-symbolic-offset @var{max-offset}
10288 @itemx set print max-symbolic-offset unlimited
10289 @cindex maximum value for offset of closest symbol
10290 Tell @value{GDBN} to only display the symbolic form of an address if the
10291 offset between the closest earlier symbol and the address is less than
10292 @var{max-offset}. The default is @code{unlimited}, which tells @value{GDBN}
10293 to always print the symbolic form of an address if any symbol precedes
10294 it. Zero is equivalent to @code{unlimited}.
10296 @item show print max-symbolic-offset
10297 Ask how large the maximum offset is that @value{GDBN} prints in a
10301 @cindex wild pointer, interpreting
10302 @cindex pointer, finding referent
10303 If you have a pointer and you are not sure where it points, try
10304 @samp{set print symbol-filename on}. Then you can determine the name
10305 and source file location of the variable where it points, using
10306 @samp{p/a @var{pointer}}. This interprets the address in symbolic form.
10307 For example, here @value{GDBN} shows that a variable @code{ptt} points
10308 at another variable @code{t}, defined in @file{hi2.c}:
10311 (@value{GDBP}) set print symbol-filename on
10312 (@value{GDBP}) p/a ptt
10313 $4 = 0xe008 <t in hi2.c>
10317 @emph{Warning:} For pointers that point to a local variable, @samp{p/a}
10318 does not show the symbol name and filename of the referent, even with
10319 the appropriate @code{set print} options turned on.
10322 You can also enable @samp{/a}-like formatting all the time using
10323 @samp{set print symbol on}:
10326 @item set print symbol on
10327 Tell @value{GDBN} to print the symbol corresponding to an address, if
10330 @item set print symbol off
10331 Tell @value{GDBN} not to print the symbol corresponding to an
10332 address. In this mode, @value{GDBN} will still print the symbol
10333 corresponding to pointers to functions. This is the default.
10335 @item show print symbol
10336 Show whether @value{GDBN} will display the symbol corresponding to an
10340 Other settings control how different kinds of objects are printed:
10343 @item set print array
10344 @itemx set print array on
10345 @cindex pretty print arrays
10346 Pretty print arrays. This format is more convenient to read,
10347 but uses more space. The default is off.
10349 @item set print array off
10350 Return to compressed format for arrays.
10352 @item show print array
10353 Show whether compressed or pretty format is selected for displaying
10356 @cindex print array indexes
10357 @item set print array-indexes
10358 @itemx set print array-indexes on
10359 Print the index of each element when displaying arrays. May be more
10360 convenient to locate a given element in the array or quickly find the
10361 index of a given element in that printed array. The default is off.
10363 @item set print array-indexes off
10364 Stop printing element indexes when displaying arrays.
10366 @item show print array-indexes
10367 Show whether the index of each element is printed when displaying
10370 @item set print elements @var{number-of-elements}
10371 @itemx set print elements unlimited
10372 @cindex number of array elements to print
10373 @cindex limit on number of printed array elements
10374 Set a limit on how many elements of an array @value{GDBN} will print.
10375 If @value{GDBN} is printing a large array, it stops printing after it has
10376 printed the number of elements set by the @code{set print elements} command.
10377 This limit also applies to the display of strings.
10378 When @value{GDBN} starts, this limit is set to 200.
10379 Setting @var{number-of-elements} to @code{unlimited} or zero means
10380 that the number of elements to print is unlimited.
10382 @item show print elements
10383 Display the number of elements of a large array that @value{GDBN} will print.
10384 If the number is 0, then the printing is unlimited.
10386 @item set print frame-arguments @var{value}
10387 @kindex set print frame-arguments
10388 @cindex printing frame argument values
10389 @cindex print all frame argument values
10390 @cindex print frame argument values for scalars only
10391 @cindex do not print frame argument values
10392 This command allows to control how the values of arguments are printed
10393 when the debugger prints a frame (@pxref{Frames}). The possible
10398 The values of all arguments are printed.
10401 Print the value of an argument only if it is a scalar. The value of more
10402 complex arguments such as arrays, structures, unions, etc, is replaced
10403 by @code{@dots{}}. This is the default. Here is an example where
10404 only scalar arguments are shown:
10407 #1 0x08048361 in call_me (i=3, s=@dots{}, ss=0xbf8d508c, u=@dots{}, e=green)
10412 None of the argument values are printed. Instead, the value of each argument
10413 is replaced by @code{@dots{}}. In this case, the example above now becomes:
10416 #1 0x08048361 in call_me (i=@dots{}, s=@dots{}, ss=@dots{}, u=@dots{}, e=@dots{})
10421 By default, only scalar arguments are printed. This command can be used
10422 to configure the debugger to print the value of all arguments, regardless
10423 of their type. However, it is often advantageous to not print the value
10424 of more complex parameters. For instance, it reduces the amount of
10425 information printed in each frame, making the backtrace more readable.
10426 Also, it improves performance when displaying Ada frames, because
10427 the computation of large arguments can sometimes be CPU-intensive,
10428 especially in large applications. Setting @code{print frame-arguments}
10429 to @code{scalars} (the default) or @code{none} avoids this computation,
10430 thus speeding up the display of each Ada frame.
10432 @item show print frame-arguments
10433 Show how the value of arguments should be displayed when printing a frame.
10435 @item set print raw frame-arguments on
10436 Print frame arguments in raw, non pretty-printed, form.
10438 @item set print raw frame-arguments off
10439 Print frame arguments in pretty-printed form, if there is a pretty-printer
10440 for the value (@pxref{Pretty Printing}),
10441 otherwise print the value in raw form.
10442 This is the default.
10444 @item show print raw frame-arguments
10445 Show whether to print frame arguments in raw form.
10447 @anchor{set print entry-values}
10448 @item set print entry-values @var{value}
10449 @kindex set print entry-values
10450 Set printing of frame argument values at function entry. In some cases
10451 @value{GDBN} can determine the value of function argument which was passed by
10452 the function caller, even if the value was modified inside the called function
10453 and therefore is different. With optimized code, the current value could be
10454 unavailable, but the entry value may still be known.
10456 The default value is @code{default} (see below for its description). Older
10457 @value{GDBN} behaved as with the setting @code{no}. Compilers not supporting
10458 this feature will behave in the @code{default} setting the same way as with the
10461 This functionality is currently supported only by DWARF 2 debugging format and
10462 the compiler has to produce @samp{DW_TAG_call_site} tags. With
10463 @value{NGCC}, you need to specify @option{-O -g} during compilation, to get
10466 The @var{value} parameter can be one of the following:
10470 Print only actual parameter values, never print values from function entry
10474 #0 different (val=6)
10475 #0 lost (val=<optimized out>)
10477 #0 invalid (val=<optimized out>)
10481 Print only parameter values from function entry point. The actual parameter
10482 values are never printed.
10484 #0 equal (val@@entry=5)
10485 #0 different (val@@entry=5)
10486 #0 lost (val@@entry=5)
10487 #0 born (val@@entry=<optimized out>)
10488 #0 invalid (val@@entry=<optimized out>)
10492 Print only parameter values from function entry point. If value from function
10493 entry point is not known while the actual value is known, print the actual
10494 value for such parameter.
10496 #0 equal (val@@entry=5)
10497 #0 different (val@@entry=5)
10498 #0 lost (val@@entry=5)
10500 #0 invalid (val@@entry=<optimized out>)
10504 Print actual parameter values. If actual parameter value is not known while
10505 value from function entry point is known, print the entry point value for such
10509 #0 different (val=6)
10510 #0 lost (val@@entry=5)
10512 #0 invalid (val=<optimized out>)
10516 Always print both the actual parameter value and its value from function entry
10517 point, even if values of one or both are not available due to compiler
10520 #0 equal (val=5, val@@entry=5)
10521 #0 different (val=6, val@@entry=5)
10522 #0 lost (val=<optimized out>, val@@entry=5)
10523 #0 born (val=10, val@@entry=<optimized out>)
10524 #0 invalid (val=<optimized out>, val@@entry=<optimized out>)
10528 Print the actual parameter value if it is known and also its value from
10529 function entry point if it is known. If neither is known, print for the actual
10530 value @code{<optimized out>}. If not in MI mode (@pxref{GDB/MI}) and if both
10531 values are known and identical, print the shortened
10532 @code{param=param@@entry=VALUE} notation.
10534 #0 equal (val=val@@entry=5)
10535 #0 different (val=6, val@@entry=5)
10536 #0 lost (val@@entry=5)
10538 #0 invalid (val=<optimized out>)
10542 Always print the actual parameter value. Print also its value from function
10543 entry point, but only if it is known. If not in MI mode (@pxref{GDB/MI}) and
10544 if both values are known and identical, print the shortened
10545 @code{param=param@@entry=VALUE} notation.
10547 #0 equal (val=val@@entry=5)
10548 #0 different (val=6, val@@entry=5)
10549 #0 lost (val=<optimized out>, val@@entry=5)
10551 #0 invalid (val=<optimized out>)
10555 For analysis messages on possible failures of frame argument values at function
10556 entry resolution see @ref{set debug entry-values}.
10558 @item show print entry-values
10559 Show the method being used for printing of frame argument values at function
10562 @item set print repeats @var{number-of-repeats}
10563 @itemx set print repeats unlimited
10564 @cindex repeated array elements
10565 Set the threshold for suppressing display of repeated array
10566 elements. When the number of consecutive identical elements of an
10567 array exceeds the threshold, @value{GDBN} prints the string
10568 @code{"<repeats @var{n} times>"}, where @var{n} is the number of
10569 identical repetitions, instead of displaying the identical elements
10570 themselves. Setting the threshold to @code{unlimited} or zero will
10571 cause all elements to be individually printed. The default threshold
10574 @item show print repeats
10575 Display the current threshold for printing repeated identical
10578 @item set print max-depth @var{depth}
10579 @item set print max-depth unlimited
10580 @cindex printing nested structures
10581 Set the threshold after which nested structures are replaced with
10582 ellipsis, this can make visualising deeply nested structures easier.
10584 For example, given this C code
10587 typedef struct s1 @{ int a; @} s1;
10588 typedef struct s2 @{ s1 b; @} s2;
10589 typedef struct s3 @{ s2 c; @} s3;
10590 typedef struct s4 @{ s3 d; @} s4;
10592 s4 var = @{ @{ @{ @{ 3 @} @} @} @};
10595 The following table shows how different values of @var{depth} will
10596 effect how @code{var} is printed by @value{GDBN}:
10598 @multitable @columnfractions .3 .7
10599 @headitem @var{depth} setting @tab Result of @samp{p var}
10601 @tab @code{$1 = @{d = @{c = @{b = @{a = 3@}@}@}@}}
10603 @tab @code{$1 = @{...@}}
10605 @tab @code{$1 = @{d = @{...@}@}}
10607 @tab @code{$1 = @{d = @{c = @{...@}@}@}}
10609 @tab @code{$1 = @{d = @{c = @{b = @{...@}@}@}@}}
10611 @tab @code{$1 = @{d = @{c = @{b = @{a = 3@}@}@}@}}
10614 To see the contents of structures that have been hidden the user can
10615 either increase the print max-depth, or they can print the elements of
10616 the structure that are visible, for example
10619 (gdb) set print max-depth 2
10621 $1 = @{d = @{c = @{...@}@}@}
10623 $2 = @{c = @{b = @{...@}@}@}
10625 $3 = @{b = @{a = 3@}@}
10628 The pattern used to replace nested structures varies based on
10629 language, for most languages @code{@{...@}} is used, but Fortran uses
10632 @item show print max-depth
10633 Display the current threshold after which nested structures are
10634 replaces with ellipsis.
10636 @item set print null-stop
10637 @cindex @sc{null} elements in arrays
10638 Cause @value{GDBN} to stop printing the characters of an array when the first
10639 @sc{null} is encountered. This is useful when large arrays actually
10640 contain only short strings.
10641 The default is off.
10643 @item show print null-stop
10644 Show whether @value{GDBN} stops printing an array on the first
10645 @sc{null} character.
10647 @item set print pretty on
10648 @cindex print structures in indented form
10649 @cindex indentation in structure display
10650 Cause @value{GDBN} to print structures in an indented format with one member
10651 per line, like this:
10666 @item set print pretty off
10667 Cause @value{GDBN} to print structures in a compact format, like this:
10671 $1 = @{next = 0x0, flags = @{sweet = 1, sour = 1@}, \
10672 meat = 0x54 "Pork"@}
10677 This is the default format.
10679 @item show print pretty
10680 Show which format @value{GDBN} is using to print structures.
10682 @item set print sevenbit-strings on
10683 @cindex eight-bit characters in strings
10684 @cindex octal escapes in strings
10685 Print using only seven-bit characters; if this option is set,
10686 @value{GDBN} displays any eight-bit characters (in strings or
10687 character values) using the notation @code{\}@var{nnn}. This setting is
10688 best if you are working in English (@sc{ascii}) and you use the
10689 high-order bit of characters as a marker or ``meta'' bit.
10691 @item set print sevenbit-strings off
10692 Print full eight-bit characters. This allows the use of more
10693 international character sets, and is the default.
10695 @item show print sevenbit-strings
10696 Show whether or not @value{GDBN} is printing only seven-bit characters.
10698 @item set print union on
10699 @cindex unions in structures, printing
10700 Tell @value{GDBN} to print unions which are contained in structures
10701 and other unions. This is the default setting.
10703 @item set print union off
10704 Tell @value{GDBN} not to print unions which are contained in
10705 structures and other unions. @value{GDBN} will print @code{"@{...@}"}
10708 @item show print union
10709 Ask @value{GDBN} whether or not it will print unions which are contained in
10710 structures and other unions.
10712 For example, given the declarations
10715 typedef enum @{Tree, Bug@} Species;
10716 typedef enum @{Big_tree, Acorn, Seedling@} Tree_forms;
10717 typedef enum @{Caterpillar, Cocoon, Butterfly@}
10728 struct thing foo = @{Tree, @{Acorn@}@};
10732 with @code{set print union on} in effect @samp{p foo} would print
10735 $1 = @{it = Tree, form = @{tree = Acorn, bug = Cocoon@}@}
10739 and with @code{set print union off} in effect it would print
10742 $1 = @{it = Tree, form = @{...@}@}
10746 @code{set print union} affects programs written in C-like languages
10752 These settings are of interest when debugging C@t{++} programs:
10755 @cindex demangling C@t{++} names
10756 @item set print demangle
10757 @itemx set print demangle on
10758 Print C@t{++} names in their source form rather than in the encoded
10759 (``mangled'') form passed to the assembler and linker for type-safe
10760 linkage. The default is on.
10762 @item show print demangle
10763 Show whether C@t{++} names are printed in mangled or demangled form.
10765 @item set print asm-demangle
10766 @itemx set print asm-demangle on
10767 Print C@t{++} names in their source form rather than their mangled form, even
10768 in assembler code printouts such as instruction disassemblies.
10769 The default is off.
10771 @item show print asm-demangle
10772 Show whether C@t{++} names in assembly listings are printed in mangled
10775 @cindex C@t{++} symbol decoding style
10776 @cindex symbol decoding style, C@t{++}
10777 @kindex set demangle-style
10778 @item set demangle-style @var{style}
10779 Choose among several encoding schemes used by different compilers to represent
10780 C@t{++} names. If you omit @var{style}, you will see a list of possible
10781 formats. The default value is @var{auto}, which lets @value{GDBN} choose a
10782 decoding style by inspecting your program.
10784 @item show demangle-style
10785 Display the encoding style currently in use for decoding C@t{++} symbols.
10787 @item set print object
10788 @itemx set print object on
10789 @cindex derived type of an object, printing
10790 @cindex display derived types
10791 When displaying a pointer to an object, identify the @emph{actual}
10792 (derived) type of the object rather than the @emph{declared} type, using
10793 the virtual function table. Note that the virtual function table is
10794 required---this feature can only work for objects that have run-time
10795 type identification; a single virtual method in the object's declared
10796 type is sufficient. Note that this setting is also taken into account when
10797 working with variable objects via MI (@pxref{GDB/MI}).
10799 @item set print object off
10800 Display only the declared type of objects, without reference to the
10801 virtual function table. This is the default setting.
10803 @item show print object
10804 Show whether actual, or declared, object types are displayed.
10806 @item set print static-members
10807 @itemx set print static-members on
10808 @cindex static members of C@t{++} objects
10809 Print static members when displaying a C@t{++} object. The default is on.
10811 @item set print static-members off
10812 Do not print static members when displaying a C@t{++} object.
10814 @item show print static-members
10815 Show whether C@t{++} static members are printed or not.
10817 @item set print pascal_static-members
10818 @itemx set print pascal_static-members on
10819 @cindex static members of Pascal objects
10820 @cindex Pascal objects, static members display
10821 Print static members when displaying a Pascal object. The default is on.
10823 @item set print pascal_static-members off
10824 Do not print static members when displaying a Pascal object.
10826 @item show print pascal_static-members
10827 Show whether Pascal static members are printed or not.
10829 @c These don't work with HP ANSI C++ yet.
10830 @item set print vtbl
10831 @itemx set print vtbl on
10832 @cindex pretty print C@t{++} virtual function tables
10833 @cindex virtual functions (C@t{++}) display
10834 @cindex VTBL display
10835 Pretty print C@t{++} virtual function tables. The default is off.
10836 (The @code{vtbl} commands do not work on programs compiled with the HP
10837 ANSI C@t{++} compiler (@code{aCC}).)
10839 @item set print vtbl off
10840 Do not pretty print C@t{++} virtual function tables.
10842 @item show print vtbl
10843 Show whether C@t{++} virtual function tables are pretty printed, or not.
10846 @node Pretty Printing
10847 @section Pretty Printing
10849 @value{GDBN} provides a mechanism to allow pretty-printing of values using
10850 Python code. It greatly simplifies the display of complex objects. This
10851 mechanism works for both MI and the CLI.
10854 * Pretty-Printer Introduction:: Introduction to pretty-printers
10855 * Pretty-Printer Example:: An example pretty-printer
10856 * Pretty-Printer Commands:: Pretty-printer commands
10859 @node Pretty-Printer Introduction
10860 @subsection Pretty-Printer Introduction
10862 When @value{GDBN} prints a value, it first sees if there is a pretty-printer
10863 registered for the value. If there is then @value{GDBN} invokes the
10864 pretty-printer to print the value. Otherwise the value is printed normally.
10866 Pretty-printers are normally named. This makes them easy to manage.
10867 The @samp{info pretty-printer} command will list all the installed
10868 pretty-printers with their names.
10869 If a pretty-printer can handle multiple data types, then its
10870 @dfn{subprinters} are the printers for the individual data types.
10871 Each such subprinter has its own name.
10872 The format of the name is @var{printer-name};@var{subprinter-name}.
10874 Pretty-printers are installed by @dfn{registering} them with @value{GDBN}.
10875 Typically they are automatically loaded and registered when the corresponding
10876 debug information is loaded, thus making them available without having to
10877 do anything special.
10879 There are three places where a pretty-printer can be registered.
10883 Pretty-printers registered globally are available when debugging
10887 Pretty-printers registered with a program space are available only
10888 when debugging that program.
10889 @xref{Progspaces In Python}, for more details on program spaces in Python.
10892 Pretty-printers registered with an objfile are loaded and unloaded
10893 with the corresponding objfile (e.g., shared library).
10894 @xref{Objfiles In Python}, for more details on objfiles in Python.
10897 @xref{Selecting Pretty-Printers}, for further information on how
10898 pretty-printers are selected,
10900 @xref{Writing a Pretty-Printer}, for implementing pretty printers
10903 @node Pretty-Printer Example
10904 @subsection Pretty-Printer Example
10906 Here is how a C@t{++} @code{std::string} looks without a pretty-printer:
10909 (@value{GDBP}) print s
10911 static npos = 4294967295,
10913 <std::allocator<char>> = @{
10914 <__gnu_cxx::new_allocator<char>> = @{
10915 <No data fields>@}, <No data fields>
10917 members of std::basic_string<char, std::char_traits<char>,
10918 std::allocator<char> >::_Alloc_hider:
10919 _M_p = 0x804a014 "abcd"
10924 With a pretty-printer for @code{std::string} only the contents are printed:
10927 (@value{GDBP}) print s
10931 @node Pretty-Printer Commands
10932 @subsection Pretty-Printer Commands
10933 @cindex pretty-printer commands
10936 @kindex info pretty-printer
10937 @item info pretty-printer [@var{object-regexp} [@var{name-regexp}]]
10938 Print the list of installed pretty-printers.
10939 This includes disabled pretty-printers, which are marked as such.
10941 @var{object-regexp} is a regular expression matching the objects
10942 whose pretty-printers to list.
10943 Objects can be @code{global}, the program space's file
10944 (@pxref{Progspaces In Python}),
10945 and the object files within that program space (@pxref{Objfiles In Python}).
10946 @xref{Selecting Pretty-Printers}, for details on how @value{GDBN}
10947 looks up a printer from these three objects.
10949 @var{name-regexp} is a regular expression matching the name of the printers
10952 @kindex disable pretty-printer
10953 @item disable pretty-printer [@var{object-regexp} [@var{name-regexp}]]
10954 Disable pretty-printers matching @var{object-regexp} and @var{name-regexp}.
10955 A disabled pretty-printer is not forgotten, it may be enabled again later.
10957 @kindex enable pretty-printer
10958 @item enable pretty-printer [@var{object-regexp} [@var{name-regexp}]]
10959 Enable pretty-printers matching @var{object-regexp} and @var{name-regexp}.
10964 Suppose we have three pretty-printers installed: one from library1.so
10965 named @code{foo} that prints objects of type @code{foo}, and
10966 another from library2.so named @code{bar} that prints two types of objects,
10967 @code{bar1} and @code{bar2}.
10970 (gdb) info pretty-printer
10977 (gdb) info pretty-printer library2
10982 (gdb) disable pretty-printer library1
10984 2 of 3 printers enabled
10985 (gdb) info pretty-printer
10992 (gdb) disable pretty-printer library2 bar;bar1
10994 1 of 3 printers enabled
10995 (gdb) info pretty-printer library2
11002 (gdb) disable pretty-printer library2 bar
11004 0 of 3 printers enabled
11005 (gdb) info pretty-printer library2
11014 Note that for @code{bar} the entire printer can be disabled,
11015 as can each individual subprinter.
11017 @node Value History
11018 @section Value History
11020 @cindex value history
11021 @cindex history of values printed by @value{GDBN}
11022 Values printed by the @code{print} command are saved in the @value{GDBN}
11023 @dfn{value history}. This allows you to refer to them in other expressions.
11024 Values are kept until the symbol table is re-read or discarded
11025 (for example with the @code{file} or @code{symbol-file} commands).
11026 When the symbol table changes, the value history is discarded,
11027 since the values may contain pointers back to the types defined in the
11032 @cindex history number
11033 The values printed are given @dfn{history numbers} by which you can
11034 refer to them. These are successive integers starting with one.
11035 @code{print} shows you the history number assigned to a value by
11036 printing @samp{$@var{num} = } before the value; here @var{num} is the
11039 To refer to any previous value, use @samp{$} followed by the value's
11040 history number. The way @code{print} labels its output is designed to
11041 remind you of this. Just @code{$} refers to the most recent value in
11042 the history, and @code{$$} refers to the value before that.
11043 @code{$$@var{n}} refers to the @var{n}th value from the end; @code{$$2}
11044 is the value just prior to @code{$$}, @code{$$1} is equivalent to
11045 @code{$$}, and @code{$$0} is equivalent to @code{$}.
11047 For example, suppose you have just printed a pointer to a structure and
11048 want to see the contents of the structure. It suffices to type
11054 If you have a chain of structures where the component @code{next} points
11055 to the next one, you can print the contents of the next one with this:
11062 You can print successive links in the chain by repeating this
11063 command---which you can do by just typing @key{RET}.
11065 Note that the history records values, not expressions. If the value of
11066 @code{x} is 4 and you type these commands:
11074 then the value recorded in the value history by the @code{print} command
11075 remains 4 even though the value of @code{x} has changed.
11078 @kindex show values
11080 Print the last ten values in the value history, with their item numbers.
11081 This is like @samp{p@ $$9} repeated ten times, except that @code{show
11082 values} does not change the history.
11084 @item show values @var{n}
11085 Print ten history values centered on history item number @var{n}.
11087 @item show values +
11088 Print ten history values just after the values last printed. If no more
11089 values are available, @code{show values +} produces no display.
11092 Pressing @key{RET} to repeat @code{show values @var{n}} has exactly the
11093 same effect as @samp{show values +}.
11095 @node Convenience Vars
11096 @section Convenience Variables
11098 @cindex convenience variables
11099 @cindex user-defined variables
11100 @value{GDBN} provides @dfn{convenience variables} that you can use within
11101 @value{GDBN} to hold on to a value and refer to it later. These variables
11102 exist entirely within @value{GDBN}; they are not part of your program, and
11103 setting a convenience variable has no direct effect on further execution
11104 of your program. That is why you can use them freely.
11106 Convenience variables are prefixed with @samp{$}. Any name preceded by
11107 @samp{$} can be used for a convenience variable, unless it is one of
11108 the predefined machine-specific register names (@pxref{Registers, ,Registers}).
11109 (Value history references, in contrast, are @emph{numbers} preceded
11110 by @samp{$}. @xref{Value History, ,Value History}.)
11112 You can save a value in a convenience variable with an assignment
11113 expression, just as you would set a variable in your program.
11117 set $foo = *object_ptr
11121 would save in @code{$foo} the value contained in the object pointed to by
11124 Using a convenience variable for the first time creates it, but its
11125 value is @code{void} until you assign a new value. You can alter the
11126 value with another assignment at any time.
11128 Convenience variables have no fixed types. You can assign a convenience
11129 variable any type of value, including structures and arrays, even if
11130 that variable already has a value of a different type. The convenience
11131 variable, when used as an expression, has the type of its current value.
11134 @kindex show convenience
11135 @cindex show all user variables and functions
11136 @item show convenience
11137 Print a list of convenience variables used so far, and their values,
11138 as well as a list of the convenience functions.
11139 Abbreviated @code{show conv}.
11141 @kindex init-if-undefined
11142 @cindex convenience variables, initializing
11143 @item init-if-undefined $@var{variable} = @var{expression}
11144 Set a convenience variable if it has not already been set. This is useful
11145 for user-defined commands that keep some state. It is similar, in concept,
11146 to using local static variables with initializers in C (except that
11147 convenience variables are global). It can also be used to allow users to
11148 override default values used in a command script.
11150 If the variable is already defined then the expression is not evaluated so
11151 any side-effects do not occur.
11154 One of the ways to use a convenience variable is as a counter to be
11155 incremented or a pointer to be advanced. For example, to print
11156 a field from successive elements of an array of structures:
11160 print bar[$i++]->contents
11164 Repeat that command by typing @key{RET}.
11166 Some convenience variables are created automatically by @value{GDBN} and given
11167 values likely to be useful.
11170 @vindex $_@r{, convenience variable}
11172 The variable @code{$_} is automatically set by the @code{x} command to
11173 the last address examined (@pxref{Memory, ,Examining Memory}). Other
11174 commands which provide a default address for @code{x} to examine also
11175 set @code{$_} to that address; these commands include @code{info line}
11176 and @code{info breakpoint}. The type of @code{$_} is @code{void *}
11177 except when set by the @code{x} command, in which case it is a pointer
11178 to the type of @code{$__}.
11180 @vindex $__@r{, convenience variable}
11182 The variable @code{$__} is automatically set by the @code{x} command
11183 to the value found in the last address examined. Its type is chosen
11184 to match the format in which the data was printed.
11187 @vindex $_exitcode@r{, convenience variable}
11188 When the program being debugged terminates normally, @value{GDBN}
11189 automatically sets this variable to the exit code of the program, and
11190 resets @code{$_exitsignal} to @code{void}.
11193 @vindex $_exitsignal@r{, convenience variable}
11194 When the program being debugged dies due to an uncaught signal,
11195 @value{GDBN} automatically sets this variable to that signal's number,
11196 and resets @code{$_exitcode} to @code{void}.
11198 To distinguish between whether the program being debugged has exited
11199 (i.e., @code{$_exitcode} is not @code{void}) or signalled (i.e.,
11200 @code{$_exitsignal} is not @code{void}), the convenience function
11201 @code{$_isvoid} can be used (@pxref{Convenience Funs,, Convenience
11202 Functions}). For example, considering the following source code:
11205 #include <signal.h>
11208 main (int argc, char *argv[])
11215 A valid way of telling whether the program being debugged has exited
11216 or signalled would be:
11219 (@value{GDBP}) define has_exited_or_signalled
11220 Type commands for definition of ``has_exited_or_signalled''.
11221 End with a line saying just ``end''.
11222 >if $_isvoid ($_exitsignal)
11223 >echo The program has exited\n
11225 >echo The program has signalled\n
11231 Program terminated with signal SIGALRM, Alarm clock.
11232 The program no longer exists.
11233 (@value{GDBP}) has_exited_or_signalled
11234 The program has signalled
11237 As can be seen, @value{GDBN} correctly informs that the program being
11238 debugged has signalled, since it calls @code{raise} and raises a
11239 @code{SIGALRM} signal. If the program being debugged had not called
11240 @code{raise}, then @value{GDBN} would report a normal exit:
11243 (@value{GDBP}) has_exited_or_signalled
11244 The program has exited
11248 The variable @code{$_exception} is set to the exception object being
11249 thrown at an exception-related catchpoint. @xref{Set Catchpoints}.
11252 @itemx $_probe_arg0@dots{}$_probe_arg11
11253 Arguments to a static probe. @xref{Static Probe Points}.
11256 @vindex $_sdata@r{, inspect, convenience variable}
11257 The variable @code{$_sdata} contains extra collected static tracepoint
11258 data. @xref{Tracepoint Actions,,Tracepoint Action Lists}. Note that
11259 @code{$_sdata} could be empty, if not inspecting a trace buffer, or
11260 if extra static tracepoint data has not been collected.
11263 @vindex $_siginfo@r{, convenience variable}
11264 The variable @code{$_siginfo} contains extra signal information
11265 (@pxref{extra signal information}). Note that @code{$_siginfo}
11266 could be empty, if the application has not yet received any signals.
11267 For example, it will be empty before you execute the @code{run} command.
11270 @vindex $_tlb@r{, convenience variable}
11271 The variable @code{$_tlb} is automatically set when debugging
11272 applications running on MS-Windows in native mode or connected to
11273 gdbserver that supports the @code{qGetTIBAddr} request.
11274 @xref{General Query Packets}.
11275 This variable contains the address of the thread information block.
11278 The number of the current inferior. @xref{Inferiors and
11279 Programs, ,Debugging Multiple Inferiors and Programs}.
11282 The thread number of the current thread. @xref{thread numbers}.
11285 The global number of the current thread. @xref{global thread numbers}.
11289 @vindex $_gdb_major@r{, convenience variable}
11290 @vindex $_gdb_minor@r{, convenience variable}
11291 The major and minor version numbers of the running @value{GDBN}.
11292 Development snapshots and pretest versions have their minor version
11293 incremented by one; thus, @value{GDBN} pretest 9.11.90 will produce
11294 the value 12 for @code{$_gdb_minor}. These variables allow you to
11295 write scripts that work with different versions of @value{GDBN}
11296 without errors caused by features unavailable in some of those
11300 @node Convenience Funs
11301 @section Convenience Functions
11303 @cindex convenience functions
11304 @value{GDBN} also supplies some @dfn{convenience functions}. These
11305 have a syntax similar to convenience variables. A convenience
11306 function can be used in an expression just like an ordinary function;
11307 however, a convenience function is implemented internally to
11310 These functions do not require @value{GDBN} to be configured with
11311 @code{Python} support, which means that they are always available.
11315 @item $_isvoid (@var{expr})
11316 @findex $_isvoid@r{, convenience function}
11317 Return one if the expression @var{expr} is @code{void}. Otherwise it
11320 A @code{void} expression is an expression where the type of the result
11321 is @code{void}. For example, you can examine a convenience variable
11322 (see @ref{Convenience Vars,, Convenience Variables}) to check whether
11326 (@value{GDBP}) print $_exitcode
11328 (@value{GDBP}) print $_isvoid ($_exitcode)
11331 Starting program: ./a.out
11332 [Inferior 1 (process 29572) exited normally]
11333 (@value{GDBP}) print $_exitcode
11335 (@value{GDBP}) print $_isvoid ($_exitcode)
11339 In the example above, we used @code{$_isvoid} to check whether
11340 @code{$_exitcode} is @code{void} before and after the execution of the
11341 program being debugged. Before the execution there is no exit code to
11342 be examined, therefore @code{$_exitcode} is @code{void}. After the
11343 execution the program being debugged returned zero, therefore
11344 @code{$_exitcode} is zero, which means that it is not @code{void}
11347 The @code{void} expression can also be a call of a function from the
11348 program being debugged. For example, given the following function:
11357 The result of calling it inside @value{GDBN} is @code{void}:
11360 (@value{GDBP}) print foo ()
11362 (@value{GDBP}) print $_isvoid (foo ())
11364 (@value{GDBP}) set $v = foo ()
11365 (@value{GDBP}) print $v
11367 (@value{GDBP}) print $_isvoid ($v)
11373 These functions require @value{GDBN} to be configured with
11374 @code{Python} support.
11378 @item $_memeq(@var{buf1}, @var{buf2}, @var{length})
11379 @findex $_memeq@r{, convenience function}
11380 Returns one if the @var{length} bytes at the addresses given by
11381 @var{buf1} and @var{buf2} are equal.
11382 Otherwise it returns zero.
11384 @item $_regex(@var{str}, @var{regex})
11385 @findex $_regex@r{, convenience function}
11386 Returns one if the string @var{str} matches the regular expression
11387 @var{regex}. Otherwise it returns zero.
11388 The syntax of the regular expression is that specified by @code{Python}'s
11389 regular expression support.
11391 @item $_streq(@var{str1}, @var{str2})
11392 @findex $_streq@r{, convenience function}
11393 Returns one if the strings @var{str1} and @var{str2} are equal.
11394 Otherwise it returns zero.
11396 @item $_strlen(@var{str})
11397 @findex $_strlen@r{, convenience function}
11398 Returns the length of string @var{str}.
11400 @item $_caller_is(@var{name}@r{[}, @var{number_of_frames}@r{]})
11401 @findex $_caller_is@r{, convenience function}
11402 Returns one if the calling function's name is equal to @var{name}.
11403 Otherwise it returns zero.
11405 If the optional argument @var{number_of_frames} is provided,
11406 it is the number of frames up in the stack to look.
11414 at testsuite/gdb.python/py-caller-is.c:21
11415 #1 0x00000000004005a0 in middle_func ()
11416 at testsuite/gdb.python/py-caller-is.c:27
11417 #2 0x00000000004005ab in top_func ()
11418 at testsuite/gdb.python/py-caller-is.c:33
11419 #3 0x00000000004005b6 in main ()
11420 at testsuite/gdb.python/py-caller-is.c:39
11421 (gdb) print $_caller_is ("middle_func")
11423 (gdb) print $_caller_is ("top_func", 2)
11427 @item $_caller_matches(@var{regexp}@r{[}, @var{number_of_frames}@r{]})
11428 @findex $_caller_matches@r{, convenience function}
11429 Returns one if the calling function's name matches the regular expression
11430 @var{regexp}. Otherwise it returns zero.
11432 If the optional argument @var{number_of_frames} is provided,
11433 it is the number of frames up in the stack to look.
11436 @item $_any_caller_is(@var{name}@r{[}, @var{number_of_frames}@r{]})
11437 @findex $_any_caller_is@r{, convenience function}
11438 Returns one if any calling function's name is equal to @var{name}.
11439 Otherwise it returns zero.
11441 If the optional argument @var{number_of_frames} is provided,
11442 it is the number of frames up in the stack to look.
11445 This function differs from @code{$_caller_is} in that this function
11446 checks all stack frames from the immediate caller to the frame specified
11447 by @var{number_of_frames}, whereas @code{$_caller_is} only checks the
11448 frame specified by @var{number_of_frames}.
11450 @item $_any_caller_matches(@var{regexp}@r{[}, @var{number_of_frames}@r{]})
11451 @findex $_any_caller_matches@r{, convenience function}
11452 Returns one if any calling function's name matches the regular expression
11453 @var{regexp}. Otherwise it returns zero.
11455 If the optional argument @var{number_of_frames} is provided,
11456 it is the number of frames up in the stack to look.
11459 This function differs from @code{$_caller_matches} in that this function
11460 checks all stack frames from the immediate caller to the frame specified
11461 by @var{number_of_frames}, whereas @code{$_caller_matches} only checks the
11462 frame specified by @var{number_of_frames}.
11464 @item $_as_string(@var{value})
11465 @findex $_as_string@r{, convenience function}
11466 Return the string representation of @var{value}.
11468 This function is useful to obtain the textual label (enumerator) of an
11469 enumeration value. For example, assuming the variable @var{node} is of
11470 an enumerated type:
11473 (gdb) printf "Visiting node of type %s\n", $_as_string(node)
11474 Visiting node of type NODE_INTEGER
11477 @item $_cimag(@var{value})
11478 @itemx $_creal(@var{value})
11479 @findex $_cimag@r{, convenience function}
11480 @findex $_creal@r{, convenience function}
11481 Return the imaginary (@code{$_cimag}) or real (@code{$_creal}) part of
11482 the complex number @var{value}.
11484 The type of the imaginary or real part depends on the type of the
11485 complex number, e.g., using @code{$_cimag} on a @code{float complex}
11486 will return an imaginary part of type @code{float}.
11490 @value{GDBN} provides the ability to list and get help on
11491 convenience functions.
11494 @item help function
11495 @kindex help function
11496 @cindex show all convenience functions
11497 Print a list of all convenience functions.
11504 You can refer to machine register contents, in expressions, as variables
11505 with names starting with @samp{$}. The names of registers are different
11506 for each machine; use @code{info registers} to see the names used on
11510 @kindex info registers
11511 @item info registers
11512 Print the names and values of all registers except floating-point
11513 and vector registers (in the selected stack frame).
11515 @kindex info all-registers
11516 @cindex floating point registers
11517 @item info all-registers
11518 Print the names and values of all registers, including floating-point
11519 and vector registers (in the selected stack frame).
11521 @item info registers @var{reggroup} @dots{}
11522 Print the name and value of the registers in each of the specified
11523 @var{reggroup}s. The @var{reggoup} can be any of those returned by
11524 @code{maint print reggroups} (@pxref{Maintenance Commands}).
11526 @item info registers @var{regname} @dots{}
11527 Print the @dfn{relativized} value of each specified register @var{regname}.
11528 As discussed in detail below, register values are normally relative to
11529 the selected stack frame. The @var{regname} may be any register name valid on
11530 the machine you are using, with or without the initial @samp{$}.
11533 @anchor{standard registers}
11534 @cindex stack pointer register
11535 @cindex program counter register
11536 @cindex process status register
11537 @cindex frame pointer register
11538 @cindex standard registers
11539 @value{GDBN} has four ``standard'' register names that are available (in
11540 expressions) on most machines---whenever they do not conflict with an
11541 architecture's canonical mnemonics for registers. The register names
11542 @code{$pc} and @code{$sp} are used for the program counter register and
11543 the stack pointer. @code{$fp} is used for a register that contains a
11544 pointer to the current stack frame, and @code{$ps} is used for a
11545 register that contains the processor status. For example,
11546 you could print the program counter in hex with
11553 or print the instruction to be executed next with
11560 or add four to the stack pointer@footnote{This is a way of removing
11561 one word from the stack, on machines where stacks grow downward in
11562 memory (most machines, nowadays). This assumes that the innermost
11563 stack frame is selected; setting @code{$sp} is not allowed when other
11564 stack frames are selected. To pop entire frames off the stack,
11565 regardless of machine architecture, use @code{return};
11566 see @ref{Returning, ,Returning from a Function}.} with
11572 Whenever possible, these four standard register names are available on
11573 your machine even though the machine has different canonical mnemonics,
11574 so long as there is no conflict. The @code{info registers} command
11575 shows the canonical names. For example, on the SPARC, @code{info
11576 registers} displays the processor status register as @code{$psr} but you
11577 can also refer to it as @code{$ps}; and on x86-based machines @code{$ps}
11578 is an alias for the @sc{eflags} register.
11580 @value{GDBN} always considers the contents of an ordinary register as an
11581 integer when the register is examined in this way. Some machines have
11582 special registers which can hold nothing but floating point; these
11583 registers are considered to have floating point values. There is no way
11584 to refer to the contents of an ordinary register as floating point value
11585 (although you can @emph{print} it as a floating point value with
11586 @samp{print/f $@var{regname}}).
11588 Some registers have distinct ``raw'' and ``virtual'' data formats. This
11589 means that the data format in which the register contents are saved by
11590 the operating system is not the same one that your program normally
11591 sees. For example, the registers of the 68881 floating point
11592 coprocessor are always saved in ``extended'' (raw) format, but all C
11593 programs expect to work with ``double'' (virtual) format. In such
11594 cases, @value{GDBN} normally works with the virtual format only (the format
11595 that makes sense for your program), but the @code{info registers} command
11596 prints the data in both formats.
11598 @cindex SSE registers (x86)
11599 @cindex MMX registers (x86)
11600 Some machines have special registers whose contents can be interpreted
11601 in several different ways. For example, modern x86-based machines
11602 have SSE and MMX registers that can hold several values packed
11603 together in several different formats. @value{GDBN} refers to such
11604 registers in @code{struct} notation:
11607 (@value{GDBP}) print $xmm1
11609 v4_float = @{0, 3.43859137e-038, 1.54142831e-044, 1.821688e-044@},
11610 v2_double = @{9.92129282474342e-303, 2.7585945287983262e-313@},
11611 v16_int8 = "\000\000\000\000\3706;\001\v\000\000\000\r\000\000",
11612 v8_int16 = @{0, 0, 14072, 315, 11, 0, 13, 0@},
11613 v4_int32 = @{0, 20657912, 11, 13@},
11614 v2_int64 = @{88725056443645952, 55834574859@},
11615 uint128 = 0x0000000d0000000b013b36f800000000
11620 To set values of such registers, you need to tell @value{GDBN} which
11621 view of the register you wish to change, as if you were assigning
11622 value to a @code{struct} member:
11625 (@value{GDBP}) set $xmm1.uint128 = 0x000000000000000000000000FFFFFFFF
11628 Normally, register values are relative to the selected stack frame
11629 (@pxref{Selection, ,Selecting a Frame}). This means that you get the
11630 value that the register would contain if all stack frames farther in
11631 were exited and their saved registers restored. In order to see the
11632 true contents of hardware registers, you must select the innermost
11633 frame (with @samp{frame 0}).
11635 @cindex caller-saved registers
11636 @cindex call-clobbered registers
11637 @cindex volatile registers
11638 @cindex <not saved> values
11639 Usually ABIs reserve some registers as not needed to be saved by the
11640 callee (a.k.a.: ``caller-saved'', ``call-clobbered'' or ``volatile''
11641 registers). It may therefore not be possible for @value{GDBN} to know
11642 the value a register had before the call (in other words, in the outer
11643 frame), if the register value has since been changed by the callee.
11644 @value{GDBN} tries to deduce where the inner frame saved
11645 (``callee-saved'') registers, from the debug info, unwind info, or the
11646 machine code generated by your compiler. If some register is not
11647 saved, and @value{GDBN} knows the register is ``caller-saved'' (via
11648 its own knowledge of the ABI, or because the debug/unwind info
11649 explicitly says the register's value is undefined), @value{GDBN}
11650 displays @w{@samp{<not saved>}} as the register's value. With targets
11651 that @value{GDBN} has no knowledge of the register saving convention,
11652 if a register was not saved by the callee, then its value and location
11653 in the outer frame are assumed to be the same of the inner frame.
11654 This is usually harmless, because if the register is call-clobbered,
11655 the caller either does not care what is in the register after the
11656 call, or has code to restore the value that it does care about. Note,
11657 however, that if you change such a register in the outer frame, you
11658 may also be affecting the inner frame. Also, the more ``outer'' the
11659 frame is you're looking at, the more likely a call-clobbered
11660 register's value is to be wrong, in the sense that it doesn't actually
11661 represent the value the register had just before the call.
11663 @node Floating Point Hardware
11664 @section Floating Point Hardware
11665 @cindex floating point
11667 Depending on the configuration, @value{GDBN} may be able to give
11668 you more information about the status of the floating point hardware.
11673 Display hardware-dependent information about the floating
11674 point unit. The exact contents and layout vary depending on the
11675 floating point chip. Currently, @samp{info float} is supported on
11676 the ARM and x86 machines.
11680 @section Vector Unit
11681 @cindex vector unit
11683 Depending on the configuration, @value{GDBN} may be able to give you
11684 more information about the status of the vector unit.
11687 @kindex info vector
11689 Display information about the vector unit. The exact contents and
11690 layout vary depending on the hardware.
11693 @node OS Information
11694 @section Operating System Auxiliary Information
11695 @cindex OS information
11697 @value{GDBN} provides interfaces to useful OS facilities that can help
11698 you debug your program.
11700 @cindex auxiliary vector
11701 @cindex vector, auxiliary
11702 Some operating systems supply an @dfn{auxiliary vector} to programs at
11703 startup. This is akin to the arguments and environment that you
11704 specify for a program, but contains a system-dependent variety of
11705 binary values that tell system libraries important details about the
11706 hardware, operating system, and process. Each value's purpose is
11707 identified by an integer tag; the meanings are well-known but system-specific.
11708 Depending on the configuration and operating system facilities,
11709 @value{GDBN} may be able to show you this information. For remote
11710 targets, this functionality may further depend on the remote stub's
11711 support of the @samp{qXfer:auxv:read} packet, see
11712 @ref{qXfer auxiliary vector read}.
11717 Display the auxiliary vector of the inferior, which can be either a
11718 live process or a core dump file. @value{GDBN} prints each tag value
11719 numerically, and also shows names and text descriptions for recognized
11720 tags. Some values in the vector are numbers, some bit masks, and some
11721 pointers to strings or other data. @value{GDBN} displays each value in the
11722 most appropriate form for a recognized tag, and in hexadecimal for
11723 an unrecognized tag.
11726 On some targets, @value{GDBN} can access operating system-specific
11727 information and show it to you. The types of information available
11728 will differ depending on the type of operating system running on the
11729 target. The mechanism used to fetch the data is described in
11730 @ref{Operating System Information}. For remote targets, this
11731 functionality depends on the remote stub's support of the
11732 @samp{qXfer:osdata:read} packet, see @ref{qXfer osdata read}.
11736 @item info os @var{infotype}
11738 Display OS information of the requested type.
11740 On @sc{gnu}/Linux, the following values of @var{infotype} are valid:
11742 @anchor{linux info os infotypes}
11744 @kindex info os cpus
11746 Display the list of all CPUs/cores. For each CPU/core, @value{GDBN} prints
11747 the available fields from /proc/cpuinfo. For each supported architecture
11748 different fields are available. Two common entries are processor which gives
11749 CPU number and bogomips; a system constant that is calculated during
11750 kernel initialization.
11752 @kindex info os files
11754 Display the list of open file descriptors on the target. For each
11755 file descriptor, @value{GDBN} prints the identifier of the process
11756 owning the descriptor, the command of the owning process, the value
11757 of the descriptor, and the target of the descriptor.
11759 @kindex info os modules
11761 Display the list of all loaded kernel modules on the target. For each
11762 module, @value{GDBN} prints the module name, the size of the module in
11763 bytes, the number of times the module is used, the dependencies of the
11764 module, the status of the module, and the address of the loaded module
11767 @kindex info os msg
11769 Display the list of all System V message queues on the target. For each
11770 message queue, @value{GDBN} prints the message queue key, the message
11771 queue identifier, the access permissions, the current number of bytes
11772 on the queue, the current number of messages on the queue, the processes
11773 that last sent and received a message on the queue, the user and group
11774 of the owner and creator of the message queue, the times at which a
11775 message was last sent and received on the queue, and the time at which
11776 the message queue was last changed.
11778 @kindex info os processes
11780 Display the list of processes on the target. For each process,
11781 @value{GDBN} prints the process identifier, the name of the user, the
11782 command corresponding to the process, and the list of processor cores
11783 that the process is currently running on. (To understand what these
11784 properties mean, for this and the following info types, please consult
11785 the general @sc{gnu}/Linux documentation.)
11787 @kindex info os procgroups
11789 Display the list of process groups on the target. For each process,
11790 @value{GDBN} prints the identifier of the process group that it belongs
11791 to, the command corresponding to the process group leader, the process
11792 identifier, and the command line of the process. The list is sorted
11793 first by the process group identifier, then by the process identifier,
11794 so that processes belonging to the same process group are grouped together
11795 and the process group leader is listed first.
11797 @kindex info os semaphores
11799 Display the list of all System V semaphore sets on the target. For each
11800 semaphore set, @value{GDBN} prints the semaphore set key, the semaphore
11801 set identifier, the access permissions, the number of semaphores in the
11802 set, the user and group of the owner and creator of the semaphore set,
11803 and the times at which the semaphore set was operated upon and changed.
11805 @kindex info os shm
11807 Display the list of all System V shared-memory regions on the target.
11808 For each shared-memory region, @value{GDBN} prints the region key,
11809 the shared-memory identifier, the access permissions, the size of the
11810 region, the process that created the region, the process that last
11811 attached to or detached from the region, the current number of live
11812 attaches to the region, and the times at which the region was last
11813 attached to, detach from, and changed.
11815 @kindex info os sockets
11817 Display the list of Internet-domain sockets on the target. For each
11818 socket, @value{GDBN} prints the address and port of the local and
11819 remote endpoints, the current state of the connection, the creator of
11820 the socket, the IP address family of the socket, and the type of the
11823 @kindex info os threads
11825 Display the list of threads running on the target. For each thread,
11826 @value{GDBN} prints the identifier of the process that the thread
11827 belongs to, the command of the process, the thread identifier, and the
11828 processor core that it is currently running on. The main thread of a
11829 process is not listed.
11833 If @var{infotype} is omitted, then list the possible values for
11834 @var{infotype} and the kind of OS information available for each
11835 @var{infotype}. If the target does not return a list of possible
11836 types, this command will report an error.
11839 @node Memory Region Attributes
11840 @section Memory Region Attributes
11841 @cindex memory region attributes
11843 @dfn{Memory region attributes} allow you to describe special handling
11844 required by regions of your target's memory. @value{GDBN} uses
11845 attributes to determine whether to allow certain types of memory
11846 accesses; whether to use specific width accesses; and whether to cache
11847 target memory. By default the description of memory regions is
11848 fetched from the target (if the current target supports this), but the
11849 user can override the fetched regions.
11851 Defined memory regions can be individually enabled and disabled. When a
11852 memory region is disabled, @value{GDBN} uses the default attributes when
11853 accessing memory in that region. Similarly, if no memory regions have
11854 been defined, @value{GDBN} uses the default attributes when accessing
11857 When a memory region is defined, it is given a number to identify it;
11858 to enable, disable, or remove a memory region, you specify that number.
11862 @item mem @var{lower} @var{upper} @var{attributes}@dots{}
11863 Define a memory region bounded by @var{lower} and @var{upper} with
11864 attributes @var{attributes}@dots{}, and add it to the list of regions
11865 monitored by @value{GDBN}. Note that @var{upper} == 0 is a special
11866 case: it is treated as the target's maximum memory address.
11867 (0xffff on 16 bit targets, 0xffffffff on 32 bit targets, etc.)
11870 Discard any user changes to the memory regions and use target-supplied
11871 regions, if available, or no regions if the target does not support.
11874 @item delete mem @var{nums}@dots{}
11875 Remove memory regions @var{nums}@dots{} from the list of regions
11876 monitored by @value{GDBN}.
11878 @kindex disable mem
11879 @item disable mem @var{nums}@dots{}
11880 Disable monitoring of memory regions @var{nums}@dots{}.
11881 A disabled memory region is not forgotten.
11882 It may be enabled again later.
11885 @item enable mem @var{nums}@dots{}
11886 Enable monitoring of memory regions @var{nums}@dots{}.
11890 Print a table of all defined memory regions, with the following columns
11894 @item Memory Region Number
11895 @item Enabled or Disabled.
11896 Enabled memory regions are marked with @samp{y}.
11897 Disabled memory regions are marked with @samp{n}.
11900 The address defining the inclusive lower bound of the memory region.
11903 The address defining the exclusive upper bound of the memory region.
11906 The list of attributes set for this memory region.
11911 @subsection Attributes
11913 @subsubsection Memory Access Mode
11914 The access mode attributes set whether @value{GDBN} may make read or
11915 write accesses to a memory region.
11917 While these attributes prevent @value{GDBN} from performing invalid
11918 memory accesses, they do nothing to prevent the target system, I/O DMA,
11919 etc.@: from accessing memory.
11923 Memory is read only.
11925 Memory is write only.
11927 Memory is read/write. This is the default.
11930 @subsubsection Memory Access Size
11931 The access size attribute tells @value{GDBN} to use specific sized
11932 accesses in the memory region. Often memory mapped device registers
11933 require specific sized accesses. If no access size attribute is
11934 specified, @value{GDBN} may use accesses of any size.
11938 Use 8 bit memory accesses.
11940 Use 16 bit memory accesses.
11942 Use 32 bit memory accesses.
11944 Use 64 bit memory accesses.
11947 @c @subsubsection Hardware/Software Breakpoints
11948 @c The hardware/software breakpoint attributes set whether @value{GDBN}
11949 @c will use hardware or software breakpoints for the internal breakpoints
11950 @c used by the step, next, finish, until, etc. commands.
11954 @c Always use hardware breakpoints
11955 @c @item swbreak (default)
11958 @subsubsection Data Cache
11959 The data cache attributes set whether @value{GDBN} will cache target
11960 memory. While this generally improves performance by reducing debug
11961 protocol overhead, it can lead to incorrect results because @value{GDBN}
11962 does not know about volatile variables or memory mapped device
11967 Enable @value{GDBN} to cache target memory.
11969 Disable @value{GDBN} from caching target memory. This is the default.
11972 @subsection Memory Access Checking
11973 @value{GDBN} can be instructed to refuse accesses to memory that is
11974 not explicitly described. This can be useful if accessing such
11975 regions has undesired effects for a specific target, or to provide
11976 better error checking. The following commands control this behaviour.
11979 @kindex set mem inaccessible-by-default
11980 @item set mem inaccessible-by-default [on|off]
11981 If @code{on} is specified, make @value{GDBN} treat memory not
11982 explicitly described by the memory ranges as non-existent and refuse accesses
11983 to such memory. The checks are only performed if there's at least one
11984 memory range defined. If @code{off} is specified, make @value{GDBN}
11985 treat the memory not explicitly described by the memory ranges as RAM.
11986 The default value is @code{on}.
11987 @kindex show mem inaccessible-by-default
11988 @item show mem inaccessible-by-default
11989 Show the current handling of accesses to unknown memory.
11993 @c @subsubsection Memory Write Verification
11994 @c The memory write verification attributes set whether @value{GDBN}
11995 @c will re-reads data after each write to verify the write was successful.
11999 @c @item noverify (default)
12002 @node Dump/Restore Files
12003 @section Copy Between Memory and a File
12004 @cindex dump/restore files
12005 @cindex append data to a file
12006 @cindex dump data to a file
12007 @cindex restore data from a file
12009 You can use the commands @code{dump}, @code{append}, and
12010 @code{restore} to copy data between target memory and a file. The
12011 @code{dump} and @code{append} commands write data to a file, and the
12012 @code{restore} command reads data from a file back into the inferior's
12013 memory. Files may be in binary, Motorola S-record, Intel hex,
12014 Tektronix Hex, or Verilog Hex format; however, @value{GDBN} can only
12015 append to binary files, and cannot read from Verilog Hex files.
12020 @item dump @r{[}@var{format}@r{]} memory @var{filename} @var{start_addr} @var{end_addr}
12021 @itemx dump @r{[}@var{format}@r{]} value @var{filename} @var{expr}
12022 Dump the contents of memory from @var{start_addr} to @var{end_addr},
12023 or the value of @var{expr}, to @var{filename} in the given format.
12025 The @var{format} parameter may be any one of:
12032 Motorola S-record format.
12034 Tektronix Hex format.
12036 Verilog Hex format.
12039 @value{GDBN} uses the same definitions of these formats as the
12040 @sc{gnu} binary utilities, like @samp{objdump} and @samp{objcopy}. If
12041 @var{format} is omitted, @value{GDBN} dumps the data in raw binary
12045 @item append @r{[}binary@r{]} memory @var{filename} @var{start_addr} @var{end_addr}
12046 @itemx append @r{[}binary@r{]} value @var{filename} @var{expr}
12047 Append the contents of memory from @var{start_addr} to @var{end_addr},
12048 or the value of @var{expr}, to the file @var{filename}, in raw binary form.
12049 (@value{GDBN} can only append data to files in raw binary form.)
12052 @item restore @var{filename} @r{[}binary@r{]} @var{bias} @var{start} @var{end}
12053 Restore the contents of file @var{filename} into memory. The
12054 @code{restore} command can automatically recognize any known @sc{bfd}
12055 file format, except for raw binary. To restore a raw binary file you
12056 must specify the optional keyword @code{binary} after the filename.
12058 If @var{bias} is non-zero, its value will be added to the addresses
12059 contained in the file. Binary files always start at address zero, so
12060 they will be restored at address @var{bias}. Other bfd files have
12061 a built-in location; they will be restored at offset @var{bias}
12062 from that location.
12064 If @var{start} and/or @var{end} are non-zero, then only data between
12065 file offset @var{start} and file offset @var{end} will be restored.
12066 These offsets are relative to the addresses in the file, before
12067 the @var{bias} argument is applied.
12071 @node Core File Generation
12072 @section How to Produce a Core File from Your Program
12073 @cindex dump core from inferior
12075 A @dfn{core file} or @dfn{core dump} is a file that records the memory
12076 image of a running process and its process status (register values
12077 etc.). Its primary use is post-mortem debugging of a program that
12078 crashed while it ran outside a debugger. A program that crashes
12079 automatically produces a core file, unless this feature is disabled by
12080 the user. @xref{Files}, for information on invoking @value{GDBN} in
12081 the post-mortem debugging mode.
12083 Occasionally, you may wish to produce a core file of the program you
12084 are debugging in order to preserve a snapshot of its state.
12085 @value{GDBN} has a special command for that.
12089 @kindex generate-core-file
12090 @item generate-core-file [@var{file}]
12091 @itemx gcore [@var{file}]
12092 Produce a core dump of the inferior process. The optional argument
12093 @var{file} specifies the file name where to put the core dump. If not
12094 specified, the file name defaults to @file{core.@var{pid}}, where
12095 @var{pid} is the inferior process ID.
12097 Note that this command is implemented only for some systems (as of
12098 this writing, @sc{gnu}/Linux, FreeBSD, Solaris, and S390).
12100 On @sc{gnu}/Linux, this command can take into account the value of the
12101 file @file{/proc/@var{pid}/coredump_filter} when generating the core
12102 dump (@pxref{set use-coredump-filter}), and by default honors the
12103 @code{VM_DONTDUMP} flag for mappings where it is present in the file
12104 @file{/proc/@var{pid}/smaps} (@pxref{set dump-excluded-mappings}).
12106 @kindex set use-coredump-filter
12107 @anchor{set use-coredump-filter}
12108 @item set use-coredump-filter on
12109 @itemx set use-coredump-filter off
12110 Enable or disable the use of the file
12111 @file{/proc/@var{pid}/coredump_filter} when generating core dump
12112 files. This file is used by the Linux kernel to decide what types of
12113 memory mappings will be dumped or ignored when generating a core dump
12114 file. @var{pid} is the process ID of a currently running process.
12116 To make use of this feature, you have to write in the
12117 @file{/proc/@var{pid}/coredump_filter} file a value, in hexadecimal,
12118 which is a bit mask representing the memory mapping types. If a bit
12119 is set in the bit mask, then the memory mappings of the corresponding
12120 types will be dumped; otherwise, they will be ignored. This
12121 configuration is inherited by child processes. For more information
12122 about the bits that can be set in the
12123 @file{/proc/@var{pid}/coredump_filter} file, please refer to the
12124 manpage of @code{core(5)}.
12126 By default, this option is @code{on}. If this option is turned
12127 @code{off}, @value{GDBN} does not read the @file{coredump_filter} file
12128 and instead uses the same default value as the Linux kernel in order
12129 to decide which pages will be dumped in the core dump file. This
12130 value is currently @code{0x33}, which means that bits @code{0}
12131 (anonymous private mappings), @code{1} (anonymous shared mappings),
12132 @code{4} (ELF headers) and @code{5} (private huge pages) are active.
12133 This will cause these memory mappings to be dumped automatically.
12135 @kindex set dump-excluded-mappings
12136 @anchor{set dump-excluded-mappings}
12137 @item set dump-excluded-mappings on
12138 @itemx set dump-excluded-mappings off
12139 If @code{on} is specified, @value{GDBN} will dump memory mappings
12140 marked with the @code{VM_DONTDUMP} flag. This flag is represented in
12141 the file @file{/proc/@var{pid}/smaps} with the acronym @code{dd}.
12143 The default value is @code{off}.
12146 @node Character Sets
12147 @section Character Sets
12148 @cindex character sets
12150 @cindex translating between character sets
12151 @cindex host character set
12152 @cindex target character set
12154 If the program you are debugging uses a different character set to
12155 represent characters and strings than the one @value{GDBN} uses itself,
12156 @value{GDBN} can automatically translate between the character sets for
12157 you. The character set @value{GDBN} uses we call the @dfn{host
12158 character set}; the one the inferior program uses we call the
12159 @dfn{target character set}.
12161 For example, if you are running @value{GDBN} on a @sc{gnu}/Linux system, which
12162 uses the ISO Latin 1 character set, but you are using @value{GDBN}'s
12163 remote protocol (@pxref{Remote Debugging}) to debug a program
12164 running on an IBM mainframe, which uses the @sc{ebcdic} character set,
12165 then the host character set is Latin-1, and the target character set is
12166 @sc{ebcdic}. If you give @value{GDBN} the command @code{set
12167 target-charset EBCDIC-US}, then @value{GDBN} translates between
12168 @sc{ebcdic} and Latin 1 as you print character or string values, or use
12169 character and string literals in expressions.
12171 @value{GDBN} has no way to automatically recognize which character set
12172 the inferior program uses; you must tell it, using the @code{set
12173 target-charset} command, described below.
12175 Here are the commands for controlling @value{GDBN}'s character set
12179 @item set target-charset @var{charset}
12180 @kindex set target-charset
12181 Set the current target character set to @var{charset}. To display the
12182 list of supported target character sets, type
12183 @kbd{@w{set target-charset @key{TAB}@key{TAB}}}.
12185 @item set host-charset @var{charset}
12186 @kindex set host-charset
12187 Set the current host character set to @var{charset}.
12189 By default, @value{GDBN} uses a host character set appropriate to the
12190 system it is running on; you can override that default using the
12191 @code{set host-charset} command. On some systems, @value{GDBN} cannot
12192 automatically determine the appropriate host character set. In this
12193 case, @value{GDBN} uses @samp{UTF-8}.
12195 @value{GDBN} can only use certain character sets as its host character
12196 set. If you type @kbd{@w{set host-charset @key{TAB}@key{TAB}}},
12197 @value{GDBN} will list the host character sets it supports.
12199 @item set charset @var{charset}
12200 @kindex set charset
12201 Set the current host and target character sets to @var{charset}. As
12202 above, if you type @kbd{@w{set charset @key{TAB}@key{TAB}}},
12203 @value{GDBN} will list the names of the character sets that can be used
12204 for both host and target.
12207 @kindex show charset
12208 Show the names of the current host and target character sets.
12210 @item show host-charset
12211 @kindex show host-charset
12212 Show the name of the current host character set.
12214 @item show target-charset
12215 @kindex show target-charset
12216 Show the name of the current target character set.
12218 @item set target-wide-charset @var{charset}
12219 @kindex set target-wide-charset
12220 Set the current target's wide character set to @var{charset}. This is
12221 the character set used by the target's @code{wchar_t} type. To
12222 display the list of supported wide character sets, type
12223 @kbd{@w{set target-wide-charset @key{TAB}@key{TAB}}}.
12225 @item show target-wide-charset
12226 @kindex show target-wide-charset
12227 Show the name of the current target's wide character set.
12230 Here is an example of @value{GDBN}'s character set support in action.
12231 Assume that the following source code has been placed in the file
12232 @file{charset-test.c}:
12238 = @{72, 101, 108, 108, 111, 44, 32, 119,
12239 111, 114, 108, 100, 33, 10, 0@};
12240 char ibm1047_hello[]
12241 = @{200, 133, 147, 147, 150, 107, 64, 166,
12242 150, 153, 147, 132, 90, 37, 0@};
12246 printf ("Hello, world!\n");
12250 In this program, @code{ascii_hello} and @code{ibm1047_hello} are arrays
12251 containing the string @samp{Hello, world!} followed by a newline,
12252 encoded in the @sc{ascii} and @sc{ibm1047} character sets.
12254 We compile the program, and invoke the debugger on it:
12257 $ gcc -g charset-test.c -o charset-test
12258 $ gdb -nw charset-test
12259 GNU gdb 2001-12-19-cvs
12260 Copyright 2001 Free Software Foundation, Inc.
12265 We can use the @code{show charset} command to see what character sets
12266 @value{GDBN} is currently using to interpret and display characters and
12270 (@value{GDBP}) show charset
12271 The current host and target character set is `ISO-8859-1'.
12275 For the sake of printing this manual, let's use @sc{ascii} as our
12276 initial character set:
12278 (@value{GDBP}) set charset ASCII
12279 (@value{GDBP}) show charset
12280 The current host and target character set is `ASCII'.
12284 Let's assume that @sc{ascii} is indeed the correct character set for our
12285 host system --- in other words, let's assume that if @value{GDBN} prints
12286 characters using the @sc{ascii} character set, our terminal will display
12287 them properly. Since our current target character set is also
12288 @sc{ascii}, the contents of @code{ascii_hello} print legibly:
12291 (@value{GDBP}) print ascii_hello
12292 $1 = 0x401698 "Hello, world!\n"
12293 (@value{GDBP}) print ascii_hello[0]
12298 @value{GDBN} uses the target character set for character and string
12299 literals you use in expressions:
12302 (@value{GDBP}) print '+'
12307 The @sc{ascii} character set uses the number 43 to encode the @samp{+}
12310 @value{GDBN} relies on the user to tell it which character set the
12311 target program uses. If we print @code{ibm1047_hello} while our target
12312 character set is still @sc{ascii}, we get jibberish:
12315 (@value{GDBP}) print ibm1047_hello
12316 $4 = 0x4016a8 "\310\205\223\223\226k@@\246\226\231\223\204Z%"
12317 (@value{GDBP}) print ibm1047_hello[0]
12322 If we invoke the @code{set target-charset} followed by @key{TAB}@key{TAB},
12323 @value{GDBN} tells us the character sets it supports:
12326 (@value{GDBP}) set target-charset
12327 ASCII EBCDIC-US IBM1047 ISO-8859-1
12328 (@value{GDBP}) set target-charset
12331 We can select @sc{ibm1047} as our target character set, and examine the
12332 program's strings again. Now the @sc{ascii} string is wrong, but
12333 @value{GDBN} translates the contents of @code{ibm1047_hello} from the
12334 target character set, @sc{ibm1047}, to the host character set,
12335 @sc{ascii}, and they display correctly:
12338 (@value{GDBP}) set target-charset IBM1047
12339 (@value{GDBP}) show charset
12340 The current host character set is `ASCII'.
12341 The current target character set is `IBM1047'.
12342 (@value{GDBP}) print ascii_hello
12343 $6 = 0x401698 "\110\145%%?\054\040\167?\162%\144\041\012"
12344 (@value{GDBP}) print ascii_hello[0]
12346 (@value{GDBP}) print ibm1047_hello
12347 $8 = 0x4016a8 "Hello, world!\n"
12348 (@value{GDBP}) print ibm1047_hello[0]
12353 As above, @value{GDBN} uses the target character set for character and
12354 string literals you use in expressions:
12357 (@value{GDBP}) print '+'
12362 The @sc{ibm1047} character set uses the number 78 to encode the @samp{+}
12365 @node Caching Target Data
12366 @section Caching Data of Targets
12367 @cindex caching data of targets
12369 @value{GDBN} caches data exchanged between the debugger and a target.
12370 Each cache is associated with the address space of the inferior.
12371 @xref{Inferiors and Programs}, about inferior and address space.
12372 Such caching generally improves performance in remote debugging
12373 (@pxref{Remote Debugging}), because it reduces the overhead of the
12374 remote protocol by bundling memory reads and writes into large chunks.
12375 Unfortunately, simply caching everything would lead to incorrect results,
12376 since @value{GDBN} does not necessarily know anything about volatile
12377 values, memory-mapped I/O addresses, etc. Furthermore, in non-stop mode
12378 (@pxref{Non-Stop Mode}) memory can be changed @emph{while} a gdb command
12380 Therefore, by default, @value{GDBN} only caches data
12381 known to be on the stack@footnote{In non-stop mode, it is moderately
12382 rare for a running thread to modify the stack of a stopped thread
12383 in a way that would interfere with a backtrace, and caching of
12384 stack reads provides a significant speed up of remote backtraces.} or
12385 in the code segment.
12386 Other regions of memory can be explicitly marked as
12387 cacheable; @pxref{Memory Region Attributes}.
12390 @kindex set remotecache
12391 @item set remotecache on
12392 @itemx set remotecache off
12393 This option no longer does anything; it exists for compatibility
12396 @kindex show remotecache
12397 @item show remotecache
12398 Show the current state of the obsolete remotecache flag.
12400 @kindex set stack-cache
12401 @item set stack-cache on
12402 @itemx set stack-cache off
12403 Enable or disable caching of stack accesses. When @code{on}, use
12404 caching. By default, this option is @code{on}.
12406 @kindex show stack-cache
12407 @item show stack-cache
12408 Show the current state of data caching for memory accesses.
12410 @kindex set code-cache
12411 @item set code-cache on
12412 @itemx set code-cache off
12413 Enable or disable caching of code segment accesses. When @code{on},
12414 use caching. By default, this option is @code{on}. This improves
12415 performance of disassembly in remote debugging.
12417 @kindex show code-cache
12418 @item show code-cache
12419 Show the current state of target memory cache for code segment
12422 @kindex info dcache
12423 @item info dcache @r{[}line@r{]}
12424 Print the information about the performance of data cache of the
12425 current inferior's address space. The information displayed
12426 includes the dcache width and depth, and for each cache line, its
12427 number, address, and how many times it was referenced. This
12428 command is useful for debugging the data cache operation.
12430 If a line number is specified, the contents of that line will be
12433 @item set dcache size @var{size}
12434 @cindex dcache size
12435 @kindex set dcache size
12436 Set maximum number of entries in dcache (dcache depth above).
12438 @item set dcache line-size @var{line-size}
12439 @cindex dcache line-size
12440 @kindex set dcache line-size
12441 Set number of bytes each dcache entry caches (dcache width above).
12442 Must be a power of 2.
12444 @item show dcache size
12445 @kindex show dcache size
12446 Show maximum number of dcache entries. @xref{Caching Target Data, info dcache}.
12448 @item show dcache line-size
12449 @kindex show dcache line-size
12450 Show default size of dcache lines.
12454 @node Searching Memory
12455 @section Search Memory
12456 @cindex searching memory
12458 Memory can be searched for a particular sequence of bytes with the
12459 @code{find} command.
12463 @item find @r{[}/@var{sn}@r{]} @var{start_addr}, +@var{len}, @var{val1} @r{[}, @var{val2}, @dots{}@r{]}
12464 @itemx find @r{[}/@var{sn}@r{]} @var{start_addr}, @var{end_addr}, @var{val1} @r{[}, @var{val2}, @dots{}@r{]}
12465 Search memory for the sequence of bytes specified by @var{val1}, @var{val2},
12466 etc. The search begins at address @var{start_addr} and continues for either
12467 @var{len} bytes or through to @var{end_addr} inclusive.
12470 @var{s} and @var{n} are optional parameters.
12471 They may be specified in either order, apart or together.
12474 @item @var{s}, search query size
12475 The size of each search query value.
12481 halfwords (two bytes)
12485 giant words (eight bytes)
12488 All values are interpreted in the current language.
12489 This means, for example, that if the current source language is C/C@t{++}
12490 then searching for the string ``hello'' includes the trailing '\0'.
12491 The null terminator can be removed from searching by using casts,
12492 e.g.: @samp{@{char[5]@}"hello"}.
12494 If the value size is not specified, it is taken from the
12495 value's type in the current language.
12496 This is useful when one wants to specify the search
12497 pattern as a mixture of types.
12498 Note that this means, for example, that in the case of C-like languages
12499 a search for an untyped 0x42 will search for @samp{(int) 0x42}
12500 which is typically four bytes.
12502 @item @var{n}, maximum number of finds
12503 The maximum number of matches to print. The default is to print all finds.
12506 You can use strings as search values. Quote them with double-quotes
12508 The string value is copied into the search pattern byte by byte,
12509 regardless of the endianness of the target and the size specification.
12511 The address of each match found is printed as well as a count of the
12512 number of matches found.
12514 The address of the last value found is stored in convenience variable
12516 A count of the number of matches is stored in @samp{$numfound}.
12518 For example, if stopped at the @code{printf} in this function:
12524 static char hello[] = "hello-hello";
12525 static struct @{ char c; short s; int i; @}
12526 __attribute__ ((packed)) mixed
12527 = @{ 'c', 0x1234, 0x87654321 @};
12528 printf ("%s\n", hello);
12533 you get during debugging:
12536 (gdb) find &hello[0], +sizeof(hello), "hello"
12537 0x804956d <hello.1620+6>
12539 (gdb) find &hello[0], +sizeof(hello), 'h', 'e', 'l', 'l', 'o'
12540 0x8049567 <hello.1620>
12541 0x804956d <hello.1620+6>
12543 (gdb) find &hello[0], +sizeof(hello), @{char[5]@}"hello"
12544 0x8049567 <hello.1620>
12545 0x804956d <hello.1620+6>
12547 (gdb) find /b1 &hello[0], +sizeof(hello), 'h', 0x65, 'l'
12548 0x8049567 <hello.1620>
12550 (gdb) find &mixed, +sizeof(mixed), (char) 'c', (short) 0x1234, (int) 0x87654321
12551 0x8049560 <mixed.1625>
12553 (gdb) print $numfound
12556 $2 = (void *) 0x8049560
12560 @section Value Sizes
12562 Whenever @value{GDBN} prints a value memory will be allocated within
12563 @value{GDBN} to hold the contents of the value. It is possible in
12564 some languages with dynamic typing systems, that an invalid program
12565 may indicate a value that is incorrectly large, this in turn may cause
12566 @value{GDBN} to try and allocate an overly large ammount of memory.
12569 @kindex set max-value-size
12570 @item set max-value-size @var{bytes}
12571 @itemx set max-value-size unlimited
12572 Set the maximum size of memory that @value{GDBN} will allocate for the
12573 contents of a value to @var{bytes}, trying to display a value that
12574 requires more memory than that will result in an error.
12576 Setting this variable does not effect values that have already been
12577 allocated within @value{GDBN}, only future allocations.
12579 There's a minimum size that @code{max-value-size} can be set to in
12580 order that @value{GDBN} can still operate correctly, this minimum is
12581 currently 16 bytes.
12583 The limit applies to the results of some subexpressions as well as to
12584 complete expressions. For example, an expression denoting a simple
12585 integer component, such as @code{x.y.z}, may fail if the size of
12586 @var{x.y} is dynamic and exceeds @var{bytes}. On the other hand,
12587 @value{GDBN} is sometimes clever; the expression @code{A[i]}, where
12588 @var{A} is an array variable with non-constant size, will generally
12589 succeed regardless of the bounds on @var{A}, as long as the component
12590 size is less than @var{bytes}.
12592 The default value of @code{max-value-size} is currently 64k.
12594 @kindex show max-value-size
12595 @item show max-value-size
12596 Show the maximum size of memory, in bytes, that @value{GDBN} will
12597 allocate for the contents of a value.
12600 @node Optimized Code
12601 @chapter Debugging Optimized Code
12602 @cindex optimized code, debugging
12603 @cindex debugging optimized code
12605 Almost all compilers support optimization. With optimization
12606 disabled, the compiler generates assembly code that corresponds
12607 directly to your source code, in a simplistic way. As the compiler
12608 applies more powerful optimizations, the generated assembly code
12609 diverges from your original source code. With help from debugging
12610 information generated by the compiler, @value{GDBN} can map from
12611 the running program back to constructs from your original source.
12613 @value{GDBN} is more accurate with optimization disabled. If you
12614 can recompile without optimization, it is easier to follow the
12615 progress of your program during debugging. But, there are many cases
12616 where you may need to debug an optimized version.
12618 When you debug a program compiled with @samp{-g -O}, remember that the
12619 optimizer has rearranged your code; the debugger shows you what is
12620 really there. Do not be too surprised when the execution path does not
12621 exactly match your source file! An extreme example: if you define a
12622 variable, but never use it, @value{GDBN} never sees that
12623 variable---because the compiler optimizes it out of existence.
12625 Some things do not work as well with @samp{-g -O} as with just
12626 @samp{-g}, particularly on machines with instruction scheduling. If in
12627 doubt, recompile with @samp{-g} alone, and if this fixes the problem,
12628 please report it to us as a bug (including a test case!).
12629 @xref{Variables}, for more information about debugging optimized code.
12632 * Inline Functions:: How @value{GDBN} presents inlining
12633 * Tail Call Frames:: @value{GDBN} analysis of jumps to functions
12636 @node Inline Functions
12637 @section Inline Functions
12638 @cindex inline functions, debugging
12640 @dfn{Inlining} is an optimization that inserts a copy of the function
12641 body directly at each call site, instead of jumping to a shared
12642 routine. @value{GDBN} displays inlined functions just like
12643 non-inlined functions. They appear in backtraces. You can view their
12644 arguments and local variables, step into them with @code{step}, skip
12645 them with @code{next}, and escape from them with @code{finish}.
12646 You can check whether a function was inlined by using the
12647 @code{info frame} command.
12649 For @value{GDBN} to support inlined functions, the compiler must
12650 record information about inlining in the debug information ---
12651 @value{NGCC} using the @sc{dwarf 2} format does this, and several
12652 other compilers do also. @value{GDBN} only supports inlined functions
12653 when using @sc{dwarf 2}. Versions of @value{NGCC} before 4.1
12654 do not emit two required attributes (@samp{DW_AT_call_file} and
12655 @samp{DW_AT_call_line}); @value{GDBN} does not display inlined
12656 function calls with earlier versions of @value{NGCC}. It instead
12657 displays the arguments and local variables of inlined functions as
12658 local variables in the caller.
12660 The body of an inlined function is directly included at its call site;
12661 unlike a non-inlined function, there are no instructions devoted to
12662 the call. @value{GDBN} still pretends that the call site and the
12663 start of the inlined function are different instructions. Stepping to
12664 the call site shows the call site, and then stepping again shows
12665 the first line of the inlined function, even though no additional
12666 instructions are executed.
12668 This makes source-level debugging much clearer; you can see both the
12669 context of the call and then the effect of the call. Only stepping by
12670 a single instruction using @code{stepi} or @code{nexti} does not do
12671 this; single instruction steps always show the inlined body.
12673 There are some ways that @value{GDBN} does not pretend that inlined
12674 function calls are the same as normal calls:
12678 Setting breakpoints at the call site of an inlined function may not
12679 work, because the call site does not contain any code. @value{GDBN}
12680 may incorrectly move the breakpoint to the next line of the enclosing
12681 function, after the call. This limitation will be removed in a future
12682 version of @value{GDBN}; until then, set a breakpoint on an earlier line
12683 or inside the inlined function instead.
12686 @value{GDBN} cannot locate the return value of inlined calls after
12687 using the @code{finish} command. This is a limitation of compiler-generated
12688 debugging information; after @code{finish}, you can step to the next line
12689 and print a variable where your program stored the return value.
12693 @node Tail Call Frames
12694 @section Tail Call Frames
12695 @cindex tail call frames, debugging
12697 Function @code{B} can call function @code{C} in its very last statement. In
12698 unoptimized compilation the call of @code{C} is immediately followed by return
12699 instruction at the end of @code{B} code. Optimizing compiler may replace the
12700 call and return in function @code{B} into one jump to function @code{C}
12701 instead. Such use of a jump instruction is called @dfn{tail call}.
12703 During execution of function @code{C}, there will be no indication in the
12704 function call stack frames that it was tail-called from @code{B}. If function
12705 @code{A} regularly calls function @code{B} which tail-calls function @code{C},
12706 then @value{GDBN} will see @code{A} as the caller of @code{C}. However, in
12707 some cases @value{GDBN} can determine that @code{C} was tail-called from
12708 @code{B}, and it will then create fictitious call frame for that, with the
12709 return address set up as if @code{B} called @code{C} normally.
12711 This functionality is currently supported only by DWARF 2 debugging format and
12712 the compiler has to produce @samp{DW_TAG_call_site} tags. With
12713 @value{NGCC}, you need to specify @option{-O -g} during compilation, to get
12716 @kbd{info frame} command (@pxref{Frame Info}) will indicate the tail call frame
12717 kind by text @code{tail call frame} such as in this sample @value{GDBN} output:
12721 0x40066b <b(int, double)+11>: jmp 0x400640 <c(int, double)>
12723 Stack level 1, frame at 0x7fffffffda30:
12724 rip = 0x40066d in b (amd64-entry-value.cc:59); saved rip 0x4004c5
12725 tail call frame, caller of frame at 0x7fffffffda30
12726 source language c++.
12727 Arglist at unknown address.
12728 Locals at unknown address, Previous frame's sp is 0x7fffffffda30
12731 The detection of all the possible code path executions can find them ambiguous.
12732 There is no execution history stored (possible @ref{Reverse Execution} is never
12733 used for this purpose) and the last known caller could have reached the known
12734 callee by multiple different jump sequences. In such case @value{GDBN} still
12735 tries to show at least all the unambiguous top tail callers and all the
12736 unambiguous bottom tail calees, if any.
12739 @anchor{set debug entry-values}
12740 @item set debug entry-values
12741 @kindex set debug entry-values
12742 When set to on, enables printing of analysis messages for both frame argument
12743 values at function entry and tail calls. It will show all the possible valid
12744 tail calls code paths it has considered. It will also print the intersection
12745 of them with the final unambiguous (possibly partial or even empty) code path
12748 @item show debug entry-values
12749 @kindex show debug entry-values
12750 Show the current state of analysis messages printing for both frame argument
12751 values at function entry and tail calls.
12754 The analysis messages for tail calls can for example show why the virtual tail
12755 call frame for function @code{c} has not been recognized (due to the indirect
12756 reference by variable @code{x}):
12759 static void __attribute__((noinline, noclone)) c (void);
12760 void (*x) (void) = c;
12761 static void __attribute__((noinline, noclone)) a (void) @{ x++; @}
12762 static void __attribute__((noinline, noclone)) c (void) @{ a (); @}
12763 int main (void) @{ x (); return 0; @}
12765 Breakpoint 1, DW_OP_entry_value resolving cannot find
12766 DW_TAG_call_site 0x40039a in main
12768 3 static void __attribute__((noinline, noclone)) a (void) @{ x++; @}
12771 #1 0x000000000040039a in main () at t.c:5
12774 Another possibility is an ambiguous virtual tail call frames resolution:
12778 static void __attribute__((noinline, noclone)) f (void) @{ i++; @}
12779 static void __attribute__((noinline, noclone)) e (void) @{ f (); @}
12780 static void __attribute__((noinline, noclone)) d (void) @{ f (); @}
12781 static void __attribute__((noinline, noclone)) c (void) @{ d (); @}
12782 static void __attribute__((noinline, noclone)) b (void)
12783 @{ if (i) c (); else e (); @}
12784 static void __attribute__((noinline, noclone)) a (void) @{ b (); @}
12785 int main (void) @{ a (); return 0; @}
12787 tailcall: initial: 0x4004d2(a) 0x4004ce(b) 0x4004b2(c) 0x4004a2(d)
12788 tailcall: compare: 0x4004d2(a) 0x4004cc(b) 0x400492(e)
12789 tailcall: reduced: 0x4004d2(a) |
12792 #1 0x00000000004004d2 in a () at t.c:8
12793 #2 0x0000000000400395 in main () at t.c:9
12796 @set CALLSEQ1A @code{main@value{ARROW}a@value{ARROW}b@value{ARROW}c@value{ARROW}d@value{ARROW}f}
12797 @set CALLSEQ2A @code{main@value{ARROW}a@value{ARROW}b@value{ARROW}e@value{ARROW}f}
12799 @c Convert CALLSEQ#A to CALLSEQ#B depending on HAVE_MAKEINFO_CLICK.
12800 @ifset HAVE_MAKEINFO_CLICK
12801 @set ARROW @click{}
12802 @set CALLSEQ1B @clicksequence{@value{CALLSEQ1A}}
12803 @set CALLSEQ2B @clicksequence{@value{CALLSEQ2A}}
12805 @ifclear HAVE_MAKEINFO_CLICK
12807 @set CALLSEQ1B @value{CALLSEQ1A}
12808 @set CALLSEQ2B @value{CALLSEQ2A}
12811 Frames #0 and #2 are real, #1 is a virtual tail call frame.
12812 The code can have possible execution paths @value{CALLSEQ1B} or
12813 @value{CALLSEQ2B}, @value{GDBN} cannot find which one from the inferior state.
12815 @code{initial:} state shows some random possible calling sequence @value{GDBN}
12816 has found. It then finds another possible calling sequcen - that one is
12817 prefixed by @code{compare:}. The non-ambiguous intersection of these two is
12818 printed as the @code{reduced:} calling sequence. That one could have many
12819 futher @code{compare:} and @code{reduced:} statements as long as there remain
12820 any non-ambiguous sequence entries.
12822 For the frame of function @code{b} in both cases there are different possible
12823 @code{$pc} values (@code{0x4004cc} or @code{0x4004ce}), therefore this frame is
12824 also ambigous. The only non-ambiguous frame is the one for function @code{a},
12825 therefore this one is displayed to the user while the ambiguous frames are
12828 There can be also reasons why printing of frame argument values at function
12833 static void __attribute__((noinline, noclone)) c (int i) @{ v++; @}
12834 static void __attribute__((noinline, noclone)) a (int i);
12835 static void __attribute__((noinline, noclone)) b (int i) @{ a (i); @}
12836 static void __attribute__((noinline, noclone)) a (int i)
12837 @{ if (i) b (i - 1); else c (0); @}
12838 int main (void) @{ a (5); return 0; @}
12841 #0 c (i=i@@entry=0) at t.c:2
12842 #1 0x0000000000400428 in a (DW_OP_entry_value resolving has found
12843 function "a" at 0x400420 can call itself via tail calls
12844 i=<optimized out>) at t.c:6
12845 #2 0x000000000040036e in main () at t.c:7
12848 @value{GDBN} cannot find out from the inferior state if and how many times did
12849 function @code{a} call itself (via function @code{b}) as these calls would be
12850 tail calls. Such tail calls would modify thue @code{i} variable, therefore
12851 @value{GDBN} cannot be sure the value it knows would be right - @value{GDBN}
12852 prints @code{<optimized out>} instead.
12855 @chapter C Preprocessor Macros
12857 Some languages, such as C and C@t{++}, provide a way to define and invoke
12858 ``preprocessor macros'' which expand into strings of tokens.
12859 @value{GDBN} can evaluate expressions containing macro invocations, show
12860 the result of macro expansion, and show a macro's definition, including
12861 where it was defined.
12863 You may need to compile your program specially to provide @value{GDBN}
12864 with information about preprocessor macros. Most compilers do not
12865 include macros in their debugging information, even when you compile
12866 with the @option{-g} flag. @xref{Compilation}.
12868 A program may define a macro at one point, remove that definition later,
12869 and then provide a different definition after that. Thus, at different
12870 points in the program, a macro may have different definitions, or have
12871 no definition at all. If there is a current stack frame, @value{GDBN}
12872 uses the macros in scope at that frame's source code line. Otherwise,
12873 @value{GDBN} uses the macros in scope at the current listing location;
12876 Whenever @value{GDBN} evaluates an expression, it always expands any
12877 macro invocations present in the expression. @value{GDBN} also provides
12878 the following commands for working with macros explicitly.
12882 @kindex macro expand
12883 @cindex macro expansion, showing the results of preprocessor
12884 @cindex preprocessor macro expansion, showing the results of
12885 @cindex expanding preprocessor macros
12886 @item macro expand @var{expression}
12887 @itemx macro exp @var{expression}
12888 Show the results of expanding all preprocessor macro invocations in
12889 @var{expression}. Since @value{GDBN} simply expands macros, but does
12890 not parse the result, @var{expression} need not be a valid expression;
12891 it can be any string of tokens.
12894 @item macro expand-once @var{expression}
12895 @itemx macro exp1 @var{expression}
12896 @cindex expand macro once
12897 @i{(This command is not yet implemented.)} Show the results of
12898 expanding those preprocessor macro invocations that appear explicitly in
12899 @var{expression}. Macro invocations appearing in that expansion are
12900 left unchanged. This command allows you to see the effect of a
12901 particular macro more clearly, without being confused by further
12902 expansions. Since @value{GDBN} simply expands macros, but does not
12903 parse the result, @var{expression} need not be a valid expression; it
12904 can be any string of tokens.
12907 @cindex macro definition, showing
12908 @cindex definition of a macro, showing
12909 @cindex macros, from debug info
12910 @item info macro [-a|-all] [--] @var{macro}
12911 Show the current definition or all definitions of the named @var{macro},
12912 and describe the source location or compiler command-line where that
12913 definition was established. The optional double dash is to signify the end of
12914 argument processing and the beginning of @var{macro} for non C-like macros where
12915 the macro may begin with a hyphen.
12917 @kindex info macros
12918 @item info macros @var{location}
12919 Show all macro definitions that are in effect at the location specified
12920 by @var{location}, and describe the source location or compiler
12921 command-line where those definitions were established.
12923 @kindex macro define
12924 @cindex user-defined macros
12925 @cindex defining macros interactively
12926 @cindex macros, user-defined
12927 @item macro define @var{macro} @var{replacement-list}
12928 @itemx macro define @var{macro}(@var{arglist}) @var{replacement-list}
12929 Introduce a definition for a preprocessor macro named @var{macro},
12930 invocations of which are replaced by the tokens given in
12931 @var{replacement-list}. The first form of this command defines an
12932 ``object-like'' macro, which takes no arguments; the second form
12933 defines a ``function-like'' macro, which takes the arguments given in
12936 A definition introduced by this command is in scope in every
12937 expression evaluated in @value{GDBN}, until it is removed with the
12938 @code{macro undef} command, described below. The definition overrides
12939 all definitions for @var{macro} present in the program being debugged,
12940 as well as any previous user-supplied definition.
12942 @kindex macro undef
12943 @item macro undef @var{macro}
12944 Remove any user-supplied definition for the macro named @var{macro}.
12945 This command only affects definitions provided with the @code{macro
12946 define} command, described above; it cannot remove definitions present
12947 in the program being debugged.
12951 List all the macros defined using the @code{macro define} command.
12954 @cindex macros, example of debugging with
12955 Here is a transcript showing the above commands in action. First, we
12956 show our source files:
12961 #include "sample.h"
12964 #define ADD(x) (M + x)
12969 printf ("Hello, world!\n");
12971 printf ("We're so creative.\n");
12973 printf ("Goodbye, world!\n");
12980 Now, we compile the program using the @sc{gnu} C compiler,
12981 @value{NGCC}. We pass the @option{-gdwarf-2}@footnote{This is the
12982 minimum. Recent versions of @value{NGCC} support @option{-gdwarf-3}
12983 and @option{-gdwarf-4}; we recommend always choosing the most recent
12984 version of DWARF.} @emph{and} @option{-g3} flags to ensure the compiler
12985 includes information about preprocessor macros in the debugging
12989 $ gcc -gdwarf-2 -g3 sample.c -o sample
12993 Now, we start @value{GDBN} on our sample program:
12997 GNU gdb 2002-05-06-cvs
12998 Copyright 2002 Free Software Foundation, Inc.
12999 GDB is free software, @dots{}
13003 We can expand macros and examine their definitions, even when the
13004 program is not running. @value{GDBN} uses the current listing position
13005 to decide which macro definitions are in scope:
13008 (@value{GDBP}) list main
13011 5 #define ADD(x) (M + x)
13016 10 printf ("Hello, world!\n");
13018 12 printf ("We're so creative.\n");
13019 (@value{GDBP}) info macro ADD
13020 Defined at /home/jimb/gdb/macros/play/sample.c:5
13021 #define ADD(x) (M + x)
13022 (@value{GDBP}) info macro Q
13023 Defined at /home/jimb/gdb/macros/play/sample.h:1
13024 included at /home/jimb/gdb/macros/play/sample.c:2
13026 (@value{GDBP}) macro expand ADD(1)
13027 expands to: (42 + 1)
13028 (@value{GDBP}) macro expand-once ADD(1)
13029 expands to: once (M + 1)
13033 In the example above, note that @code{macro expand-once} expands only
13034 the macro invocation explicit in the original text --- the invocation of
13035 @code{ADD} --- but does not expand the invocation of the macro @code{M},
13036 which was introduced by @code{ADD}.
13038 Once the program is running, @value{GDBN} uses the macro definitions in
13039 force at the source line of the current stack frame:
13042 (@value{GDBP}) break main
13043 Breakpoint 1 at 0x8048370: file sample.c, line 10.
13045 Starting program: /home/jimb/gdb/macros/play/sample
13047 Breakpoint 1, main () at sample.c:10
13048 10 printf ("Hello, world!\n");
13052 At line 10, the definition of the macro @code{N} at line 9 is in force:
13055 (@value{GDBP}) info macro N
13056 Defined at /home/jimb/gdb/macros/play/sample.c:9
13058 (@value{GDBP}) macro expand N Q M
13059 expands to: 28 < 42
13060 (@value{GDBP}) print N Q M
13065 As we step over directives that remove @code{N}'s definition, and then
13066 give it a new definition, @value{GDBN} finds the definition (or lack
13067 thereof) in force at each point:
13070 (@value{GDBP}) next
13072 12 printf ("We're so creative.\n");
13073 (@value{GDBP}) info macro N
13074 The symbol `N' has no definition as a C/C++ preprocessor macro
13075 at /home/jimb/gdb/macros/play/sample.c:12
13076 (@value{GDBP}) next
13078 14 printf ("Goodbye, world!\n");
13079 (@value{GDBP}) info macro N
13080 Defined at /home/jimb/gdb/macros/play/sample.c:13
13082 (@value{GDBP}) macro expand N Q M
13083 expands to: 1729 < 42
13084 (@value{GDBP}) print N Q M
13089 In addition to source files, macros can be defined on the compilation command
13090 line using the @option{-D@var{name}=@var{value}} syntax. For macros defined in
13091 such a way, @value{GDBN} displays the location of their definition as line zero
13092 of the source file submitted to the compiler.
13095 (@value{GDBP}) info macro __STDC__
13096 Defined at /home/jimb/gdb/macros/play/sample.c:0
13103 @chapter Tracepoints
13104 @c This chapter is based on the documentation written by Michael
13105 @c Snyder, David Taylor, Jim Blandy, and Elena Zannoni.
13107 @cindex tracepoints
13108 In some applications, it is not feasible for the debugger to interrupt
13109 the program's execution long enough for the developer to learn
13110 anything helpful about its behavior. If the program's correctness
13111 depends on its real-time behavior, delays introduced by a debugger
13112 might cause the program to change its behavior drastically, or perhaps
13113 fail, even when the code itself is correct. It is useful to be able
13114 to observe the program's behavior without interrupting it.
13116 Using @value{GDBN}'s @code{trace} and @code{collect} commands, you can
13117 specify locations in the program, called @dfn{tracepoints}, and
13118 arbitrary expressions to evaluate when those tracepoints are reached.
13119 Later, using the @code{tfind} command, you can examine the values
13120 those expressions had when the program hit the tracepoints. The
13121 expressions may also denote objects in memory---structures or arrays,
13122 for example---whose values @value{GDBN} should record; while visiting
13123 a particular tracepoint, you may inspect those objects as if they were
13124 in memory at that moment. However, because @value{GDBN} records these
13125 values without interacting with you, it can do so quickly and
13126 unobtrusively, hopefully not disturbing the program's behavior.
13128 The tracepoint facility is currently available only for remote
13129 targets. @xref{Targets}. In addition, your remote target must know
13130 how to collect trace data. This functionality is implemented in the
13131 remote stub; however, none of the stubs distributed with @value{GDBN}
13132 support tracepoints as of this writing. The format of the remote
13133 packets used to implement tracepoints are described in @ref{Tracepoint
13136 It is also possible to get trace data from a file, in a manner reminiscent
13137 of corefiles; you specify the filename, and use @code{tfind} to search
13138 through the file. @xref{Trace Files}, for more details.
13140 This chapter describes the tracepoint commands and features.
13143 * Set Tracepoints::
13144 * Analyze Collected Data::
13145 * Tracepoint Variables::
13149 @node Set Tracepoints
13150 @section Commands to Set Tracepoints
13152 Before running such a @dfn{trace experiment}, an arbitrary number of
13153 tracepoints can be set. A tracepoint is actually a special type of
13154 breakpoint (@pxref{Set Breaks}), so you can manipulate it using
13155 standard breakpoint commands. For instance, as with breakpoints,
13156 tracepoint numbers are successive integers starting from one, and many
13157 of the commands associated with tracepoints take the tracepoint number
13158 as their argument, to identify which tracepoint to work on.
13160 For each tracepoint, you can specify, in advance, some arbitrary set
13161 of data that you want the target to collect in the trace buffer when
13162 it hits that tracepoint. The collected data can include registers,
13163 local variables, or global data. Later, you can use @value{GDBN}
13164 commands to examine the values these data had at the time the
13165 tracepoint was hit.
13167 Tracepoints do not support every breakpoint feature. Ignore counts on
13168 tracepoints have no effect, and tracepoints cannot run @value{GDBN}
13169 commands when they are hit. Tracepoints may not be thread-specific
13172 @cindex fast tracepoints
13173 Some targets may support @dfn{fast tracepoints}, which are inserted in
13174 a different way (such as with a jump instead of a trap), that is
13175 faster but possibly restricted in where they may be installed.
13177 @cindex static tracepoints
13178 @cindex markers, static tracepoints
13179 @cindex probing markers, static tracepoints
13180 Regular and fast tracepoints are dynamic tracing facilities, meaning
13181 that they can be used to insert tracepoints at (almost) any location
13182 in the target. Some targets may also support controlling @dfn{static
13183 tracepoints} from @value{GDBN}. With static tracing, a set of
13184 instrumentation points, also known as @dfn{markers}, are embedded in
13185 the target program, and can be activated or deactivated by name or
13186 address. These are usually placed at locations which facilitate
13187 investigating what the target is actually doing. @value{GDBN}'s
13188 support for static tracing includes being able to list instrumentation
13189 points, and attach them with @value{GDBN} defined high level
13190 tracepoints that expose the whole range of convenience of
13191 @value{GDBN}'s tracepoints support. Namely, support for collecting
13192 registers values and values of global or local (to the instrumentation
13193 point) variables; tracepoint conditions and trace state variables.
13194 The act of installing a @value{GDBN} static tracepoint on an
13195 instrumentation point, or marker, is referred to as @dfn{probing} a
13196 static tracepoint marker.
13198 @code{gdbserver} supports tracepoints on some target systems.
13199 @xref{Server,,Tracepoints support in @code{gdbserver}}.
13201 This section describes commands to set tracepoints and associated
13202 conditions and actions.
13205 * Create and Delete Tracepoints::
13206 * Enable and Disable Tracepoints::
13207 * Tracepoint Passcounts::
13208 * Tracepoint Conditions::
13209 * Trace State Variables::
13210 * Tracepoint Actions::
13211 * Listing Tracepoints::
13212 * Listing Static Tracepoint Markers::
13213 * Starting and Stopping Trace Experiments::
13214 * Tracepoint Restrictions::
13217 @node Create and Delete Tracepoints
13218 @subsection Create and Delete Tracepoints
13221 @cindex set tracepoint
13223 @item trace @var{location}
13224 The @code{trace} command is very similar to the @code{break} command.
13225 Its argument @var{location} can be any valid location.
13226 @xref{Specify Location}. The @code{trace} command defines a tracepoint,
13227 which is a point in the target program where the debugger will briefly stop,
13228 collect some data, and then allow the program to continue. Setting a tracepoint
13229 or changing its actions takes effect immediately if the remote stub
13230 supports the @samp{InstallInTrace} feature (@pxref{install tracepoint
13232 If remote stub doesn't support the @samp{InstallInTrace} feature, all
13233 these changes don't take effect until the next @code{tstart}
13234 command, and once a trace experiment is running, further changes will
13235 not have any effect until the next trace experiment starts. In addition,
13236 @value{GDBN} supports @dfn{pending tracepoints}---tracepoints whose
13237 address is not yet resolved. (This is similar to pending breakpoints.)
13238 Pending tracepoints are not downloaded to the target and not installed
13239 until they are resolved. The resolution of pending tracepoints requires
13240 @value{GDBN} support---when debugging with the remote target, and
13241 @value{GDBN} disconnects from the remote stub (@pxref{disconnected
13242 tracing}), pending tracepoints can not be resolved (and downloaded to
13243 the remote stub) while @value{GDBN} is disconnected.
13245 Here are some examples of using the @code{trace} command:
13248 (@value{GDBP}) @b{trace foo.c:121} // a source file and line number
13250 (@value{GDBP}) @b{trace +2} // 2 lines forward
13252 (@value{GDBP}) @b{trace my_function} // first source line of function
13254 (@value{GDBP}) @b{trace *my_function} // EXACT start address of function
13256 (@value{GDBP}) @b{trace *0x2117c4} // an address
13260 You can abbreviate @code{trace} as @code{tr}.
13262 @item trace @var{location} if @var{cond}
13263 Set a tracepoint with condition @var{cond}; evaluate the expression
13264 @var{cond} each time the tracepoint is reached, and collect data only
13265 if the value is nonzero---that is, if @var{cond} evaluates as true.
13266 @xref{Tracepoint Conditions, ,Tracepoint Conditions}, for more
13267 information on tracepoint conditions.
13269 @item ftrace @var{location} [ if @var{cond} ]
13270 @cindex set fast tracepoint
13271 @cindex fast tracepoints, setting
13273 The @code{ftrace} command sets a fast tracepoint. For targets that
13274 support them, fast tracepoints will use a more efficient but possibly
13275 less general technique to trigger data collection, such as a jump
13276 instruction instead of a trap, or some sort of hardware support. It
13277 may not be possible to create a fast tracepoint at the desired
13278 location, in which case the command will exit with an explanatory
13281 @value{GDBN} handles arguments to @code{ftrace} exactly as for
13284 On 32-bit x86-architecture systems, fast tracepoints normally need to
13285 be placed at an instruction that is 5 bytes or longer, but can be
13286 placed at 4-byte instructions if the low 64K of memory of the target
13287 program is available to install trampolines. Some Unix-type systems,
13288 such as @sc{gnu}/Linux, exclude low addresses from the program's
13289 address space; but for instance with the Linux kernel it is possible
13290 to let @value{GDBN} use this area by doing a @command{sysctl} command
13291 to set the @code{mmap_min_addr} kernel parameter, as in
13294 sudo sysctl -w vm.mmap_min_addr=32768
13298 which sets the low address to 32K, which leaves plenty of room for
13299 trampolines. The minimum address should be set to a page boundary.
13301 @item strace @var{location} [ if @var{cond} ]
13302 @cindex set static tracepoint
13303 @cindex static tracepoints, setting
13304 @cindex probe static tracepoint marker
13306 The @code{strace} command sets a static tracepoint. For targets that
13307 support it, setting a static tracepoint probes a static
13308 instrumentation point, or marker, found at @var{location}. It may not
13309 be possible to set a static tracepoint at the desired location, in
13310 which case the command will exit with an explanatory message.
13312 @value{GDBN} handles arguments to @code{strace} exactly as for
13313 @code{trace}, with the addition that the user can also specify
13314 @code{-m @var{marker}} as @var{location}. This probes the marker
13315 identified by the @var{marker} string identifier. This identifier
13316 depends on the static tracepoint backend library your program is
13317 using. You can find all the marker identifiers in the @samp{ID} field
13318 of the @code{info static-tracepoint-markers} command output.
13319 @xref{Listing Static Tracepoint Markers,,Listing Static Tracepoint
13320 Markers}. For example, in the following small program using the UST
13326 trace_mark(ust, bar33, "str %s", "FOOBAZ");
13331 the marker id is composed of joining the first two arguments to the
13332 @code{trace_mark} call with a slash, which translates to:
13335 (@value{GDBP}) info static-tracepoint-markers
13336 Cnt Enb ID Address What
13337 1 n ust/bar33 0x0000000000400ddc in main at stexample.c:22
13343 so you may probe the marker above with:
13346 (@value{GDBP}) strace -m ust/bar33
13349 Static tracepoints accept an extra collect action --- @code{collect
13350 $_sdata}. This collects arbitrary user data passed in the probe point
13351 call to the tracing library. In the UST example above, you'll see
13352 that the third argument to @code{trace_mark} is a printf-like format
13353 string. The user data is then the result of running that formating
13354 string against the following arguments. Note that @code{info
13355 static-tracepoint-markers} command output lists that format string in
13356 the @samp{Data:} field.
13358 You can inspect this data when analyzing the trace buffer, by printing
13359 the $_sdata variable like any other variable available to
13360 @value{GDBN}. @xref{Tracepoint Actions,,Tracepoint Action Lists}.
13363 @cindex last tracepoint number
13364 @cindex recent tracepoint number
13365 @cindex tracepoint number
13366 The convenience variable @code{$tpnum} records the tracepoint number
13367 of the most recently set tracepoint.
13369 @kindex delete tracepoint
13370 @cindex tracepoint deletion
13371 @item delete tracepoint @r{[}@var{num}@r{]}
13372 Permanently delete one or more tracepoints. With no argument, the
13373 default is to delete all tracepoints. Note that the regular
13374 @code{delete} command can remove tracepoints also.
13379 (@value{GDBP}) @b{delete trace 1 2 3} // remove three tracepoints
13381 (@value{GDBP}) @b{delete trace} // remove all tracepoints
13385 You can abbreviate this command as @code{del tr}.
13388 @node Enable and Disable Tracepoints
13389 @subsection Enable and Disable Tracepoints
13391 These commands are deprecated; they are equivalent to plain @code{disable} and @code{enable}.
13394 @kindex disable tracepoint
13395 @item disable tracepoint @r{[}@var{num}@r{]}
13396 Disable tracepoint @var{num}, or all tracepoints if no argument
13397 @var{num} is given. A disabled tracepoint will have no effect during
13398 a trace experiment, but it is not forgotten. You can re-enable
13399 a disabled tracepoint using the @code{enable tracepoint} command.
13400 If the command is issued during a trace experiment and the debug target
13401 has support for disabling tracepoints during a trace experiment, then the
13402 change will be effective immediately. Otherwise, it will be applied to the
13403 next trace experiment.
13405 @kindex enable tracepoint
13406 @item enable tracepoint @r{[}@var{num}@r{]}
13407 Enable tracepoint @var{num}, or all tracepoints. If this command is
13408 issued during a trace experiment and the debug target supports enabling
13409 tracepoints during a trace experiment, then the enabled tracepoints will
13410 become effective immediately. Otherwise, they will become effective the
13411 next time a trace experiment is run.
13414 @node Tracepoint Passcounts
13415 @subsection Tracepoint Passcounts
13419 @cindex tracepoint pass count
13420 @item passcount @r{[}@var{n} @r{[}@var{num}@r{]]}
13421 Set the @dfn{passcount} of a tracepoint. The passcount is a way to
13422 automatically stop a trace experiment. If a tracepoint's passcount is
13423 @var{n}, then the trace experiment will be automatically stopped on
13424 the @var{n}'th time that tracepoint is hit. If the tracepoint number
13425 @var{num} is not specified, the @code{passcount} command sets the
13426 passcount of the most recently defined tracepoint. If no passcount is
13427 given, the trace experiment will run until stopped explicitly by the
13433 (@value{GDBP}) @b{passcount 5 2} // Stop on the 5th execution of
13434 @exdent @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @code{// tracepoint 2}
13436 (@value{GDBP}) @b{passcount 12} // Stop on the 12th execution of the
13437 @exdent @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @code{// most recently defined tracepoint.}
13438 (@value{GDBP}) @b{trace foo}
13439 (@value{GDBP}) @b{pass 3}
13440 (@value{GDBP}) @b{trace bar}
13441 (@value{GDBP}) @b{pass 2}
13442 (@value{GDBP}) @b{trace baz}
13443 (@value{GDBP}) @b{pass 1} // Stop tracing when foo has been
13444 @exdent @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @code{// executed 3 times OR when bar has}
13445 @exdent @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @code{// been executed 2 times}
13446 @exdent @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @code{// OR when baz has been executed 1 time.}
13450 @node Tracepoint Conditions
13451 @subsection Tracepoint Conditions
13452 @cindex conditional tracepoints
13453 @cindex tracepoint conditions
13455 The simplest sort of tracepoint collects data every time your program
13456 reaches a specified place. You can also specify a @dfn{condition} for
13457 a tracepoint. A condition is just a Boolean expression in your
13458 programming language (@pxref{Expressions, ,Expressions}). A
13459 tracepoint with a condition evaluates the expression each time your
13460 program reaches it, and data collection happens only if the condition
13463 Tracepoint conditions can be specified when a tracepoint is set, by
13464 using @samp{if} in the arguments to the @code{trace} command.
13465 @xref{Create and Delete Tracepoints, ,Setting Tracepoints}. They can
13466 also be set or changed at any time with the @code{condition} command,
13467 just as with breakpoints.
13469 Unlike breakpoint conditions, @value{GDBN} does not actually evaluate
13470 the conditional expression itself. Instead, @value{GDBN} encodes the
13471 expression into an agent expression (@pxref{Agent Expressions})
13472 suitable for execution on the target, independently of @value{GDBN}.
13473 Global variables become raw memory locations, locals become stack
13474 accesses, and so forth.
13476 For instance, suppose you have a function that is usually called
13477 frequently, but should not be called after an error has occurred. You
13478 could use the following tracepoint command to collect data about calls
13479 of that function that happen while the error code is propagating
13480 through the program; an unconditional tracepoint could end up
13481 collecting thousands of useless trace frames that you would have to
13485 (@value{GDBP}) @kbd{trace normal_operation if errcode > 0}
13488 @node Trace State Variables
13489 @subsection Trace State Variables
13490 @cindex trace state variables
13492 A @dfn{trace state variable} is a special type of variable that is
13493 created and managed by target-side code. The syntax is the same as
13494 that for GDB's convenience variables (a string prefixed with ``$''),
13495 but they are stored on the target. They must be created explicitly,
13496 using a @code{tvariable} command. They are always 64-bit signed
13499 Trace state variables are remembered by @value{GDBN}, and downloaded
13500 to the target along with tracepoint information when the trace
13501 experiment starts. There are no intrinsic limits on the number of
13502 trace state variables, beyond memory limitations of the target.
13504 @cindex convenience variables, and trace state variables
13505 Although trace state variables are managed by the target, you can use
13506 them in print commands and expressions as if they were convenience
13507 variables; @value{GDBN} will get the current value from the target
13508 while the trace experiment is running. Trace state variables share
13509 the same namespace as other ``$'' variables, which means that you
13510 cannot have trace state variables with names like @code{$23} or
13511 @code{$pc}, nor can you have a trace state variable and a convenience
13512 variable with the same name.
13516 @item tvariable $@var{name} [ = @var{expression} ]
13518 The @code{tvariable} command creates a new trace state variable named
13519 @code{$@var{name}}, and optionally gives it an initial value of
13520 @var{expression}. The @var{expression} is evaluated when this command is
13521 entered; the result will be converted to an integer if possible,
13522 otherwise @value{GDBN} will report an error. A subsequent
13523 @code{tvariable} command specifying the same name does not create a
13524 variable, but instead assigns the supplied initial value to the
13525 existing variable of that name, overwriting any previous initial
13526 value. The default initial value is 0.
13528 @item info tvariables
13529 @kindex info tvariables
13530 List all the trace state variables along with their initial values.
13531 Their current values may also be displayed, if the trace experiment is
13534 @item delete tvariable @r{[} $@var{name} @dots{} @r{]}
13535 @kindex delete tvariable
13536 Delete the given trace state variables, or all of them if no arguments
13541 @node Tracepoint Actions
13542 @subsection Tracepoint Action Lists
13546 @cindex tracepoint actions
13547 @item actions @r{[}@var{num}@r{]}
13548 This command will prompt for a list of actions to be taken when the
13549 tracepoint is hit. If the tracepoint number @var{num} is not
13550 specified, this command sets the actions for the one that was most
13551 recently defined (so that you can define a tracepoint and then say
13552 @code{actions} without bothering about its number). You specify the
13553 actions themselves on the following lines, one action at a time, and
13554 terminate the actions list with a line containing just @code{end}. So
13555 far, the only defined actions are @code{collect}, @code{teval}, and
13556 @code{while-stepping}.
13558 @code{actions} is actually equivalent to @code{commands} (@pxref{Break
13559 Commands, ,Breakpoint Command Lists}), except that only the defined
13560 actions are allowed; any other @value{GDBN} command is rejected.
13562 @cindex remove actions from a tracepoint
13563 To remove all actions from a tracepoint, type @samp{actions @var{num}}
13564 and follow it immediately with @samp{end}.
13567 (@value{GDBP}) @b{collect @var{data}} // collect some data
13569 (@value{GDBP}) @b{while-stepping 5} // single-step 5 times, collect data
13571 (@value{GDBP}) @b{end} // signals the end of actions.
13574 In the following example, the action list begins with @code{collect}
13575 commands indicating the things to be collected when the tracepoint is
13576 hit. Then, in order to single-step and collect additional data
13577 following the tracepoint, a @code{while-stepping} command is used,
13578 followed by the list of things to be collected after each step in a
13579 sequence of single steps. The @code{while-stepping} command is
13580 terminated by its own separate @code{end} command. Lastly, the action
13581 list is terminated by an @code{end} command.
13584 (@value{GDBP}) @b{trace foo}
13585 (@value{GDBP}) @b{actions}
13586 Enter actions for tracepoint 1, one per line:
13589 > while-stepping 12
13590 > collect $pc, arr[i]
13595 @kindex collect @r{(tracepoints)}
13596 @item collect@r{[}/@var{mods}@r{]} @var{expr1}, @var{expr2}, @dots{}
13597 Collect values of the given expressions when the tracepoint is hit.
13598 This command accepts a comma-separated list of any valid expressions.
13599 In addition to global, static, or local variables, the following
13600 special arguments are supported:
13604 Collect all registers.
13607 Collect all function arguments.
13610 Collect all local variables.
13613 Collect the return address. This is helpful if you want to see more
13616 @emph{Note:} The return address location can not always be reliably
13617 determined up front, and the wrong address / registers may end up
13618 collected instead. On some architectures the reliability is higher
13619 for tracepoints at function entry, while on others it's the opposite.
13620 When this happens, backtracing will stop because the return address is
13621 found unavailable (unless another collect rule happened to match it).
13624 Collects the number of arguments from the static probe at which the
13625 tracepoint is located.
13626 @xref{Static Probe Points}.
13628 @item $_probe_arg@var{n}
13629 @var{n} is an integer between 0 and 11. Collects the @var{n}th argument
13630 from the static probe at which the tracepoint is located.
13631 @xref{Static Probe Points}.
13634 @vindex $_sdata@r{, collect}
13635 Collect static tracepoint marker specific data. Only available for
13636 static tracepoints. @xref{Tracepoint Actions,,Tracepoint Action
13637 Lists}. On the UST static tracepoints library backend, an
13638 instrumentation point resembles a @code{printf} function call. The
13639 tracing library is able to collect user specified data formatted to a
13640 character string using the format provided by the programmer that
13641 instrumented the program. Other backends have similar mechanisms.
13642 Here's an example of a UST marker call:
13645 const char master_name[] = "$your_name";
13646 trace_mark(channel1, marker1, "hello %s", master_name)
13649 In this case, collecting @code{$_sdata} collects the string
13650 @samp{hello $yourname}. When analyzing the trace buffer, you can
13651 inspect @samp{$_sdata} like any other variable available to
13655 You can give several consecutive @code{collect} commands, each one
13656 with a single argument, or one @code{collect} command with several
13657 arguments separated by commas; the effect is the same.
13659 The optional @var{mods} changes the usual handling of the arguments.
13660 @code{s} requests that pointers to chars be handled as strings, in
13661 particular collecting the contents of the memory being pointed at, up
13662 to the first zero. The upper bound is by default the value of the
13663 @code{print elements} variable; if @code{s} is followed by a decimal
13664 number, that is the upper bound instead. So for instance
13665 @samp{collect/s25 mystr} collects as many as 25 characters at
13668 The command @code{info scope} (@pxref{Symbols, info scope}) is
13669 particularly useful for figuring out what data to collect.
13671 @kindex teval @r{(tracepoints)}
13672 @item teval @var{expr1}, @var{expr2}, @dots{}
13673 Evaluate the given expressions when the tracepoint is hit. This
13674 command accepts a comma-separated list of expressions. The results
13675 are discarded, so this is mainly useful for assigning values to trace
13676 state variables (@pxref{Trace State Variables}) without adding those
13677 values to the trace buffer, as would be the case if the @code{collect}
13680 @kindex while-stepping @r{(tracepoints)}
13681 @item while-stepping @var{n}
13682 Perform @var{n} single-step instruction traces after the tracepoint,
13683 collecting new data after each step. The @code{while-stepping}
13684 command is followed by the list of what to collect while stepping
13685 (followed by its own @code{end} command):
13688 > while-stepping 12
13689 > collect $regs, myglobal
13695 Note that @code{$pc} is not automatically collected by
13696 @code{while-stepping}; you need to explicitly collect that register if
13697 you need it. You may abbreviate @code{while-stepping} as @code{ws} or
13700 @item set default-collect @var{expr1}, @var{expr2}, @dots{}
13701 @kindex set default-collect
13702 @cindex default collection action
13703 This variable is a list of expressions to collect at each tracepoint
13704 hit. It is effectively an additional @code{collect} action prepended
13705 to every tracepoint action list. The expressions are parsed
13706 individually for each tracepoint, so for instance a variable named
13707 @code{xyz} may be interpreted as a global for one tracepoint, and a
13708 local for another, as appropriate to the tracepoint's location.
13710 @item show default-collect
13711 @kindex show default-collect
13712 Show the list of expressions that are collected by default at each
13717 @node Listing Tracepoints
13718 @subsection Listing Tracepoints
13721 @kindex info tracepoints @r{[}@var{n}@dots{}@r{]}
13722 @kindex info tp @r{[}@var{n}@dots{}@r{]}
13723 @cindex information about tracepoints
13724 @item info tracepoints @r{[}@var{num}@dots{}@r{]}
13725 Display information about the tracepoint @var{num}. If you don't
13726 specify a tracepoint number, displays information about all the
13727 tracepoints defined so far. The format is similar to that used for
13728 @code{info breakpoints}; in fact, @code{info tracepoints} is the same
13729 command, simply restricting itself to tracepoints.
13731 A tracepoint's listing may include additional information specific to
13736 its passcount as given by the @code{passcount @var{n}} command
13739 the state about installed on target of each location
13743 (@value{GDBP}) @b{info trace}
13744 Num Type Disp Enb Address What
13745 1 tracepoint keep y 0x0804ab57 in foo() at main.cxx:7
13747 collect globfoo, $regs
13752 2 tracepoint keep y <MULTIPLE>
13754 2.1 y 0x0804859c in func4 at change-loc.h:35
13755 installed on target
13756 2.2 y 0xb7ffc480 in func4 at change-loc.h:35
13757 installed on target
13758 2.3 y <PENDING> set_tracepoint
13759 3 tracepoint keep y 0x080485b1 in foo at change-loc.c:29
13760 not installed on target
13765 This command can be abbreviated @code{info tp}.
13768 @node Listing Static Tracepoint Markers
13769 @subsection Listing Static Tracepoint Markers
13772 @kindex info static-tracepoint-markers
13773 @cindex information about static tracepoint markers
13774 @item info static-tracepoint-markers
13775 Display information about all static tracepoint markers defined in the
13778 For each marker, the following columns are printed:
13782 An incrementing counter, output to help readability. This is not a
13785 The marker ID, as reported by the target.
13786 @item Enabled or Disabled
13787 Probed markers are tagged with @samp{y}. @samp{n} identifies marks
13788 that are not enabled.
13790 Where the marker is in your program, as a memory address.
13792 Where the marker is in the source for your program, as a file and line
13793 number. If the debug information included in the program does not
13794 allow @value{GDBN} to locate the source of the marker, this column
13795 will be left blank.
13799 In addition, the following information may be printed for each marker:
13803 User data passed to the tracing library by the marker call. In the
13804 UST backend, this is the format string passed as argument to the
13806 @item Static tracepoints probing the marker
13807 The list of static tracepoints attached to the marker.
13811 (@value{GDBP}) info static-tracepoint-markers
13812 Cnt ID Enb Address What
13813 1 ust/bar2 y 0x0000000000400e1a in main at stexample.c:25
13814 Data: number1 %d number2 %d
13815 Probed by static tracepoints: #2
13816 2 ust/bar33 n 0x0000000000400c87 in main at stexample.c:24
13822 @node Starting and Stopping Trace Experiments
13823 @subsection Starting and Stopping Trace Experiments
13826 @kindex tstart [ @var{notes} ]
13827 @cindex start a new trace experiment
13828 @cindex collected data discarded
13830 This command starts the trace experiment, and begins collecting data.
13831 It has the side effect of discarding all the data collected in the
13832 trace buffer during the previous trace experiment. If any arguments
13833 are supplied, they are taken as a note and stored with the trace
13834 experiment's state. The notes may be arbitrary text, and are
13835 especially useful with disconnected tracing in a multi-user context;
13836 the notes can explain what the trace is doing, supply user contact
13837 information, and so forth.
13839 @kindex tstop [ @var{notes} ]
13840 @cindex stop a running trace experiment
13842 This command stops the trace experiment. If any arguments are
13843 supplied, they are recorded with the experiment as a note. This is
13844 useful if you are stopping a trace started by someone else, for
13845 instance if the trace is interfering with the system's behavior and
13846 needs to be stopped quickly.
13848 @strong{Note}: a trace experiment and data collection may stop
13849 automatically if any tracepoint's passcount is reached
13850 (@pxref{Tracepoint Passcounts}), or if the trace buffer becomes full.
13853 @cindex status of trace data collection
13854 @cindex trace experiment, status of
13856 This command displays the status of the current trace data
13860 Here is an example of the commands we described so far:
13863 (@value{GDBP}) @b{trace gdb_c_test}
13864 (@value{GDBP}) @b{actions}
13865 Enter actions for tracepoint #1, one per line.
13866 > collect $regs,$locals,$args
13867 > while-stepping 11
13871 (@value{GDBP}) @b{tstart}
13872 [time passes @dots{}]
13873 (@value{GDBP}) @b{tstop}
13876 @anchor{disconnected tracing}
13877 @cindex disconnected tracing
13878 You can choose to continue running the trace experiment even if
13879 @value{GDBN} disconnects from the target, voluntarily or
13880 involuntarily. For commands such as @code{detach}, the debugger will
13881 ask what you want to do with the trace. But for unexpected
13882 terminations (@value{GDBN} crash, network outage), it would be
13883 unfortunate to lose hard-won trace data, so the variable
13884 @code{disconnected-tracing} lets you decide whether the trace should
13885 continue running without @value{GDBN}.
13888 @item set disconnected-tracing on
13889 @itemx set disconnected-tracing off
13890 @kindex set disconnected-tracing
13891 Choose whether a tracing run should continue to run if @value{GDBN}
13892 has disconnected from the target. Note that @code{detach} or
13893 @code{quit} will ask you directly what to do about a running trace no
13894 matter what this variable's setting, so the variable is mainly useful
13895 for handling unexpected situations, such as loss of the network.
13897 @item show disconnected-tracing
13898 @kindex show disconnected-tracing
13899 Show the current choice for disconnected tracing.
13903 When you reconnect to the target, the trace experiment may or may not
13904 still be running; it might have filled the trace buffer in the
13905 meantime, or stopped for one of the other reasons. If it is running,
13906 it will continue after reconnection.
13908 Upon reconnection, the target will upload information about the
13909 tracepoints in effect. @value{GDBN} will then compare that
13910 information to the set of tracepoints currently defined, and attempt
13911 to match them up, allowing for the possibility that the numbers may
13912 have changed due to creation and deletion in the meantime. If one of
13913 the target's tracepoints does not match any in @value{GDBN}, the
13914 debugger will create a new tracepoint, so that you have a number with
13915 which to specify that tracepoint. This matching-up process is
13916 necessarily heuristic, and it may result in useless tracepoints being
13917 created; you may simply delete them if they are of no use.
13919 @cindex circular trace buffer
13920 If your target agent supports a @dfn{circular trace buffer}, then you
13921 can run a trace experiment indefinitely without filling the trace
13922 buffer; when space runs out, the agent deletes already-collected trace
13923 frames, oldest first, until there is enough room to continue
13924 collecting. This is especially useful if your tracepoints are being
13925 hit too often, and your trace gets terminated prematurely because the
13926 buffer is full. To ask for a circular trace buffer, simply set
13927 @samp{circular-trace-buffer} to on. You can set this at any time,
13928 including during tracing; if the agent can do it, it will change
13929 buffer handling on the fly, otherwise it will not take effect until
13933 @item set circular-trace-buffer on
13934 @itemx set circular-trace-buffer off
13935 @kindex set circular-trace-buffer
13936 Choose whether a tracing run should use a linear or circular buffer
13937 for trace data. A linear buffer will not lose any trace data, but may
13938 fill up prematurely, while a circular buffer will discard old trace
13939 data, but it will have always room for the latest tracepoint hits.
13941 @item show circular-trace-buffer
13942 @kindex show circular-trace-buffer
13943 Show the current choice for the trace buffer. Note that this may not
13944 match the agent's current buffer handling, nor is it guaranteed to
13945 match the setting that might have been in effect during a past run,
13946 for instance if you are looking at frames from a trace file.
13951 @item set trace-buffer-size @var{n}
13952 @itemx set trace-buffer-size unlimited
13953 @kindex set trace-buffer-size
13954 Request that the target use a trace buffer of @var{n} bytes. Not all
13955 targets will honor the request; they may have a compiled-in size for
13956 the trace buffer, or some other limitation. Set to a value of
13957 @code{unlimited} or @code{-1} to let the target use whatever size it
13958 likes. This is also the default.
13960 @item show trace-buffer-size
13961 @kindex show trace-buffer-size
13962 Show the current requested size for the trace buffer. Note that this
13963 will only match the actual size if the target supports size-setting,
13964 and was able to handle the requested size. For instance, if the
13965 target can only change buffer size between runs, this variable will
13966 not reflect the change until the next run starts. Use @code{tstatus}
13967 to get a report of the actual buffer size.
13971 @item set trace-user @var{text}
13972 @kindex set trace-user
13974 @item show trace-user
13975 @kindex show trace-user
13977 @item set trace-notes @var{text}
13978 @kindex set trace-notes
13979 Set the trace run's notes.
13981 @item show trace-notes
13982 @kindex show trace-notes
13983 Show the trace run's notes.
13985 @item set trace-stop-notes @var{text}
13986 @kindex set trace-stop-notes
13987 Set the trace run's stop notes. The handling of the note is as for
13988 @code{tstop} arguments; the set command is convenient way to fix a
13989 stop note that is mistaken or incomplete.
13991 @item show trace-stop-notes
13992 @kindex show trace-stop-notes
13993 Show the trace run's stop notes.
13997 @node Tracepoint Restrictions
13998 @subsection Tracepoint Restrictions
14000 @cindex tracepoint restrictions
14001 There are a number of restrictions on the use of tracepoints. As
14002 described above, tracepoint data gathering occurs on the target
14003 without interaction from @value{GDBN}. Thus the full capabilities of
14004 the debugger are not available during data gathering, and then at data
14005 examination time, you will be limited by only having what was
14006 collected. The following items describe some common problems, but it
14007 is not exhaustive, and you may run into additional difficulties not
14013 Tracepoint expressions are intended to gather objects (lvalues). Thus
14014 the full flexibility of GDB's expression evaluator is not available.
14015 You cannot call functions, cast objects to aggregate types, access
14016 convenience variables or modify values (except by assignment to trace
14017 state variables). Some language features may implicitly call
14018 functions (for instance Objective-C fields with accessors), and therefore
14019 cannot be collected either.
14022 Collection of local variables, either individually or in bulk with
14023 @code{$locals} or @code{$args}, during @code{while-stepping} may
14024 behave erratically. The stepping action may enter a new scope (for
14025 instance by stepping into a function), or the location of the variable
14026 may change (for instance it is loaded into a register). The
14027 tracepoint data recorded uses the location information for the
14028 variables that is correct for the tracepoint location. When the
14029 tracepoint is created, it is not possible, in general, to determine
14030 where the steps of a @code{while-stepping} sequence will advance the
14031 program---particularly if a conditional branch is stepped.
14034 Collection of an incompletely-initialized or partially-destroyed object
14035 may result in something that @value{GDBN} cannot display, or displays
14036 in a misleading way.
14039 When @value{GDBN} displays a pointer to character it automatically
14040 dereferences the pointer to also display characters of the string
14041 being pointed to. However, collecting the pointer during tracing does
14042 not automatically collect the string. You need to explicitly
14043 dereference the pointer and provide size information if you want to
14044 collect not only the pointer, but the memory pointed to. For example,
14045 @code{*ptr@@50} can be used to collect the 50 element array pointed to
14049 It is not possible to collect a complete stack backtrace at a
14050 tracepoint. Instead, you may collect the registers and a few hundred
14051 bytes from the stack pointer with something like @code{*(unsigned char *)$esp@@300}
14052 (adjust to use the name of the actual stack pointer register on your
14053 target architecture, and the amount of stack you wish to capture).
14054 Then the @code{backtrace} command will show a partial backtrace when
14055 using a trace frame. The number of stack frames that can be examined
14056 depends on the sizes of the frames in the collected stack. Note that
14057 if you ask for a block so large that it goes past the bottom of the
14058 stack, the target agent may report an error trying to read from an
14062 If you do not collect registers at a tracepoint, @value{GDBN} can
14063 infer that the value of @code{$pc} must be the same as the address of
14064 the tracepoint and use that when you are looking at a trace frame
14065 for that tracepoint. However, this cannot work if the tracepoint has
14066 multiple locations (for instance if it was set in a function that was
14067 inlined), or if it has a @code{while-stepping} loop. In those cases
14068 @value{GDBN} will warn you that it can't infer @code{$pc}, and default
14073 @node Analyze Collected Data
14074 @section Using the Collected Data
14076 After the tracepoint experiment ends, you use @value{GDBN} commands
14077 for examining the trace data. The basic idea is that each tracepoint
14078 collects a trace @dfn{snapshot} every time it is hit and another
14079 snapshot every time it single-steps. All these snapshots are
14080 consecutively numbered from zero and go into a buffer, and you can
14081 examine them later. The way you examine them is to @dfn{focus} on a
14082 specific trace snapshot. When the remote stub is focused on a trace
14083 snapshot, it will respond to all @value{GDBN} requests for memory and
14084 registers by reading from the buffer which belongs to that snapshot,
14085 rather than from @emph{real} memory or registers of the program being
14086 debugged. This means that @strong{all} @value{GDBN} commands
14087 (@code{print}, @code{info registers}, @code{backtrace}, etc.) will
14088 behave as if we were currently debugging the program state as it was
14089 when the tracepoint occurred. Any requests for data that are not in
14090 the buffer will fail.
14093 * tfind:: How to select a trace snapshot
14094 * tdump:: How to display all data for a snapshot
14095 * save tracepoints:: How to save tracepoints for a future run
14099 @subsection @code{tfind @var{n}}
14102 @cindex select trace snapshot
14103 @cindex find trace snapshot
14104 The basic command for selecting a trace snapshot from the buffer is
14105 @code{tfind @var{n}}, which finds trace snapshot number @var{n},
14106 counting from zero. If no argument @var{n} is given, the next
14107 snapshot is selected.
14109 Here are the various forms of using the @code{tfind} command.
14113 Find the first snapshot in the buffer. This is a synonym for
14114 @code{tfind 0} (since 0 is the number of the first snapshot).
14117 Stop debugging trace snapshots, resume @emph{live} debugging.
14120 Same as @samp{tfind none}.
14123 No argument means find the next trace snapshot or find the first
14124 one if no trace snapshot is selected.
14127 Find the previous trace snapshot before the current one. This permits
14128 retracing earlier steps.
14130 @item tfind tracepoint @var{num}
14131 Find the next snapshot associated with tracepoint @var{num}. Search
14132 proceeds forward from the last examined trace snapshot. If no
14133 argument @var{num} is given, it means find the next snapshot collected
14134 for the same tracepoint as the current snapshot.
14136 @item tfind pc @var{addr}
14137 Find the next snapshot associated with the value @var{addr} of the
14138 program counter. Search proceeds forward from the last examined trace
14139 snapshot. If no argument @var{addr} is given, it means find the next
14140 snapshot with the same value of PC as the current snapshot.
14142 @item tfind outside @var{addr1}, @var{addr2}
14143 Find the next snapshot whose PC is outside the given range of
14144 addresses (exclusive).
14146 @item tfind range @var{addr1}, @var{addr2}
14147 Find the next snapshot whose PC is between @var{addr1} and
14148 @var{addr2} (inclusive).
14150 @item tfind line @r{[}@var{file}:@r{]}@var{n}
14151 Find the next snapshot associated with the source line @var{n}. If
14152 the optional argument @var{file} is given, refer to line @var{n} in
14153 that source file. Search proceeds forward from the last examined
14154 trace snapshot. If no argument @var{n} is given, it means find the
14155 next line other than the one currently being examined; thus saying
14156 @code{tfind line} repeatedly can appear to have the same effect as
14157 stepping from line to line in a @emph{live} debugging session.
14160 The default arguments for the @code{tfind} commands are specifically
14161 designed to make it easy to scan through the trace buffer. For
14162 instance, @code{tfind} with no argument selects the next trace
14163 snapshot, and @code{tfind -} with no argument selects the previous
14164 trace snapshot. So, by giving one @code{tfind} command, and then
14165 simply hitting @key{RET} repeatedly you can examine all the trace
14166 snapshots in order. Or, by saying @code{tfind -} and then hitting
14167 @key{RET} repeatedly you can examine the snapshots in reverse order.
14168 The @code{tfind line} command with no argument selects the snapshot
14169 for the next source line executed. The @code{tfind pc} command with
14170 no argument selects the next snapshot with the same program counter
14171 (PC) as the current frame. The @code{tfind tracepoint} command with
14172 no argument selects the next trace snapshot collected by the same
14173 tracepoint as the current one.
14175 In addition to letting you scan through the trace buffer manually,
14176 these commands make it easy to construct @value{GDBN} scripts that
14177 scan through the trace buffer and print out whatever collected data
14178 you are interested in. Thus, if we want to examine the PC, FP, and SP
14179 registers from each trace frame in the buffer, we can say this:
14182 (@value{GDBP}) @b{tfind start}
14183 (@value{GDBP}) @b{while ($trace_frame != -1)}
14184 > printf "Frame %d, PC = %08X, SP = %08X, FP = %08X\n", \
14185 $trace_frame, $pc, $sp, $fp
14189 Frame 0, PC = 0020DC64, SP = 0030BF3C, FP = 0030BF44
14190 Frame 1, PC = 0020DC6C, SP = 0030BF38, FP = 0030BF44
14191 Frame 2, PC = 0020DC70, SP = 0030BF34, FP = 0030BF44
14192 Frame 3, PC = 0020DC74, SP = 0030BF30, FP = 0030BF44
14193 Frame 4, PC = 0020DC78, SP = 0030BF2C, FP = 0030BF44
14194 Frame 5, PC = 0020DC7C, SP = 0030BF28, FP = 0030BF44
14195 Frame 6, PC = 0020DC80, SP = 0030BF24, FP = 0030BF44
14196 Frame 7, PC = 0020DC84, SP = 0030BF20, FP = 0030BF44
14197 Frame 8, PC = 0020DC88, SP = 0030BF1C, FP = 0030BF44
14198 Frame 9, PC = 0020DC8E, SP = 0030BF18, FP = 0030BF44
14199 Frame 10, PC = 00203F6C, SP = 0030BE3C, FP = 0030BF14
14202 Or, if we want to examine the variable @code{X} at each source line in
14206 (@value{GDBP}) @b{tfind start}
14207 (@value{GDBP}) @b{while ($trace_frame != -1)}
14208 > printf "Frame %d, X == %d\n", $trace_frame, X
14218 @subsection @code{tdump}
14220 @cindex dump all data collected at tracepoint
14221 @cindex tracepoint data, display
14223 This command takes no arguments. It prints all the data collected at
14224 the current trace snapshot.
14227 (@value{GDBP}) @b{trace 444}
14228 (@value{GDBP}) @b{actions}
14229 Enter actions for tracepoint #2, one per line:
14230 > collect $regs, $locals, $args, gdb_long_test
14233 (@value{GDBP}) @b{tstart}
14235 (@value{GDBP}) @b{tfind line 444}
14236 #0 gdb_test (p1=0x11, p2=0x22, p3=0x33, p4=0x44, p5=0x55, p6=0x66)
14238 444 printp( "%s: arguments = 0x%X 0x%X 0x%X 0x%X 0x%X 0x%X\n", )
14240 (@value{GDBP}) @b{tdump}
14241 Data collected at tracepoint 2, trace frame 1:
14242 d0 0xc4aa0085 -995491707
14246 d4 0x71aea3d 119204413
14249 d7 0x380035 3670069
14250 a0 0x19e24a 1696330
14251 a1 0x3000668 50333288
14253 a3 0x322000 3284992
14254 a4 0x3000698 50333336
14255 a5 0x1ad3cc 1758156
14256 fp 0x30bf3c 0x30bf3c
14257 sp 0x30bf34 0x30bf34
14259 pc 0x20b2c8 0x20b2c8
14263 p = 0x20e5b4 "gdb-test"
14270 gdb_long_test = 17 '\021'
14275 @code{tdump} works by scanning the tracepoint's current collection
14276 actions and printing the value of each expression listed. So
14277 @code{tdump} can fail, if after a run, you change the tracepoint's
14278 actions to mention variables that were not collected during the run.
14280 Also, for tracepoints with @code{while-stepping} loops, @code{tdump}
14281 uses the collected value of @code{$pc} to distinguish between trace
14282 frames that were collected at the tracepoint hit, and frames that were
14283 collected while stepping. This allows it to correctly choose whether
14284 to display the basic list of collections, or the collections from the
14285 body of the while-stepping loop. However, if @code{$pc} was not collected,
14286 then @code{tdump} will always attempt to dump using the basic collection
14287 list, and may fail if a while-stepping frame does not include all the
14288 same data that is collected at the tracepoint hit.
14289 @c This is getting pretty arcane, example would be good.
14291 @node save tracepoints
14292 @subsection @code{save tracepoints @var{filename}}
14293 @kindex save tracepoints
14294 @kindex save-tracepoints
14295 @cindex save tracepoints for future sessions
14297 This command saves all current tracepoint definitions together with
14298 their actions and passcounts, into a file @file{@var{filename}}
14299 suitable for use in a later debugging session. To read the saved
14300 tracepoint definitions, use the @code{source} command (@pxref{Command
14301 Files}). The @w{@code{save-tracepoints}} command is a deprecated
14302 alias for @w{@code{save tracepoints}}
14304 @node Tracepoint Variables
14305 @section Convenience Variables for Tracepoints
14306 @cindex tracepoint variables
14307 @cindex convenience variables for tracepoints
14310 @vindex $trace_frame
14311 @item (int) $trace_frame
14312 The current trace snapshot (a.k.a.@: @dfn{frame}) number, or -1 if no
14313 snapshot is selected.
14315 @vindex $tracepoint
14316 @item (int) $tracepoint
14317 The tracepoint for the current trace snapshot.
14319 @vindex $trace_line
14320 @item (int) $trace_line
14321 The line number for the current trace snapshot.
14323 @vindex $trace_file
14324 @item (char []) $trace_file
14325 The source file for the current trace snapshot.
14327 @vindex $trace_func
14328 @item (char []) $trace_func
14329 The name of the function containing @code{$tracepoint}.
14332 Note: @code{$trace_file} is not suitable for use in @code{printf},
14333 use @code{output} instead.
14335 Here's a simple example of using these convenience variables for
14336 stepping through all the trace snapshots and printing some of their
14337 data. Note that these are not the same as trace state variables,
14338 which are managed by the target.
14341 (@value{GDBP}) @b{tfind start}
14343 (@value{GDBP}) @b{while $trace_frame != -1}
14344 > output $trace_file
14345 > printf ", line %d (tracepoint #%d)\n", $trace_line, $tracepoint
14351 @section Using Trace Files
14352 @cindex trace files
14354 In some situations, the target running a trace experiment may no
14355 longer be available; perhaps it crashed, or the hardware was needed
14356 for a different activity. To handle these cases, you can arrange to
14357 dump the trace data into a file, and later use that file as a source
14358 of trace data, via the @code{target tfile} command.
14363 @item tsave [ -r ] @var{filename}
14364 @itemx tsave [-ctf] @var{dirname}
14365 Save the trace data to @var{filename}. By default, this command
14366 assumes that @var{filename} refers to the host filesystem, so if
14367 necessary @value{GDBN} will copy raw trace data up from the target and
14368 then save it. If the target supports it, you can also supply the
14369 optional argument @code{-r} (``remote'') to direct the target to save
14370 the data directly into @var{filename} in its own filesystem, which may be
14371 more efficient if the trace buffer is very large. (Note, however, that
14372 @code{target tfile} can only read from files accessible to the host.)
14373 By default, this command will save trace frame in tfile format.
14374 You can supply the optional argument @code{-ctf} to save data in CTF
14375 format. The @dfn{Common Trace Format} (CTF) is proposed as a trace format
14376 that can be shared by multiple debugging and tracing tools. Please go to
14377 @indicateurl{http://www.efficios.com/ctf} to get more information.
14379 @kindex target tfile
14383 @item target tfile @var{filename}
14384 @itemx target ctf @var{dirname}
14385 Use the file named @var{filename} or directory named @var{dirname} as
14386 a source of trace data. Commands that examine data work as they do with
14387 a live target, but it is not possible to run any new trace experiments.
14388 @code{tstatus} will report the state of the trace run at the moment
14389 the data was saved, as well as the current trace frame you are examining.
14390 Both @var{filename} and @var{dirname} must be on a filesystem accessible to
14394 (@value{GDBP}) target ctf ctf.ctf
14395 (@value{GDBP}) tfind
14396 Found trace frame 0, tracepoint 2
14397 39 ++a; /* set tracepoint 1 here */
14398 (@value{GDBP}) tdump
14399 Data collected at tracepoint 2, trace frame 0:
14403 c = @{"123", "456", "789", "123", "456", "789"@}
14404 d = @{@{@{a = 1, b = 2@}, @{a = 3, b = 4@}@}, @{@{a = 5, b = 6@}, @{a = 7, b = 8@}@}@}
14412 @chapter Debugging Programs That Use Overlays
14415 If your program is too large to fit completely in your target system's
14416 memory, you can sometimes use @dfn{overlays} to work around this
14417 problem. @value{GDBN} provides some support for debugging programs that
14421 * How Overlays Work:: A general explanation of overlays.
14422 * Overlay Commands:: Managing overlays in @value{GDBN}.
14423 * Automatic Overlay Debugging:: @value{GDBN} can find out which overlays are
14424 mapped by asking the inferior.
14425 * Overlay Sample Program:: A sample program using overlays.
14428 @node How Overlays Work
14429 @section How Overlays Work
14430 @cindex mapped overlays
14431 @cindex unmapped overlays
14432 @cindex load address, overlay's
14433 @cindex mapped address
14434 @cindex overlay area
14436 Suppose you have a computer whose instruction address space is only 64
14437 kilobytes long, but which has much more memory which can be accessed by
14438 other means: special instructions, segment registers, or memory
14439 management hardware, for example. Suppose further that you want to
14440 adapt a program which is larger than 64 kilobytes to run on this system.
14442 One solution is to identify modules of your program which are relatively
14443 independent, and need not call each other directly; call these modules
14444 @dfn{overlays}. Separate the overlays from the main program, and place
14445 their machine code in the larger memory. Place your main program in
14446 instruction memory, but leave at least enough space there to hold the
14447 largest overlay as well.
14449 Now, to call a function located in an overlay, you must first copy that
14450 overlay's machine code from the large memory into the space set aside
14451 for it in the instruction memory, and then jump to its entry point
14454 @c NB: In the below the mapped area's size is greater or equal to the
14455 @c size of all overlays. This is intentional to remind the developer
14456 @c that overlays don't necessarily need to be the same size.
14460 Data Instruction Larger
14461 Address Space Address Space Address Space
14462 +-----------+ +-----------+ +-----------+
14464 +-----------+ +-----------+ +-----------+<-- overlay 1
14465 | program | | main | .----| overlay 1 | load address
14466 | variables | | program | | +-----------+
14467 | and heap | | | | | |
14468 +-----------+ | | | +-----------+<-- overlay 2
14469 | | +-----------+ | | | load address
14470 +-----------+ | | | .-| overlay 2 |
14472 mapped --->+-----------+ | | +-----------+
14473 address | | | | | |
14474 | overlay | <-' | | |
14475 | area | <---' +-----------+<-- overlay 3
14476 | | <---. | | load address
14477 +-----------+ `--| overlay 3 |
14484 @anchor{A code overlay}A code overlay
14488 The diagram (@pxref{A code overlay}) shows a system with separate data
14489 and instruction address spaces. To map an overlay, the program copies
14490 its code from the larger address space to the instruction address space.
14491 Since the overlays shown here all use the same mapped address, only one
14492 may be mapped at a time. For a system with a single address space for
14493 data and instructions, the diagram would be similar, except that the
14494 program variables and heap would share an address space with the main
14495 program and the overlay area.
14497 An overlay loaded into instruction memory and ready for use is called a
14498 @dfn{mapped} overlay; its @dfn{mapped address} is its address in the
14499 instruction memory. An overlay not present (or only partially present)
14500 in instruction memory is called @dfn{unmapped}; its @dfn{load address}
14501 is its address in the larger memory. The mapped address is also called
14502 the @dfn{virtual memory address}, or @dfn{VMA}; the load address is also
14503 called the @dfn{load memory address}, or @dfn{LMA}.
14505 Unfortunately, overlays are not a completely transparent way to adapt a
14506 program to limited instruction memory. They introduce a new set of
14507 global constraints you must keep in mind as you design your program:
14512 Before calling or returning to a function in an overlay, your program
14513 must make sure that overlay is actually mapped. Otherwise, the call or
14514 return will transfer control to the right address, but in the wrong
14515 overlay, and your program will probably crash.
14518 If the process of mapping an overlay is expensive on your system, you
14519 will need to choose your overlays carefully to minimize their effect on
14520 your program's performance.
14523 The executable file you load onto your system must contain each
14524 overlay's instructions, appearing at the overlay's load address, not its
14525 mapped address. However, each overlay's instructions must be relocated
14526 and its symbols defined as if the overlay were at its mapped address.
14527 You can use GNU linker scripts to specify different load and relocation
14528 addresses for pieces of your program; see @ref{Overlay Description,,,
14529 ld.info, Using ld: the GNU linker}.
14532 The procedure for loading executable files onto your system must be able
14533 to load their contents into the larger address space as well as the
14534 instruction and data spaces.
14538 The overlay system described above is rather simple, and could be
14539 improved in many ways:
14544 If your system has suitable bank switch registers or memory management
14545 hardware, you could use those facilities to make an overlay's load area
14546 contents simply appear at their mapped address in instruction space.
14547 This would probably be faster than copying the overlay to its mapped
14548 area in the usual way.
14551 If your overlays are small enough, you could set aside more than one
14552 overlay area, and have more than one overlay mapped at a time.
14555 You can use overlays to manage data, as well as instructions. In
14556 general, data overlays are even less transparent to your design than
14557 code overlays: whereas code overlays only require care when you call or
14558 return to functions, data overlays require care every time you access
14559 the data. Also, if you change the contents of a data overlay, you
14560 must copy its contents back out to its load address before you can copy a
14561 different data overlay into the same mapped area.
14566 @node Overlay Commands
14567 @section Overlay Commands
14569 To use @value{GDBN}'s overlay support, each overlay in your program must
14570 correspond to a separate section of the executable file. The section's
14571 virtual memory address and load memory address must be the overlay's
14572 mapped and load addresses. Identifying overlays with sections allows
14573 @value{GDBN} to determine the appropriate address of a function or
14574 variable, depending on whether the overlay is mapped or not.
14576 @value{GDBN}'s overlay commands all start with the word @code{overlay};
14577 you can abbreviate this as @code{ov} or @code{ovly}. The commands are:
14582 Disable @value{GDBN}'s overlay support. When overlay support is
14583 disabled, @value{GDBN} assumes that all functions and variables are
14584 always present at their mapped addresses. By default, @value{GDBN}'s
14585 overlay support is disabled.
14587 @item overlay manual
14588 @cindex manual overlay debugging
14589 Enable @dfn{manual} overlay debugging. In this mode, @value{GDBN}
14590 relies on you to tell it which overlays are mapped, and which are not,
14591 using the @code{overlay map-overlay} and @code{overlay unmap-overlay}
14592 commands described below.
14594 @item overlay map-overlay @var{overlay}
14595 @itemx overlay map @var{overlay}
14596 @cindex map an overlay
14597 Tell @value{GDBN} that @var{overlay} is now mapped; @var{overlay} must
14598 be the name of the object file section containing the overlay. When an
14599 overlay is mapped, @value{GDBN} assumes it can find the overlay's
14600 functions and variables at their mapped addresses. @value{GDBN} assumes
14601 that any other overlays whose mapped ranges overlap that of
14602 @var{overlay} are now unmapped.
14604 @item overlay unmap-overlay @var{overlay}
14605 @itemx overlay unmap @var{overlay}
14606 @cindex unmap an overlay
14607 Tell @value{GDBN} that @var{overlay} is no longer mapped; @var{overlay}
14608 must be the name of the object file section containing the overlay.
14609 When an overlay is unmapped, @value{GDBN} assumes it can find the
14610 overlay's functions and variables at their load addresses.
14613 Enable @dfn{automatic} overlay debugging. In this mode, @value{GDBN}
14614 consults a data structure the overlay manager maintains in the inferior
14615 to see which overlays are mapped. For details, see @ref{Automatic
14616 Overlay Debugging}.
14618 @item overlay load-target
14619 @itemx overlay load
14620 @cindex reloading the overlay table
14621 Re-read the overlay table from the inferior. Normally, @value{GDBN}
14622 re-reads the table @value{GDBN} automatically each time the inferior
14623 stops, so this command should only be necessary if you have changed the
14624 overlay mapping yourself using @value{GDBN}. This command is only
14625 useful when using automatic overlay debugging.
14627 @item overlay list-overlays
14628 @itemx overlay list
14629 @cindex listing mapped overlays
14630 Display a list of the overlays currently mapped, along with their mapped
14631 addresses, load addresses, and sizes.
14635 Normally, when @value{GDBN} prints a code address, it includes the name
14636 of the function the address falls in:
14639 (@value{GDBP}) print main
14640 $3 = @{int ()@} 0x11a0 <main>
14643 When overlay debugging is enabled, @value{GDBN} recognizes code in
14644 unmapped overlays, and prints the names of unmapped functions with
14645 asterisks around them. For example, if @code{foo} is a function in an
14646 unmapped overlay, @value{GDBN} prints it this way:
14649 (@value{GDBP}) overlay list
14650 No sections are mapped.
14651 (@value{GDBP}) print foo
14652 $5 = @{int (int)@} 0x100000 <*foo*>
14655 When @code{foo}'s overlay is mapped, @value{GDBN} prints the function's
14659 (@value{GDBP}) overlay list
14660 Section .ov.foo.text, loaded at 0x100000 - 0x100034,
14661 mapped at 0x1016 - 0x104a
14662 (@value{GDBP}) print foo
14663 $6 = @{int (int)@} 0x1016 <foo>
14666 When overlay debugging is enabled, @value{GDBN} can find the correct
14667 address for functions and variables in an overlay, whether or not the
14668 overlay is mapped. This allows most @value{GDBN} commands, like
14669 @code{break} and @code{disassemble}, to work normally, even on unmapped
14670 code. However, @value{GDBN}'s breakpoint support has some limitations:
14674 @cindex breakpoints in overlays
14675 @cindex overlays, setting breakpoints in
14676 You can set breakpoints in functions in unmapped overlays, as long as
14677 @value{GDBN} can write to the overlay at its load address.
14679 @value{GDBN} can not set hardware or simulator-based breakpoints in
14680 unmapped overlays. However, if you set a breakpoint at the end of your
14681 overlay manager (and tell @value{GDBN} which overlays are now mapped, if
14682 you are using manual overlay management), @value{GDBN} will re-set its
14683 breakpoints properly.
14687 @node Automatic Overlay Debugging
14688 @section Automatic Overlay Debugging
14689 @cindex automatic overlay debugging
14691 @value{GDBN} can automatically track which overlays are mapped and which
14692 are not, given some simple co-operation from the overlay manager in the
14693 inferior. If you enable automatic overlay debugging with the
14694 @code{overlay auto} command (@pxref{Overlay Commands}), @value{GDBN}
14695 looks in the inferior's memory for certain variables describing the
14696 current state of the overlays.
14698 Here are the variables your overlay manager must define to support
14699 @value{GDBN}'s automatic overlay debugging:
14703 @item @code{_ovly_table}:
14704 This variable must be an array of the following structures:
14709 /* The overlay's mapped address. */
14712 /* The size of the overlay, in bytes. */
14713 unsigned long size;
14715 /* The overlay's load address. */
14718 /* Non-zero if the overlay is currently mapped;
14720 unsigned long mapped;
14724 @item @code{_novlys}:
14725 This variable must be a four-byte signed integer, holding the total
14726 number of elements in @code{_ovly_table}.
14730 To decide whether a particular overlay is mapped or not, @value{GDBN}
14731 looks for an entry in @w{@code{_ovly_table}} whose @code{vma} and
14732 @code{lma} members equal the VMA and LMA of the overlay's section in the
14733 executable file. When @value{GDBN} finds a matching entry, it consults
14734 the entry's @code{mapped} member to determine whether the overlay is
14737 In addition, your overlay manager may define a function called
14738 @code{_ovly_debug_event}. If this function is defined, @value{GDBN}
14739 will silently set a breakpoint there. If the overlay manager then
14740 calls this function whenever it has changed the overlay table, this
14741 will enable @value{GDBN} to accurately keep track of which overlays
14742 are in program memory, and update any breakpoints that may be set
14743 in overlays. This will allow breakpoints to work even if the
14744 overlays are kept in ROM or other non-writable memory while they
14745 are not being executed.
14747 @node Overlay Sample Program
14748 @section Overlay Sample Program
14749 @cindex overlay example program
14751 When linking a program which uses overlays, you must place the overlays
14752 at their load addresses, while relocating them to run at their mapped
14753 addresses. To do this, you must write a linker script (@pxref{Overlay
14754 Description,,, ld.info, Using ld: the GNU linker}). Unfortunately,
14755 since linker scripts are specific to a particular host system, target
14756 architecture, and target memory layout, this manual cannot provide
14757 portable sample code demonstrating @value{GDBN}'s overlay support.
14759 However, the @value{GDBN} source distribution does contain an overlaid
14760 program, with linker scripts for a few systems, as part of its test
14761 suite. The program consists of the following files from
14762 @file{gdb/testsuite/gdb.base}:
14766 The main program file.
14768 A simple overlay manager, used by @file{overlays.c}.
14773 Overlay modules, loaded and used by @file{overlays.c}.
14776 Linker scripts for linking the test program on the @code{d10v-elf}
14777 and @code{m32r-elf} targets.
14780 You can build the test program using the @code{d10v-elf} GCC
14781 cross-compiler like this:
14784 $ d10v-elf-gcc -g -c overlays.c
14785 $ d10v-elf-gcc -g -c ovlymgr.c
14786 $ d10v-elf-gcc -g -c foo.c
14787 $ d10v-elf-gcc -g -c bar.c
14788 $ d10v-elf-gcc -g -c baz.c
14789 $ d10v-elf-gcc -g -c grbx.c
14790 $ d10v-elf-gcc -g overlays.o ovlymgr.o foo.o bar.o \
14791 baz.o grbx.o -Wl,-Td10v.ld -o overlays
14794 The build process is identical for any other architecture, except that
14795 you must substitute the appropriate compiler and linker script for the
14796 target system for @code{d10v-elf-gcc} and @code{d10v.ld}.
14800 @chapter Using @value{GDBN} with Different Languages
14803 Although programming languages generally have common aspects, they are
14804 rarely expressed in the same manner. For instance, in ANSI C,
14805 dereferencing a pointer @code{p} is accomplished by @code{*p}, but in
14806 Modula-2, it is accomplished by @code{p^}. Values can also be
14807 represented (and displayed) differently. Hex numbers in C appear as
14808 @samp{0x1ae}, while in Modula-2 they appear as @samp{1AEH}.
14810 @cindex working language
14811 Language-specific information is built into @value{GDBN} for some languages,
14812 allowing you to express operations like the above in your program's
14813 native language, and allowing @value{GDBN} to output values in a manner
14814 consistent with the syntax of your program's native language. The
14815 language you use to build expressions is called the @dfn{working
14819 * Setting:: Switching between source languages
14820 * Show:: Displaying the language
14821 * Checks:: Type and range checks
14822 * Supported Languages:: Supported languages
14823 * Unsupported Languages:: Unsupported languages
14827 @section Switching Between Source Languages
14829 There are two ways to control the working language---either have @value{GDBN}
14830 set it automatically, or select it manually yourself. You can use the
14831 @code{set language} command for either purpose. On startup, @value{GDBN}
14832 defaults to setting the language automatically. The working language is
14833 used to determine how expressions you type are interpreted, how values
14836 In addition to the working language, every source file that
14837 @value{GDBN} knows about has its own working language. For some object
14838 file formats, the compiler might indicate which language a particular
14839 source file is in. However, most of the time @value{GDBN} infers the
14840 language from the name of the file. The language of a source file
14841 controls whether C@t{++} names are demangled---this way @code{backtrace} can
14842 show each frame appropriately for its own language. There is no way to
14843 set the language of a source file from within @value{GDBN}, but you can
14844 set the language associated with a filename extension. @xref{Show, ,
14845 Displaying the Language}.
14847 This is most commonly a problem when you use a program, such
14848 as @code{cfront} or @code{f2c}, that generates C but is written in
14849 another language. In that case, make the
14850 program use @code{#line} directives in its C output; that way
14851 @value{GDBN} will know the correct language of the source code of the original
14852 program, and will display that source code, not the generated C code.
14855 * Filenames:: Filename extensions and languages.
14856 * Manually:: Setting the working language manually
14857 * Automatically:: Having @value{GDBN} infer the source language
14861 @subsection List of Filename Extensions and Languages
14863 If a source file name ends in one of the following extensions, then
14864 @value{GDBN} infers that its language is the one indicated.
14882 C@t{++} source file
14888 Objective-C source file
14892 Fortran source file
14895 Modula-2 source file
14899 Assembler source file. This actually behaves almost like C, but
14900 @value{GDBN} does not skip over function prologues when stepping.
14903 In addition, you may set the language associated with a filename
14904 extension. @xref{Show, , Displaying the Language}.
14907 @subsection Setting the Working Language
14909 If you allow @value{GDBN} to set the language automatically,
14910 expressions are interpreted the same way in your debugging session and
14913 @kindex set language
14914 If you wish, you may set the language manually. To do this, issue the
14915 command @samp{set language @var{lang}}, where @var{lang} is the name of
14916 a language, such as
14917 @code{c} or @code{modula-2}.
14918 For a list of the supported languages, type @samp{set language}.
14920 Setting the language manually prevents @value{GDBN} from updating the working
14921 language automatically. This can lead to confusion if you try
14922 to debug a program when the working language is not the same as the
14923 source language, when an expression is acceptable to both
14924 languages---but means different things. For instance, if the current
14925 source file were written in C, and @value{GDBN} was parsing Modula-2, a
14933 might not have the effect you intended. In C, this means to add
14934 @code{b} and @code{c} and place the result in @code{a}. The result
14935 printed would be the value of @code{a}. In Modula-2, this means to compare
14936 @code{a} to the result of @code{b+c}, yielding a @code{BOOLEAN} value.
14938 @node Automatically
14939 @subsection Having @value{GDBN} Infer the Source Language
14941 To have @value{GDBN} set the working language automatically, use
14942 @samp{set language local} or @samp{set language auto}. @value{GDBN}
14943 then infers the working language. That is, when your program stops in a
14944 frame (usually by encountering a breakpoint), @value{GDBN} sets the
14945 working language to the language recorded for the function in that
14946 frame. If the language for a frame is unknown (that is, if the function
14947 or block corresponding to the frame was defined in a source file that
14948 does not have a recognized extension), the current working language is
14949 not changed, and @value{GDBN} issues a warning.
14951 This may not seem necessary for most programs, which are written
14952 entirely in one source language. However, program modules and libraries
14953 written in one source language can be used by a main program written in
14954 a different source language. Using @samp{set language auto} in this
14955 case frees you from having to set the working language manually.
14958 @section Displaying the Language
14960 The following commands help you find out which language is the
14961 working language, and also what language source files were written in.
14964 @item show language
14965 @anchor{show language}
14966 @kindex show language
14967 Display the current working language. This is the
14968 language you can use with commands such as @code{print} to
14969 build and compute expressions that may involve variables in your program.
14972 @kindex info frame@r{, show the source language}
14973 Display the source language for this frame. This language becomes the
14974 working language if you use an identifier from this frame.
14975 @xref{Frame Info, ,Information about a Frame}, to identify the other
14976 information listed here.
14979 @kindex info source@r{, show the source language}
14980 Display the source language of this source file.
14981 @xref{Symbols, ,Examining the Symbol Table}, to identify the other
14982 information listed here.
14985 In unusual circumstances, you may have source files with extensions
14986 not in the standard list. You can then set the extension associated
14987 with a language explicitly:
14990 @item set extension-language @var{ext} @var{language}
14991 @kindex set extension-language
14992 Tell @value{GDBN} that source files with extension @var{ext} are to be
14993 assumed as written in the source language @var{language}.
14995 @item info extensions
14996 @kindex info extensions
14997 List all the filename extensions and the associated languages.
15001 @section Type and Range Checking
15003 Some languages are designed to guard you against making seemingly common
15004 errors through a series of compile- and run-time checks. These include
15005 checking the type of arguments to functions and operators and making
15006 sure mathematical overflows are caught at run time. Checks such as
15007 these help to ensure a program's correctness once it has been compiled
15008 by eliminating type mismatches and providing active checks for range
15009 errors when your program is running.
15011 By default @value{GDBN} checks for these errors according to the
15012 rules of the current source language. Although @value{GDBN} does not check
15013 the statements in your program, it can check expressions entered directly
15014 into @value{GDBN} for evaluation via the @code{print} command, for example.
15017 * Type Checking:: An overview of type checking
15018 * Range Checking:: An overview of range checking
15021 @cindex type checking
15022 @cindex checks, type
15023 @node Type Checking
15024 @subsection An Overview of Type Checking
15026 Some languages, such as C and C@t{++}, are strongly typed, meaning that the
15027 arguments to operators and functions have to be of the correct type,
15028 otherwise an error occurs. These checks prevent type mismatch
15029 errors from ever causing any run-time problems. For example,
15032 int klass::my_method(char *b) @{ return b ? 1 : 2; @}
15034 (@value{GDBP}) print obj.my_method (0)
15037 (@value{GDBP}) print obj.my_method (0x1234)
15038 Cannot resolve method klass::my_method to any overloaded instance
15041 The second example fails because in C@t{++} the integer constant
15042 @samp{0x1234} is not type-compatible with the pointer parameter type.
15044 For the expressions you use in @value{GDBN} commands, you can tell
15045 @value{GDBN} to not enforce strict type checking or
15046 to treat any mismatches as errors and abandon the expression;
15047 When type checking is disabled, @value{GDBN} successfully evaluates
15048 expressions like the second example above.
15050 Even if type checking is off, there may be other reasons
15051 related to type that prevent @value{GDBN} from evaluating an expression.
15052 For instance, @value{GDBN} does not know how to add an @code{int} and
15053 a @code{struct foo}. These particular type errors have nothing to do
15054 with the language in use and usually arise from expressions which make
15055 little sense to evaluate anyway.
15057 @value{GDBN} provides some additional commands for controlling type checking:
15059 @kindex set check type
15060 @kindex show check type
15062 @item set check type on
15063 @itemx set check type off
15064 Set strict type checking on or off. If any type mismatches occur in
15065 evaluating an expression while type checking is on, @value{GDBN} prints a
15066 message and aborts evaluation of the expression.
15068 @item show check type
15069 Show the current setting of type checking and whether @value{GDBN}
15070 is enforcing strict type checking rules.
15073 @cindex range checking
15074 @cindex checks, range
15075 @node Range Checking
15076 @subsection An Overview of Range Checking
15078 In some languages (such as Modula-2), it is an error to exceed the
15079 bounds of a type; this is enforced with run-time checks. Such range
15080 checking is meant to ensure program correctness by making sure
15081 computations do not overflow, or indices on an array element access do
15082 not exceed the bounds of the array.
15084 For expressions you use in @value{GDBN} commands, you can tell
15085 @value{GDBN} to treat range errors in one of three ways: ignore them,
15086 always treat them as errors and abandon the expression, or issue
15087 warnings but evaluate the expression anyway.
15089 A range error can result from numerical overflow, from exceeding an
15090 array index bound, or when you type a constant that is not a member
15091 of any type. Some languages, however, do not treat overflows as an
15092 error. In many implementations of C, mathematical overflow causes the
15093 result to ``wrap around'' to lower values---for example, if @var{m} is
15094 the largest integer value, and @var{s} is the smallest, then
15097 @var{m} + 1 @result{} @var{s}
15100 This, too, is specific to individual languages, and in some cases
15101 specific to individual compilers or machines. @xref{Supported Languages, ,
15102 Supported Languages}, for further details on specific languages.
15104 @value{GDBN} provides some additional commands for controlling the range checker:
15106 @kindex set check range
15107 @kindex show check range
15109 @item set check range auto
15110 Set range checking on or off based on the current working language.
15111 @xref{Supported Languages, ,Supported Languages}, for the default settings for
15114 @item set check range on
15115 @itemx set check range off
15116 Set range checking on or off, overriding the default setting for the
15117 current working language. A warning is issued if the setting does not
15118 match the language default. If a range error occurs and range checking is on,
15119 then a message is printed and evaluation of the expression is aborted.
15121 @item set check range warn
15122 Output messages when the @value{GDBN} range checker detects a range error,
15123 but attempt to evaluate the expression anyway. Evaluating the
15124 expression may still be impossible for other reasons, such as accessing
15125 memory that the process does not own (a typical example from many Unix
15129 Show the current setting of the range checker, and whether or not it is
15130 being set automatically by @value{GDBN}.
15133 @node Supported Languages
15134 @section Supported Languages
15136 @value{GDBN} supports C, C@t{++}, D, Go, Objective-C, Fortran,
15137 OpenCL C, Pascal, Rust, assembly, Modula-2, and Ada.
15138 @c This is false ...
15139 Some @value{GDBN} features may be used in expressions regardless of the
15140 language you use: the @value{GDBN} @code{@@} and @code{::} operators,
15141 and the @samp{@{type@}addr} construct (@pxref{Expressions,
15142 ,Expressions}) can be used with the constructs of any supported
15145 The following sections detail to what degree each source language is
15146 supported by @value{GDBN}. These sections are not meant to be language
15147 tutorials or references, but serve only as a reference guide to what the
15148 @value{GDBN} expression parser accepts, and what input and output
15149 formats should look like for different languages. There are many good
15150 books written on each of these languages; please look to these for a
15151 language reference or tutorial.
15154 * C:: C and C@t{++}
15157 * Objective-C:: Objective-C
15158 * OpenCL C:: OpenCL C
15159 * Fortran:: Fortran
15162 * Modula-2:: Modula-2
15167 @subsection C and C@t{++}
15169 @cindex C and C@t{++}
15170 @cindex expressions in C or C@t{++}
15172 Since C and C@t{++} are so closely related, many features of @value{GDBN} apply
15173 to both languages. Whenever this is the case, we discuss those languages
15177 @cindex @code{g++}, @sc{gnu} C@t{++} compiler
15178 @cindex @sc{gnu} C@t{++}
15179 The C@t{++} debugging facilities are jointly implemented by the C@t{++}
15180 compiler and @value{GDBN}. Therefore, to debug your C@t{++} code
15181 effectively, you must compile your C@t{++} programs with a supported
15182 C@t{++} compiler, such as @sc{gnu} @code{g++}, or the HP ANSI C@t{++}
15183 compiler (@code{aCC}).
15186 * C Operators:: C and C@t{++} operators
15187 * C Constants:: C and C@t{++} constants
15188 * C Plus Plus Expressions:: C@t{++} expressions
15189 * C Defaults:: Default settings for C and C@t{++}
15190 * C Checks:: C and C@t{++} type and range checks
15191 * Debugging C:: @value{GDBN} and C
15192 * Debugging C Plus Plus:: @value{GDBN} features for C@t{++}
15193 * Decimal Floating Point:: Numbers in Decimal Floating Point format
15197 @subsubsection C and C@t{++} Operators
15199 @cindex C and C@t{++} operators
15201 Operators must be defined on values of specific types. For instance,
15202 @code{+} is defined on numbers, but not on structures. Operators are
15203 often defined on groups of types.
15205 For the purposes of C and C@t{++}, the following definitions hold:
15210 @emph{Integral types} include @code{int} with any of its storage-class
15211 specifiers; @code{char}; @code{enum}; and, for C@t{++}, @code{bool}.
15214 @emph{Floating-point types} include @code{float}, @code{double}, and
15215 @code{long double} (if supported by the target platform).
15218 @emph{Pointer types} include all types defined as @code{(@var{type} *)}.
15221 @emph{Scalar types} include all of the above.
15226 The following operators are supported. They are listed here
15227 in order of increasing precedence:
15231 The comma or sequencing operator. Expressions in a comma-separated list
15232 are evaluated from left to right, with the result of the entire
15233 expression being the last expression evaluated.
15236 Assignment. The value of an assignment expression is the value
15237 assigned. Defined on scalar types.
15240 Used in an expression of the form @w{@code{@var{a} @var{op}= @var{b}}},
15241 and translated to @w{@code{@var{a} = @var{a op b}}}.
15242 @w{@code{@var{op}=}} and @code{=} have the same precedence. The operator
15243 @var{op} is any one of the operators @code{|}, @code{^}, @code{&},
15244 @code{<<}, @code{>>}, @code{+}, @code{-}, @code{*}, @code{/}, @code{%}.
15247 The ternary operator. @code{@var{a} ? @var{b} : @var{c}} can be thought
15248 of as: if @var{a} then @var{b} else @var{c}. The argument @var{a}
15249 should be of an integral type.
15252 Logical @sc{or}. Defined on integral types.
15255 Logical @sc{and}. Defined on integral types.
15258 Bitwise @sc{or}. Defined on integral types.
15261 Bitwise exclusive-@sc{or}. Defined on integral types.
15264 Bitwise @sc{and}. Defined on integral types.
15267 Equality and inequality. Defined on scalar types. The value of these
15268 expressions is 0 for false and non-zero for true.
15270 @item <@r{, }>@r{, }<=@r{, }>=
15271 Less than, greater than, less than or equal, greater than or equal.
15272 Defined on scalar types. The value of these expressions is 0 for false
15273 and non-zero for true.
15276 left shift, and right shift. Defined on integral types.
15279 The @value{GDBN} ``artificial array'' operator (@pxref{Expressions, ,Expressions}).
15282 Addition and subtraction. Defined on integral types, floating-point types and
15285 @item *@r{, }/@r{, }%
15286 Multiplication, division, and modulus. Multiplication and division are
15287 defined on integral and floating-point types. Modulus is defined on
15291 Increment and decrement. When appearing before a variable, the
15292 operation is performed before the variable is used in an expression;
15293 when appearing after it, the variable's value is used before the
15294 operation takes place.
15297 Pointer dereferencing. Defined on pointer types. Same precedence as
15301 Address operator. Defined on variables. Same precedence as @code{++}.
15303 For debugging C@t{++}, @value{GDBN} implements a use of @samp{&} beyond what is
15304 allowed in the C@t{++} language itself: you can use @samp{&(&@var{ref})}
15305 to examine the address
15306 where a C@t{++} reference variable (declared with @samp{&@var{ref}}) is
15310 Negative. Defined on integral and floating-point types. Same
15311 precedence as @code{++}.
15314 Logical negation. Defined on integral types. Same precedence as
15318 Bitwise complement operator. Defined on integral types. Same precedence as
15323 Structure member, and pointer-to-structure member. For convenience,
15324 @value{GDBN} regards the two as equivalent, choosing whether to dereference a
15325 pointer based on the stored type information.
15326 Defined on @code{struct} and @code{union} data.
15329 Dereferences of pointers to members.
15332 Array indexing. @code{@var{a}[@var{i}]} is defined as
15333 @code{*(@var{a}+@var{i})}. Same precedence as @code{->}.
15336 Function parameter list. Same precedence as @code{->}.
15339 C@t{++} scope resolution operator. Defined on @code{struct}, @code{union},
15340 and @code{class} types.
15343 Doubled colons also represent the @value{GDBN} scope operator
15344 (@pxref{Expressions, ,Expressions}). Same precedence as @code{::},
15348 If an operator is redefined in the user code, @value{GDBN} usually
15349 attempts to invoke the redefined version instead of using the operator's
15350 predefined meaning.
15353 @subsubsection C and C@t{++} Constants
15355 @cindex C and C@t{++} constants
15357 @value{GDBN} allows you to express the constants of C and C@t{++} in the
15362 Integer constants are a sequence of digits. Octal constants are
15363 specified by a leading @samp{0} (i.e.@: zero), and hexadecimal constants
15364 by a leading @samp{0x} or @samp{0X}. Constants may also end with a letter
15365 @samp{l}, specifying that the constant should be treated as a
15369 Floating point constants are a sequence of digits, followed by a decimal
15370 point, followed by a sequence of digits, and optionally followed by an
15371 exponent. An exponent is of the form:
15372 @samp{@w{e@r{[[}+@r{]|}-@r{]}@var{nnn}}}, where @var{nnn} is another
15373 sequence of digits. The @samp{+} is optional for positive exponents.
15374 A floating-point constant may also end with a letter @samp{f} or
15375 @samp{F}, specifying that the constant should be treated as being of
15376 the @code{float} (as opposed to the default @code{double}) type; or with
15377 a letter @samp{l} or @samp{L}, which specifies a @code{long double}
15381 Enumerated constants consist of enumerated identifiers, or their
15382 integral equivalents.
15385 Character constants are a single character surrounded by single quotes
15386 (@code{'}), or a number---the ordinal value of the corresponding character
15387 (usually its @sc{ascii} value). Within quotes, the single character may
15388 be represented by a letter or by @dfn{escape sequences}, which are of
15389 the form @samp{\@var{nnn}}, where @var{nnn} is the octal representation
15390 of the character's ordinal value; or of the form @samp{\@var{x}}, where
15391 @samp{@var{x}} is a predefined special character---for example,
15392 @samp{\n} for newline.
15394 Wide character constants can be written by prefixing a character
15395 constant with @samp{L}, as in C. For example, @samp{L'x'} is the wide
15396 form of @samp{x}. The target wide character set is used when
15397 computing the value of this constant (@pxref{Character Sets}).
15400 String constants are a sequence of character constants surrounded by
15401 double quotes (@code{"}). Any valid character constant (as described
15402 above) may appear. Double quotes within the string must be preceded by
15403 a backslash, so for instance @samp{"a\"b'c"} is a string of five
15406 Wide string constants can be written by prefixing a string constant
15407 with @samp{L}, as in C. The target wide character set is used when
15408 computing the value of this constant (@pxref{Character Sets}).
15411 Pointer constants are an integral value. You can also write pointers
15412 to constants using the C operator @samp{&}.
15415 Array constants are comma-separated lists surrounded by braces @samp{@{}
15416 and @samp{@}}; for example, @samp{@{1,2,3@}} is a three-element array of
15417 integers, @samp{@{@{1,2@}, @{3,4@}, @{5,6@}@}} is a three-by-two array,
15418 and @samp{@{&"hi", &"there", &"fred"@}} is a three-element array of pointers.
15421 @node C Plus Plus Expressions
15422 @subsubsection C@t{++} Expressions
15424 @cindex expressions in C@t{++}
15425 @value{GDBN} expression handling can interpret most C@t{++} expressions.
15427 @cindex debugging C@t{++} programs
15428 @cindex C@t{++} compilers
15429 @cindex debug formats and C@t{++}
15430 @cindex @value{NGCC} and C@t{++}
15432 @emph{Warning:} @value{GDBN} can only debug C@t{++} code if you use
15433 the proper compiler and the proper debug format. Currently,
15434 @value{GDBN} works best when debugging C@t{++} code that is compiled
15435 with the most recent version of @value{NGCC} possible. The DWARF
15436 debugging format is preferred; @value{NGCC} defaults to this on most
15437 popular platforms. Other compilers and/or debug formats are likely to
15438 work badly or not at all when using @value{GDBN} to debug C@t{++}
15439 code. @xref{Compilation}.
15444 @cindex member functions
15446 Member function calls are allowed; you can use expressions like
15449 count = aml->GetOriginal(x, y)
15452 @vindex this@r{, inside C@t{++} member functions}
15453 @cindex namespace in C@t{++}
15455 While a member function is active (in the selected stack frame), your
15456 expressions have the same namespace available as the member function;
15457 that is, @value{GDBN} allows implicit references to the class instance
15458 pointer @code{this} following the same rules as C@t{++}. @code{using}
15459 declarations in the current scope are also respected by @value{GDBN}.
15461 @cindex call overloaded functions
15462 @cindex overloaded functions, calling
15463 @cindex type conversions in C@t{++}
15465 You can call overloaded functions; @value{GDBN} resolves the function
15466 call to the right definition, with some restrictions. @value{GDBN} does not
15467 perform overload resolution involving user-defined type conversions,
15468 calls to constructors, or instantiations of templates that do not exist
15469 in the program. It also cannot handle ellipsis argument lists or
15472 It does perform integral conversions and promotions, floating-point
15473 promotions, arithmetic conversions, pointer conversions, conversions of
15474 class objects to base classes, and standard conversions such as those of
15475 functions or arrays to pointers; it requires an exact match on the
15476 number of function arguments.
15478 Overload resolution is always performed, unless you have specified
15479 @code{set overload-resolution off}. @xref{Debugging C Plus Plus,
15480 ,@value{GDBN} Features for C@t{++}}.
15482 You must specify @code{set overload-resolution off} in order to use an
15483 explicit function signature to call an overloaded function, as in
15485 p 'foo(char,int)'('x', 13)
15488 The @value{GDBN} command-completion facility can simplify this;
15489 see @ref{Completion, ,Command Completion}.
15491 @cindex reference declarations
15493 @value{GDBN} understands variables declared as C@t{++} lvalue or rvalue
15494 references; you can use them in expressions just as you do in C@t{++}
15495 source---they are automatically dereferenced.
15497 In the parameter list shown when @value{GDBN} displays a frame, the values of
15498 reference variables are not displayed (unlike other variables); this
15499 avoids clutter, since references are often used for large structures.
15500 The @emph{address} of a reference variable is always shown, unless
15501 you have specified @samp{set print address off}.
15504 @value{GDBN} supports the C@t{++} name resolution operator @code{::}---your
15505 expressions can use it just as expressions in your program do. Since
15506 one scope may be defined in another, you can use @code{::} repeatedly if
15507 necessary, for example in an expression like
15508 @samp{@var{scope1}::@var{scope2}::@var{name}}. @value{GDBN} also allows
15509 resolving name scope by reference to source files, in both C and C@t{++}
15510 debugging (@pxref{Variables, ,Program Variables}).
15513 @value{GDBN} performs argument-dependent lookup, following the C@t{++}
15518 @subsubsection C and C@t{++} Defaults
15520 @cindex C and C@t{++} defaults
15522 If you allow @value{GDBN} to set range checking automatically, it
15523 defaults to @code{off} whenever the working language changes to
15524 C or C@t{++}. This happens regardless of whether you or @value{GDBN}
15525 selects the working language.
15527 If you allow @value{GDBN} to set the language automatically, it
15528 recognizes source files whose names end with @file{.c}, @file{.C}, or
15529 @file{.cc}, etc, and when @value{GDBN} enters code compiled from one of
15530 these files, it sets the working language to C or C@t{++}.
15531 @xref{Automatically, ,Having @value{GDBN} Infer the Source Language},
15532 for further details.
15535 @subsubsection C and C@t{++} Type and Range Checks
15537 @cindex C and C@t{++} checks
15539 By default, when @value{GDBN} parses C or C@t{++} expressions, strict type
15540 checking is used. However, if you turn type checking off, @value{GDBN}
15541 will allow certain non-standard conversions, such as promoting integer
15542 constants to pointers.
15544 Range checking, if turned on, is done on mathematical operations. Array
15545 indices are not checked, since they are often used to index a pointer
15546 that is not itself an array.
15549 @subsubsection @value{GDBN} and C
15551 The @code{set print union} and @code{show print union} commands apply to
15552 the @code{union} type. When set to @samp{on}, any @code{union} that is
15553 inside a @code{struct} or @code{class} is also printed. Otherwise, it
15554 appears as @samp{@{...@}}.
15556 The @code{@@} operator aids in the debugging of dynamic arrays, formed
15557 with pointers and a memory allocation function. @xref{Expressions,
15560 @node Debugging C Plus Plus
15561 @subsubsection @value{GDBN} Features for C@t{++}
15563 @cindex commands for C@t{++}
15565 Some @value{GDBN} commands are particularly useful with C@t{++}, and some are
15566 designed specifically for use with C@t{++}. Here is a summary:
15569 @cindex break in overloaded functions
15570 @item @r{breakpoint menus}
15571 When you want a breakpoint in a function whose name is overloaded,
15572 @value{GDBN} has the capability to display a menu of possible breakpoint
15573 locations to help you specify which function definition you want.
15574 @xref{Ambiguous Expressions,,Ambiguous Expressions}.
15576 @cindex overloading in C@t{++}
15577 @item rbreak @var{regex}
15578 Setting breakpoints using regular expressions is helpful for setting
15579 breakpoints on overloaded functions that are not members of any special
15581 @xref{Set Breaks, ,Setting Breakpoints}.
15583 @cindex C@t{++} exception handling
15585 @itemx catch rethrow
15587 Debug C@t{++} exception handling using these commands. @xref{Set
15588 Catchpoints, , Setting Catchpoints}.
15590 @cindex inheritance
15591 @item ptype @var{typename}
15592 Print inheritance relationships as well as other information for type
15594 @xref{Symbols, ,Examining the Symbol Table}.
15596 @item info vtbl @var{expression}.
15597 The @code{info vtbl} command can be used to display the virtual
15598 method tables of the object computed by @var{expression}. This shows
15599 one entry per virtual table; there may be multiple virtual tables when
15600 multiple inheritance is in use.
15602 @cindex C@t{++} demangling
15603 @item demangle @var{name}
15604 Demangle @var{name}.
15605 @xref{Symbols}, for a more complete description of the @code{demangle} command.
15607 @cindex C@t{++} symbol display
15608 @item set print demangle
15609 @itemx show print demangle
15610 @itemx set print asm-demangle
15611 @itemx show print asm-demangle
15612 Control whether C@t{++} symbols display in their source form, both when
15613 displaying code as C@t{++} source and when displaying disassemblies.
15614 @xref{Print Settings, ,Print Settings}.
15616 @item set print object
15617 @itemx show print object
15618 Choose whether to print derived (actual) or declared types of objects.
15619 @xref{Print Settings, ,Print Settings}.
15621 @item set print vtbl
15622 @itemx show print vtbl
15623 Control the format for printing virtual function tables.
15624 @xref{Print Settings, ,Print Settings}.
15625 (The @code{vtbl} commands do not work on programs compiled with the HP
15626 ANSI C@t{++} compiler (@code{aCC}).)
15628 @kindex set overload-resolution
15629 @cindex overloaded functions, overload resolution
15630 @item set overload-resolution on
15631 Enable overload resolution for C@t{++} expression evaluation. The default
15632 is on. For overloaded functions, @value{GDBN} evaluates the arguments
15633 and searches for a function whose signature matches the argument types,
15634 using the standard C@t{++} conversion rules (see @ref{C Plus Plus
15635 Expressions, ,C@t{++} Expressions}, for details).
15636 If it cannot find a match, it emits a message.
15638 @item set overload-resolution off
15639 Disable overload resolution for C@t{++} expression evaluation. For
15640 overloaded functions that are not class member functions, @value{GDBN}
15641 chooses the first function of the specified name that it finds in the
15642 symbol table, whether or not its arguments are of the correct type. For
15643 overloaded functions that are class member functions, @value{GDBN}
15644 searches for a function whose signature @emph{exactly} matches the
15647 @kindex show overload-resolution
15648 @item show overload-resolution
15649 Show the current setting of overload resolution.
15651 @item @r{Overloaded symbol names}
15652 You can specify a particular definition of an overloaded symbol, using
15653 the same notation that is used to declare such symbols in C@t{++}: type
15654 @code{@var{symbol}(@var{types})} rather than just @var{symbol}. You can
15655 also use the @value{GDBN} command-line word completion facilities to list the
15656 available choices, or to finish the type list for you.
15657 @xref{Completion,, Command Completion}, for details on how to do this.
15659 @item @r{Breakpoints in functions with ABI tags}
15661 The GNU C@t{++} compiler introduced the notion of ABI ``tags'', which
15662 correspond to changes in the ABI of a type, function, or variable that
15663 would not otherwise be reflected in a mangled name. See
15664 @url{https://developers.redhat.com/blog/2015/02/05/gcc5-and-the-c11-abi/}
15667 The ABI tags are visible in C@t{++} demangled names. For example, a
15668 function that returns a std::string:
15671 std::string function(int);
15675 when compiled for the C++11 ABI is marked with the @code{cxx11} ABI
15676 tag, and @value{GDBN} displays the symbol like this:
15679 function[abi:cxx11](int)
15682 You can set a breakpoint on such functions simply as if they had no
15686 (gdb) b function(int)
15687 Breakpoint 2 at 0x40060d: file main.cc, line 10.
15688 (gdb) info breakpoints
15689 Num Type Disp Enb Address What
15690 1 breakpoint keep y 0x0040060d in function[abi:cxx11](int)
15694 On the rare occasion you need to disambiguate between different ABI
15695 tags, you can do so by simply including the ABI tag in the function
15699 (@value{GDBP}) b ambiguous[abi:other_tag](int)
15703 @node Decimal Floating Point
15704 @subsubsection Decimal Floating Point format
15705 @cindex decimal floating point format
15707 @value{GDBN} can examine, set and perform computations with numbers in
15708 decimal floating point format, which in the C language correspond to the
15709 @code{_Decimal32}, @code{_Decimal64} and @code{_Decimal128} types as
15710 specified by the extension to support decimal floating-point arithmetic.
15712 There are two encodings in use, depending on the architecture: BID (Binary
15713 Integer Decimal) for x86 and x86-64, and DPD (Densely Packed Decimal) for
15714 PowerPC and S/390. @value{GDBN} will use the appropriate encoding for the
15717 Because of a limitation in @file{libdecnumber}, the library used by @value{GDBN}
15718 to manipulate decimal floating point numbers, it is not possible to convert
15719 (using a cast, for example) integers wider than 32-bit to decimal float.
15721 In addition, in order to imitate @value{GDBN}'s behaviour with binary floating
15722 point computations, error checking in decimal float operations ignores
15723 underflow, overflow and divide by zero exceptions.
15725 In the PowerPC architecture, @value{GDBN} provides a set of pseudo-registers
15726 to inspect @code{_Decimal128} values stored in floating point registers.
15727 See @ref{PowerPC,,PowerPC} for more details.
15733 @value{GDBN} can be used to debug programs written in D and compiled with
15734 GDC, LDC or DMD compilers. Currently @value{GDBN} supports only one D
15735 specific feature --- dynamic arrays.
15740 @cindex Go (programming language)
15741 @value{GDBN} can be used to debug programs written in Go and compiled with
15742 @file{gccgo} or @file{6g} compilers.
15744 Here is a summary of the Go-specific features and restrictions:
15747 @cindex current Go package
15748 @item The current Go package
15749 The name of the current package does not need to be specified when
15750 specifying global variables and functions.
15752 For example, given the program:
15756 var myglob = "Shall we?"
15762 When stopped inside @code{main} either of these work:
15766 (gdb) p main.myglob
15769 @cindex builtin Go types
15770 @item Builtin Go types
15771 The @code{string} type is recognized by @value{GDBN} and is printed
15774 @cindex builtin Go functions
15775 @item Builtin Go functions
15776 The @value{GDBN} expression parser recognizes the @code{unsafe.Sizeof}
15777 function and handles it internally.
15779 @cindex restrictions on Go expressions
15780 @item Restrictions on Go expressions
15781 All Go operators are supported except @code{&^}.
15782 The Go @code{_} ``blank identifier'' is not supported.
15783 Automatic dereferencing of pointers is not supported.
15787 @subsection Objective-C
15789 @cindex Objective-C
15790 This section provides information about some commands and command
15791 options that are useful for debugging Objective-C code. See also
15792 @ref{Symbols, info classes}, and @ref{Symbols, info selectors}, for a
15793 few more commands specific to Objective-C support.
15796 * Method Names in Commands::
15797 * The Print Command with Objective-C::
15800 @node Method Names in Commands
15801 @subsubsection Method Names in Commands
15803 The following commands have been extended to accept Objective-C method
15804 names as line specifications:
15806 @kindex clear@r{, and Objective-C}
15807 @kindex break@r{, and Objective-C}
15808 @kindex info line@r{, and Objective-C}
15809 @kindex jump@r{, and Objective-C}
15810 @kindex list@r{, and Objective-C}
15814 @item @code{info line}
15819 A fully qualified Objective-C method name is specified as
15822 -[@var{Class} @var{methodName}]
15825 where the minus sign is used to indicate an instance method and a
15826 plus sign (not shown) is used to indicate a class method. The class
15827 name @var{Class} and method name @var{methodName} are enclosed in
15828 brackets, similar to the way messages are specified in Objective-C
15829 source code. For example, to set a breakpoint at the @code{create}
15830 instance method of class @code{Fruit} in the program currently being
15834 break -[Fruit create]
15837 To list ten program lines around the @code{initialize} class method,
15841 list +[NSText initialize]
15844 In the current version of @value{GDBN}, the plus or minus sign is
15845 required. In future versions of @value{GDBN}, the plus or minus
15846 sign will be optional, but you can use it to narrow the search. It
15847 is also possible to specify just a method name:
15853 You must specify the complete method name, including any colons. If
15854 your program's source files contain more than one @code{create} method,
15855 you'll be presented with a numbered list of classes that implement that
15856 method. Indicate your choice by number, or type @samp{0} to exit if
15859 As another example, to clear a breakpoint established at the
15860 @code{makeKeyAndOrderFront:} method of the @code{NSWindow} class, enter:
15863 clear -[NSWindow makeKeyAndOrderFront:]
15866 @node The Print Command with Objective-C
15867 @subsubsection The Print Command With Objective-C
15868 @cindex Objective-C, print objects
15869 @kindex print-object
15870 @kindex po @r{(@code{print-object})}
15872 The print command has also been extended to accept methods. For example:
15875 print -[@var{object} hash]
15878 @cindex print an Objective-C object description
15879 @cindex @code{_NSPrintForDebugger}, and printing Objective-C objects
15881 will tell @value{GDBN} to send the @code{hash} message to @var{object}
15882 and print the result. Also, an additional command has been added,
15883 @code{print-object} or @code{po} for short, which is meant to print
15884 the description of an object. However, this command may only work
15885 with certain Objective-C libraries that have a particular hook
15886 function, @code{_NSPrintForDebugger}, defined.
15889 @subsection OpenCL C
15892 This section provides information about @value{GDBN}s OpenCL C support.
15895 * OpenCL C Datatypes::
15896 * OpenCL C Expressions::
15897 * OpenCL C Operators::
15900 @node OpenCL C Datatypes
15901 @subsubsection OpenCL C Datatypes
15903 @cindex OpenCL C Datatypes
15904 @value{GDBN} supports the builtin scalar and vector datatypes specified
15905 by OpenCL 1.1. In addition the half- and double-precision floating point
15906 data types of the @code{cl_khr_fp16} and @code{cl_khr_fp64} OpenCL
15907 extensions are also known to @value{GDBN}.
15909 @node OpenCL C Expressions
15910 @subsubsection OpenCL C Expressions
15912 @cindex OpenCL C Expressions
15913 @value{GDBN} supports accesses to vector components including the access as
15914 lvalue where possible. Since OpenCL C is based on C99 most C expressions
15915 supported by @value{GDBN} can be used as well.
15917 @node OpenCL C Operators
15918 @subsubsection OpenCL C Operators
15920 @cindex OpenCL C Operators
15921 @value{GDBN} supports the operators specified by OpenCL 1.1 for scalar and
15925 @subsection Fortran
15926 @cindex Fortran-specific support in @value{GDBN}
15928 @value{GDBN} can be used to debug programs written in Fortran, but it
15929 currently supports only the features of Fortran 77 language.
15931 @cindex trailing underscore, in Fortran symbols
15932 Some Fortran compilers (@sc{gnu} Fortran 77 and Fortran 95 compilers
15933 among them) append an underscore to the names of variables and
15934 functions. When you debug programs compiled by those compilers, you
15935 will need to refer to variables and functions with a trailing
15939 * Fortran Operators:: Fortran operators and expressions
15940 * Fortran Defaults:: Default settings for Fortran
15941 * Special Fortran Commands:: Special @value{GDBN} commands for Fortran
15944 @node Fortran Operators
15945 @subsubsection Fortran Operators and Expressions
15947 @cindex Fortran operators and expressions
15949 Operators must be defined on values of specific types. For instance,
15950 @code{+} is defined on numbers, but not on characters or other non-
15951 arithmetic types. Operators are often defined on groups of types.
15955 The exponentiation operator. It raises the first operand to the power
15959 The range operator. Normally used in the form of array(low:high) to
15960 represent a section of array.
15963 The access component operator. Normally used to access elements in derived
15964 types. Also suitable for unions. As unions aren't part of regular Fortran,
15965 this can only happen when accessing a register that uses a gdbarch-defined
15969 @node Fortran Defaults
15970 @subsubsection Fortran Defaults
15972 @cindex Fortran Defaults
15974 Fortran symbols are usually case-insensitive, so @value{GDBN} by
15975 default uses case-insensitive matches for Fortran symbols. You can
15976 change that with the @samp{set case-insensitive} command, see
15977 @ref{Symbols}, for the details.
15979 @node Special Fortran Commands
15980 @subsubsection Special Fortran Commands
15982 @cindex Special Fortran commands
15984 @value{GDBN} has some commands to support Fortran-specific features,
15985 such as displaying common blocks.
15988 @cindex @code{COMMON} blocks, Fortran
15989 @kindex info common
15990 @item info common @r{[}@var{common-name}@r{]}
15991 This command prints the values contained in the Fortran @code{COMMON}
15992 block whose name is @var{common-name}. With no argument, the names of
15993 all @code{COMMON} blocks visible at the current program location are
16000 @cindex Pascal support in @value{GDBN}, limitations
16001 Debugging Pascal programs which use sets, subranges, file variables, or
16002 nested functions does not currently work. @value{GDBN} does not support
16003 entering expressions, printing values, or similar features using Pascal
16006 The Pascal-specific command @code{set print pascal_static-members}
16007 controls whether static members of Pascal objects are displayed.
16008 @xref{Print Settings, pascal_static-members}.
16013 @value{GDBN} supports the @url{https://www.rust-lang.org/, Rust
16014 Programming Language}. Type- and value-printing, and expression
16015 parsing, are reasonably complete. However, there are a few
16016 peculiarities and holes to be aware of.
16020 Linespecs (@pxref{Specify Location}) are never relative to the current
16021 crate. Instead, they act as if there were a global namespace of
16022 crates, somewhat similar to the way @code{extern crate} behaves.
16024 That is, if @value{GDBN} is stopped at a breakpoint in a function in
16025 crate @samp{A}, module @samp{B}, then @code{break B::f} will attempt
16026 to set a breakpoint in a function named @samp{f} in a crate named
16029 As a consequence of this approach, linespecs also cannot refer to
16030 items using @samp{self::} or @samp{super::}.
16033 Because @value{GDBN} implements Rust name-lookup semantics in
16034 expressions, it will sometimes prepend the current crate to a name.
16035 For example, if @value{GDBN} is stopped at a breakpoint in the crate
16036 @samp{K}, then @code{print ::x::y} will try to find the symbol
16039 However, since it is useful to be able to refer to other crates when
16040 debugging, @value{GDBN} provides the @code{extern} extension to
16041 circumvent this. To use the extension, just put @code{extern} before
16042 a path expression to refer to the otherwise unavailable ``global''
16045 In the above example, if you wanted to refer to the symbol @samp{y} in
16046 the crate @samp{x}, you would use @code{print extern x::y}.
16049 The Rust expression evaluator does not support ``statement-like''
16050 expressions such as @code{if} or @code{match}, or lambda expressions.
16053 Tuple expressions are not implemented.
16056 The Rust expression evaluator does not currently implement the
16057 @code{Drop} trait. Objects that may be created by the evaluator will
16058 never be destroyed.
16061 @value{GDBN} does not implement type inference for generics. In order
16062 to call generic functions or otherwise refer to generic items, you
16063 will have to specify the type parameters manually.
16066 @value{GDBN} currently uses the C@t{++} demangler for Rust. In most
16067 cases this does not cause any problems. However, in an expression
16068 context, completing a generic function name will give syntactically
16069 invalid results. This happens because Rust requires the @samp{::}
16070 operator between the function name and its generic arguments. For
16071 example, @value{GDBN} might provide a completion like
16072 @code{crate::f<u32>}, where the parser would require
16073 @code{crate::f::<u32>}.
16076 As of this writing, the Rust compiler (version 1.8) has a few holes in
16077 the debugging information it generates. These holes prevent certain
16078 features from being implemented by @value{GDBN}:
16082 Method calls cannot be made via traits.
16085 Operator overloading is not implemented.
16088 When debugging in a monomorphized function, you cannot use the generic
16092 The type @code{Self} is not available.
16095 @code{use} statements are not available, so some names may not be
16096 available in the crate.
16101 @subsection Modula-2
16103 @cindex Modula-2, @value{GDBN} support
16105 The extensions made to @value{GDBN} to support Modula-2 only support
16106 output from the @sc{gnu} Modula-2 compiler (which is currently being
16107 developed). Other Modula-2 compilers are not currently supported, and
16108 attempting to debug executables produced by them is most likely
16109 to give an error as @value{GDBN} reads in the executable's symbol
16112 @cindex expressions in Modula-2
16114 * M2 Operators:: Built-in operators
16115 * Built-In Func/Proc:: Built-in functions and procedures
16116 * M2 Constants:: Modula-2 constants
16117 * M2 Types:: Modula-2 types
16118 * M2 Defaults:: Default settings for Modula-2
16119 * Deviations:: Deviations from standard Modula-2
16120 * M2 Checks:: Modula-2 type and range checks
16121 * M2 Scope:: The scope operators @code{::} and @code{.}
16122 * GDB/M2:: @value{GDBN} and Modula-2
16126 @subsubsection Operators
16127 @cindex Modula-2 operators
16129 Operators must be defined on values of specific types. For instance,
16130 @code{+} is defined on numbers, but not on structures. Operators are
16131 often defined on groups of types. For the purposes of Modula-2, the
16132 following definitions hold:
16137 @emph{Integral types} consist of @code{INTEGER}, @code{CARDINAL}, and
16141 @emph{Character types} consist of @code{CHAR} and its subranges.
16144 @emph{Floating-point types} consist of @code{REAL}.
16147 @emph{Pointer types} consist of anything declared as @code{POINTER TO
16151 @emph{Scalar types} consist of all of the above.
16154 @emph{Set types} consist of @code{SET} and @code{BITSET} types.
16157 @emph{Boolean types} consist of @code{BOOLEAN}.
16161 The following operators are supported, and appear in order of
16162 increasing precedence:
16166 Function argument or array index separator.
16169 Assignment. The value of @var{var} @code{:=} @var{value} is
16173 Less than, greater than on integral, floating-point, or enumerated
16177 Less than or equal to, greater than or equal to
16178 on integral, floating-point and enumerated types, or set inclusion on
16179 set types. Same precedence as @code{<}.
16181 @item =@r{, }<>@r{, }#
16182 Equality and two ways of expressing inequality, valid on scalar types.
16183 Same precedence as @code{<}. In @value{GDBN} scripts, only @code{<>} is
16184 available for inequality, since @code{#} conflicts with the script
16188 Set membership. Defined on set types and the types of their members.
16189 Same precedence as @code{<}.
16192 Boolean disjunction. Defined on boolean types.
16195 Boolean conjunction. Defined on boolean types.
16198 The @value{GDBN} ``artificial array'' operator (@pxref{Expressions, ,Expressions}).
16201 Addition and subtraction on integral and floating-point types, or union
16202 and difference on set types.
16205 Multiplication on integral and floating-point types, or set intersection
16209 Division on floating-point types, or symmetric set difference on set
16210 types. Same precedence as @code{*}.
16213 Integer division and remainder. Defined on integral types. Same
16214 precedence as @code{*}.
16217 Negative. Defined on @code{INTEGER} and @code{REAL} data.
16220 Pointer dereferencing. Defined on pointer types.
16223 Boolean negation. Defined on boolean types. Same precedence as
16227 @code{RECORD} field selector. Defined on @code{RECORD} data. Same
16228 precedence as @code{^}.
16231 Array indexing. Defined on @code{ARRAY} data. Same precedence as @code{^}.
16234 Procedure argument list. Defined on @code{PROCEDURE} objects. Same precedence
16238 @value{GDBN} and Modula-2 scope operators.
16242 @emph{Warning:} Set expressions and their operations are not yet supported, so @value{GDBN}
16243 treats the use of the operator @code{IN}, or the use of operators
16244 @code{+}, @code{-}, @code{*}, @code{/}, @code{=}, , @code{<>}, @code{#},
16245 @code{<=}, and @code{>=} on sets as an error.
16249 @node Built-In Func/Proc
16250 @subsubsection Built-in Functions and Procedures
16251 @cindex Modula-2 built-ins
16253 Modula-2 also makes available several built-in procedures and functions.
16254 In describing these, the following metavariables are used:
16259 represents an @code{ARRAY} variable.
16262 represents a @code{CHAR} constant or variable.
16265 represents a variable or constant of integral type.
16268 represents an identifier that belongs to a set. Generally used in the
16269 same function with the metavariable @var{s}. The type of @var{s} should
16270 be @code{SET OF @var{mtype}} (where @var{mtype} is the type of @var{m}).
16273 represents a variable or constant of integral or floating-point type.
16276 represents a variable or constant of floating-point type.
16282 represents a variable.
16285 represents a variable or constant of one of many types. See the
16286 explanation of the function for details.
16289 All Modula-2 built-in procedures also return a result, described below.
16293 Returns the absolute value of @var{n}.
16296 If @var{c} is a lower case letter, it returns its upper case
16297 equivalent, otherwise it returns its argument.
16300 Returns the character whose ordinal value is @var{i}.
16303 Decrements the value in the variable @var{v} by one. Returns the new value.
16305 @item DEC(@var{v},@var{i})
16306 Decrements the value in the variable @var{v} by @var{i}. Returns the
16309 @item EXCL(@var{m},@var{s})
16310 Removes the element @var{m} from the set @var{s}. Returns the new
16313 @item FLOAT(@var{i})
16314 Returns the floating point equivalent of the integer @var{i}.
16316 @item HIGH(@var{a})
16317 Returns the index of the last member of @var{a}.
16320 Increments the value in the variable @var{v} by one. Returns the new value.
16322 @item INC(@var{v},@var{i})
16323 Increments the value in the variable @var{v} by @var{i}. Returns the
16326 @item INCL(@var{m},@var{s})
16327 Adds the element @var{m} to the set @var{s} if it is not already
16328 there. Returns the new set.
16331 Returns the maximum value of the type @var{t}.
16334 Returns the minimum value of the type @var{t}.
16337 Returns boolean TRUE if @var{i} is an odd number.
16340 Returns the ordinal value of its argument. For example, the ordinal
16341 value of a character is its @sc{ascii} value (on machines supporting
16342 the @sc{ascii} character set). The argument @var{x} must be of an
16343 ordered type, which include integral, character and enumerated types.
16345 @item SIZE(@var{x})
16346 Returns the size of its argument. The argument @var{x} can be a
16347 variable or a type.
16349 @item TRUNC(@var{r})
16350 Returns the integral part of @var{r}.
16352 @item TSIZE(@var{x})
16353 Returns the size of its argument. The argument @var{x} can be a
16354 variable or a type.
16356 @item VAL(@var{t},@var{i})
16357 Returns the member of the type @var{t} whose ordinal value is @var{i}.
16361 @emph{Warning:} Sets and their operations are not yet supported, so
16362 @value{GDBN} treats the use of procedures @code{INCL} and @code{EXCL} as
16366 @cindex Modula-2 constants
16368 @subsubsection Constants
16370 @value{GDBN} allows you to express the constants of Modula-2 in the following
16376 Integer constants are simply a sequence of digits. When used in an
16377 expression, a constant is interpreted to be type-compatible with the
16378 rest of the expression. Hexadecimal integers are specified by a
16379 trailing @samp{H}, and octal integers by a trailing @samp{B}.
16382 Floating point constants appear as a sequence of digits, followed by a
16383 decimal point and another sequence of digits. An optional exponent can
16384 then be specified, in the form @samp{E@r{[}+@r{|}-@r{]}@var{nnn}}, where
16385 @samp{@r{[}+@r{|}-@r{]}@var{nnn}} is the desired exponent. All of the
16386 digits of the floating point constant must be valid decimal (base 10)
16390 Character constants consist of a single character enclosed by a pair of
16391 like quotes, either single (@code{'}) or double (@code{"}). They may
16392 also be expressed by their ordinal value (their @sc{ascii} value, usually)
16393 followed by a @samp{C}.
16396 String constants consist of a sequence of characters enclosed by a
16397 pair of like quotes, either single (@code{'}) or double (@code{"}).
16398 Escape sequences in the style of C are also allowed. @xref{C
16399 Constants, ,C and C@t{++} Constants}, for a brief explanation of escape
16403 Enumerated constants consist of an enumerated identifier.
16406 Boolean constants consist of the identifiers @code{TRUE} and
16410 Pointer constants consist of integral values only.
16413 Set constants are not yet supported.
16417 @subsubsection Modula-2 Types
16418 @cindex Modula-2 types
16420 Currently @value{GDBN} can print the following data types in Modula-2
16421 syntax: array types, record types, set types, pointer types, procedure
16422 types, enumerated types, subrange types and base types. You can also
16423 print the contents of variables declared using these type.
16424 This section gives a number of simple source code examples together with
16425 sample @value{GDBN} sessions.
16427 The first example contains the following section of code:
16436 and you can request @value{GDBN} to interrogate the type and value of
16437 @code{r} and @code{s}.
16440 (@value{GDBP}) print s
16442 (@value{GDBP}) ptype s
16444 (@value{GDBP}) print r
16446 (@value{GDBP}) ptype r
16451 Likewise if your source code declares @code{s} as:
16455 s: SET ['A'..'Z'] ;
16459 then you may query the type of @code{s} by:
16462 (@value{GDBP}) ptype s
16463 type = SET ['A'..'Z']
16467 Note that at present you cannot interactively manipulate set
16468 expressions using the debugger.
16470 The following example shows how you might declare an array in Modula-2
16471 and how you can interact with @value{GDBN} to print its type and contents:
16475 s: ARRAY [-10..10] OF CHAR ;
16479 (@value{GDBP}) ptype s
16480 ARRAY [-10..10] OF CHAR
16483 Note that the array handling is not yet complete and although the type
16484 is printed correctly, expression handling still assumes that all
16485 arrays have a lower bound of zero and not @code{-10} as in the example
16488 Here are some more type related Modula-2 examples:
16492 colour = (blue, red, yellow, green) ;
16493 t = [blue..yellow] ;
16501 The @value{GDBN} interaction shows how you can query the data type
16502 and value of a variable.
16505 (@value{GDBP}) print s
16507 (@value{GDBP}) ptype t
16508 type = [blue..yellow]
16512 In this example a Modula-2 array is declared and its contents
16513 displayed. Observe that the contents are written in the same way as
16514 their @code{C} counterparts.
16518 s: ARRAY [1..5] OF CARDINAL ;
16524 (@value{GDBP}) print s
16525 $1 = @{1, 0, 0, 0, 0@}
16526 (@value{GDBP}) ptype s
16527 type = ARRAY [1..5] OF CARDINAL
16530 The Modula-2 language interface to @value{GDBN} also understands
16531 pointer types as shown in this example:
16535 s: POINTER TO ARRAY [1..5] OF CARDINAL ;
16542 and you can request that @value{GDBN} describes the type of @code{s}.
16545 (@value{GDBP}) ptype s
16546 type = POINTER TO ARRAY [1..5] OF CARDINAL
16549 @value{GDBN} handles compound types as we can see in this example.
16550 Here we combine array types, record types, pointer types and subrange
16561 myarray = ARRAY myrange OF CARDINAL ;
16562 myrange = [-2..2] ;
16564 s: POINTER TO ARRAY myrange OF foo ;
16568 and you can ask @value{GDBN} to describe the type of @code{s} as shown
16572 (@value{GDBP}) ptype s
16573 type = POINTER TO ARRAY [-2..2] OF foo = RECORD
16576 f3 : ARRAY [-2..2] OF CARDINAL;
16581 @subsubsection Modula-2 Defaults
16582 @cindex Modula-2 defaults
16584 If type and range checking are set automatically by @value{GDBN}, they
16585 both default to @code{on} whenever the working language changes to
16586 Modula-2. This happens regardless of whether you or @value{GDBN}
16587 selected the working language.
16589 If you allow @value{GDBN} to set the language automatically, then entering
16590 code compiled from a file whose name ends with @file{.mod} sets the
16591 working language to Modula-2. @xref{Automatically, ,Having @value{GDBN}
16592 Infer the Source Language}, for further details.
16595 @subsubsection Deviations from Standard Modula-2
16596 @cindex Modula-2, deviations from
16598 A few changes have been made to make Modula-2 programs easier to debug.
16599 This is done primarily via loosening its type strictness:
16603 Unlike in standard Modula-2, pointer constants can be formed by
16604 integers. This allows you to modify pointer variables during
16605 debugging. (In standard Modula-2, the actual address contained in a
16606 pointer variable is hidden from you; it can only be modified
16607 through direct assignment to another pointer variable or expression that
16608 returned a pointer.)
16611 C escape sequences can be used in strings and characters to represent
16612 non-printable characters. @value{GDBN} prints out strings with these
16613 escape sequences embedded. Single non-printable characters are
16614 printed using the @samp{CHR(@var{nnn})} format.
16617 The assignment operator (@code{:=}) returns the value of its right-hand
16621 All built-in procedures both modify @emph{and} return their argument.
16625 @subsubsection Modula-2 Type and Range Checks
16626 @cindex Modula-2 checks
16629 @emph{Warning:} in this release, @value{GDBN} does not yet perform type or
16632 @c FIXME remove warning when type/range checks added
16634 @value{GDBN} considers two Modula-2 variables type equivalent if:
16638 They are of types that have been declared equivalent via a @code{TYPE
16639 @var{t1} = @var{t2}} statement
16642 They have been declared on the same line. (Note: This is true of the
16643 @sc{gnu} Modula-2 compiler, but it may not be true of other compilers.)
16646 As long as type checking is enabled, any attempt to combine variables
16647 whose types are not equivalent is an error.
16649 Range checking is done on all mathematical operations, assignment, array
16650 index bounds, and all built-in functions and procedures.
16653 @subsubsection The Scope Operators @code{::} and @code{.}
16655 @cindex @code{.}, Modula-2 scope operator
16656 @cindex colon, doubled as scope operator
16658 @vindex colon-colon@r{, in Modula-2}
16659 @c Info cannot handle :: but TeX can.
16662 @vindex ::@r{, in Modula-2}
16665 There are a few subtle differences between the Modula-2 scope operator
16666 (@code{.}) and the @value{GDBN} scope operator (@code{::}). The two have
16671 @var{module} . @var{id}
16672 @var{scope} :: @var{id}
16676 where @var{scope} is the name of a module or a procedure,
16677 @var{module} the name of a module, and @var{id} is any declared
16678 identifier within your program, except another module.
16680 Using the @code{::} operator makes @value{GDBN} search the scope
16681 specified by @var{scope} for the identifier @var{id}. If it is not
16682 found in the specified scope, then @value{GDBN} searches all scopes
16683 enclosing the one specified by @var{scope}.
16685 Using the @code{.} operator makes @value{GDBN} search the current scope for
16686 the identifier specified by @var{id} that was imported from the
16687 definition module specified by @var{module}. With this operator, it is
16688 an error if the identifier @var{id} was not imported from definition
16689 module @var{module}, or if @var{id} is not an identifier in
16693 @subsubsection @value{GDBN} and Modula-2
16695 Some @value{GDBN} commands have little use when debugging Modula-2 programs.
16696 Five subcommands of @code{set print} and @code{show print} apply
16697 specifically to C and C@t{++}: @samp{vtbl}, @samp{demangle},
16698 @samp{asm-demangle}, @samp{object}, and @samp{union}. The first four
16699 apply to C@t{++}, and the last to the C @code{union} type, which has no direct
16700 analogue in Modula-2.
16702 The @code{@@} operator (@pxref{Expressions, ,Expressions}), while available
16703 with any language, is not useful with Modula-2. Its
16704 intent is to aid the debugging of @dfn{dynamic arrays}, which cannot be
16705 created in Modula-2 as they can in C or C@t{++}. However, because an
16706 address can be specified by an integral constant, the construct
16707 @samp{@{@var{type}@}@var{adrexp}} is still useful.
16709 @cindex @code{#} in Modula-2
16710 In @value{GDBN} scripts, the Modula-2 inequality operator @code{#} is
16711 interpreted as the beginning of a comment. Use @code{<>} instead.
16717 The extensions made to @value{GDBN} for Ada only support
16718 output from the @sc{gnu} Ada (GNAT) compiler.
16719 Other Ada compilers are not currently supported, and
16720 attempting to debug executables produced by them is most likely
16724 @cindex expressions in Ada
16726 * Ada Mode Intro:: General remarks on the Ada syntax
16727 and semantics supported by Ada mode
16729 * Omissions from Ada:: Restrictions on the Ada expression syntax.
16730 * Additions to Ada:: Extensions of the Ada expression syntax.
16731 * Overloading support for Ada:: Support for expressions involving overloaded
16733 * Stopping Before Main Program:: Debugging the program during elaboration.
16734 * Ada Exceptions:: Ada Exceptions
16735 * Ada Tasks:: Listing and setting breakpoints in tasks.
16736 * Ada Tasks and Core Files:: Tasking Support when Debugging Core Files
16737 * Ravenscar Profile:: Tasking Support when using the Ravenscar
16739 * Ada Settings:: New settable GDB parameters for Ada.
16740 * Ada Glitches:: Known peculiarities of Ada mode.
16743 @node Ada Mode Intro
16744 @subsubsection Introduction
16745 @cindex Ada mode, general
16747 The Ada mode of @value{GDBN} supports a fairly large subset of Ada expression
16748 syntax, with some extensions.
16749 The philosophy behind the design of this subset is
16753 That @value{GDBN} should provide basic literals and access to operations for
16754 arithmetic, dereferencing, field selection, indexing, and subprogram calls,
16755 leaving more sophisticated computations to subprograms written into the
16756 program (which therefore may be called from @value{GDBN}).
16759 That type safety and strict adherence to Ada language restrictions
16760 are not particularly important to the @value{GDBN} user.
16763 That brevity is important to the @value{GDBN} user.
16766 Thus, for brevity, the debugger acts as if all names declared in
16767 user-written packages are directly visible, even if they are not visible
16768 according to Ada rules, thus making it unnecessary to fully qualify most
16769 names with their packages, regardless of context. Where this causes
16770 ambiguity, @value{GDBN} asks the user's intent.
16772 The debugger will start in Ada mode if it detects an Ada main program.
16773 As for other languages, it will enter Ada mode when stopped in a program that
16774 was translated from an Ada source file.
16776 While in Ada mode, you may use `@t{--}' for comments. This is useful
16777 mostly for documenting command files. The standard @value{GDBN} comment
16778 (@samp{#}) still works at the beginning of a line in Ada mode, but not in the
16779 middle (to allow based literals).
16781 @node Omissions from Ada
16782 @subsubsection Omissions from Ada
16783 @cindex Ada, omissions from
16785 Here are the notable omissions from the subset:
16789 Only a subset of the attributes are supported:
16793 @t{'First}, @t{'Last}, and @t{'Length}
16794 on array objects (not on types and subtypes).
16797 @t{'Min} and @t{'Max}.
16800 @t{'Pos} and @t{'Val}.
16806 @t{'Range} on array objects (not subtypes), but only as the right
16807 operand of the membership (@code{in}) operator.
16810 @t{'Access}, @t{'Unchecked_Access}, and
16811 @t{'Unrestricted_Access} (a GNAT extension).
16819 @code{Characters.Latin_1} are not available and
16820 concatenation is not implemented. Thus, escape characters in strings are
16821 not currently available.
16824 Equality tests (@samp{=} and @samp{/=}) on arrays test for bitwise
16825 equality of representations. They will generally work correctly
16826 for strings and arrays whose elements have integer or enumeration types.
16827 They may not work correctly for arrays whose element
16828 types have user-defined equality, for arrays of real values
16829 (in particular, IEEE-conformant floating point, because of negative
16830 zeroes and NaNs), and for arrays whose elements contain unused bits with
16831 indeterminate values.
16834 The other component-by-component array operations (@code{and}, @code{or},
16835 @code{xor}, @code{not}, and relational tests other than equality)
16836 are not implemented.
16839 @cindex array aggregates (Ada)
16840 @cindex record aggregates (Ada)
16841 @cindex aggregates (Ada)
16842 There is limited support for array and record aggregates. They are
16843 permitted only on the right sides of assignments, as in these examples:
16846 (@value{GDBP}) set An_Array := (1, 2, 3, 4, 5, 6)
16847 (@value{GDBP}) set An_Array := (1, others => 0)
16848 (@value{GDBP}) set An_Array := (0|4 => 1, 1..3 => 2, 5 => 6)
16849 (@value{GDBP}) set A_2D_Array := ((1, 2, 3), (4, 5, 6), (7, 8, 9))
16850 (@value{GDBP}) set A_Record := (1, "Peter", True);
16851 (@value{GDBP}) set A_Record := (Name => "Peter", Id => 1, Alive => True)
16855 discriminant's value by assigning an aggregate has an
16856 undefined effect if that discriminant is used within the record.
16857 However, you can first modify discriminants by directly assigning to
16858 them (which normally would not be allowed in Ada), and then performing an
16859 aggregate assignment. For example, given a variable @code{A_Rec}
16860 declared to have a type such as:
16863 type Rec (Len : Small_Integer := 0) is record
16865 Vals : IntArray (1 .. Len);
16869 you can assign a value with a different size of @code{Vals} with two
16873 (@value{GDBP}) set A_Rec.Len := 4
16874 (@value{GDBP}) set A_Rec := (Id => 42, Vals => (1, 2, 3, 4))
16877 As this example also illustrates, @value{GDBN} is very loose about the usual
16878 rules concerning aggregates. You may leave out some of the
16879 components of an array or record aggregate (such as the @code{Len}
16880 component in the assignment to @code{A_Rec} above); they will retain their
16881 original values upon assignment. You may freely use dynamic values as
16882 indices in component associations. You may even use overlapping or
16883 redundant component associations, although which component values are
16884 assigned in such cases is not defined.
16887 Calls to dispatching subprograms are not implemented.
16890 The overloading algorithm is much more limited (i.e., less selective)
16891 than that of real Ada. It makes only limited use of the context in
16892 which a subexpression appears to resolve its meaning, and it is much
16893 looser in its rules for allowing type matches. As a result, some
16894 function calls will be ambiguous, and the user will be asked to choose
16895 the proper resolution.
16898 The @code{new} operator is not implemented.
16901 Entry calls are not implemented.
16904 Aside from printing, arithmetic operations on the native VAX floating-point
16905 formats are not supported.
16908 It is not possible to slice a packed array.
16911 The names @code{True} and @code{False}, when not part of a qualified name,
16912 are interpreted as if implicitly prefixed by @code{Standard}, regardless of
16914 Should your program
16915 redefine these names in a package or procedure (at best a dubious practice),
16916 you will have to use fully qualified names to access their new definitions.
16919 @node Additions to Ada
16920 @subsubsection Additions to Ada
16921 @cindex Ada, deviations from
16923 As it does for other languages, @value{GDBN} makes certain generic
16924 extensions to Ada (@pxref{Expressions}):
16928 If the expression @var{E} is a variable residing in memory (typically
16929 a local variable or array element) and @var{N} is a positive integer,
16930 then @code{@var{E}@@@var{N}} displays the values of @var{E} and the
16931 @var{N}-1 adjacent variables following it in memory as an array. In
16932 Ada, this operator is generally not necessary, since its prime use is
16933 in displaying parts of an array, and slicing will usually do this in
16934 Ada. However, there are occasional uses when debugging programs in
16935 which certain debugging information has been optimized away.
16938 @code{@var{B}::@var{var}} means ``the variable named @var{var} that
16939 appears in function or file @var{B}.'' When @var{B} is a file name,
16940 you must typically surround it in single quotes.
16943 The expression @code{@{@var{type}@} @var{addr}} means ``the variable of type
16944 @var{type} that appears at address @var{addr}.''
16947 A name starting with @samp{$} is a convenience variable
16948 (@pxref{Convenience Vars}) or a machine register (@pxref{Registers}).
16951 In addition, @value{GDBN} provides a few other shortcuts and outright
16952 additions specific to Ada:
16956 The assignment statement is allowed as an expression, returning
16957 its right-hand operand as its value. Thus, you may enter
16960 (@value{GDBP}) set x := y + 3
16961 (@value{GDBP}) print A(tmp := y + 1)
16965 The semicolon is allowed as an ``operator,'' returning as its value
16966 the value of its right-hand operand.
16967 This allows, for example,
16968 complex conditional breaks:
16971 (@value{GDBP}) break f
16972 (@value{GDBP}) condition 1 (report(i); k += 1; A(k) > 100)
16976 Rather than use catenation and symbolic character names to introduce special
16977 characters into strings, one may instead use a special bracket notation,
16978 which is also used to print strings. A sequence of characters of the form
16979 @samp{["@var{XX}"]} within a string or character literal denotes the
16980 (single) character whose numeric encoding is @var{XX} in hexadecimal. The
16981 sequence of characters @samp{["""]} also denotes a single quotation mark
16982 in strings. For example,
16984 "One line.["0a"]Next line.["0a"]"
16987 contains an ASCII newline character (@code{Ada.Characters.Latin_1.LF})
16991 The subtype used as a prefix for the attributes @t{'Pos}, @t{'Min}, and
16992 @t{'Max} is optional (and is ignored in any case). For example, it is valid
16996 (@value{GDBP}) print 'max(x, y)
17000 When printing arrays, @value{GDBN} uses positional notation when the
17001 array has a lower bound of 1, and uses a modified named notation otherwise.
17002 For example, a one-dimensional array of three integers with a lower bound
17003 of 3 might print as
17010 That is, in contrast to valid Ada, only the first component has a @code{=>}
17014 You may abbreviate attributes in expressions with any unique,
17015 multi-character subsequence of
17016 their names (an exact match gets preference).
17017 For example, you may use @t{a'len}, @t{a'gth}, or @t{a'lh}
17018 in place of @t{a'length}.
17021 @cindex quoting Ada internal identifiers
17022 Since Ada is case-insensitive, the debugger normally maps identifiers you type
17023 to lower case. The GNAT compiler uses upper-case characters for
17024 some of its internal identifiers, which are normally of no interest to users.
17025 For the rare occasions when you actually have to look at them,
17026 enclose them in angle brackets to avoid the lower-case mapping.
17029 (@value{GDBP}) print <JMPBUF_SAVE>[0]
17033 Printing an object of class-wide type or dereferencing an
17034 access-to-class-wide value will display all the components of the object's
17035 specific type (as indicated by its run-time tag). Likewise, component
17036 selection on such a value will operate on the specific type of the
17041 @node Overloading support for Ada
17042 @subsubsection Overloading support for Ada
17043 @cindex overloading, Ada
17045 The debugger supports limited overloading. Given a subprogram call in which
17046 the function symbol has multiple definitions, it will use the number of
17047 actual parameters and some information about their types to attempt to narrow
17048 the set of definitions. It also makes very limited use of context, preferring
17049 procedures to functions in the context of the @code{call} command, and
17050 functions to procedures elsewhere.
17052 If, after narrowing, the set of matching definitions still contains more than
17053 one definition, @value{GDBN} will display a menu to query which one it should
17057 (@value{GDBP}) print f(1)
17058 Multiple matches for f
17060 [1] foo.f (integer) return boolean at foo.adb:23
17061 [2] foo.f (foo.new_integer) return boolean at foo.adb:28
17065 In this case, just select one menu entry either to cancel expression evaluation
17066 (type @kbd{0} and press @key{RET}) or to continue evaluation with a specific
17067 instance (type the corresponding number and press @key{RET}).
17069 Here are a couple of commands to customize @value{GDBN}'s behavior in this
17074 @kindex set ada print-signatures
17075 @item set ada print-signatures
17076 Control whether parameter types and return types are displayed in overloads
17077 selection menus. It is @code{on} by default.
17078 @xref{Overloading support for Ada}.
17080 @kindex show ada print-signatures
17081 @item show ada print-signatures
17082 Show the current setting for displaying parameter types and return types in
17083 overloads selection menu.
17084 @xref{Overloading support for Ada}.
17088 @node Stopping Before Main Program
17089 @subsubsection Stopping at the Very Beginning
17091 @cindex breakpointing Ada elaboration code
17092 It is sometimes necessary to debug the program during elaboration, and
17093 before reaching the main procedure.
17094 As defined in the Ada Reference
17095 Manual, the elaboration code is invoked from a procedure called
17096 @code{adainit}. To run your program up to the beginning of
17097 elaboration, simply use the following two commands:
17098 @code{tbreak adainit} and @code{run}.
17100 @node Ada Exceptions
17101 @subsubsection Ada Exceptions
17103 A command is provided to list all Ada exceptions:
17106 @kindex info exceptions
17107 @item info exceptions
17108 @itemx info exceptions @var{regexp}
17109 The @code{info exceptions} command allows you to list all Ada exceptions
17110 defined within the program being debugged, as well as their addresses.
17111 With a regular expression, @var{regexp}, as argument, only those exceptions
17112 whose names match @var{regexp} are listed.
17115 Below is a small example, showing how the command can be used, first
17116 without argument, and next with a regular expression passed as an
17120 (@value{GDBP}) info exceptions
17121 All defined Ada exceptions:
17122 constraint_error: 0x613da0
17123 program_error: 0x613d20
17124 storage_error: 0x613ce0
17125 tasking_error: 0x613ca0
17126 const.aint_global_e: 0x613b00
17127 (@value{GDBP}) info exceptions const.aint
17128 All Ada exceptions matching regular expression "const.aint":
17129 constraint_error: 0x613da0
17130 const.aint_global_e: 0x613b00
17133 It is also possible to ask @value{GDBN} to stop your program's execution
17134 when an exception is raised. For more details, see @ref{Set Catchpoints}.
17137 @subsubsection Extensions for Ada Tasks
17138 @cindex Ada, tasking
17140 Support for Ada tasks is analogous to that for threads (@pxref{Threads}).
17141 @value{GDBN} provides the following task-related commands:
17146 This command shows a list of current Ada tasks, as in the following example:
17153 (@value{GDBP}) info tasks
17154 ID TID P-ID Pri State Name
17155 1 8088000 0 15 Child Activation Wait main_task
17156 2 80a4000 1 15 Accept Statement b
17157 3 809a800 1 15 Child Activation Wait a
17158 * 4 80ae800 3 15 Runnable c
17163 In this listing, the asterisk before the last task indicates it to be the
17164 task currently being inspected.
17168 Represents @value{GDBN}'s internal task number.
17174 The parent's task ID (@value{GDBN}'s internal task number).
17177 The base priority of the task.
17180 Current state of the task.
17184 The task has been created but has not been activated. It cannot be
17188 The task is not blocked for any reason known to Ada. (It may be waiting
17189 for a mutex, though.) It is conceptually "executing" in normal mode.
17192 The task is terminated, in the sense of ARM 9.3 (5). Any dependents
17193 that were waiting on terminate alternatives have been awakened and have
17194 terminated themselves.
17196 @item Child Activation Wait
17197 The task is waiting for created tasks to complete activation.
17199 @item Accept Statement
17200 The task is waiting on an accept or selective wait statement.
17202 @item Waiting on entry call
17203 The task is waiting on an entry call.
17205 @item Async Select Wait
17206 The task is waiting to start the abortable part of an asynchronous
17210 The task is waiting on a select statement with only a delay
17213 @item Child Termination Wait
17214 The task is sleeping having completed a master within itself, and is
17215 waiting for the tasks dependent on that master to become terminated or
17216 waiting on a terminate Phase.
17218 @item Wait Child in Term Alt
17219 The task is sleeping waiting for tasks on terminate alternatives to
17220 finish terminating.
17222 @item Accepting RV with @var{taskno}
17223 The task is accepting a rendez-vous with the task @var{taskno}.
17227 Name of the task in the program.
17231 @kindex info task @var{taskno}
17232 @item info task @var{taskno}
17233 This command shows detailled informations on the specified task, as in
17234 the following example:
17239 (@value{GDBP}) info tasks
17240 ID TID P-ID Pri State Name
17241 1 8077880 0 15 Child Activation Wait main_task
17242 * 2 807c468 1 15 Runnable task_1
17243 (@value{GDBP}) info task 2
17244 Ada Task: 0x807c468
17248 Parent: 1 (main_task)
17254 @kindex task@r{ (Ada)}
17255 @cindex current Ada task ID
17256 This command prints the ID of the current task.
17262 (@value{GDBP}) info tasks
17263 ID TID P-ID Pri State Name
17264 1 8077870 0 15 Child Activation Wait main_task
17265 * 2 807c458 1 15 Runnable t
17266 (@value{GDBP}) task
17267 [Current task is 2]
17270 @item task @var{taskno}
17271 @cindex Ada task switching
17272 This command is like the @code{thread @var{thread-id}}
17273 command (@pxref{Threads}). It switches the context of debugging
17274 from the current task to the given task.
17280 (@value{GDBP}) info tasks
17281 ID TID P-ID Pri State Name
17282 1 8077870 0 15 Child Activation Wait main_task
17283 * 2 807c458 1 15 Runnable t
17284 (@value{GDBP}) task 1
17285 [Switching to task 1]
17286 #0 0x8067726 in pthread_cond_wait ()
17288 #0 0x8067726 in pthread_cond_wait ()
17289 #1 0x8056714 in system.os_interface.pthread_cond_wait ()
17290 #2 0x805cb63 in system.task_primitives.operations.sleep ()
17291 #3 0x806153e in system.tasking.stages.activate_tasks ()
17292 #4 0x804aacc in un () at un.adb:5
17295 @item break @var{location} task @var{taskno}
17296 @itemx break @var{location} task @var{taskno} if @dots{}
17297 @cindex breakpoints and tasks, in Ada
17298 @cindex task breakpoints, in Ada
17299 @kindex break @dots{} task @var{taskno}@r{ (Ada)}
17300 These commands are like the @code{break @dots{} thread @dots{}}
17301 command (@pxref{Thread Stops}). The
17302 @var{location} argument specifies source lines, as described
17303 in @ref{Specify Location}.
17305 Use the qualifier @samp{task @var{taskno}} with a breakpoint command
17306 to specify that you only want @value{GDBN} to stop the program when a
17307 particular Ada task reaches this breakpoint. The @var{taskno} is one of the
17308 numeric task identifiers assigned by @value{GDBN}, shown in the first
17309 column of the @samp{info tasks} display.
17311 If you do not specify @samp{task @var{taskno}} when you set a
17312 breakpoint, the breakpoint applies to @emph{all} tasks of your
17315 You can use the @code{task} qualifier on conditional breakpoints as
17316 well; in this case, place @samp{task @var{taskno}} before the
17317 breakpoint condition (before the @code{if}).
17325 (@value{GDBP}) info tasks
17326 ID TID P-ID Pri State Name
17327 1 140022020 0 15 Child Activation Wait main_task
17328 2 140045060 1 15 Accept/Select Wait t2
17329 3 140044840 1 15 Runnable t1
17330 * 4 140056040 1 15 Runnable t3
17331 (@value{GDBP}) b 15 task 2
17332 Breakpoint 5 at 0x120044cb0: file test_task_debug.adb, line 15.
17333 (@value{GDBP}) cont
17338 Breakpoint 5, test_task_debug () at test_task_debug.adb:15
17340 (@value{GDBP}) info tasks
17341 ID TID P-ID Pri State Name
17342 1 140022020 0 15 Child Activation Wait main_task
17343 * 2 140045060 1 15 Runnable t2
17344 3 140044840 1 15 Runnable t1
17345 4 140056040 1 15 Delay Sleep t3
17349 @node Ada Tasks and Core Files
17350 @subsubsection Tasking Support when Debugging Core Files
17351 @cindex Ada tasking and core file debugging
17353 When inspecting a core file, as opposed to debugging a live program,
17354 tasking support may be limited or even unavailable, depending on
17355 the platform being used.
17356 For instance, on x86-linux, the list of tasks is available, but task
17357 switching is not supported.
17359 On certain platforms, the debugger needs to perform some
17360 memory writes in order to provide Ada tasking support. When inspecting
17361 a core file, this means that the core file must be opened with read-write
17362 privileges, using the command @samp{"set write on"} (@pxref{Patching}).
17363 Under these circumstances, you should make a backup copy of the core
17364 file before inspecting it with @value{GDBN}.
17366 @node Ravenscar Profile
17367 @subsubsection Tasking Support when using the Ravenscar Profile
17368 @cindex Ravenscar Profile
17370 The @dfn{Ravenscar Profile} is a subset of the Ada tasking features,
17371 specifically designed for systems with safety-critical real-time
17375 @kindex set ravenscar task-switching on
17376 @cindex task switching with program using Ravenscar Profile
17377 @item set ravenscar task-switching on
17378 Allows task switching when debugging a program that uses the Ravenscar
17379 Profile. This is the default.
17381 @kindex set ravenscar task-switching off
17382 @item set ravenscar task-switching off
17383 Turn off task switching when debugging a program that uses the Ravenscar
17384 Profile. This is mostly intended to disable the code that adds support
17385 for the Ravenscar Profile, in case a bug in either @value{GDBN} or in
17386 the Ravenscar runtime is preventing @value{GDBN} from working properly.
17387 To be effective, this command should be run before the program is started.
17389 @kindex show ravenscar task-switching
17390 @item show ravenscar task-switching
17391 Show whether it is possible to switch from task to task in a program
17392 using the Ravenscar Profile.
17397 @subsubsection Ada Settings
17398 @cindex Ada settings
17401 @kindex set varsize-limit
17402 @item set varsize-limit @var{size}
17403 Prevent @value{GDBN} from attempting to evaluate objects whose size
17404 is above the given limit (@var{size}) when those sizes are computed
17405 from run-time quantities. This is typically the case when the object
17406 has a variable size, such as an array whose bounds are not known at
17407 compile time for example. Setting @var{size} to @code{unlimited}
17408 removes the size limitation. By default, the limit is about 65KB.
17410 The purpose of having such a limit is to prevent @value{GDBN} from
17411 trying to grab enormous chunks of virtual memory when asked to evaluate
17412 a quantity whose bounds have been corrupted or have not yet been fully
17413 initialized. The limit applies to the results of some subexpressions
17414 as well as to complete expressions. For example, an expression denoting
17415 a simple integer component, such as @code{x.y.z}, may fail if the size of
17416 @code{x.y} is variable and exceeds @code{size}. On the other hand,
17417 @value{GDBN} is sometimes clever; the expression @code{A(i)}, where
17418 @code{A} is an array variable with non-constant size, will generally
17419 succeed regardless of the bounds on @code{A}, as long as the component
17420 size is less than @var{size}.
17422 @kindex show varsize-limit
17423 @item show varsize-limit
17424 Show the limit on types whose size is determined by run-time quantities.
17428 @subsubsection Known Peculiarities of Ada Mode
17429 @cindex Ada, problems
17431 Besides the omissions listed previously (@pxref{Omissions from Ada}),
17432 we know of several problems with and limitations of Ada mode in
17434 some of which will be fixed with planned future releases of the debugger
17435 and the GNU Ada compiler.
17439 Static constants that the compiler chooses not to materialize as objects in
17440 storage are invisible to the debugger.
17443 Named parameter associations in function argument lists are ignored (the
17444 argument lists are treated as positional).
17447 Many useful library packages are currently invisible to the debugger.
17450 Fixed-point arithmetic, conversions, input, and output is carried out using
17451 floating-point arithmetic, and may give results that only approximate those on
17455 The GNAT compiler never generates the prefix @code{Standard} for any of
17456 the standard symbols defined by the Ada language. @value{GDBN} knows about
17457 this: it will strip the prefix from names when you use it, and will never
17458 look for a name you have so qualified among local symbols, nor match against
17459 symbols in other packages or subprograms. If you have
17460 defined entities anywhere in your program other than parameters and
17461 local variables whose simple names match names in @code{Standard},
17462 GNAT's lack of qualification here can cause confusion. When this happens,
17463 you can usually resolve the confusion
17464 by qualifying the problematic names with package
17465 @code{Standard} explicitly.
17468 Older versions of the compiler sometimes generate erroneous debugging
17469 information, resulting in the debugger incorrectly printing the value
17470 of affected entities. In some cases, the debugger is able to work
17471 around an issue automatically. In other cases, the debugger is able
17472 to work around the issue, but the work-around has to be specifically
17475 @kindex set ada trust-PAD-over-XVS
17476 @kindex show ada trust-PAD-over-XVS
17479 @item set ada trust-PAD-over-XVS on
17480 Configure GDB to strictly follow the GNAT encoding when computing the
17481 value of Ada entities, particularly when @code{PAD} and @code{PAD___XVS}
17482 types are involved (see @code{ada/exp_dbug.ads} in the GCC sources for
17483 a complete description of the encoding used by the GNAT compiler).
17484 This is the default.
17486 @item set ada trust-PAD-over-XVS off
17487 This is related to the encoding using by the GNAT compiler. If @value{GDBN}
17488 sometimes prints the wrong value for certain entities, changing @code{ada
17489 trust-PAD-over-XVS} to @code{off} activates a work-around which may fix
17490 the issue. It is always safe to set @code{ada trust-PAD-over-XVS} to
17491 @code{off}, but this incurs a slight performance penalty, so it is
17492 recommended to leave this setting to @code{on} unless necessary.
17496 @cindex GNAT descriptive types
17497 @cindex GNAT encoding
17498 Internally, the debugger also relies on the compiler following a number
17499 of conventions known as the @samp{GNAT Encoding}, all documented in
17500 @file{gcc/ada/exp_dbug.ads} in the GCC sources. This encoding describes
17501 how the debugging information should be generated for certain types.
17502 In particular, this convention makes use of @dfn{descriptive types},
17503 which are artificial types generated purely to help the debugger.
17505 These encodings were defined at a time when the debugging information
17506 format used was not powerful enough to describe some of the more complex
17507 types available in Ada. Since DWARF allows us to express nearly all
17508 Ada features, the long-term goal is to slowly replace these descriptive
17509 types by their pure DWARF equivalent. To facilitate that transition,
17510 a new maintenance option is available to force the debugger to ignore
17511 those descriptive types. It allows the user to quickly evaluate how
17512 well @value{GDBN} works without them.
17516 @kindex maint ada set ignore-descriptive-types
17517 @item maintenance ada set ignore-descriptive-types [on|off]
17518 Control whether the debugger should ignore descriptive types.
17519 The default is not to ignore descriptives types (@code{off}).
17521 @kindex maint ada show ignore-descriptive-types
17522 @item maintenance ada show ignore-descriptive-types
17523 Show if descriptive types are ignored by @value{GDBN}.
17527 @node Unsupported Languages
17528 @section Unsupported Languages
17530 @cindex unsupported languages
17531 @cindex minimal language
17532 In addition to the other fully-supported programming languages,
17533 @value{GDBN} also provides a pseudo-language, called @code{minimal}.
17534 It does not represent a real programming language, but provides a set
17535 of capabilities close to what the C or assembly languages provide.
17536 This should allow most simple operations to be performed while debugging
17537 an application that uses a language currently not supported by @value{GDBN}.
17539 If the language is set to @code{auto}, @value{GDBN} will automatically
17540 select this language if the current frame corresponds to an unsupported
17544 @chapter Examining the Symbol Table
17546 The commands described in this chapter allow you to inquire about the
17547 symbols (names of variables, functions and types) defined in your
17548 program. This information is inherent in the text of your program and
17549 does not change as your program executes. @value{GDBN} finds it in your
17550 program's symbol table, in the file indicated when you started @value{GDBN}
17551 (@pxref{File Options, ,Choosing Files}), or by one of the
17552 file-management commands (@pxref{Files, ,Commands to Specify Files}).
17554 @cindex symbol names
17555 @cindex names of symbols
17556 @cindex quoting names
17557 @anchor{quoting names}
17558 Occasionally, you may need to refer to symbols that contain unusual
17559 characters, which @value{GDBN} ordinarily treats as word delimiters. The
17560 most frequent case is in referring to static variables in other
17561 source files (@pxref{Variables,,Program Variables}). File names
17562 are recorded in object files as debugging symbols, but @value{GDBN} would
17563 ordinarily parse a typical file name, like @file{foo.c}, as the three words
17564 @samp{foo} @samp{.} @samp{c}. To allow @value{GDBN} to recognize
17565 @samp{foo.c} as a single symbol, enclose it in single quotes; for example,
17572 looks up the value of @code{x} in the scope of the file @file{foo.c}.
17575 @cindex case-insensitive symbol names
17576 @cindex case sensitivity in symbol names
17577 @kindex set case-sensitive
17578 @item set case-sensitive on
17579 @itemx set case-sensitive off
17580 @itemx set case-sensitive auto
17581 Normally, when @value{GDBN} looks up symbols, it matches their names
17582 with case sensitivity determined by the current source language.
17583 Occasionally, you may wish to control that. The command @code{set
17584 case-sensitive} lets you do that by specifying @code{on} for
17585 case-sensitive matches or @code{off} for case-insensitive ones. If
17586 you specify @code{auto}, case sensitivity is reset to the default
17587 suitable for the source language. The default is case-sensitive
17588 matches for all languages except for Fortran, for which the default is
17589 case-insensitive matches.
17591 @kindex show case-sensitive
17592 @item show case-sensitive
17593 This command shows the current setting of case sensitivity for symbols
17596 @kindex set print type methods
17597 @item set print type methods
17598 @itemx set print type methods on
17599 @itemx set print type methods off
17600 Normally, when @value{GDBN} prints a class, it displays any methods
17601 declared in that class. You can control this behavior either by
17602 passing the appropriate flag to @code{ptype}, or using @command{set
17603 print type methods}. Specifying @code{on} will cause @value{GDBN} to
17604 display the methods; this is the default. Specifying @code{off} will
17605 cause @value{GDBN} to omit the methods.
17607 @kindex show print type methods
17608 @item show print type methods
17609 This command shows the current setting of method display when printing
17612 @kindex set print type nested-type-limit
17613 @item set print type nested-type-limit @var{limit}
17614 @itemx set print type nested-type-limit unlimited
17615 Set the limit of displayed nested types that the type printer will
17616 show. A @var{limit} of @code{unlimited} or @code{-1} will show all
17617 nested definitions. By default, the type printer will not show any nested
17618 types defined in classes.
17620 @kindex show print type nested-type-limit
17621 @item show print type nested-type-limit
17622 This command shows the current display limit of nested types when
17625 @kindex set print type typedefs
17626 @item set print type typedefs
17627 @itemx set print type typedefs on
17628 @itemx set print type typedefs off
17630 Normally, when @value{GDBN} prints a class, it displays any typedefs
17631 defined in that class. You can control this behavior either by
17632 passing the appropriate flag to @code{ptype}, or using @command{set
17633 print type typedefs}. Specifying @code{on} will cause @value{GDBN} to
17634 display the typedef definitions; this is the default. Specifying
17635 @code{off} will cause @value{GDBN} to omit the typedef definitions.
17636 Note that this controls whether the typedef definition itself is
17637 printed, not whether typedef names are substituted when printing other
17640 @kindex show print type typedefs
17641 @item show print type typedefs
17642 This command shows the current setting of typedef display when
17645 @kindex info address
17646 @cindex address of a symbol
17647 @item info address @var{symbol}
17648 Describe where the data for @var{symbol} is stored. For a register
17649 variable, this says which register it is kept in. For a non-register
17650 local variable, this prints the stack-frame offset at which the variable
17653 Note the contrast with @samp{print &@var{symbol}}, which does not work
17654 at all for a register variable, and for a stack local variable prints
17655 the exact address of the current instantiation of the variable.
17657 @kindex info symbol
17658 @cindex symbol from address
17659 @cindex closest symbol and offset for an address
17660 @item info symbol @var{addr}
17661 Print the name of a symbol which is stored at the address @var{addr}.
17662 If no symbol is stored exactly at @var{addr}, @value{GDBN} prints the
17663 nearest symbol and an offset from it:
17666 (@value{GDBP}) info symbol 0x54320
17667 _initialize_vx + 396 in section .text
17671 This is the opposite of the @code{info address} command. You can use
17672 it to find out the name of a variable or a function given its address.
17674 For dynamically linked executables, the name of executable or shared
17675 library containing the symbol is also printed:
17678 (@value{GDBP}) info symbol 0x400225
17679 _start + 5 in section .text of /tmp/a.out
17680 (@value{GDBP}) info symbol 0x2aaaac2811cf
17681 __read_nocancel + 6 in section .text of /usr/lib64/libc.so.6
17686 @item demangle @r{[}-l @var{language}@r{]} @r{[}@var{--}@r{]} @var{name}
17687 Demangle @var{name}.
17688 If @var{language} is provided it is the name of the language to demangle
17689 @var{name} in. Otherwise @var{name} is demangled in the current language.
17691 The @samp{--} option specifies the end of options,
17692 and is useful when @var{name} begins with a dash.
17694 The parameter @code{demangle-style} specifies how to interpret the kind
17695 of mangling used. @xref{Print Settings}.
17698 @item whatis[/@var{flags}] [@var{arg}]
17699 Print the data type of @var{arg}, which can be either an expression
17700 or a name of a data type. With no argument, print the data type of
17701 @code{$}, the last value in the value history.
17703 If @var{arg} is an expression (@pxref{Expressions, ,Expressions}), it
17704 is not actually evaluated, and any side-effecting operations (such as
17705 assignments or function calls) inside it do not take place.
17707 If @var{arg} is a variable or an expression, @code{whatis} prints its
17708 literal type as it is used in the source code. If the type was
17709 defined using a @code{typedef}, @code{whatis} will @emph{not} print
17710 the data type underlying the @code{typedef}. If the type of the
17711 variable or the expression is a compound data type, such as
17712 @code{struct} or @code{class}, @code{whatis} never prints their
17713 fields or methods. It just prints the @code{struct}/@code{class}
17714 name (a.k.a.@: its @dfn{tag}). If you want to see the members of
17715 such a compound data type, use @code{ptype}.
17717 If @var{arg} is a type name that was defined using @code{typedef},
17718 @code{whatis} @dfn{unrolls} only one level of that @code{typedef}.
17719 Unrolling means that @code{whatis} will show the underlying type used
17720 in the @code{typedef} declaration of @var{arg}. However, if that
17721 underlying type is also a @code{typedef}, @code{whatis} will not
17724 For C code, the type names may also have the form @samp{class
17725 @var{class-name}}, @samp{struct @var{struct-tag}}, @samp{union
17726 @var{union-tag}} or @samp{enum @var{enum-tag}}.
17728 @var{flags} can be used to modify how the type is displayed.
17729 Available flags are:
17733 Display in ``raw'' form. Normally, @value{GDBN} substitutes template
17734 parameters and typedefs defined in a class when printing the class'
17735 members. The @code{/r} flag disables this.
17738 Do not print methods defined in the class.
17741 Print methods defined in the class. This is the default, but the flag
17742 exists in case you change the default with @command{set print type methods}.
17745 Do not print typedefs defined in the class. Note that this controls
17746 whether the typedef definition itself is printed, not whether typedef
17747 names are substituted when printing other types.
17750 Print typedefs defined in the class. This is the default, but the flag
17751 exists in case you change the default with @command{set print type typedefs}.
17754 Print the offsets and sizes of fields in a struct, similar to what the
17755 @command{pahole} tool does. This option implies the @code{/tm} flags.
17757 For example, given the following declarations:
17793 Issuing a @kbd{ptype /o struct tuv} command would print:
17796 (@value{GDBP}) ptype /o struct tuv
17797 /* offset | size */ type = struct tuv @{
17798 /* 0 | 4 */ int a1;
17799 /* XXX 4-byte hole */
17800 /* 8 | 8 */ char *a2;
17801 /* 16 | 4 */ int a3;
17803 /* total size (bytes): 24 */
17807 Notice the format of the first column of comments. There, you can
17808 find two parts separated by the @samp{|} character: the @emph{offset},
17809 which indicates where the field is located inside the struct, in
17810 bytes, and the @emph{size} of the field. Another interesting line is
17811 the marker of a @emph{hole} in the struct, indicating that it may be
17812 possible to pack the struct and make it use less space by reorganizing
17815 It is also possible to print offsets inside an union:
17818 (@value{GDBP}) ptype /o union qwe
17819 /* offset | size */ type = union qwe @{
17820 /* 24 */ struct tuv @{
17821 /* 0 | 4 */ int a1;
17822 /* XXX 4-byte hole */
17823 /* 8 | 8 */ char *a2;
17824 /* 16 | 4 */ int a3;
17826 /* total size (bytes): 24 */
17828 /* 40 */ struct xyz @{
17829 /* 0 | 4 */ int f1;
17830 /* 4 | 1 */ char f2;
17831 /* XXX 3-byte hole */
17832 /* 8 | 8 */ void *f3;
17833 /* 16 | 24 */ struct tuv @{
17834 /* 16 | 4 */ int a1;
17835 /* XXX 4-byte hole */
17836 /* 24 | 8 */ char *a2;
17837 /* 32 | 4 */ int a3;
17839 /* total size (bytes): 24 */
17842 /* total size (bytes): 40 */
17845 /* total size (bytes): 40 */
17849 In this case, since @code{struct tuv} and @code{struct xyz} occupy the
17850 same space (because we are dealing with an union), the offset is not
17851 printed for them. However, you can still examine the offset of each
17852 of these structures' fields.
17854 Another useful scenario is printing the offsets of a struct containing
17858 (@value{GDBP}) ptype /o struct tyu
17859 /* offset | size */ type = struct tyu @{
17860 /* 0:31 | 4 */ int a1 : 1;
17861 /* 0:28 | 4 */ int a2 : 3;
17862 /* 0: 5 | 4 */ int a3 : 23;
17863 /* 3: 3 | 1 */ signed char a4 : 2;
17864 /* XXX 3-bit hole */
17865 /* XXX 4-byte hole */
17866 /* 8 | 8 */ int64_t a5;
17867 /* 16: 0 | 4 */ int a6 : 5;
17868 /* 16: 5 | 8 */ int64_t a7 : 3;
17869 "/* XXX 7-byte padding */
17871 /* total size (bytes): 24 */
17875 Note how the offset information is now extended to also include the
17876 first bit of the bitfield.
17880 @item ptype[/@var{flags}] [@var{arg}]
17881 @code{ptype} accepts the same arguments as @code{whatis}, but prints a
17882 detailed description of the type, instead of just the name of the type.
17883 @xref{Expressions, ,Expressions}.
17885 Contrary to @code{whatis}, @code{ptype} always unrolls any
17886 @code{typedef}s in its argument declaration, whether the argument is
17887 a variable, expression, or a data type. This means that @code{ptype}
17888 of a variable or an expression will not print literally its type as
17889 present in the source code---use @code{whatis} for that. @code{typedef}s at
17890 the pointer or reference targets are also unrolled. Only @code{typedef}s of
17891 fields, methods and inner @code{class typedef}s of @code{struct}s,
17892 @code{class}es and @code{union}s are not unrolled even with @code{ptype}.
17894 For example, for this variable declaration:
17897 typedef double real_t;
17898 struct complex @{ real_t real; double imag; @};
17899 typedef struct complex complex_t;
17901 real_t *real_pointer_var;
17905 the two commands give this output:
17909 (@value{GDBP}) whatis var
17911 (@value{GDBP}) ptype var
17912 type = struct complex @{
17916 (@value{GDBP}) whatis complex_t
17917 type = struct complex
17918 (@value{GDBP}) whatis struct complex
17919 type = struct complex
17920 (@value{GDBP}) ptype struct complex
17921 type = struct complex @{
17925 (@value{GDBP}) whatis real_pointer_var
17927 (@value{GDBP}) ptype real_pointer_var
17933 As with @code{whatis}, using @code{ptype} without an argument refers to
17934 the type of @code{$}, the last value in the value history.
17936 @cindex incomplete type
17937 Sometimes, programs use opaque data types or incomplete specifications
17938 of complex data structure. If the debug information included in the
17939 program does not allow @value{GDBN} to display a full declaration of
17940 the data type, it will say @samp{<incomplete type>}. For example,
17941 given these declarations:
17945 struct foo *fooptr;
17949 but no definition for @code{struct foo} itself, @value{GDBN} will say:
17952 (@value{GDBP}) ptype foo
17953 $1 = <incomplete type>
17957 ``Incomplete type'' is C terminology for data types that are not
17958 completely specified.
17960 @cindex unknown type
17961 Othertimes, information about a variable's type is completely absent
17962 from the debug information included in the program. This most often
17963 happens when the program or library where the variable is defined
17964 includes no debug information at all. @value{GDBN} knows the variable
17965 exists from inspecting the linker/loader symbol table (e.g., the ELF
17966 dynamic symbol table), but such symbols do not contain type
17967 information. Inspecting the type of a (global) variable for which
17968 @value{GDBN} has no type information shows:
17971 (@value{GDBP}) ptype var
17972 type = <data variable, no debug info>
17975 @xref{Variables, no debug info variables}, for how to print the values
17979 @item info types @var{regexp}
17981 Print a brief description of all types whose names match the regular
17982 expression @var{regexp} (or all types in your program, if you supply
17983 no argument). Each complete typename is matched as though it were a
17984 complete line; thus, @samp{i type value} gives information on all
17985 types in your program whose names include the string @code{value}, but
17986 @samp{i type ^value$} gives information only on types whose complete
17987 name is @code{value}.
17989 In programs using different languages, @value{GDBN} chooses the syntax
17990 to print the type description according to the
17991 @samp{set language} value: using @samp{set language auto}
17992 (see @ref{Automatically, ,Set Language Automatically}) means to use the
17993 language of the type, other values mean to use
17994 the manually specified language (see @ref{Manually, ,Set Language Manually}).
17996 This command differs from @code{ptype} in two ways: first, like
17997 @code{whatis}, it does not print a detailed description; second, it
17998 lists all source files and line numbers where a type is defined.
18000 @kindex info type-printers
18001 @item info type-printers
18002 Versions of @value{GDBN} that ship with Python scripting enabled may
18003 have ``type printers'' available. When using @command{ptype} or
18004 @command{whatis}, these printers are consulted when the name of a type
18005 is needed. @xref{Type Printing API}, for more information on writing
18008 @code{info type-printers} displays all the available type printers.
18010 @kindex enable type-printer
18011 @kindex disable type-printer
18012 @item enable type-printer @var{name}@dots{}
18013 @item disable type-printer @var{name}@dots{}
18014 These commands can be used to enable or disable type printers.
18017 @cindex local variables
18018 @item info scope @var{location}
18019 List all the variables local to a particular scope. This command
18020 accepts a @var{location} argument---a function name, a source line, or
18021 an address preceded by a @samp{*}, and prints all the variables local
18022 to the scope defined by that location. (@xref{Specify Location}, for
18023 details about supported forms of @var{location}.) For example:
18026 (@value{GDBP}) @b{info scope command_line_handler}
18027 Scope for command_line_handler:
18028 Symbol rl is an argument at stack/frame offset 8, length 4.
18029 Symbol linebuffer is in static storage at address 0x150a18, length 4.
18030 Symbol linelength is in static storage at address 0x150a1c, length 4.
18031 Symbol p is a local variable in register $esi, length 4.
18032 Symbol p1 is a local variable in register $ebx, length 4.
18033 Symbol nline is a local variable in register $edx, length 4.
18034 Symbol repeat is a local variable at frame offset -8, length 4.
18038 This command is especially useful for determining what data to collect
18039 during a @dfn{trace experiment}, see @ref{Tracepoint Actions,
18042 @kindex info source
18044 Show information about the current source file---that is, the source file for
18045 the function containing the current point of execution:
18048 the name of the source file, and the directory containing it,
18050 the directory it was compiled in,
18052 its length, in lines,
18054 which programming language it is written in,
18056 if the debug information provides it, the program that compiled the file
18057 (which may include, e.g., the compiler version and command line arguments),
18059 whether the executable includes debugging information for that file, and
18060 if so, what format the information is in (e.g., STABS, Dwarf 2, etc.), and
18062 whether the debugging information includes information about
18063 preprocessor macros.
18067 @kindex info sources
18069 Print the names of all source files in your program for which there is
18070 debugging information, organized into two lists: files whose symbols
18071 have already been read, and files whose symbols will be read when needed.
18073 @kindex info functions
18074 @item info functions [-q]
18075 Print the names and data types of all defined functions.
18076 Similarly to @samp{info types}, this command groups its output by source
18077 files and annotates each function definition with its source line
18080 In programs using different languages, @value{GDBN} chooses the syntax
18081 to print the function name and type according to the
18082 @samp{set language} value: using @samp{set language auto}
18083 (see @ref{Automatically, ,Set Language Automatically}) means to use the
18084 language of the function, other values mean to use
18085 the manually specified language (see @ref{Manually, ,Set Language Manually}).
18087 The optional flag @samp{-q}, which stands for @samp{quiet}, disables
18088 printing header information and messages explaining why no functions
18091 @item info functions [-q] [-t @var{type_regexp}] [@var{regexp}]
18092 Like @samp{info functions}, but only print the names and data types
18093 of the functions selected with the provided regexp(s).
18095 If @var{regexp} is provided, print only the functions whose names
18096 match the regular expression @var{regexp}.
18097 Thus, @samp{info fun step} finds all functions whose
18098 names include @code{step}; @samp{info fun ^step} finds those whose names
18099 start with @code{step}. If a function name contains characters that
18100 conflict with the regular expression language (e.g.@:
18101 @samp{operator*()}), they may be quoted with a backslash.
18103 If @var{type_regexp} is provided, print only the functions whose
18104 types, as printed by the @code{whatis} command, match
18105 the regular expression @var{type_regexp}.
18106 If @var{type_regexp} contains space(s), it should be enclosed in
18107 quote characters. If needed, use backslash to escape the meaning
18108 of special characters or quotes.
18109 Thus, @samp{info fun -t '^int ('} finds the functions that return
18110 an integer; @samp{info fun -t '(.*int.*'} finds the functions that
18111 have an argument type containing int; @samp{info fun -t '^int (' ^step}
18112 finds the functions whose names start with @code{step} and that return
18115 If both @var{regexp} and @var{type_regexp} are provided, a function
18116 is printed only if its name matches @var{regexp} and its type matches
18120 @kindex info variables
18121 @item info variables [-q]
18122 Print the names and data types of all variables that are defined
18123 outside of functions (i.e.@: excluding local variables).
18124 The printed variables are grouped by source files and annotated with
18125 their respective source line numbers.
18127 In programs using different languages, @value{GDBN} chooses the syntax
18128 to print the variable name and type according to the
18129 @samp{set language} value: using @samp{set language auto}
18130 (see @ref{Automatically, ,Set Language Automatically}) means to use the
18131 language of the variable, other values mean to use
18132 the manually specified language (see @ref{Manually, ,Set Language Manually}).
18134 The optional flag @samp{-q}, which stands for @samp{quiet}, disables
18135 printing header information and messages explaining why no variables
18138 @item info variables [-q] [-t @var{type_regexp}] [@var{regexp}]
18139 Like @kbd{info variables}, but only print the variables selected
18140 with the provided regexp(s).
18142 If @var{regexp} is provided, print only the variables whose names
18143 match the regular expression @var{regexp}.
18145 If @var{type_regexp} is provided, print only the variables whose
18146 types, as printed by the @code{whatis} command, match
18147 the regular expression @var{type_regexp}.
18148 If @var{type_regexp} contains space(s), it should be enclosed in
18149 quote characters. If needed, use backslash to escape the meaning
18150 of special characters or quotes.
18152 If both @var{regexp} and @var{type_regexp} are provided, an argument
18153 is printed only if its name matches @var{regexp} and its type matches
18156 @kindex info classes
18157 @cindex Objective-C, classes and selectors
18159 @itemx info classes @var{regexp}
18160 Display all Objective-C classes in your program, or
18161 (with the @var{regexp} argument) all those matching a particular regular
18164 @kindex info selectors
18165 @item info selectors
18166 @itemx info selectors @var{regexp}
18167 Display all Objective-C selectors in your program, or
18168 (with the @var{regexp} argument) all those matching a particular regular
18172 This was never implemented.
18173 @kindex info methods
18175 @itemx info methods @var{regexp}
18176 The @code{info methods} command permits the user to examine all defined
18177 methods within C@t{++} program, or (with the @var{regexp} argument) a
18178 specific set of methods found in the various C@t{++} classes. Many
18179 C@t{++} classes provide a large number of methods. Thus, the output
18180 from the @code{ptype} command can be overwhelming and hard to use. The
18181 @code{info-methods} command filters the methods, printing only those
18182 which match the regular-expression @var{regexp}.
18185 @cindex opaque data types
18186 @kindex set opaque-type-resolution
18187 @item set opaque-type-resolution on
18188 Tell @value{GDBN} to resolve opaque types. An opaque type is a type
18189 declared as a pointer to a @code{struct}, @code{class}, or
18190 @code{union}---for example, @code{struct MyType *}---that is used in one
18191 source file although the full declaration of @code{struct MyType} is in
18192 another source file. The default is on.
18194 A change in the setting of this subcommand will not take effect until
18195 the next time symbols for a file are loaded.
18197 @item set opaque-type-resolution off
18198 Tell @value{GDBN} not to resolve opaque types. In this case, the type
18199 is printed as follows:
18201 @{<no data fields>@}
18204 @kindex show opaque-type-resolution
18205 @item show opaque-type-resolution
18206 Show whether opaque types are resolved or not.
18208 @kindex set print symbol-loading
18209 @cindex print messages when symbols are loaded
18210 @item set print symbol-loading
18211 @itemx set print symbol-loading full
18212 @itemx set print symbol-loading brief
18213 @itemx set print symbol-loading off
18214 The @code{set print symbol-loading} command allows you to control the
18215 printing of messages when @value{GDBN} loads symbol information.
18216 By default a message is printed for the executable and one for each
18217 shared library, and normally this is what you want. However, when
18218 debugging apps with large numbers of shared libraries these messages
18220 When set to @code{brief} a message is printed for each executable,
18221 and when @value{GDBN} loads a collection of shared libraries at once
18222 it will only print one message regardless of the number of shared
18223 libraries. When set to @code{off} no messages are printed.
18225 @kindex show print symbol-loading
18226 @item show print symbol-loading
18227 Show whether messages will be printed when a @value{GDBN} command
18228 entered from the keyboard causes symbol information to be loaded.
18230 @kindex maint print symbols
18231 @cindex symbol dump
18232 @kindex maint print psymbols
18233 @cindex partial symbol dump
18234 @kindex maint print msymbols
18235 @cindex minimal symbol dump
18236 @item maint print symbols @r{[}-pc @var{address}@r{]} @r{[}@var{filename}@r{]}
18237 @itemx maint print symbols @r{[}-objfile @var{objfile}@r{]} @r{[}-source @var{source}@r{]} @r{[}--@r{]} @r{[}@var{filename}@r{]}
18238 @itemx maint print psymbols @r{[}-objfile @var{objfile}@r{]} @r{[}-pc @var{address}@r{]} @r{[}--@r{]} @r{[}@var{filename}@r{]}
18239 @itemx maint print psymbols @r{[}-objfile @var{objfile}@r{]} @r{[}-source @var{source}@r{]} @r{[}--@r{]} @r{[}@var{filename}@r{]}
18240 @itemx maint print msymbols @r{[}-objfile @var{objfile}@r{]} @r{[}--@r{]} @r{[}@var{filename}@r{]}
18241 Write a dump of debugging symbol data into the file @var{filename} or
18242 the terminal if @var{filename} is unspecified.
18243 If @code{-objfile @var{objfile}} is specified, only dump symbols for
18245 If @code{-pc @var{address}} is specified, only dump symbols for the file
18246 with code at that address. Note that @var{address} may be a symbol like
18248 If @code{-source @var{source}} is specified, only dump symbols for that
18251 These commands are used to debug the @value{GDBN} symbol-reading code.
18252 These commands do not modify internal @value{GDBN} state, therefore
18253 @samp{maint print symbols} will only print symbols for already expanded symbol
18255 You can use the command @code{info sources} to find out which files these are.
18256 If you use @samp{maint print psymbols} instead, the dump shows information
18257 about symbols that @value{GDBN} only knows partially---that is, symbols
18258 defined in files that @value{GDBN} has skimmed, but not yet read completely.
18259 Finally, @samp{maint print msymbols} just dumps ``minimal symbols'', e.g.,
18262 @xref{Files, ,Commands to Specify Files}, for a discussion of how
18263 @value{GDBN} reads symbols (in the description of @code{symbol-file}).
18265 @kindex maint info symtabs
18266 @kindex maint info psymtabs
18267 @cindex listing @value{GDBN}'s internal symbol tables
18268 @cindex symbol tables, listing @value{GDBN}'s internal
18269 @cindex full symbol tables, listing @value{GDBN}'s internal
18270 @cindex partial symbol tables, listing @value{GDBN}'s internal
18271 @item maint info symtabs @r{[} @var{regexp} @r{]}
18272 @itemx maint info psymtabs @r{[} @var{regexp} @r{]}
18274 List the @code{struct symtab} or @code{struct partial_symtab}
18275 structures whose names match @var{regexp}. If @var{regexp} is not
18276 given, list them all. The output includes expressions which you can
18277 copy into a @value{GDBN} debugging this one to examine a particular
18278 structure in more detail. For example:
18281 (@value{GDBP}) maint info psymtabs dwarf2read
18282 @{ objfile /home/gnu/build/gdb/gdb
18283 ((struct objfile *) 0x82e69d0)
18284 @{ psymtab /home/gnu/src/gdb/dwarf2read.c
18285 ((struct partial_symtab *) 0x8474b10)
18288 text addresses 0x814d3c8 -- 0x8158074
18289 globals (* (struct partial_symbol **) 0x8507a08 @@ 9)
18290 statics (* (struct partial_symbol **) 0x40e95b78 @@ 2882)
18291 dependencies (none)
18294 (@value{GDBP}) maint info symtabs
18298 We see that there is one partial symbol table whose filename contains
18299 the string @samp{dwarf2read}, belonging to the @samp{gdb} executable;
18300 and we see that @value{GDBN} has not read in any symtabs yet at all.
18301 If we set a breakpoint on a function, that will cause @value{GDBN} to
18302 read the symtab for the compilation unit containing that function:
18305 (@value{GDBP}) break dwarf2_psymtab_to_symtab
18306 Breakpoint 1 at 0x814e5da: file /home/gnu/src/gdb/dwarf2read.c,
18308 (@value{GDBP}) maint info symtabs
18309 @{ objfile /home/gnu/build/gdb/gdb
18310 ((struct objfile *) 0x82e69d0)
18311 @{ symtab /home/gnu/src/gdb/dwarf2read.c
18312 ((struct symtab *) 0x86c1f38)
18315 blockvector ((struct blockvector *) 0x86c1bd0) (primary)
18316 linetable ((struct linetable *) 0x8370fa0)
18317 debugformat DWARF 2
18323 @kindex maint info line-table
18324 @cindex listing @value{GDBN}'s internal line tables
18325 @cindex line tables, listing @value{GDBN}'s internal
18326 @item maint info line-table @r{[} @var{regexp} @r{]}
18328 List the @code{struct linetable} from all @code{struct symtab}
18329 instances whose name matches @var{regexp}. If @var{regexp} is not
18330 given, list the @code{struct linetable} from all @code{struct symtab}.
18332 @kindex maint set symbol-cache-size
18333 @cindex symbol cache size
18334 @item maint set symbol-cache-size @var{size}
18335 Set the size of the symbol cache to @var{size}.
18336 The default size is intended to be good enough for debugging
18337 most applications. This option exists to allow for experimenting
18338 with different sizes.
18340 @kindex maint show symbol-cache-size
18341 @item maint show symbol-cache-size
18342 Show the size of the symbol cache.
18344 @kindex maint print symbol-cache
18345 @cindex symbol cache, printing its contents
18346 @item maint print symbol-cache
18347 Print the contents of the symbol cache.
18348 This is useful when debugging symbol cache issues.
18350 @kindex maint print symbol-cache-statistics
18351 @cindex symbol cache, printing usage statistics
18352 @item maint print symbol-cache-statistics
18353 Print symbol cache usage statistics.
18354 This helps determine how well the cache is being utilized.
18356 @kindex maint flush-symbol-cache
18357 @cindex symbol cache, flushing
18358 @item maint flush-symbol-cache
18359 Flush the contents of the symbol cache, all entries are removed.
18360 This command is useful when debugging the symbol cache.
18361 It is also useful when collecting performance data.
18366 @chapter Altering Execution
18368 Once you think you have found an error in your program, you might want to
18369 find out for certain whether correcting the apparent error would lead to
18370 correct results in the rest of the run. You can find the answer by
18371 experiment, using the @value{GDBN} features for altering execution of the
18374 For example, you can store new values into variables or memory
18375 locations, give your program a signal, restart it at a different
18376 address, or even return prematurely from a function.
18379 * Assignment:: Assignment to variables
18380 * Jumping:: Continuing at a different address
18381 * Signaling:: Giving your program a signal
18382 * Returning:: Returning from a function
18383 * Calling:: Calling your program's functions
18384 * Patching:: Patching your program
18385 * Compiling and Injecting Code:: Compiling and injecting code in @value{GDBN}
18389 @section Assignment to Variables
18392 @cindex setting variables
18393 To alter the value of a variable, evaluate an assignment expression.
18394 @xref{Expressions, ,Expressions}. For example,
18401 stores the value 4 into the variable @code{x}, and then prints the
18402 value of the assignment expression (which is 4).
18403 @xref{Languages, ,Using @value{GDBN} with Different Languages}, for more
18404 information on operators in supported languages.
18406 @kindex set variable
18407 @cindex variables, setting
18408 If you are not interested in seeing the value of the assignment, use the
18409 @code{set} command instead of the @code{print} command. @code{set} is
18410 really the same as @code{print} except that the expression's value is
18411 not printed and is not put in the value history (@pxref{Value History,
18412 ,Value History}). The expression is evaluated only for its effects.
18414 If the beginning of the argument string of the @code{set} command
18415 appears identical to a @code{set} subcommand, use the @code{set
18416 variable} command instead of just @code{set}. This command is identical
18417 to @code{set} except for its lack of subcommands. For example, if your
18418 program has a variable @code{width}, you get an error if you try to set
18419 a new value with just @samp{set width=13}, because @value{GDBN} has the
18420 command @code{set width}:
18423 (@value{GDBP}) whatis width
18425 (@value{GDBP}) p width
18427 (@value{GDBP}) set width=47
18428 Invalid syntax in expression.
18432 The invalid expression, of course, is @samp{=47}. In
18433 order to actually set the program's variable @code{width}, use
18436 (@value{GDBP}) set var width=47
18439 Because the @code{set} command has many subcommands that can conflict
18440 with the names of program variables, it is a good idea to use the
18441 @code{set variable} command instead of just @code{set}. For example, if
18442 your program has a variable @code{g}, you run into problems if you try
18443 to set a new value with just @samp{set g=4}, because @value{GDBN} has
18444 the command @code{set gnutarget}, abbreviated @code{set g}:
18448 (@value{GDBP}) whatis g
18452 (@value{GDBP}) set g=4
18456 The program being debugged has been started already.
18457 Start it from the beginning? (y or n) y
18458 Starting program: /home/smith/cc_progs/a.out
18459 "/home/smith/cc_progs/a.out": can't open to read symbols:
18460 Invalid bfd target.
18461 (@value{GDBP}) show g
18462 The current BFD target is "=4".
18467 The program variable @code{g} did not change, and you silently set the
18468 @code{gnutarget} to an invalid value. In order to set the variable
18472 (@value{GDBP}) set var g=4
18475 @value{GDBN} allows more implicit conversions in assignments than C; you can
18476 freely store an integer value into a pointer variable or vice versa,
18477 and you can convert any structure to any other structure that is the
18478 same length or shorter.
18479 @comment FIXME: how do structs align/pad in these conversions?
18480 @comment /doc@cygnus.com 18dec1990
18482 To store values into arbitrary places in memory, use the @samp{@{@dots{}@}}
18483 construct to generate a value of specified type at a specified address
18484 (@pxref{Expressions, ,Expressions}). For example, @code{@{int@}0x83040} refers
18485 to memory location @code{0x83040} as an integer (which implies a certain size
18486 and representation in memory), and
18489 set @{int@}0x83040 = 4
18493 stores the value 4 into that memory location.
18496 @section Continuing at a Different Address
18498 Ordinarily, when you continue your program, you do so at the place where
18499 it stopped, with the @code{continue} command. You can instead continue at
18500 an address of your own choosing, with the following commands:
18504 @kindex j @r{(@code{jump})}
18505 @item jump @var{location}
18506 @itemx j @var{location}
18507 Resume execution at @var{location}. Execution stops again immediately
18508 if there is a breakpoint there. @xref{Specify Location}, for a description
18509 of the different forms of @var{location}. It is common
18510 practice to use the @code{tbreak} command in conjunction with
18511 @code{jump}. @xref{Set Breaks, ,Setting Breakpoints}.
18513 The @code{jump} command does not change the current stack frame, or
18514 the stack pointer, or the contents of any memory location or any
18515 register other than the program counter. If @var{location} is in
18516 a different function from the one currently executing, the results may
18517 be bizarre if the two functions expect different patterns of arguments or
18518 of local variables. For this reason, the @code{jump} command requests
18519 confirmation if the specified line is not in the function currently
18520 executing. However, even bizarre results are predictable if you are
18521 well acquainted with the machine-language code of your program.
18524 On many systems, you can get much the same effect as the @code{jump}
18525 command by storing a new value into the register @code{$pc}. The
18526 difference is that this does not start your program running; it only
18527 changes the address of where it @emph{will} run when you continue. For
18535 makes the next @code{continue} command or stepping command execute at
18536 address @code{0x485}, rather than at the address where your program stopped.
18537 @xref{Continuing and Stepping, ,Continuing and Stepping}.
18539 The most common occasion to use the @code{jump} command is to back
18540 up---perhaps with more breakpoints set---over a portion of a program
18541 that has already executed, in order to examine its execution in more
18546 @section Giving your Program a Signal
18547 @cindex deliver a signal to a program
18551 @item signal @var{signal}
18552 Resume execution where your program is stopped, but immediately give it the
18553 signal @var{signal}. The @var{signal} can be the name or the number of a
18554 signal. For example, on many systems @code{signal 2} and @code{signal
18555 SIGINT} are both ways of sending an interrupt signal.
18557 Alternatively, if @var{signal} is zero, continue execution without
18558 giving a signal. This is useful when your program stopped on account of
18559 a signal and would ordinarily see the signal when resumed with the
18560 @code{continue} command; @samp{signal 0} causes it to resume without a
18563 @emph{Note:} When resuming a multi-threaded program, @var{signal} is
18564 delivered to the currently selected thread, not the thread that last
18565 reported a stop. This includes the situation where a thread was
18566 stopped due to a signal. So if you want to continue execution
18567 suppressing the signal that stopped a thread, you should select that
18568 same thread before issuing the @samp{signal 0} command. If you issue
18569 the @samp{signal 0} command with another thread as the selected one,
18570 @value{GDBN} detects that and asks for confirmation.
18572 Invoking the @code{signal} command is not the same as invoking the
18573 @code{kill} utility from the shell. Sending a signal with @code{kill}
18574 causes @value{GDBN} to decide what to do with the signal depending on
18575 the signal handling tables (@pxref{Signals}). The @code{signal} command
18576 passes the signal directly to your program.
18578 @code{signal} does not repeat when you press @key{RET} a second time
18579 after executing the command.
18581 @kindex queue-signal
18582 @item queue-signal @var{signal}
18583 Queue @var{signal} to be delivered immediately to the current thread
18584 when execution of the thread resumes. The @var{signal} can be the name or
18585 the number of a signal. For example, on many systems @code{signal 2} and
18586 @code{signal SIGINT} are both ways of sending an interrupt signal.
18587 The handling of the signal must be set to pass the signal to the program,
18588 otherwise @value{GDBN} will report an error.
18589 You can control the handling of signals from @value{GDBN} with the
18590 @code{handle} command (@pxref{Signals}).
18592 Alternatively, if @var{signal} is zero, any currently queued signal
18593 for the current thread is discarded and when execution resumes no signal
18594 will be delivered. This is useful when your program stopped on account
18595 of a signal and would ordinarily see the signal when resumed with the
18596 @code{continue} command.
18598 This command differs from the @code{signal} command in that the signal
18599 is just queued, execution is not resumed. And @code{queue-signal} cannot
18600 be used to pass a signal whose handling state has been set to @code{nopass}
18605 @xref{stepping into signal handlers}, for information on how stepping
18606 commands behave when the thread has a signal queued.
18609 @section Returning from a Function
18612 @cindex returning from a function
18615 @itemx return @var{expression}
18616 You can cancel execution of a function call with the @code{return}
18617 command. If you give an
18618 @var{expression} argument, its value is used as the function's return
18622 When you use @code{return}, @value{GDBN} discards the selected stack frame
18623 (and all frames within it). You can think of this as making the
18624 discarded frame return prematurely. If you wish to specify a value to
18625 be returned, give that value as the argument to @code{return}.
18627 This pops the selected stack frame (@pxref{Selection, ,Selecting a
18628 Frame}), and any other frames inside of it, leaving its caller as the
18629 innermost remaining frame. That frame becomes selected. The
18630 specified value is stored in the registers used for returning values
18633 The @code{return} command does not resume execution; it leaves the
18634 program stopped in the state that would exist if the function had just
18635 returned. In contrast, the @code{finish} command (@pxref{Continuing
18636 and Stepping, ,Continuing and Stepping}) resumes execution until the
18637 selected stack frame returns naturally.
18639 @value{GDBN} needs to know how the @var{expression} argument should be set for
18640 the inferior. The concrete registers assignment depends on the OS ABI and the
18641 type being returned by the selected stack frame. For example it is common for
18642 OS ABI to return floating point values in FPU registers while integer values in
18643 CPU registers. Still some ABIs return even floating point values in CPU
18644 registers. Larger integer widths (such as @code{long long int}) also have
18645 specific placement rules. @value{GDBN} already knows the OS ABI from its
18646 current target so it needs to find out also the type being returned to make the
18647 assignment into the right register(s).
18649 Normally, the selected stack frame has debug info. @value{GDBN} will always
18650 use the debug info instead of the implicit type of @var{expression} when the
18651 debug info is available. For example, if you type @kbd{return -1}, and the
18652 function in the current stack frame is declared to return a @code{long long
18653 int}, @value{GDBN} transparently converts the implicit @code{int} value of -1
18654 into a @code{long long int}:
18657 Breakpoint 1, func () at gdb.base/return-nodebug.c:29
18659 (@value{GDBP}) return -1
18660 Make func return now? (y or n) y
18661 #0 0x004004f6 in main () at gdb.base/return-nodebug.c:43
18662 43 printf ("result=%lld\n", func ());
18666 However, if the selected stack frame does not have a debug info, e.g., if the
18667 function was compiled without debug info, @value{GDBN} has to find out the type
18668 to return from user. Specifying a different type by mistake may set the value
18669 in different inferior registers than the caller code expects. For example,
18670 typing @kbd{return -1} with its implicit type @code{int} would set only a part
18671 of a @code{long long int} result for a debug info less function (on 32-bit
18672 architectures). Therefore the user is required to specify the return type by
18673 an appropriate cast explicitly:
18676 Breakpoint 2, 0x0040050b in func ()
18677 (@value{GDBP}) return -1
18678 Return value type not available for selected stack frame.
18679 Please use an explicit cast of the value to return.
18680 (@value{GDBP}) return (long long int) -1
18681 Make selected stack frame return now? (y or n) y
18682 #0 0x00400526 in main ()
18687 @section Calling Program Functions
18690 @cindex calling functions
18691 @cindex inferior functions, calling
18692 @item print @var{expr}
18693 Evaluate the expression @var{expr} and display the resulting value.
18694 The expression may include calls to functions in the program being
18698 @item call @var{expr}
18699 Evaluate the expression @var{expr} without displaying @code{void}
18702 You can use this variant of the @code{print} command if you want to
18703 execute a function from your program that does not return anything
18704 (a.k.a.@: @dfn{a void function}), but without cluttering the output
18705 with @code{void} returned values that @value{GDBN} will otherwise
18706 print. If the result is not void, it is printed and saved in the
18710 It is possible for the function you call via the @code{print} or
18711 @code{call} command to generate a signal (e.g., if there's a bug in
18712 the function, or if you passed it incorrect arguments). What happens
18713 in that case is controlled by the @code{set unwindonsignal} command.
18715 Similarly, with a C@t{++} program it is possible for the function you
18716 call via the @code{print} or @code{call} command to generate an
18717 exception that is not handled due to the constraints of the dummy
18718 frame. In this case, any exception that is raised in the frame, but has
18719 an out-of-frame exception handler will not be found. GDB builds a
18720 dummy-frame for the inferior function call, and the unwinder cannot
18721 seek for exception handlers outside of this dummy-frame. What happens
18722 in that case is controlled by the
18723 @code{set unwind-on-terminating-exception} command.
18726 @item set unwindonsignal
18727 @kindex set unwindonsignal
18728 @cindex unwind stack in called functions
18729 @cindex call dummy stack unwinding
18730 Set unwinding of the stack if a signal is received while in a function
18731 that @value{GDBN} called in the program being debugged. If set to on,
18732 @value{GDBN} unwinds the stack it created for the call and restores
18733 the context to what it was before the call. If set to off (the
18734 default), @value{GDBN} stops in the frame where the signal was
18737 @item show unwindonsignal
18738 @kindex show unwindonsignal
18739 Show the current setting of stack unwinding in the functions called by
18742 @item set unwind-on-terminating-exception
18743 @kindex set unwind-on-terminating-exception
18744 @cindex unwind stack in called functions with unhandled exceptions
18745 @cindex call dummy stack unwinding on unhandled exception.
18746 Set unwinding of the stack if a C@t{++} exception is raised, but left
18747 unhandled while in a function that @value{GDBN} called in the program being
18748 debugged. If set to on (the default), @value{GDBN} unwinds the stack
18749 it created for the call and restores the context to what it was before
18750 the call. If set to off, @value{GDBN} the exception is delivered to
18751 the default C@t{++} exception handler and the inferior terminated.
18753 @item show unwind-on-terminating-exception
18754 @kindex show unwind-on-terminating-exception
18755 Show the current setting of stack unwinding in the functions called by
18758 @item set may-call-functions
18759 @kindex set may-call-functions
18760 @cindex disabling calling functions in the program
18761 @cindex calling functions in the program, disabling
18762 Set permission to call functions in the program.
18763 This controls whether @value{GDBN} will attempt to call functions in
18764 the program, such as with expressions in the @code{print} command. It
18765 defaults to @code{on}.
18767 To call a function in the program, @value{GDBN} has to temporarily
18768 modify the state of the inferior. This has potentially undesired side
18769 effects. Also, having @value{GDBN} call nested functions is likely to
18770 be erroneous and may even crash the program being debugged. You can
18771 avoid such hazards by forbidding @value{GDBN} from calling functions
18772 in the program being debugged. If calling functions in the program
18773 is forbidden, GDB will throw an error when a command (such as printing
18774 an expression) starts a function call in the program.
18776 @item show may-call-functions
18777 @kindex show may-call-functions
18778 Show permission to call functions in the program.
18782 @subsection Calling functions with no debug info
18784 @cindex no debug info functions
18785 Sometimes, a function you wish to call is missing debug information.
18786 In such case, @value{GDBN} does not know the type of the function,
18787 including the types of the function's parameters. To avoid calling
18788 the inferior function incorrectly, which could result in the called
18789 function functioning erroneously and even crash, @value{GDBN} refuses
18790 to call the function unless you tell it the type of the function.
18792 For prototyped (i.e.@: ANSI/ISO style) functions, there are two ways
18793 to do that. The simplest is to cast the call to the function's
18794 declared return type. For example:
18797 (@value{GDBP}) p getenv ("PATH")
18798 'getenv' has unknown return type; cast the call to its declared return type
18799 (@value{GDBP}) p (char *) getenv ("PATH")
18800 $1 = 0x7fffffffe7ba "/usr/local/bin:/"...
18803 Casting the return type of a no-debug function is equivalent to
18804 casting the function to a pointer to a prototyped function that has a
18805 prototype that matches the types of the passed-in arguments, and
18806 calling that. I.e., the call above is equivalent to:
18809 (@value{GDBP}) p ((char * (*) (const char *)) getenv) ("PATH")
18813 and given this prototyped C or C++ function with float parameters:
18816 float multiply (float v1, float v2) @{ return v1 * v2; @}
18820 these calls are equivalent:
18823 (@value{GDBP}) p (float) multiply (2.0f, 3.0f)
18824 (@value{GDBP}) p ((float (*) (float, float)) multiply) (2.0f, 3.0f)
18827 If the function you wish to call is declared as unprototyped (i.e.@:
18828 old K&R style), you must use the cast-to-function-pointer syntax, so
18829 that @value{GDBN} knows that it needs to apply default argument
18830 promotions (promote float arguments to double). @xref{ABI, float
18831 promotion}. For example, given this unprototyped C function with
18832 float parameters, and no debug info:
18836 multiply_noproto (v1, v2)
18844 you call it like this:
18847 (@value{GDBP}) p ((float (*) ()) multiply_noproto) (2.0f, 3.0f)
18851 @section Patching Programs
18853 @cindex patching binaries
18854 @cindex writing into executables
18855 @cindex writing into corefiles
18857 By default, @value{GDBN} opens the file containing your program's
18858 executable code (or the corefile) read-only. This prevents accidental
18859 alterations to machine code; but it also prevents you from intentionally
18860 patching your program's binary.
18862 If you'd like to be able to patch the binary, you can specify that
18863 explicitly with the @code{set write} command. For example, you might
18864 want to turn on internal debugging flags, or even to make emergency
18870 @itemx set write off
18871 If you specify @samp{set write on}, @value{GDBN} opens executable and
18872 core files for both reading and writing; if you specify @kbd{set write
18873 off} (the default), @value{GDBN} opens them read-only.
18875 If you have already loaded a file, you must load it again (using the
18876 @code{exec-file} or @code{core-file} command) after changing @code{set
18877 write}, for your new setting to take effect.
18881 Display whether executable files and core files are opened for writing
18882 as well as reading.
18885 @node Compiling and Injecting Code
18886 @section Compiling and injecting code in @value{GDBN}
18887 @cindex injecting code
18888 @cindex writing into executables
18889 @cindex compiling code
18891 @value{GDBN} supports on-demand compilation and code injection into
18892 programs running under @value{GDBN}. GCC 5.0 or higher built with
18893 @file{libcc1.so} must be installed for this functionality to be enabled.
18894 This functionality is implemented with the following commands.
18897 @kindex compile code
18898 @item compile code @var{source-code}
18899 @itemx compile code -raw @var{--} @var{source-code}
18900 Compile @var{source-code} with the compiler language found as the current
18901 language in @value{GDBN} (@pxref{Languages}). If compilation and
18902 injection is not supported with the current language specified in
18903 @value{GDBN}, or the compiler does not support this feature, an error
18904 message will be printed. If @var{source-code} compiles and links
18905 successfully, @value{GDBN} will load the object-code emitted,
18906 and execute it within the context of the currently selected inferior.
18907 It is important to note that the compiled code is executed immediately.
18908 After execution, the compiled code is removed from @value{GDBN} and any
18909 new types or variables you have defined will be deleted.
18911 The command allows you to specify @var{source-code} in two ways.
18912 The simplest method is to provide a single line of code to the command.
18916 compile code printf ("hello world\n");
18919 If you specify options on the command line as well as source code, they
18920 may conflict. The @samp{--} delimiter can be used to separate options
18921 from actual source code. E.g.:
18924 compile code -r -- printf ("hello world\n");
18927 Alternatively you can enter source code as multiple lines of text. To
18928 enter this mode, invoke the @samp{compile code} command without any text
18929 following the command. This will start the multiple-line editor and
18930 allow you to type as many lines of source code as required. When you
18931 have completed typing, enter @samp{end} on its own line to exit the
18936 >printf ("hello\n");
18937 >printf ("world\n");
18941 Specifying @samp{-raw}, prohibits @value{GDBN} from wrapping the
18942 provided @var{source-code} in a callable scope. In this case, you must
18943 specify the entry point of the code by defining a function named
18944 @code{_gdb_expr_}. The @samp{-raw} code cannot access variables of the
18945 inferior. Using @samp{-raw} option may be needed for example when
18946 @var{source-code} requires @samp{#include} lines which may conflict with
18947 inferior symbols otherwise.
18949 @kindex compile file
18950 @item compile file @var{filename}
18951 @itemx compile file -raw @var{filename}
18952 Like @code{compile code}, but take the source code from @var{filename}.
18955 compile file /home/user/example.c
18960 @item compile print @var{expr}
18961 @itemx compile print /@var{f} @var{expr}
18962 Compile and execute @var{expr} with the compiler language found as the
18963 current language in @value{GDBN} (@pxref{Languages}). By default the
18964 value of @var{expr} is printed in a format appropriate to its data type;
18965 you can choose a different format by specifying @samp{/@var{f}}, where
18966 @var{f} is a letter specifying the format; see @ref{Output Formats,,Output
18969 @item compile print
18970 @itemx compile print /@var{f}
18971 @cindex reprint the last value
18972 Alternatively you can enter the expression (source code producing it) as
18973 multiple lines of text. To enter this mode, invoke the @samp{compile print}
18974 command without any text following the command. This will start the
18975 multiple-line editor.
18979 The process of compiling and injecting the code can be inspected using:
18982 @anchor{set debug compile}
18983 @item set debug compile
18984 @cindex compile command debugging info
18985 Turns on or off display of @value{GDBN} process of compiling and
18986 injecting the code. The default is off.
18988 @item show debug compile
18989 Displays the current state of displaying @value{GDBN} process of
18990 compiling and injecting the code.
18992 @anchor{set debug compile-cplus-types}
18993 @item set debug compile-cplus-types
18994 @cindex compile C@t{++} type conversion
18995 Turns on or off the display of C@t{++} type conversion debugging information.
18996 The default is off.
18998 @item show debug compile-cplus-types
18999 Displays the current state of displaying debugging information for
19000 C@t{++} type conversion.
19003 @subsection Compilation options for the @code{compile} command
19005 @value{GDBN} needs to specify the right compilation options for the code
19006 to be injected, in part to make its ABI compatible with the inferior
19007 and in part to make the injected code compatible with @value{GDBN}'s
19011 The options used, in increasing precedence:
19014 @item target architecture and OS options (@code{gdbarch})
19015 These options depend on target processor type and target operating
19016 system, usually they specify at least 32-bit (@code{-m32}) or 64-bit
19017 (@code{-m64}) compilation option.
19019 @item compilation options recorded in the target
19020 @value{NGCC} (since version 4.7) stores the options used for compilation
19021 into @code{DW_AT_producer} part of DWARF debugging information according
19022 to the @value{NGCC} option @code{-grecord-gcc-switches}. One has to
19023 explicitly specify @code{-g} during inferior compilation otherwise
19024 @value{NGCC} produces no DWARF. This feature is only relevant for
19025 platforms where @code{-g} produces DWARF by default, otherwise one may
19026 try to enforce DWARF by using @code{-gdwarf-4}.
19028 @item compilation options set by @code{set compile-args}
19032 You can override compilation options using the following command:
19035 @item set compile-args
19036 @cindex compile command options override
19037 Set compilation options used for compiling and injecting code with the
19038 @code{compile} commands. These options override any conflicting ones
19039 from the target architecture and/or options stored during inferior
19042 @item show compile-args
19043 Displays the current state of compilation options override.
19044 This does not show all the options actually used during compilation,
19045 use @ref{set debug compile} for that.
19048 @subsection Caveats when using the @code{compile} command
19050 There are a few caveats to keep in mind when using the @code{compile}
19051 command. As the caveats are different per language, the table below
19052 highlights specific issues on a per language basis.
19055 @item C code examples and caveats
19056 When the language in @value{GDBN} is set to @samp{C}, the compiler will
19057 attempt to compile the source code with a @samp{C} compiler. The source
19058 code provided to the @code{compile} command will have much the same
19059 access to variables and types as it normally would if it were part of
19060 the program currently being debugged in @value{GDBN}.
19062 Below is a sample program that forms the basis of the examples that
19063 follow. This program has been compiled and loaded into @value{GDBN},
19064 much like any other normal debugging session.
19067 void function1 (void)
19070 printf ("function 1\n");
19073 void function2 (void)
19088 For the purposes of the examples in this section, the program above has
19089 been compiled, loaded into @value{GDBN}, stopped at the function
19090 @code{main}, and @value{GDBN} is awaiting input from the user.
19092 To access variables and types for any program in @value{GDBN}, the
19093 program must be compiled and packaged with debug information. The
19094 @code{compile} command is not an exception to this rule. Without debug
19095 information, you can still use the @code{compile} command, but you will
19096 be very limited in what variables and types you can access.
19098 So with that in mind, the example above has been compiled with debug
19099 information enabled. The @code{compile} command will have access to
19100 all variables and types (except those that may have been optimized
19101 out). Currently, as @value{GDBN} has stopped the program in the
19102 @code{main} function, the @code{compile} command would have access to
19103 the variable @code{k}. You could invoke the @code{compile} command
19104 and type some source code to set the value of @code{k}. You can also
19105 read it, or do anything with that variable you would normally do in
19106 @code{C}. Be aware that changes to inferior variables in the
19107 @code{compile} command are persistent. In the following example:
19110 compile code k = 3;
19114 the variable @code{k} is now 3. It will retain that value until
19115 something else in the example program changes it, or another
19116 @code{compile} command changes it.
19118 Normal scope and access rules apply to source code compiled and
19119 injected by the @code{compile} command. In the example, the variables
19120 @code{j} and @code{k} are not accessible yet, because the program is
19121 currently stopped in the @code{main} function, where these variables
19122 are not in scope. Therefore, the following command
19125 compile code j = 3;
19129 will result in a compilation error message.
19131 Once the program is continued, execution will bring these variables in
19132 scope, and they will become accessible; then the code you specify via
19133 the @code{compile} command will be able to access them.
19135 You can create variables and types with the @code{compile} command as
19136 part of your source code. Variables and types that are created as part
19137 of the @code{compile} command are not visible to the rest of the program for
19138 the duration of its run. This example is valid:
19141 compile code int ff = 5; printf ("ff is %d\n", ff);
19144 However, if you were to type the following into @value{GDBN} after that
19145 command has completed:
19148 compile code printf ("ff is %d\n'', ff);
19152 a compiler error would be raised as the variable @code{ff} no longer
19153 exists. Object code generated and injected by the @code{compile}
19154 command is removed when its execution ends. Caution is advised
19155 when assigning to program variables values of variables created by the
19156 code submitted to the @code{compile} command. This example is valid:
19159 compile code int ff = 5; k = ff;
19162 The value of the variable @code{ff} is assigned to @code{k}. The variable
19163 @code{k} does not require the existence of @code{ff} to maintain the value
19164 it has been assigned. However, pointers require particular care in
19165 assignment. If the source code compiled with the @code{compile} command
19166 changed the address of a pointer in the example program, perhaps to a
19167 variable created in the @code{compile} command, that pointer would point
19168 to an invalid location when the command exits. The following example
19169 would likely cause issues with your debugged program:
19172 compile code int ff = 5; p = &ff;
19175 In this example, @code{p} would point to @code{ff} when the
19176 @code{compile} command is executing the source code provided to it.
19177 However, as variables in the (example) program persist with their
19178 assigned values, the variable @code{p} would point to an invalid
19179 location when the command exists. A general rule should be followed
19180 in that you should either assign @code{NULL} to any assigned pointers,
19181 or restore a valid location to the pointer before the command exits.
19183 Similar caution must be exercised with any structs, unions, and typedefs
19184 defined in @code{compile} command. Types defined in the @code{compile}
19185 command will no longer be available in the next @code{compile} command.
19186 Therefore, if you cast a variable to a type defined in the
19187 @code{compile} command, care must be taken to ensure that any future
19188 need to resolve the type can be achieved.
19191 (gdb) compile code static struct a @{ int a; @} v = @{ 42 @}; argv = &v;
19192 (gdb) compile code printf ("%d\n", ((struct a *) argv)->a);
19193 gdb command line:1:36: error: dereferencing pointer to incomplete type ‘struct a’
19194 Compilation failed.
19195 (gdb) compile code struct a @{ int a; @}; printf ("%d\n", ((struct a *) argv)->a);
19199 Variables that have been optimized away by the compiler are not
19200 accessible to the code submitted to the @code{compile} command.
19201 Access to those variables will generate a compiler error which @value{GDBN}
19202 will print to the console.
19205 @subsection Compiler search for the @code{compile} command
19207 @value{GDBN} needs to find @value{NGCC} for the inferior being debugged
19208 which may not be obvious for remote targets of different architecture
19209 than where @value{GDBN} is running. Environment variable @code{PATH} on
19210 @value{GDBN} host is searched for @value{NGCC} binary matching the
19211 target architecture and operating system. This search can be overriden
19212 by @code{set compile-gcc} @value{GDBN} command below. @code{PATH} is
19213 taken from shell that executed @value{GDBN}, it is not the value set by
19214 @value{GDBN} command @code{set environment}). @xref{Environment}.
19217 Specifically @code{PATH} is searched for binaries matching regular expression
19218 @code{@var{arch}(-[^-]*)?-@var{os}-gcc} according to the inferior target being
19219 debugged. @var{arch} is processor name --- multiarch is supported, so for
19220 example both @code{i386} and @code{x86_64} targets look for pattern
19221 @code{(x86_64|i.86)} and both @code{s390} and @code{s390x} targets look
19222 for pattern @code{s390x?}. @var{os} is currently supported only for
19223 pattern @code{linux(-gnu)?}.
19225 On Posix hosts the compiler driver @value{GDBN} needs to find also
19226 shared library @file{libcc1.so} from the compiler. It is searched in
19227 default shared library search path (overridable with usual environment
19228 variable @code{LD_LIBRARY_PATH}), unrelated to @code{PATH} or @code{set
19229 compile-gcc} settings. Contrary to it @file{libcc1plugin.so} is found
19230 according to the installation of the found compiler --- as possibly
19231 specified by the @code{set compile-gcc} command.
19234 @item set compile-gcc
19235 @cindex compile command driver filename override
19236 Set compilation command used for compiling and injecting code with the
19237 @code{compile} commands. If this option is not set (it is set to
19238 an empty string), the search described above will occur --- that is the
19241 @item show compile-gcc
19242 Displays the current compile command @value{NGCC} driver filename.
19243 If set, it is the main command @command{gcc}, found usually for example
19244 under name @file{x86_64-linux-gnu-gcc}.
19248 @chapter @value{GDBN} Files
19250 @value{GDBN} needs to know the file name of the program to be debugged,
19251 both in order to read its symbol table and in order to start your
19252 program. To debug a core dump of a previous run, you must also tell
19253 @value{GDBN} the name of the core dump file.
19256 * Files:: Commands to specify files
19257 * File Caching:: Information about @value{GDBN}'s file caching
19258 * Separate Debug Files:: Debugging information in separate files
19259 * MiniDebugInfo:: Debugging information in a special section
19260 * Index Files:: Index files speed up GDB
19261 * Symbol Errors:: Errors reading symbol files
19262 * Data Files:: GDB data files
19266 @section Commands to Specify Files
19268 @cindex symbol table
19269 @cindex core dump file
19271 You may want to specify executable and core dump file names. The usual
19272 way to do this is at start-up time, using the arguments to
19273 @value{GDBN}'s start-up commands (@pxref{Invocation, , Getting In and
19274 Out of @value{GDBN}}).
19276 Occasionally it is necessary to change to a different file during a
19277 @value{GDBN} session. Or you may run @value{GDBN} and forget to
19278 specify a file you want to use. Or you are debugging a remote target
19279 via @code{gdbserver} (@pxref{Server, file, Using the @code{gdbserver}
19280 Program}). In these situations the @value{GDBN} commands to specify
19281 new files are useful.
19284 @cindex executable file
19286 @item file @var{filename}
19287 Use @var{filename} as the program to be debugged. It is read for its
19288 symbols and for the contents of pure memory. It is also the program
19289 executed when you use the @code{run} command. If you do not specify a
19290 directory and the file is not found in the @value{GDBN} working directory,
19291 @value{GDBN} uses the environment variable @code{PATH} as a list of
19292 directories to search, just as the shell does when looking for a program
19293 to run. You can change the value of this variable, for both @value{GDBN}
19294 and your program, using the @code{path} command.
19296 @cindex unlinked object files
19297 @cindex patching object files
19298 You can load unlinked object @file{.o} files into @value{GDBN} using
19299 the @code{file} command. You will not be able to ``run'' an object
19300 file, but you can disassemble functions and inspect variables. Also,
19301 if the underlying BFD functionality supports it, you could use
19302 @kbd{gdb -write} to patch object files using this technique. Note
19303 that @value{GDBN} can neither interpret nor modify relocations in this
19304 case, so branches and some initialized variables will appear to go to
19305 the wrong place. But this feature is still handy from time to time.
19308 @code{file} with no argument makes @value{GDBN} discard any information it
19309 has on both executable file and the symbol table.
19312 @item exec-file @r{[} @var{filename} @r{]}
19313 Specify that the program to be run (but not the symbol table) is found
19314 in @var{filename}. @value{GDBN} searches the environment variable @code{PATH}
19315 if necessary to locate your program. Omitting @var{filename} means to
19316 discard information on the executable file.
19318 @kindex symbol-file
19319 @item symbol-file @r{[} @var{filename} @r{[} -o @var{offset} @r{]]}
19320 Read symbol table information from file @var{filename}. @code{PATH} is
19321 searched when necessary. Use the @code{file} command to get both symbol
19322 table and program to run from the same file.
19324 If an optional @var{offset} is specified, it is added to the start
19325 address of each section in the symbol file. This is useful if the
19326 program is relocated at runtime, such as the Linux kernel with kASLR
19329 @code{symbol-file} with no argument clears out @value{GDBN} information on your
19330 program's symbol table.
19332 The @code{symbol-file} command causes @value{GDBN} to forget the contents of
19333 some breakpoints and auto-display expressions. This is because they may
19334 contain pointers to the internal data recording symbols and data types,
19335 which are part of the old symbol table data being discarded inside
19338 @code{symbol-file} does not repeat if you press @key{RET} again after
19341 When @value{GDBN} is configured for a particular environment, it
19342 understands debugging information in whatever format is the standard
19343 generated for that environment; you may use either a @sc{gnu} compiler, or
19344 other compilers that adhere to the local conventions.
19345 Best results are usually obtained from @sc{gnu} compilers; for example,
19346 using @code{@value{NGCC}} you can generate debugging information for
19349 For most kinds of object files, with the exception of old SVR3 systems
19350 using COFF, the @code{symbol-file} command does not normally read the
19351 symbol table in full right away. Instead, it scans the symbol table
19352 quickly to find which source files and which symbols are present. The
19353 details are read later, one source file at a time, as they are needed.
19355 The purpose of this two-stage reading strategy is to make @value{GDBN}
19356 start up faster. For the most part, it is invisible except for
19357 occasional pauses while the symbol table details for a particular source
19358 file are being read. (The @code{set verbose} command can turn these
19359 pauses into messages if desired. @xref{Messages/Warnings, ,Optional
19360 Warnings and Messages}.)
19362 We have not implemented the two-stage strategy for COFF yet. When the
19363 symbol table is stored in COFF format, @code{symbol-file} reads the
19364 symbol table data in full right away. Note that ``stabs-in-COFF''
19365 still does the two-stage strategy, since the debug info is actually
19369 @cindex reading symbols immediately
19370 @cindex symbols, reading immediately
19371 @item symbol-file @r{[} -readnow @r{]} @var{filename}
19372 @itemx file @r{[} -readnow @r{]} @var{filename}
19373 You can override the @value{GDBN} two-stage strategy for reading symbol
19374 tables by using the @samp{-readnow} option with any of the commands that
19375 load symbol table information, if you want to be sure @value{GDBN} has the
19376 entire symbol table available.
19378 @cindex @code{-readnever}, option for symbol-file command
19379 @cindex never read symbols
19380 @cindex symbols, never read
19381 @item symbol-file @r{[} -readnever @r{]} @var{filename}
19382 @itemx file @r{[} -readnever @r{]} @var{filename}
19383 You can instruct @value{GDBN} to never read the symbolic information
19384 contained in @var{filename} by using the @samp{-readnever} option.
19385 @xref{--readnever}.
19387 @c FIXME: for now no mention of directories, since this seems to be in
19388 @c flux. 13mar1992 status is that in theory GDB would look either in
19389 @c current dir or in same dir as myprog; but issues like competing
19390 @c GDB's, or clutter in system dirs, mean that in practice right now
19391 @c only current dir is used. FFish says maybe a special GDB hierarchy
19392 @c (eg rooted in val of env var GDBSYMS) could exist for mappable symbol
19396 @item core-file @r{[}@var{filename}@r{]}
19398 Specify the whereabouts of a core dump file to be used as the ``contents
19399 of memory''. Traditionally, core files contain only some parts of the
19400 address space of the process that generated them; @value{GDBN} can access the
19401 executable file itself for other parts.
19403 @code{core-file} with no argument specifies that no core file is
19406 Note that the core file is ignored when your program is actually running
19407 under @value{GDBN}. So, if you have been running your program and you
19408 wish to debug a core file instead, you must kill the subprocess in which
19409 the program is running. To do this, use the @code{kill} command
19410 (@pxref{Kill Process, ,Killing the Child Process}).
19412 @kindex add-symbol-file
19413 @cindex dynamic linking
19414 @item add-symbol-file @var{filename} @r{[} -readnow @r{|} -readnever @r{]} @r{[} -o @var{offset} @r{]} @r{[} @var{textaddress} @r{]} @r{[} -s @var{section} @var{address} @dots{} @r{]}
19415 The @code{add-symbol-file} command reads additional symbol table
19416 information from the file @var{filename}. You would use this command
19417 when @var{filename} has been dynamically loaded (by some other means)
19418 into the program that is running. The @var{textaddress} parameter gives
19419 the memory address at which the file's text section has been loaded.
19420 You can additionally specify the base address of other sections using
19421 an arbitrary number of @samp{-s @var{section} @var{address}} pairs.
19422 If a section is omitted, @value{GDBN} will use its default addresses
19423 as found in @var{filename}. Any @var{address} or @var{textaddress}
19424 can be given as an expression.
19426 If an optional @var{offset} is specified, it is added to the start
19427 address of each section, except those for which the address was
19428 specified explicitly.
19430 The symbol table of the file @var{filename} is added to the symbol table
19431 originally read with the @code{symbol-file} command. You can use the
19432 @code{add-symbol-file} command any number of times; the new symbol data
19433 thus read is kept in addition to the old.
19435 Changes can be reverted using the command @code{remove-symbol-file}.
19437 @cindex relocatable object files, reading symbols from
19438 @cindex object files, relocatable, reading symbols from
19439 @cindex reading symbols from relocatable object files
19440 @cindex symbols, reading from relocatable object files
19441 @cindex @file{.o} files, reading symbols from
19442 Although @var{filename} is typically a shared library file, an
19443 executable file, or some other object file which has been fully
19444 relocated for loading into a process, you can also load symbolic
19445 information from relocatable @file{.o} files, as long as:
19449 the file's symbolic information refers only to linker symbols defined in
19450 that file, not to symbols defined by other object files,
19452 every section the file's symbolic information refers to has actually
19453 been loaded into the inferior, as it appears in the file, and
19455 you can determine the address at which every section was loaded, and
19456 provide these to the @code{add-symbol-file} command.
19460 Some embedded operating systems, like Sun Chorus and VxWorks, can load
19461 relocatable files into an already running program; such systems
19462 typically make the requirements above easy to meet. However, it's
19463 important to recognize that many native systems use complex link
19464 procedures (@code{.linkonce} section factoring and C@t{++} constructor table
19465 assembly, for example) that make the requirements difficult to meet. In
19466 general, one cannot assume that using @code{add-symbol-file} to read a
19467 relocatable object file's symbolic information will have the same effect
19468 as linking the relocatable object file into the program in the normal
19471 @code{add-symbol-file} does not repeat if you press @key{RET} after using it.
19473 @kindex remove-symbol-file
19474 @item remove-symbol-file @var{filename}
19475 @item remove-symbol-file -a @var{address}
19476 Remove a symbol file added via the @code{add-symbol-file} command. The
19477 file to remove can be identified by its @var{filename} or by an @var{address}
19478 that lies within the boundaries of this symbol file in memory. Example:
19481 (gdb) add-symbol-file /home/user/gdb/mylib.so 0x7ffff7ff9480
19482 add symbol table from file "/home/user/gdb/mylib.so" at
19483 .text_addr = 0x7ffff7ff9480
19485 Reading symbols from /home/user/gdb/mylib.so...done.
19486 (gdb) remove-symbol-file -a 0x7ffff7ff9480
19487 Remove symbol table from file "/home/user/gdb/mylib.so"? (y or n) y
19492 @code{remove-symbol-file} does not repeat if you press @key{RET} after using it.
19494 @kindex add-symbol-file-from-memory
19495 @cindex @code{syscall DSO}
19496 @cindex load symbols from memory
19497 @item add-symbol-file-from-memory @var{address}
19498 Load symbols from the given @var{address} in a dynamically loaded
19499 object file whose image is mapped directly into the inferior's memory.
19500 For example, the Linux kernel maps a @code{syscall DSO} into each
19501 process's address space; this DSO provides kernel-specific code for
19502 some system calls. The argument can be any expression whose
19503 evaluation yields the address of the file's shared object file header.
19504 For this command to work, you must have used @code{symbol-file} or
19505 @code{exec-file} commands in advance.
19508 @item section @var{section} @var{addr}
19509 The @code{section} command changes the base address of the named
19510 @var{section} of the exec file to @var{addr}. This can be used if the
19511 exec file does not contain section addresses, (such as in the
19512 @code{a.out} format), or when the addresses specified in the file
19513 itself are wrong. Each section must be changed separately. The
19514 @code{info files} command, described below, lists all the sections and
19518 @kindex info target
19521 @code{info files} and @code{info target} are synonymous; both print the
19522 current target (@pxref{Targets, ,Specifying a Debugging Target}),
19523 including the names of the executable and core dump files currently in
19524 use by @value{GDBN}, and the files from which symbols were loaded. The
19525 command @code{help target} lists all possible targets rather than
19528 @kindex maint info sections
19529 @item maint info sections
19530 Another command that can give you extra information about program sections
19531 is @code{maint info sections}. In addition to the section information
19532 displayed by @code{info files}, this command displays the flags and file
19533 offset of each section in the executable and core dump files. In addition,
19534 @code{maint info sections} provides the following command options (which
19535 may be arbitrarily combined):
19539 Display sections for all loaded object files, including shared libraries.
19540 @item @var{sections}
19541 Display info only for named @var{sections}.
19542 @item @var{section-flags}
19543 Display info only for sections for which @var{section-flags} are true.
19544 The section flags that @value{GDBN} currently knows about are:
19547 Section will have space allocated in the process when loaded.
19548 Set for all sections except those containing debug information.
19550 Section will be loaded from the file into the child process memory.
19551 Set for pre-initialized code and data, clear for @code{.bss} sections.
19553 Section needs to be relocated before loading.
19555 Section cannot be modified by the child process.
19557 Section contains executable code only.
19559 Section contains data only (no executable code).
19561 Section will reside in ROM.
19563 Section contains data for constructor/destructor lists.
19565 Section is not empty.
19567 An instruction to the linker to not output the section.
19568 @item COFF_SHARED_LIBRARY
19569 A notification to the linker that the section contains
19570 COFF shared library information.
19572 Section contains common symbols.
19575 @kindex set trust-readonly-sections
19576 @cindex read-only sections
19577 @item set trust-readonly-sections on
19578 Tell @value{GDBN} that readonly sections in your object file
19579 really are read-only (i.e.@: that their contents will not change).
19580 In that case, @value{GDBN} can fetch values from these sections
19581 out of the object file, rather than from the target program.
19582 For some targets (notably embedded ones), this can be a significant
19583 enhancement to debugging performance.
19585 The default is off.
19587 @item set trust-readonly-sections off
19588 Tell @value{GDBN} not to trust readonly sections. This means that
19589 the contents of the section might change while the program is running,
19590 and must therefore be fetched from the target when needed.
19592 @item show trust-readonly-sections
19593 Show the current setting of trusting readonly sections.
19596 All file-specifying commands allow both absolute and relative file names
19597 as arguments. @value{GDBN} always converts the file name to an absolute file
19598 name and remembers it that way.
19600 @cindex shared libraries
19601 @anchor{Shared Libraries}
19602 @value{GDBN} supports @sc{gnu}/Linux, MS-Windows, SunOS,
19603 Darwin/Mach-O, SVr4, IBM RS/6000 AIX, QNX Neutrino, FDPIC (FR-V), and
19604 DSBT (TIC6X) shared libraries.
19606 On MS-Windows @value{GDBN} must be linked with the Expat library to support
19607 shared libraries. @xref{Expat}.
19609 @value{GDBN} automatically loads symbol definitions from shared libraries
19610 when you use the @code{run} command, or when you examine a core file.
19611 (Before you issue the @code{run} command, @value{GDBN} does not understand
19612 references to a function in a shared library, however---unless you are
19613 debugging a core file).
19615 @c FIXME: some @value{GDBN} release may permit some refs to undef
19616 @c FIXME...symbols---eg in a break cmd---assuming they are from a shared
19617 @c FIXME...lib; check this from time to time when updating manual
19619 There are times, however, when you may wish to not automatically load
19620 symbol definitions from shared libraries, such as when they are
19621 particularly large or there are many of them.
19623 To control the automatic loading of shared library symbols, use the
19627 @kindex set auto-solib-add
19628 @item set auto-solib-add @var{mode}
19629 If @var{mode} is @code{on}, symbols from all shared object libraries
19630 will be loaded automatically when the inferior begins execution, you
19631 attach to an independently started inferior, or when the dynamic linker
19632 informs @value{GDBN} that a new library has been loaded. If @var{mode}
19633 is @code{off}, symbols must be loaded manually, using the
19634 @code{sharedlibrary} command. The default value is @code{on}.
19636 @cindex memory used for symbol tables
19637 If your program uses lots of shared libraries with debug info that
19638 takes large amounts of memory, you can decrease the @value{GDBN}
19639 memory footprint by preventing it from automatically loading the
19640 symbols from shared libraries. To that end, type @kbd{set
19641 auto-solib-add off} before running the inferior, then load each
19642 library whose debug symbols you do need with @kbd{sharedlibrary
19643 @var{regexp}}, where @var{regexp} is a regular expression that matches
19644 the libraries whose symbols you want to be loaded.
19646 @kindex show auto-solib-add
19647 @item show auto-solib-add
19648 Display the current autoloading mode.
19651 @cindex load shared library
19652 To explicitly load shared library symbols, use the @code{sharedlibrary}
19656 @kindex info sharedlibrary
19658 @item info share @var{regex}
19659 @itemx info sharedlibrary @var{regex}
19660 Print the names of the shared libraries which are currently loaded
19661 that match @var{regex}. If @var{regex} is omitted then print
19662 all shared libraries that are loaded.
19665 @item info dll @var{regex}
19666 This is an alias of @code{info sharedlibrary}.
19668 @kindex sharedlibrary
19670 @item sharedlibrary @var{regex}
19671 @itemx share @var{regex}
19672 Load shared object library symbols for files matching a
19673 Unix regular expression.
19674 As with files loaded automatically, it only loads shared libraries
19675 required by your program for a core file or after typing @code{run}. If
19676 @var{regex} is omitted all shared libraries required by your program are
19679 @item nosharedlibrary
19680 @kindex nosharedlibrary
19681 @cindex unload symbols from shared libraries
19682 Unload all shared object library symbols. This discards all symbols
19683 that have been loaded from all shared libraries. Symbols from shared
19684 libraries that were loaded by explicit user requests are not
19688 Sometimes you may wish that @value{GDBN} stops and gives you control
19689 when any of shared library events happen. The best way to do this is
19690 to use @code{catch load} and @code{catch unload} (@pxref{Set
19693 @value{GDBN} also supports the the @code{set stop-on-solib-events}
19694 command for this. This command exists for historical reasons. It is
19695 less useful than setting a catchpoint, because it does not allow for
19696 conditions or commands as a catchpoint does.
19699 @item set stop-on-solib-events
19700 @kindex set stop-on-solib-events
19701 This command controls whether @value{GDBN} should give you control
19702 when the dynamic linker notifies it about some shared library event.
19703 The most common event of interest is loading or unloading of a new
19706 @item show stop-on-solib-events
19707 @kindex show stop-on-solib-events
19708 Show whether @value{GDBN} stops and gives you control when shared
19709 library events happen.
19712 Shared libraries are also supported in many cross or remote debugging
19713 configurations. @value{GDBN} needs to have access to the target's libraries;
19714 this can be accomplished either by providing copies of the libraries
19715 on the host system, or by asking @value{GDBN} to automatically retrieve the
19716 libraries from the target. If copies of the target libraries are
19717 provided, they need to be the same as the target libraries, although the
19718 copies on the target can be stripped as long as the copies on the host are
19721 @cindex where to look for shared libraries
19722 For remote debugging, you need to tell @value{GDBN} where the target
19723 libraries are, so that it can load the correct copies---otherwise, it
19724 may try to load the host's libraries. @value{GDBN} has two variables
19725 to specify the search directories for target libraries.
19728 @cindex prefix for executable and shared library file names
19729 @cindex system root, alternate
19730 @kindex set solib-absolute-prefix
19731 @kindex set sysroot
19732 @item set sysroot @var{path}
19733 Use @var{path} as the system root for the program being debugged. Any
19734 absolute shared library paths will be prefixed with @var{path}; many
19735 runtime loaders store the absolute paths to the shared library in the
19736 target program's memory. When starting processes remotely, and when
19737 attaching to already-running processes (local or remote), their
19738 executable filenames will be prefixed with @var{path} if reported to
19739 @value{GDBN} as absolute by the operating system. If you use
19740 @code{set sysroot} to find executables and shared libraries, they need
19741 to be laid out in the same way that they are on the target, with
19742 e.g.@: a @file{/bin}, @file{/lib} and @file{/usr/lib} hierarchy under
19745 If @var{path} starts with the sequence @file{target:} and the target
19746 system is remote then @value{GDBN} will retrieve the target binaries
19747 from the remote system. This is only supported when using a remote
19748 target that supports the @code{remote get} command (@pxref{File
19749 Transfer,,Sending files to a remote system}). The part of @var{path}
19750 following the initial @file{target:} (if present) is used as system
19751 root prefix on the remote file system. If @var{path} starts with the
19752 sequence @file{remote:} this is converted to the sequence
19753 @file{target:} by @code{set sysroot}@footnote{Historically the
19754 functionality to retrieve binaries from the remote system was
19755 provided by prefixing @var{path} with @file{remote:}}. If you want
19756 to specify a local system root using a directory that happens to be
19757 named @file{target:} or @file{remote:}, you need to use some
19758 equivalent variant of the name like @file{./target:}.
19760 For targets with an MS-DOS based filesystem, such as MS-Windows and
19761 SymbianOS, @value{GDBN} tries prefixing a few variants of the target
19762 absolute file name with @var{path}. But first, on Unix hosts,
19763 @value{GDBN} converts all backslash directory separators into forward
19764 slashes, because the backslash is not a directory separator on Unix:
19767 c:\foo\bar.dll @result{} c:/foo/bar.dll
19770 Then, @value{GDBN} attempts prefixing the target file name with
19771 @var{path}, and looks for the resulting file name in the host file
19775 c:/foo/bar.dll @result{} /path/to/sysroot/c:/foo/bar.dll
19778 If that does not find the binary, @value{GDBN} tries removing
19779 the @samp{:} character from the drive spec, both for convenience, and,
19780 for the case of the host file system not supporting file names with
19784 c:/foo/bar.dll @result{} /path/to/sysroot/c/foo/bar.dll
19787 This makes it possible to have a system root that mirrors a target
19788 with more than one drive. E.g., you may want to setup your local
19789 copies of the target system shared libraries like so (note @samp{c} vs
19793 @file{/path/to/sysroot/c/sys/bin/foo.dll}
19794 @file{/path/to/sysroot/c/sys/bin/bar.dll}
19795 @file{/path/to/sysroot/z/sys/bin/bar.dll}
19799 and point the system root at @file{/path/to/sysroot}, so that
19800 @value{GDBN} can find the correct copies of both
19801 @file{c:\sys\bin\foo.dll}, and @file{z:\sys\bin\bar.dll}.
19803 If that still does not find the binary, @value{GDBN} tries
19804 removing the whole drive spec from the target file name:
19807 c:/foo/bar.dll @result{} /path/to/sysroot/foo/bar.dll
19810 This last lookup makes it possible to not care about the drive name,
19811 if you don't want or need to.
19813 The @code{set solib-absolute-prefix} command is an alias for @code{set
19816 @cindex default system root
19817 @cindex @samp{--with-sysroot}
19818 You can set the default system root by using the configure-time
19819 @samp{--with-sysroot} option. If the system root is inside
19820 @value{GDBN}'s configured binary prefix (set with @samp{--prefix} or
19821 @samp{--exec-prefix}), then the default system root will be updated
19822 automatically if the installed @value{GDBN} is moved to a new
19825 @kindex show sysroot
19827 Display the current executable and shared library prefix.
19829 @kindex set solib-search-path
19830 @item set solib-search-path @var{path}
19831 If this variable is set, @var{path} is a colon-separated list of
19832 directories to search for shared libraries. @samp{solib-search-path}
19833 is used after @samp{sysroot} fails to locate the library, or if the
19834 path to the library is relative instead of absolute. If you want to
19835 use @samp{solib-search-path} instead of @samp{sysroot}, be sure to set
19836 @samp{sysroot} to a nonexistent directory to prevent @value{GDBN} from
19837 finding your host's libraries. @samp{sysroot} is preferred; setting
19838 it to a nonexistent directory may interfere with automatic loading
19839 of shared library symbols.
19841 @kindex show solib-search-path
19842 @item show solib-search-path
19843 Display the current shared library search path.
19845 @cindex DOS file-name semantics of file names.
19846 @kindex set target-file-system-kind (unix|dos-based|auto)
19847 @kindex show target-file-system-kind
19848 @item set target-file-system-kind @var{kind}
19849 Set assumed file system kind for target reported file names.
19851 Shared library file names as reported by the target system may not
19852 make sense as is on the system @value{GDBN} is running on. For
19853 example, when remote debugging a target that has MS-DOS based file
19854 system semantics, from a Unix host, the target may be reporting to
19855 @value{GDBN} a list of loaded shared libraries with file names such as
19856 @file{c:\Windows\kernel32.dll}. On Unix hosts, there's no concept of
19857 drive letters, so the @samp{c:\} prefix is not normally understood as
19858 indicating an absolute file name, and neither is the backslash
19859 normally considered a directory separator character. In that case,
19860 the native file system would interpret this whole absolute file name
19861 as a relative file name with no directory components. This would make
19862 it impossible to point @value{GDBN} at a copy of the remote target's
19863 shared libraries on the host using @code{set sysroot}, and impractical
19864 with @code{set solib-search-path}. Setting
19865 @code{target-file-system-kind} to @code{dos-based} tells @value{GDBN}
19866 to interpret such file names similarly to how the target would, and to
19867 map them to file names valid on @value{GDBN}'s native file system
19868 semantics. The value of @var{kind} can be @code{"auto"}, in addition
19869 to one of the supported file system kinds. In that case, @value{GDBN}
19870 tries to determine the appropriate file system variant based on the
19871 current target's operating system (@pxref{ABI, ,Configuring the
19872 Current ABI}). The supported file system settings are:
19876 Instruct @value{GDBN} to assume the target file system is of Unix
19877 kind. Only file names starting the forward slash (@samp{/}) character
19878 are considered absolute, and the directory separator character is also
19882 Instruct @value{GDBN} to assume the target file system is DOS based.
19883 File names starting with either a forward slash, or a drive letter
19884 followed by a colon (e.g., @samp{c:}), are considered absolute, and
19885 both the slash (@samp{/}) and the backslash (@samp{\\}) characters are
19886 considered directory separators.
19889 Instruct @value{GDBN} to use the file system kind associated with the
19890 target operating system (@pxref{ABI, ,Configuring the Current ABI}).
19891 This is the default.
19895 @cindex file name canonicalization
19896 @cindex base name differences
19897 When processing file names provided by the user, @value{GDBN}
19898 frequently needs to compare them to the file names recorded in the
19899 program's debug info. Normally, @value{GDBN} compares just the
19900 @dfn{base names} of the files as strings, which is reasonably fast
19901 even for very large programs. (The base name of a file is the last
19902 portion of its name, after stripping all the leading directories.)
19903 This shortcut in comparison is based upon the assumption that files
19904 cannot have more than one base name. This is usually true, but
19905 references to files that use symlinks or similar filesystem
19906 facilities violate that assumption. If your program records files
19907 using such facilities, or if you provide file names to @value{GDBN}
19908 using symlinks etc., you can set @code{basenames-may-differ} to
19909 @code{true} to instruct @value{GDBN} to completely canonicalize each
19910 pair of file names it needs to compare. This will make file-name
19911 comparisons accurate, but at a price of a significant slowdown.
19914 @item set basenames-may-differ
19915 @kindex set basenames-may-differ
19916 Set whether a source file may have multiple base names.
19918 @item show basenames-may-differ
19919 @kindex show basenames-may-differ
19920 Show whether a source file may have multiple base names.
19924 @section File Caching
19925 @cindex caching of opened files
19926 @cindex caching of bfd objects
19928 To speed up file loading, and reduce memory usage, @value{GDBN} will
19929 reuse the @code{bfd} objects used to track open files. @xref{Top, ,
19930 BFD, bfd, The Binary File Descriptor Library}. The following commands
19931 allow visibility and control of the caching behavior.
19934 @kindex maint info bfds
19935 @item maint info bfds
19936 This prints information about each @code{bfd} object that is known to
19939 @kindex maint set bfd-sharing
19940 @kindex maint show bfd-sharing
19941 @kindex bfd caching
19942 @item maint set bfd-sharing
19943 @item maint show bfd-sharing
19944 Control whether @code{bfd} objects can be shared. When sharing is
19945 enabled @value{GDBN} reuses already open @code{bfd} objects rather
19946 than reopening the same file. Turning sharing off does not cause
19947 already shared @code{bfd} objects to be unshared, but all future files
19948 that are opened will create a new @code{bfd} object. Similarly,
19949 re-enabling sharing does not cause multiple existing @code{bfd}
19950 objects to be collapsed into a single shared @code{bfd} object.
19952 @kindex set debug bfd-cache @var{level}
19953 @kindex bfd caching
19954 @item set debug bfd-cache @var{level}
19955 Turns on debugging of the bfd cache, setting the level to @var{level}.
19957 @kindex show debug bfd-cache
19958 @kindex bfd caching
19959 @item show debug bfd-cache
19960 Show the current debugging level of the bfd cache.
19963 @node Separate Debug Files
19964 @section Debugging Information in Separate Files
19965 @cindex separate debugging information files
19966 @cindex debugging information in separate files
19967 @cindex @file{.debug} subdirectories
19968 @cindex debugging information directory, global
19969 @cindex global debugging information directories
19970 @cindex build ID, and separate debugging files
19971 @cindex @file{.build-id} directory
19973 @value{GDBN} allows you to put a program's debugging information in a
19974 file separate from the executable itself, in a way that allows
19975 @value{GDBN} to find and load the debugging information automatically.
19976 Since debugging information can be very large---sometimes larger
19977 than the executable code itself---some systems distribute debugging
19978 information for their executables in separate files, which users can
19979 install only when they need to debug a problem.
19981 @value{GDBN} supports two ways of specifying the separate debug info
19986 The executable contains a @dfn{debug link} that specifies the name of
19987 the separate debug info file. The separate debug file's name is
19988 usually @file{@var{executable}.debug}, where @var{executable} is the
19989 name of the corresponding executable file without leading directories
19990 (e.g., @file{ls.debug} for @file{/usr/bin/ls}). In addition, the
19991 debug link specifies a 32-bit @dfn{Cyclic Redundancy Check} (CRC)
19992 checksum for the debug file, which @value{GDBN} uses to validate that
19993 the executable and the debug file came from the same build.
19996 The executable contains a @dfn{build ID}, a unique bit string that is
19997 also present in the corresponding debug info file. (This is supported
19998 only on some operating systems, when using the ELF or PE file formats
19999 for binary files and the @sc{gnu} Binutils.) For more details about
20000 this feature, see the description of the @option{--build-id}
20001 command-line option in @ref{Options, , Command Line Options, ld,
20002 The GNU Linker}. The debug info file's name is not specified
20003 explicitly by the build ID, but can be computed from the build ID, see
20007 Depending on the way the debug info file is specified, @value{GDBN}
20008 uses two different methods of looking for the debug file:
20012 For the ``debug link'' method, @value{GDBN} looks up the named file in
20013 the directory of the executable file, then in a subdirectory of that
20014 directory named @file{.debug}, and finally under each one of the
20015 global debug directories, in a subdirectory whose name is identical to
20016 the leading directories of the executable's absolute file name. (On
20017 MS-Windows/MS-DOS, the drive letter of the executable's leading
20018 directories is converted to a one-letter subdirectory, i.e.@:
20019 @file{d:/usr/bin/} is converted to @file{/d/usr/bin/}, because Windows
20020 filesystems disallow colons in file names.)
20023 For the ``build ID'' method, @value{GDBN} looks in the
20024 @file{.build-id} subdirectory of each one of the global debug directories for
20025 a file named @file{@var{nn}/@var{nnnnnnnn}.debug}, where @var{nn} are the
20026 first 2 hex characters of the build ID bit string, and @var{nnnnnnnn}
20027 are the rest of the bit string. (Real build ID strings are 32 or more
20028 hex characters, not 10.)
20031 So, for example, suppose you ask @value{GDBN} to debug
20032 @file{/usr/bin/ls}, which has a debug link that specifies the
20033 file @file{ls.debug}, and a build ID whose value in hex is
20034 @code{abcdef1234}. If the list of the global debug directories includes
20035 @file{/usr/lib/debug}, then @value{GDBN} will look for the following
20036 debug information files, in the indicated order:
20040 @file{/usr/lib/debug/.build-id/ab/cdef1234.debug}
20042 @file{/usr/bin/ls.debug}
20044 @file{/usr/bin/.debug/ls.debug}
20046 @file{/usr/lib/debug/usr/bin/ls.debug}.
20049 @anchor{debug-file-directory}
20050 Global debugging info directories default to what is set by @value{GDBN}
20051 configure option @option{--with-separate-debug-dir}. During @value{GDBN} run
20052 you can also set the global debugging info directories, and view the list
20053 @value{GDBN} is currently using.
20057 @kindex set debug-file-directory
20058 @item set debug-file-directory @var{directories}
20059 Set the directories which @value{GDBN} searches for separate debugging
20060 information files to @var{directory}. Multiple path components can be set
20061 concatenating them by a path separator.
20063 @kindex show debug-file-directory
20064 @item show debug-file-directory
20065 Show the directories @value{GDBN} searches for separate debugging
20070 @cindex @code{.gnu_debuglink} sections
20071 @cindex debug link sections
20072 A debug link is a special section of the executable file named
20073 @code{.gnu_debuglink}. The section must contain:
20077 A filename, with any leading directory components removed, followed by
20080 zero to three bytes of padding, as needed to reach the next four-byte
20081 boundary within the section, and
20083 a four-byte CRC checksum, stored in the same endianness used for the
20084 executable file itself. The checksum is computed on the debugging
20085 information file's full contents by the function given below, passing
20086 zero as the @var{crc} argument.
20089 Any executable file format can carry a debug link, as long as it can
20090 contain a section named @code{.gnu_debuglink} with the contents
20093 @cindex @code{.note.gnu.build-id} sections
20094 @cindex build ID sections
20095 The build ID is a special section in the executable file (and in other
20096 ELF binary files that @value{GDBN} may consider). This section is
20097 often named @code{.note.gnu.build-id}, but that name is not mandatory.
20098 It contains unique identification for the built files---the ID remains
20099 the same across multiple builds of the same build tree. The default
20100 algorithm SHA1 produces 160 bits (40 hexadecimal characters) of the
20101 content for the build ID string. The same section with an identical
20102 value is present in the original built binary with symbols, in its
20103 stripped variant, and in the separate debugging information file.
20105 The debugging information file itself should be an ordinary
20106 executable, containing a full set of linker symbols, sections, and
20107 debugging information. The sections of the debugging information file
20108 should have the same names, addresses, and sizes as the original file,
20109 but they need not contain any data---much like a @code{.bss} section
20110 in an ordinary executable.
20112 The @sc{gnu} binary utilities (Binutils) package includes the
20113 @samp{objcopy} utility that can produce
20114 the separated executable / debugging information file pairs using the
20115 following commands:
20118 @kbd{objcopy --only-keep-debug foo foo.debug}
20123 These commands remove the debugging
20124 information from the executable file @file{foo} and place it in the file
20125 @file{foo.debug}. You can use the first, second or both methods to link the
20130 The debug link method needs the following additional command to also leave
20131 behind a debug link in @file{foo}:
20134 @kbd{objcopy --add-gnu-debuglink=foo.debug foo}
20137 Ulrich Drepper's @file{elfutils} package, starting with version 0.53, contains
20138 a version of the @code{strip} command such that the command @kbd{strip foo -f
20139 foo.debug} has the same functionality as the two @code{objcopy} commands and
20140 the @code{ln -s} command above, together.
20143 Build ID gets embedded into the main executable using @code{ld --build-id} or
20144 the @value{NGCC} counterpart @code{gcc -Wl,--build-id}. Build ID support plus
20145 compatibility fixes for debug files separation are present in @sc{gnu} binary
20146 utilities (Binutils) package since version 2.18.
20151 @cindex CRC algorithm definition
20152 The CRC used in @code{.gnu_debuglink} is the CRC-32 defined in
20153 IEEE 802.3 using the polynomial:
20155 @c TexInfo requires naked braces for multi-digit exponents for Tex
20156 @c output, but this causes HTML output to barf. HTML has to be set using
20157 @c raw commands. So we end up having to specify this equation in 2
20162 <em>x</em><sup>32</sup> + <em>x</em><sup>26</sup> + <em>x</em><sup>23</sup> + <em>x</em><sup>22</sup> + <em>x</em><sup>16</sup> + <em>x</em><sup>12</sup> + <em>x</em><sup>11</sup>
20163 + <em>x</em><sup>10</sup> + <em>x</em><sup>8</sup> + <em>x</em><sup>7</sup> + <em>x</em><sup>5</sup> + <em>x</em><sup>4</sup> + <em>x</em><sup>2</sup> + <em>x</em> + 1
20169 @math{x^{32} + x^{26} + x^{23} + x^{22} + x^{16} + x^{12} + x^{11}}
20170 @math{+ x^{10} + x^8 + x^7 + x^5 + x^4 + x^2 + x + 1}
20174 The function is computed byte at a time, taking the least
20175 significant bit of each byte first. The initial pattern
20176 @code{0xffffffff} is used, to ensure leading zeros affect the CRC and
20177 the final result is inverted to ensure trailing zeros also affect the
20180 @emph{Note:} This is the same CRC polynomial as used in handling the
20181 @dfn{Remote Serial Protocol} @code{qCRC} packet (@pxref{qCRC packet}).
20182 However in the case of the Remote Serial Protocol, the CRC is computed
20183 @emph{most} significant bit first, and the result is not inverted, so
20184 trailing zeros have no effect on the CRC value.
20186 To complete the description, we show below the code of the function
20187 which produces the CRC used in @code{.gnu_debuglink}. Inverting the
20188 initially supplied @code{crc} argument means that an initial call to
20189 this function passing in zero will start computing the CRC using
20192 @kindex gnu_debuglink_crc32
20195 gnu_debuglink_crc32 (unsigned long crc,
20196 unsigned char *buf, size_t len)
20198 static const unsigned long crc32_table[256] =
20200 0x00000000, 0x77073096, 0xee0e612c, 0x990951ba, 0x076dc419,
20201 0x706af48f, 0xe963a535, 0x9e6495a3, 0x0edb8832, 0x79dcb8a4,
20202 0xe0d5e91e, 0x97d2d988, 0x09b64c2b, 0x7eb17cbd, 0xe7b82d07,
20203 0x90bf1d91, 0x1db71064, 0x6ab020f2, 0xf3b97148, 0x84be41de,
20204 0x1adad47d, 0x6ddde4eb, 0xf4d4b551, 0x83d385c7, 0x136c9856,
20205 0x646ba8c0, 0xfd62f97a, 0x8a65c9ec, 0x14015c4f, 0x63066cd9,
20206 0xfa0f3d63, 0x8d080df5, 0x3b6e20c8, 0x4c69105e, 0xd56041e4,
20207 0xa2677172, 0x3c03e4d1, 0x4b04d447, 0xd20d85fd, 0xa50ab56b,
20208 0x35b5a8fa, 0x42b2986c, 0xdbbbc9d6, 0xacbcf940, 0x32d86ce3,
20209 0x45df5c75, 0xdcd60dcf, 0xabd13d59, 0x26d930ac, 0x51de003a,
20210 0xc8d75180, 0xbfd06116, 0x21b4f4b5, 0x56b3c423, 0xcfba9599,
20211 0xb8bda50f, 0x2802b89e, 0x5f058808, 0xc60cd9b2, 0xb10be924,
20212 0x2f6f7c87, 0x58684c11, 0xc1611dab, 0xb6662d3d, 0x76dc4190,
20213 0x01db7106, 0x98d220bc, 0xefd5102a, 0x71b18589, 0x06b6b51f,
20214 0x9fbfe4a5, 0xe8b8d433, 0x7807c9a2, 0x0f00f934, 0x9609a88e,
20215 0xe10e9818, 0x7f6a0dbb, 0x086d3d2d, 0x91646c97, 0xe6635c01,
20216 0x6b6b51f4, 0x1c6c6162, 0x856530d8, 0xf262004e, 0x6c0695ed,
20217 0x1b01a57b, 0x8208f4c1, 0xf50fc457, 0x65b0d9c6, 0x12b7e950,
20218 0x8bbeb8ea, 0xfcb9887c, 0x62dd1ddf, 0x15da2d49, 0x8cd37cf3,
20219 0xfbd44c65, 0x4db26158, 0x3ab551ce, 0xa3bc0074, 0xd4bb30e2,
20220 0x4adfa541, 0x3dd895d7, 0xa4d1c46d, 0xd3d6f4fb, 0x4369e96a,
20221 0x346ed9fc, 0xad678846, 0xda60b8d0, 0x44042d73, 0x33031de5,
20222 0xaa0a4c5f, 0xdd0d7cc9, 0x5005713c, 0x270241aa, 0xbe0b1010,
20223 0xc90c2086, 0x5768b525, 0x206f85b3, 0xb966d409, 0xce61e49f,
20224 0x5edef90e, 0x29d9c998, 0xb0d09822, 0xc7d7a8b4, 0x59b33d17,
20225 0x2eb40d81, 0xb7bd5c3b, 0xc0ba6cad, 0xedb88320, 0x9abfb3b6,
20226 0x03b6e20c, 0x74b1d29a, 0xead54739, 0x9dd277af, 0x04db2615,
20227 0x73dc1683, 0xe3630b12, 0x94643b84, 0x0d6d6a3e, 0x7a6a5aa8,
20228 0xe40ecf0b, 0x9309ff9d, 0x0a00ae27, 0x7d079eb1, 0xf00f9344,
20229 0x8708a3d2, 0x1e01f268, 0x6906c2fe, 0xf762575d, 0x806567cb,
20230 0x196c3671, 0x6e6b06e7, 0xfed41b76, 0x89d32be0, 0x10da7a5a,
20231 0x67dd4acc, 0xf9b9df6f, 0x8ebeeff9, 0x17b7be43, 0x60b08ed5,
20232 0xd6d6a3e8, 0xa1d1937e, 0x38d8c2c4, 0x4fdff252, 0xd1bb67f1,
20233 0xa6bc5767, 0x3fb506dd, 0x48b2364b, 0xd80d2bda, 0xaf0a1b4c,
20234 0x36034af6, 0x41047a60, 0xdf60efc3, 0xa867df55, 0x316e8eef,
20235 0x4669be79, 0xcb61b38c, 0xbc66831a, 0x256fd2a0, 0x5268e236,
20236 0xcc0c7795, 0xbb0b4703, 0x220216b9, 0x5505262f, 0xc5ba3bbe,
20237 0xb2bd0b28, 0x2bb45a92, 0x5cb36a04, 0xc2d7ffa7, 0xb5d0cf31,
20238 0x2cd99e8b, 0x5bdeae1d, 0x9b64c2b0, 0xec63f226, 0x756aa39c,
20239 0x026d930a, 0x9c0906a9, 0xeb0e363f, 0x72076785, 0x05005713,
20240 0x95bf4a82, 0xe2b87a14, 0x7bb12bae, 0x0cb61b38, 0x92d28e9b,
20241 0xe5d5be0d, 0x7cdcefb7, 0x0bdbdf21, 0x86d3d2d4, 0xf1d4e242,
20242 0x68ddb3f8, 0x1fda836e, 0x81be16cd, 0xf6b9265b, 0x6fb077e1,
20243 0x18b74777, 0x88085ae6, 0xff0f6a70, 0x66063bca, 0x11010b5c,
20244 0x8f659eff, 0xf862ae69, 0x616bffd3, 0x166ccf45, 0xa00ae278,
20245 0xd70dd2ee, 0x4e048354, 0x3903b3c2, 0xa7672661, 0xd06016f7,
20246 0x4969474d, 0x3e6e77db, 0xaed16a4a, 0xd9d65adc, 0x40df0b66,
20247 0x37d83bf0, 0xa9bcae53, 0xdebb9ec5, 0x47b2cf7f, 0x30b5ffe9,
20248 0xbdbdf21c, 0xcabac28a, 0x53b39330, 0x24b4a3a6, 0xbad03605,
20249 0xcdd70693, 0x54de5729, 0x23d967bf, 0xb3667a2e, 0xc4614ab8,
20250 0x5d681b02, 0x2a6f2b94, 0xb40bbe37, 0xc30c8ea1, 0x5a05df1b,
20253 unsigned char *end;
20255 crc = ~crc & 0xffffffff;
20256 for (end = buf + len; buf < end; ++buf)
20257 crc = crc32_table[(crc ^ *buf) & 0xff] ^ (crc >> 8);
20258 return ~crc & 0xffffffff;
20263 This computation does not apply to the ``build ID'' method.
20265 @node MiniDebugInfo
20266 @section Debugging information in a special section
20267 @cindex separate debug sections
20268 @cindex @samp{.gnu_debugdata} section
20270 Some systems ship pre-built executables and libraries that have a
20271 special @samp{.gnu_debugdata} section. This feature is called
20272 @dfn{MiniDebugInfo}. This section holds an LZMA-compressed object and
20273 is used to supply extra symbols for backtraces.
20275 The intent of this section is to provide extra minimal debugging
20276 information for use in simple backtraces. It is not intended to be a
20277 replacement for full separate debugging information (@pxref{Separate
20278 Debug Files}). The example below shows the intended use; however,
20279 @value{GDBN} does not currently put restrictions on what sort of
20280 debugging information might be included in the section.
20282 @value{GDBN} has support for this extension. If the section exists,
20283 then it is used provided that no other source of debugging information
20284 can be found, and that @value{GDBN} was configured with LZMA support.
20286 This section can be easily created using @command{objcopy} and other
20287 standard utilities:
20290 # Extract the dynamic symbols from the main binary, there is no need
20291 # to also have these in the normal symbol table.
20292 nm -D @var{binary} --format=posix --defined-only \
20293 | awk '@{ print $1 @}' | sort > dynsyms
20295 # Extract all the text (i.e. function) symbols from the debuginfo.
20296 # (Note that we actually also accept "D" symbols, for the benefit
20297 # of platforms like PowerPC64 that use function descriptors.)
20298 nm @var{binary} --format=posix --defined-only \
20299 | awk '@{ if ($2 == "T" || $2 == "t" || $2 == "D") print $1 @}' \
20302 # Keep all the function symbols not already in the dynamic symbol
20304 comm -13 dynsyms funcsyms > keep_symbols
20306 # Separate full debug info into debug binary.
20307 objcopy --only-keep-debug @var{binary} debug
20309 # Copy the full debuginfo, keeping only a minimal set of symbols and
20310 # removing some unnecessary sections.
20311 objcopy -S --remove-section .gdb_index --remove-section .comment \
20312 --keep-symbols=keep_symbols debug mini_debuginfo
20314 # Drop the full debug info from the original binary.
20315 strip --strip-all -R .comment @var{binary}
20317 # Inject the compressed data into the .gnu_debugdata section of the
20320 objcopy --add-section .gnu_debugdata=mini_debuginfo.xz @var{binary}
20324 @section Index Files Speed Up @value{GDBN}
20325 @cindex index files
20326 @cindex @samp{.gdb_index} section
20328 When @value{GDBN} finds a symbol file, it scans the symbols in the
20329 file in order to construct an internal symbol table. This lets most
20330 @value{GDBN} operations work quickly---at the cost of a delay early
20331 on. For large programs, this delay can be quite lengthy, so
20332 @value{GDBN} provides a way to build an index, which speeds up
20335 For convenience, @value{GDBN} comes with a program,
20336 @command{gdb-add-index}, which can be used to add the index to a
20337 symbol file. It takes the symbol file as its only argument:
20340 $ gdb-add-index symfile
20343 @xref{gdb-add-index}.
20345 It is also possible to do the work manually. Here is what
20346 @command{gdb-add-index} does behind the curtains.
20348 The index is stored as a section in the symbol file. @value{GDBN} can
20349 write the index to a file, then you can put it into the symbol file
20350 using @command{objcopy}.
20352 To create an index file, use the @code{save gdb-index} command:
20355 @item save gdb-index [-dwarf-5] @var{directory}
20356 @kindex save gdb-index
20357 Create index files for all symbol files currently known by
20358 @value{GDBN}. For each known @var{symbol-file}, this command by
20359 default creates it produces a single file
20360 @file{@var{symbol-file}.gdb-index}. If you invoke this command with
20361 the @option{-dwarf-5} option, it produces 2 files:
20362 @file{@var{symbol-file}.debug_names} and
20363 @file{@var{symbol-file}.debug_str}. The files are created in the
20364 given @var{directory}.
20367 Once you have created an index file you can merge it into your symbol
20368 file, here named @file{symfile}, using @command{objcopy}:
20371 $ objcopy --add-section .gdb_index=symfile.gdb-index \
20372 --set-section-flags .gdb_index=readonly symfile symfile
20375 Or for @code{-dwarf-5}:
20378 $ objcopy --dump-section .debug_str=symfile.debug_str.new symfile
20379 $ cat symfile.debug_str >>symfile.debug_str.new
20380 $ objcopy --add-section .debug_names=symfile.gdb-index \
20381 --set-section-flags .debug_names=readonly \
20382 --update-section .debug_str=symfile.debug_str.new symfile symfile
20385 @value{GDBN} will normally ignore older versions of @file{.gdb_index}
20386 sections that have been deprecated. Usually they are deprecated because
20387 they are missing a new feature or have performance issues.
20388 To tell @value{GDBN} to use a deprecated index section anyway
20389 specify @code{set use-deprecated-index-sections on}.
20390 The default is @code{off}.
20391 This can speed up startup, but may result in some functionality being lost.
20392 @xref{Index Section Format}.
20394 @emph{Warning:} Setting @code{use-deprecated-index-sections} to @code{on}
20395 must be done before gdb reads the file. The following will not work:
20398 $ gdb -ex "set use-deprecated-index-sections on" <program>
20401 Instead you must do, for example,
20404 $ gdb -iex "set use-deprecated-index-sections on" <program>
20407 There are currently some limitation on indices. They only work when
20408 for DWARF debugging information, not stabs. And, they do not
20409 currently work for programs using Ada.
20411 @subsection Automatic symbol index cache
20413 @cindex automatic symbol index cache
20414 It is possible for @value{GDBN} to automatically save a copy of this index in a
20415 cache on disk and retrieve it from there when loading the same binary in the
20416 future. This feature can be turned on with @kbd{set index-cache on}. The
20417 following commands can be used to tweak the behavior of the index cache.
20421 @kindex set index-cache
20422 @item set index-cache on
20423 @itemx set index-cache off
20424 Enable or disable the use of the symbol index cache.
20426 @item set index-cache directory @var{directory}
20427 @kindex show index-cache
20428 @itemx show index-cache directory
20429 Set/show the directory where index files will be saved.
20431 The default value for this directory depends on the host platform. On
20432 most systems, the index is cached in the @file{gdb} subdirectory of
20433 the directory pointed to by the @env{XDG_CACHE_HOME} environment
20434 variable, if it is defined, else in the @file{.cache/gdb} subdirectory
20435 of your home directory. However, on some systems, the default may
20436 differ according to local convention.
20438 There is no limit on the disk space used by index cache. It is perfectly safe
20439 to delete the content of that directory to free up disk space.
20441 @item show index-cache stats
20442 Print the number of cache hits and misses since the launch of @value{GDBN}.
20446 @node Symbol Errors
20447 @section Errors Reading Symbol Files
20449 While reading a symbol file, @value{GDBN} occasionally encounters problems,
20450 such as symbol types it does not recognize, or known bugs in compiler
20451 output. By default, @value{GDBN} does not notify you of such problems, since
20452 they are relatively common and primarily of interest to people
20453 debugging compilers. If you are interested in seeing information
20454 about ill-constructed symbol tables, you can either ask @value{GDBN} to print
20455 only one message about each such type of problem, no matter how many
20456 times the problem occurs; or you can ask @value{GDBN} to print more messages,
20457 to see how many times the problems occur, with the @code{set
20458 complaints} command (@pxref{Messages/Warnings, ,Optional Warnings and
20461 The messages currently printed, and their meanings, include:
20464 @item inner block not inside outer block in @var{symbol}
20466 The symbol information shows where symbol scopes begin and end
20467 (such as at the start of a function or a block of statements). This
20468 error indicates that an inner scope block is not fully contained
20469 in its outer scope blocks.
20471 @value{GDBN} circumvents the problem by treating the inner block as if it had
20472 the same scope as the outer block. In the error message, @var{symbol}
20473 may be shown as ``@code{(don't know)}'' if the outer block is not a
20476 @item block at @var{address} out of order
20478 The symbol information for symbol scope blocks should occur in
20479 order of increasing addresses. This error indicates that it does not
20482 @value{GDBN} does not circumvent this problem, and has trouble
20483 locating symbols in the source file whose symbols it is reading. (You
20484 can often determine what source file is affected by specifying
20485 @code{set verbose on}. @xref{Messages/Warnings, ,Optional Warnings and
20488 @item bad block start address patched
20490 The symbol information for a symbol scope block has a start address
20491 smaller than the address of the preceding source line. This is known
20492 to occur in the SunOS 4.1.1 (and earlier) C compiler.
20494 @value{GDBN} circumvents the problem by treating the symbol scope block as
20495 starting on the previous source line.
20497 @item bad string table offset in symbol @var{n}
20500 Symbol number @var{n} contains a pointer into the string table which is
20501 larger than the size of the string table.
20503 @value{GDBN} circumvents the problem by considering the symbol to have the
20504 name @code{foo}, which may cause other problems if many symbols end up
20507 @item unknown symbol type @code{0x@var{nn}}
20509 The symbol information contains new data types that @value{GDBN} does
20510 not yet know how to read. @code{0x@var{nn}} is the symbol type of the
20511 uncomprehended information, in hexadecimal.
20513 @value{GDBN} circumvents the error by ignoring this symbol information.
20514 This usually allows you to debug your program, though certain symbols
20515 are not accessible. If you encounter such a problem and feel like
20516 debugging it, you can debug @code{@value{GDBP}} with itself, breakpoint
20517 on @code{complain}, then go up to the function @code{read_dbx_symtab}
20518 and examine @code{*bufp} to see the symbol.
20520 @item stub type has NULL name
20522 @value{GDBN} could not find the full definition for a struct or class.
20524 @item const/volatile indicator missing (ok if using g++ v1.x), got@dots{}
20525 The symbol information for a C@t{++} member function is missing some
20526 information that recent versions of the compiler should have output for
20529 @item info mismatch between compiler and debugger
20531 @value{GDBN} could not parse a type specification output by the compiler.
20536 @section GDB Data Files
20538 @cindex prefix for data files
20539 @value{GDBN} will sometimes read an auxiliary data file. These files
20540 are kept in a directory known as the @dfn{data directory}.
20542 You can set the data directory's name, and view the name @value{GDBN}
20543 is currently using.
20546 @kindex set data-directory
20547 @item set data-directory @var{directory}
20548 Set the directory which @value{GDBN} searches for auxiliary data files
20549 to @var{directory}.
20551 @kindex show data-directory
20552 @item show data-directory
20553 Show the directory @value{GDBN} searches for auxiliary data files.
20556 @cindex default data directory
20557 @cindex @samp{--with-gdb-datadir}
20558 You can set the default data directory by using the configure-time
20559 @samp{--with-gdb-datadir} option. If the data directory is inside
20560 @value{GDBN}'s configured binary prefix (set with @samp{--prefix} or
20561 @samp{--exec-prefix}), then the default data directory will be updated
20562 automatically if the installed @value{GDBN} is moved to a new
20565 The data directory may also be specified with the
20566 @code{--data-directory} command line option.
20567 @xref{Mode Options}.
20570 @chapter Specifying a Debugging Target
20572 @cindex debugging target
20573 A @dfn{target} is the execution environment occupied by your program.
20575 Often, @value{GDBN} runs in the same host environment as your program;
20576 in that case, the debugging target is specified as a side effect when
20577 you use the @code{file} or @code{core} commands. When you need more
20578 flexibility---for example, running @value{GDBN} on a physically separate
20579 host, or controlling a standalone system over a serial port or a
20580 realtime system over a TCP/IP connection---you can use the @code{target}
20581 command to specify one of the target types configured for @value{GDBN}
20582 (@pxref{Target Commands, ,Commands for Managing Targets}).
20584 @cindex target architecture
20585 It is possible to build @value{GDBN} for several different @dfn{target
20586 architectures}. When @value{GDBN} is built like that, you can choose
20587 one of the available architectures with the @kbd{set architecture}
20591 @kindex set architecture
20592 @kindex show architecture
20593 @item set architecture @var{arch}
20594 This command sets the current target architecture to @var{arch}. The
20595 value of @var{arch} can be @code{"auto"}, in addition to one of the
20596 supported architectures.
20598 @item show architecture
20599 Show the current target architecture.
20601 @item set processor
20603 @kindex set processor
20604 @kindex show processor
20605 These are alias commands for, respectively, @code{set architecture}
20606 and @code{show architecture}.
20610 * Active Targets:: Active targets
20611 * Target Commands:: Commands for managing targets
20612 * Byte Order:: Choosing target byte order
20615 @node Active Targets
20616 @section Active Targets
20618 @cindex stacking targets
20619 @cindex active targets
20620 @cindex multiple targets
20622 There are multiple classes of targets such as: processes, executable files or
20623 recording sessions. Core files belong to the process class, making core file
20624 and process mutually exclusive. Otherwise, @value{GDBN} can work concurrently
20625 on multiple active targets, one in each class. This allows you to (for
20626 example) start a process and inspect its activity, while still having access to
20627 the executable file after the process finishes. Or if you start process
20628 recording (@pxref{Reverse Execution}) and @code{reverse-step} there, you are
20629 presented a virtual layer of the recording target, while the process target
20630 remains stopped at the chronologically last point of the process execution.
20632 Use the @code{core-file} and @code{exec-file} commands to select a new core
20633 file or executable target (@pxref{Files, ,Commands to Specify Files}). To
20634 specify as a target a process that is already running, use the @code{attach}
20635 command (@pxref{Attach, ,Debugging an Already-running Process}).
20637 @node Target Commands
20638 @section Commands for Managing Targets
20641 @item target @var{type} @var{parameters}
20642 Connects the @value{GDBN} host environment to a target machine or
20643 process. A target is typically a protocol for talking to debugging
20644 facilities. You use the argument @var{type} to specify the type or
20645 protocol of the target machine.
20647 Further @var{parameters} are interpreted by the target protocol, but
20648 typically include things like device names or host names to connect
20649 with, process numbers, and baud rates.
20651 The @code{target} command does not repeat if you press @key{RET} again
20652 after executing the command.
20654 @kindex help target
20656 Displays the names of all targets available. To display targets
20657 currently selected, use either @code{info target} or @code{info files}
20658 (@pxref{Files, ,Commands to Specify Files}).
20660 @item help target @var{name}
20661 Describe a particular target, including any parameters necessary to
20664 @kindex set gnutarget
20665 @item set gnutarget @var{args}
20666 @value{GDBN} uses its own library BFD to read your files. @value{GDBN}
20667 knows whether it is reading an @dfn{executable},
20668 a @dfn{core}, or a @dfn{.o} file; however, you can specify the file format
20669 with the @code{set gnutarget} command. Unlike most @code{target} commands,
20670 with @code{gnutarget} the @code{target} refers to a program, not a machine.
20673 @emph{Warning:} To specify a file format with @code{set gnutarget},
20674 you must know the actual BFD name.
20678 @xref{Files, , Commands to Specify Files}.
20680 @kindex show gnutarget
20681 @item show gnutarget
20682 Use the @code{show gnutarget} command to display what file format
20683 @code{gnutarget} is set to read. If you have not set @code{gnutarget},
20684 @value{GDBN} will determine the file format for each file automatically,
20685 and @code{show gnutarget} displays @samp{The current BFD target is "auto"}.
20688 @cindex common targets
20689 Here are some common targets (available, or not, depending on the GDB
20694 @item target exec @var{program}
20695 @cindex executable file target
20696 An executable file. @samp{target exec @var{program}} is the same as
20697 @samp{exec-file @var{program}}.
20699 @item target core @var{filename}
20700 @cindex core dump file target
20701 A core dump file. @samp{target core @var{filename}} is the same as
20702 @samp{core-file @var{filename}}.
20704 @item target remote @var{medium}
20705 @cindex remote target
20706 A remote system connected to @value{GDBN} via a serial line or network
20707 connection. This command tells @value{GDBN} to use its own remote
20708 protocol over @var{medium} for debugging. @xref{Remote Debugging}.
20710 For example, if you have a board connected to @file{/dev/ttya} on the
20711 machine running @value{GDBN}, you could say:
20714 target remote /dev/ttya
20717 @code{target remote} supports the @code{load} command. This is only
20718 useful if you have some other way of getting the stub to the target
20719 system, and you can put it somewhere in memory where it won't get
20720 clobbered by the download.
20722 @item target sim @r{[}@var{simargs}@r{]} @dots{}
20723 @cindex built-in simulator target
20724 Builtin CPU simulator. @value{GDBN} includes simulators for most architectures.
20732 works; however, you cannot assume that a specific memory map, device
20733 drivers, or even basic I/O is available, although some simulators do
20734 provide these. For info about any processor-specific simulator details,
20735 see the appropriate section in @ref{Embedded Processors, ,Embedded
20738 @item target native
20739 @cindex native target
20740 Setup for local/native process debugging. Useful to make the
20741 @code{run} command spawn native processes (likewise @code{attach},
20742 etc.@:) even when @code{set auto-connect-native-target} is @code{off}
20743 (@pxref{set auto-connect-native-target}).
20747 Different targets are available on different configurations of @value{GDBN};
20748 your configuration may have more or fewer targets.
20750 Many remote targets require you to download the executable's code once
20751 you've successfully established a connection. You may wish to control
20752 various aspects of this process.
20757 @kindex set hash@r{, for remote monitors}
20758 @cindex hash mark while downloading
20759 This command controls whether a hash mark @samp{#} is displayed while
20760 downloading a file to the remote monitor. If on, a hash mark is
20761 displayed after each S-record is successfully downloaded to the
20765 @kindex show hash@r{, for remote monitors}
20766 Show the current status of displaying the hash mark.
20768 @item set debug monitor
20769 @kindex set debug monitor
20770 @cindex display remote monitor communications
20771 Enable or disable display of communications messages between
20772 @value{GDBN} and the remote monitor.
20774 @item show debug monitor
20775 @kindex show debug monitor
20776 Show the current status of displaying communications between
20777 @value{GDBN} and the remote monitor.
20782 @kindex load @var{filename} @var{offset}
20783 @item load @var{filename} @var{offset}
20785 Depending on what remote debugging facilities are configured into
20786 @value{GDBN}, the @code{load} command may be available. Where it exists, it
20787 is meant to make @var{filename} (an executable) available for debugging
20788 on the remote system---by downloading, or dynamic linking, for example.
20789 @code{load} also records the @var{filename} symbol table in @value{GDBN}, like
20790 the @code{add-symbol-file} command.
20792 If your @value{GDBN} does not have a @code{load} command, attempting to
20793 execute it gets the error message ``@code{You can't do that when your
20794 target is @dots{}}''
20796 The file is loaded at whatever address is specified in the executable.
20797 For some object file formats, you can specify the load address when you
20798 link the program; for other formats, like a.out, the object file format
20799 specifies a fixed address.
20800 @c FIXME! This would be a good place for an xref to the GNU linker doc.
20802 It is also possible to tell @value{GDBN} to load the executable file at a
20803 specific offset described by the optional argument @var{offset}. When
20804 @var{offset} is provided, @var{filename} must also be provided.
20806 Depending on the remote side capabilities, @value{GDBN} may be able to
20807 load programs into flash memory.
20809 @code{load} does not repeat if you press @key{RET} again after using it.
20814 @kindex flash-erase
20816 @anchor{flash-erase}
20818 Erases all known flash memory regions on the target.
20823 @section Choosing Target Byte Order
20825 @cindex choosing target byte order
20826 @cindex target byte order
20828 Some types of processors, such as the @acronym{MIPS}, PowerPC, and Renesas SH,
20829 offer the ability to run either big-endian or little-endian byte
20830 orders. Usually the executable or symbol will include a bit to
20831 designate the endian-ness, and you will not need to worry about
20832 which to use. However, you may still find it useful to adjust
20833 @value{GDBN}'s idea of processor endian-ness manually.
20837 @item set endian big
20838 Instruct @value{GDBN} to assume the target is big-endian.
20840 @item set endian little
20841 Instruct @value{GDBN} to assume the target is little-endian.
20843 @item set endian auto
20844 Instruct @value{GDBN} to use the byte order associated with the
20848 Display @value{GDBN}'s current idea of the target byte order.
20852 If the @code{set endian auto} mode is in effect and no executable has
20853 been selected, then the endianness used is the last one chosen either
20854 by one of the @code{set endian big} and @code{set endian little}
20855 commands or by inferring from the last executable used. If no
20856 endianness has been previously chosen, then the default for this mode
20857 is inferred from the target @value{GDBN} has been built for, and is
20858 @code{little} if the name of the target CPU has an @code{el} suffix
20859 and @code{big} otherwise.
20861 Note that these commands merely adjust interpretation of symbolic
20862 data on the host, and that they have absolutely no effect on the
20866 @node Remote Debugging
20867 @chapter Debugging Remote Programs
20868 @cindex remote debugging
20870 If you are trying to debug a program running on a machine that cannot run
20871 @value{GDBN} in the usual way, it is often useful to use remote debugging.
20872 For example, you might use remote debugging on an operating system kernel,
20873 or on a small system which does not have a general purpose operating system
20874 powerful enough to run a full-featured debugger.
20876 Some configurations of @value{GDBN} have special serial or TCP/IP interfaces
20877 to make this work with particular debugging targets. In addition,
20878 @value{GDBN} comes with a generic serial protocol (specific to @value{GDBN},
20879 but not specific to any particular target system) which you can use if you
20880 write the remote stubs---the code that runs on the remote system to
20881 communicate with @value{GDBN}.
20883 Other remote targets may be available in your
20884 configuration of @value{GDBN}; use @code{help target} to list them.
20887 * Connecting:: Connecting to a remote target
20888 * File Transfer:: Sending files to a remote system
20889 * Server:: Using the gdbserver program
20890 * Remote Configuration:: Remote configuration
20891 * Remote Stub:: Implementing a remote stub
20895 @section Connecting to a Remote Target
20896 @cindex remote debugging, connecting
20897 @cindex @code{gdbserver}, connecting
20898 @cindex remote debugging, types of connections
20899 @cindex @code{gdbserver}, types of connections
20900 @cindex @code{gdbserver}, @code{target remote} mode
20901 @cindex @code{gdbserver}, @code{target extended-remote} mode
20903 This section describes how to connect to a remote target, including the
20904 types of connections and their differences, how to set up executable and
20905 symbol files on the host and target, and the commands used for
20906 connecting to and disconnecting from the remote target.
20908 @subsection Types of Remote Connections
20910 @value{GDBN} supports two types of remote connections, @code{target remote}
20911 mode and @code{target extended-remote} mode. Note that many remote targets
20912 support only @code{target remote} mode. There are several major
20913 differences between the two types of connections, enumerated here:
20917 @cindex remote debugging, detach and program exit
20918 @item Result of detach or program exit
20919 @strong{With target remote mode:} When the debugged program exits or you
20920 detach from it, @value{GDBN} disconnects from the target. When using
20921 @code{gdbserver}, @code{gdbserver} will exit.
20923 @strong{With target extended-remote mode:} When the debugged program exits or
20924 you detach from it, @value{GDBN} remains connected to the target, even
20925 though no program is running. You can rerun the program, attach to a
20926 running program, or use @code{monitor} commands specific to the target.
20928 When using @code{gdbserver} in this case, it does not exit unless it was
20929 invoked using the @option{--once} option. If the @option{--once} option
20930 was not used, you can ask @code{gdbserver} to exit using the
20931 @code{monitor exit} command (@pxref{Monitor Commands for gdbserver}).
20933 @item Specifying the program to debug
20934 For both connection types you use the @code{file} command to specify the
20935 program on the host system. If you are using @code{gdbserver} there are
20936 some differences in how to specify the location of the program on the
20939 @strong{With target remote mode:} You must either specify the program to debug
20940 on the @code{gdbserver} command line or use the @option{--attach} option
20941 (@pxref{Attaching to a program,,Attaching to a Running Program}).
20943 @cindex @option{--multi}, @code{gdbserver} option
20944 @strong{With target extended-remote mode:} You may specify the program to debug
20945 on the @code{gdbserver} command line, or you can load the program or attach
20946 to it using @value{GDBN} commands after connecting to @code{gdbserver}.
20948 @anchor{--multi Option in Types of Remote Connnections}
20949 You can start @code{gdbserver} without supplying an initial command to run
20950 or process ID to attach. To do this, use the @option{--multi} command line
20951 option. Then you can connect using @code{target extended-remote} and start
20952 the program you want to debug (see below for details on using the
20953 @code{run} command in this scenario). Note that the conditions under which
20954 @code{gdbserver} terminates depend on how @value{GDBN} connects to it
20955 (@code{target remote} or @code{target extended-remote}). The
20956 @option{--multi} option to @code{gdbserver} has no influence on that.
20958 @item The @code{run} command
20959 @strong{With target remote mode:} The @code{run} command is not
20960 supported. Once a connection has been established, you can use all
20961 the usual @value{GDBN} commands to examine and change data. The
20962 remote program is already running, so you can use commands like
20963 @kbd{step} and @kbd{continue}.
20965 @strong{With target extended-remote mode:} The @code{run} command is
20966 supported. The @code{run} command uses the value set by
20967 @code{set remote exec-file} (@pxref{set remote exec-file}) to select
20968 the program to run. Command line arguments are supported, except for
20969 wildcard expansion and I/O redirection (@pxref{Arguments}).
20971 If you specify the program to debug on the command line, then the
20972 @code{run} command is not required to start execution, and you can
20973 resume using commands like @kbd{step} and @kbd{continue} as with
20974 @code{target remote} mode.
20976 @anchor{Attaching in Types of Remote Connections}
20978 @strong{With target remote mode:} The @value{GDBN} command @code{attach} is
20979 not supported. To attach to a running program using @code{gdbserver}, you
20980 must use the @option{--attach} option (@pxref{Running gdbserver}).
20982 @strong{With target extended-remote mode:} To attach to a running program,
20983 you may use the @code{attach} command after the connection has been
20984 established. If you are using @code{gdbserver}, you may also invoke
20985 @code{gdbserver} using the @option{--attach} option
20986 (@pxref{Running gdbserver}).
20990 @anchor{Host and target files}
20991 @subsection Host and Target Files
20992 @cindex remote debugging, symbol files
20993 @cindex symbol files, remote debugging
20995 @value{GDBN}, running on the host, needs access to symbol and debugging
20996 information for your program running on the target. This requires
20997 access to an unstripped copy of your program, and possibly any associated
20998 symbol files. Note that this section applies equally to both @code{target
20999 remote} mode and @code{target extended-remote} mode.
21001 Some remote targets (@pxref{qXfer executable filename read}, and
21002 @pxref{Host I/O Packets}) allow @value{GDBN} to access program files over
21003 the same connection used to communicate with @value{GDBN}. With such a
21004 target, if the remote program is unstripped, the only command you need is
21005 @code{target remote} (or @code{target extended-remote}).
21007 If the remote program is stripped, or the target does not support remote
21008 program file access, start up @value{GDBN} using the name of the local
21009 unstripped copy of your program as the first argument, or use the
21010 @code{file} command. Use @code{set sysroot} to specify the location (on
21011 the host) of target libraries (unless your @value{GDBN} was compiled with
21012 the correct sysroot using @code{--with-sysroot}). Alternatively, you
21013 may use @code{set solib-search-path} to specify how @value{GDBN} locates
21016 The symbol file and target libraries must exactly match the executable
21017 and libraries on the target, with one exception: the files on the host
21018 system should not be stripped, even if the files on the target system
21019 are. Mismatched or missing files will lead to confusing results
21020 during debugging. On @sc{gnu}/Linux targets, mismatched or missing
21021 files may also prevent @code{gdbserver} from debugging multi-threaded
21024 @subsection Remote Connection Commands
21025 @cindex remote connection commands
21026 @value{GDBN} can communicate with the target over a serial line, a
21027 local Unix domain socket, or
21028 over an @acronym{IP} network using @acronym{TCP} or @acronym{UDP}. In
21029 each case, @value{GDBN} uses the same protocol for debugging your
21030 program; only the medium carrying the debugging packets varies. The
21031 @code{target remote} and @code{target extended-remote} commands
21032 establish a connection to the target. Both commands accept the same
21033 arguments, which indicate the medium to use:
21037 @item target remote @var{serial-device}
21038 @itemx target extended-remote @var{serial-device}
21039 @cindex serial line, @code{target remote}
21040 Use @var{serial-device} to communicate with the target. For example,
21041 to use a serial line connected to the device named @file{/dev/ttyb}:
21044 target remote /dev/ttyb
21047 If you're using a serial line, you may want to give @value{GDBN} the
21048 @samp{--baud} option, or use the @code{set serial baud} command
21049 (@pxref{Remote Configuration, set serial baud}) before the
21050 @code{target} command.
21052 @item target remote @var{local-socket}
21053 @itemx target extended-remote @var{local-socket}
21054 @cindex local socket, @code{target remote}
21055 @cindex Unix domain socket
21056 Use @var{local-socket} to communicate with the target. For example,
21057 to use a local Unix domain socket bound to the file system entry @file{/tmp/gdb-socket0}:
21060 target remote /tmp/gdb-socket0
21063 Note that this command has the same form as the command to connect
21064 to a serial line. @value{GDBN} will automatically determine which
21065 kind of file you have specified and will make the appropriate kind
21067 This feature is not available if the host system does not support
21068 Unix domain sockets.
21070 @item target remote @code{@var{host}:@var{port}}
21071 @itemx target remote @code{@var{[host]}:@var{port}}
21072 @itemx target remote @code{tcp:@var{host}:@var{port}}
21073 @itemx target remote @code{tcp:@var{[host]}:@var{port}}
21074 @itemx target remote @code{tcp4:@var{host}:@var{port}}
21075 @itemx target remote @code{tcp6:@var{host}:@var{port}}
21076 @itemx target remote @code{tcp6:@var{[host]}:@var{port}}
21077 @itemx target extended-remote @code{@var{host}:@var{port}}
21078 @itemx target extended-remote @code{@var{[host]}:@var{port}}
21079 @itemx target extended-remote @code{tcp:@var{host}:@var{port}}
21080 @itemx target extended-remote @code{tcp:@var{[host]}:@var{port}}
21081 @itemx target extended-remote @code{tcp4:@var{host}:@var{port}}
21082 @itemx target extended-remote @code{tcp6:@var{host}:@var{port}}
21083 @itemx target extended-remote @code{tcp6:@var{[host]}:@var{port}}
21084 @cindex @acronym{TCP} port, @code{target remote}
21085 Debug using a @acronym{TCP} connection to @var{port} on @var{host}.
21086 The @var{host} may be either a host name, a numeric @acronym{IPv4}
21087 address, or a numeric @acronym{IPv6} address (with or without the
21088 square brackets to separate the address from the port); @var{port}
21089 must be a decimal number. The @var{host} could be the target machine
21090 itself, if it is directly connected to the net, or it might be a
21091 terminal server which in turn has a serial line to the target.
21093 For example, to connect to port 2828 on a terminal server named
21097 target remote manyfarms:2828
21100 To connect to port 2828 on a terminal server whose address is
21101 @code{2001:0db8:85a3:0000:0000:8a2e:0370:7334}, you can either use the
21102 square bracket syntax:
21105 target remote [2001:0db8:85a3:0000:0000:8a2e:0370:7334]:2828
21109 or explicitly specify the @acronym{IPv6} protocol:
21112 target remote tcp6:2001:0db8:85a3:0000:0000:8a2e:0370:7334:2828
21115 This last example may be confusing to the reader, because there is no
21116 visible separation between the hostname and the port number.
21117 Therefore, we recommend the user to provide @acronym{IPv6} addresses
21118 using square brackets for clarity. However, it is important to
21119 mention that for @value{GDBN} there is no ambiguity: the number after
21120 the last colon is considered to be the port number.
21122 If your remote target is actually running on the same machine as your
21123 debugger session (e.g.@: a simulator for your target running on the
21124 same host), you can omit the hostname. For example, to connect to
21125 port 1234 on your local machine:
21128 target remote :1234
21132 Note that the colon is still required here.
21134 @item target remote @code{udp:@var{host}:@var{port}}
21135 @itemx target remote @code{udp:@var{[host]}:@var{port}}
21136 @itemx target remote @code{udp4:@var{host}:@var{port}}
21137 @itemx target remote @code{udp6:@var{[host]}:@var{port}}
21138 @itemx target extended-remote @code{udp:@var{host}:@var{port}}
21139 @itemx target extended-remote @code{udp:@var{host}:@var{port}}
21140 @itemx target extended-remote @code{udp:@var{[host]}:@var{port}}
21141 @itemx target extended-remote @code{udp4:@var{host}:@var{port}}
21142 @itemx target extended-remote @code{udp6:@var{host}:@var{port}}
21143 @itemx target extended-remote @code{udp6:@var{[host]}:@var{port}}
21144 @cindex @acronym{UDP} port, @code{target remote}
21145 Debug using @acronym{UDP} packets to @var{port} on @var{host}. For example, to
21146 connect to @acronym{UDP} port 2828 on a terminal server named @code{manyfarms}:
21149 target remote udp:manyfarms:2828
21152 When using a @acronym{UDP} connection for remote debugging, you should
21153 keep in mind that the `U' stands for ``Unreliable''. @acronym{UDP}
21154 can silently drop packets on busy or unreliable networks, which will
21155 cause havoc with your debugging session.
21157 @item target remote | @var{command}
21158 @itemx target extended-remote | @var{command}
21159 @cindex pipe, @code{target remote} to
21160 Run @var{command} in the background and communicate with it using a
21161 pipe. The @var{command} is a shell command, to be parsed and expanded
21162 by the system's command shell, @code{/bin/sh}; it should expect remote
21163 protocol packets on its standard input, and send replies on its
21164 standard output. You could use this to run a stand-alone simulator
21165 that speaks the remote debugging protocol, to make net connections
21166 using programs like @code{ssh}, or for other similar tricks.
21168 If @var{command} closes its standard output (perhaps by exiting),
21169 @value{GDBN} will try to send it a @code{SIGTERM} signal. (If the
21170 program has already exited, this will have no effect.)
21174 @cindex interrupting remote programs
21175 @cindex remote programs, interrupting
21176 Whenever @value{GDBN} is waiting for the remote program, if you type the
21177 interrupt character (often @kbd{Ctrl-c}), @value{GDBN} attempts to stop the
21178 program. This may or may not succeed, depending in part on the hardware
21179 and the serial drivers the remote system uses. If you type the
21180 interrupt character once again, @value{GDBN} displays this prompt:
21183 Interrupted while waiting for the program.
21184 Give up (and stop debugging it)? (y or n)
21187 In @code{target remote} mode, if you type @kbd{y}, @value{GDBN} abandons
21188 the remote debugging session. (If you decide you want to try again later,
21189 you can use @kbd{target remote} again to connect once more.) If you type
21190 @kbd{n}, @value{GDBN} goes back to waiting.
21192 In @code{target extended-remote} mode, typing @kbd{n} will leave
21193 @value{GDBN} connected to the target.
21196 @kindex detach (remote)
21198 When you have finished debugging the remote program, you can use the
21199 @code{detach} command to release it from @value{GDBN} control.
21200 Detaching from the target normally resumes its execution, but the results
21201 will depend on your particular remote stub. After the @code{detach}
21202 command in @code{target remote} mode, @value{GDBN} is free to connect to
21203 another target. In @code{target extended-remote} mode, @value{GDBN} is
21204 still connected to the target.
21208 The @code{disconnect} command closes the connection to the target, and
21209 the target is generally not resumed. It will wait for @value{GDBN}
21210 (this instance or another one) to connect and continue debugging. After
21211 the @code{disconnect} command, @value{GDBN} is again free to connect to
21214 @cindex send command to remote monitor
21215 @cindex extend @value{GDBN} for remote targets
21216 @cindex add new commands for external monitor
21218 @item monitor @var{cmd}
21219 This command allows you to send arbitrary commands directly to the
21220 remote monitor. Since @value{GDBN} doesn't care about the commands it
21221 sends like this, this command is the way to extend @value{GDBN}---you
21222 can add new commands that only the external monitor will understand
21226 @node File Transfer
21227 @section Sending files to a remote system
21228 @cindex remote target, file transfer
21229 @cindex file transfer
21230 @cindex sending files to remote systems
21232 Some remote targets offer the ability to transfer files over the same
21233 connection used to communicate with @value{GDBN}. This is convenient
21234 for targets accessible through other means, e.g.@: @sc{gnu}/Linux systems
21235 running @code{gdbserver} over a network interface. For other targets,
21236 e.g.@: embedded devices with only a single serial port, this may be
21237 the only way to upload or download files.
21239 Not all remote targets support these commands.
21243 @item remote put @var{hostfile} @var{targetfile}
21244 Copy file @var{hostfile} from the host system (the machine running
21245 @value{GDBN}) to @var{targetfile} on the target system.
21248 @item remote get @var{targetfile} @var{hostfile}
21249 Copy file @var{targetfile} from the target system to @var{hostfile}
21250 on the host system.
21252 @kindex remote delete
21253 @item remote delete @var{targetfile}
21254 Delete @var{targetfile} from the target system.
21259 @section Using the @code{gdbserver} Program
21262 @cindex remote connection without stubs
21263 @code{gdbserver} is a control program for Unix-like systems, which
21264 allows you to connect your program with a remote @value{GDBN} via
21265 @code{target remote} or @code{target extended-remote}---but without
21266 linking in the usual debugging stub.
21268 @code{gdbserver} is not a complete replacement for the debugging stubs,
21269 because it requires essentially the same operating-system facilities
21270 that @value{GDBN} itself does. In fact, a system that can run
21271 @code{gdbserver} to connect to a remote @value{GDBN} could also run
21272 @value{GDBN} locally! @code{gdbserver} is sometimes useful nevertheless,
21273 because it is a much smaller program than @value{GDBN} itself. It is
21274 also easier to port than all of @value{GDBN}, so you may be able to get
21275 started more quickly on a new system by using @code{gdbserver}.
21276 Finally, if you develop code for real-time systems, you may find that
21277 the tradeoffs involved in real-time operation make it more convenient to
21278 do as much development work as possible on another system, for example
21279 by cross-compiling. You can use @code{gdbserver} to make a similar
21280 choice for debugging.
21282 @value{GDBN} and @code{gdbserver} communicate via either a serial line
21283 or a TCP connection, using the standard @value{GDBN} remote serial
21287 @emph{Warning:} @code{gdbserver} does not have any built-in security.
21288 Do not run @code{gdbserver} connected to any public network; a
21289 @value{GDBN} connection to @code{gdbserver} provides access to the
21290 target system with the same privileges as the user running
21294 @anchor{Running gdbserver}
21295 @subsection Running @code{gdbserver}
21296 @cindex arguments, to @code{gdbserver}
21297 @cindex @code{gdbserver}, command-line arguments
21299 Run @code{gdbserver} on the target system. You need a copy of the
21300 program you want to debug, including any libraries it requires.
21301 @code{gdbserver} does not need your program's symbol table, so you can
21302 strip the program if necessary to save space. @value{GDBN} on the host
21303 system does all the symbol handling.
21305 To use the server, you must tell it how to communicate with @value{GDBN};
21306 the name of your program; and the arguments for your program. The usual
21310 target> gdbserver @var{comm} @var{program} [ @var{args} @dots{} ]
21313 @var{comm} is either a device name (to use a serial line), or a TCP
21314 hostname and portnumber, or @code{-} or @code{stdio} to use
21315 stdin/stdout of @code{gdbserver}.
21316 For example, to debug Emacs with the argument
21317 @samp{foo.txt} and communicate with @value{GDBN} over the serial port
21321 target> gdbserver /dev/com1 emacs foo.txt
21324 @code{gdbserver} waits passively for the host @value{GDBN} to communicate
21327 To use a TCP connection instead of a serial line:
21330 target> gdbserver host:2345 emacs foo.txt
21333 The only difference from the previous example is the first argument,
21334 specifying that you are communicating with the host @value{GDBN} via
21335 TCP. The @samp{host:2345} argument means that @code{gdbserver} is to
21336 expect a TCP connection from machine @samp{host} to local TCP port 2345.
21337 (Currently, the @samp{host} part is ignored.) You can choose any number
21338 you want for the port number as long as it does not conflict with any
21339 TCP ports already in use on the target system (for example, @code{23} is
21340 reserved for @code{telnet}).@footnote{If you choose a port number that
21341 conflicts with another service, @code{gdbserver} prints an error message
21342 and exits.} You must use the same port number with the host @value{GDBN}
21343 @code{target remote} command.
21345 The @code{stdio} connection is useful when starting @code{gdbserver}
21349 (gdb) target remote | ssh -T hostname gdbserver - hello
21352 The @samp{-T} option to ssh is provided because we don't need a remote pty,
21353 and we don't want escape-character handling. Ssh does this by default when
21354 a command is provided, the flag is provided to make it explicit.
21355 You could elide it if you want to.
21357 Programs started with stdio-connected gdbserver have @file{/dev/null} for
21358 @code{stdin}, and @code{stdout},@code{stderr} are sent back to gdb for
21359 display through a pipe connected to gdbserver.
21360 Both @code{stdout} and @code{stderr} use the same pipe.
21362 @anchor{Attaching to a program}
21363 @subsubsection Attaching to a Running Program
21364 @cindex attach to a program, @code{gdbserver}
21365 @cindex @option{--attach}, @code{gdbserver} option
21367 On some targets, @code{gdbserver} can also attach to running programs.
21368 This is accomplished via the @code{--attach} argument. The syntax is:
21371 target> gdbserver --attach @var{comm} @var{pid}
21374 @var{pid} is the process ID of a currently running process. It isn't
21375 necessary to point @code{gdbserver} at a binary for the running process.
21377 In @code{target extended-remote} mode, you can also attach using the
21378 @value{GDBN} attach command
21379 (@pxref{Attaching in Types of Remote Connections}).
21382 You can debug processes by name instead of process ID if your target has the
21383 @code{pidof} utility:
21386 target> gdbserver --attach @var{comm} `pidof @var{program}`
21389 In case more than one copy of @var{program} is running, or @var{program}
21390 has multiple threads, most versions of @code{pidof} support the
21391 @code{-s} option to only return the first process ID.
21393 @subsubsection TCP port allocation lifecycle of @code{gdbserver}
21395 This section applies only when @code{gdbserver} is run to listen on a TCP
21398 @code{gdbserver} normally terminates after all of its debugged processes have
21399 terminated in @kbd{target remote} mode. On the other hand, for @kbd{target
21400 extended-remote}, @code{gdbserver} stays running even with no processes left.
21401 @value{GDBN} normally terminates the spawned debugged process on its exit,
21402 which normally also terminates @code{gdbserver} in the @kbd{target remote}
21403 mode. Therefore, when the connection drops unexpectedly, and @value{GDBN}
21404 cannot ask @code{gdbserver} to kill its debugged processes, @code{gdbserver}
21405 stays running even in the @kbd{target remote} mode.
21407 When @code{gdbserver} stays running, @value{GDBN} can connect to it again later.
21408 Such reconnecting is useful for features like @ref{disconnected tracing}. For
21409 completeness, at most one @value{GDBN} can be connected at a time.
21411 @cindex @option{--once}, @code{gdbserver} option
21412 By default, @code{gdbserver} keeps the listening TCP port open, so that
21413 subsequent connections are possible. However, if you start @code{gdbserver}
21414 with the @option{--once} option, it will stop listening for any further
21415 connection attempts after connecting to the first @value{GDBN} session. This
21416 means no further connections to @code{gdbserver} will be possible after the
21417 first one. It also means @code{gdbserver} will terminate after the first
21418 connection with remote @value{GDBN} has closed, even for unexpectedly closed
21419 connections and even in the @kbd{target extended-remote} mode. The
21420 @option{--once} option allows reusing the same port number for connecting to
21421 multiple instances of @code{gdbserver} running on the same host, since each
21422 instance closes its port after the first connection.
21424 @anchor{Other Command-Line Arguments for gdbserver}
21425 @subsubsection Other Command-Line Arguments for @code{gdbserver}
21427 You can use the @option{--multi} option to start @code{gdbserver} without
21428 specifying a program to debug or a process to attach to. Then you can
21429 attach in @code{target extended-remote} mode and run or attach to a
21430 program. For more information,
21431 @pxref{--multi Option in Types of Remote Connnections}.
21433 @cindex @option{--debug}, @code{gdbserver} option
21434 The @option{--debug} option tells @code{gdbserver} to display extra
21435 status information about the debugging process.
21436 @cindex @option{--remote-debug}, @code{gdbserver} option
21437 The @option{--remote-debug} option tells @code{gdbserver} to display
21438 remote protocol debug output.
21439 @cindex @option{--debug-file}, @code{gdbserver} option
21440 @cindex @code{gdbserver}, send all debug output to a single file
21441 The @option{--debug-file=@var{filename}} option tells @code{gdbserver} to
21442 write any debug output to the given @var{filename}. These options are intended
21443 for @code{gdbserver} development and for bug reports to the developers.
21445 @cindex @option{--debug-format}, @code{gdbserver} option
21446 The @option{--debug-format=option1[,option2,...]} option tells
21447 @code{gdbserver} to include additional information in each output.
21448 Possible options are:
21452 Turn off all extra information in debugging output.
21454 Turn on all extra information in debugging output.
21456 Include a timestamp in each line of debugging output.
21459 Options are processed in order. Thus, for example, if @option{none}
21460 appears last then no additional information is added to debugging output.
21462 @cindex @option{--wrapper}, @code{gdbserver} option
21463 The @option{--wrapper} option specifies a wrapper to launch programs
21464 for debugging. The option should be followed by the name of the
21465 wrapper, then any command-line arguments to pass to the wrapper, then
21466 @kbd{--} indicating the end of the wrapper arguments.
21468 @code{gdbserver} runs the specified wrapper program with a combined
21469 command line including the wrapper arguments, then the name of the
21470 program to debug, then any arguments to the program. The wrapper
21471 runs until it executes your program, and then @value{GDBN} gains control.
21473 You can use any program that eventually calls @code{execve} with
21474 its arguments as a wrapper. Several standard Unix utilities do
21475 this, e.g.@: @code{env} and @code{nohup}. Any Unix shell script ending
21476 with @code{exec "$@@"} will also work.
21478 For example, you can use @code{env} to pass an environment variable to
21479 the debugged program, without setting the variable in @code{gdbserver}'s
21483 $ gdbserver --wrapper env LD_PRELOAD=libtest.so -- :2222 ./testprog
21486 @cindex @option{--selftest}
21487 The @option{--selftest} option runs the self tests in @code{gdbserver}:
21490 $ gdbserver --selftest
21491 Ran 2 unit tests, 0 failed
21494 These tests are disabled in release.
21495 @subsection Connecting to @code{gdbserver}
21497 The basic procedure for connecting to the remote target is:
21501 Run @value{GDBN} on the host system.
21504 Make sure you have the necessary symbol files
21505 (@pxref{Host and target files}).
21506 Load symbols for your application using the @code{file} command before you
21507 connect. Use @code{set sysroot} to locate target libraries (unless your
21508 @value{GDBN} was compiled with the correct sysroot using
21509 @code{--with-sysroot}).
21512 Connect to your target (@pxref{Connecting,,Connecting to a Remote Target}).
21513 For TCP connections, you must start up @code{gdbserver} prior to using
21514 the @code{target} command. Otherwise you may get an error whose
21515 text depends on the host system, but which usually looks something like
21516 @samp{Connection refused}. Don't use the @code{load}
21517 command in @value{GDBN} when using @code{target remote} mode, since the
21518 program is already on the target.
21522 @anchor{Monitor Commands for gdbserver}
21523 @subsection Monitor Commands for @code{gdbserver}
21524 @cindex monitor commands, for @code{gdbserver}
21526 During a @value{GDBN} session using @code{gdbserver}, you can use the
21527 @code{monitor} command to send special requests to @code{gdbserver}.
21528 Here are the available commands.
21532 List the available monitor commands.
21534 @item monitor set debug 0
21535 @itemx monitor set debug 1
21536 Disable or enable general debugging messages.
21538 @item monitor set remote-debug 0
21539 @itemx monitor set remote-debug 1
21540 Disable or enable specific debugging messages associated with the remote
21541 protocol (@pxref{Remote Protocol}).
21543 @item monitor set debug-file filename
21544 @itemx monitor set debug-file
21545 Send any debug output to the given file, or to stderr.
21547 @item monitor set debug-format option1@r{[},option2,...@r{]}
21548 Specify additional text to add to debugging messages.
21549 Possible options are:
21553 Turn off all extra information in debugging output.
21555 Turn on all extra information in debugging output.
21557 Include a timestamp in each line of debugging output.
21560 Options are processed in order. Thus, for example, if @option{none}
21561 appears last then no additional information is added to debugging output.
21563 @item monitor set libthread-db-search-path [PATH]
21564 @cindex gdbserver, search path for @code{libthread_db}
21565 When this command is issued, @var{path} is a colon-separated list of
21566 directories to search for @code{libthread_db} (@pxref{Threads,,set
21567 libthread-db-search-path}). If you omit @var{path},
21568 @samp{libthread-db-search-path} will be reset to its default value.
21570 The special entry @samp{$pdir} for @samp{libthread-db-search-path} is
21571 not supported in @code{gdbserver}.
21574 Tell gdbserver to exit immediately. This command should be followed by
21575 @code{disconnect} to close the debugging session. @code{gdbserver} will
21576 detach from any attached processes and kill any processes it created.
21577 Use @code{monitor exit} to terminate @code{gdbserver} at the end
21578 of a multi-process mode debug session.
21582 @subsection Tracepoints support in @code{gdbserver}
21583 @cindex tracepoints support in @code{gdbserver}
21585 On some targets, @code{gdbserver} supports tracepoints, fast
21586 tracepoints and static tracepoints.
21588 For fast or static tracepoints to work, a special library called the
21589 @dfn{in-process agent} (IPA), must be loaded in the inferior process.
21590 This library is built and distributed as an integral part of
21591 @code{gdbserver}. In addition, support for static tracepoints
21592 requires building the in-process agent library with static tracepoints
21593 support. At present, the UST (LTTng Userspace Tracer,
21594 @url{http://lttng.org/ust}) tracing engine is supported. This support
21595 is automatically available if UST development headers are found in the
21596 standard include path when @code{gdbserver} is built, or if
21597 @code{gdbserver} was explicitly configured using @option{--with-ust}
21598 to point at such headers. You can explicitly disable the support
21599 using @option{--with-ust=no}.
21601 There are several ways to load the in-process agent in your program:
21604 @item Specifying it as dependency at link time
21606 You can link your program dynamically with the in-process agent
21607 library. On most systems, this is accomplished by adding
21608 @code{-linproctrace} to the link command.
21610 @item Using the system's preloading mechanisms
21612 You can force loading the in-process agent at startup time by using
21613 your system's support for preloading shared libraries. Many Unixes
21614 support the concept of preloading user defined libraries. In most
21615 cases, you do that by specifying @code{LD_PRELOAD=libinproctrace.so}
21616 in the environment. See also the description of @code{gdbserver}'s
21617 @option{--wrapper} command line option.
21619 @item Using @value{GDBN} to force loading the agent at run time
21621 On some systems, you can force the inferior to load a shared library,
21622 by calling a dynamic loader function in the inferior that takes care
21623 of dynamically looking up and loading a shared library. On most Unix
21624 systems, the function is @code{dlopen}. You'll use the @code{call}
21625 command for that. For example:
21628 (@value{GDBP}) call dlopen ("libinproctrace.so", ...)
21631 Note that on most Unix systems, for the @code{dlopen} function to be
21632 available, the program needs to be linked with @code{-ldl}.
21635 On systems that have a userspace dynamic loader, like most Unix
21636 systems, when you connect to @code{gdbserver} using @code{target
21637 remote}, you'll find that the program is stopped at the dynamic
21638 loader's entry point, and no shared library has been loaded in the
21639 program's address space yet, including the in-process agent. In that
21640 case, before being able to use any of the fast or static tracepoints
21641 features, you need to let the loader run and load the shared
21642 libraries. The simplest way to do that is to run the program to the
21643 main procedure. E.g., if debugging a C or C@t{++} program, start
21644 @code{gdbserver} like so:
21647 $ gdbserver :9999 myprogram
21650 Start GDB and connect to @code{gdbserver} like so, and run to main:
21654 (@value{GDBP}) target remote myhost:9999
21655 0x00007f215893ba60 in ?? () from /lib64/ld-linux-x86-64.so.2
21656 (@value{GDBP}) b main
21657 (@value{GDBP}) continue
21660 The in-process tracing agent library should now be loaded into the
21661 process; you can confirm it with the @code{info sharedlibrary}
21662 command, which will list @file{libinproctrace.so} as loaded in the
21663 process. You are now ready to install fast tracepoints, list static
21664 tracepoint markers, probe static tracepoints markers, and start
21667 @node Remote Configuration
21668 @section Remote Configuration
21671 @kindex show remote
21672 This section documents the configuration options available when
21673 debugging remote programs. For the options related to the File I/O
21674 extensions of the remote protocol, see @ref{system,
21675 system-call-allowed}.
21678 @item set remoteaddresssize @var{bits}
21679 @cindex address size for remote targets
21680 @cindex bits in remote address
21681 Set the maximum size of address in a memory packet to the specified
21682 number of bits. @value{GDBN} will mask off the address bits above
21683 that number, when it passes addresses to the remote target. The
21684 default value is the number of bits in the target's address.
21686 @item show remoteaddresssize
21687 Show the current value of remote address size in bits.
21689 @item set serial baud @var{n}
21690 @cindex baud rate for remote targets
21691 Set the baud rate for the remote serial I/O to @var{n} baud. The
21692 value is used to set the speed of the serial port used for debugging
21695 @item show serial baud
21696 Show the current speed of the remote connection.
21698 @item set serial parity @var{parity}
21699 Set the parity for the remote serial I/O. Supported values of @var{parity} are:
21700 @code{even}, @code{none}, and @code{odd}. The default is @code{none}.
21702 @item show serial parity
21703 Show the current parity of the serial port.
21705 @item set remotebreak
21706 @cindex interrupt remote programs
21707 @cindex BREAK signal instead of Ctrl-C
21708 @anchor{set remotebreak}
21709 If set to on, @value{GDBN} sends a @code{BREAK} signal to the remote
21710 when you type @kbd{Ctrl-c} to interrupt the program running
21711 on the remote. If set to off, @value{GDBN} sends the @samp{Ctrl-C}
21712 character instead. The default is off, since most remote systems
21713 expect to see @samp{Ctrl-C} as the interrupt signal.
21715 @item show remotebreak
21716 Show whether @value{GDBN} sends @code{BREAK} or @samp{Ctrl-C} to
21717 interrupt the remote program.
21719 @item set remoteflow on
21720 @itemx set remoteflow off
21721 @kindex set remoteflow
21722 Enable or disable hardware flow control (@code{RTS}/@code{CTS})
21723 on the serial port used to communicate to the remote target.
21725 @item show remoteflow
21726 @kindex show remoteflow
21727 Show the current setting of hardware flow control.
21729 @item set remotelogbase @var{base}
21730 Set the base (a.k.a.@: radix) of logging serial protocol
21731 communications to @var{base}. Supported values of @var{base} are:
21732 @code{ascii}, @code{octal}, and @code{hex}. The default is
21735 @item show remotelogbase
21736 Show the current setting of the radix for logging remote serial
21739 @item set remotelogfile @var{file}
21740 @cindex record serial communications on file
21741 Record remote serial communications on the named @var{file}. The
21742 default is not to record at all.
21744 @item show remotelogfile
21745 Show the current setting of the file name on which to record the
21746 serial communications.
21748 @item set remotetimeout @var{num}
21749 @cindex timeout for serial communications
21750 @cindex remote timeout
21751 Set the timeout limit to wait for the remote target to respond to
21752 @var{num} seconds. The default is 2 seconds.
21754 @item show remotetimeout
21755 Show the current number of seconds to wait for the remote target
21758 @cindex limit hardware breakpoints and watchpoints
21759 @cindex remote target, limit break- and watchpoints
21760 @anchor{set remote hardware-watchpoint-limit}
21761 @anchor{set remote hardware-breakpoint-limit}
21762 @item set remote hardware-watchpoint-limit @var{limit}
21763 @itemx set remote hardware-breakpoint-limit @var{limit}
21764 Restrict @value{GDBN} to using @var{limit} remote hardware watchpoints
21765 or breakpoints. The @var{limit} can be set to 0 to disable hardware
21766 watchpoints or breakpoints, and @code{unlimited} for unlimited
21767 watchpoints or breakpoints.
21769 @item show remote hardware-watchpoint-limit
21770 @itemx show remote hardware-breakpoint-limit
21771 Show the current limit for the number of hardware watchpoints or
21772 breakpoints that @value{GDBN} can use.
21774 @cindex limit hardware watchpoints length
21775 @cindex remote target, limit watchpoints length
21776 @anchor{set remote hardware-watchpoint-length-limit}
21777 @item set remote hardware-watchpoint-length-limit @var{limit}
21778 Restrict @value{GDBN} to using @var{limit} bytes for the maximum
21779 length of a remote hardware watchpoint. A @var{limit} of 0 disables
21780 hardware watchpoints and @code{unlimited} allows watchpoints of any
21783 @item show remote hardware-watchpoint-length-limit
21784 Show the current limit (in bytes) of the maximum length of
21785 a remote hardware watchpoint.
21787 @item set remote exec-file @var{filename}
21788 @itemx show remote exec-file
21789 @anchor{set remote exec-file}
21790 @cindex executable file, for remote target
21791 Select the file used for @code{run} with @code{target
21792 extended-remote}. This should be set to a filename valid on the
21793 target system. If it is not set, the target will use a default
21794 filename (e.g.@: the last program run).
21796 @item set remote interrupt-sequence
21797 @cindex interrupt remote programs
21798 @cindex select Ctrl-C, BREAK or BREAK-g
21799 Allow the user to select one of @samp{Ctrl-C}, a @code{BREAK} or
21800 @samp{BREAK-g} as the
21801 sequence to the remote target in order to interrupt the execution.
21802 @samp{Ctrl-C} is a default. Some system prefers @code{BREAK} which
21803 is high level of serial line for some certain time.
21804 Linux kernel prefers @samp{BREAK-g}, a.k.a Magic SysRq g.
21805 It is @code{BREAK} signal followed by character @code{g}.
21807 @item show interrupt-sequence
21808 Show which of @samp{Ctrl-C}, @code{BREAK} or @code{BREAK-g}
21809 is sent by @value{GDBN} to interrupt the remote program.
21810 @code{BREAK-g} is BREAK signal followed by @code{g} and
21811 also known as Magic SysRq g.
21813 @item set remote interrupt-on-connect
21814 @cindex send interrupt-sequence on start
21815 Specify whether interrupt-sequence is sent to remote target when
21816 @value{GDBN} connects to it. This is mostly needed when you debug
21817 Linux kernel. Linux kernel expects @code{BREAK} followed by @code{g}
21818 which is known as Magic SysRq g in order to connect @value{GDBN}.
21820 @item show interrupt-on-connect
21821 Show whether interrupt-sequence is sent
21822 to remote target when @value{GDBN} connects to it.
21826 @item set tcp auto-retry on
21827 @cindex auto-retry, for remote TCP target
21828 Enable auto-retry for remote TCP connections. This is useful if the remote
21829 debugging agent is launched in parallel with @value{GDBN}; there is a race
21830 condition because the agent may not become ready to accept the connection
21831 before @value{GDBN} attempts to connect. When auto-retry is
21832 enabled, if the initial attempt to connect fails, @value{GDBN} reattempts
21833 to establish the connection using the timeout specified by
21834 @code{set tcp connect-timeout}.
21836 @item set tcp auto-retry off
21837 Do not auto-retry failed TCP connections.
21839 @item show tcp auto-retry
21840 Show the current auto-retry setting.
21842 @item set tcp connect-timeout @var{seconds}
21843 @itemx set tcp connect-timeout unlimited
21844 @cindex connection timeout, for remote TCP target
21845 @cindex timeout, for remote target connection
21846 Set the timeout for establishing a TCP connection to the remote target to
21847 @var{seconds}. The timeout affects both polling to retry failed connections
21848 (enabled by @code{set tcp auto-retry on}) and waiting for connections
21849 that are merely slow to complete, and represents an approximate cumulative
21850 value. If @var{seconds} is @code{unlimited}, there is no timeout and
21851 @value{GDBN} will keep attempting to establish a connection forever,
21852 unless interrupted with @kbd{Ctrl-c}. The default is 15 seconds.
21854 @item show tcp connect-timeout
21855 Show the current connection timeout setting.
21858 @cindex remote packets, enabling and disabling
21859 The @value{GDBN} remote protocol autodetects the packets supported by
21860 your debugging stub. If you need to override the autodetection, you
21861 can use these commands to enable or disable individual packets. Each
21862 packet can be set to @samp{on} (the remote target supports this
21863 packet), @samp{off} (the remote target does not support this packet),
21864 or @samp{auto} (detect remote target support for this packet). They
21865 all default to @samp{auto}. For more information about each packet,
21866 see @ref{Remote Protocol}.
21868 During normal use, you should not have to use any of these commands.
21869 If you do, that may be a bug in your remote debugging stub, or a bug
21870 in @value{GDBN}. You may want to report the problem to the
21871 @value{GDBN} developers.
21873 For each packet @var{name}, the command to enable or disable the
21874 packet is @code{set remote @var{name}-packet}. The available settings
21877 @multitable @columnfractions 0.28 0.32 0.25
21880 @tab Related Features
21882 @item @code{fetch-register}
21884 @tab @code{info registers}
21886 @item @code{set-register}
21890 @item @code{binary-download}
21892 @tab @code{load}, @code{set}
21894 @item @code{read-aux-vector}
21895 @tab @code{qXfer:auxv:read}
21896 @tab @code{info auxv}
21898 @item @code{symbol-lookup}
21899 @tab @code{qSymbol}
21900 @tab Detecting multiple threads
21902 @item @code{attach}
21903 @tab @code{vAttach}
21906 @item @code{verbose-resume}
21908 @tab Stepping or resuming multiple threads
21914 @item @code{software-breakpoint}
21918 @item @code{hardware-breakpoint}
21922 @item @code{write-watchpoint}
21926 @item @code{read-watchpoint}
21930 @item @code{access-watchpoint}
21934 @item @code{pid-to-exec-file}
21935 @tab @code{qXfer:exec-file:read}
21936 @tab @code{attach}, @code{run}
21938 @item @code{target-features}
21939 @tab @code{qXfer:features:read}
21940 @tab @code{set architecture}
21942 @item @code{library-info}
21943 @tab @code{qXfer:libraries:read}
21944 @tab @code{info sharedlibrary}
21946 @item @code{memory-map}
21947 @tab @code{qXfer:memory-map:read}
21948 @tab @code{info mem}
21950 @item @code{read-sdata-object}
21951 @tab @code{qXfer:sdata:read}
21952 @tab @code{print $_sdata}
21954 @item @code{read-spu-object}
21955 @tab @code{qXfer:spu:read}
21956 @tab @code{info spu}
21958 @item @code{write-spu-object}
21959 @tab @code{qXfer:spu:write}
21960 @tab @code{info spu}
21962 @item @code{read-siginfo-object}
21963 @tab @code{qXfer:siginfo:read}
21964 @tab @code{print $_siginfo}
21966 @item @code{write-siginfo-object}
21967 @tab @code{qXfer:siginfo:write}
21968 @tab @code{set $_siginfo}
21970 @item @code{threads}
21971 @tab @code{qXfer:threads:read}
21972 @tab @code{info threads}
21974 @item @code{get-thread-local-@*storage-address}
21975 @tab @code{qGetTLSAddr}
21976 @tab Displaying @code{__thread} variables
21978 @item @code{get-thread-information-block-address}
21979 @tab @code{qGetTIBAddr}
21980 @tab Display MS-Windows Thread Information Block.
21982 @item @code{search-memory}
21983 @tab @code{qSearch:memory}
21986 @item @code{supported-packets}
21987 @tab @code{qSupported}
21988 @tab Remote communications parameters
21990 @item @code{catch-syscalls}
21991 @tab @code{QCatchSyscalls}
21992 @tab @code{catch syscall}
21994 @item @code{pass-signals}
21995 @tab @code{QPassSignals}
21996 @tab @code{handle @var{signal}}
21998 @item @code{program-signals}
21999 @tab @code{QProgramSignals}
22000 @tab @code{handle @var{signal}}
22002 @item @code{hostio-close-packet}
22003 @tab @code{vFile:close}
22004 @tab @code{remote get}, @code{remote put}
22006 @item @code{hostio-open-packet}
22007 @tab @code{vFile:open}
22008 @tab @code{remote get}, @code{remote put}
22010 @item @code{hostio-pread-packet}
22011 @tab @code{vFile:pread}
22012 @tab @code{remote get}, @code{remote put}
22014 @item @code{hostio-pwrite-packet}
22015 @tab @code{vFile:pwrite}
22016 @tab @code{remote get}, @code{remote put}
22018 @item @code{hostio-unlink-packet}
22019 @tab @code{vFile:unlink}
22020 @tab @code{remote delete}
22022 @item @code{hostio-readlink-packet}
22023 @tab @code{vFile:readlink}
22026 @item @code{hostio-fstat-packet}
22027 @tab @code{vFile:fstat}
22030 @item @code{hostio-setfs-packet}
22031 @tab @code{vFile:setfs}
22034 @item @code{noack-packet}
22035 @tab @code{QStartNoAckMode}
22036 @tab Packet acknowledgment
22038 @item @code{osdata}
22039 @tab @code{qXfer:osdata:read}
22040 @tab @code{info os}
22042 @item @code{query-attached}
22043 @tab @code{qAttached}
22044 @tab Querying remote process attach state.
22046 @item @code{trace-buffer-size}
22047 @tab @code{QTBuffer:size}
22048 @tab @code{set trace-buffer-size}
22050 @item @code{trace-status}
22051 @tab @code{qTStatus}
22052 @tab @code{tstatus}
22054 @item @code{traceframe-info}
22055 @tab @code{qXfer:traceframe-info:read}
22056 @tab Traceframe info
22058 @item @code{install-in-trace}
22059 @tab @code{InstallInTrace}
22060 @tab Install tracepoint in tracing
22062 @item @code{disable-randomization}
22063 @tab @code{QDisableRandomization}
22064 @tab @code{set disable-randomization}
22066 @item @code{startup-with-shell}
22067 @tab @code{QStartupWithShell}
22068 @tab @code{set startup-with-shell}
22070 @item @code{environment-hex-encoded}
22071 @tab @code{QEnvironmentHexEncoded}
22072 @tab @code{set environment}
22074 @item @code{environment-unset}
22075 @tab @code{QEnvironmentUnset}
22076 @tab @code{unset environment}
22078 @item @code{environment-reset}
22079 @tab @code{QEnvironmentReset}
22080 @tab @code{Reset the inferior environment (i.e., unset user-set variables)}
22082 @item @code{set-working-dir}
22083 @tab @code{QSetWorkingDir}
22084 @tab @code{set cwd}
22086 @item @code{conditional-breakpoints-packet}
22087 @tab @code{Z0 and Z1}
22088 @tab @code{Support for target-side breakpoint condition evaluation}
22090 @item @code{multiprocess-extensions}
22091 @tab @code{multiprocess extensions}
22092 @tab Debug multiple processes and remote process PID awareness
22094 @item @code{swbreak-feature}
22095 @tab @code{swbreak stop reason}
22098 @item @code{hwbreak-feature}
22099 @tab @code{hwbreak stop reason}
22102 @item @code{fork-event-feature}
22103 @tab @code{fork stop reason}
22106 @item @code{vfork-event-feature}
22107 @tab @code{vfork stop reason}
22110 @item @code{exec-event-feature}
22111 @tab @code{exec stop reason}
22114 @item @code{thread-events}
22115 @tab @code{QThreadEvents}
22116 @tab Tracking thread lifetime.
22118 @item @code{no-resumed-stop-reply}
22119 @tab @code{no resumed thread left stop reply}
22120 @tab Tracking thread lifetime.
22125 @section Implementing a Remote Stub
22127 @cindex debugging stub, example
22128 @cindex remote stub, example
22129 @cindex stub example, remote debugging
22130 The stub files provided with @value{GDBN} implement the target side of the
22131 communication protocol, and the @value{GDBN} side is implemented in the
22132 @value{GDBN} source file @file{remote.c}. Normally, you can simply allow
22133 these subroutines to communicate, and ignore the details. (If you're
22134 implementing your own stub file, you can still ignore the details: start
22135 with one of the existing stub files. @file{sparc-stub.c} is the best
22136 organized, and therefore the easiest to read.)
22138 @cindex remote serial debugging, overview
22139 To debug a program running on another machine (the debugging
22140 @dfn{target} machine), you must first arrange for all the usual
22141 prerequisites for the program to run by itself. For example, for a C
22146 A startup routine to set up the C runtime environment; these usually
22147 have a name like @file{crt0}. The startup routine may be supplied by
22148 your hardware supplier, or you may have to write your own.
22151 A C subroutine library to support your program's
22152 subroutine calls, notably managing input and output.
22155 A way of getting your program to the other machine---for example, a
22156 download program. These are often supplied by the hardware
22157 manufacturer, but you may have to write your own from hardware
22161 The next step is to arrange for your program to use a serial port to
22162 communicate with the machine where @value{GDBN} is running (the @dfn{host}
22163 machine). In general terms, the scheme looks like this:
22167 @value{GDBN} already understands how to use this protocol; when everything
22168 else is set up, you can simply use the @samp{target remote} command
22169 (@pxref{Targets,,Specifying a Debugging Target}).
22171 @item On the target,
22172 you must link with your program a few special-purpose subroutines that
22173 implement the @value{GDBN} remote serial protocol. The file containing these
22174 subroutines is called a @dfn{debugging stub}.
22176 On certain remote targets, you can use an auxiliary program
22177 @code{gdbserver} instead of linking a stub into your program.
22178 @xref{Server,,Using the @code{gdbserver} Program}, for details.
22181 The debugging stub is specific to the architecture of the remote
22182 machine; for example, use @file{sparc-stub.c} to debug programs on
22185 @cindex remote serial stub list
22186 These working remote stubs are distributed with @value{GDBN}:
22191 @cindex @file{i386-stub.c}
22194 For Intel 386 and compatible architectures.
22197 @cindex @file{m68k-stub.c}
22198 @cindex Motorola 680x0
22200 For Motorola 680x0 architectures.
22203 @cindex @file{sh-stub.c}
22206 For Renesas SH architectures.
22209 @cindex @file{sparc-stub.c}
22211 For @sc{sparc} architectures.
22213 @item sparcl-stub.c
22214 @cindex @file{sparcl-stub.c}
22217 For Fujitsu @sc{sparclite} architectures.
22221 The @file{README} file in the @value{GDBN} distribution may list other
22222 recently added stubs.
22225 * Stub Contents:: What the stub can do for you
22226 * Bootstrapping:: What you must do for the stub
22227 * Debug Session:: Putting it all together
22230 @node Stub Contents
22231 @subsection What the Stub Can Do for You
22233 @cindex remote serial stub
22234 The debugging stub for your architecture supplies these three
22238 @item set_debug_traps
22239 @findex set_debug_traps
22240 @cindex remote serial stub, initialization
22241 This routine arranges for @code{handle_exception} to run when your
22242 program stops. You must call this subroutine explicitly in your
22243 program's startup code.
22245 @item handle_exception
22246 @findex handle_exception
22247 @cindex remote serial stub, main routine
22248 This is the central workhorse, but your program never calls it
22249 explicitly---the setup code arranges for @code{handle_exception} to
22250 run when a trap is triggered.
22252 @code{handle_exception} takes control when your program stops during
22253 execution (for example, on a breakpoint), and mediates communications
22254 with @value{GDBN} on the host machine. This is where the communications
22255 protocol is implemented; @code{handle_exception} acts as the @value{GDBN}
22256 representative on the target machine. It begins by sending summary
22257 information on the state of your program, then continues to execute,
22258 retrieving and transmitting any information @value{GDBN} needs, until you
22259 execute a @value{GDBN} command that makes your program resume; at that point,
22260 @code{handle_exception} returns control to your own code on the target
22264 @cindex @code{breakpoint} subroutine, remote
22265 Use this auxiliary subroutine to make your program contain a
22266 breakpoint. Depending on the particular situation, this may be the only
22267 way for @value{GDBN} to get control. For instance, if your target
22268 machine has some sort of interrupt button, you won't need to call this;
22269 pressing the interrupt button transfers control to
22270 @code{handle_exception}---in effect, to @value{GDBN}. On some machines,
22271 simply receiving characters on the serial port may also trigger a trap;
22272 again, in that situation, you don't need to call @code{breakpoint} from
22273 your own program---simply running @samp{target remote} from the host
22274 @value{GDBN} session gets control.
22276 Call @code{breakpoint} if none of these is true, or if you simply want
22277 to make certain your program stops at a predetermined point for the
22278 start of your debugging session.
22281 @node Bootstrapping
22282 @subsection What You Must Do for the Stub
22284 @cindex remote stub, support routines
22285 The debugging stubs that come with @value{GDBN} are set up for a particular
22286 chip architecture, but they have no information about the rest of your
22287 debugging target machine.
22289 First of all you need to tell the stub how to communicate with the
22293 @item int getDebugChar()
22294 @findex getDebugChar
22295 Write this subroutine to read a single character from the serial port.
22296 It may be identical to @code{getchar} for your target system; a
22297 different name is used to allow you to distinguish the two if you wish.
22299 @item void putDebugChar(int)
22300 @findex putDebugChar
22301 Write this subroutine to write a single character to the serial port.
22302 It may be identical to @code{putchar} for your target system; a
22303 different name is used to allow you to distinguish the two if you wish.
22306 @cindex control C, and remote debugging
22307 @cindex interrupting remote targets
22308 If you want @value{GDBN} to be able to stop your program while it is
22309 running, you need to use an interrupt-driven serial driver, and arrange
22310 for it to stop when it receives a @code{^C} (@samp{\003}, the control-C
22311 character). That is the character which @value{GDBN} uses to tell the
22312 remote system to stop.
22314 Getting the debugging target to return the proper status to @value{GDBN}
22315 probably requires changes to the standard stub; one quick and dirty way
22316 is to just execute a breakpoint instruction (the ``dirty'' part is that
22317 @value{GDBN} reports a @code{SIGTRAP} instead of a @code{SIGINT}).
22319 Other routines you need to supply are:
22322 @item void exceptionHandler (int @var{exception_number}, void *@var{exception_address})
22323 @findex exceptionHandler
22324 Write this function to install @var{exception_address} in the exception
22325 handling tables. You need to do this because the stub does not have any
22326 way of knowing what the exception handling tables on your target system
22327 are like (for example, the processor's table might be in @sc{rom},
22328 containing entries which point to a table in @sc{ram}).
22329 The @var{exception_number} specifies the exception which should be changed;
22330 its meaning is architecture-dependent (for example, different numbers
22331 might represent divide by zero, misaligned access, etc). When this
22332 exception occurs, control should be transferred directly to
22333 @var{exception_address}, and the processor state (stack, registers,
22334 and so on) should be just as it is when a processor exception occurs. So if
22335 you want to use a jump instruction to reach @var{exception_address}, it
22336 should be a simple jump, not a jump to subroutine.
22338 For the 386, @var{exception_address} should be installed as an interrupt
22339 gate so that interrupts are masked while the handler runs. The gate
22340 should be at privilege level 0 (the most privileged level). The
22341 @sc{sparc} and 68k stubs are able to mask interrupts themselves without
22342 help from @code{exceptionHandler}.
22344 @item void flush_i_cache()
22345 @findex flush_i_cache
22346 On @sc{sparc} and @sc{sparclite} only, write this subroutine to flush the
22347 instruction cache, if any, on your target machine. If there is no
22348 instruction cache, this subroutine may be a no-op.
22350 On target machines that have instruction caches, @value{GDBN} requires this
22351 function to make certain that the state of your program is stable.
22355 You must also make sure this library routine is available:
22358 @item void *memset(void *, int, int)
22360 This is the standard library function @code{memset} that sets an area of
22361 memory to a known value. If you have one of the free versions of
22362 @code{libc.a}, @code{memset} can be found there; otherwise, you must
22363 either obtain it from your hardware manufacturer, or write your own.
22366 If you do not use the GNU C compiler, you may need other standard
22367 library subroutines as well; this varies from one stub to another,
22368 but in general the stubs are likely to use any of the common library
22369 subroutines which @code{@value{NGCC}} generates as inline code.
22372 @node Debug Session
22373 @subsection Putting it All Together
22375 @cindex remote serial debugging summary
22376 In summary, when your program is ready to debug, you must follow these
22381 Make sure you have defined the supporting low-level routines
22382 (@pxref{Bootstrapping,,What You Must Do for the Stub}):
22384 @code{getDebugChar}, @code{putDebugChar},
22385 @code{flush_i_cache}, @code{memset}, @code{exceptionHandler}.
22389 Insert these lines in your program's startup code, before the main
22390 procedure is called:
22397 On some machines, when a breakpoint trap is raised, the hardware
22398 automatically makes the PC point to the instruction after the
22399 breakpoint. If your machine doesn't do that, you may need to adjust
22400 @code{handle_exception} to arrange for it to return to the instruction
22401 after the breakpoint on this first invocation, so that your program
22402 doesn't keep hitting the initial breakpoint instead of making
22406 For the 680x0 stub only, you need to provide a variable called
22407 @code{exceptionHook}. Normally you just use:
22410 void (*exceptionHook)() = 0;
22414 but if before calling @code{set_debug_traps}, you set it to point to a
22415 function in your program, that function is called when
22416 @code{@value{GDBN}} continues after stopping on a trap (for example, bus
22417 error). The function indicated by @code{exceptionHook} is called with
22418 one parameter: an @code{int} which is the exception number.
22421 Compile and link together: your program, the @value{GDBN} debugging stub for
22422 your target architecture, and the supporting subroutines.
22425 Make sure you have a serial connection between your target machine and
22426 the @value{GDBN} host, and identify the serial port on the host.
22429 @c The "remote" target now provides a `load' command, so we should
22430 @c document that. FIXME.
22431 Download your program to your target machine (or get it there by
22432 whatever means the manufacturer provides), and start it.
22435 Start @value{GDBN} on the host, and connect to the target
22436 (@pxref{Connecting,,Connecting to a Remote Target}).
22440 @node Configurations
22441 @chapter Configuration-Specific Information
22443 While nearly all @value{GDBN} commands are available for all native and
22444 cross versions of the debugger, there are some exceptions. This chapter
22445 describes things that are only available in certain configurations.
22447 There are three major categories of configurations: native
22448 configurations, where the host and target are the same, embedded
22449 operating system configurations, which are usually the same for several
22450 different processor architectures, and bare embedded processors, which
22451 are quite different from each other.
22456 * Embedded Processors::
22463 This section describes details specific to particular native
22467 * BSD libkvm Interface:: Debugging BSD kernel memory images
22468 * Process Information:: Process information
22469 * DJGPP Native:: Features specific to the DJGPP port
22470 * Cygwin Native:: Features specific to the Cygwin port
22471 * Hurd Native:: Features specific to @sc{gnu} Hurd
22472 * Darwin:: Features specific to Darwin
22473 * FreeBSD:: Features specific to FreeBSD
22476 @node BSD libkvm Interface
22477 @subsection BSD libkvm Interface
22480 @cindex kernel memory image
22481 @cindex kernel crash dump
22483 BSD-derived systems (FreeBSD/NetBSD/OpenBSD) have a kernel memory
22484 interface that provides a uniform interface for accessing kernel virtual
22485 memory images, including live systems and crash dumps. @value{GDBN}
22486 uses this interface to allow you to debug live kernels and kernel crash
22487 dumps on many native BSD configurations. This is implemented as a
22488 special @code{kvm} debugging target. For debugging a live system, load
22489 the currently running kernel into @value{GDBN} and connect to the
22493 (@value{GDBP}) @b{target kvm}
22496 For debugging crash dumps, provide the file name of the crash dump as an
22500 (@value{GDBP}) @b{target kvm /var/crash/bsd.0}
22503 Once connected to the @code{kvm} target, the following commands are
22509 Set current context from the @dfn{Process Control Block} (PCB) address.
22512 Set current context from proc address. This command isn't available on
22513 modern FreeBSD systems.
22516 @node Process Information
22517 @subsection Process Information
22519 @cindex examine process image
22520 @cindex process info via @file{/proc}
22522 Some operating systems provide interfaces to fetch additional
22523 information about running processes beyond memory and per-thread
22524 register state. If @value{GDBN} is configured for an operating system
22525 with a supported interface, the command @code{info proc} is available
22526 to report information about the process running your program, or about
22527 any process running on your system.
22529 One supported interface is a facility called @samp{/proc} that can be
22530 used to examine the image of a running process using file-system
22531 subroutines. This facility is supported on @sc{gnu}/Linux and Solaris
22534 On FreeBSD systems, system control nodes are used to query process
22537 In addition, some systems may provide additional process information
22538 in core files. Note that a core file may include a subset of the
22539 information available from a live process. Process information is
22540 currently avaiable from cores created on @sc{gnu}/Linux and FreeBSD
22547 @itemx info proc @var{process-id}
22548 Summarize available information about a process. If a
22549 process ID is specified by @var{process-id}, display information about
22550 that process; otherwise display information about the program being
22551 debugged. The summary includes the debugged process ID, the command
22552 line used to invoke it, its current working directory, and its
22553 executable file's absolute file name.
22555 On some systems, @var{process-id} can be of the form
22556 @samp{[@var{pid}]/@var{tid}} which specifies a certain thread ID
22557 within a process. If the optional @var{pid} part is missing, it means
22558 a thread from the process being debugged (the leading @samp{/} still
22559 needs to be present, or else @value{GDBN} will interpret the number as
22560 a process ID rather than a thread ID).
22562 @item info proc cmdline
22563 @cindex info proc cmdline
22564 Show the original command line of the process. This command is
22565 supported on @sc{gnu}/Linux and FreeBSD.
22567 @item info proc cwd
22568 @cindex info proc cwd
22569 Show the current working directory of the process. This command is
22570 supported on @sc{gnu}/Linux and FreeBSD.
22572 @item info proc exe
22573 @cindex info proc exe
22574 Show the name of executable of the process. This command is supported
22575 on @sc{gnu}/Linux and FreeBSD.
22577 @item info proc files
22578 @cindex info proc files
22579 Show the file descriptors open by the process. For each open file
22580 descriptor, @value{GDBN} shows its number, type (file, directory,
22581 character device, socket), file pointer offset, and the name of the
22582 resource open on the descriptor. The resource name can be a file name
22583 (for files, directories, and devices) or a protocol followed by socket
22584 address (for network connections). This command is supported on
22587 This example shows the open file descriptors for a process using a
22588 tty for standard input and output as well as two network sockets:
22591 (gdb) info proc files 22136
22595 FD Type Offset Flags Name
22596 text file - r-------- /usr/bin/ssh
22597 ctty chr - rw------- /dev/pts/20
22598 cwd dir - r-------- /usr/home/john
22599 root dir - r-------- /
22600 0 chr 0x32933a4 rw------- /dev/pts/20
22601 1 chr 0x32933a4 rw------- /dev/pts/20
22602 2 chr 0x32933a4 rw------- /dev/pts/20
22603 3 socket 0x0 rw----n-- tcp4 10.0.1.2:53014 -> 10.0.1.10:22
22604 4 socket 0x0 rw------- unix stream:/tmp/ssh-FIt89oAzOn5f/agent.2456
22607 @item info proc mappings
22608 @cindex memory address space mappings
22609 Report the memory address space ranges accessible in a process. On
22610 Solaris and FreeBSD systems, each memory range includes information on
22611 whether the process has read, write, or execute access rights to each
22612 range. On @sc{gnu}/Linux and FreeBSD systems, each memory range
22613 includes the object file which is mapped to that range.
22615 @item info proc stat
22616 @itemx info proc status
22617 @cindex process detailed status information
22618 Show additional process-related information, including the user ID and
22619 group ID; virtual memory usage; the signals that are pending, blocked,
22620 and ignored; its TTY; its consumption of system and user time; its
22621 stack size; its @samp{nice} value; etc. These commands are supported
22622 on @sc{gnu}/Linux and FreeBSD.
22624 For @sc{gnu}/Linux systems, see the @samp{proc} man page for more
22625 information (type @kbd{man 5 proc} from your shell prompt).
22627 For FreeBSD systems, @code{info proc stat} is an alias for @code{info
22630 @item info proc all
22631 Show all the information about the process described under all of the
22632 above @code{info proc} subcommands.
22635 @comment These sub-options of 'info proc' were not included when
22636 @comment procfs.c was re-written. Keep their descriptions around
22637 @comment against the day when someone finds the time to put them back in.
22638 @kindex info proc times
22639 @item info proc times
22640 Starting time, user CPU time, and system CPU time for your program and
22643 @kindex info proc id
22645 Report on the process IDs related to your program: its own process ID,
22646 the ID of its parent, the process group ID, and the session ID.
22649 @item set procfs-trace
22650 @kindex set procfs-trace
22651 @cindex @code{procfs} API calls
22652 This command enables and disables tracing of @code{procfs} API calls.
22654 @item show procfs-trace
22655 @kindex show procfs-trace
22656 Show the current state of @code{procfs} API call tracing.
22658 @item set procfs-file @var{file}
22659 @kindex set procfs-file
22660 Tell @value{GDBN} to write @code{procfs} API trace to the named
22661 @var{file}. @value{GDBN} appends the trace info to the previous
22662 contents of the file. The default is to display the trace on the
22665 @item show procfs-file
22666 @kindex show procfs-file
22667 Show the file to which @code{procfs} API trace is written.
22669 @item proc-trace-entry
22670 @itemx proc-trace-exit
22671 @itemx proc-untrace-entry
22672 @itemx proc-untrace-exit
22673 @kindex proc-trace-entry
22674 @kindex proc-trace-exit
22675 @kindex proc-untrace-entry
22676 @kindex proc-untrace-exit
22677 These commands enable and disable tracing of entries into and exits
22678 from the @code{syscall} interface.
22681 @kindex info pidlist
22682 @cindex process list, QNX Neutrino
22683 For QNX Neutrino only, this command displays the list of all the
22684 processes and all the threads within each process.
22687 @kindex info meminfo
22688 @cindex mapinfo list, QNX Neutrino
22689 For QNX Neutrino only, this command displays the list of all mapinfos.
22693 @subsection Features for Debugging @sc{djgpp} Programs
22694 @cindex @sc{djgpp} debugging
22695 @cindex native @sc{djgpp} debugging
22696 @cindex MS-DOS-specific commands
22699 @sc{djgpp} is a port of the @sc{gnu} development tools to MS-DOS and
22700 MS-Windows. @sc{djgpp} programs are 32-bit protected-mode programs
22701 that use the @dfn{DPMI} (DOS Protected-Mode Interface) API to run on
22702 top of real-mode DOS systems and their emulations.
22704 @value{GDBN} supports native debugging of @sc{djgpp} programs, and
22705 defines a few commands specific to the @sc{djgpp} port. This
22706 subsection describes those commands.
22711 This is a prefix of @sc{djgpp}-specific commands which print
22712 information about the target system and important OS structures.
22715 @cindex MS-DOS system info
22716 @cindex free memory information (MS-DOS)
22717 @item info dos sysinfo
22718 This command displays assorted information about the underlying
22719 platform: the CPU type and features, the OS version and flavor, the
22720 DPMI version, and the available conventional and DPMI memory.
22725 @cindex segment descriptor tables
22726 @cindex descriptor tables display
22728 @itemx info dos ldt
22729 @itemx info dos idt
22730 These 3 commands display entries from, respectively, Global, Local,
22731 and Interrupt Descriptor Tables (GDT, LDT, and IDT). The descriptor
22732 tables are data structures which store a descriptor for each segment
22733 that is currently in use. The segment's selector is an index into a
22734 descriptor table; the table entry for that index holds the
22735 descriptor's base address and limit, and its attributes and access
22738 A typical @sc{djgpp} program uses 3 segments: a code segment, a data
22739 segment (used for both data and the stack), and a DOS segment (which
22740 allows access to DOS/BIOS data structures and absolute addresses in
22741 conventional memory). However, the DPMI host will usually define
22742 additional segments in order to support the DPMI environment.
22744 @cindex garbled pointers
22745 These commands allow to display entries from the descriptor tables.
22746 Without an argument, all entries from the specified table are
22747 displayed. An argument, which should be an integer expression, means
22748 display a single entry whose index is given by the argument. For
22749 example, here's a convenient way to display information about the
22750 debugged program's data segment:
22753 @exdent @code{(@value{GDBP}) info dos ldt $ds}
22754 @exdent @code{0x13f: base=0x11970000 limit=0x0009ffff 32-Bit Data (Read/Write, Exp-up)}
22758 This comes in handy when you want to see whether a pointer is outside
22759 the data segment's limit (i.e.@: @dfn{garbled}).
22761 @cindex page tables display (MS-DOS)
22763 @itemx info dos pte
22764 These two commands display entries from, respectively, the Page
22765 Directory and the Page Tables. Page Directories and Page Tables are
22766 data structures which control how virtual memory addresses are mapped
22767 into physical addresses. A Page Table includes an entry for every
22768 page of memory that is mapped into the program's address space; there
22769 may be several Page Tables, each one holding up to 4096 entries. A
22770 Page Directory has up to 4096 entries, one each for every Page Table
22771 that is currently in use.
22773 Without an argument, @kbd{info dos pde} displays the entire Page
22774 Directory, and @kbd{info dos pte} displays all the entries in all of
22775 the Page Tables. An argument, an integer expression, given to the
22776 @kbd{info dos pde} command means display only that entry from the Page
22777 Directory table. An argument given to the @kbd{info dos pte} command
22778 means display entries from a single Page Table, the one pointed to by
22779 the specified entry in the Page Directory.
22781 @cindex direct memory access (DMA) on MS-DOS
22782 These commands are useful when your program uses @dfn{DMA} (Direct
22783 Memory Access), which needs physical addresses to program the DMA
22786 These commands are supported only with some DPMI servers.
22788 @cindex physical address from linear address
22789 @item info dos address-pte @var{addr}
22790 This command displays the Page Table entry for a specified linear
22791 address. The argument @var{addr} is a linear address which should
22792 already have the appropriate segment's base address added to it,
22793 because this command accepts addresses which may belong to @emph{any}
22794 segment. For example, here's how to display the Page Table entry for
22795 the page where a variable @code{i} is stored:
22798 @exdent @code{(@value{GDBP}) info dos address-pte __djgpp_base_address + (char *)&i}
22799 @exdent @code{Page Table entry for address 0x11a00d30:}
22800 @exdent @code{Base=0x02698000 Dirty Acc. Not-Cached Write-Back Usr Read-Write +0xd30}
22804 This says that @code{i} is stored at offset @code{0xd30} from the page
22805 whose physical base address is @code{0x02698000}, and shows all the
22806 attributes of that page.
22808 Note that you must cast the addresses of variables to a @code{char *},
22809 since otherwise the value of @code{__djgpp_base_address}, the base
22810 address of all variables and functions in a @sc{djgpp} program, will
22811 be added using the rules of C pointer arithmetics: if @code{i} is
22812 declared an @code{int}, @value{GDBN} will add 4 times the value of
22813 @code{__djgpp_base_address} to the address of @code{i}.
22815 Here's another example, it displays the Page Table entry for the
22819 @exdent @code{(@value{GDBP}) info dos address-pte *((unsigned *)&_go32_info_block + 3)}
22820 @exdent @code{Page Table entry for address 0x29110:}
22821 @exdent @code{Base=0x00029000 Dirty Acc. Not-Cached Write-Back Usr Read-Write +0x110}
22825 (The @code{+ 3} offset is because the transfer buffer's address is the
22826 3rd member of the @code{_go32_info_block} structure.) The output
22827 clearly shows that this DPMI server maps the addresses in conventional
22828 memory 1:1, i.e.@: the physical (@code{0x00029000} + @code{0x110}) and
22829 linear (@code{0x29110}) addresses are identical.
22831 This command is supported only with some DPMI servers.
22834 @cindex DOS serial data link, remote debugging
22835 In addition to native debugging, the DJGPP port supports remote
22836 debugging via a serial data link. The following commands are specific
22837 to remote serial debugging in the DJGPP port of @value{GDBN}.
22840 @kindex set com1base
22841 @kindex set com1irq
22842 @kindex set com2base
22843 @kindex set com2irq
22844 @kindex set com3base
22845 @kindex set com3irq
22846 @kindex set com4base
22847 @kindex set com4irq
22848 @item set com1base @var{addr}
22849 This command sets the base I/O port address of the @file{COM1} serial
22852 @item set com1irq @var{irq}
22853 This command sets the @dfn{Interrupt Request} (@code{IRQ}) line to use
22854 for the @file{COM1} serial port.
22856 There are similar commands @samp{set com2base}, @samp{set com3irq},
22857 etc.@: for setting the port address and the @code{IRQ} lines for the
22860 @kindex show com1base
22861 @kindex show com1irq
22862 @kindex show com2base
22863 @kindex show com2irq
22864 @kindex show com3base
22865 @kindex show com3irq
22866 @kindex show com4base
22867 @kindex show com4irq
22868 The related commands @samp{show com1base}, @samp{show com1irq} etc.@:
22869 display the current settings of the base address and the @code{IRQ}
22870 lines used by the COM ports.
22873 @kindex info serial
22874 @cindex DOS serial port status
22875 This command prints the status of the 4 DOS serial ports. For each
22876 port, it prints whether it's active or not, its I/O base address and
22877 IRQ number, whether it uses a 16550-style FIFO, its baudrate, and the
22878 counts of various errors encountered so far.
22882 @node Cygwin Native
22883 @subsection Features for Debugging MS Windows PE Executables
22884 @cindex MS Windows debugging
22885 @cindex native Cygwin debugging
22886 @cindex Cygwin-specific commands
22888 @value{GDBN} supports native debugging of MS Windows programs, including
22889 DLLs with and without symbolic debugging information.
22891 @cindex Ctrl-BREAK, MS-Windows
22892 @cindex interrupt debuggee on MS-Windows
22893 MS-Windows programs that call @code{SetConsoleMode} to switch off the
22894 special meaning of the @samp{Ctrl-C} keystroke cannot be interrupted
22895 by typing @kbd{C-c}. For this reason, @value{GDBN} on MS-Windows
22896 supports @kbd{C-@key{BREAK}} as an alternative interrupt key
22897 sequence, which can be used to interrupt the debuggee even if it
22900 There are various additional Cygwin-specific commands, described in
22901 this section. Working with DLLs that have no debugging symbols is
22902 described in @ref{Non-debug DLL Symbols}.
22907 This is a prefix of MS Windows-specific commands which print
22908 information about the target system and important OS structures.
22910 @item info w32 selector
22911 This command displays information returned by
22912 the Win32 API @code{GetThreadSelectorEntry} function.
22913 It takes an optional argument that is evaluated to
22914 a long value to give the information about this given selector.
22915 Without argument, this command displays information
22916 about the six segment registers.
22918 @item info w32 thread-information-block
22919 This command displays thread specific information stored in the
22920 Thread Information Block (readable on the X86 CPU family using @code{$fs}
22921 selector for 32-bit programs and @code{$gs} for 64-bit programs).
22923 @kindex signal-event
22924 @item signal-event @var{id}
22925 This command signals an event with user-provided @var{id}. Used to resume
22926 crashing process when attached to it using MS-Windows JIT debugging (AeDebug).
22928 To use it, create or edit the following keys in
22929 @code{HKLM\SOFTWARE\Microsoft\Windows NT\CurrentVersion\AeDebug} and/or
22930 @code{HKLM\SOFTWARE\Wow6432Node\Microsoft\Windows NT\CurrentVersion\AeDebug}
22931 (for x86_64 versions):
22935 @code{Debugger} (REG_SZ) --- a command to launch the debugger.
22936 Suggested command is: @code{@var{fully-qualified-path-to-gdb.exe} -ex
22937 "attach %ld" -ex "signal-event %ld" -ex "continue"}.
22939 The first @code{%ld} will be replaced by the process ID of the
22940 crashing process, the second @code{%ld} will be replaced by the ID of
22941 the event that blocks the crashing process, waiting for @value{GDBN}
22945 @code{Auto} (REG_SZ) --- either @code{1} or @code{0}. @code{1} will
22946 make the system run debugger specified by the Debugger key
22947 automatically, @code{0} will cause a dialog box with ``OK'' and
22948 ``Cancel'' buttons to appear, which allows the user to either
22949 terminate the crashing process (OK) or debug it (Cancel).
22952 @kindex set cygwin-exceptions
22953 @cindex debugging the Cygwin DLL
22954 @cindex Cygwin DLL, debugging
22955 @item set cygwin-exceptions @var{mode}
22956 If @var{mode} is @code{on}, @value{GDBN} will break on exceptions that
22957 happen inside the Cygwin DLL. If @var{mode} is @code{off},
22958 @value{GDBN} will delay recognition of exceptions, and may ignore some
22959 exceptions which seem to be caused by internal Cygwin DLL
22960 ``bookkeeping''. This option is meant primarily for debugging the
22961 Cygwin DLL itself; the default value is @code{off} to avoid annoying
22962 @value{GDBN} users with false @code{SIGSEGV} signals.
22964 @kindex show cygwin-exceptions
22965 @item show cygwin-exceptions
22966 Displays whether @value{GDBN} will break on exceptions that happen
22967 inside the Cygwin DLL itself.
22969 @kindex set new-console
22970 @item set new-console @var{mode}
22971 If @var{mode} is @code{on} the debuggee will
22972 be started in a new console on next start.
22973 If @var{mode} is @code{off}, the debuggee will
22974 be started in the same console as the debugger.
22976 @kindex show new-console
22977 @item show new-console
22978 Displays whether a new console is used
22979 when the debuggee is started.
22981 @kindex set new-group
22982 @item set new-group @var{mode}
22983 This boolean value controls whether the debuggee should
22984 start a new group or stay in the same group as the debugger.
22985 This affects the way the Windows OS handles
22988 @kindex show new-group
22989 @item show new-group
22990 Displays current value of new-group boolean.
22992 @kindex set debugevents
22993 @item set debugevents
22994 This boolean value adds debug output concerning kernel events related
22995 to the debuggee seen by the debugger. This includes events that
22996 signal thread and process creation and exit, DLL loading and
22997 unloading, console interrupts, and debugging messages produced by the
22998 Windows @code{OutputDebugString} API call.
23000 @kindex set debugexec
23001 @item set debugexec
23002 This boolean value adds debug output concerning execute events
23003 (such as resume thread) seen by the debugger.
23005 @kindex set debugexceptions
23006 @item set debugexceptions
23007 This boolean value adds debug output concerning exceptions in the
23008 debuggee seen by the debugger.
23010 @kindex set debugmemory
23011 @item set debugmemory
23012 This boolean value adds debug output concerning debuggee memory reads
23013 and writes by the debugger.
23017 This boolean values specifies whether the debuggee is called
23018 via a shell or directly (default value is on).
23022 Displays if the debuggee will be started with a shell.
23027 * Non-debug DLL Symbols:: Support for DLLs without debugging symbols
23030 @node Non-debug DLL Symbols
23031 @subsubsection Support for DLLs without Debugging Symbols
23032 @cindex DLLs with no debugging symbols
23033 @cindex Minimal symbols and DLLs
23035 Very often on windows, some of the DLLs that your program relies on do
23036 not include symbolic debugging information (for example,
23037 @file{kernel32.dll}). When @value{GDBN} doesn't recognize any debugging
23038 symbols in a DLL, it relies on the minimal amount of symbolic
23039 information contained in the DLL's export table. This section
23040 describes working with such symbols, known internally to @value{GDBN} as
23041 ``minimal symbols''.
23043 Note that before the debugged program has started execution, no DLLs
23044 will have been loaded. The easiest way around this problem is simply to
23045 start the program --- either by setting a breakpoint or letting the
23046 program run once to completion.
23048 @subsubsection DLL Name Prefixes
23050 In keeping with the naming conventions used by the Microsoft debugging
23051 tools, DLL export symbols are made available with a prefix based on the
23052 DLL name, for instance @code{KERNEL32!CreateFileA}. The plain name is
23053 also entered into the symbol table, so @code{CreateFileA} is often
23054 sufficient. In some cases there will be name clashes within a program
23055 (particularly if the executable itself includes full debugging symbols)
23056 necessitating the use of the fully qualified name when referring to the
23057 contents of the DLL. Use single-quotes around the name to avoid the
23058 exclamation mark (``!'') being interpreted as a language operator.
23060 Note that the internal name of the DLL may be all upper-case, even
23061 though the file name of the DLL is lower-case, or vice-versa. Since
23062 symbols within @value{GDBN} are @emph{case-sensitive} this may cause
23063 some confusion. If in doubt, try the @code{info functions} and
23064 @code{info variables} commands or even @code{maint print msymbols}
23065 (@pxref{Symbols}). Here's an example:
23068 (@value{GDBP}) info function CreateFileA
23069 All functions matching regular expression "CreateFileA":
23071 Non-debugging symbols:
23072 0x77e885f4 CreateFileA
23073 0x77e885f4 KERNEL32!CreateFileA
23077 (@value{GDBP}) info function !
23078 All functions matching regular expression "!":
23080 Non-debugging symbols:
23081 0x6100114c cygwin1!__assert
23082 0x61004034 cygwin1!_dll_crt0@@0
23083 0x61004240 cygwin1!dll_crt0(per_process *)
23087 @subsubsection Working with Minimal Symbols
23089 Symbols extracted from a DLL's export table do not contain very much
23090 type information. All that @value{GDBN} can do is guess whether a symbol
23091 refers to a function or variable depending on the linker section that
23092 contains the symbol. Also note that the actual contents of the memory
23093 contained in a DLL are not available unless the program is running. This
23094 means that you cannot examine the contents of a variable or disassemble
23095 a function within a DLL without a running program.
23097 Variables are generally treated as pointers and dereferenced
23098 automatically. For this reason, it is often necessary to prefix a
23099 variable name with the address-of operator (``&'') and provide explicit
23100 type information in the command. Here's an example of the type of
23104 (@value{GDBP}) print 'cygwin1!__argv'
23105 'cygwin1!__argv' has unknown type; cast it to its declared type
23109 (@value{GDBP}) x 'cygwin1!__argv'
23110 'cygwin1!__argv' has unknown type; cast it to its declared type
23113 And two possible solutions:
23116 (@value{GDBP}) print ((char **)'cygwin1!__argv')[0]
23117 $2 = 0x22fd98 "/cygdrive/c/mydirectory/myprogram"
23121 (@value{GDBP}) x/2x &'cygwin1!__argv'
23122 0x610c0aa8 <cygwin1!__argv>: 0x10021608 0x00000000
23123 (@value{GDBP}) x/x 0x10021608
23124 0x10021608: 0x0022fd98
23125 (@value{GDBP}) x/s 0x0022fd98
23126 0x22fd98: "/cygdrive/c/mydirectory/myprogram"
23129 Setting a break point within a DLL is possible even before the program
23130 starts execution. However, under these circumstances, @value{GDBN} can't
23131 examine the initial instructions of the function in order to skip the
23132 function's frame set-up code. You can work around this by using ``*&''
23133 to set the breakpoint at a raw memory address:
23136 (@value{GDBP}) break *&'python22!PyOS_Readline'
23137 Breakpoint 1 at 0x1e04eff0
23140 The author of these extensions is not entirely convinced that setting a
23141 break point within a shared DLL like @file{kernel32.dll} is completely
23145 @subsection Commands Specific to @sc{gnu} Hurd Systems
23146 @cindex @sc{gnu} Hurd debugging
23148 This subsection describes @value{GDBN} commands specific to the
23149 @sc{gnu} Hurd native debugging.
23154 @kindex set signals@r{, Hurd command}
23155 @kindex set sigs@r{, Hurd command}
23156 This command toggles the state of inferior signal interception by
23157 @value{GDBN}. Mach exceptions, such as breakpoint traps, are not
23158 affected by this command. @code{sigs} is a shorthand alias for
23163 @kindex show signals@r{, Hurd command}
23164 @kindex show sigs@r{, Hurd command}
23165 Show the current state of intercepting inferior's signals.
23167 @item set signal-thread
23168 @itemx set sigthread
23169 @kindex set signal-thread
23170 @kindex set sigthread
23171 This command tells @value{GDBN} which thread is the @code{libc} signal
23172 thread. That thread is run when a signal is delivered to a running
23173 process. @code{set sigthread} is the shorthand alias of @code{set
23176 @item show signal-thread
23177 @itemx show sigthread
23178 @kindex show signal-thread
23179 @kindex show sigthread
23180 These two commands show which thread will run when the inferior is
23181 delivered a signal.
23184 @kindex set stopped@r{, Hurd command}
23185 This commands tells @value{GDBN} that the inferior process is stopped,
23186 as with the @code{SIGSTOP} signal. The stopped process can be
23187 continued by delivering a signal to it.
23190 @kindex show stopped@r{, Hurd command}
23191 This command shows whether @value{GDBN} thinks the debuggee is
23194 @item set exceptions
23195 @kindex set exceptions@r{, Hurd command}
23196 Use this command to turn off trapping of exceptions in the inferior.
23197 When exception trapping is off, neither breakpoints nor
23198 single-stepping will work. To restore the default, set exception
23201 @item show exceptions
23202 @kindex show exceptions@r{, Hurd command}
23203 Show the current state of trapping exceptions in the inferior.
23205 @item set task pause
23206 @kindex set task@r{, Hurd commands}
23207 @cindex task attributes (@sc{gnu} Hurd)
23208 @cindex pause current task (@sc{gnu} Hurd)
23209 This command toggles task suspension when @value{GDBN} has control.
23210 Setting it to on takes effect immediately, and the task is suspended
23211 whenever @value{GDBN} gets control. Setting it to off will take
23212 effect the next time the inferior is continued. If this option is set
23213 to off, you can use @code{set thread default pause on} or @code{set
23214 thread pause on} (see below) to pause individual threads.
23216 @item show task pause
23217 @kindex show task@r{, Hurd commands}
23218 Show the current state of task suspension.
23220 @item set task detach-suspend-count
23221 @cindex task suspend count
23222 @cindex detach from task, @sc{gnu} Hurd
23223 This command sets the suspend count the task will be left with when
23224 @value{GDBN} detaches from it.
23226 @item show task detach-suspend-count
23227 Show the suspend count the task will be left with when detaching.
23229 @item set task exception-port
23230 @itemx set task excp
23231 @cindex task exception port, @sc{gnu} Hurd
23232 This command sets the task exception port to which @value{GDBN} will
23233 forward exceptions. The argument should be the value of the @dfn{send
23234 rights} of the task. @code{set task excp} is a shorthand alias.
23236 @item set noninvasive
23237 @cindex noninvasive task options
23238 This command switches @value{GDBN} to a mode that is the least
23239 invasive as far as interfering with the inferior is concerned. This
23240 is the same as using @code{set task pause}, @code{set exceptions}, and
23241 @code{set signals} to values opposite to the defaults.
23243 @item info send-rights
23244 @itemx info receive-rights
23245 @itemx info port-rights
23246 @itemx info port-sets
23247 @itemx info dead-names
23250 @cindex send rights, @sc{gnu} Hurd
23251 @cindex receive rights, @sc{gnu} Hurd
23252 @cindex port rights, @sc{gnu} Hurd
23253 @cindex port sets, @sc{gnu} Hurd
23254 @cindex dead names, @sc{gnu} Hurd
23255 These commands display information about, respectively, send rights,
23256 receive rights, port rights, port sets, and dead names of a task.
23257 There are also shorthand aliases: @code{info ports} for @code{info
23258 port-rights} and @code{info psets} for @code{info port-sets}.
23260 @item set thread pause
23261 @kindex set thread@r{, Hurd command}
23262 @cindex thread properties, @sc{gnu} Hurd
23263 @cindex pause current thread (@sc{gnu} Hurd)
23264 This command toggles current thread suspension when @value{GDBN} has
23265 control. Setting it to on takes effect immediately, and the current
23266 thread is suspended whenever @value{GDBN} gets control. Setting it to
23267 off will take effect the next time the inferior is continued.
23268 Normally, this command has no effect, since when @value{GDBN} has
23269 control, the whole task is suspended. However, if you used @code{set
23270 task pause off} (see above), this command comes in handy to suspend
23271 only the current thread.
23273 @item show thread pause
23274 @kindex show thread@r{, Hurd command}
23275 This command shows the state of current thread suspension.
23277 @item set thread run
23278 This command sets whether the current thread is allowed to run.
23280 @item show thread run
23281 Show whether the current thread is allowed to run.
23283 @item set thread detach-suspend-count
23284 @cindex thread suspend count, @sc{gnu} Hurd
23285 @cindex detach from thread, @sc{gnu} Hurd
23286 This command sets the suspend count @value{GDBN} will leave on a
23287 thread when detaching. This number is relative to the suspend count
23288 found by @value{GDBN} when it notices the thread; use @code{set thread
23289 takeover-suspend-count} to force it to an absolute value.
23291 @item show thread detach-suspend-count
23292 Show the suspend count @value{GDBN} will leave on the thread when
23295 @item set thread exception-port
23296 @itemx set thread excp
23297 Set the thread exception port to which to forward exceptions. This
23298 overrides the port set by @code{set task exception-port} (see above).
23299 @code{set thread excp} is the shorthand alias.
23301 @item set thread takeover-suspend-count
23302 Normally, @value{GDBN}'s thread suspend counts are relative to the
23303 value @value{GDBN} finds when it notices each thread. This command
23304 changes the suspend counts to be absolute instead.
23306 @item set thread default
23307 @itemx show thread default
23308 @cindex thread default settings, @sc{gnu} Hurd
23309 Each of the above @code{set thread} commands has a @code{set thread
23310 default} counterpart (e.g., @code{set thread default pause}, @code{set
23311 thread default exception-port}, etc.). The @code{thread default}
23312 variety of commands sets the default thread properties for all
23313 threads; you can then change the properties of individual threads with
23314 the non-default commands.
23321 @value{GDBN} provides the following commands specific to the Darwin target:
23324 @item set debug darwin @var{num}
23325 @kindex set debug darwin
23326 When set to a non zero value, enables debugging messages specific to
23327 the Darwin support. Higher values produce more verbose output.
23329 @item show debug darwin
23330 @kindex show debug darwin
23331 Show the current state of Darwin messages.
23333 @item set debug mach-o @var{num}
23334 @kindex set debug mach-o
23335 When set to a non zero value, enables debugging messages while
23336 @value{GDBN} is reading Darwin object files. (@dfn{Mach-O} is the
23337 file format used on Darwin for object and executable files.) Higher
23338 values produce more verbose output. This is a command to diagnose
23339 problems internal to @value{GDBN} and should not be needed in normal
23342 @item show debug mach-o
23343 @kindex show debug mach-o
23344 Show the current state of Mach-O file messages.
23346 @item set mach-exceptions on
23347 @itemx set mach-exceptions off
23348 @kindex set mach-exceptions
23349 On Darwin, faults are first reported as a Mach exception and are then
23350 mapped to a Posix signal. Use this command to turn on trapping of
23351 Mach exceptions in the inferior. This might be sometimes useful to
23352 better understand the cause of a fault. The default is off.
23354 @item show mach-exceptions
23355 @kindex show mach-exceptions
23356 Show the current state of exceptions trapping.
23360 @subsection FreeBSD
23363 When the ABI of a system call is changed in the FreeBSD kernel, this
23364 is implemented by leaving a compatibility system call using the old
23365 ABI at the existing number and allocating a new system call number for
23366 the version using the new ABI. As a convenience, when a system call
23367 is caught by name (@pxref{catch syscall}), compatibility system calls
23370 For example, FreeBSD 12 introduced a new variant of the @code{kevent}
23371 system call and catching the @code{kevent} system call by name catches
23375 (@value{GDBP}) catch syscall kevent
23376 Catchpoint 1 (syscalls 'freebsd11_kevent' [363] 'kevent' [560])
23382 @section Embedded Operating Systems
23384 This section describes configurations involving the debugging of
23385 embedded operating systems that are available for several different
23388 @value{GDBN} includes the ability to debug programs running on
23389 various real-time operating systems.
23391 @node Embedded Processors
23392 @section Embedded Processors
23394 This section goes into details specific to particular embedded
23397 @cindex send command to simulator
23398 Whenever a specific embedded processor has a simulator, @value{GDBN}
23399 allows to send an arbitrary command to the simulator.
23402 @item sim @var{command}
23403 @kindex sim@r{, a command}
23404 Send an arbitrary @var{command} string to the simulator. Consult the
23405 documentation for the specific simulator in use for information about
23406 acceptable commands.
23411 * ARC:: Synopsys ARC
23413 * M68K:: Motorola M68K
23414 * MicroBlaze:: Xilinx MicroBlaze
23415 * MIPS Embedded:: MIPS Embedded
23416 * OpenRISC 1000:: OpenRISC 1000 (or1k)
23417 * PowerPC Embedded:: PowerPC Embedded
23420 * Super-H:: Renesas Super-H
23424 @subsection Synopsys ARC
23425 @cindex Synopsys ARC
23426 @cindex ARC specific commands
23432 @value{GDBN} provides the following ARC-specific commands:
23435 @item set debug arc
23436 @kindex set debug arc
23437 Control the level of ARC specific debug messages. Use 0 for no messages (the
23438 default), 1 for debug messages, and 2 for even more debug messages.
23440 @item show debug arc
23441 @kindex show debug arc
23442 Show the level of ARC specific debugging in operation.
23444 @item maint print arc arc-instruction @var{address}
23445 @kindex maint print arc arc-instruction
23446 Print internal disassembler information about instruction at a given address.
23453 @value{GDBN} provides the following ARM-specific commands:
23456 @item set arm disassembler
23458 This commands selects from a list of disassembly styles. The
23459 @code{"std"} style is the standard style.
23461 @item show arm disassembler
23463 Show the current disassembly style.
23465 @item set arm apcs32
23466 @cindex ARM 32-bit mode
23467 This command toggles ARM operation mode between 32-bit and 26-bit.
23469 @item show arm apcs32
23470 Display the current usage of the ARM 32-bit mode.
23472 @item set arm fpu @var{fputype}
23473 This command sets the ARM floating-point unit (FPU) type. The
23474 argument @var{fputype} can be one of these:
23478 Determine the FPU type by querying the OS ABI.
23480 Software FPU, with mixed-endian doubles on little-endian ARM
23483 GCC-compiled FPA co-processor.
23485 Software FPU with pure-endian doubles.
23491 Show the current type of the FPU.
23494 This command forces @value{GDBN} to use the specified ABI.
23497 Show the currently used ABI.
23499 @item set arm fallback-mode (arm|thumb|auto)
23500 @value{GDBN} uses the symbol table, when available, to determine
23501 whether instructions are ARM or Thumb. This command controls
23502 @value{GDBN}'s default behavior when the symbol table is not
23503 available. The default is @samp{auto}, which causes @value{GDBN} to
23504 use the current execution mode (from the @code{T} bit in the @code{CPSR}
23507 @item show arm fallback-mode
23508 Show the current fallback instruction mode.
23510 @item set arm force-mode (arm|thumb|auto)
23511 This command overrides use of the symbol table to determine whether
23512 instructions are ARM or Thumb. The default is @samp{auto}, which
23513 causes @value{GDBN} to use the symbol table and then the setting
23514 of @samp{set arm fallback-mode}.
23516 @item show arm force-mode
23517 Show the current forced instruction mode.
23519 @item set debug arm
23520 Toggle whether to display ARM-specific debugging messages from the ARM
23521 target support subsystem.
23523 @item show debug arm
23524 Show whether ARM-specific debugging messages are enabled.
23528 @item target sim @r{[}@var{simargs}@r{]} @dots{}
23529 The @value{GDBN} ARM simulator accepts the following optional arguments.
23532 @item --swi-support=@var{type}
23533 Tell the simulator which SWI interfaces to support. The argument
23534 @var{type} may be a comma separated list of the following values.
23535 The default value is @code{all}.
23550 The Motorola m68k configuration includes ColdFire support.
23553 @subsection MicroBlaze
23554 @cindex Xilinx MicroBlaze
23555 @cindex XMD, Xilinx Microprocessor Debugger
23557 The MicroBlaze is a soft-core processor supported on various Xilinx
23558 FPGAs, such as Spartan or Virtex series. Boards with these processors
23559 usually have JTAG ports which connect to a host system running the Xilinx
23560 Embedded Development Kit (EDK) or Software Development Kit (SDK).
23561 This host system is used to download the configuration bitstream to
23562 the target FPGA. The Xilinx Microprocessor Debugger (XMD) program
23563 communicates with the target board using the JTAG interface and
23564 presents a @code{gdbserver} interface to the board. By default
23565 @code{xmd} uses port @code{1234}. (While it is possible to change
23566 this default port, it requires the use of undocumented @code{xmd}
23567 commands. Contact Xilinx support if you need to do this.)
23569 Use these GDB commands to connect to the MicroBlaze target processor.
23572 @item target remote :1234
23573 Use this command to connect to the target if you are running @value{GDBN}
23574 on the same system as @code{xmd}.
23576 @item target remote @var{xmd-host}:1234
23577 Use this command to connect to the target if it is connected to @code{xmd}
23578 running on a different system named @var{xmd-host}.
23581 Use this command to download a program to the MicroBlaze target.
23583 @item set debug microblaze @var{n}
23584 Enable MicroBlaze-specific debugging messages if non-zero.
23586 @item show debug microblaze @var{n}
23587 Show MicroBlaze-specific debugging level.
23590 @node MIPS Embedded
23591 @subsection @acronym{MIPS} Embedded
23594 @value{GDBN} supports these special commands for @acronym{MIPS} targets:
23597 @item set mipsfpu double
23598 @itemx set mipsfpu single
23599 @itemx set mipsfpu none
23600 @itemx set mipsfpu auto
23601 @itemx show mipsfpu
23602 @kindex set mipsfpu
23603 @kindex show mipsfpu
23604 @cindex @acronym{MIPS} remote floating point
23605 @cindex floating point, @acronym{MIPS} remote
23606 If your target board does not support the @acronym{MIPS} floating point
23607 coprocessor, you should use the command @samp{set mipsfpu none} (if you
23608 need this, you may wish to put the command in your @value{GDBN} init
23609 file). This tells @value{GDBN} how to find the return value of
23610 functions which return floating point values. It also allows
23611 @value{GDBN} to avoid saving the floating point registers when calling
23612 functions on the board. If you are using a floating point coprocessor
23613 with only single precision floating point support, as on the @sc{r4650}
23614 processor, use the command @samp{set mipsfpu single}. The default
23615 double precision floating point coprocessor may be selected using
23616 @samp{set mipsfpu double}.
23618 In previous versions the only choices were double precision or no
23619 floating point, so @samp{set mipsfpu on} will select double precision
23620 and @samp{set mipsfpu off} will select no floating point.
23622 As usual, you can inquire about the @code{mipsfpu} variable with
23623 @samp{show mipsfpu}.
23626 @node OpenRISC 1000
23627 @subsection OpenRISC 1000
23628 @cindex OpenRISC 1000
23631 The OpenRISC 1000 provides a free RISC instruction set architecture. It is
23632 mainly provided as a soft-core which can run on Xilinx, Altera and other
23635 @value{GDBN} for OpenRISC supports the below commands when connecting to
23643 Runs the builtin CPU simulator which can run very basic
23644 programs but does not support most hardware functions like MMU.
23645 For more complex use cases the user is advised to run an external
23646 target, and connect using @samp{target remote}.
23648 Example: @code{target sim}
23650 @item set debug or1k
23651 Toggle whether to display OpenRISC-specific debugging messages from the
23652 OpenRISC target support subsystem.
23654 @item show debug or1k
23655 Show whether OpenRISC-specific debugging messages are enabled.
23658 @node PowerPC Embedded
23659 @subsection PowerPC Embedded
23661 @cindex DVC register
23662 @value{GDBN} supports using the DVC (Data Value Compare) register to
23663 implement in hardware simple hardware watchpoint conditions of the form:
23666 (@value{GDBP}) watch @var{ADDRESS|VARIABLE} \
23667 if @var{ADDRESS|VARIABLE} == @var{CONSTANT EXPRESSION}
23670 The DVC register will be automatically used when @value{GDBN} detects
23671 such pattern in a condition expression, and the created watchpoint uses one
23672 debug register (either the @code{exact-watchpoints} option is on and the
23673 variable is scalar, or the variable has a length of one byte). This feature
23674 is available in native @value{GDBN} running on a Linux kernel version 2.6.34
23677 When running on PowerPC embedded processors, @value{GDBN} automatically uses
23678 ranged hardware watchpoints, unless the @code{exact-watchpoints} option is on,
23679 in which case watchpoints using only one debug register are created when
23680 watching variables of scalar types.
23682 You can create an artificial array to watch an arbitrary memory
23683 region using one of the following commands (@pxref{Expressions}):
23686 (@value{GDBP}) watch *((char *) @var{address})@@@var{length}
23687 (@value{GDBP}) watch @{char[@var{length}]@} @var{address}
23690 PowerPC embedded processors support masked watchpoints. See the discussion
23691 about the @code{mask} argument in @ref{Set Watchpoints}.
23693 @cindex ranged breakpoint
23694 PowerPC embedded processors support hardware accelerated
23695 @dfn{ranged breakpoints}. A ranged breakpoint stops execution of
23696 the inferior whenever it executes an instruction at any address within
23697 the range it specifies. To set a ranged breakpoint in @value{GDBN},
23698 use the @code{break-range} command.
23700 @value{GDBN} provides the following PowerPC-specific commands:
23703 @kindex break-range
23704 @item break-range @var{start-location}, @var{end-location}
23705 Set a breakpoint for an address range given by
23706 @var{start-location} and @var{end-location}, which can specify a function name,
23707 a line number, an offset of lines from the current line or from the start
23708 location, or an address of an instruction (see @ref{Specify Location},
23709 for a list of all the possible ways to specify a @var{location}.)
23710 The breakpoint will stop execution of the inferior whenever it
23711 executes an instruction at any address within the specified range,
23712 (including @var{start-location} and @var{end-location}.)
23714 @kindex set powerpc
23715 @item set powerpc soft-float
23716 @itemx show powerpc soft-float
23717 Force @value{GDBN} to use (or not use) a software floating point calling
23718 convention. By default, @value{GDBN} selects the calling convention based
23719 on the selected architecture and the provided executable file.
23721 @item set powerpc vector-abi
23722 @itemx show powerpc vector-abi
23723 Force @value{GDBN} to use the specified calling convention for vector
23724 arguments and return values. The valid options are @samp{auto};
23725 @samp{generic}, to avoid vector registers even if they are present;
23726 @samp{altivec}, to use AltiVec registers; and @samp{spe} to use SPE
23727 registers. By default, @value{GDBN} selects the calling convention
23728 based on the selected architecture and the provided executable file.
23730 @item set powerpc exact-watchpoints
23731 @itemx show powerpc exact-watchpoints
23732 Allow @value{GDBN} to use only one debug register when watching a variable
23733 of scalar type, thus assuming that the variable is accessed through the
23734 address of its first byte.
23739 @subsection Atmel AVR
23742 When configured for debugging the Atmel AVR, @value{GDBN} supports the
23743 following AVR-specific commands:
23746 @item info io_registers
23747 @kindex info io_registers@r{, AVR}
23748 @cindex I/O registers (Atmel AVR)
23749 This command displays information about the AVR I/O registers. For
23750 each register, @value{GDBN} prints its number and value.
23757 When configured for debugging CRIS, @value{GDBN} provides the
23758 following CRIS-specific commands:
23761 @item set cris-version @var{ver}
23762 @cindex CRIS version
23763 Set the current CRIS version to @var{ver}, either @samp{10} or @samp{32}.
23764 The CRIS version affects register names and sizes. This command is useful in
23765 case autodetection of the CRIS version fails.
23767 @item show cris-version
23768 Show the current CRIS version.
23770 @item set cris-dwarf2-cfi
23771 @cindex DWARF-2 CFI and CRIS
23772 Set the usage of DWARF-2 CFI for CRIS debugging. The default is @samp{on}.
23773 Change to @samp{off} when using @code{gcc-cris} whose version is below
23776 @item show cris-dwarf2-cfi
23777 Show the current state of using DWARF-2 CFI.
23779 @item set cris-mode @var{mode}
23781 Set the current CRIS mode to @var{mode}. It should only be changed when
23782 debugging in guru mode, in which case it should be set to
23783 @samp{guru} (the default is @samp{normal}).
23785 @item show cris-mode
23786 Show the current CRIS mode.
23790 @subsection Renesas Super-H
23793 For the Renesas Super-H processor, @value{GDBN} provides these
23797 @item set sh calling-convention @var{convention}
23798 @kindex set sh calling-convention
23799 Set the calling-convention used when calling functions from @value{GDBN}.
23800 Allowed values are @samp{gcc}, which is the default setting, and @samp{renesas}.
23801 With the @samp{gcc} setting, functions are called using the @value{NGCC} calling
23802 convention. If the DWARF-2 information of the called function specifies
23803 that the function follows the Renesas calling convention, the function
23804 is called using the Renesas calling convention. If the calling convention
23805 is set to @samp{renesas}, the Renesas calling convention is always used,
23806 regardless of the DWARF-2 information. This can be used to override the
23807 default of @samp{gcc} if debug information is missing, or the compiler
23808 does not emit the DWARF-2 calling convention entry for a function.
23810 @item show sh calling-convention
23811 @kindex show sh calling-convention
23812 Show the current calling convention setting.
23817 @node Architectures
23818 @section Architectures
23820 This section describes characteristics of architectures that affect
23821 all uses of @value{GDBN} with the architecture, both native and cross.
23828 * HPPA:: HP PA architecture
23829 * SPU:: Cell Broadband Engine SPU architecture
23837 @subsection AArch64
23838 @cindex AArch64 support
23840 When @value{GDBN} is debugging the AArch64 architecture, it provides the
23841 following special commands:
23844 @item set debug aarch64
23845 @kindex set debug aarch64
23846 This command determines whether AArch64 architecture-specific debugging
23847 messages are to be displayed.
23849 @item show debug aarch64
23850 Show whether AArch64 debugging messages are displayed.
23854 @subsubsection AArch64 SVE.
23855 @cindex AArch64 SVE.
23857 When @value{GDBN} is debugging the AArch64 architecture, if the Scalable Vector
23858 Extension (SVE) is present, then @value{GDBN} will provide the vector registers
23859 @code{$z0} through @code{$z31}, vector predicate registers @code{$p0} through
23860 @code{$p15}, and the @code{$ffr} register. In addition, the pseudo register
23861 @code{$vg} will be provided. This is the vector granule for the current thread
23862 and represents the number of 64-bit chunks in an SVE @code{z} register.
23864 If the vector length changes, then the @code{$vg} register will be updated,
23865 but the lengths of the @code{z} and @code{p} registers will not change. This
23866 is a known limitation of @value{GDBN} and does not affect the execution of the
23871 @subsection x86 Architecture-specific Issues
23874 @item set struct-convention @var{mode}
23875 @kindex set struct-convention
23876 @cindex struct return convention
23877 @cindex struct/union returned in registers
23878 Set the convention used by the inferior to return @code{struct}s and
23879 @code{union}s from functions to @var{mode}. Possible values of
23880 @var{mode} are @code{"pcc"}, @code{"reg"}, and @code{"default"} (the
23881 default). @code{"default"} or @code{"pcc"} means that @code{struct}s
23882 are returned on the stack, while @code{"reg"} means that a
23883 @code{struct} or a @code{union} whose size is 1, 2, 4, or 8 bytes will
23884 be returned in a register.
23886 @item show struct-convention
23887 @kindex show struct-convention
23888 Show the current setting of the convention to return @code{struct}s
23893 @subsubsection Intel @dfn{Memory Protection Extensions} (MPX).
23894 @cindex Intel Memory Protection Extensions (MPX).
23896 Memory Protection Extension (MPX) adds the bound registers @samp{BND0}
23897 @footnote{The register named with capital letters represent the architecture
23898 registers.} through @samp{BND3}. Bound registers store a pair of 64-bit values
23899 which are the lower bound and upper bound. Bounds are effective addresses or
23900 memory locations. The upper bounds are architecturally represented in 1's
23901 complement form. A bound having lower bound = 0, and upper bound = 0
23902 (1's complement of all bits set) will allow access to the entire address space.
23904 @samp{BND0} through @samp{BND3} are represented in @value{GDBN} as @samp{bnd0raw}
23905 through @samp{bnd3raw}. Pseudo registers @samp{bnd0} through @samp{bnd3}
23906 display the upper bound performing the complement of one operation on the
23907 upper bound value, i.e.@ when upper bound in @samp{bnd0raw} is 0 in the
23908 @value{GDBN} @samp{bnd0} it will be @code{0xfff@dots{}}. In this sense it
23909 can also be noted that the upper bounds are inclusive.
23911 As an example, assume that the register BND0 holds bounds for a pointer having
23912 access allowed for the range between 0x32 and 0x71. The values present on
23913 bnd0raw and bnd registers are presented as follows:
23916 bnd0raw = @{0x32, 0xffffffff8e@}
23917 bnd0 = @{lbound = 0x32, ubound = 0x71@} : size 64
23920 This way the raw value can be accessed via bnd0raw@dots{}bnd3raw. Any
23921 change on bnd0@dots{}bnd3 or bnd0raw@dots{}bnd3raw is reflect on its
23922 counterpart. When the bnd0@dots{}bnd3 registers are displayed via
23923 Python, the display includes the memory size, in bits, accessible to
23926 Bounds can also be stored in bounds tables, which are stored in
23927 application memory. These tables store bounds for pointers by specifying
23928 the bounds pointer's value along with its bounds. Evaluating and changing
23929 bounds located in bound tables is therefore interesting while investigating
23930 bugs on MPX context. @value{GDBN} provides commands for this purpose:
23933 @item show mpx bound @var{pointer}
23934 @kindex show mpx bound
23935 Display bounds of the given @var{pointer}.
23937 @item set mpx bound @var{pointer}, @var{lbound}, @var{ubound}
23938 @kindex set mpx bound
23939 Set the bounds of a pointer in the bound table.
23940 This command takes three parameters: @var{pointer} is the pointers
23941 whose bounds are to be changed, @var{lbound} and @var{ubound} are new values
23942 for lower and upper bounds respectively.
23945 When you call an inferior function on an Intel MPX enabled program,
23946 GDB sets the inferior's bound registers to the init (disabled) state
23947 before calling the function. As a consequence, bounds checks for the
23948 pointer arguments passed to the function will always pass.
23950 This is necessary because when you call an inferior function, the
23951 program is usually in the middle of the execution of other function.
23952 Since at that point bound registers are in an arbitrary state, not
23953 clearing them would lead to random bound violations in the called
23956 You can still examine the influence of the bound registers on the
23957 execution of the called function by stopping the execution of the
23958 called function at its prologue, setting bound registers, and
23959 continuing the execution. For example:
23963 Breakpoint 2 at 0x4009de: file i386-mpx-call.c, line 47.
23964 $ print upper (a, b, c, d, 1)
23965 Breakpoint 2, upper (a=0x0, b=0x6e0000005b, c=0x0, d=0x0, len=48)....
23967 @{lbound = 0x0, ubound = ffffffff@} : size -1
23970 At this last step the value of bnd0 can be changed for investigation of bound
23971 violations caused along the execution of the call. In order to know how to
23972 set the bound registers or bound table for the call consult the ABI.
23977 See the following section.
23980 @subsection @acronym{MIPS}
23982 @cindex stack on Alpha
23983 @cindex stack on @acronym{MIPS}
23984 @cindex Alpha stack
23985 @cindex @acronym{MIPS} stack
23986 Alpha- and @acronym{MIPS}-based computers use an unusual stack frame, which
23987 sometimes requires @value{GDBN} to search backward in the object code to
23988 find the beginning of a function.
23990 @cindex response time, @acronym{MIPS} debugging
23991 To improve response time (especially for embedded applications, where
23992 @value{GDBN} may be restricted to a slow serial line for this search)
23993 you may want to limit the size of this search, using one of these
23997 @cindex @code{heuristic-fence-post} (Alpha, @acronym{MIPS})
23998 @item set heuristic-fence-post @var{limit}
23999 Restrict @value{GDBN} to examining at most @var{limit} bytes in its
24000 search for the beginning of a function. A value of @var{0} (the
24001 default) means there is no limit. However, except for @var{0}, the
24002 larger the limit the more bytes @code{heuristic-fence-post} must search
24003 and therefore the longer it takes to run. You should only need to use
24004 this command when debugging a stripped executable.
24006 @item show heuristic-fence-post
24007 Display the current limit.
24011 These commands are available @emph{only} when @value{GDBN} is configured
24012 for debugging programs on Alpha or @acronym{MIPS} processors.
24014 Several @acronym{MIPS}-specific commands are available when debugging @acronym{MIPS}
24018 @item set mips abi @var{arg}
24019 @kindex set mips abi
24020 @cindex set ABI for @acronym{MIPS}
24021 Tell @value{GDBN} which @acronym{MIPS} ABI is used by the inferior. Possible
24022 values of @var{arg} are:
24026 The default ABI associated with the current binary (this is the
24036 @item show mips abi
24037 @kindex show mips abi
24038 Show the @acronym{MIPS} ABI used by @value{GDBN} to debug the inferior.
24040 @item set mips compression @var{arg}
24041 @kindex set mips compression
24042 @cindex code compression, @acronym{MIPS}
24043 Tell @value{GDBN} which @acronym{MIPS} compressed
24044 @acronym{ISA, Instruction Set Architecture} encoding is used by the
24045 inferior. @value{GDBN} uses this for code disassembly and other
24046 internal interpretation purposes. This setting is only referred to
24047 when no executable has been associated with the debugging session or
24048 the executable does not provide information about the encoding it uses.
24049 Otherwise this setting is automatically updated from information
24050 provided by the executable.
24052 Possible values of @var{arg} are @samp{mips16} and @samp{micromips}.
24053 The default compressed @acronym{ISA} encoding is @samp{mips16}, as
24054 executables containing @acronym{MIPS16} code frequently are not
24055 identified as such.
24057 This setting is ``sticky''; that is, it retains its value across
24058 debugging sessions until reset either explicitly with this command or
24059 implicitly from an executable.
24061 The compiler and/or assembler typically add symbol table annotations to
24062 identify functions compiled for the @acronym{MIPS16} or
24063 @acronym{microMIPS} @acronym{ISA}s. If these function-scope annotations
24064 are present, @value{GDBN} uses them in preference to the global
24065 compressed @acronym{ISA} encoding setting.
24067 @item show mips compression
24068 @kindex show mips compression
24069 Show the @acronym{MIPS} compressed @acronym{ISA} encoding used by
24070 @value{GDBN} to debug the inferior.
24073 @itemx show mipsfpu
24074 @xref{MIPS Embedded, set mipsfpu}.
24076 @item set mips mask-address @var{arg}
24077 @kindex set mips mask-address
24078 @cindex @acronym{MIPS} addresses, masking
24079 This command determines whether the most-significant 32 bits of 64-bit
24080 @acronym{MIPS} addresses are masked off. The argument @var{arg} can be
24081 @samp{on}, @samp{off}, or @samp{auto}. The latter is the default
24082 setting, which lets @value{GDBN} determine the correct value.
24084 @item show mips mask-address
24085 @kindex show mips mask-address
24086 Show whether the upper 32 bits of @acronym{MIPS} addresses are masked off or
24089 @item set remote-mips64-transfers-32bit-regs
24090 @kindex set remote-mips64-transfers-32bit-regs
24091 This command controls compatibility with 64-bit @acronym{MIPS} targets that
24092 transfer data in 32-bit quantities. If you have an old @acronym{MIPS} 64 target
24093 that transfers 32 bits for some registers, like @sc{sr} and @sc{fsr},
24094 and 64 bits for other registers, set this option to @samp{on}.
24096 @item show remote-mips64-transfers-32bit-regs
24097 @kindex show remote-mips64-transfers-32bit-regs
24098 Show the current setting of compatibility with older @acronym{MIPS} 64 targets.
24100 @item set debug mips
24101 @kindex set debug mips
24102 This command turns on and off debugging messages for the @acronym{MIPS}-specific
24103 target code in @value{GDBN}.
24105 @item show debug mips
24106 @kindex show debug mips
24107 Show the current setting of @acronym{MIPS} debugging messages.
24113 @cindex HPPA support
24115 When @value{GDBN} is debugging the HP PA architecture, it provides the
24116 following special commands:
24119 @item set debug hppa
24120 @kindex set debug hppa
24121 This command determines whether HPPA architecture-specific debugging
24122 messages are to be displayed.
24124 @item show debug hppa
24125 Show whether HPPA debugging messages are displayed.
24127 @item maint print unwind @var{address}
24128 @kindex maint print unwind@r{, HPPA}
24129 This command displays the contents of the unwind table entry at the
24130 given @var{address}.
24136 @subsection Cell Broadband Engine SPU architecture
24137 @cindex Cell Broadband Engine
24140 When @value{GDBN} is debugging the Cell Broadband Engine SPU architecture,
24141 it provides the following special commands:
24144 @item info spu event
24146 Display SPU event facility status. Shows current event mask
24147 and pending event status.
24149 @item info spu signal
24150 Display SPU signal notification facility status. Shows pending
24151 signal-control word and signal notification mode of both signal
24152 notification channels.
24154 @item info spu mailbox
24155 Display SPU mailbox facility status. Shows all pending entries,
24156 in order of processing, in each of the SPU Write Outbound,
24157 SPU Write Outbound Interrupt, and SPU Read Inbound mailboxes.
24160 Display MFC DMA status. Shows all pending commands in the MFC
24161 DMA queue. For each entry, opcode, tag, class IDs, effective
24162 and local store addresses and transfer size are shown.
24164 @item info spu proxydma
24165 Display MFC Proxy-DMA status. Shows all pending commands in the MFC
24166 Proxy-DMA queue. For each entry, opcode, tag, class IDs, effective
24167 and local store addresses and transfer size are shown.
24171 When @value{GDBN} is debugging a combined PowerPC/SPU application
24172 on the Cell Broadband Engine, it provides in addition the following
24176 @item set spu stop-on-load @var{arg}
24178 Set whether to stop for new SPE threads. When set to @code{on}, @value{GDBN}
24179 will give control to the user when a new SPE thread enters its @code{main}
24180 function. The default is @code{off}.
24182 @item show spu stop-on-load
24184 Show whether to stop for new SPE threads.
24186 @item set spu auto-flush-cache @var{arg}
24187 Set whether to automatically flush the software-managed cache. When set to
24188 @code{on}, @value{GDBN} will automatically cause the SPE software-managed
24189 cache to be flushed whenever SPE execution stops. This provides a consistent
24190 view of PowerPC memory that is accessed via the cache. If an application
24191 does not use the software-managed cache, this option has no effect.
24193 @item show spu auto-flush-cache
24194 Show whether to automatically flush the software-managed cache.
24199 @subsection PowerPC
24200 @cindex PowerPC architecture
24202 When @value{GDBN} is debugging the PowerPC architecture, it provides a set of
24203 pseudo-registers to enable inspection of 128-bit wide Decimal Floating Point
24204 numbers stored in the floating point registers. These values must be stored
24205 in two consecutive registers, always starting at an even register like
24206 @code{f0} or @code{f2}.
24208 The pseudo-registers go from @code{$dl0} through @code{$dl15}, and are formed
24209 by joining the even/odd register pairs @code{f0} and @code{f1} for @code{$dl0},
24210 @code{f2} and @code{f3} for @code{$dl1} and so on.
24212 For POWER7 processors, @value{GDBN} provides a set of pseudo-registers, the 64-bit
24213 wide Extended Floating Point Registers (@samp{f32} through @samp{f63}).
24216 @subsection Nios II
24217 @cindex Nios II architecture
24219 When @value{GDBN} is debugging the Nios II architecture,
24220 it provides the following special commands:
24224 @item set debug nios2
24225 @kindex set debug nios2
24226 This command turns on and off debugging messages for the Nios II
24227 target code in @value{GDBN}.
24229 @item show debug nios2
24230 @kindex show debug nios2
24231 Show the current setting of Nios II debugging messages.
24235 @subsection Sparc64
24236 @cindex Sparc64 support
24237 @cindex Application Data Integrity
24238 @subsubsection ADI Support
24240 The M7 processor supports an Application Data Integrity (ADI) feature that
24241 detects invalid data accesses. When software allocates memory and enables
24242 ADI on the allocated memory, it chooses a 4-bit version number, sets the
24243 version in the upper 4 bits of the 64-bit pointer to that data, and stores
24244 the 4-bit version in every cacheline of that data. Hardware saves the latter
24245 in spare bits in the cache and memory hierarchy. On each load and store,
24246 the processor compares the upper 4 VA (virtual address) bits to the
24247 cacheline's version. If there is a mismatch, the processor generates a
24248 version mismatch trap which can be either precise or disrupting. The trap
24249 is an error condition which the kernel delivers to the process as a SIGSEGV
24252 Note that only 64-bit applications can use ADI and need to be built with
24255 Values of the ADI version tags, which are in granularity of a
24256 cacheline (64 bytes), can be viewed or modified.
24260 @kindex adi examine
24261 @item adi (examine | x) [ / @var{n} ] @var{addr}
24263 The @code{adi examine} command displays the value of one ADI version tag per
24266 @var{n} is a decimal integer specifying the number in bytes; the default
24267 is 1. It specifies how much ADI version information, at the ratio of 1:ADI
24268 block size, to display.
24270 @var{addr} is the address in user address space where you want @value{GDBN}
24271 to begin displaying the ADI version tags.
24273 Below is an example of displaying ADI versions of variable "shmaddr".
24276 (@value{GDBP}) adi x/100 shmaddr
24277 0xfff800010002c000: 0 0
24281 @item adi (assign | a) [ / @var{n} ] @var{addr} = @var{tag}
24283 The @code{adi assign} command is used to assign new ADI version tag
24286 @var{n} is a decimal integer specifying the number in bytes;
24287 the default is 1. It specifies how much ADI version information, at the
24288 ratio of 1:ADI block size, to modify.
24290 @var{addr} is the address in user address space where you want @value{GDBN}
24291 to begin modifying the ADI version tags.
24293 @var{tag} is the new ADI version tag.
24295 For example, do the following to modify then verify ADI versions of
24296 variable "shmaddr":
24299 (@value{GDBP}) adi a/100 shmaddr = 7
24300 (@value{GDBP}) adi x/100 shmaddr
24301 0xfff800010002c000: 7 7
24308 @cindex S12Z support
24310 When @value{GDBN} is debugging the S12Z architecture,
24311 it provides the following special command:
24314 @item maint info bdccsr
24315 @kindex maint info bdccsr@r{, S12Z}
24316 This command displays the current value of the microprocessor's
24321 @node Controlling GDB
24322 @chapter Controlling @value{GDBN}
24324 You can alter the way @value{GDBN} interacts with you by using the
24325 @code{set} command. For commands controlling how @value{GDBN} displays
24326 data, see @ref{Print Settings, ,Print Settings}. Other settings are
24331 * Editing:: Command editing
24332 * Command History:: Command history
24333 * Screen Size:: Screen size
24334 * Output Styling:: Output styling
24335 * Numbers:: Numbers
24336 * ABI:: Configuring the current ABI
24337 * Auto-loading:: Automatically loading associated files
24338 * Messages/Warnings:: Optional warnings and messages
24339 * Debugging Output:: Optional messages about internal happenings
24340 * Other Misc Settings:: Other Miscellaneous Settings
24348 @value{GDBN} indicates its readiness to read a command by printing a string
24349 called the @dfn{prompt}. This string is normally @samp{(@value{GDBP})}. You
24350 can change the prompt string with the @code{set prompt} command. For
24351 instance, when debugging @value{GDBN} with @value{GDBN}, it is useful to change
24352 the prompt in one of the @value{GDBN} sessions so that you can always tell
24353 which one you are talking to.
24355 @emph{Note:} @code{set prompt} does not add a space for you after the
24356 prompt you set. This allows you to set a prompt which ends in a space
24357 or a prompt that does not.
24361 @item set prompt @var{newprompt}
24362 Directs @value{GDBN} to use @var{newprompt} as its prompt string henceforth.
24364 @kindex show prompt
24366 Prints a line of the form: @samp{Gdb's prompt is: @var{your-prompt}}
24369 Versions of @value{GDBN} that ship with Python scripting enabled have
24370 prompt extensions. The commands for interacting with these extensions
24374 @kindex set extended-prompt
24375 @item set extended-prompt @var{prompt}
24376 Set an extended prompt that allows for substitutions.
24377 @xref{gdb.prompt}, for a list of escape sequences that can be used for
24378 substitution. Any escape sequences specified as part of the prompt
24379 string are replaced with the corresponding strings each time the prompt
24385 set extended-prompt Current working directory: \w (gdb)
24388 Note that when an extended-prompt is set, it takes control of the
24389 @var{prompt_hook} hook. @xref{prompt_hook}, for further information.
24391 @kindex show extended-prompt
24392 @item show extended-prompt
24393 Prints the extended prompt. Any escape sequences specified as part of
24394 the prompt string with @code{set extended-prompt}, are replaced with the
24395 corresponding strings each time the prompt is displayed.
24399 @section Command Editing
24401 @cindex command line editing
24403 @value{GDBN} reads its input commands via the @dfn{Readline} interface. This
24404 @sc{gnu} library provides consistent behavior for programs which provide a
24405 command line interface to the user. Advantages are @sc{gnu} Emacs-style
24406 or @dfn{vi}-style inline editing of commands, @code{csh}-like history
24407 substitution, and a storage and recall of command history across
24408 debugging sessions.
24410 You may control the behavior of command line editing in @value{GDBN} with the
24411 command @code{set}.
24414 @kindex set editing
24417 @itemx set editing on
24418 Enable command line editing (enabled by default).
24420 @item set editing off
24421 Disable command line editing.
24423 @kindex show editing
24425 Show whether command line editing is enabled.
24428 @ifset SYSTEM_READLINE
24429 @xref{Command Line Editing, , , rluserman, GNU Readline Library},
24431 @ifclear SYSTEM_READLINE
24432 @xref{Command Line Editing},
24434 for more details about the Readline
24435 interface. Users unfamiliar with @sc{gnu} Emacs or @code{vi} are
24436 encouraged to read that chapter.
24438 @node Command History
24439 @section Command History
24440 @cindex command history
24442 @value{GDBN} can keep track of the commands you type during your
24443 debugging sessions, so that you can be certain of precisely what
24444 happened. Use these commands to manage the @value{GDBN} command
24447 @value{GDBN} uses the @sc{gnu} History library, a part of the Readline
24448 package, to provide the history facility.
24449 @ifset SYSTEM_READLINE
24450 @xref{Using History Interactively, , , history, GNU History Library},
24452 @ifclear SYSTEM_READLINE
24453 @xref{Using History Interactively},
24455 for the detailed description of the History library.
24457 To issue a command to @value{GDBN} without affecting certain aspects of
24458 the state which is seen by users, prefix it with @samp{server }
24459 (@pxref{Server Prefix}). This
24460 means that this command will not affect the command history, nor will it
24461 affect @value{GDBN}'s notion of which command to repeat if @key{RET} is
24462 pressed on a line by itself.
24464 @cindex @code{server}, command prefix
24465 The server prefix does not affect the recording of values into the value
24466 history; to print a value without recording it into the value history,
24467 use the @code{output} command instead of the @code{print} command.
24469 Here is the description of @value{GDBN} commands related to command
24473 @cindex history substitution
24474 @cindex history file
24475 @kindex set history filename
24476 @cindex @env{GDBHISTFILE}, environment variable
24477 @item set history filename @var{fname}
24478 Set the name of the @value{GDBN} command history file to @var{fname}.
24479 This is the file where @value{GDBN} reads an initial command history
24480 list, and where it writes the command history from this session when it
24481 exits. You can access this list through history expansion or through
24482 the history command editing characters listed below. This file defaults
24483 to the value of the environment variable @code{GDBHISTFILE}, or to
24484 @file{./.gdb_history} (@file{./_gdb_history} on MS-DOS) if this variable
24487 @cindex save command history
24488 @kindex set history save
24489 @item set history save
24490 @itemx set history save on
24491 Record command history in a file, whose name may be specified with the
24492 @code{set history filename} command. By default, this option is disabled.
24494 @item set history save off
24495 Stop recording command history in a file.
24497 @cindex history size
24498 @kindex set history size
24499 @cindex @env{GDBHISTSIZE}, environment variable
24500 @item set history size @var{size}
24501 @itemx set history size unlimited
24502 Set the number of commands which @value{GDBN} keeps in its history list.
24503 This defaults to the value of the environment variable @env{GDBHISTSIZE}, or
24504 to 256 if this variable is not set. Non-numeric values of @env{GDBHISTSIZE}
24505 are ignored. If @var{size} is @code{unlimited} or if @env{GDBHISTSIZE} is
24506 either a negative number or the empty string, then the number of commands
24507 @value{GDBN} keeps in the history list is unlimited.
24509 @cindex remove duplicate history
24510 @kindex set history remove-duplicates
24511 @item set history remove-duplicates @var{count}
24512 @itemx set history remove-duplicates unlimited
24513 Control the removal of duplicate history entries in the command history list.
24514 If @var{count} is non-zero, @value{GDBN} will look back at the last @var{count}
24515 history entries and remove the first entry that is a duplicate of the current
24516 entry being added to the command history list. If @var{count} is
24517 @code{unlimited} then this lookbehind is unbounded. If @var{count} is 0, then
24518 removal of duplicate history entries is disabled.
24520 Only history entries added during the current session are considered for
24521 removal. This option is set to 0 by default.
24525 History expansion assigns special meaning to the character @kbd{!}.
24526 @ifset SYSTEM_READLINE
24527 @xref{Event Designators, , , history, GNU History Library},
24529 @ifclear SYSTEM_READLINE
24530 @xref{Event Designators},
24534 @cindex history expansion, turn on/off
24535 Since @kbd{!} is also the logical not operator in C, history expansion
24536 is off by default. If you decide to enable history expansion with the
24537 @code{set history expansion on} command, you may sometimes need to
24538 follow @kbd{!} (when it is used as logical not, in an expression) with
24539 a space or a tab to prevent it from being expanded. The readline
24540 history facilities do not attempt substitution on the strings
24541 @kbd{!=} and @kbd{!(}, even when history expansion is enabled.
24543 The commands to control history expansion are:
24546 @item set history expansion on
24547 @itemx set history expansion
24548 @kindex set history expansion
24549 Enable history expansion. History expansion is off by default.
24551 @item set history expansion off
24552 Disable history expansion.
24555 @kindex show history
24557 @itemx show history filename
24558 @itemx show history save
24559 @itemx show history size
24560 @itemx show history expansion
24561 These commands display the state of the @value{GDBN} history parameters.
24562 @code{show history} by itself displays all four states.
24567 @kindex show commands
24568 @cindex show last commands
24569 @cindex display command history
24570 @item show commands
24571 Display the last ten commands in the command history.
24573 @item show commands @var{n}
24574 Print ten commands centered on command number @var{n}.
24576 @item show commands +
24577 Print ten commands just after the commands last printed.
24581 @section Screen Size
24582 @cindex size of screen
24583 @cindex screen size
24586 @cindex pauses in output
24588 Certain commands to @value{GDBN} may produce large amounts of
24589 information output to the screen. To help you read all of it,
24590 @value{GDBN} pauses and asks you for input at the end of each page of
24591 output. Type @key{RET} when you want to see one more page of output,
24592 @kbd{q} to discard the remaining output, or @kbd{c} to continue
24593 without paging for the rest of the current command. Also, the screen
24594 width setting determines when to wrap lines of output. Depending on
24595 what is being printed, @value{GDBN} tries to break the line at a
24596 readable place, rather than simply letting it overflow onto the
24599 Normally @value{GDBN} knows the size of the screen from the terminal
24600 driver software. For example, on Unix @value{GDBN} uses the termcap data base
24601 together with the value of the @code{TERM} environment variable and the
24602 @code{stty rows} and @code{stty cols} settings. If this is not correct,
24603 you can override it with the @code{set height} and @code{set
24610 @kindex show height
24611 @item set height @var{lpp}
24612 @itemx set height unlimited
24614 @itemx set width @var{cpl}
24615 @itemx set width unlimited
24617 These @code{set} commands specify a screen height of @var{lpp} lines and
24618 a screen width of @var{cpl} characters. The associated @code{show}
24619 commands display the current settings.
24621 If you specify a height of either @code{unlimited} or zero lines,
24622 @value{GDBN} does not pause during output no matter how long the
24623 output is. This is useful if output is to a file or to an editor
24626 Likewise, you can specify @samp{set width unlimited} or @samp{set
24627 width 0} to prevent @value{GDBN} from wrapping its output.
24629 @item set pagination on
24630 @itemx set pagination off
24631 @kindex set pagination
24632 Turn the output pagination on or off; the default is on. Turning
24633 pagination off is the alternative to @code{set height unlimited}. Note that
24634 running @value{GDBN} with the @option{--batch} option (@pxref{Mode
24635 Options, -batch}) also automatically disables pagination.
24637 @item show pagination
24638 @kindex show pagination
24639 Show the current pagination mode.
24642 @node Output Styling
24643 @section Output Styling
24649 @value{GDBN} can style its output on a capable terminal. This is
24650 enabled by default on most systems, but disabled by default when in
24651 batch mode (@pxref{Mode Options}). Various style settings are available;
24652 and styles can also be disabled entirely.
24655 @item set style enabled @samp{on|off}
24656 Enable or disable all styling. The default is host-dependent, with
24657 most hosts defaulting to @samp{on}.
24659 @item show style enabled
24660 Show the current state of styling.
24662 @item set style sources @samp{on|off}
24663 Enable or disable source code styling. This affects whether source
24664 code, such as the output of the @code{list} command, is styled. Note
24665 that source styling only works if styling in general is enabled, and
24666 if @value{GDBN} was linked with the GNU Source Highlight library. The
24667 default is @samp{on}.
24669 @item show style sources
24670 Show the current state of source code styling.
24673 Subcommands of @code{set style} control specific forms of styling.
24674 These subcommands all follow the same pattern: each style-able object
24675 can be styled with a foreground color, a background color, and an
24678 For example, the style of file names can be controlled using the
24679 @code{set style filename} group of commands:
24682 @item set style filename background @var{color}
24683 Set the background to @var{color}. Valid colors are @samp{none}
24684 (meaning the terminal's default color), @samp{black}, @samp{red},
24685 @samp{green}, @samp{yellow}, @samp{blue}, @samp{magenta}, @samp{cyan},
24688 @item set style filename foreground @var{color}
24689 Set the foreground to @var{color}. Valid colors are @samp{none}
24690 (meaning the terminal's default color), @samp{black}, @samp{red},
24691 @samp{green}, @samp{yellow}, @samp{blue}, @samp{magenta}, @samp{cyan},
24694 @item set style filename intensity @var{value}
24695 Set the intensity to @var{value}. Valid intensities are @samp{normal}
24696 (the default), @samp{bold}, and @samp{dim}.
24699 The style-able objects are:
24702 Control the styling of file names. By default, this style's
24703 foreground color is green.
24706 Control the styling of function names. These are managed with the
24707 @code{set style function} family of commands. By default, this
24708 style's foreground color is yellow.
24711 Control the styling of variable names. These are managed with the
24712 @code{set style variable} family of commands. By default, this style's
24713 foreground color is cyan.
24716 Control the styling of addresses. These are managed with the
24717 @code{set style address} family of commands. By default, this style's
24718 foreground color is blue.
24723 @cindex number representation
24724 @cindex entering numbers
24726 You can always enter numbers in octal, decimal, or hexadecimal in
24727 @value{GDBN} by the usual conventions: octal numbers begin with
24728 @samp{0}, decimal numbers end with @samp{.}, and hexadecimal numbers
24729 begin with @samp{0x}. Numbers that neither begin with @samp{0} or
24730 @samp{0x}, nor end with a @samp{.} are, by default, entered in base
24731 10; likewise, the default display for numbers---when no particular
24732 format is specified---is base 10. You can change the default base for
24733 both input and output with the commands described below.
24736 @kindex set input-radix
24737 @item set input-radix @var{base}
24738 Set the default base for numeric input. Supported choices
24739 for @var{base} are decimal 8, 10, or 16. The base must itself be
24740 specified either unambiguously or using the current input radix; for
24744 set input-radix 012
24745 set input-radix 10.
24746 set input-radix 0xa
24750 sets the input base to decimal. On the other hand, @samp{set input-radix 10}
24751 leaves the input radix unchanged, no matter what it was, since
24752 @samp{10}, being without any leading or trailing signs of its base, is
24753 interpreted in the current radix. Thus, if the current radix is 16,
24754 @samp{10} is interpreted in hex, i.e.@: as 16 decimal, which doesn't
24757 @kindex set output-radix
24758 @item set output-radix @var{base}
24759 Set the default base for numeric display. Supported choices
24760 for @var{base} are decimal 8, 10, or 16. The base must itself be
24761 specified either unambiguously or using the current input radix.
24763 @kindex show input-radix
24764 @item show input-radix
24765 Display the current default base for numeric input.
24767 @kindex show output-radix
24768 @item show output-radix
24769 Display the current default base for numeric display.
24771 @item set radix @r{[}@var{base}@r{]}
24775 These commands set and show the default base for both input and output
24776 of numbers. @code{set radix} sets the radix of input and output to
24777 the same base; without an argument, it resets the radix back to its
24778 default value of 10.
24783 @section Configuring the Current ABI
24785 @value{GDBN} can determine the @dfn{ABI} (Application Binary Interface) of your
24786 application automatically. However, sometimes you need to override its
24787 conclusions. Use these commands to manage @value{GDBN}'s view of the
24793 @cindex Newlib OS ABI and its influence on the longjmp handling
24795 One @value{GDBN} configuration can debug binaries for multiple operating
24796 system targets, either via remote debugging or native emulation.
24797 @value{GDBN} will autodetect the @dfn{OS ABI} (Operating System ABI) in use,
24798 but you can override its conclusion using the @code{set osabi} command.
24799 One example where this is useful is in debugging of binaries which use
24800 an alternate C library (e.g.@: @sc{uClibc} for @sc{gnu}/Linux) which does
24801 not have the same identifying marks that the standard C library for your
24804 When @value{GDBN} is debugging the AArch64 architecture, it provides a
24805 ``Newlib'' OS ABI. This is useful for handling @code{setjmp} and
24806 @code{longjmp} when debugging binaries that use the @sc{newlib} C library.
24807 The ``Newlib'' OS ABI can be selected by @code{set osabi Newlib}.
24811 Show the OS ABI currently in use.
24814 With no argument, show the list of registered available OS ABI's.
24816 @item set osabi @var{abi}
24817 Set the current OS ABI to @var{abi}.
24820 @cindex float promotion
24822 Generally, the way that an argument of type @code{float} is passed to a
24823 function depends on whether the function is prototyped. For a prototyped
24824 (i.e.@: ANSI/ISO style) function, @code{float} arguments are passed unchanged,
24825 according to the architecture's convention for @code{float}. For unprototyped
24826 (i.e.@: K&R style) functions, @code{float} arguments are first promoted to type
24827 @code{double} and then passed.
24829 Unfortunately, some forms of debug information do not reliably indicate whether
24830 a function is prototyped. If @value{GDBN} calls a function that is not marked
24831 as prototyped, it consults @kbd{set coerce-float-to-double}.
24834 @kindex set coerce-float-to-double
24835 @item set coerce-float-to-double
24836 @itemx set coerce-float-to-double on
24837 Arguments of type @code{float} will be promoted to @code{double} when passed
24838 to an unprototyped function. This is the default setting.
24840 @item set coerce-float-to-double off
24841 Arguments of type @code{float} will be passed directly to unprototyped
24844 @kindex show coerce-float-to-double
24845 @item show coerce-float-to-double
24846 Show the current setting of promoting @code{float} to @code{double}.
24850 @kindex show cp-abi
24851 @value{GDBN} needs to know the ABI used for your program's C@t{++}
24852 objects. The correct C@t{++} ABI depends on which C@t{++} compiler was
24853 used to build your application. @value{GDBN} only fully supports
24854 programs with a single C@t{++} ABI; if your program contains code using
24855 multiple C@t{++} ABI's or if @value{GDBN} can not identify your
24856 program's ABI correctly, you can tell @value{GDBN} which ABI to use.
24857 Currently supported ABI's include ``gnu-v2'', for @code{g++} versions
24858 before 3.0, ``gnu-v3'', for @code{g++} versions 3.0 and later, and
24859 ``hpaCC'' for the HP ANSI C@t{++} compiler. Other C@t{++} compilers may
24860 use the ``gnu-v2'' or ``gnu-v3'' ABI's as well. The default setting is
24865 Show the C@t{++} ABI currently in use.
24868 With no argument, show the list of supported C@t{++} ABI's.
24870 @item set cp-abi @var{abi}
24871 @itemx set cp-abi auto
24872 Set the current C@t{++} ABI to @var{abi}, or return to automatic detection.
24876 @section Automatically loading associated files
24877 @cindex auto-loading
24879 @value{GDBN} sometimes reads files with commands and settings automatically,
24880 without being explicitly told so by the user. We call this feature
24881 @dfn{auto-loading}. While auto-loading is useful for automatically adapting
24882 @value{GDBN} to the needs of your project, it can sometimes produce unexpected
24883 results or introduce security risks (e.g., if the file comes from untrusted
24887 * Init File in the Current Directory:: @samp{set/show/info auto-load local-gdbinit}
24888 * libthread_db.so.1 file:: @samp{set/show/info auto-load libthread-db}
24890 * Auto-loading safe path:: @samp{set/show/info auto-load safe-path}
24891 * Auto-loading verbose mode:: @samp{set/show debug auto-load}
24894 There are various kinds of files @value{GDBN} can automatically load.
24895 In addition to these files, @value{GDBN} supports auto-loading code written
24896 in various extension languages. @xref{Auto-loading extensions}.
24898 Note that loading of these associated files (including the local @file{.gdbinit}
24899 file) requires accordingly configured @code{auto-load safe-path}
24900 (@pxref{Auto-loading safe path}).
24902 For these reasons, @value{GDBN} includes commands and options to let you
24903 control when to auto-load files and which files should be auto-loaded.
24906 @anchor{set auto-load off}
24907 @kindex set auto-load off
24908 @item set auto-load off
24909 Globally disable loading of all auto-loaded files.
24910 You may want to use this command with the @samp{-iex} option
24911 (@pxref{Option -init-eval-command}) such as:
24913 $ @kbd{gdb -iex "set auto-load off" untrusted-executable corefile}
24916 Be aware that system init file (@pxref{System-wide configuration})
24917 and init files from your home directory (@pxref{Home Directory Init File})
24918 still get read (as they come from generally trusted directories).
24919 To prevent @value{GDBN} from auto-loading even those init files, use the
24920 @option{-nx} option (@pxref{Mode Options}), in addition to
24921 @code{set auto-load no}.
24923 @anchor{show auto-load}
24924 @kindex show auto-load
24925 @item show auto-load
24926 Show whether auto-loading of each specific @samp{auto-load} file(s) is enabled
24930 (gdb) show auto-load
24931 gdb-scripts: Auto-loading of canned sequences of commands scripts is on.
24932 libthread-db: Auto-loading of inferior specific libthread_db is on.
24933 local-gdbinit: Auto-loading of .gdbinit script from current directory
24935 python-scripts: Auto-loading of Python scripts is on.
24936 safe-path: List of directories from which it is safe to auto-load files
24937 is $debugdir:$datadir/auto-load.
24938 scripts-directory: List of directories from which to load auto-loaded scripts
24939 is $debugdir:$datadir/auto-load.
24942 @anchor{info auto-load}
24943 @kindex info auto-load
24944 @item info auto-load
24945 Print whether each specific @samp{auto-load} file(s) have been auto-loaded or
24949 (gdb) info auto-load
24952 Yes /home/user/gdb/gdb-gdb.gdb
24953 libthread-db: No auto-loaded libthread-db.
24954 local-gdbinit: Local .gdbinit file "/home/user/gdb/.gdbinit" has been
24958 Yes /home/user/gdb/gdb-gdb.py
24962 These are @value{GDBN} control commands for the auto-loading:
24964 @multitable @columnfractions .5 .5
24965 @item @xref{set auto-load off}.
24966 @tab Disable auto-loading globally.
24967 @item @xref{show auto-load}.
24968 @tab Show setting of all kinds of files.
24969 @item @xref{info auto-load}.
24970 @tab Show state of all kinds of files.
24971 @item @xref{set auto-load gdb-scripts}.
24972 @tab Control for @value{GDBN} command scripts.
24973 @item @xref{show auto-load gdb-scripts}.
24974 @tab Show setting of @value{GDBN} command scripts.
24975 @item @xref{info auto-load gdb-scripts}.
24976 @tab Show state of @value{GDBN} command scripts.
24977 @item @xref{set auto-load python-scripts}.
24978 @tab Control for @value{GDBN} Python scripts.
24979 @item @xref{show auto-load python-scripts}.
24980 @tab Show setting of @value{GDBN} Python scripts.
24981 @item @xref{info auto-load python-scripts}.
24982 @tab Show state of @value{GDBN} Python scripts.
24983 @item @xref{set auto-load guile-scripts}.
24984 @tab Control for @value{GDBN} Guile scripts.
24985 @item @xref{show auto-load guile-scripts}.
24986 @tab Show setting of @value{GDBN} Guile scripts.
24987 @item @xref{info auto-load guile-scripts}.
24988 @tab Show state of @value{GDBN} Guile scripts.
24989 @item @xref{set auto-load scripts-directory}.
24990 @tab Control for @value{GDBN} auto-loaded scripts location.
24991 @item @xref{show auto-load scripts-directory}.
24992 @tab Show @value{GDBN} auto-loaded scripts location.
24993 @item @xref{add-auto-load-scripts-directory}.
24994 @tab Add directory for auto-loaded scripts location list.
24995 @item @xref{set auto-load local-gdbinit}.
24996 @tab Control for init file in the current directory.
24997 @item @xref{show auto-load local-gdbinit}.
24998 @tab Show setting of init file in the current directory.
24999 @item @xref{info auto-load local-gdbinit}.
25000 @tab Show state of init file in the current directory.
25001 @item @xref{set auto-load libthread-db}.
25002 @tab Control for thread debugging library.
25003 @item @xref{show auto-load libthread-db}.
25004 @tab Show setting of thread debugging library.
25005 @item @xref{info auto-load libthread-db}.
25006 @tab Show state of thread debugging library.
25007 @item @xref{set auto-load safe-path}.
25008 @tab Control directories trusted for automatic loading.
25009 @item @xref{show auto-load safe-path}.
25010 @tab Show directories trusted for automatic loading.
25011 @item @xref{add-auto-load-safe-path}.
25012 @tab Add directory trusted for automatic loading.
25015 @node Init File in the Current Directory
25016 @subsection Automatically loading init file in the current directory
25017 @cindex auto-loading init file in the current directory
25019 By default, @value{GDBN} reads and executes the canned sequences of commands
25020 from init file (if any) in the current working directory,
25021 see @ref{Init File in the Current Directory during Startup}.
25023 Note that loading of this local @file{.gdbinit} file also requires accordingly
25024 configured @code{auto-load safe-path} (@pxref{Auto-loading safe path}).
25027 @anchor{set auto-load local-gdbinit}
25028 @kindex set auto-load local-gdbinit
25029 @item set auto-load local-gdbinit [on|off]
25030 Enable or disable the auto-loading of canned sequences of commands
25031 (@pxref{Sequences}) found in init file in the current directory.
25033 @anchor{show auto-load local-gdbinit}
25034 @kindex show auto-load local-gdbinit
25035 @item show auto-load local-gdbinit
25036 Show whether auto-loading of canned sequences of commands from init file in the
25037 current directory is enabled or disabled.
25039 @anchor{info auto-load local-gdbinit}
25040 @kindex info auto-load local-gdbinit
25041 @item info auto-load local-gdbinit
25042 Print whether canned sequences of commands from init file in the
25043 current directory have been auto-loaded.
25046 @node libthread_db.so.1 file
25047 @subsection Automatically loading thread debugging library
25048 @cindex auto-loading libthread_db.so.1
25050 This feature is currently present only on @sc{gnu}/Linux native hosts.
25052 @value{GDBN} reads in some cases thread debugging library from places specific
25053 to the inferior (@pxref{set libthread-db-search-path}).
25055 The special @samp{libthread-db-search-path} entry @samp{$sdir} is processed
25056 without checking this @samp{set auto-load libthread-db} switch as system
25057 libraries have to be trusted in general. In all other cases of
25058 @samp{libthread-db-search-path} entries @value{GDBN} checks first if @samp{set
25059 auto-load libthread-db} is enabled before trying to open such thread debugging
25062 Note that loading of this debugging library also requires accordingly configured
25063 @code{auto-load safe-path} (@pxref{Auto-loading safe path}).
25066 @anchor{set auto-load libthread-db}
25067 @kindex set auto-load libthread-db
25068 @item set auto-load libthread-db [on|off]
25069 Enable or disable the auto-loading of inferior specific thread debugging library.
25071 @anchor{show auto-load libthread-db}
25072 @kindex show auto-load libthread-db
25073 @item show auto-load libthread-db
25074 Show whether auto-loading of inferior specific thread debugging library is
25075 enabled or disabled.
25077 @anchor{info auto-load libthread-db}
25078 @kindex info auto-load libthread-db
25079 @item info auto-load libthread-db
25080 Print the list of all loaded inferior specific thread debugging libraries and
25081 for each such library print list of inferior @var{pid}s using it.
25084 @node Auto-loading safe path
25085 @subsection Security restriction for auto-loading
25086 @cindex auto-loading safe-path
25088 As the files of inferior can come from untrusted source (such as submitted by
25089 an application user) @value{GDBN} does not always load any files automatically.
25090 @value{GDBN} provides the @samp{set auto-load safe-path} setting to list
25091 directories trusted for loading files not explicitly requested by user.
25092 Each directory can also be a shell wildcard pattern.
25094 If the path is not set properly you will see a warning and the file will not
25099 Reading symbols from /home/user/gdb/gdb...done.
25100 warning: File "/home/user/gdb/gdb-gdb.gdb" auto-loading has been
25101 declined by your `auto-load safe-path' set
25102 to "$debugdir:$datadir/auto-load".
25103 warning: File "/home/user/gdb/gdb-gdb.py" auto-loading has been
25104 declined by your `auto-load safe-path' set
25105 to "$debugdir:$datadir/auto-load".
25109 To instruct @value{GDBN} to go ahead and use the init files anyway,
25110 invoke @value{GDBN} like this:
25113 $ gdb -q -iex "set auto-load safe-path /home/user/gdb" ./gdb
25116 The list of trusted directories is controlled by the following commands:
25119 @anchor{set auto-load safe-path}
25120 @kindex set auto-load safe-path
25121 @item set auto-load safe-path @r{[}@var{directories}@r{]}
25122 Set the list of directories (and their subdirectories) trusted for automatic
25123 loading and execution of scripts. You can also enter a specific trusted file.
25124 Each directory can also be a shell wildcard pattern; wildcards do not match
25125 directory separator - see @code{FNM_PATHNAME} for system function @code{fnmatch}
25126 (@pxref{Wildcard Matching, fnmatch, , libc, GNU C Library Reference Manual}).
25127 If you omit @var{directories}, @samp{auto-load safe-path} will be reset to
25128 its default value as specified during @value{GDBN} compilation.
25130 The list of directories uses path separator (@samp{:} on GNU and Unix
25131 systems, @samp{;} on MS-Windows and MS-DOS) to separate directories, similarly
25132 to the @env{PATH} environment variable.
25134 @anchor{show auto-load safe-path}
25135 @kindex show auto-load safe-path
25136 @item show auto-load safe-path
25137 Show the list of directories trusted for automatic loading and execution of
25140 @anchor{add-auto-load-safe-path}
25141 @kindex add-auto-load-safe-path
25142 @item add-auto-load-safe-path
25143 Add an entry (or list of entries) to the list of directories trusted for
25144 automatic loading and execution of scripts. Multiple entries may be delimited
25145 by the host platform path separator in use.
25148 This variable defaults to what @code{--with-auto-load-dir} has been configured
25149 to (@pxref{with-auto-load-dir}). @file{$debugdir} and @file{$datadir}
25150 substitution applies the same as for @ref{set auto-load scripts-directory}.
25151 The default @code{set auto-load safe-path} value can be also overriden by
25152 @value{GDBN} configuration option @option{--with-auto-load-safe-path}.
25154 Setting this variable to @file{/} disables this security protection,
25155 corresponding @value{GDBN} configuration option is
25156 @option{--without-auto-load-safe-path}.
25157 This variable is supposed to be set to the system directories writable by the
25158 system superuser only. Users can add their source directories in init files in
25159 their home directories (@pxref{Home Directory Init File}). See also deprecated
25160 init file in the current directory
25161 (@pxref{Init File in the Current Directory during Startup}).
25163 To force @value{GDBN} to load the files it declined to load in the previous
25164 example, you could use one of the following ways:
25167 @item @file{~/.gdbinit}: @samp{add-auto-load-safe-path ~/src/gdb}
25168 Specify this trusted directory (or a file) as additional component of the list.
25169 You have to specify also any existing directories displayed by
25170 by @samp{show auto-load safe-path} (such as @samp{/usr:/bin} in this example).
25172 @item @kbd{gdb -iex "set auto-load safe-path /usr:/bin:~/src/gdb" @dots{}}
25173 Specify this directory as in the previous case but just for a single
25174 @value{GDBN} session.
25176 @item @kbd{gdb -iex "set auto-load safe-path /" @dots{}}
25177 Disable auto-loading safety for a single @value{GDBN} session.
25178 This assumes all the files you debug during this @value{GDBN} session will come
25179 from trusted sources.
25181 @item @kbd{./configure --without-auto-load-safe-path}
25182 During compilation of @value{GDBN} you may disable any auto-loading safety.
25183 This assumes all the files you will ever debug with this @value{GDBN} come from
25187 On the other hand you can also explicitly forbid automatic files loading which
25188 also suppresses any such warning messages:
25191 @item @kbd{gdb -iex "set auto-load no" @dots{}}
25192 You can use @value{GDBN} command-line option for a single @value{GDBN} session.
25194 @item @file{~/.gdbinit}: @samp{set auto-load no}
25195 Disable auto-loading globally for the user
25196 (@pxref{Home Directory Init File}). While it is improbable, you could also
25197 use system init file instead (@pxref{System-wide configuration}).
25200 This setting applies to the file names as entered by user. If no entry matches
25201 @value{GDBN} tries as a last resort to also resolve all the file names into
25202 their canonical form (typically resolving symbolic links) and compare the
25203 entries again. @value{GDBN} already canonicalizes most of the filenames on its
25204 own before starting the comparison so a canonical form of directories is
25205 recommended to be entered.
25207 @node Auto-loading verbose mode
25208 @subsection Displaying files tried for auto-load
25209 @cindex auto-loading verbose mode
25211 For better visibility of all the file locations where you can place scripts to
25212 be auto-loaded with inferior --- or to protect yourself against accidental
25213 execution of untrusted scripts --- @value{GDBN} provides a feature for printing
25214 all the files attempted to be loaded. Both existing and non-existing files may
25217 For example the list of directories from which it is safe to auto-load files
25218 (@pxref{Auto-loading safe path}) applies also to canonicalized filenames which
25219 may not be too obvious while setting it up.
25222 (gdb) set debug auto-load on
25223 (gdb) file ~/src/t/true
25224 auto-load: Loading canned sequences of commands script "/tmp/true-gdb.gdb"
25225 for objfile "/tmp/true".
25226 auto-load: Updating directories of "/usr:/opt".
25227 auto-load: Using directory "/usr".
25228 auto-load: Using directory "/opt".
25229 warning: File "/tmp/true-gdb.gdb" auto-loading has been declined
25230 by your `auto-load safe-path' set to "/usr:/opt".
25234 @anchor{set debug auto-load}
25235 @kindex set debug auto-load
25236 @item set debug auto-load [on|off]
25237 Set whether to print the filenames attempted to be auto-loaded.
25239 @anchor{show debug auto-load}
25240 @kindex show debug auto-load
25241 @item show debug auto-load
25242 Show whether printing of the filenames attempted to be auto-loaded is turned
25246 @node Messages/Warnings
25247 @section Optional Warnings and Messages
25249 @cindex verbose operation
25250 @cindex optional warnings
25251 By default, @value{GDBN} is silent about its inner workings. If you are
25252 running on a slow machine, you may want to use the @code{set verbose}
25253 command. This makes @value{GDBN} tell you when it does a lengthy
25254 internal operation, so you will not think it has crashed.
25256 Currently, the messages controlled by @code{set verbose} are those
25257 which announce that the symbol table for a source file is being read;
25258 see @code{symbol-file} in @ref{Files, ,Commands to Specify Files}.
25261 @kindex set verbose
25262 @item set verbose on
25263 Enables @value{GDBN} output of certain informational messages.
25265 @item set verbose off
25266 Disables @value{GDBN} output of certain informational messages.
25268 @kindex show verbose
25270 Displays whether @code{set verbose} is on or off.
25273 By default, if @value{GDBN} encounters bugs in the symbol table of an
25274 object file, it is silent; but if you are debugging a compiler, you may
25275 find this information useful (@pxref{Symbol Errors, ,Errors Reading
25280 @kindex set complaints
25281 @item set complaints @var{limit}
25282 Permits @value{GDBN} to output @var{limit} complaints about each type of
25283 unusual symbols before becoming silent about the problem. Set
25284 @var{limit} to zero to suppress all complaints; set it to a large number
25285 to prevent complaints from being suppressed.
25287 @kindex show complaints
25288 @item show complaints
25289 Displays how many symbol complaints @value{GDBN} is permitted to produce.
25293 @anchor{confirmation requests}
25294 By default, @value{GDBN} is cautious, and asks what sometimes seems to be a
25295 lot of stupid questions to confirm certain commands. For example, if
25296 you try to run a program which is already running:
25300 The program being debugged has been started already.
25301 Start it from the beginning? (y or n)
25304 If you are willing to unflinchingly face the consequences of your own
25305 commands, you can disable this ``feature'':
25309 @kindex set confirm
25311 @cindex confirmation
25312 @cindex stupid questions
25313 @item set confirm off
25314 Disables confirmation requests. Note that running @value{GDBN} with
25315 the @option{--batch} option (@pxref{Mode Options, -batch}) also
25316 automatically disables confirmation requests.
25318 @item set confirm on
25319 Enables confirmation requests (the default).
25321 @kindex show confirm
25323 Displays state of confirmation requests.
25327 @cindex command tracing
25328 If you need to debug user-defined commands or sourced files you may find it
25329 useful to enable @dfn{command tracing}. In this mode each command will be
25330 printed as it is executed, prefixed with one or more @samp{+} symbols, the
25331 quantity denoting the call depth of each command.
25334 @kindex set trace-commands
25335 @cindex command scripts, debugging
25336 @item set trace-commands on
25337 Enable command tracing.
25338 @item set trace-commands off
25339 Disable command tracing.
25340 @item show trace-commands
25341 Display the current state of command tracing.
25344 @node Debugging Output
25345 @section Optional Messages about Internal Happenings
25346 @cindex optional debugging messages
25348 @value{GDBN} has commands that enable optional debugging messages from
25349 various @value{GDBN} subsystems; normally these commands are of
25350 interest to @value{GDBN} maintainers, or when reporting a bug. This
25351 section documents those commands.
25354 @kindex set exec-done-display
25355 @item set exec-done-display
25356 Turns on or off the notification of asynchronous commands'
25357 completion. When on, @value{GDBN} will print a message when an
25358 asynchronous command finishes its execution. The default is off.
25359 @kindex show exec-done-display
25360 @item show exec-done-display
25361 Displays the current setting of asynchronous command completion
25364 @cindex ARM AArch64
25365 @item set debug aarch64
25366 Turns on or off display of debugging messages related to ARM AArch64.
25367 The default is off.
25369 @item show debug aarch64
25370 Displays the current state of displaying debugging messages related to
25372 @cindex gdbarch debugging info
25373 @cindex architecture debugging info
25374 @item set debug arch
25375 Turns on or off display of gdbarch debugging info. The default is off
25376 @item show debug arch
25377 Displays the current state of displaying gdbarch debugging info.
25378 @item set debug aix-solib
25379 @cindex AIX shared library debugging
25380 Control display of debugging messages from the AIX shared library
25381 support module. The default is off.
25382 @item show debug aix-thread
25383 Show the current state of displaying AIX shared library debugging messages.
25384 @item set debug aix-thread
25385 @cindex AIX threads
25386 Display debugging messages about inner workings of the AIX thread
25388 @item show debug aix-thread
25389 Show the current state of AIX thread debugging info display.
25390 @item set debug check-physname
25392 Check the results of the ``physname'' computation. When reading DWARF
25393 debugging information for C@t{++}, @value{GDBN} attempts to compute
25394 each entity's name. @value{GDBN} can do this computation in two
25395 different ways, depending on exactly what information is present.
25396 When enabled, this setting causes @value{GDBN} to compute the names
25397 both ways and display any discrepancies.
25398 @item show debug check-physname
25399 Show the current state of ``physname'' checking.
25400 @item set debug coff-pe-read
25401 @cindex COFF/PE exported symbols
25402 Control display of debugging messages related to reading of COFF/PE
25403 exported symbols. The default is off.
25404 @item show debug coff-pe-read
25405 Displays the current state of displaying debugging messages related to
25406 reading of COFF/PE exported symbols.
25407 @item set debug dwarf-die
25409 Dump DWARF DIEs after they are read in.
25410 The value is the number of nesting levels to print.
25411 A value of zero turns off the display.
25412 @item show debug dwarf-die
25413 Show the current state of DWARF DIE debugging.
25414 @item set debug dwarf-line
25415 @cindex DWARF Line Tables
25416 Turns on or off display of debugging messages related to reading
25417 DWARF line tables. The default is 0 (off).
25418 A value of 1 provides basic information.
25419 A value greater than 1 provides more verbose information.
25420 @item show debug dwarf-line
25421 Show the current state of DWARF line table debugging.
25422 @item set debug dwarf-read
25423 @cindex DWARF Reading
25424 Turns on or off display of debugging messages related to reading
25425 DWARF debug info. The default is 0 (off).
25426 A value of 1 provides basic information.
25427 A value greater than 1 provides more verbose information.
25428 @item show debug dwarf-read
25429 Show the current state of DWARF reader debugging.
25430 @item set debug displaced
25431 @cindex displaced stepping debugging info
25432 Turns on or off display of @value{GDBN} debugging info for the
25433 displaced stepping support. The default is off.
25434 @item show debug displaced
25435 Displays the current state of displaying @value{GDBN} debugging info
25436 related to displaced stepping.
25437 @item set debug event
25438 @cindex event debugging info
25439 Turns on or off display of @value{GDBN} event debugging info. The
25441 @item show debug event
25442 Displays the current state of displaying @value{GDBN} event debugging
25444 @item set debug expression
25445 @cindex expression debugging info
25446 Turns on or off display of debugging info about @value{GDBN}
25447 expression parsing. The default is off.
25448 @item show debug expression
25449 Displays the current state of displaying debugging info about
25450 @value{GDBN} expression parsing.
25451 @item set debug fbsd-lwp
25452 @cindex FreeBSD LWP debug messages
25453 Turns on or off debugging messages from the FreeBSD LWP debug support.
25454 @item show debug fbsd-lwp
25455 Show the current state of FreeBSD LWP debugging messages.
25456 @item set debug fbsd-nat
25457 @cindex FreeBSD native target debug messages
25458 Turns on or off debugging messages from the FreeBSD native target.
25459 @item show debug fbsd-nat
25460 Show the current state of FreeBSD native target debugging messages.
25461 @item set debug frame
25462 @cindex frame debugging info
25463 Turns on or off display of @value{GDBN} frame debugging info. The
25465 @item show debug frame
25466 Displays the current state of displaying @value{GDBN} frame debugging
25468 @item set debug gnu-nat
25469 @cindex @sc{gnu}/Hurd debug messages
25470 Turn on or off debugging messages from the @sc{gnu}/Hurd debug support.
25471 @item show debug gnu-nat
25472 Show the current state of @sc{gnu}/Hurd debugging messages.
25473 @item set debug infrun
25474 @cindex inferior debugging info
25475 Turns on or off display of @value{GDBN} debugging info for running the inferior.
25476 The default is off. @file{infrun.c} contains GDB's runtime state machine used
25477 for implementing operations such as single-stepping the inferior.
25478 @item show debug infrun
25479 Displays the current state of @value{GDBN} inferior debugging.
25480 @item set debug jit
25481 @cindex just-in-time compilation, debugging messages
25482 Turn on or off debugging messages from JIT debug support.
25483 @item show debug jit
25484 Displays the current state of @value{GDBN} JIT debugging.
25485 @item set debug lin-lwp
25486 @cindex @sc{gnu}/Linux LWP debug messages
25487 @cindex Linux lightweight processes
25488 Turn on or off debugging messages from the Linux LWP debug support.
25489 @item show debug lin-lwp
25490 Show the current state of Linux LWP debugging messages.
25491 @item set debug linux-namespaces
25492 @cindex @sc{gnu}/Linux namespaces debug messages
25493 Turn on or off debugging messages from the Linux namespaces debug support.
25494 @item show debug linux-namespaces
25495 Show the current state of Linux namespaces debugging messages.
25496 @item set debug mach-o
25497 @cindex Mach-O symbols processing
25498 Control display of debugging messages related to Mach-O symbols
25499 processing. The default is off.
25500 @item show debug mach-o
25501 Displays the current state of displaying debugging messages related to
25502 reading of COFF/PE exported symbols.
25503 @item set debug notification
25504 @cindex remote async notification debugging info
25505 Turn on or off debugging messages about remote async notification.
25506 The default is off.
25507 @item show debug notification
25508 Displays the current state of remote async notification debugging messages.
25509 @item set debug observer
25510 @cindex observer debugging info
25511 Turns on or off display of @value{GDBN} observer debugging. This
25512 includes info such as the notification of observable events.
25513 @item show debug observer
25514 Displays the current state of observer debugging.
25515 @item set debug overload
25516 @cindex C@t{++} overload debugging info
25517 Turns on or off display of @value{GDBN} C@t{++} overload debugging
25518 info. This includes info such as ranking of functions, etc. The default
25520 @item show debug overload
25521 Displays the current state of displaying @value{GDBN} C@t{++} overload
25523 @cindex expression parser, debugging info
25524 @cindex debug expression parser
25525 @item set debug parser
25526 Turns on or off the display of expression parser debugging output.
25527 Internally, this sets the @code{yydebug} variable in the expression
25528 parser. @xref{Tracing, , Tracing Your Parser, bison, Bison}, for
25529 details. The default is off.
25530 @item show debug parser
25531 Show the current state of expression parser debugging.
25532 @cindex packets, reporting on stdout
25533 @cindex serial connections, debugging
25534 @cindex debug remote protocol
25535 @cindex remote protocol debugging
25536 @cindex display remote packets
25537 @item set debug remote
25538 Turns on or off display of reports on all packets sent back and forth across
25539 the serial line to the remote machine. The info is printed on the
25540 @value{GDBN} standard output stream. The default is off.
25541 @item show debug remote
25542 Displays the state of display of remote packets.
25544 @item set debug separate-debug-file
25545 Turns on or off display of debug output about separate debug file search.
25546 @item show debug separate-debug-file
25547 Displays the state of separate debug file search debug output.
25549 @item set debug serial
25550 Turns on or off display of @value{GDBN} serial debugging info. The
25552 @item show debug serial
25553 Displays the current state of displaying @value{GDBN} serial debugging
25555 @item set debug solib-frv
25556 @cindex FR-V shared-library debugging
25557 Turn on or off debugging messages for FR-V shared-library code.
25558 @item show debug solib-frv
25559 Display the current state of FR-V shared-library code debugging
25561 @item set debug symbol-lookup
25562 @cindex symbol lookup
25563 Turns on or off display of debugging messages related to symbol lookup.
25564 The default is 0 (off).
25565 A value of 1 provides basic information.
25566 A value greater than 1 provides more verbose information.
25567 @item show debug symbol-lookup
25568 Show the current state of symbol lookup debugging messages.
25569 @item set debug symfile
25570 @cindex symbol file functions
25571 Turns on or off display of debugging messages related to symbol file functions.
25572 The default is off. @xref{Files}.
25573 @item show debug symfile
25574 Show the current state of symbol file debugging messages.
25575 @item set debug symtab-create
25576 @cindex symbol table creation
25577 Turns on or off display of debugging messages related to symbol table creation.
25578 The default is 0 (off).
25579 A value of 1 provides basic information.
25580 A value greater than 1 provides more verbose information.
25581 @item show debug symtab-create
25582 Show the current state of symbol table creation debugging.
25583 @item set debug target
25584 @cindex target debugging info
25585 Turns on or off display of @value{GDBN} target debugging info. This info
25586 includes what is going on at the target level of GDB, as it happens. The
25587 default is 0. Set it to 1 to track events, and to 2 to also track the
25588 value of large memory transfers.
25589 @item show debug target
25590 Displays the current state of displaying @value{GDBN} target debugging
25592 @item set debug timestamp
25593 @cindex timestampping debugging info
25594 Turns on or off display of timestamps with @value{GDBN} debugging info.
25595 When enabled, seconds and microseconds are displayed before each debugging
25597 @item show debug timestamp
25598 Displays the current state of displaying timestamps with @value{GDBN}
25600 @item set debug varobj
25601 @cindex variable object debugging info
25602 Turns on or off display of @value{GDBN} variable object debugging
25603 info. The default is off.
25604 @item show debug varobj
25605 Displays the current state of displaying @value{GDBN} variable object
25607 @item set debug xml
25608 @cindex XML parser debugging
25609 Turn on or off debugging messages for built-in XML parsers.
25610 @item show debug xml
25611 Displays the current state of XML debugging messages.
25614 @node Other Misc Settings
25615 @section Other Miscellaneous Settings
25616 @cindex miscellaneous settings
25619 @kindex set interactive-mode
25620 @item set interactive-mode
25621 If @code{on}, forces @value{GDBN} to assume that GDB was started
25622 in a terminal. In practice, this means that @value{GDBN} should wait
25623 for the user to answer queries generated by commands entered at
25624 the command prompt. If @code{off}, forces @value{GDBN} to operate
25625 in the opposite mode, and it uses the default answers to all queries.
25626 If @code{auto} (the default), @value{GDBN} tries to determine whether
25627 its standard input is a terminal, and works in interactive-mode if it
25628 is, non-interactively otherwise.
25630 In the vast majority of cases, the debugger should be able to guess
25631 correctly which mode should be used. But this setting can be useful
25632 in certain specific cases, such as running a MinGW @value{GDBN}
25633 inside a cygwin window.
25635 @kindex show interactive-mode
25636 @item show interactive-mode
25637 Displays whether the debugger is operating in interactive mode or not.
25640 @node Extending GDB
25641 @chapter Extending @value{GDBN}
25642 @cindex extending GDB
25644 @value{GDBN} provides several mechanisms for extension.
25645 @value{GDBN} also provides the ability to automatically load
25646 extensions when it reads a file for debugging. This allows the
25647 user to automatically customize @value{GDBN} for the program
25651 * Sequences:: Canned Sequences of @value{GDBN} Commands
25652 * Python:: Extending @value{GDBN} using Python
25653 * Guile:: Extending @value{GDBN} using Guile
25654 * Auto-loading extensions:: Automatically loading extensions
25655 * Multiple Extension Languages:: Working with multiple extension languages
25656 * Aliases:: Creating new spellings of existing commands
25659 To facilitate the use of extension languages, @value{GDBN} is capable
25660 of evaluating the contents of a file. When doing so, @value{GDBN}
25661 can recognize which extension language is being used by looking at
25662 the filename extension. Files with an unrecognized filename extension
25663 are always treated as a @value{GDBN} Command Files.
25664 @xref{Command Files,, Command files}.
25666 You can control how @value{GDBN} evaluates these files with the following
25670 @kindex set script-extension
25671 @kindex show script-extension
25672 @item set script-extension off
25673 All scripts are always evaluated as @value{GDBN} Command Files.
25675 @item set script-extension soft
25676 The debugger determines the scripting language based on filename
25677 extension. If this scripting language is supported, @value{GDBN}
25678 evaluates the script using that language. Otherwise, it evaluates
25679 the file as a @value{GDBN} Command File.
25681 @item set script-extension strict
25682 The debugger determines the scripting language based on filename
25683 extension, and evaluates the script using that language. If the
25684 language is not supported, then the evaluation fails.
25686 @item show script-extension
25687 Display the current value of the @code{script-extension} option.
25692 @section Canned Sequences of Commands
25694 Aside from breakpoint commands (@pxref{Break Commands, ,Breakpoint
25695 Command Lists}), @value{GDBN} provides two ways to store sequences of
25696 commands for execution as a unit: user-defined commands and command
25700 * Define:: How to define your own commands
25701 * Hooks:: Hooks for user-defined commands
25702 * Command Files:: How to write scripts of commands to be stored in a file
25703 * Output:: Commands for controlled output
25704 * Auto-loading sequences:: Controlling auto-loaded command files
25708 @subsection User-defined Commands
25710 @cindex user-defined command
25711 @cindex arguments, to user-defined commands
25712 A @dfn{user-defined command} is a sequence of @value{GDBN} commands to
25713 which you assign a new name as a command. This is done with the
25714 @code{define} command. User commands may accept an unlimited number of arguments
25715 separated by whitespace. Arguments are accessed within the user command
25716 via @code{$arg0@dots{}$argN}. A trivial example:
25720 print $arg0 + $arg1 + $arg2
25725 To execute the command use:
25732 This defines the command @code{adder}, which prints the sum of
25733 its three arguments. Note the arguments are text substitutions, so they may
25734 reference variables, use complex expressions, or even perform inferior
25737 @cindex argument count in user-defined commands
25738 @cindex how many arguments (user-defined commands)
25739 In addition, @code{$argc} may be used to find out how many arguments have
25745 print $arg0 + $arg1
25748 print $arg0 + $arg1 + $arg2
25753 Combining with the @code{eval} command (@pxref{eval}) makes it easier
25754 to process a variable number of arguments:
25761 eval "set $sum = $sum + $arg%d", $i
25771 @item define @var{commandname}
25772 Define a command named @var{commandname}. If there is already a command
25773 by that name, you are asked to confirm that you want to redefine it.
25774 The argument @var{commandname} may be a bare command name consisting of letters,
25775 numbers, dashes, and underscores. It may also start with any predefined
25776 prefix command. For example, @samp{define target my-target} creates
25777 a user-defined @samp{target my-target} command.
25779 The definition of the command is made up of other @value{GDBN} command lines,
25780 which are given following the @code{define} command. The end of these
25781 commands is marked by a line containing @code{end}.
25784 @kindex end@r{ (user-defined commands)}
25785 @item document @var{commandname}
25786 Document the user-defined command @var{commandname}, so that it can be
25787 accessed by @code{help}. The command @var{commandname} must already be
25788 defined. This command reads lines of documentation just as @code{define}
25789 reads the lines of the command definition, ending with @code{end}.
25790 After the @code{document} command is finished, @code{help} on command
25791 @var{commandname} displays the documentation you have written.
25793 You may use the @code{document} command again to change the
25794 documentation of a command. Redefining the command with @code{define}
25795 does not change the documentation.
25797 @kindex dont-repeat
25798 @cindex don't repeat command
25800 Used inside a user-defined command, this tells @value{GDBN} that this
25801 command should not be repeated when the user hits @key{RET}
25802 (@pxref{Command Syntax, repeat last command}).
25804 @kindex help user-defined
25805 @item help user-defined
25806 List all user-defined commands and all python commands defined in class
25807 COMAND_USER. The first line of the documentation or docstring is
25812 @itemx show user @var{commandname}
25813 Display the @value{GDBN} commands used to define @var{commandname} (but
25814 not its documentation). If no @var{commandname} is given, display the
25815 definitions for all user-defined commands.
25816 This does not work for user-defined python commands.
25818 @cindex infinite recursion in user-defined commands
25819 @kindex show max-user-call-depth
25820 @kindex set max-user-call-depth
25821 @item show max-user-call-depth
25822 @itemx set max-user-call-depth
25823 The value of @code{max-user-call-depth} controls how many recursion
25824 levels are allowed in user-defined commands before @value{GDBN} suspects an
25825 infinite recursion and aborts the command.
25826 This does not apply to user-defined python commands.
25829 In addition to the above commands, user-defined commands frequently
25830 use control flow commands, described in @ref{Command Files}.
25832 When user-defined commands are executed, the
25833 commands of the definition are not printed. An error in any command
25834 stops execution of the user-defined command.
25836 If used interactively, commands that would ask for confirmation proceed
25837 without asking when used inside a user-defined command. Many @value{GDBN}
25838 commands that normally print messages to say what they are doing omit the
25839 messages when used in a user-defined command.
25842 @subsection User-defined Command Hooks
25843 @cindex command hooks
25844 @cindex hooks, for commands
25845 @cindex hooks, pre-command
25848 You may define @dfn{hooks}, which are a special kind of user-defined
25849 command. Whenever you run the command @samp{foo}, if the user-defined
25850 command @samp{hook-foo} exists, it is executed (with no arguments)
25851 before that command.
25853 @cindex hooks, post-command
25855 A hook may also be defined which is run after the command you executed.
25856 Whenever you run the command @samp{foo}, if the user-defined command
25857 @samp{hookpost-foo} exists, it is executed (with no arguments) after
25858 that command. Post-execution hooks may exist simultaneously with
25859 pre-execution hooks, for the same command.
25861 It is valid for a hook to call the command which it hooks. If this
25862 occurs, the hook is not re-executed, thereby avoiding infinite recursion.
25864 @c It would be nice if hookpost could be passed a parameter indicating
25865 @c if the command it hooks executed properly or not. FIXME!
25867 @kindex stop@r{, a pseudo-command}
25868 In addition, a pseudo-command, @samp{stop} exists. Defining
25869 (@samp{hook-stop}) makes the associated commands execute every time
25870 execution stops in your program: before breakpoint commands are run,
25871 displays are printed, or the stack frame is printed.
25873 For example, to ignore @code{SIGALRM} signals while
25874 single-stepping, but treat them normally during normal execution,
25879 handle SIGALRM nopass
25883 handle SIGALRM pass
25886 define hook-continue
25887 handle SIGALRM pass
25891 As a further example, to hook at the beginning and end of the @code{echo}
25892 command, and to add extra text to the beginning and end of the message,
25900 define hookpost-echo
25904 (@value{GDBP}) echo Hello World
25905 <<<---Hello World--->>>
25910 You can define a hook for any single-word command in @value{GDBN}, but
25911 not for command aliases; you should define a hook for the basic command
25912 name, e.g.@: @code{backtrace} rather than @code{bt}.
25913 @c FIXME! So how does Joe User discover whether a command is an alias
25915 You can hook a multi-word command by adding @code{hook-} or
25916 @code{hookpost-} to the last word of the command, e.g.@:
25917 @samp{define target hook-remote} to add a hook to @samp{target remote}.
25919 If an error occurs during the execution of your hook, execution of
25920 @value{GDBN} commands stops and @value{GDBN} issues a prompt
25921 (before the command that you actually typed had a chance to run).
25923 If you try to define a hook which does not match any known command, you
25924 get a warning from the @code{define} command.
25926 @node Command Files
25927 @subsection Command Files
25929 @cindex command files
25930 @cindex scripting commands
25931 A command file for @value{GDBN} is a text file made of lines that are
25932 @value{GDBN} commands. Comments (lines starting with @kbd{#}) may
25933 also be included. An empty line in a command file does nothing; it
25934 does not mean to repeat the last command, as it would from the
25937 You can request the execution of a command file with the @code{source}
25938 command. Note that the @code{source} command is also used to evaluate
25939 scripts that are not Command Files. The exact behavior can be configured
25940 using the @code{script-extension} setting.
25941 @xref{Extending GDB,, Extending GDB}.
25945 @cindex execute commands from a file
25946 @item source [-s] [-v] @var{filename}
25947 Execute the command file @var{filename}.
25950 The lines in a command file are generally executed sequentially,
25951 unless the order of execution is changed by one of the
25952 @emph{flow-control commands} described below. The commands are not
25953 printed as they are executed. An error in any command terminates
25954 execution of the command file and control is returned to the console.
25956 @value{GDBN} first searches for @var{filename} in the current directory.
25957 If the file is not found there, and @var{filename} does not specify a
25958 directory, then @value{GDBN} also looks for the file on the source search path
25959 (specified with the @samp{directory} command);
25960 except that @file{$cdir} is not searched because the compilation directory
25961 is not relevant to scripts.
25963 If @code{-s} is specified, then @value{GDBN} searches for @var{filename}
25964 on the search path even if @var{filename} specifies a directory.
25965 The search is done by appending @var{filename} to each element of the
25966 search path. So, for example, if @var{filename} is @file{mylib/myscript}
25967 and the search path contains @file{/home/user} then @value{GDBN} will
25968 look for the script @file{/home/user/mylib/myscript}.
25969 The search is also done if @var{filename} is an absolute path.
25970 For example, if @var{filename} is @file{/tmp/myscript} and
25971 the search path contains @file{/home/user} then @value{GDBN} will
25972 look for the script @file{/home/user/tmp/myscript}.
25973 For DOS-like systems, if @var{filename} contains a drive specification,
25974 it is stripped before concatenation. For example, if @var{filename} is
25975 @file{d:myscript} and the search path contains @file{c:/tmp} then @value{GDBN}
25976 will look for the script @file{c:/tmp/myscript}.
25978 If @code{-v}, for verbose mode, is given then @value{GDBN} displays
25979 each command as it is executed. The option must be given before
25980 @var{filename}, and is interpreted as part of the filename anywhere else.
25982 Commands that would ask for confirmation if used interactively proceed
25983 without asking when used in a command file. Many @value{GDBN} commands that
25984 normally print messages to say what they are doing omit the messages
25985 when called from command files.
25987 @value{GDBN} also accepts command input from standard input. In this
25988 mode, normal output goes to standard output and error output goes to
25989 standard error. Errors in a command file supplied on standard input do
25990 not terminate execution of the command file---execution continues with
25994 gdb < cmds > log 2>&1
25997 (The syntax above will vary depending on the shell used.) This example
25998 will execute commands from the file @file{cmds}. All output and errors
25999 would be directed to @file{log}.
26001 Since commands stored on command files tend to be more general than
26002 commands typed interactively, they frequently need to deal with
26003 complicated situations, such as different or unexpected values of
26004 variables and symbols, changes in how the program being debugged is
26005 built, etc. @value{GDBN} provides a set of flow-control commands to
26006 deal with these complexities. Using these commands, you can write
26007 complex scripts that loop over data structures, execute commands
26008 conditionally, etc.
26015 This command allows to include in your script conditionally executed
26016 commands. The @code{if} command takes a single argument, which is an
26017 expression to evaluate. It is followed by a series of commands that
26018 are executed only if the expression is true (its value is nonzero).
26019 There can then optionally be an @code{else} line, followed by a series
26020 of commands that are only executed if the expression was false. The
26021 end of the list is marked by a line containing @code{end}.
26025 This command allows to write loops. Its syntax is similar to
26026 @code{if}: the command takes a single argument, which is an expression
26027 to evaluate, and must be followed by the commands to execute, one per
26028 line, terminated by an @code{end}. These commands are called the
26029 @dfn{body} of the loop. The commands in the body of @code{while} are
26030 executed repeatedly as long as the expression evaluates to true.
26034 This command exits the @code{while} loop in whose body it is included.
26035 Execution of the script continues after that @code{while}s @code{end}
26038 @kindex loop_continue
26039 @item loop_continue
26040 This command skips the execution of the rest of the body of commands
26041 in the @code{while} loop in whose body it is included. Execution
26042 branches to the beginning of the @code{while} loop, where it evaluates
26043 the controlling expression.
26045 @kindex end@r{ (if/else/while commands)}
26047 Terminate the block of commands that are the body of @code{if},
26048 @code{else}, or @code{while} flow-control commands.
26053 @subsection Commands for Controlled Output
26055 During the execution of a command file or a user-defined command, normal
26056 @value{GDBN} output is suppressed; the only output that appears is what is
26057 explicitly printed by the commands in the definition. This section
26058 describes three commands useful for generating exactly the output you
26063 @item echo @var{text}
26064 @c I do not consider backslash-space a standard C escape sequence
26065 @c because it is not in ANSI.
26066 Print @var{text}. Nonprinting characters can be included in
26067 @var{text} using C escape sequences, such as @samp{\n} to print a
26068 newline. @strong{No newline is printed unless you specify one.}
26069 In addition to the standard C escape sequences, a backslash followed
26070 by a space stands for a space. This is useful for displaying a
26071 string with spaces at the beginning or the end, since leading and
26072 trailing spaces are otherwise trimmed from all arguments.
26073 To print @samp{@w{ }and foo =@w{ }}, use the command
26074 @samp{echo \@w{ }and foo = \@w{ }}.
26076 A backslash at the end of @var{text} can be used, as in C, to continue
26077 the command onto subsequent lines. For example,
26080 echo This is some text\n\
26081 which is continued\n\
26082 onto several lines.\n
26085 produces the same output as
26088 echo This is some text\n
26089 echo which is continued\n
26090 echo onto several lines.\n
26094 @item output @var{expression}
26095 Print the value of @var{expression} and nothing but that value: no
26096 newlines, no @samp{$@var{nn} = }. The value is not entered in the
26097 value history either. @xref{Expressions, ,Expressions}, for more information
26100 @item output/@var{fmt} @var{expression}
26101 Print the value of @var{expression} in format @var{fmt}. You can use
26102 the same formats as for @code{print}. @xref{Output Formats,,Output
26103 Formats}, for more information.
26106 @item printf @var{template}, @var{expressions}@dots{}
26107 Print the values of one or more @var{expressions} under the control of
26108 the string @var{template}. To print several values, make
26109 @var{expressions} be a comma-separated list of individual expressions,
26110 which may be either numbers or pointers. Their values are printed as
26111 specified by @var{template}, exactly as a C program would do by
26112 executing the code below:
26115 printf (@var{template}, @var{expressions}@dots{});
26118 As in @code{C} @code{printf}, ordinary characters in @var{template}
26119 are printed verbatim, while @dfn{conversion specification} introduced
26120 by the @samp{%} character cause subsequent @var{expressions} to be
26121 evaluated, their values converted and formatted according to type and
26122 style information encoded in the conversion specifications, and then
26125 For example, you can print two values in hex like this:
26128 printf "foo, bar-foo = 0x%x, 0x%x\n", foo, bar-foo
26131 @code{printf} supports all the standard @code{C} conversion
26132 specifications, including the flags and modifiers between the @samp{%}
26133 character and the conversion letter, with the following exceptions:
26137 The argument-ordering modifiers, such as @samp{2$}, are not supported.
26140 The modifier @samp{*} is not supported for specifying precision or
26144 The @samp{'} flag (for separation of digits into groups according to
26145 @code{LC_NUMERIC'}) is not supported.
26148 The type modifiers @samp{hh}, @samp{j}, @samp{t}, and @samp{z} are not
26152 The conversion letter @samp{n} (as in @samp{%n}) is not supported.
26155 The conversion letters @samp{a} and @samp{A} are not supported.
26159 Note that the @samp{ll} type modifier is supported only if the
26160 underlying @code{C} implementation used to build @value{GDBN} supports
26161 the @code{long long int} type, and the @samp{L} type modifier is
26162 supported only if @code{long double} type is available.
26164 As in @code{C}, @code{printf} supports simple backslash-escape
26165 sequences, such as @code{\n}, @samp{\t}, @samp{\\}, @samp{\"},
26166 @samp{\a}, and @samp{\f}, that consist of backslash followed by a
26167 single character. Octal and hexadecimal escape sequences are not
26170 Additionally, @code{printf} supports conversion specifications for DFP
26171 (@dfn{Decimal Floating Point}) types using the following length modifiers
26172 together with a floating point specifier.
26177 @samp{H} for printing @code{Decimal32} types.
26180 @samp{D} for printing @code{Decimal64} types.
26183 @samp{DD} for printing @code{Decimal128} types.
26186 If the underlying @code{C} implementation used to build @value{GDBN} has
26187 support for the three length modifiers for DFP types, other modifiers
26188 such as width and precision will also be available for @value{GDBN} to use.
26190 In case there is no such @code{C} support, no additional modifiers will be
26191 available and the value will be printed in the standard way.
26193 Here's an example of printing DFP types using the above conversion letters:
26195 printf "D32: %Hf - D64: %Df - D128: %DDf\n",1.2345df,1.2E10dd,1.2E1dl
26200 @item eval @var{template}, @var{expressions}@dots{}
26201 Convert the values of one or more @var{expressions} under the control of
26202 the string @var{template} to a command line, and call it.
26206 @node Auto-loading sequences
26207 @subsection Controlling auto-loading native @value{GDBN} scripts
26208 @cindex native script auto-loading
26210 When a new object file is read (for example, due to the @code{file}
26211 command, or because the inferior has loaded a shared library),
26212 @value{GDBN} will look for the command file @file{@var{objfile}-gdb.gdb}.
26213 @xref{Auto-loading extensions}.
26215 Auto-loading can be enabled or disabled,
26216 and the list of auto-loaded scripts can be printed.
26219 @anchor{set auto-load gdb-scripts}
26220 @kindex set auto-load gdb-scripts
26221 @item set auto-load gdb-scripts [on|off]
26222 Enable or disable the auto-loading of canned sequences of commands scripts.
26224 @anchor{show auto-load gdb-scripts}
26225 @kindex show auto-load gdb-scripts
26226 @item show auto-load gdb-scripts
26227 Show whether auto-loading of canned sequences of commands scripts is enabled or
26230 @anchor{info auto-load gdb-scripts}
26231 @kindex info auto-load gdb-scripts
26232 @cindex print list of auto-loaded canned sequences of commands scripts
26233 @item info auto-load gdb-scripts [@var{regexp}]
26234 Print the list of all canned sequences of commands scripts that @value{GDBN}
26238 If @var{regexp} is supplied only canned sequences of commands scripts with
26239 matching names are printed.
26241 @c Python docs live in a separate file.
26242 @include python.texi
26244 @c Guile docs live in a separate file.
26245 @include guile.texi
26247 @node Auto-loading extensions
26248 @section Auto-loading extensions
26249 @cindex auto-loading extensions
26251 @value{GDBN} provides two mechanisms for automatically loading extensions
26252 when a new object file is read (for example, due to the @code{file}
26253 command, or because the inferior has loaded a shared library):
26254 @file{@var{objfile}-gdb.@var{ext}} and the @code{.debug_gdb_scripts}
26255 section of modern file formats like ELF.
26258 * objfile-gdb.ext file: objfile-gdbdotext file. The @file{@var{objfile}-gdb.@var{ext}} file
26259 * .debug_gdb_scripts section: dotdebug_gdb_scripts section. The @code{.debug_gdb_scripts} section
26260 * Which flavor to choose?::
26263 The auto-loading feature is useful for supplying application-specific
26264 debugging commands and features.
26266 Auto-loading can be enabled or disabled,
26267 and the list of auto-loaded scripts can be printed.
26268 See the @samp{auto-loading} section of each extension language
26269 for more information.
26270 For @value{GDBN} command files see @ref{Auto-loading sequences}.
26271 For Python files see @ref{Python Auto-loading}.
26273 Note that loading of this script file also requires accordingly configured
26274 @code{auto-load safe-path} (@pxref{Auto-loading safe path}).
26276 @node objfile-gdbdotext file
26277 @subsection The @file{@var{objfile}-gdb.@var{ext}} file
26278 @cindex @file{@var{objfile}-gdb.gdb}
26279 @cindex @file{@var{objfile}-gdb.py}
26280 @cindex @file{@var{objfile}-gdb.scm}
26282 When a new object file is read, @value{GDBN} looks for a file named
26283 @file{@var{objfile}-gdb.@var{ext}} (we call it @var{script-name} below),
26284 where @var{objfile} is the object file's name and
26285 where @var{ext} is the file extension for the extension language:
26288 @item @file{@var{objfile}-gdb.gdb}
26289 GDB's own command language
26290 @item @file{@var{objfile}-gdb.py}
26292 @item @file{@var{objfile}-gdb.scm}
26296 @var{script-name} is formed by ensuring that the file name of @var{objfile}
26297 is absolute, following all symlinks, and resolving @code{.} and @code{..}
26298 components, and appending the @file{-gdb.@var{ext}} suffix.
26299 If this file exists and is readable, @value{GDBN} will evaluate it as a
26300 script in the specified extension language.
26302 If this file does not exist, then @value{GDBN} will look for
26303 @var{script-name} file in all of the directories as specified below.
26305 Note that loading of these files requires an accordingly configured
26306 @code{auto-load safe-path} (@pxref{Auto-loading safe path}).
26308 For object files using @file{.exe} suffix @value{GDBN} tries to load first the
26309 scripts normally according to its @file{.exe} filename. But if no scripts are
26310 found @value{GDBN} also tries script filenames matching the object file without
26311 its @file{.exe} suffix. This @file{.exe} stripping is case insensitive and it
26312 is attempted on any platform. This makes the script filenames compatible
26313 between Unix and MS-Windows hosts.
26316 @anchor{set auto-load scripts-directory}
26317 @kindex set auto-load scripts-directory
26318 @item set auto-load scripts-directory @r{[}@var{directories}@r{]}
26319 Control @value{GDBN} auto-loaded scripts location. Multiple directory entries
26320 may be delimited by the host platform path separator in use
26321 (@samp{:} on Unix, @samp{;} on MS-Windows and MS-DOS).
26323 Each entry here needs to be covered also by the security setting
26324 @code{set auto-load safe-path} (@pxref{set auto-load safe-path}).
26326 @anchor{with-auto-load-dir}
26327 This variable defaults to @file{$debugdir:$datadir/auto-load}. The default
26328 @code{set auto-load safe-path} value can be also overriden by @value{GDBN}
26329 configuration option @option{--with-auto-load-dir}.
26331 Any reference to @file{$debugdir} will get replaced by
26332 @var{debug-file-directory} value (@pxref{Separate Debug Files}) and any
26333 reference to @file{$datadir} will get replaced by @var{data-directory} which is
26334 determined at @value{GDBN} startup (@pxref{Data Files}). @file{$debugdir} and
26335 @file{$datadir} must be placed as a directory component --- either alone or
26336 delimited by @file{/} or @file{\} directory separators, depending on the host
26339 The list of directories uses path separator (@samp{:} on GNU and Unix
26340 systems, @samp{;} on MS-Windows and MS-DOS) to separate directories, similarly
26341 to the @env{PATH} environment variable.
26343 @anchor{show auto-load scripts-directory}
26344 @kindex show auto-load scripts-directory
26345 @item show auto-load scripts-directory
26346 Show @value{GDBN} auto-loaded scripts location.
26348 @anchor{add-auto-load-scripts-directory}
26349 @kindex add-auto-load-scripts-directory
26350 @item add-auto-load-scripts-directory @r{[}@var{directories}@dots{}@r{]}
26351 Add an entry (or list of entries) to the list of auto-loaded scripts locations.
26352 Multiple entries may be delimited by the host platform path separator in use.
26355 @value{GDBN} does not track which files it has already auto-loaded this way.
26356 @value{GDBN} will load the associated script every time the corresponding
26357 @var{objfile} is opened.
26358 So your @file{-gdb.@var{ext}} file should be careful to avoid errors if it
26359 is evaluated more than once.
26361 @node dotdebug_gdb_scripts section
26362 @subsection The @code{.debug_gdb_scripts} section
26363 @cindex @code{.debug_gdb_scripts} section
26365 For systems using file formats like ELF and COFF,
26366 when @value{GDBN} loads a new object file
26367 it will look for a special section named @code{.debug_gdb_scripts}.
26368 If this section exists, its contents is a list of null-terminated entries
26369 specifying scripts to load. Each entry begins with a non-null prefix byte that
26370 specifies the kind of entry, typically the extension language and whether the
26371 script is in a file or inlined in @code{.debug_gdb_scripts}.
26373 The following entries are supported:
26376 @item SECTION_SCRIPT_ID_PYTHON_FILE = 1
26377 @item SECTION_SCRIPT_ID_SCHEME_FILE = 3
26378 @item SECTION_SCRIPT_ID_PYTHON_TEXT = 4
26379 @item SECTION_SCRIPT_ID_SCHEME_TEXT = 6
26382 @subsubsection Script File Entries
26384 If the entry specifies a file, @value{GDBN} will look for the file first
26385 in the current directory and then along the source search path
26386 (@pxref{Source Path, ,Specifying Source Directories}),
26387 except that @file{$cdir} is not searched, since the compilation
26388 directory is not relevant to scripts.
26390 File entries can be placed in section @code{.debug_gdb_scripts} with,
26391 for example, this GCC macro for Python scripts.
26394 /* Note: The "MS" section flags are to remove duplicates. */
26395 #define DEFINE_GDB_PY_SCRIPT(script_name) \
26397 .pushsection \".debug_gdb_scripts\", \"MS\",@@progbits,1\n\
26398 .byte 1 /* Python */\n\
26399 .asciz \"" script_name "\"\n\
26405 For Guile scripts, replace @code{.byte 1} with @code{.byte 3}.
26406 Then one can reference the macro in a header or source file like this:
26409 DEFINE_GDB_PY_SCRIPT ("my-app-scripts.py")
26412 The script name may include directories if desired.
26414 Note that loading of this script file also requires accordingly configured
26415 @code{auto-load safe-path} (@pxref{Auto-loading safe path}).
26417 If the macro invocation is put in a header, any application or library
26418 using this header will get a reference to the specified script,
26419 and with the use of @code{"MS"} attributes on the section, the linker
26420 will remove duplicates.
26422 @subsubsection Script Text Entries
26424 Script text entries allow to put the executable script in the entry
26425 itself instead of loading it from a file.
26426 The first line of the entry, everything after the prefix byte and up to
26427 the first newline (@code{0xa}) character, is the script name, and must not
26428 contain any kind of space character, e.g., spaces or tabs.
26429 The rest of the entry, up to the trailing null byte, is the script to
26430 execute in the specified language. The name needs to be unique among
26431 all script names, as @value{GDBN} executes each script only once based
26434 Here is an example from file @file{py-section-script.c} in the @value{GDBN}
26438 #include "symcat.h"
26439 #include "gdb/section-scripts.h"
26441 ".pushsection \".debug_gdb_scripts\", \"MS\",@@progbits,1\n"
26442 ".byte " XSTRING (SECTION_SCRIPT_ID_PYTHON_TEXT) "\n"
26443 ".ascii \"gdb.inlined-script\\n\"\n"
26444 ".ascii \"class test_cmd (gdb.Command):\\n\"\n"
26445 ".ascii \" def __init__ (self):\\n\"\n"
26446 ".ascii \" super (test_cmd, self).__init__ ("
26447 "\\\"test-cmd\\\", gdb.COMMAND_OBSCURE)\\n\"\n"
26448 ".ascii \" def invoke (self, arg, from_tty):\\n\"\n"
26449 ".ascii \" print (\\\"test-cmd output, arg = %s\\\" % arg)\\n\"\n"
26450 ".ascii \"test_cmd ()\\n\"\n"
26456 Loading of inlined scripts requires a properly configured
26457 @code{auto-load safe-path} (@pxref{Auto-loading safe path}).
26458 The path to specify in @code{auto-load safe-path} is the path of the file
26459 containing the @code{.debug_gdb_scripts} section.
26461 @node Which flavor to choose?
26462 @subsection Which flavor to choose?
26464 Given the multiple ways of auto-loading extensions, it might not always
26465 be clear which one to choose. This section provides some guidance.
26468 Benefits of the @file{-gdb.@var{ext}} way:
26472 Can be used with file formats that don't support multiple sections.
26475 Ease of finding scripts for public libraries.
26477 Scripts specified in the @code{.debug_gdb_scripts} section are searched for
26478 in the source search path.
26479 For publicly installed libraries, e.g., @file{libstdc++}, there typically
26480 isn't a source directory in which to find the script.
26483 Doesn't require source code additions.
26487 Benefits of the @code{.debug_gdb_scripts} way:
26491 Works with static linking.
26493 Scripts for libraries done the @file{-gdb.@var{ext}} way require an objfile to
26494 trigger their loading. When an application is statically linked the only
26495 objfile available is the executable, and it is cumbersome to attach all the
26496 scripts from all the input libraries to the executable's
26497 @file{-gdb.@var{ext}} script.
26500 Works with classes that are entirely inlined.
26502 Some classes can be entirely inlined, and thus there may not be an associated
26503 shared library to attach a @file{-gdb.@var{ext}} script to.
26506 Scripts needn't be copied out of the source tree.
26508 In some circumstances, apps can be built out of large collections of internal
26509 libraries, and the build infrastructure necessary to install the
26510 @file{-gdb.@var{ext}} scripts in a place where @value{GDBN} can find them is
26511 cumbersome. It may be easier to specify the scripts in the
26512 @code{.debug_gdb_scripts} section as relative paths, and add a path to the
26513 top of the source tree to the source search path.
26516 @node Multiple Extension Languages
26517 @section Multiple Extension Languages
26519 The Guile and Python extension languages do not share any state,
26520 and generally do not interfere with each other.
26521 There are some things to be aware of, however.
26523 @subsection Python comes first
26525 Python was @value{GDBN}'s first extension language, and to avoid breaking
26526 existing behaviour Python comes first. This is generally solved by the
26527 ``first one wins'' principle. @value{GDBN} maintains a list of enabled
26528 extension languages, and when it makes a call to an extension language,
26529 (say to pretty-print a value), it tries each in turn until an extension
26530 language indicates it has performed the request (e.g., has returned the
26531 pretty-printed form of a value).
26532 This extends to errors while performing such requests: If an error happens
26533 while, for example, trying to pretty-print an object then the error is
26534 reported and any following extension languages are not tried.
26537 @section Creating new spellings of existing commands
26538 @cindex aliases for commands
26540 It is often useful to define alternate spellings of existing commands.
26541 For example, if a new @value{GDBN} command defined in Python has
26542 a long name to type, it is handy to have an abbreviated version of it
26543 that involves less typing.
26545 @value{GDBN} itself uses aliases. For example @samp{s} is an alias
26546 of the @samp{step} command even though it is otherwise an ambiguous
26547 abbreviation of other commands like @samp{set} and @samp{show}.
26549 Aliases are also used to provide shortened or more common versions
26550 of multi-word commands. For example, @value{GDBN} provides the
26551 @samp{tty} alias of the @samp{set inferior-tty} command.
26553 You can define a new alias with the @samp{alias} command.
26558 @item alias [-a] [--] @var{ALIAS} = @var{COMMAND}
26562 @var{ALIAS} specifies the name of the new alias.
26563 Each word of @var{ALIAS} must consist of letters, numbers, dashes and
26566 @var{COMMAND} specifies the name of an existing command
26567 that is being aliased.
26569 The @samp{-a} option specifies that the new alias is an abbreviation
26570 of the command. Abbreviations are not shown in command
26571 lists displayed by the @samp{help} command.
26573 The @samp{--} option specifies the end of options,
26574 and is useful when @var{ALIAS} begins with a dash.
26576 Here is a simple example showing how to make an abbreviation
26577 of a command so that there is less to type.
26578 Suppose you were tired of typing @samp{disas}, the current
26579 shortest unambiguous abbreviation of the @samp{disassemble} command
26580 and you wanted an even shorter version named @samp{di}.
26581 The following will accomplish this.
26584 (gdb) alias -a di = disas
26587 Note that aliases are different from user-defined commands.
26588 With a user-defined command, you also need to write documentation
26589 for it with the @samp{document} command.
26590 An alias automatically picks up the documentation of the existing command.
26592 Here is an example where we make @samp{elms} an abbreviation of
26593 @samp{elements} in the @samp{set print elements} command.
26594 This is to show that you can make an abbreviation of any part
26598 (gdb) alias -a set print elms = set print elements
26599 (gdb) alias -a show print elms = show print elements
26600 (gdb) set p elms 20
26602 Limit on string chars or array elements to print is 200.
26605 Note that if you are defining an alias of a @samp{set} command,
26606 and you want to have an alias for the corresponding @samp{show}
26607 command, then you need to define the latter separately.
26609 Unambiguously abbreviated commands are allowed in @var{COMMAND} and
26610 @var{ALIAS}, just as they are normally.
26613 (gdb) alias -a set pr elms = set p ele
26616 Finally, here is an example showing the creation of a one word
26617 alias for a more complex command.
26618 This creates alias @samp{spe} of the command @samp{set print elements}.
26621 (gdb) alias spe = set print elements
26626 @chapter Command Interpreters
26627 @cindex command interpreters
26629 @value{GDBN} supports multiple command interpreters, and some command
26630 infrastructure to allow users or user interface writers to switch
26631 between interpreters or run commands in other interpreters.
26633 @value{GDBN} currently supports two command interpreters, the console
26634 interpreter (sometimes called the command-line interpreter or @sc{cli})
26635 and the machine interface interpreter (or @sc{gdb/mi}). This manual
26636 describes both of these interfaces in great detail.
26638 By default, @value{GDBN} will start with the console interpreter.
26639 However, the user may choose to start @value{GDBN} with another
26640 interpreter by specifying the @option{-i} or @option{--interpreter}
26641 startup options. Defined interpreters include:
26645 @cindex console interpreter
26646 The traditional console or command-line interpreter. This is the most often
26647 used interpreter with @value{GDBN}. With no interpreter specified at runtime,
26648 @value{GDBN} will use this interpreter.
26651 @cindex mi interpreter
26652 The newest @sc{gdb/mi} interface (currently @code{mi3}). Used primarily
26653 by programs wishing to use @value{GDBN} as a backend for a debugger GUI
26654 or an IDE. For more information, see @ref{GDB/MI, ,The @sc{gdb/mi}
26658 @cindex mi3 interpreter
26659 The @sc{gdb/mi} interface introduced in @value{GDBN} 9.1.
26662 @cindex mi2 interpreter
26663 The @sc{gdb/mi} interface introduced in @value{GDBN} 6.0.
26666 @cindex mi1 interpreter
26667 The @sc{gdb/mi} interface introduced in @value{GDBN} 5.1.
26671 @cindex invoke another interpreter
26673 @kindex interpreter-exec
26674 You may execute commands in any interpreter from the current
26675 interpreter using the appropriate command. If you are running the
26676 console interpreter, simply use the @code{interpreter-exec} command:
26679 interpreter-exec mi "-data-list-register-names"
26682 @sc{gdb/mi} has a similar command, although it is only available in versions of
26683 @value{GDBN} which support @sc{gdb/mi} version 2 (or greater).
26685 Note that @code{interpreter-exec} only changes the interpreter for the
26686 duration of the specified command. It does not change the interpreter
26689 @cindex start a new independent interpreter
26691 Although you may only choose a single interpreter at startup, it is
26692 possible to run an independent interpreter on a specified input/output
26693 device (usually a tty).
26695 For example, consider a debugger GUI or IDE that wants to provide a
26696 @value{GDBN} console view. It may do so by embedding a terminal
26697 emulator widget in its GUI, starting @value{GDBN} in the traditional
26698 command-line mode with stdin/stdout/stderr redirected to that
26699 terminal, and then creating an MI interpreter running on a specified
26700 input/output device. The console interpreter created by @value{GDBN}
26701 at startup handles commands the user types in the terminal widget,
26702 while the GUI controls and synchronizes state with @value{GDBN} using
26703 the separate MI interpreter.
26705 To start a new secondary @dfn{user interface} running MI, use the
26706 @code{new-ui} command:
26709 @cindex new user interface
26711 new-ui @var{interpreter} @var{tty}
26714 The @var{interpreter} parameter specifies the interpreter to run.
26715 This accepts the same values as the @code{interpreter-exec} command.
26716 For example, @samp{console}, @samp{mi}, @samp{mi2}, etc. The
26717 @var{tty} parameter specifies the name of the bidirectional file the
26718 interpreter uses for input/output, usually the name of a
26719 pseudoterminal slave on Unix systems. For example:
26722 (@value{GDBP}) new-ui mi /dev/pts/9
26726 runs an MI interpreter on @file{/dev/pts/9}.
26729 @chapter @value{GDBN} Text User Interface
26731 @cindex Text User Interface
26734 * TUI Overview:: TUI overview
26735 * TUI Keys:: TUI key bindings
26736 * TUI Single Key Mode:: TUI single key mode
26737 * TUI Commands:: TUI-specific commands
26738 * TUI Configuration:: TUI configuration variables
26741 The @value{GDBN} Text User Interface (TUI) is a terminal
26742 interface which uses the @code{curses} library to show the source
26743 file, the assembly output, the program registers and @value{GDBN}
26744 commands in separate text windows. The TUI mode is supported only
26745 on platforms where a suitable version of the @code{curses} library
26748 The TUI mode is enabled by default when you invoke @value{GDBN} as
26749 @samp{@value{GDBP} -tui}.
26750 You can also switch in and out of TUI mode while @value{GDBN} runs by
26751 using various TUI commands and key bindings, such as @command{tui
26752 enable} or @kbd{C-x C-a}. @xref{TUI Commands, ,TUI Commands}, and
26753 @ref{TUI Keys, ,TUI Key Bindings}.
26756 @section TUI Overview
26758 In TUI mode, @value{GDBN} can display several text windows:
26762 This window is the @value{GDBN} command window with the @value{GDBN}
26763 prompt and the @value{GDBN} output. The @value{GDBN} input is still
26764 managed using readline.
26767 The source window shows the source file of the program. The current
26768 line and active breakpoints are displayed in this window.
26771 The assembly window shows the disassembly output of the program.
26774 This window shows the processor registers. Registers are highlighted
26775 when their values change.
26778 The source and assembly windows show the current program position
26779 by highlighting the current line and marking it with a @samp{>} marker.
26780 Breakpoints are indicated with two markers. The first marker
26781 indicates the breakpoint type:
26785 Breakpoint which was hit at least once.
26788 Breakpoint which was never hit.
26791 Hardware breakpoint which was hit at least once.
26794 Hardware breakpoint which was never hit.
26797 The second marker indicates whether the breakpoint is enabled or not:
26801 Breakpoint is enabled.
26804 Breakpoint is disabled.
26807 The source, assembly and register windows are updated when the current
26808 thread changes, when the frame changes, or when the program counter
26811 These windows are not all visible at the same time. The command
26812 window is always visible. The others can be arranged in several
26823 source and assembly,
26826 source and registers, or
26829 assembly and registers.
26832 A status line above the command window shows the following information:
26836 Indicates the current @value{GDBN} target.
26837 (@pxref{Targets, ,Specifying a Debugging Target}).
26840 Gives the current process or thread number.
26841 When no process is being debugged, this field is set to @code{No process}.
26844 Gives the current function name for the selected frame.
26845 The name is demangled if demangling is turned on (@pxref{Print Settings}).
26846 When there is no symbol corresponding to the current program counter,
26847 the string @code{??} is displayed.
26850 Indicates the current line number for the selected frame.
26851 When the current line number is not known, the string @code{??} is displayed.
26854 Indicates the current program counter address.
26858 @section TUI Key Bindings
26859 @cindex TUI key bindings
26861 The TUI installs several key bindings in the readline keymaps
26862 @ifset SYSTEM_READLINE
26863 (@pxref{Command Line Editing, , , rluserman, GNU Readline Library}).
26865 @ifclear SYSTEM_READLINE
26866 (@pxref{Command Line Editing}).
26868 The following key bindings are installed for both TUI mode and the
26869 @value{GDBN} standard mode.
26878 Enter or leave the TUI mode. When leaving the TUI mode,
26879 the curses window management stops and @value{GDBN} operates using
26880 its standard mode, writing on the terminal directly. When reentering
26881 the TUI mode, control is given back to the curses windows.
26882 The screen is then refreshed.
26886 Use a TUI layout with only one window. The layout will
26887 either be @samp{source} or @samp{assembly}. When the TUI mode
26888 is not active, it will switch to the TUI mode.
26890 Think of this key binding as the Emacs @kbd{C-x 1} binding.
26894 Use a TUI layout with at least two windows. When the current
26895 layout already has two windows, the next layout with two windows is used.
26896 When a new layout is chosen, one window will always be common to the
26897 previous layout and the new one.
26899 Think of it as the Emacs @kbd{C-x 2} binding.
26903 Change the active window. The TUI associates several key bindings
26904 (like scrolling and arrow keys) with the active window. This command
26905 gives the focus to the next TUI window.
26907 Think of it as the Emacs @kbd{C-x o} binding.
26911 Switch in and out of the TUI SingleKey mode that binds single
26912 keys to @value{GDBN} commands (@pxref{TUI Single Key Mode}).
26915 The following key bindings only work in the TUI mode:
26920 Scroll the active window one page up.
26924 Scroll the active window one page down.
26928 Scroll the active window one line up.
26932 Scroll the active window one line down.
26936 Scroll the active window one column left.
26940 Scroll the active window one column right.
26944 Refresh the screen.
26947 Because the arrow keys scroll the active window in the TUI mode, they
26948 are not available for their normal use by readline unless the command
26949 window has the focus. When another window is active, you must use
26950 other readline key bindings such as @kbd{C-p}, @kbd{C-n}, @kbd{C-b}
26951 and @kbd{C-f} to control the command window.
26953 @node TUI Single Key Mode
26954 @section TUI Single Key Mode
26955 @cindex TUI single key mode
26957 The TUI also provides a @dfn{SingleKey} mode, which binds several
26958 frequently used @value{GDBN} commands to single keys. Type @kbd{C-x s} to
26959 switch into this mode, where the following key bindings are used:
26962 @kindex c @r{(SingleKey TUI key)}
26966 @kindex d @r{(SingleKey TUI key)}
26970 @kindex f @r{(SingleKey TUI key)}
26974 @kindex n @r{(SingleKey TUI key)}
26978 @kindex o @r{(SingleKey TUI key)}
26980 nexti. The shortcut letter @samp{o} stands for ``step Over''.
26982 @kindex q @r{(SingleKey TUI key)}
26984 exit the SingleKey mode.
26986 @kindex r @r{(SingleKey TUI key)}
26990 @kindex s @r{(SingleKey TUI key)}
26994 @kindex i @r{(SingleKey TUI key)}
26996 stepi. The shortcut letter @samp{i} stands for ``step Into''.
26998 @kindex u @r{(SingleKey TUI key)}
27002 @kindex v @r{(SingleKey TUI key)}
27006 @kindex w @r{(SingleKey TUI key)}
27011 Other keys temporarily switch to the @value{GDBN} command prompt.
27012 The key that was pressed is inserted in the editing buffer so that
27013 it is possible to type most @value{GDBN} commands without interaction
27014 with the TUI SingleKey mode. Once the command is entered the TUI
27015 SingleKey mode is restored. The only way to permanently leave
27016 this mode is by typing @kbd{q} or @kbd{C-x s}.
27020 @section TUI-specific Commands
27021 @cindex TUI commands
27023 The TUI has specific commands to control the text windows.
27024 These commands are always available, even when @value{GDBN} is not in
27025 the TUI mode. When @value{GDBN} is in the standard mode, most
27026 of these commands will automatically switch to the TUI mode.
27028 Note that if @value{GDBN}'s @code{stdout} is not connected to a
27029 terminal, or @value{GDBN} has been started with the machine interface
27030 interpreter (@pxref{GDB/MI, ,The @sc{gdb/mi} Interface}), most of
27031 these commands will fail with an error, because it would not be
27032 possible or desirable to enable curses window management.
27037 Activate TUI mode. The last active TUI window layout will be used if
27038 TUI mode has prevsiouly been used in the current debugging session,
27039 otherwise a default layout is used.
27042 @kindex tui disable
27043 Disable TUI mode, returning to the console interpreter.
27047 List and give the size of all displayed windows.
27049 @item layout @var{name}
27051 Changes which TUI windows are displayed. In each layout the command
27052 window is always displayed, the @var{name} parameter controls which
27053 additional windows are displayed, and can be any of the following:
27057 Display the next layout.
27060 Display the previous layout.
27063 Display the source and command windows.
27066 Display the assembly and command windows.
27069 Display the source, assembly, and command windows.
27072 When in @code{src} layout display the register, source, and command
27073 windows. When in @code{asm} or @code{split} layout display the
27074 register, assembler, and command windows.
27077 @item focus @var{name}
27079 Changes which TUI window is currently active for scrolling. The
27080 @var{name} parameter can be any of the following:
27084 Make the next window active for scrolling.
27087 Make the previous window active for scrolling.
27090 Make the source window active for scrolling.
27093 Make the assembly window active for scrolling.
27096 Make the register window active for scrolling.
27099 Make the command window active for scrolling.
27104 Refresh the screen. This is similar to typing @kbd{C-L}.
27106 @item tui reg @var{group}
27108 Changes the register group displayed in the tui register window to
27109 @var{group}. If the register window is not currently displayed this
27110 command will cause the register window to be displayed. The list of
27111 register groups, as well as their order is target specific. The
27112 following groups are available on most targets:
27115 Repeatedly selecting this group will cause the display to cycle
27116 through all of the available register groups.
27119 Repeatedly selecting this group will cause the display to cycle
27120 through all of the available register groups in the reverse order to
27124 Display the general registers.
27126 Display the floating point registers.
27128 Display the system registers.
27130 Display the vector registers.
27132 Display all registers.
27137 Update the source window and the current execution point.
27139 @item winheight @var{name} +@var{count}
27140 @itemx winheight @var{name} -@var{count}
27142 Change the height of the window @var{name} by @var{count}
27143 lines. Positive counts increase the height, while negative counts
27144 decrease it. The @var{name} parameter can be one of @code{src} (the
27145 source window), @code{cmd} (the command window), @code{asm} (the
27146 disassembly window), or @code{regs} (the register display window).
27149 @node TUI Configuration
27150 @section TUI Configuration Variables
27151 @cindex TUI configuration variables
27153 Several configuration variables control the appearance of TUI windows.
27156 @item set tui border-kind @var{kind}
27157 @kindex set tui border-kind
27158 Select the border appearance for the source, assembly and register windows.
27159 The possible values are the following:
27162 Use a space character to draw the border.
27165 Use @sc{ascii} characters @samp{+}, @samp{-} and @samp{|} to draw the border.
27168 Use the Alternate Character Set to draw the border. The border is
27169 drawn using character line graphics if the terminal supports them.
27172 @item set tui border-mode @var{mode}
27173 @kindex set tui border-mode
27174 @itemx set tui active-border-mode @var{mode}
27175 @kindex set tui active-border-mode
27176 Select the display attributes for the borders of the inactive windows
27177 or the active window. The @var{mode} can be one of the following:
27180 Use normal attributes to display the border.
27186 Use reverse video mode.
27189 Use half bright mode.
27191 @item half-standout
27192 Use half bright and standout mode.
27195 Use extra bright or bold mode.
27197 @item bold-standout
27198 Use extra bright or bold and standout mode.
27201 @item set tui tab-width @var{nchars}
27202 @kindex set tui tab-width
27204 Set the width of tab stops to be @var{nchars} characters. This
27205 setting affects the display of TAB characters in the source and
27210 @chapter Using @value{GDBN} under @sc{gnu} Emacs
27213 @cindex @sc{gnu} Emacs
27214 A special interface allows you to use @sc{gnu} Emacs to view (and
27215 edit) the source files for the program you are debugging with
27218 To use this interface, use the command @kbd{M-x gdb} in Emacs. Give the
27219 executable file you want to debug as an argument. This command starts
27220 @value{GDBN} as a subprocess of Emacs, with input and output through a newly
27221 created Emacs buffer.
27222 @c (Do not use the @code{-tui} option to run @value{GDBN} from Emacs.)
27224 Running @value{GDBN} under Emacs can be just like running @value{GDBN} normally except for two
27229 All ``terminal'' input and output goes through an Emacs buffer, called
27232 This applies both to @value{GDBN} commands and their output, and to the input
27233 and output done by the program you are debugging.
27235 This is useful because it means that you can copy the text of previous
27236 commands and input them again; you can even use parts of the output
27239 All the facilities of Emacs' Shell mode are available for interacting
27240 with your program. In particular, you can send signals the usual
27241 way---for example, @kbd{C-c C-c} for an interrupt, @kbd{C-c C-z} for a
27245 @value{GDBN} displays source code through Emacs.
27247 Each time @value{GDBN} displays a stack frame, Emacs automatically finds the
27248 source file for that frame and puts an arrow (@samp{=>}) at the
27249 left margin of the current line. Emacs uses a separate buffer for
27250 source display, and splits the screen to show both your @value{GDBN} session
27253 Explicit @value{GDBN} @code{list} or search commands still produce output as
27254 usual, but you probably have no reason to use them from Emacs.
27257 We call this @dfn{text command mode}. Emacs 22.1, and later, also uses
27258 a graphical mode, enabled by default, which provides further buffers
27259 that can control the execution and describe the state of your program.
27260 @xref{GDB Graphical Interface,,, Emacs, The @sc{gnu} Emacs Manual}.
27262 If you specify an absolute file name when prompted for the @kbd{M-x
27263 gdb} argument, then Emacs sets your current working directory to where
27264 your program resides. If you only specify the file name, then Emacs
27265 sets your current working directory to the directory associated
27266 with the previous buffer. In this case, @value{GDBN} may find your
27267 program by searching your environment's @code{PATH} variable, but on
27268 some operating systems it might not find the source. So, although the
27269 @value{GDBN} input and output session proceeds normally, the auxiliary
27270 buffer does not display the current source and line of execution.
27272 The initial working directory of @value{GDBN} is printed on the top
27273 line of the GUD buffer and this serves as a default for the commands
27274 that specify files for @value{GDBN} to operate on. @xref{Files,
27275 ,Commands to Specify Files}.
27277 By default, @kbd{M-x gdb} calls the program called @file{gdb}. If you
27278 need to call @value{GDBN} by a different name (for example, if you
27279 keep several configurations around, with different names) you can
27280 customize the Emacs variable @code{gud-gdb-command-name} to run the
27283 In the GUD buffer, you can use these special Emacs commands in
27284 addition to the standard Shell mode commands:
27288 Describe the features of Emacs' GUD Mode.
27291 Execute to another source line, like the @value{GDBN} @code{step} command; also
27292 update the display window to show the current file and location.
27295 Execute to next source line in this function, skipping all function
27296 calls, like the @value{GDBN} @code{next} command. Then update the display window
27297 to show the current file and location.
27300 Execute one instruction, like the @value{GDBN} @code{stepi} command; update
27301 display window accordingly.
27304 Execute until exit from the selected stack frame, like the @value{GDBN}
27305 @code{finish} command.
27308 Continue execution of your program, like the @value{GDBN} @code{continue}
27312 Go up the number of frames indicated by the numeric argument
27313 (@pxref{Arguments, , Numeric Arguments, Emacs, The @sc{gnu} Emacs Manual}),
27314 like the @value{GDBN} @code{up} command.
27317 Go down the number of frames indicated by the numeric argument, like the
27318 @value{GDBN} @code{down} command.
27321 In any source file, the Emacs command @kbd{C-x @key{SPC}} (@code{gud-break})
27322 tells @value{GDBN} to set a breakpoint on the source line point is on.
27324 In text command mode, if you type @kbd{M-x speedbar}, Emacs displays a
27325 separate frame which shows a backtrace when the GUD buffer is current.
27326 Move point to any frame in the stack and type @key{RET} to make it
27327 become the current frame and display the associated source in the
27328 source buffer. Alternatively, click @kbd{Mouse-2} to make the
27329 selected frame become the current one. In graphical mode, the
27330 speedbar displays watch expressions.
27332 If you accidentally delete the source-display buffer, an easy way to get
27333 it back is to type the command @code{f} in the @value{GDBN} buffer, to
27334 request a frame display; when you run under Emacs, this recreates
27335 the source buffer if necessary to show you the context of the current
27338 The source files displayed in Emacs are in ordinary Emacs buffers
27339 which are visiting the source files in the usual way. You can edit
27340 the files with these buffers if you wish; but keep in mind that @value{GDBN}
27341 communicates with Emacs in terms of line numbers. If you add or
27342 delete lines from the text, the line numbers that @value{GDBN} knows cease
27343 to correspond properly with the code.
27345 A more detailed description of Emacs' interaction with @value{GDBN} is
27346 given in the Emacs manual (@pxref{Debuggers,,, Emacs, The @sc{gnu}
27350 @chapter The @sc{gdb/mi} Interface
27352 @unnumberedsec Function and Purpose
27354 @cindex @sc{gdb/mi}, its purpose
27355 @sc{gdb/mi} is a line based machine oriented text interface to
27356 @value{GDBN} and is activated by specifying using the
27357 @option{--interpreter} command line option (@pxref{Mode Options}). It
27358 is specifically intended to support the development of systems which
27359 use the debugger as just one small component of a larger system.
27361 This chapter is a specification of the @sc{gdb/mi} interface. It is written
27362 in the form of a reference manual.
27364 Note that @sc{gdb/mi} is still under construction, so some of the
27365 features described below are incomplete and subject to change
27366 (@pxref{GDB/MI Development and Front Ends, , @sc{gdb/mi} Development and Front Ends}).
27368 @unnumberedsec Notation and Terminology
27370 @cindex notational conventions, for @sc{gdb/mi}
27371 This chapter uses the following notation:
27375 @code{|} separates two alternatives.
27378 @code{[ @var{something} ]} indicates that @var{something} is optional:
27379 it may or may not be given.
27382 @code{( @var{group} )*} means that @var{group} inside the parentheses
27383 may repeat zero or more times.
27386 @code{( @var{group} )+} means that @var{group} inside the parentheses
27387 may repeat one or more times.
27390 @code{"@var{string}"} means a literal @var{string}.
27394 @heading Dependencies
27398 * GDB/MI General Design::
27399 * GDB/MI Command Syntax::
27400 * GDB/MI Compatibility with CLI::
27401 * GDB/MI Development and Front Ends::
27402 * GDB/MI Output Records::
27403 * GDB/MI Simple Examples::
27404 * GDB/MI Command Description Format::
27405 * GDB/MI Breakpoint Commands::
27406 * GDB/MI Catchpoint Commands::
27407 * GDB/MI Program Context::
27408 * GDB/MI Thread Commands::
27409 * GDB/MI Ada Tasking Commands::
27410 * GDB/MI Program Execution::
27411 * GDB/MI Stack Manipulation::
27412 * GDB/MI Variable Objects::
27413 * GDB/MI Data Manipulation::
27414 * GDB/MI Tracepoint Commands::
27415 * GDB/MI Symbol Query::
27416 * GDB/MI File Commands::
27418 * GDB/MI Kod Commands::
27419 * GDB/MI Memory Overlay Commands::
27420 * GDB/MI Signal Handling Commands::
27422 * GDB/MI Target Manipulation::
27423 * GDB/MI File Transfer Commands::
27424 * GDB/MI Ada Exceptions Commands::
27425 * GDB/MI Support Commands::
27426 * GDB/MI Miscellaneous Commands::
27429 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27430 @node GDB/MI General Design
27431 @section @sc{gdb/mi} General Design
27432 @cindex GDB/MI General Design
27434 Interaction of a @sc{GDB/MI} frontend with @value{GDBN} involves three
27435 parts---commands sent to @value{GDBN}, responses to those commands
27436 and notifications. Each command results in exactly one response,
27437 indicating either successful completion of the command, or an error.
27438 For the commands that do not resume the target, the response contains the
27439 requested information. For the commands that resume the target, the
27440 response only indicates whether the target was successfully resumed.
27441 Notifications is the mechanism for reporting changes in the state of the
27442 target, or in @value{GDBN} state, that cannot conveniently be associated with
27443 a command and reported as part of that command response.
27445 The important examples of notifications are:
27449 Exec notifications. These are used to report changes in
27450 target state---when a target is resumed, or stopped. It would not
27451 be feasible to include this information in response of resuming
27452 commands, because one resume commands can result in multiple events in
27453 different threads. Also, quite some time may pass before any event
27454 happens in the target, while a frontend needs to know whether the resuming
27455 command itself was successfully executed.
27458 Console output, and status notifications. Console output
27459 notifications are used to report output of CLI commands, as well as
27460 diagnostics for other commands. Status notifications are used to
27461 report the progress of a long-running operation. Naturally, including
27462 this information in command response would mean no output is produced
27463 until the command is finished, which is undesirable.
27466 General notifications. Commands may have various side effects on
27467 the @value{GDBN} or target state beyond their official purpose. For example,
27468 a command may change the selected thread. Although such changes can
27469 be included in command response, using notification allows for more
27470 orthogonal frontend design.
27474 There's no guarantee that whenever an MI command reports an error,
27475 @value{GDBN} or the target are in any specific state, and especially,
27476 the state is not reverted to the state before the MI command was
27477 processed. Therefore, whenever an MI command results in an error,
27478 we recommend that the frontend refreshes all the information shown in
27479 the user interface.
27483 * Context management::
27484 * Asynchronous and non-stop modes::
27488 @node Context management
27489 @subsection Context management
27491 @subsubsection Threads and Frames
27493 In most cases when @value{GDBN} accesses the target, this access is
27494 done in context of a specific thread and frame (@pxref{Frames}).
27495 Often, even when accessing global data, the target requires that a thread
27496 be specified. The CLI interface maintains the selected thread and frame,
27497 and supplies them to target on each command. This is convenient,
27498 because a command line user would not want to specify that information
27499 explicitly on each command, and because user interacts with
27500 @value{GDBN} via a single terminal, so no confusion is possible as
27501 to what thread and frame are the current ones.
27503 In the case of MI, the concept of selected thread and frame is less
27504 useful. First, a frontend can easily remember this information
27505 itself. Second, a graphical frontend can have more than one window,
27506 each one used for debugging a different thread, and the frontend might
27507 want to access additional threads for internal purposes. This
27508 increases the risk that by relying on implicitly selected thread, the
27509 frontend may be operating on a wrong one. Therefore, each MI command
27510 should explicitly specify which thread and frame to operate on. To
27511 make it possible, each MI command accepts the @samp{--thread} and
27512 @samp{--frame} options, the value to each is @value{GDBN} global
27513 identifier for thread and frame to operate on.
27515 Usually, each top-level window in a frontend allows the user to select
27516 a thread and a frame, and remembers the user selection for further
27517 operations. However, in some cases @value{GDBN} may suggest that the
27518 current thread or frame be changed. For example, when stopping on a
27519 breakpoint it is reasonable to switch to the thread where breakpoint is
27520 hit. For another example, if the user issues the CLI @samp{thread} or
27521 @samp{frame} commands via the frontend, it is desirable to change the
27522 frontend's selection to the one specified by user. @value{GDBN}
27523 communicates the suggestion to change current thread and frame using the
27524 @samp{=thread-selected} notification.
27526 Note that historically, MI shares the selected thread with CLI, so
27527 frontends used the @code{-thread-select} to execute commands in the
27528 right context. However, getting this to work right is cumbersome. The
27529 simplest way is for frontend to emit @code{-thread-select} command
27530 before every command. This doubles the number of commands that need
27531 to be sent. The alternative approach is to suppress @code{-thread-select}
27532 if the selected thread in @value{GDBN} is supposed to be identical to the
27533 thread the frontend wants to operate on. However, getting this
27534 optimization right can be tricky. In particular, if the frontend
27535 sends several commands to @value{GDBN}, and one of the commands changes the
27536 selected thread, then the behaviour of subsequent commands will
27537 change. So, a frontend should either wait for response from such
27538 problematic commands, or explicitly add @code{-thread-select} for
27539 all subsequent commands. No frontend is known to do this exactly
27540 right, so it is suggested to just always pass the @samp{--thread} and
27541 @samp{--frame} options.
27543 @subsubsection Language
27545 The execution of several commands depends on which language is selected.
27546 By default, the current language (@pxref{show language}) is used.
27547 But for commands known to be language-sensitive, it is recommended
27548 to use the @samp{--language} option. This option takes one argument,
27549 which is the name of the language to use while executing the command.
27553 -data-evaluate-expression --language c "sizeof (void*)"
27558 The valid language names are the same names accepted by the
27559 @samp{set language} command (@pxref{Manually}), excluding @samp{auto},
27560 @samp{local} or @samp{unknown}.
27562 @node Asynchronous and non-stop modes
27563 @subsection Asynchronous command execution and non-stop mode
27565 On some targets, @value{GDBN} is capable of processing MI commands
27566 even while the target is running. This is called @dfn{asynchronous
27567 command execution} (@pxref{Background Execution}). The frontend may
27568 specify a preferrence for asynchronous execution using the
27569 @code{-gdb-set mi-async 1} command, which should be emitted before
27570 either running the executable or attaching to the target. After the
27571 frontend has started the executable or attached to the target, it can
27572 find if asynchronous execution is enabled using the
27573 @code{-list-target-features} command.
27576 @item -gdb-set mi-async on
27577 @item -gdb-set mi-async off
27578 Set whether MI is in asynchronous mode.
27580 When @code{off}, which is the default, MI execution commands (e.g.,
27581 @code{-exec-continue}) are foreground commands, and @value{GDBN} waits
27582 for the program to stop before processing further commands.
27584 When @code{on}, MI execution commands are background execution
27585 commands (e.g., @code{-exec-continue} becomes the equivalent of the
27586 @code{c&} CLI command), and so @value{GDBN} is capable of processing
27587 MI commands even while the target is running.
27589 @item -gdb-show mi-async
27590 Show whether MI asynchronous mode is enabled.
27593 Note: In @value{GDBN} version 7.7 and earlier, this option was called
27594 @code{target-async} instead of @code{mi-async}, and it had the effect
27595 of both putting MI in asynchronous mode and making CLI background
27596 commands possible. CLI background commands are now always possible
27597 ``out of the box'' if the target supports them. The old spelling is
27598 kept as a deprecated alias for backwards compatibility.
27600 Even if @value{GDBN} can accept a command while target is running,
27601 many commands that access the target do not work when the target is
27602 running. Therefore, asynchronous command execution is most useful
27603 when combined with non-stop mode (@pxref{Non-Stop Mode}). Then,
27604 it is possible to examine the state of one thread, while other threads
27607 When a given thread is running, MI commands that try to access the
27608 target in the context of that thread may not work, or may work only on
27609 some targets. In particular, commands that try to operate on thread's
27610 stack will not work, on any target. Commands that read memory, or
27611 modify breakpoints, may work or not work, depending on the target. Note
27612 that even commands that operate on global state, such as @code{print},
27613 @code{set}, and breakpoint commands, still access the target in the
27614 context of a specific thread, so frontend should try to find a
27615 stopped thread and perform the operation on that thread (using the
27616 @samp{--thread} option).
27618 Which commands will work in the context of a running thread is
27619 highly target dependent. However, the two commands
27620 @code{-exec-interrupt}, to stop a thread, and @code{-thread-info},
27621 to find the state of a thread, will always work.
27623 @node Thread groups
27624 @subsection Thread groups
27625 @value{GDBN} may be used to debug several processes at the same time.
27626 On some platfroms, @value{GDBN} may support debugging of several
27627 hardware systems, each one having several cores with several different
27628 processes running on each core. This section describes the MI
27629 mechanism to support such debugging scenarios.
27631 The key observation is that regardless of the structure of the
27632 target, MI can have a global list of threads, because most commands that
27633 accept the @samp{--thread} option do not need to know what process that
27634 thread belongs to. Therefore, it is not necessary to introduce
27635 neither additional @samp{--process} option, nor an notion of the
27636 current process in the MI interface. The only strictly new feature
27637 that is required is the ability to find how the threads are grouped
27640 To allow the user to discover such grouping, and to support arbitrary
27641 hierarchy of machines/cores/processes, MI introduces the concept of a
27642 @dfn{thread group}. Thread group is a collection of threads and other
27643 thread groups. A thread group always has a string identifier, a type,
27644 and may have additional attributes specific to the type. A new
27645 command, @code{-list-thread-groups}, returns the list of top-level
27646 thread groups, which correspond to processes that @value{GDBN} is
27647 debugging at the moment. By passing an identifier of a thread group
27648 to the @code{-list-thread-groups} command, it is possible to obtain
27649 the members of specific thread group.
27651 To allow the user to easily discover processes, and other objects, he
27652 wishes to debug, a concept of @dfn{available thread group} is
27653 introduced. Available thread group is an thread group that
27654 @value{GDBN} is not debugging, but that can be attached to, using the
27655 @code{-target-attach} command. The list of available top-level thread
27656 groups can be obtained using @samp{-list-thread-groups --available}.
27657 In general, the content of a thread group may be only retrieved only
27658 after attaching to that thread group.
27660 Thread groups are related to inferiors (@pxref{Inferiors and
27661 Programs}). Each inferior corresponds to a thread group of a special
27662 type @samp{process}, and some additional operations are permitted on
27663 such thread groups.
27665 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27666 @node GDB/MI Command Syntax
27667 @section @sc{gdb/mi} Command Syntax
27670 * GDB/MI Input Syntax::
27671 * GDB/MI Output Syntax::
27674 @node GDB/MI Input Syntax
27675 @subsection @sc{gdb/mi} Input Syntax
27677 @cindex input syntax for @sc{gdb/mi}
27678 @cindex @sc{gdb/mi}, input syntax
27680 @item @var{command} @expansion{}
27681 @code{@var{cli-command} | @var{mi-command}}
27683 @item @var{cli-command} @expansion{}
27684 @code{[ @var{token} ] @var{cli-command} @var{nl}}, where
27685 @var{cli-command} is any existing @value{GDBN} CLI command.
27687 @item @var{mi-command} @expansion{}
27688 @code{[ @var{token} ] "-" @var{operation} ( " " @var{option} )*
27689 @code{[} " --" @code{]} ( " " @var{parameter} )* @var{nl}}
27691 @item @var{token} @expansion{}
27692 "any sequence of digits"
27694 @item @var{option} @expansion{}
27695 @code{"-" @var{parameter} [ " " @var{parameter} ]}
27697 @item @var{parameter} @expansion{}
27698 @code{@var{non-blank-sequence} | @var{c-string}}
27700 @item @var{operation} @expansion{}
27701 @emph{any of the operations described in this chapter}
27703 @item @var{non-blank-sequence} @expansion{}
27704 @emph{anything, provided it doesn't contain special characters such as
27705 "-", @var{nl}, """ and of course " "}
27707 @item @var{c-string} @expansion{}
27708 @code{""" @var{seven-bit-iso-c-string-content} """}
27710 @item @var{nl} @expansion{}
27719 The CLI commands are still handled by the @sc{mi} interpreter; their
27720 output is described below.
27723 The @code{@var{token}}, when present, is passed back when the command
27727 Some @sc{mi} commands accept optional arguments as part of the parameter
27728 list. Each option is identified by a leading @samp{-} (dash) and may be
27729 followed by an optional argument parameter. Options occur first in the
27730 parameter list and can be delimited from normal parameters using
27731 @samp{--} (this is useful when some parameters begin with a dash).
27738 We want easy access to the existing CLI syntax (for debugging).
27741 We want it to be easy to spot a @sc{mi} operation.
27744 @node GDB/MI Output Syntax
27745 @subsection @sc{gdb/mi} Output Syntax
27747 @cindex output syntax of @sc{gdb/mi}
27748 @cindex @sc{gdb/mi}, output syntax
27749 The output from @sc{gdb/mi} consists of zero or more out-of-band records
27750 followed, optionally, by a single result record. This result record
27751 is for the most recent command. The sequence of output records is
27752 terminated by @samp{(gdb)}.
27754 If an input command was prefixed with a @code{@var{token}} then the
27755 corresponding output for that command will also be prefixed by that same
27759 @item @var{output} @expansion{}
27760 @code{( @var{out-of-band-record} )* [ @var{result-record} ] "(gdb)" @var{nl}}
27762 @item @var{result-record} @expansion{}
27763 @code{ [ @var{token} ] "^" @var{result-class} ( "," @var{result} )* @var{nl}}
27765 @item @var{out-of-band-record} @expansion{}
27766 @code{@var{async-record} | @var{stream-record}}
27768 @item @var{async-record} @expansion{}
27769 @code{@var{exec-async-output} | @var{status-async-output} | @var{notify-async-output}}
27771 @item @var{exec-async-output} @expansion{}
27772 @code{[ @var{token} ] "*" @var{async-output nl}}
27774 @item @var{status-async-output} @expansion{}
27775 @code{[ @var{token} ] "+" @var{async-output nl}}
27777 @item @var{notify-async-output} @expansion{}
27778 @code{[ @var{token} ] "=" @var{async-output nl}}
27780 @item @var{async-output} @expansion{}
27781 @code{@var{async-class} ( "," @var{result} )*}
27783 @item @var{result-class} @expansion{}
27784 @code{"done" | "running" | "connected" | "error" | "exit"}
27786 @item @var{async-class} @expansion{}
27787 @code{"stopped" | @var{others}} (where @var{others} will be added
27788 depending on the needs---this is still in development).
27790 @item @var{result} @expansion{}
27791 @code{ @var{variable} "=" @var{value}}
27793 @item @var{variable} @expansion{}
27794 @code{ @var{string} }
27796 @item @var{value} @expansion{}
27797 @code{ @var{const} | @var{tuple} | @var{list} }
27799 @item @var{const} @expansion{}
27800 @code{@var{c-string}}
27802 @item @var{tuple} @expansion{}
27803 @code{ "@{@}" | "@{" @var{result} ( "," @var{result} )* "@}" }
27805 @item @var{list} @expansion{}
27806 @code{ "[]" | "[" @var{value} ( "," @var{value} )* "]" | "["
27807 @var{result} ( "," @var{result} )* "]" }
27809 @item @var{stream-record} @expansion{}
27810 @code{@var{console-stream-output} | @var{target-stream-output} | @var{log-stream-output}}
27812 @item @var{console-stream-output} @expansion{}
27813 @code{"~" @var{c-string nl}}
27815 @item @var{target-stream-output} @expansion{}
27816 @code{"@@" @var{c-string nl}}
27818 @item @var{log-stream-output} @expansion{}
27819 @code{"&" @var{c-string nl}}
27821 @item @var{nl} @expansion{}
27824 @item @var{token} @expansion{}
27825 @emph{any sequence of digits}.
27833 All output sequences end in a single line containing a period.
27836 The @code{@var{token}} is from the corresponding request. Note that
27837 for all async output, while the token is allowed by the grammar and
27838 may be output by future versions of @value{GDBN} for select async
27839 output messages, it is generally omitted. Frontends should treat
27840 all async output as reporting general changes in the state of the
27841 target and there should be no need to associate async output to any
27845 @cindex status output in @sc{gdb/mi}
27846 @var{status-async-output} contains on-going status information about the
27847 progress of a slow operation. It can be discarded. All status output is
27848 prefixed by @samp{+}.
27851 @cindex async output in @sc{gdb/mi}
27852 @var{exec-async-output} contains asynchronous state change on the target
27853 (stopped, started, disappeared). All async output is prefixed by
27857 @cindex notify output in @sc{gdb/mi}
27858 @var{notify-async-output} contains supplementary information that the
27859 client should handle (e.g., a new breakpoint information). All notify
27860 output is prefixed by @samp{=}.
27863 @cindex console output in @sc{gdb/mi}
27864 @var{console-stream-output} is output that should be displayed as is in the
27865 console. It is the textual response to a CLI command. All the console
27866 output is prefixed by @samp{~}.
27869 @cindex target output in @sc{gdb/mi}
27870 @var{target-stream-output} is the output produced by the target program.
27871 All the target output is prefixed by @samp{@@}.
27874 @cindex log output in @sc{gdb/mi}
27875 @var{log-stream-output} is output text coming from @value{GDBN}'s internals, for
27876 instance messages that should be displayed as part of an error log. All
27877 the log output is prefixed by @samp{&}.
27880 @cindex list output in @sc{gdb/mi}
27881 New @sc{gdb/mi} commands should only output @var{lists} containing
27887 @xref{GDB/MI Stream Records, , @sc{gdb/mi} Stream Records}, for more
27888 details about the various output records.
27890 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27891 @node GDB/MI Compatibility with CLI
27892 @section @sc{gdb/mi} Compatibility with CLI
27894 @cindex compatibility, @sc{gdb/mi} and CLI
27895 @cindex @sc{gdb/mi}, compatibility with CLI
27897 For the developers convenience CLI commands can be entered directly,
27898 but there may be some unexpected behaviour. For example, commands
27899 that query the user will behave as if the user replied yes, breakpoint
27900 command lists are not executed and some CLI commands, such as
27901 @code{if}, @code{when} and @code{define}, prompt for further input with
27902 @samp{>}, which is not valid MI output.
27904 This feature may be removed at some stage in the future and it is
27905 recommended that front ends use the @code{-interpreter-exec} command
27906 (@pxref{-interpreter-exec}).
27908 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27909 @node GDB/MI Development and Front Ends
27910 @section @sc{gdb/mi} Development and Front Ends
27911 @cindex @sc{gdb/mi} development
27913 The application which takes the MI output and presents the state of the
27914 program being debugged to the user is called a @dfn{front end}.
27916 Since @sc{gdb/mi} is used by a variety of front ends to @value{GDBN}, changes
27917 to the MI interface may break existing usage. This section describes how the
27918 protocol changes and how to request previous version of the protocol when it
27921 Some changes in MI need not break a carefully designed front end, and
27922 for these the MI version will remain unchanged. The following is a
27923 list of changes that may occur within one level, so front ends should
27924 parse MI output in a way that can handle them:
27928 New MI commands may be added.
27931 New fields may be added to the output of any MI command.
27934 The range of values for fields with specified values, e.g.,
27935 @code{in_scope} (@pxref{-var-update}) may be extended.
27937 @c The format of field's content e.g type prefix, may change so parse it
27938 @c at your own risk. Yes, in general?
27940 @c The order of fields may change? Shouldn't really matter but it might
27941 @c resolve inconsistencies.
27944 If the changes are likely to break front ends, the MI version level
27945 will be increased by one. The new versions of the MI protocol are not compatible
27946 with the old versions. Old versions of MI remain available, allowing front ends
27947 to keep using them until they are modified to use the latest MI version.
27949 Since @code{--interpreter=mi} always points to the latest MI version, it is
27950 recommended that front ends request a specific version of MI when launching
27951 @value{GDBN} (e.g. @code{--interpreter=mi2}) to make sure they get an
27952 interpreter with the MI version they expect.
27954 The following table gives a summary of the the released versions of the MI
27955 interface: the version number, the version of GDB in which it first appeared
27956 and the breaking changes compared to the previous version.
27958 @multitable @columnfractions .05 .05 .9
27959 @headitem MI version @tab GDB version @tab Breaking changes
27976 The @code{-environment-pwd}, @code{-environment-directory} and
27977 @code{-environment-path} commands now returns values using the MI output
27978 syntax, rather than CLI output syntax.
27981 @code{-var-list-children}'s @code{children} result field is now a list, rather
27985 @code{-var-update}'s @code{changelist} result field is now a list, rather than
27997 The output of information about multi-location breakpoints has changed in the
27998 responses to the @code{-break-insert} and @code{-break-info} commands, as well
27999 as in the @code{=breakpoint-created} and @code{=breakpoint-modified} events.
28000 The multiple locations are now placed in a @code{locations} field, whose value
28006 If your front end cannot yet migrate to a more recent version of the
28007 MI protocol, you can nevertheless selectively enable specific features
28008 available in those recent MI versions, using the following commands:
28012 @item -fix-multi-location-breakpoint-output
28013 Use the output for multi-location breakpoints which was introduced by
28014 MI 3, even when using MI versions 2 or 1. This command has no
28015 effect when using MI version 3 or later.
28019 The best way to avoid unexpected changes in MI that might break your front
28020 end is to make your project known to @value{GDBN} developers and
28021 follow development on @email{gdb@@sourceware.org} and
28022 @email{gdb-patches@@sourceware.org}.
28023 @cindex mailing lists
28025 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28026 @node GDB/MI Output Records
28027 @section @sc{gdb/mi} Output Records
28030 * GDB/MI Result Records::
28031 * GDB/MI Stream Records::
28032 * GDB/MI Async Records::
28033 * GDB/MI Breakpoint Information::
28034 * GDB/MI Frame Information::
28035 * GDB/MI Thread Information::
28036 * GDB/MI Ada Exception Information::
28039 @node GDB/MI Result Records
28040 @subsection @sc{gdb/mi} Result Records
28042 @cindex result records in @sc{gdb/mi}
28043 @cindex @sc{gdb/mi}, result records
28044 In addition to a number of out-of-band notifications, the response to a
28045 @sc{gdb/mi} command includes one of the following result indications:
28049 @item "^done" [ "," @var{results} ]
28050 The synchronous operation was successful, @code{@var{results}} are the return
28055 This result record is equivalent to @samp{^done}. Historically, it
28056 was output instead of @samp{^done} if the command has resumed the
28057 target. This behaviour is maintained for backward compatibility, but
28058 all frontends should treat @samp{^done} and @samp{^running}
28059 identically and rely on the @samp{*running} output record to determine
28060 which threads are resumed.
28064 @value{GDBN} has connected to a remote target.
28066 @item "^error" "," "msg=" @var{c-string} [ "," "code=" @var{c-string} ]
28068 The operation failed. The @code{msg=@var{c-string}} variable contains
28069 the corresponding error message.
28071 If present, the @code{code=@var{c-string}} variable provides an error
28072 code on which consumers can rely on to detect the corresponding
28073 error condition. At present, only one error code is defined:
28076 @item "undefined-command"
28077 Indicates that the command causing the error does not exist.
28082 @value{GDBN} has terminated.
28086 @node GDB/MI Stream Records
28087 @subsection @sc{gdb/mi} Stream Records
28089 @cindex @sc{gdb/mi}, stream records
28090 @cindex stream records in @sc{gdb/mi}
28091 @value{GDBN} internally maintains a number of output streams: the console, the
28092 target, and the log. The output intended for each of these streams is
28093 funneled through the @sc{gdb/mi} interface using @dfn{stream records}.
28095 Each stream record begins with a unique @dfn{prefix character} which
28096 identifies its stream (@pxref{GDB/MI Output Syntax, , @sc{gdb/mi} Output
28097 Syntax}). In addition to the prefix, each stream record contains a
28098 @code{@var{string-output}}. This is either raw text (with an implicit new
28099 line) or a quoted C string (which does not contain an implicit newline).
28102 @item "~" @var{string-output}
28103 The console output stream contains text that should be displayed in the
28104 CLI console window. It contains the textual responses to CLI commands.
28106 @item "@@" @var{string-output}
28107 The target output stream contains any textual output from the running
28108 target. This is only present when GDB's event loop is truly
28109 asynchronous, which is currently only the case for remote targets.
28111 @item "&" @var{string-output}
28112 The log stream contains debugging messages being produced by @value{GDBN}'s
28116 @node GDB/MI Async Records
28117 @subsection @sc{gdb/mi} Async Records
28119 @cindex async records in @sc{gdb/mi}
28120 @cindex @sc{gdb/mi}, async records
28121 @dfn{Async} records are used to notify the @sc{gdb/mi} client of
28122 additional changes that have occurred. Those changes can either be a
28123 consequence of @sc{gdb/mi} commands (e.g., a breakpoint modified) or a result of
28124 target activity (e.g., target stopped).
28126 The following is the list of possible async records:
28130 @item *running,thread-id="@var{thread}"
28131 The target is now running. The @var{thread} field can be the global
28132 thread ID of the the thread that is now running, and it can be
28133 @samp{all} if all threads are running. The frontend should assume
28134 that no interaction with a running thread is possible after this
28135 notification is produced. The frontend should not assume that this
28136 notification is output only once for any command. @value{GDBN} may
28137 emit this notification several times, either for different threads,
28138 because it cannot resume all threads together, or even for a single
28139 thread, if the thread must be stepped though some code before letting
28142 @item *stopped,reason="@var{reason}",thread-id="@var{id}",stopped-threads="@var{stopped}",core="@var{core}"
28143 The target has stopped. The @var{reason} field can have one of the
28147 @item breakpoint-hit
28148 A breakpoint was reached.
28149 @item watchpoint-trigger
28150 A watchpoint was triggered.
28151 @item read-watchpoint-trigger
28152 A read watchpoint was triggered.
28153 @item access-watchpoint-trigger
28154 An access watchpoint was triggered.
28155 @item function-finished
28156 An -exec-finish or similar CLI command was accomplished.
28157 @item location-reached
28158 An -exec-until or similar CLI command was accomplished.
28159 @item watchpoint-scope
28160 A watchpoint has gone out of scope.
28161 @item end-stepping-range
28162 An -exec-next, -exec-next-instruction, -exec-step, -exec-step-instruction or
28163 similar CLI command was accomplished.
28164 @item exited-signalled
28165 The inferior exited because of a signal.
28167 The inferior exited.
28168 @item exited-normally
28169 The inferior exited normally.
28170 @item signal-received
28171 A signal was received by the inferior.
28173 The inferior has stopped due to a library being loaded or unloaded.
28174 This can happen when @code{stop-on-solib-events} (@pxref{Files}) is
28175 set or when a @code{catch load} or @code{catch unload} catchpoint is
28176 in use (@pxref{Set Catchpoints}).
28178 The inferior has forked. This is reported when @code{catch fork}
28179 (@pxref{Set Catchpoints}) has been used.
28181 The inferior has vforked. This is reported in when @code{catch vfork}
28182 (@pxref{Set Catchpoints}) has been used.
28183 @item syscall-entry
28184 The inferior entered a system call. This is reported when @code{catch
28185 syscall} (@pxref{Set Catchpoints}) has been used.
28186 @item syscall-return
28187 The inferior returned from a system call. This is reported when
28188 @code{catch syscall} (@pxref{Set Catchpoints}) has been used.
28190 The inferior called @code{exec}. This is reported when @code{catch exec}
28191 (@pxref{Set Catchpoints}) has been used.
28194 The @var{id} field identifies the global thread ID of the thread
28195 that directly caused the stop -- for example by hitting a breakpoint.
28196 Depending on whether all-stop
28197 mode is in effect (@pxref{All-Stop Mode}), @value{GDBN} may either
28198 stop all threads, or only the thread that directly triggered the stop.
28199 If all threads are stopped, the @var{stopped} field will have the
28200 value of @code{"all"}. Otherwise, the value of the @var{stopped}
28201 field will be a list of thread identifiers. Presently, this list will
28202 always include a single thread, but frontend should be prepared to see
28203 several threads in the list. The @var{core} field reports the
28204 processor core on which the stop event has happened. This field may be absent
28205 if such information is not available.
28207 @item =thread-group-added,id="@var{id}"
28208 @itemx =thread-group-removed,id="@var{id}"
28209 A thread group was either added or removed. The @var{id} field
28210 contains the @value{GDBN} identifier of the thread group. When a thread
28211 group is added, it generally might not be associated with a running
28212 process. When a thread group is removed, its id becomes invalid and
28213 cannot be used in any way.
28215 @item =thread-group-started,id="@var{id}",pid="@var{pid}"
28216 A thread group became associated with a running program,
28217 either because the program was just started or the thread group
28218 was attached to a program. The @var{id} field contains the
28219 @value{GDBN} identifier of the thread group. The @var{pid} field
28220 contains process identifier, specific to the operating system.
28222 @item =thread-group-exited,id="@var{id}"[,exit-code="@var{code}"]
28223 A thread group is no longer associated with a running program,
28224 either because the program has exited, or because it was detached
28225 from. The @var{id} field contains the @value{GDBN} identifier of the
28226 thread group. The @var{code} field is the exit code of the inferior; it exists
28227 only when the inferior exited with some code.
28229 @item =thread-created,id="@var{id}",group-id="@var{gid}"
28230 @itemx =thread-exited,id="@var{id}",group-id="@var{gid}"
28231 A thread either was created, or has exited. The @var{id} field
28232 contains the global @value{GDBN} identifier of the thread. The @var{gid}
28233 field identifies the thread group this thread belongs to.
28235 @item =thread-selected,id="@var{id}"[,frame="@var{frame}"]
28236 Informs that the selected thread or frame were changed. This notification
28237 is not emitted as result of the @code{-thread-select} or
28238 @code{-stack-select-frame} commands, but is emitted whenever an MI command
28239 that is not documented to change the selected thread and frame actually
28240 changes them. In particular, invoking, directly or indirectly
28241 (via user-defined command), the CLI @code{thread} or @code{frame} commands,
28242 will generate this notification. Changing the thread or frame from another
28243 user interface (see @ref{Interpreters}) will also generate this notification.
28245 The @var{frame} field is only present if the newly selected thread is
28246 stopped. See @ref{GDB/MI Frame Information} for the format of its value.
28248 We suggest that in response to this notification, front ends
28249 highlight the selected thread and cause subsequent commands to apply to
28252 @item =library-loaded,...
28253 Reports that a new library file was loaded by the program. This
28254 notification has 5 fields---@var{id}, @var{target-name},
28255 @var{host-name}, @var{symbols-loaded} and @var{ranges}. The @var{id} field is an
28256 opaque identifier of the library. For remote debugging case,
28257 @var{target-name} and @var{host-name} fields give the name of the
28258 library file on the target, and on the host respectively. For native
28259 debugging, both those fields have the same value. The
28260 @var{symbols-loaded} field is emitted only for backward compatibility
28261 and should not be relied on to convey any useful information. The
28262 @var{thread-group} field, if present, specifies the id of the thread
28263 group in whose context the library was loaded. If the field is
28264 absent, it means the library was loaded in the context of all present
28265 thread groups. The @var{ranges} field specifies the ranges of addresses belonging
28268 @item =library-unloaded,...
28269 Reports that a library was unloaded by the program. This notification
28270 has 3 fields---@var{id}, @var{target-name} and @var{host-name} with
28271 the same meaning as for the @code{=library-loaded} notification.
28272 The @var{thread-group} field, if present, specifies the id of the
28273 thread group in whose context the library was unloaded. If the field is
28274 absent, it means the library was unloaded in the context of all present
28277 @item =traceframe-changed,num=@var{tfnum},tracepoint=@var{tpnum}
28278 @itemx =traceframe-changed,end
28279 Reports that the trace frame was changed and its new number is
28280 @var{tfnum}. The number of the tracepoint associated with this trace
28281 frame is @var{tpnum}.
28283 @item =tsv-created,name=@var{name},initial=@var{initial}
28284 Reports that the new trace state variable @var{name} is created with
28285 initial value @var{initial}.
28287 @item =tsv-deleted,name=@var{name}
28288 @itemx =tsv-deleted
28289 Reports that the trace state variable @var{name} is deleted or all
28290 trace state variables are deleted.
28292 @item =tsv-modified,name=@var{name},initial=@var{initial}[,current=@var{current}]
28293 Reports that the trace state variable @var{name} is modified with
28294 the initial value @var{initial}. The current value @var{current} of
28295 trace state variable is optional and is reported if the current
28296 value of trace state variable is known.
28298 @item =breakpoint-created,bkpt=@{...@}
28299 @itemx =breakpoint-modified,bkpt=@{...@}
28300 @itemx =breakpoint-deleted,id=@var{number}
28301 Reports that a breakpoint was created, modified, or deleted,
28302 respectively. Only user-visible breakpoints are reported to the MI
28305 The @var{bkpt} argument is of the same form as returned by the various
28306 breakpoint commands; @xref{GDB/MI Breakpoint Commands}. The
28307 @var{number} is the ordinal number of the breakpoint.
28309 Note that if a breakpoint is emitted in the result record of a
28310 command, then it will not also be emitted in an async record.
28312 @item =record-started,thread-group="@var{id}",method="@var{method}"[,format="@var{format}"]
28313 @itemx =record-stopped,thread-group="@var{id}"
28314 Execution log recording was either started or stopped on an
28315 inferior. The @var{id} is the @value{GDBN} identifier of the thread
28316 group corresponding to the affected inferior.
28318 The @var{method} field indicates the method used to record execution. If the
28319 method in use supports multiple recording formats, @var{format} will be present
28320 and contain the currently used format. @xref{Process Record and Replay},
28321 for existing method and format values.
28323 @item =cmd-param-changed,param=@var{param},value=@var{value}
28324 Reports that a parameter of the command @code{set @var{param}} is
28325 changed to @var{value}. In the multi-word @code{set} command,
28326 the @var{param} is the whole parameter list to @code{set} command.
28327 For example, In command @code{set check type on}, @var{param}
28328 is @code{check type} and @var{value} is @code{on}.
28330 @item =memory-changed,thread-group=@var{id},addr=@var{addr},len=@var{len}[,type="code"]
28331 Reports that bytes from @var{addr} to @var{data} + @var{len} were
28332 written in an inferior. The @var{id} is the identifier of the
28333 thread group corresponding to the affected inferior. The optional
28334 @code{type="code"} part is reported if the memory written to holds
28338 @node GDB/MI Breakpoint Information
28339 @subsection @sc{gdb/mi} Breakpoint Information
28341 When @value{GDBN} reports information about a breakpoint, a
28342 tracepoint, a watchpoint, or a catchpoint, it uses a tuple with the
28347 The breakpoint number.
28350 The type of the breakpoint. For ordinary breakpoints this will be
28351 @samp{breakpoint}, but many values are possible.
28354 If the type of the breakpoint is @samp{catchpoint}, then this
28355 indicates the exact type of catchpoint.
28358 This is the breakpoint disposition---either @samp{del}, meaning that
28359 the breakpoint will be deleted at the next stop, or @samp{keep},
28360 meaning that the breakpoint will not be deleted.
28363 This indicates whether the breakpoint is enabled, in which case the
28364 value is @samp{y}, or disabled, in which case the value is @samp{n}.
28365 Note that this is not the same as the field @code{enable}.
28368 The address of the breakpoint. This may be a hexidecimal number,
28369 giving the address; or the string @samp{<PENDING>}, for a pending
28370 breakpoint; or the string @samp{<MULTIPLE>}, for a breakpoint with
28371 multiple locations. This field will not be present if no address can
28372 be determined. For example, a watchpoint does not have an address.
28375 If known, the function in which the breakpoint appears.
28376 If not known, this field is not present.
28379 The name of the source file which contains this function, if known.
28380 If not known, this field is not present.
28383 The full file name of the source file which contains this function, if
28384 known. If not known, this field is not present.
28387 The line number at which this breakpoint appears, if known.
28388 If not known, this field is not present.
28391 If the source file is not known, this field may be provided. If
28392 provided, this holds the address of the breakpoint, possibly followed
28396 If this breakpoint is pending, this field is present and holds the
28397 text used to set the breakpoint, as entered by the user.
28400 Where this breakpoint's condition is evaluated, either @samp{host} or
28404 If this is a thread-specific breakpoint, then this identifies the
28405 thread in which the breakpoint can trigger.
28408 If this breakpoint is restricted to a particular Ada task, then this
28409 field will hold the task identifier.
28412 If the breakpoint is conditional, this is the condition expression.
28415 The ignore count of the breakpoint.
28418 The enable count of the breakpoint.
28420 @item traceframe-usage
28423 @item static-tracepoint-marker-string-id
28424 For a static tracepoint, the name of the static tracepoint marker.
28427 For a masked watchpoint, this is the mask.
28430 A tracepoint's pass count.
28432 @item original-location
28433 The location of the breakpoint as originally specified by the user.
28434 This field is optional.
28437 The number of times the breakpoint has been hit.
28440 This field is only given for tracepoints. This is either @samp{y},
28441 meaning that the tracepoint is installed, or @samp{n}, meaning that it
28445 Some extra data, the exact contents of which are type-dependent.
28448 This field is present if the breakpoint has multiple locations. It is also
28449 exceptionally present if the breakpoint is enabled and has a single, disabled
28452 The value is a list of locations. The format of a location is decribed below.
28456 A location in a multi-location breakpoint is represented as a tuple with the
28462 The location number as a dotted pair, like @samp{1.2}. The first digit is the
28463 number of the parent breakpoint. The second digit is the number of the
28464 location within that breakpoint.
28467 This indicates whether the location is enabled, in which case the
28468 value is @samp{y}, or disabled, in which case the value is @samp{n}.
28469 Note that this is not the same as the field @code{enable}.
28472 The address of this location as an hexidecimal number.
28475 If known, the function in which the location appears.
28476 If not known, this field is not present.
28479 The name of the source file which contains this location, if known.
28480 If not known, this field is not present.
28483 The full file name of the source file which contains this location, if
28484 known. If not known, this field is not present.
28487 The line number at which this location appears, if known.
28488 If not known, this field is not present.
28490 @item thread-groups
28491 The thread groups this location is in.
28495 For example, here is what the output of @code{-break-insert}
28496 (@pxref{GDB/MI Breakpoint Commands}) might be:
28499 -> -break-insert main
28500 <- ^done,bkpt=@{number="1",type="breakpoint",disp="keep",
28501 enabled="y",addr="0x08048564",func="main",file="myprog.c",
28502 fullname="/home/nickrob/myprog.c",line="68",thread-groups=["i1"],
28507 @node GDB/MI Frame Information
28508 @subsection @sc{gdb/mi} Frame Information
28510 Response from many MI commands includes an information about stack
28511 frame. This information is a tuple that may have the following
28516 The level of the stack frame. The innermost frame has the level of
28517 zero. This field is always present.
28520 The name of the function corresponding to the frame. This field may
28521 be absent if @value{GDBN} is unable to determine the function name.
28524 The code address for the frame. This field is always present.
28527 The name of the source files that correspond to the frame's code
28528 address. This field may be absent.
28531 The source line corresponding to the frames' code address. This field
28535 The name of the binary file (either executable or shared library) the
28536 corresponds to the frame's code address. This field may be absent.
28540 @node GDB/MI Thread Information
28541 @subsection @sc{gdb/mi} Thread Information
28543 Whenever @value{GDBN} has to report an information about a thread, it
28544 uses a tuple with the following fields. The fields are always present unless
28549 The global numeric id assigned to the thread by @value{GDBN}.
28552 The target-specific string identifying the thread.
28555 Additional information about the thread provided by the target.
28556 It is supposed to be human-readable and not interpreted by the
28557 frontend. This field is optional.
28560 The name of the thread. If the user specified a name using the
28561 @code{thread name} command, then this name is given. Otherwise, if
28562 @value{GDBN} can extract the thread name from the target, then that
28563 name is given. If @value{GDBN} cannot find the thread name, then this
28567 The execution state of the thread, either @samp{stopped} or @samp{running},
28568 depending on whether the thread is presently running.
28571 The stack frame currently executing in the thread. This field is only present
28572 if the thread is stopped. Its format is documented in
28573 @ref{GDB/MI Frame Information}.
28576 The value of this field is an integer number of the processor core the
28577 thread was last seen on. This field is optional.
28580 @node GDB/MI Ada Exception Information
28581 @subsection @sc{gdb/mi} Ada Exception Information
28583 Whenever a @code{*stopped} record is emitted because the program
28584 stopped after hitting an exception catchpoint (@pxref{Set Catchpoints}),
28585 @value{GDBN} provides the name of the exception that was raised via
28586 the @code{exception-name} field. Also, for exceptions that were raised
28587 with an exception message, @value{GDBN} provides that message via
28588 the @code{exception-message} field.
28590 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28591 @node GDB/MI Simple Examples
28592 @section Simple Examples of @sc{gdb/mi} Interaction
28593 @cindex @sc{gdb/mi}, simple examples
28595 This subsection presents several simple examples of interaction using
28596 the @sc{gdb/mi} interface. In these examples, @samp{->} means that the
28597 following line is passed to @sc{gdb/mi} as input, while @samp{<-} means
28598 the output received from @sc{gdb/mi}.
28600 Note the line breaks shown in the examples are here only for
28601 readability, they don't appear in the real output.
28603 @subheading Setting a Breakpoint
28605 Setting a breakpoint generates synchronous output which contains detailed
28606 information of the breakpoint.
28609 -> -break-insert main
28610 <- ^done,bkpt=@{number="1",type="breakpoint",disp="keep",
28611 enabled="y",addr="0x08048564",func="main",file="myprog.c",
28612 fullname="/home/nickrob/myprog.c",line="68",thread-groups=["i1"],
28617 @subheading Program Execution
28619 Program execution generates asynchronous records and MI gives the
28620 reason that execution stopped.
28626 <- *stopped,reason="breakpoint-hit",disp="keep",bkptno="1",thread-id="0",
28627 frame=@{addr="0x08048564",func="main",
28628 args=[@{name="argc",value="1"@},@{name="argv",value="0xbfc4d4d4"@}],
28629 file="myprog.c",fullname="/home/nickrob/myprog.c",line="68",
28630 arch="i386:x86_64"@}
28635 <- *stopped,reason="exited-normally"
28639 @subheading Quitting @value{GDBN}
28641 Quitting @value{GDBN} just prints the result class @samp{^exit}.
28649 Please note that @samp{^exit} is printed immediately, but it might
28650 take some time for @value{GDBN} to actually exit. During that time, @value{GDBN}
28651 performs necessary cleanups, including killing programs being debugged
28652 or disconnecting from debug hardware, so the frontend should wait till
28653 @value{GDBN} exits and should only forcibly kill @value{GDBN} if it
28654 fails to exit in reasonable time.
28656 @subheading A Bad Command
28658 Here's what happens if you pass a non-existent command:
28662 <- ^error,msg="Undefined MI command: rubbish"
28667 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28668 @node GDB/MI Command Description Format
28669 @section @sc{gdb/mi} Command Description Format
28671 The remaining sections describe blocks of commands. Each block of
28672 commands is laid out in a fashion similar to this section.
28674 @subheading Motivation
28676 The motivation for this collection of commands.
28678 @subheading Introduction
28680 A brief introduction to this collection of commands as a whole.
28682 @subheading Commands
28684 For each command in the block, the following is described:
28686 @subsubheading Synopsis
28689 -command @var{args}@dots{}
28692 @subsubheading Result
28694 @subsubheading @value{GDBN} Command
28696 The corresponding @value{GDBN} CLI command(s), if any.
28698 @subsubheading Example
28700 Example(s) formatted for readability. Some of the described commands have
28701 not been implemented yet and these are labeled N.A.@: (not available).
28704 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28705 @node GDB/MI Breakpoint Commands
28706 @section @sc{gdb/mi} Breakpoint Commands
28708 @cindex breakpoint commands for @sc{gdb/mi}
28709 @cindex @sc{gdb/mi}, breakpoint commands
28710 This section documents @sc{gdb/mi} commands for manipulating
28713 @subheading The @code{-break-after} Command
28714 @findex -break-after
28716 @subsubheading Synopsis
28719 -break-after @var{number} @var{count}
28722 The breakpoint number @var{number} is not in effect until it has been
28723 hit @var{count} times. To see how this is reflected in the output of
28724 the @samp{-break-list} command, see the description of the
28725 @samp{-break-list} command below.
28727 @subsubheading @value{GDBN} Command
28729 The corresponding @value{GDBN} command is @samp{ignore}.
28731 @subsubheading Example
28736 ^done,bkpt=@{number="1",type="breakpoint",disp="keep",
28737 enabled="y",addr="0x000100d0",func="main",file="hello.c",
28738 fullname="/home/foo/hello.c",line="5",thread-groups=["i1"],
28746 ^done,BreakpointTable=@{nr_rows="1",nr_cols="6",
28747 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
28748 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
28749 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
28750 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
28751 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
28752 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
28753 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
28754 addr="0x000100d0",func="main",file="hello.c",fullname="/home/foo/hello.c",
28755 line="5",thread-groups=["i1"],times="0",ignore="3"@}]@}
28760 @subheading The @code{-break-catch} Command
28761 @findex -break-catch
28764 @subheading The @code{-break-commands} Command
28765 @findex -break-commands
28767 @subsubheading Synopsis
28770 -break-commands @var{number} [ @var{command1} ... @var{commandN} ]
28773 Specifies the CLI commands that should be executed when breakpoint
28774 @var{number} is hit. The parameters @var{command1} to @var{commandN}
28775 are the commands. If no command is specified, any previously-set
28776 commands are cleared. @xref{Break Commands}. Typical use of this
28777 functionality is tracing a program, that is, printing of values of
28778 some variables whenever breakpoint is hit and then continuing.
28780 @subsubheading @value{GDBN} Command
28782 The corresponding @value{GDBN} command is @samp{commands}.
28784 @subsubheading Example
28789 ^done,bkpt=@{number="1",type="breakpoint",disp="keep",
28790 enabled="y",addr="0x000100d0",func="main",file="hello.c",
28791 fullname="/home/foo/hello.c",line="5",thread-groups=["i1"],
28794 -break-commands 1 "print v" "continue"
28799 @subheading The @code{-break-condition} Command
28800 @findex -break-condition
28802 @subsubheading Synopsis
28805 -break-condition @var{number} @var{expr}
28808 Breakpoint @var{number} will stop the program only if the condition in
28809 @var{expr} is true. The condition becomes part of the
28810 @samp{-break-list} output (see the description of the @samp{-break-list}
28813 @subsubheading @value{GDBN} Command
28815 The corresponding @value{GDBN} command is @samp{condition}.
28817 @subsubheading Example
28821 -break-condition 1 1
28825 ^done,BreakpointTable=@{nr_rows="1",nr_cols="6",
28826 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
28827 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
28828 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
28829 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
28830 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
28831 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
28832 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
28833 addr="0x000100d0",func="main",file="hello.c",fullname="/home/foo/hello.c",
28834 line="5",cond="1",thread-groups=["i1"],times="0",ignore="3"@}]@}
28838 @subheading The @code{-break-delete} Command
28839 @findex -break-delete
28841 @subsubheading Synopsis
28844 -break-delete ( @var{breakpoint} )+
28847 Delete the breakpoint(s) whose number(s) are specified in the argument
28848 list. This is obviously reflected in the breakpoint list.
28850 @subsubheading @value{GDBN} Command
28852 The corresponding @value{GDBN} command is @samp{delete}.
28854 @subsubheading Example
28862 ^done,BreakpointTable=@{nr_rows="0",nr_cols="6",
28863 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
28864 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
28865 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
28866 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
28867 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
28868 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
28873 @subheading The @code{-break-disable} Command
28874 @findex -break-disable
28876 @subsubheading Synopsis
28879 -break-disable ( @var{breakpoint} )+
28882 Disable the named @var{breakpoint}(s). The field @samp{enabled} in the
28883 break list is now set to @samp{n} for the named @var{breakpoint}(s).
28885 @subsubheading @value{GDBN} Command
28887 The corresponding @value{GDBN} command is @samp{disable}.
28889 @subsubheading Example
28897 ^done,BreakpointTable=@{nr_rows="1",nr_cols="6",
28898 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
28899 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
28900 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
28901 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
28902 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
28903 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
28904 body=[bkpt=@{number="2",type="breakpoint",disp="keep",enabled="n",
28905 addr="0x000100d0",func="main",file="hello.c",fullname="/home/foo/hello.c",
28906 line="5",thread-groups=["i1"],times="0"@}]@}
28910 @subheading The @code{-break-enable} Command
28911 @findex -break-enable
28913 @subsubheading Synopsis
28916 -break-enable ( @var{breakpoint} )+
28919 Enable (previously disabled) @var{breakpoint}(s).
28921 @subsubheading @value{GDBN} Command
28923 The corresponding @value{GDBN} command is @samp{enable}.
28925 @subsubheading Example
28933 ^done,BreakpointTable=@{nr_rows="1",nr_cols="6",
28934 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
28935 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
28936 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
28937 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
28938 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
28939 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
28940 body=[bkpt=@{number="2",type="breakpoint",disp="keep",enabled="y",
28941 addr="0x000100d0",func="main",file="hello.c",fullname="/home/foo/hello.c",
28942 line="5",thread-groups=["i1"],times="0"@}]@}
28946 @subheading The @code{-break-info} Command
28947 @findex -break-info
28949 @subsubheading Synopsis
28952 -break-info @var{breakpoint}
28956 Get information about a single breakpoint.
28958 The result is a table of breakpoints. @xref{GDB/MI Breakpoint
28959 Information}, for details on the format of each breakpoint in the
28962 @subsubheading @value{GDBN} Command
28964 The corresponding @value{GDBN} command is @samp{info break @var{breakpoint}}.
28966 @subsubheading Example
28969 @subheading The @code{-break-insert} Command
28970 @findex -break-insert
28971 @anchor{-break-insert}
28973 @subsubheading Synopsis
28976 -break-insert [ -t ] [ -h ] [ -f ] [ -d ] [ -a ]
28977 [ -c @var{condition} ] [ -i @var{ignore-count} ]
28978 [ -p @var{thread-id} ] [ @var{location} ]
28982 If specified, @var{location}, can be one of:
28985 @item linespec location
28986 A linespec location. @xref{Linespec Locations}.
28988 @item explicit location
28989 An explicit location. @sc{gdb/mi} explicit locations are
28990 analogous to the CLI's explicit locations using the option names
28991 listed below. @xref{Explicit Locations}.
28994 @item --source @var{filename}
28995 The source file name of the location. This option requires the use
28996 of either @samp{--function} or @samp{--line}.
28998 @item --function @var{function}
28999 The name of a function or method.
29001 @item --label @var{label}
29002 The name of a label.
29004 @item --line @var{lineoffset}
29005 An absolute or relative line offset from the start of the location.
29008 @item address location
29009 An address location, *@var{address}. @xref{Address Locations}.
29013 The possible optional parameters of this command are:
29017 Insert a temporary breakpoint.
29019 Insert a hardware breakpoint.
29021 If @var{location} cannot be parsed (for example if it
29022 refers to unknown files or functions), create a pending
29023 breakpoint. Without this flag, @value{GDBN} will report
29024 an error, and won't create a breakpoint, if @var{location}
29027 Create a disabled breakpoint.
29029 Create a tracepoint. @xref{Tracepoints}. When this parameter
29030 is used together with @samp{-h}, a fast tracepoint is created.
29031 @item -c @var{condition}
29032 Make the breakpoint conditional on @var{condition}.
29033 @item -i @var{ignore-count}
29034 Initialize the @var{ignore-count}.
29035 @item -p @var{thread-id}
29036 Restrict the breakpoint to the thread with the specified global
29040 @subsubheading Result
29042 @xref{GDB/MI Breakpoint Information}, for details on the format of the
29043 resulting breakpoint.
29045 Note: this format is open to change.
29046 @c An out-of-band breakpoint instead of part of the result?
29048 @subsubheading @value{GDBN} Command
29050 The corresponding @value{GDBN} commands are @samp{break}, @samp{tbreak},
29051 @samp{hbreak}, and @samp{thbreak}. @c and @samp{rbreak}.
29053 @subsubheading Example
29058 ^done,bkpt=@{number="1",addr="0x0001072c",file="recursive2.c",
29059 fullname="/home/foo/recursive2.c,line="4",thread-groups=["i1"],
29062 -break-insert -t foo
29063 ^done,bkpt=@{number="2",addr="0x00010774",file="recursive2.c",
29064 fullname="/home/foo/recursive2.c,line="11",thread-groups=["i1"],
29068 ^done,BreakpointTable=@{nr_rows="2",nr_cols="6",
29069 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
29070 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
29071 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
29072 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
29073 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
29074 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
29075 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
29076 addr="0x0001072c", func="main",file="recursive2.c",
29077 fullname="/home/foo/recursive2.c,"line="4",thread-groups=["i1"],
29079 bkpt=@{number="2",type="breakpoint",disp="del",enabled="y",
29080 addr="0x00010774",func="foo",file="recursive2.c",
29081 fullname="/home/foo/recursive2.c",line="11",thread-groups=["i1"],
29084 @c -break-insert -r foo.*
29085 @c ~int foo(int, int);
29086 @c ^done,bkpt=@{number="3",addr="0x00010774",file="recursive2.c,
29087 @c "fullname="/home/foo/recursive2.c",line="11",thread-groups=["i1"],
29092 @subheading The @code{-dprintf-insert} Command
29093 @findex -dprintf-insert
29095 @subsubheading Synopsis
29098 -dprintf-insert [ -t ] [ -f ] [ -d ]
29099 [ -c @var{condition} ] [ -i @var{ignore-count} ]
29100 [ -p @var{thread-id} ] [ @var{location} ] [ @var{format} ]
29105 If supplied, @var{location} may be specified the same way as for
29106 the @code{-break-insert} command. @xref{-break-insert}.
29108 The possible optional parameters of this command are:
29112 Insert a temporary breakpoint.
29114 If @var{location} cannot be parsed (for example, if it
29115 refers to unknown files or functions), create a pending
29116 breakpoint. Without this flag, @value{GDBN} will report
29117 an error, and won't create a breakpoint, if @var{location}
29120 Create a disabled breakpoint.
29121 @item -c @var{condition}
29122 Make the breakpoint conditional on @var{condition}.
29123 @item -i @var{ignore-count}
29124 Set the ignore count of the breakpoint (@pxref{Conditions, ignore count})
29125 to @var{ignore-count}.
29126 @item -p @var{thread-id}
29127 Restrict the breakpoint to the thread with the specified global
29131 @subsubheading Result
29133 @xref{GDB/MI Breakpoint Information}, for details on the format of the
29134 resulting breakpoint.
29136 @c An out-of-band breakpoint instead of part of the result?
29138 @subsubheading @value{GDBN} Command
29140 The corresponding @value{GDBN} command is @samp{dprintf}.
29142 @subsubheading Example
29146 4-dprintf-insert foo "At foo entry\n"
29147 4^done,bkpt=@{number="1",type="dprintf",disp="keep",enabled="y",
29148 addr="0x000000000040061b",func="foo",file="mi-dprintf.c",
29149 fullname="mi-dprintf.c",line="25",thread-groups=["i1"],
29150 times="0",script=@{"printf \"At foo entry\\n\"","continue"@},
29151 original-location="foo"@}
29153 5-dprintf-insert 26 "arg=%d, g=%d\n" arg g
29154 5^done,bkpt=@{number="2",type="dprintf",disp="keep",enabled="y",
29155 addr="0x000000000040062a",func="foo",file="mi-dprintf.c",
29156 fullname="mi-dprintf.c",line="26",thread-groups=["i1"],
29157 times="0",script=@{"printf \"arg=%d, g=%d\\n\", arg, g","continue"@},
29158 original-location="mi-dprintf.c:26"@}
29162 @subheading The @code{-break-list} Command
29163 @findex -break-list
29165 @subsubheading Synopsis
29171 Displays the list of inserted breakpoints, showing the following fields:
29175 number of the breakpoint
29177 type of the breakpoint: @samp{breakpoint} or @samp{watchpoint}
29179 should the breakpoint be deleted or disabled when it is hit: @samp{keep}
29182 is the breakpoint enabled or no: @samp{y} or @samp{n}
29184 memory location at which the breakpoint is set
29186 logical location of the breakpoint, expressed by function name, file
29188 @item Thread-groups
29189 list of thread groups to which this breakpoint applies
29191 number of times the breakpoint has been hit
29194 If there are no breakpoints or watchpoints, the @code{BreakpointTable}
29195 @code{body} field is an empty list.
29197 @subsubheading @value{GDBN} Command
29199 The corresponding @value{GDBN} command is @samp{info break}.
29201 @subsubheading Example
29206 ^done,BreakpointTable=@{nr_rows="2",nr_cols="6",
29207 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
29208 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
29209 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
29210 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
29211 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
29212 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
29213 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
29214 addr="0x000100d0",func="main",file="hello.c",line="5",thread-groups=["i1"],
29216 bkpt=@{number="2",type="breakpoint",disp="keep",enabled="y",
29217 addr="0x00010114",func="foo",file="hello.c",fullname="/home/foo/hello.c",
29218 line="13",thread-groups=["i1"],times="0"@}]@}
29222 Here's an example of the result when there are no breakpoints:
29227 ^done,BreakpointTable=@{nr_rows="0",nr_cols="6",
29228 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
29229 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
29230 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
29231 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
29232 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
29233 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
29238 @subheading The @code{-break-passcount} Command
29239 @findex -break-passcount
29241 @subsubheading Synopsis
29244 -break-passcount @var{tracepoint-number} @var{passcount}
29247 Set the passcount for tracepoint @var{tracepoint-number} to
29248 @var{passcount}. If the breakpoint referred to by @var{tracepoint-number}
29249 is not a tracepoint, error is emitted. This corresponds to CLI
29250 command @samp{passcount}.
29252 @subheading The @code{-break-watch} Command
29253 @findex -break-watch
29255 @subsubheading Synopsis
29258 -break-watch [ -a | -r ]
29261 Create a watchpoint. With the @samp{-a} option it will create an
29262 @dfn{access} watchpoint, i.e., a watchpoint that triggers either on a
29263 read from or on a write to the memory location. With the @samp{-r}
29264 option, the watchpoint created is a @dfn{read} watchpoint, i.e., it will
29265 trigger only when the memory location is accessed for reading. Without
29266 either of the options, the watchpoint created is a regular watchpoint,
29267 i.e., it will trigger when the memory location is accessed for writing.
29268 @xref{Set Watchpoints, , Setting Watchpoints}.
29270 Note that @samp{-break-list} will report a single list of watchpoints and
29271 breakpoints inserted.
29273 @subsubheading @value{GDBN} Command
29275 The corresponding @value{GDBN} commands are @samp{watch}, @samp{awatch}, and
29278 @subsubheading Example
29280 Setting a watchpoint on a variable in the @code{main} function:
29285 ^done,wpt=@{number="2",exp="x"@}
29290 *stopped,reason="watchpoint-trigger",wpt=@{number="2",exp="x"@},
29291 value=@{old="-268439212",new="55"@},
29292 frame=@{func="main",args=[],file="recursive2.c",
29293 fullname="/home/foo/bar/recursive2.c",line="5",arch="i386:x86_64"@}
29297 Setting a watchpoint on a variable local to a function. @value{GDBN} will stop
29298 the program execution twice: first for the variable changing value, then
29299 for the watchpoint going out of scope.
29304 ^done,wpt=@{number="5",exp="C"@}
29309 *stopped,reason="watchpoint-trigger",
29310 wpt=@{number="5",exp="C"@},value=@{old="-276895068",new="3"@},
29311 frame=@{func="callee4",args=[],
29312 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
29313 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="13",
29314 arch="i386:x86_64"@}
29319 *stopped,reason="watchpoint-scope",wpnum="5",
29320 frame=@{func="callee3",args=[@{name="strarg",
29321 value="0x11940 \"A string argument.\""@}],
29322 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
29323 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="18",
29324 arch="i386:x86_64"@}
29328 Listing breakpoints and watchpoints, at different points in the program
29329 execution. Note that once the watchpoint goes out of scope, it is
29335 ^done,wpt=@{number="2",exp="C"@}
29338 ^done,BreakpointTable=@{nr_rows="2",nr_cols="6",
29339 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
29340 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
29341 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
29342 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
29343 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
29344 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
29345 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
29346 addr="0x00010734",func="callee4",
29347 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
29348 fullname="/home/foo/devo/gdb/testsuite/gdb.mi/basics.c"line="8",thread-groups=["i1"],
29350 bkpt=@{number="2",type="watchpoint",disp="keep",
29351 enabled="y",addr="",what="C",thread-groups=["i1"],times="0"@}]@}
29356 *stopped,reason="watchpoint-trigger",wpt=@{number="2",exp="C"@},
29357 value=@{old="-276895068",new="3"@},
29358 frame=@{func="callee4",args=[],
29359 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
29360 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="13",
29361 arch="i386:x86_64"@}
29364 ^done,BreakpointTable=@{nr_rows="2",nr_cols="6",
29365 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
29366 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
29367 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
29368 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
29369 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
29370 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
29371 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
29372 addr="0x00010734",func="callee4",
29373 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
29374 fullname="/home/foo/devo/gdb/testsuite/gdb.mi/basics.c",line="8",thread-groups=["i1"],
29376 bkpt=@{number="2",type="watchpoint",disp="keep",
29377 enabled="y",addr="",what="C",thread-groups=["i1"],times="-5"@}]@}
29381 ^done,reason="watchpoint-scope",wpnum="2",
29382 frame=@{func="callee3",args=[@{name="strarg",
29383 value="0x11940 \"A string argument.\""@}],
29384 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
29385 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="18",
29386 arch="i386:x86_64"@}
29389 ^done,BreakpointTable=@{nr_rows="1",nr_cols="6",
29390 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
29391 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
29392 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
29393 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
29394 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
29395 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
29396 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
29397 addr="0x00010734",func="callee4",
29398 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
29399 fullname="/home/foo/devo/gdb/testsuite/gdb.mi/basics.c",line="8",
29400 thread-groups=["i1"],times="1"@}]@}
29405 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
29406 @node GDB/MI Catchpoint Commands
29407 @section @sc{gdb/mi} Catchpoint Commands
29409 This section documents @sc{gdb/mi} commands for manipulating
29413 * Shared Library GDB/MI Catchpoint Commands::
29414 * Ada Exception GDB/MI Catchpoint Commands::
29417 @node Shared Library GDB/MI Catchpoint Commands
29418 @subsection Shared Library @sc{gdb/mi} Catchpoints
29420 @subheading The @code{-catch-load} Command
29421 @findex -catch-load
29423 @subsubheading Synopsis
29426 -catch-load [ -t ] [ -d ] @var{regexp}
29429 Add a catchpoint for library load events. If the @samp{-t} option is used,
29430 the catchpoint is a temporary one (@pxref{Set Breaks, ,Setting
29431 Breakpoints}). If the @samp{-d} option is used, the catchpoint is created
29432 in a disabled state. The @samp{regexp} argument is a regular
29433 expression used to match the name of the loaded library.
29436 @subsubheading @value{GDBN} Command
29438 The corresponding @value{GDBN} command is @samp{catch load}.
29440 @subsubheading Example
29443 -catch-load -t foo.so
29444 ^done,bkpt=@{number="1",type="catchpoint",disp="del",enabled="y",
29445 what="load of library matching foo.so",catch-type="load",times="0"@}
29450 @subheading The @code{-catch-unload} Command
29451 @findex -catch-unload
29453 @subsubheading Synopsis
29456 -catch-unload [ -t ] [ -d ] @var{regexp}
29459 Add a catchpoint for library unload events. If the @samp{-t} option is
29460 used, the catchpoint is a temporary one (@pxref{Set Breaks, ,Setting
29461 Breakpoints}). If the @samp{-d} option is used, the catchpoint is
29462 created in a disabled state. The @samp{regexp} argument is a regular
29463 expression used to match the name of the unloaded library.
29465 @subsubheading @value{GDBN} Command
29467 The corresponding @value{GDBN} command is @samp{catch unload}.
29469 @subsubheading Example
29472 -catch-unload -d bar.so
29473 ^done,bkpt=@{number="2",type="catchpoint",disp="keep",enabled="n",
29474 what="load of library matching bar.so",catch-type="unload",times="0"@}
29478 @node Ada Exception GDB/MI Catchpoint Commands
29479 @subsection Ada Exception @sc{gdb/mi} Catchpoints
29481 The following @sc{gdb/mi} commands can be used to create catchpoints
29482 that stop the execution when Ada exceptions are being raised.
29484 @subheading The @code{-catch-assert} Command
29485 @findex -catch-assert
29487 @subsubheading Synopsis
29490 -catch-assert [ -c @var{condition}] [ -d ] [ -t ]
29493 Add a catchpoint for failed Ada assertions.
29495 The possible optional parameters for this command are:
29498 @item -c @var{condition}
29499 Make the catchpoint conditional on @var{condition}.
29501 Create a disabled catchpoint.
29503 Create a temporary catchpoint.
29506 @subsubheading @value{GDBN} Command
29508 The corresponding @value{GDBN} command is @samp{catch assert}.
29510 @subsubheading Example
29514 ^done,bkptno="5",bkpt=@{number="5",type="breakpoint",disp="keep",
29515 enabled="y",addr="0x0000000000404888",what="failed Ada assertions",
29516 thread-groups=["i1"],times="0",
29517 original-location="__gnat_debug_raise_assert_failure"@}
29521 @subheading The @code{-catch-exception} Command
29522 @findex -catch-exception
29524 @subsubheading Synopsis
29527 -catch-exception [ -c @var{condition}] [ -d ] [ -e @var{exception-name} ]
29531 Add a catchpoint stopping when Ada exceptions are raised.
29532 By default, the command stops the program when any Ada exception
29533 gets raised. But it is also possible, by using some of the
29534 optional parameters described below, to create more selective
29537 The possible optional parameters for this command are:
29540 @item -c @var{condition}
29541 Make the catchpoint conditional on @var{condition}.
29543 Create a disabled catchpoint.
29544 @item -e @var{exception-name}
29545 Only stop when @var{exception-name} is raised. This option cannot
29546 be used combined with @samp{-u}.
29548 Create a temporary catchpoint.
29550 Stop only when an unhandled exception gets raised. This option
29551 cannot be used combined with @samp{-e}.
29554 @subsubheading @value{GDBN} Command
29556 The corresponding @value{GDBN} commands are @samp{catch exception}
29557 and @samp{catch exception unhandled}.
29559 @subsubheading Example
29562 -catch-exception -e Program_Error
29563 ^done,bkptno="4",bkpt=@{number="4",type="breakpoint",disp="keep",
29564 enabled="y",addr="0x0000000000404874",
29565 what="`Program_Error' Ada exception", thread-groups=["i1"],
29566 times="0",original-location="__gnat_debug_raise_exception"@}
29570 @subheading The @code{-catch-handlers} Command
29571 @findex -catch-handlers
29573 @subsubheading Synopsis
29576 -catch-handlers [ -c @var{condition}] [ -d ] [ -e @var{exception-name} ]
29580 Add a catchpoint stopping when Ada exceptions are handled.
29581 By default, the command stops the program when any Ada exception
29582 gets handled. But it is also possible, by using some of the
29583 optional parameters described below, to create more selective
29586 The possible optional parameters for this command are:
29589 @item -c @var{condition}
29590 Make the catchpoint conditional on @var{condition}.
29592 Create a disabled catchpoint.
29593 @item -e @var{exception-name}
29594 Only stop when @var{exception-name} is handled.
29596 Create a temporary catchpoint.
29599 @subsubheading @value{GDBN} Command
29601 The corresponding @value{GDBN} command is @samp{catch handlers}.
29603 @subsubheading Example
29606 -catch-handlers -e Constraint_Error
29607 ^done,bkptno="4",bkpt=@{number="4",type="breakpoint",disp="keep",
29608 enabled="y",addr="0x0000000000402f68",
29609 what="`Constraint_Error' Ada exception handlers",thread-groups=["i1"],
29610 times="0",original-location="__gnat_begin_handler"@}
29614 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
29615 @node GDB/MI Program Context
29616 @section @sc{gdb/mi} Program Context
29618 @subheading The @code{-exec-arguments} Command
29619 @findex -exec-arguments
29622 @subsubheading Synopsis
29625 -exec-arguments @var{args}
29628 Set the inferior program arguments, to be used in the next
29631 @subsubheading @value{GDBN} Command
29633 The corresponding @value{GDBN} command is @samp{set args}.
29635 @subsubheading Example
29639 -exec-arguments -v word
29646 @subheading The @code{-exec-show-arguments} Command
29647 @findex -exec-show-arguments
29649 @subsubheading Synopsis
29652 -exec-show-arguments
29655 Print the arguments of the program.
29657 @subsubheading @value{GDBN} Command
29659 The corresponding @value{GDBN} command is @samp{show args}.
29661 @subsubheading Example
29666 @subheading The @code{-environment-cd} Command
29667 @findex -environment-cd
29669 @subsubheading Synopsis
29672 -environment-cd @var{pathdir}
29675 Set @value{GDBN}'s working directory.
29677 @subsubheading @value{GDBN} Command
29679 The corresponding @value{GDBN} command is @samp{cd}.
29681 @subsubheading Example
29685 -environment-cd /kwikemart/marge/ezannoni/flathead-dev/devo/gdb
29691 @subheading The @code{-environment-directory} Command
29692 @findex -environment-directory
29694 @subsubheading Synopsis
29697 -environment-directory [ -r ] [ @var{pathdir} ]+
29700 Add directories @var{pathdir} to beginning of search path for source files.
29701 If the @samp{-r} option is used, the search path is reset to the default
29702 search path. If directories @var{pathdir} are supplied in addition to the
29703 @samp{-r} option, the search path is first reset and then addition
29705 Multiple directories may be specified, separated by blanks. Specifying
29706 multiple directories in a single command
29707 results in the directories added to the beginning of the
29708 search path in the same order they were presented in the command.
29709 If blanks are needed as
29710 part of a directory name, double-quotes should be used around
29711 the name. In the command output, the path will show up separated
29712 by the system directory-separator character. The directory-separator
29713 character must not be used
29714 in any directory name.
29715 If no directories are specified, the current search path is displayed.
29717 @subsubheading @value{GDBN} Command
29719 The corresponding @value{GDBN} command is @samp{dir}.
29721 @subsubheading Example
29725 -environment-directory /kwikemart/marge/ezannoni/flathead-dev/devo/gdb
29726 ^done,source-path="/kwikemart/marge/ezannoni/flathead-dev/devo/gdb:$cdir:$cwd"
29728 -environment-directory ""
29729 ^done,source-path="/kwikemart/marge/ezannoni/flathead-dev/devo/gdb:$cdir:$cwd"
29731 -environment-directory -r /home/jjohnstn/src/gdb /usr/src
29732 ^done,source-path="/home/jjohnstn/src/gdb:/usr/src:$cdir:$cwd"
29734 -environment-directory -r
29735 ^done,source-path="$cdir:$cwd"
29740 @subheading The @code{-environment-path} Command
29741 @findex -environment-path
29743 @subsubheading Synopsis
29746 -environment-path [ -r ] [ @var{pathdir} ]+
29749 Add directories @var{pathdir} to beginning of search path for object files.
29750 If the @samp{-r} option is used, the search path is reset to the original
29751 search path that existed at gdb start-up. If directories @var{pathdir} are
29752 supplied in addition to the
29753 @samp{-r} option, the search path is first reset and then addition
29755 Multiple directories may be specified, separated by blanks. Specifying
29756 multiple directories in a single command
29757 results in the directories added to the beginning of the
29758 search path in the same order they were presented in the command.
29759 If blanks are needed as
29760 part of a directory name, double-quotes should be used around
29761 the name. In the command output, the path will show up separated
29762 by the system directory-separator character. The directory-separator
29763 character must not be used
29764 in any directory name.
29765 If no directories are specified, the current path is displayed.
29768 @subsubheading @value{GDBN} Command
29770 The corresponding @value{GDBN} command is @samp{path}.
29772 @subsubheading Example
29777 ^done,path="/usr/bin"
29779 -environment-path /kwikemart/marge/ezannoni/flathead-dev/ppc-eabi/gdb /bin
29780 ^done,path="/kwikemart/marge/ezannoni/flathead-dev/ppc-eabi/gdb:/bin:/usr/bin"
29782 -environment-path -r /usr/local/bin
29783 ^done,path="/usr/local/bin:/usr/bin"
29788 @subheading The @code{-environment-pwd} Command
29789 @findex -environment-pwd
29791 @subsubheading Synopsis
29797 Show the current working directory.
29799 @subsubheading @value{GDBN} Command
29801 The corresponding @value{GDBN} command is @samp{pwd}.
29803 @subsubheading Example
29808 ^done,cwd="/kwikemart/marge/ezannoni/flathead-dev/devo/gdb"
29812 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
29813 @node GDB/MI Thread Commands
29814 @section @sc{gdb/mi} Thread Commands
29817 @subheading The @code{-thread-info} Command
29818 @findex -thread-info
29820 @subsubheading Synopsis
29823 -thread-info [ @var{thread-id} ]
29826 Reports information about either a specific thread, if the
29827 @var{thread-id} parameter is present, or about all threads.
29828 @var{thread-id} is the thread's global thread ID. When printing
29829 information about all threads, also reports the global ID of the
29832 @subsubheading @value{GDBN} Command
29834 The @samp{info thread} command prints the same information
29837 @subsubheading Result
29839 The result contains the following attributes:
29843 A list of threads. The format of the elements of the list is described in
29844 @ref{GDB/MI Thread Information}.
29846 @item current-thread-id
29847 The global id of the currently selected thread. This field is omitted if there
29848 is no selected thread (for example, when the selected inferior is not running,
29849 and therefore has no threads) or if a @var{thread-id} argument was passed to
29854 @subsubheading Example
29859 @{id="2",target-id="Thread 0xb7e14b90 (LWP 21257)",
29860 frame=@{level="0",addr="0xffffe410",func="__kernel_vsyscall",
29861 args=[]@},state="running"@},
29862 @{id="1",target-id="Thread 0xb7e156b0 (LWP 21254)",
29863 frame=@{level="0",addr="0x0804891f",func="foo",
29864 args=[@{name="i",value="10"@}],
29865 file="/tmp/a.c",fullname="/tmp/a.c",line="158",arch="i386:x86_64"@},
29866 state="running"@}],
29867 current-thread-id="1"
29871 @subheading The @code{-thread-list-ids} Command
29872 @findex -thread-list-ids
29874 @subsubheading Synopsis
29880 Produces a list of the currently known global @value{GDBN} thread ids.
29881 At the end of the list it also prints the total number of such
29884 This command is retained for historical reasons, the
29885 @code{-thread-info} command should be used instead.
29887 @subsubheading @value{GDBN} Command
29889 Part of @samp{info threads} supplies the same information.
29891 @subsubheading Example
29896 ^done,thread-ids=@{thread-id="3",thread-id="2",thread-id="1"@},
29897 current-thread-id="1",number-of-threads="3"
29902 @subheading The @code{-thread-select} Command
29903 @findex -thread-select
29905 @subsubheading Synopsis
29908 -thread-select @var{thread-id}
29911 Make thread with global thread number @var{thread-id} the current
29912 thread. It prints the number of the new current thread, and the
29913 topmost frame for that thread.
29915 This command is deprecated in favor of explicitly using the
29916 @samp{--thread} option to each command.
29918 @subsubheading @value{GDBN} Command
29920 The corresponding @value{GDBN} command is @samp{thread}.
29922 @subsubheading Example
29929 *stopped,reason="end-stepping-range",thread-id="2",line="187",
29930 file="../../../devo/gdb/testsuite/gdb.threads/linux-dp.c"
29934 thread-ids=@{thread-id="3",thread-id="2",thread-id="1"@},
29935 number-of-threads="3"
29938 ^done,new-thread-id="3",
29939 frame=@{level="0",func="vprintf",
29940 args=[@{name="format",value="0x8048e9c \"%*s%c %d %c\\n\""@},
29941 @{name="arg",value="0x2"@}],file="vprintf.c",line="31",arch="i386:x86_64"@}
29945 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
29946 @node GDB/MI Ada Tasking Commands
29947 @section @sc{gdb/mi} Ada Tasking Commands
29949 @subheading The @code{-ada-task-info} Command
29950 @findex -ada-task-info
29952 @subsubheading Synopsis
29955 -ada-task-info [ @var{task-id} ]
29958 Reports information about either a specific Ada task, if the
29959 @var{task-id} parameter is present, or about all Ada tasks.
29961 @subsubheading @value{GDBN} Command
29963 The @samp{info tasks} command prints the same information
29964 about all Ada tasks (@pxref{Ada Tasks}).
29966 @subsubheading Result
29968 The result is a table of Ada tasks. The following columns are
29969 defined for each Ada task:
29973 This field exists only for the current thread. It has the value @samp{*}.
29976 The identifier that @value{GDBN} uses to refer to the Ada task.
29979 The identifier that the target uses to refer to the Ada task.
29982 The global thread identifier of the thread corresponding to the Ada
29985 This field should always exist, as Ada tasks are always implemented
29986 on top of a thread. But if @value{GDBN} cannot find this corresponding
29987 thread for any reason, the field is omitted.
29990 This field exists only when the task was created by another task.
29991 In this case, it provides the ID of the parent task.
29994 The base priority of the task.
29997 The current state of the task. For a detailed description of the
29998 possible states, see @ref{Ada Tasks}.
30001 The name of the task.
30005 @subsubheading Example
30009 ^done,tasks=@{nr_rows="3",nr_cols="8",
30010 hdr=[@{width="1",alignment="-1",col_name="current",colhdr=""@},
30011 @{width="3",alignment="1",col_name="id",colhdr="ID"@},
30012 @{width="9",alignment="1",col_name="task-id",colhdr="TID"@},
30013 @{width="4",alignment="1",col_name="thread-id",colhdr=""@},
30014 @{width="4",alignment="1",col_name="parent-id",colhdr="P-ID"@},
30015 @{width="3",alignment="1",col_name="priority",colhdr="Pri"@},
30016 @{width="22",alignment="-1",col_name="state",colhdr="State"@},
30017 @{width="1",alignment="2",col_name="name",colhdr="Name"@}],
30018 body=[@{current="*",id="1",task-id=" 644010",thread-id="1",priority="48",
30019 state="Child Termination Wait",name="main_task"@}]@}
30023 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
30024 @node GDB/MI Program Execution
30025 @section @sc{gdb/mi} Program Execution
30027 These are the asynchronous commands which generate the out-of-band
30028 record @samp{*stopped}. Currently @value{GDBN} only really executes
30029 asynchronously with remote targets and this interaction is mimicked in
30032 @subheading The @code{-exec-continue} Command
30033 @findex -exec-continue
30035 @subsubheading Synopsis
30038 -exec-continue [--reverse] [--all|--thread-group N]
30041 Resumes the execution of the inferior program, which will continue
30042 to execute until it reaches a debugger stop event. If the
30043 @samp{--reverse} option is specified, execution resumes in reverse until
30044 it reaches a stop event. Stop events may include
30047 breakpoints or watchpoints
30049 signals or exceptions
30051 the end of the process (or its beginning under @samp{--reverse})
30053 the end or beginning of a replay log if one is being used.
30055 In all-stop mode (@pxref{All-Stop
30056 Mode}), may resume only one thread, or all threads, depending on the
30057 value of the @samp{scheduler-locking} variable. If @samp{--all} is
30058 specified, all threads (in all inferiors) will be resumed. The @samp{--all} option is
30059 ignored in all-stop mode. If the @samp{--thread-group} options is
30060 specified, then all threads in that thread group are resumed.
30062 @subsubheading @value{GDBN} Command
30064 The corresponding @value{GDBN} corresponding is @samp{continue}.
30066 @subsubheading Example
30073 *stopped,reason="breakpoint-hit",disp="keep",bkptno="2",frame=@{
30074 func="foo",args=[],file="hello.c",fullname="/home/foo/bar/hello.c",
30075 line="13",arch="i386:x86_64"@}
30080 @subheading The @code{-exec-finish} Command
30081 @findex -exec-finish
30083 @subsubheading Synopsis
30086 -exec-finish [--reverse]
30089 Resumes the execution of the inferior program until the current
30090 function is exited. Displays the results returned by the function.
30091 If the @samp{--reverse} option is specified, resumes the reverse
30092 execution of the inferior program until the point where current
30093 function was called.
30095 @subsubheading @value{GDBN} Command
30097 The corresponding @value{GDBN} command is @samp{finish}.
30099 @subsubheading Example
30101 Function returning @code{void}.
30108 *stopped,reason="function-finished",frame=@{func="main",args=[],
30109 file="hello.c",fullname="/home/foo/bar/hello.c",line="7",arch="i386:x86_64"@}
30113 Function returning other than @code{void}. The name of the internal
30114 @value{GDBN} variable storing the result is printed, together with the
30121 *stopped,reason="function-finished",frame=@{addr="0x000107b0",func="foo",
30122 args=[@{name="a",value="1"],@{name="b",value="9"@}@},
30123 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30124 arch="i386:x86_64"@},
30125 gdb-result-var="$1",return-value="0"
30130 @subheading The @code{-exec-interrupt} Command
30131 @findex -exec-interrupt
30133 @subsubheading Synopsis
30136 -exec-interrupt [--all|--thread-group N]
30139 Interrupts the background execution of the target. Note how the token
30140 associated with the stop message is the one for the execution command
30141 that has been interrupted. The token for the interrupt itself only
30142 appears in the @samp{^done} output. If the user is trying to
30143 interrupt a non-running program, an error message will be printed.
30145 Note that when asynchronous execution is enabled, this command is
30146 asynchronous just like other execution commands. That is, first the
30147 @samp{^done} response will be printed, and the target stop will be
30148 reported after that using the @samp{*stopped} notification.
30150 In non-stop mode, only the context thread is interrupted by default.
30151 All threads (in all inferiors) will be interrupted if the
30152 @samp{--all} option is specified. If the @samp{--thread-group}
30153 option is specified, all threads in that group will be interrupted.
30155 @subsubheading @value{GDBN} Command
30157 The corresponding @value{GDBN} command is @samp{interrupt}.
30159 @subsubheading Example
30170 111*stopped,signal-name="SIGINT",signal-meaning="Interrupt",
30171 frame=@{addr="0x00010140",func="foo",args=[],file="try.c",
30172 fullname="/home/foo/bar/try.c",line="13",arch="i386:x86_64"@}
30177 ^error,msg="mi_cmd_exec_interrupt: Inferior not executing."
30181 @subheading The @code{-exec-jump} Command
30184 @subsubheading Synopsis
30187 -exec-jump @var{location}
30190 Resumes execution of the inferior program at the location specified by
30191 parameter. @xref{Specify Location}, for a description of the
30192 different forms of @var{location}.
30194 @subsubheading @value{GDBN} Command
30196 The corresponding @value{GDBN} command is @samp{jump}.
30198 @subsubheading Example
30201 -exec-jump foo.c:10
30202 *running,thread-id="all"
30207 @subheading The @code{-exec-next} Command
30210 @subsubheading Synopsis
30213 -exec-next [--reverse]
30216 Resumes execution of the inferior program, stopping when the beginning
30217 of the next source line is reached.
30219 If the @samp{--reverse} option is specified, resumes reverse execution
30220 of the inferior program, stopping at the beginning of the previous
30221 source line. If you issue this command on the first line of a
30222 function, it will take you back to the caller of that function, to the
30223 source line where the function was called.
30226 @subsubheading @value{GDBN} Command
30228 The corresponding @value{GDBN} command is @samp{next}.
30230 @subsubheading Example
30236 *stopped,reason="end-stepping-range",line="8",file="hello.c"
30241 @subheading The @code{-exec-next-instruction} Command
30242 @findex -exec-next-instruction
30244 @subsubheading Synopsis
30247 -exec-next-instruction [--reverse]
30250 Executes one machine instruction. If the instruction is a function
30251 call, continues until the function returns. If the program stops at an
30252 instruction in the middle of a source line, the address will be
30255 If the @samp{--reverse} option is specified, resumes reverse execution
30256 of the inferior program, stopping at the previous instruction. If the
30257 previously executed instruction was a return from another function,
30258 it will continue to execute in reverse until the call to that function
30259 (from the current stack frame) is reached.
30261 @subsubheading @value{GDBN} Command
30263 The corresponding @value{GDBN} command is @samp{nexti}.
30265 @subsubheading Example
30269 -exec-next-instruction
30273 *stopped,reason="end-stepping-range",
30274 addr="0x000100d4",line="5",file="hello.c"
30279 @subheading The @code{-exec-return} Command
30280 @findex -exec-return
30282 @subsubheading Synopsis
30288 Makes current function return immediately. Doesn't execute the inferior.
30289 Displays the new current frame.
30291 @subsubheading @value{GDBN} Command
30293 The corresponding @value{GDBN} command is @samp{return}.
30295 @subsubheading Example
30299 200-break-insert callee4
30300 200^done,bkpt=@{number="1",addr="0x00010734",
30301 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",line="8"@}
30306 000*stopped,reason="breakpoint-hit",disp="keep",bkptno="1",
30307 frame=@{func="callee4",args=[],
30308 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
30309 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="8",
30310 arch="i386:x86_64"@}
30316 111^done,frame=@{level="0",func="callee3",
30317 args=[@{name="strarg",
30318 value="0x11940 \"A string argument.\""@}],
30319 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
30320 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="18",
30321 arch="i386:x86_64"@}
30326 @subheading The @code{-exec-run} Command
30329 @subsubheading Synopsis
30332 -exec-run [ --all | --thread-group N ] [ --start ]
30335 Starts execution of the inferior from the beginning. The inferior
30336 executes until either a breakpoint is encountered or the program
30337 exits. In the latter case the output will include an exit code, if
30338 the program has exited exceptionally.
30340 When neither the @samp{--all} nor the @samp{--thread-group} option
30341 is specified, the current inferior is started. If the
30342 @samp{--thread-group} option is specified, it should refer to a thread
30343 group of type @samp{process}, and that thread group will be started.
30344 If the @samp{--all} option is specified, then all inferiors will be started.
30346 Using the @samp{--start} option instructs the debugger to stop
30347 the execution at the start of the inferior's main subprogram,
30348 following the same behavior as the @code{start} command
30349 (@pxref{Starting}).
30351 @subsubheading @value{GDBN} Command
30353 The corresponding @value{GDBN} command is @samp{run}.
30355 @subsubheading Examples
30360 ^done,bkpt=@{number="1",addr="0x0001072c",file="recursive2.c",line="4"@}
30365 *stopped,reason="breakpoint-hit",disp="keep",bkptno="1",
30366 frame=@{func="main",args=[],file="recursive2.c",
30367 fullname="/home/foo/bar/recursive2.c",line="4",arch="i386:x86_64"@}
30372 Program exited normally:
30380 *stopped,reason="exited-normally"
30385 Program exited exceptionally:
30393 *stopped,reason="exited",exit-code="01"
30397 Another way the program can terminate is if it receives a signal such as
30398 @code{SIGINT}. In this case, @sc{gdb/mi} displays this:
30402 *stopped,reason="exited-signalled",signal-name="SIGINT",
30403 signal-meaning="Interrupt"
30407 @c @subheading -exec-signal
30410 @subheading The @code{-exec-step} Command
30413 @subsubheading Synopsis
30416 -exec-step [--reverse]
30419 Resumes execution of the inferior program, stopping when the beginning
30420 of the next source line is reached, if the next source line is not a
30421 function call. If it is, stop at the first instruction of the called
30422 function. If the @samp{--reverse} option is specified, resumes reverse
30423 execution of the inferior program, stopping at the beginning of the
30424 previously executed source line.
30426 @subsubheading @value{GDBN} Command
30428 The corresponding @value{GDBN} command is @samp{step}.
30430 @subsubheading Example
30432 Stepping into a function:
30438 *stopped,reason="end-stepping-range",
30439 frame=@{func="foo",args=[@{name="a",value="10"@},
30440 @{name="b",value="0"@}],file="recursive2.c",
30441 fullname="/home/foo/bar/recursive2.c",line="11",arch="i386:x86_64"@}
30451 *stopped,reason="end-stepping-range",line="14",file="recursive2.c"
30456 @subheading The @code{-exec-step-instruction} Command
30457 @findex -exec-step-instruction
30459 @subsubheading Synopsis
30462 -exec-step-instruction [--reverse]
30465 Resumes the inferior which executes one machine instruction. If the
30466 @samp{--reverse} option is specified, resumes reverse execution of the
30467 inferior program, stopping at the previously executed instruction.
30468 The output, once @value{GDBN} has stopped, will vary depending on
30469 whether we have stopped in the middle of a source line or not. In the
30470 former case, the address at which the program stopped will be printed
30473 @subsubheading @value{GDBN} Command
30475 The corresponding @value{GDBN} command is @samp{stepi}.
30477 @subsubheading Example
30481 -exec-step-instruction
30485 *stopped,reason="end-stepping-range",
30486 frame=@{func="foo",args=[],file="try.c",
30487 fullname="/home/foo/bar/try.c",line="10",arch="i386:x86_64"@}
30489 -exec-step-instruction
30493 *stopped,reason="end-stepping-range",
30494 frame=@{addr="0x000100f4",func="foo",args=[],file="try.c",
30495 fullname="/home/foo/bar/try.c",line="10",arch="i386:x86_64"@}
30500 @subheading The @code{-exec-until} Command
30501 @findex -exec-until
30503 @subsubheading Synopsis
30506 -exec-until [ @var{location} ]
30509 Executes the inferior until the @var{location} specified in the
30510 argument is reached. If there is no argument, the inferior executes
30511 until a source line greater than the current one is reached. The
30512 reason for stopping in this case will be @samp{location-reached}.
30514 @subsubheading @value{GDBN} Command
30516 The corresponding @value{GDBN} command is @samp{until}.
30518 @subsubheading Example
30522 -exec-until recursive2.c:6
30526 *stopped,reason="location-reached",frame=@{func="main",args=[],
30527 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="6",
30528 arch="i386:x86_64"@}
30533 @subheading -file-clear
30534 Is this going away????
30537 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
30538 @node GDB/MI Stack Manipulation
30539 @section @sc{gdb/mi} Stack Manipulation Commands
30541 @subheading The @code{-enable-frame-filters} Command
30542 @findex -enable-frame-filters
30545 -enable-frame-filters
30548 @value{GDBN} allows Python-based frame filters to affect the output of
30549 the MI commands relating to stack traces. As there is no way to
30550 implement this in a fully backward-compatible way, a front end must
30551 request that this functionality be enabled.
30553 Once enabled, this feature cannot be disabled.
30555 Note that if Python support has not been compiled into @value{GDBN},
30556 this command will still succeed (and do nothing).
30558 @subheading The @code{-stack-info-frame} Command
30559 @findex -stack-info-frame
30561 @subsubheading Synopsis
30567 Get info on the selected frame.
30569 @subsubheading @value{GDBN} Command
30571 The corresponding @value{GDBN} command is @samp{info frame} or @samp{frame}
30572 (without arguments).
30574 @subsubheading Example
30579 ^done,frame=@{level="1",addr="0x0001076c",func="callee3",
30580 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
30581 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="17",
30582 arch="i386:x86_64"@}
30586 @subheading The @code{-stack-info-depth} Command
30587 @findex -stack-info-depth
30589 @subsubheading Synopsis
30592 -stack-info-depth [ @var{max-depth} ]
30595 Return the depth of the stack. If the integer argument @var{max-depth}
30596 is specified, do not count beyond @var{max-depth} frames.
30598 @subsubheading @value{GDBN} Command
30600 There's no equivalent @value{GDBN} command.
30602 @subsubheading Example
30604 For a stack with frame levels 0 through 11:
30611 -stack-info-depth 4
30614 -stack-info-depth 12
30617 -stack-info-depth 11
30620 -stack-info-depth 13
30625 @anchor{-stack-list-arguments}
30626 @subheading The @code{-stack-list-arguments} Command
30627 @findex -stack-list-arguments
30629 @subsubheading Synopsis
30632 -stack-list-arguments [ --no-frame-filters ] [ --skip-unavailable ] @var{print-values}
30633 [ @var{low-frame} @var{high-frame} ]
30636 Display a list of the arguments for the frames between @var{low-frame}
30637 and @var{high-frame} (inclusive). If @var{low-frame} and
30638 @var{high-frame} are not provided, list the arguments for the whole
30639 call stack. If the two arguments are equal, show the single frame
30640 at the corresponding level. It is an error if @var{low-frame} is
30641 larger than the actual number of frames. On the other hand,
30642 @var{high-frame} may be larger than the actual number of frames, in
30643 which case only existing frames will be returned.
30645 If @var{print-values} is 0 or @code{--no-values}, print only the names of
30646 the variables; if it is 1 or @code{--all-values}, print also their
30647 values; and if it is 2 or @code{--simple-values}, print the name,
30648 type and value for simple data types, and the name and type for arrays,
30649 structures and unions. If the option @code{--no-frame-filters} is
30650 supplied, then Python frame filters will not be executed.
30652 If the @code{--skip-unavailable} option is specified, arguments that
30653 are not available are not listed. Partially available arguments
30654 are still displayed, however.
30656 Use of this command to obtain arguments in a single frame is
30657 deprecated in favor of the @samp{-stack-list-variables} command.
30659 @subsubheading @value{GDBN} Command
30661 @value{GDBN} does not have an equivalent command. @code{gdbtk} has a
30662 @samp{gdb_get_args} command which partially overlaps with the
30663 functionality of @samp{-stack-list-arguments}.
30665 @subsubheading Example
30672 frame=@{level="0",addr="0x00010734",func="callee4",
30673 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
30674 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="8",
30675 arch="i386:x86_64"@},
30676 frame=@{level="1",addr="0x0001076c",func="callee3",
30677 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
30678 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="17",
30679 arch="i386:x86_64"@},
30680 frame=@{level="2",addr="0x0001078c",func="callee2",
30681 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
30682 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="22",
30683 arch="i386:x86_64"@},
30684 frame=@{level="3",addr="0x000107b4",func="callee1",
30685 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
30686 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="27",
30687 arch="i386:x86_64"@},
30688 frame=@{level="4",addr="0x000107e0",func="main",
30689 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
30690 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="32",
30691 arch="i386:x86_64"@}]
30693 -stack-list-arguments 0
30696 frame=@{level="0",args=[]@},
30697 frame=@{level="1",args=[name="strarg"]@},
30698 frame=@{level="2",args=[name="intarg",name="strarg"]@},
30699 frame=@{level="3",args=[name="intarg",name="strarg",name="fltarg"]@},
30700 frame=@{level="4",args=[]@}]
30702 -stack-list-arguments 1
30705 frame=@{level="0",args=[]@},
30707 args=[@{name="strarg",value="0x11940 \"A string argument.\""@}]@},
30708 frame=@{level="2",args=[
30709 @{name="intarg",value="2"@},
30710 @{name="strarg",value="0x11940 \"A string argument.\""@}]@},
30711 @{frame=@{level="3",args=[
30712 @{name="intarg",value="2"@},
30713 @{name="strarg",value="0x11940 \"A string argument.\""@},
30714 @{name="fltarg",value="3.5"@}]@},
30715 frame=@{level="4",args=[]@}]
30717 -stack-list-arguments 0 2 2
30718 ^done,stack-args=[frame=@{level="2",args=[name="intarg",name="strarg"]@}]
30720 -stack-list-arguments 1 2 2
30721 ^done,stack-args=[frame=@{level="2",
30722 args=[@{name="intarg",value="2"@},
30723 @{name="strarg",value="0x11940 \"A string argument.\""@}]@}]
30727 @c @subheading -stack-list-exception-handlers
30730 @anchor{-stack-list-frames}
30731 @subheading The @code{-stack-list-frames} Command
30732 @findex -stack-list-frames
30734 @subsubheading Synopsis
30737 -stack-list-frames [ --no-frame-filters @var{low-frame} @var{high-frame} ]
30740 List the frames currently on the stack. For each frame it displays the
30745 The frame number, 0 being the topmost frame, i.e., the innermost function.
30747 The @code{$pc} value for that frame.
30751 File name of the source file where the function lives.
30752 @item @var{fullname}
30753 The full file name of the source file where the function lives.
30755 Line number corresponding to the @code{$pc}.
30757 The shared library where this function is defined. This is only given
30758 if the frame's function is not known.
30760 Frame's architecture.
30763 If invoked without arguments, this command prints a backtrace for the
30764 whole stack. If given two integer arguments, it shows the frames whose
30765 levels are between the two arguments (inclusive). If the two arguments
30766 are equal, it shows the single frame at the corresponding level. It is
30767 an error if @var{low-frame} is larger than the actual number of
30768 frames. On the other hand, @var{high-frame} may be larger than the
30769 actual number of frames, in which case only existing frames will be
30770 returned. If the option @code{--no-frame-filters} is supplied, then
30771 Python frame filters will not be executed.
30773 @subsubheading @value{GDBN} Command
30775 The corresponding @value{GDBN} commands are @samp{backtrace} and @samp{where}.
30777 @subsubheading Example
30779 Full stack backtrace:
30785 [frame=@{level="0",addr="0x0001076c",func="foo",
30786 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="11",
30787 arch="i386:x86_64"@},
30788 frame=@{level="1",addr="0x000107a4",func="foo",
30789 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30790 arch="i386:x86_64"@},
30791 frame=@{level="2",addr="0x000107a4",func="foo",
30792 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30793 arch="i386:x86_64"@},
30794 frame=@{level="3",addr="0x000107a4",func="foo",
30795 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30796 arch="i386:x86_64"@},
30797 frame=@{level="4",addr="0x000107a4",func="foo",
30798 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30799 arch="i386:x86_64"@},
30800 frame=@{level="5",addr="0x000107a4",func="foo",
30801 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30802 arch="i386:x86_64"@},
30803 frame=@{level="6",addr="0x000107a4",func="foo",
30804 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30805 arch="i386:x86_64"@},
30806 frame=@{level="7",addr="0x000107a4",func="foo",
30807 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30808 arch="i386:x86_64"@},
30809 frame=@{level="8",addr="0x000107a4",func="foo",
30810 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30811 arch="i386:x86_64"@},
30812 frame=@{level="9",addr="0x000107a4",func="foo",
30813 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30814 arch="i386:x86_64"@},
30815 frame=@{level="10",addr="0x000107a4",func="foo",
30816 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30817 arch="i386:x86_64"@},
30818 frame=@{level="11",addr="0x00010738",func="main",
30819 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="4",
30820 arch="i386:x86_64"@}]
30824 Show frames between @var{low_frame} and @var{high_frame}:
30828 -stack-list-frames 3 5
30830 [frame=@{level="3",addr="0x000107a4",func="foo",
30831 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30832 arch="i386:x86_64"@},
30833 frame=@{level="4",addr="0x000107a4",func="foo",
30834 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30835 arch="i386:x86_64"@},
30836 frame=@{level="5",addr="0x000107a4",func="foo",
30837 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30838 arch="i386:x86_64"@}]
30842 Show a single frame:
30846 -stack-list-frames 3 3
30848 [frame=@{level="3",addr="0x000107a4",func="foo",
30849 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14",
30850 arch="i386:x86_64"@}]
30855 @subheading The @code{-stack-list-locals} Command
30856 @findex -stack-list-locals
30857 @anchor{-stack-list-locals}
30859 @subsubheading Synopsis
30862 -stack-list-locals [ --no-frame-filters ] [ --skip-unavailable ] @var{print-values}
30865 Display the local variable names for the selected frame. If
30866 @var{print-values} is 0 or @code{--no-values}, print only the names of
30867 the variables; if it is 1 or @code{--all-values}, print also their
30868 values; and if it is 2 or @code{--simple-values}, print the name,
30869 type and value for simple data types, and the name and type for arrays,
30870 structures and unions. In this last case, a frontend can immediately
30871 display the value of simple data types and create variable objects for
30872 other data types when the user wishes to explore their values in
30873 more detail. If the option @code{--no-frame-filters} is supplied, then
30874 Python frame filters will not be executed.
30876 If the @code{--skip-unavailable} option is specified, local variables
30877 that are not available are not listed. Partially available local
30878 variables are still displayed, however.
30880 This command is deprecated in favor of the
30881 @samp{-stack-list-variables} command.
30883 @subsubheading @value{GDBN} Command
30885 @samp{info locals} in @value{GDBN}, @samp{gdb_get_locals} in @code{gdbtk}.
30887 @subsubheading Example
30891 -stack-list-locals 0
30892 ^done,locals=[name="A",name="B",name="C"]
30894 -stack-list-locals --all-values
30895 ^done,locals=[@{name="A",value="1"@},@{name="B",value="2"@},
30896 @{name="C",value="@{1, 2, 3@}"@}]
30897 -stack-list-locals --simple-values
30898 ^done,locals=[@{name="A",type="int",value="1"@},
30899 @{name="B",type="int",value="2"@},@{name="C",type="int [3]"@}]
30903 @anchor{-stack-list-variables}
30904 @subheading The @code{-stack-list-variables} Command
30905 @findex -stack-list-variables
30907 @subsubheading Synopsis
30910 -stack-list-variables [ --no-frame-filters ] [ --skip-unavailable ] @var{print-values}
30913 Display the names of local variables and function arguments for the selected frame. If
30914 @var{print-values} is 0 or @code{--no-values}, print only the names of
30915 the variables; if it is 1 or @code{--all-values}, print also their
30916 values; and if it is 2 or @code{--simple-values}, print the name,
30917 type and value for simple data types, and the name and type for arrays,
30918 structures and unions. If the option @code{--no-frame-filters} is
30919 supplied, then Python frame filters will not be executed.
30921 If the @code{--skip-unavailable} option is specified, local variables
30922 and arguments that are not available are not listed. Partially
30923 available arguments and local variables are still displayed, however.
30925 @subsubheading Example
30929 -stack-list-variables --thread 1 --frame 0 --all-values
30930 ^done,variables=[@{name="x",value="11"@},@{name="s",value="@{a = 1, b = 2@}"@}]
30935 @subheading The @code{-stack-select-frame} Command
30936 @findex -stack-select-frame
30938 @subsubheading Synopsis
30941 -stack-select-frame @var{framenum}
30944 Change the selected frame. Select a different frame @var{framenum} on
30947 This command in deprecated in favor of passing the @samp{--frame}
30948 option to every command.
30950 @subsubheading @value{GDBN} Command
30952 The corresponding @value{GDBN} commands are @samp{frame}, @samp{up},
30953 @samp{down}, @samp{select-frame}, @samp{up-silent}, and @samp{down-silent}.
30955 @subsubheading Example
30959 -stack-select-frame 2
30964 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
30965 @node GDB/MI Variable Objects
30966 @section @sc{gdb/mi} Variable Objects
30970 @subheading Motivation for Variable Objects in @sc{gdb/mi}
30972 For the implementation of a variable debugger window (locals, watched
30973 expressions, etc.), we are proposing the adaptation of the existing code
30974 used by @code{Insight}.
30976 The two main reasons for that are:
30980 It has been proven in practice (it is already on its second generation).
30983 It will shorten development time (needless to say how important it is
30987 The original interface was designed to be used by Tcl code, so it was
30988 slightly changed so it could be used through @sc{gdb/mi}. This section
30989 describes the @sc{gdb/mi} operations that will be available and gives some
30990 hints about their use.
30992 @emph{Note}: In addition to the set of operations described here, we
30993 expect the @sc{gui} implementation of a variable window to require, at
30994 least, the following operations:
30997 @item @code{-gdb-show} @code{output-radix}
30998 @item @code{-stack-list-arguments}
30999 @item @code{-stack-list-locals}
31000 @item @code{-stack-select-frame}
31005 @subheading Introduction to Variable Objects
31007 @cindex variable objects in @sc{gdb/mi}
31009 Variable objects are "object-oriented" MI interface for examining and
31010 changing values of expressions. Unlike some other MI interfaces that
31011 work with expressions, variable objects are specifically designed for
31012 simple and efficient presentation in the frontend. A variable object
31013 is identified by string name. When a variable object is created, the
31014 frontend specifies the expression for that variable object. The
31015 expression can be a simple variable, or it can be an arbitrary complex
31016 expression, and can even involve CPU registers. After creating a
31017 variable object, the frontend can invoke other variable object
31018 operations---for example to obtain or change the value of a variable
31019 object, or to change display format.
31021 Variable objects have hierarchical tree structure. Any variable object
31022 that corresponds to a composite type, such as structure in C, has
31023 a number of child variable objects, for example corresponding to each
31024 element of a structure. A child variable object can itself have
31025 children, recursively. Recursion ends when we reach
31026 leaf variable objects, which always have built-in types. Child variable
31027 objects are created only by explicit request, so if a frontend
31028 is not interested in the children of a particular variable object, no
31029 child will be created.
31031 For a leaf variable object it is possible to obtain its value as a
31032 string, or set the value from a string. String value can be also
31033 obtained for a non-leaf variable object, but it's generally a string
31034 that only indicates the type of the object, and does not list its
31035 contents. Assignment to a non-leaf variable object is not allowed.
31037 A frontend does not need to read the values of all variable objects each time
31038 the program stops. Instead, MI provides an update command that lists all
31039 variable objects whose values has changed since the last update
31040 operation. This considerably reduces the amount of data that must
31041 be transferred to the frontend. As noted above, children variable
31042 objects are created on demand, and only leaf variable objects have a
31043 real value. As result, gdb will read target memory only for leaf
31044 variables that frontend has created.
31046 The automatic update is not always desirable. For example, a frontend
31047 might want to keep a value of some expression for future reference,
31048 and never update it. For another example, fetching memory is
31049 relatively slow for embedded targets, so a frontend might want
31050 to disable automatic update for the variables that are either not
31051 visible on the screen, or ``closed''. This is possible using so
31052 called ``frozen variable objects''. Such variable objects are never
31053 implicitly updated.
31055 Variable objects can be either @dfn{fixed} or @dfn{floating}. For the
31056 fixed variable object, the expression is parsed when the variable
31057 object is created, including associating identifiers to specific
31058 variables. The meaning of expression never changes. For a floating
31059 variable object the values of variables whose names appear in the
31060 expressions are re-evaluated every time in the context of the current
31061 frame. Consider this example:
31066 struct work_state state;
31073 If a fixed variable object for the @code{state} variable is created in
31074 this function, and we enter the recursive call, the variable
31075 object will report the value of @code{state} in the top-level
31076 @code{do_work} invocation. On the other hand, a floating variable
31077 object will report the value of @code{state} in the current frame.
31079 If an expression specified when creating a fixed variable object
31080 refers to a local variable, the variable object becomes bound to the
31081 thread and frame in which the variable object is created. When such
31082 variable object is updated, @value{GDBN} makes sure that the
31083 thread/frame combination the variable object is bound to still exists,
31084 and re-evaluates the variable object in context of that thread/frame.
31086 The following is the complete set of @sc{gdb/mi} operations defined to
31087 access this functionality:
31089 @multitable @columnfractions .4 .6
31090 @item @strong{Operation}
31091 @tab @strong{Description}
31093 @item @code{-enable-pretty-printing}
31094 @tab enable Python-based pretty-printing
31095 @item @code{-var-create}
31096 @tab create a variable object
31097 @item @code{-var-delete}
31098 @tab delete the variable object and/or its children
31099 @item @code{-var-set-format}
31100 @tab set the display format of this variable
31101 @item @code{-var-show-format}
31102 @tab show the display format of this variable
31103 @item @code{-var-info-num-children}
31104 @tab tells how many children this object has
31105 @item @code{-var-list-children}
31106 @tab return a list of the object's children
31107 @item @code{-var-info-type}
31108 @tab show the type of this variable object
31109 @item @code{-var-info-expression}
31110 @tab print parent-relative expression that this variable object represents
31111 @item @code{-var-info-path-expression}
31112 @tab print full expression that this variable object represents
31113 @item @code{-var-show-attributes}
31114 @tab is this variable editable? does it exist here?
31115 @item @code{-var-evaluate-expression}
31116 @tab get the value of this variable
31117 @item @code{-var-assign}
31118 @tab set the value of this variable
31119 @item @code{-var-update}
31120 @tab update the variable and its children
31121 @item @code{-var-set-frozen}
31122 @tab set frozeness attribute
31123 @item @code{-var-set-update-range}
31124 @tab set range of children to display on update
31127 In the next subsection we describe each operation in detail and suggest
31128 how it can be used.
31130 @subheading Description And Use of Operations on Variable Objects
31132 @subheading The @code{-enable-pretty-printing} Command
31133 @findex -enable-pretty-printing
31136 -enable-pretty-printing
31139 @value{GDBN} allows Python-based visualizers to affect the output of the
31140 MI variable object commands. However, because there was no way to
31141 implement this in a fully backward-compatible way, a front end must
31142 request that this functionality be enabled.
31144 Once enabled, this feature cannot be disabled.
31146 Note that if Python support has not been compiled into @value{GDBN},
31147 this command will still succeed (and do nothing).
31149 This feature is currently (as of @value{GDBN} 7.0) experimental, and
31150 may work differently in future versions of @value{GDBN}.
31152 @subheading The @code{-var-create} Command
31153 @findex -var-create
31155 @subsubheading Synopsis
31158 -var-create @{@var{name} | "-"@}
31159 @{@var{frame-addr} | "*" | "@@"@} @var{expression}
31162 This operation creates a variable object, which allows the monitoring of
31163 a variable, the result of an expression, a memory cell or a CPU
31166 The @var{name} parameter is the string by which the object can be
31167 referenced. It must be unique. If @samp{-} is specified, the varobj
31168 system will generate a string ``varNNNNNN'' automatically. It will be
31169 unique provided that one does not specify @var{name} of that format.
31170 The command fails if a duplicate name is found.
31172 The frame under which the expression should be evaluated can be
31173 specified by @var{frame-addr}. A @samp{*} indicates that the current
31174 frame should be used. A @samp{@@} indicates that a floating variable
31175 object must be created.
31177 @var{expression} is any expression valid on the current language set (must not
31178 begin with a @samp{*}), or one of the following:
31182 @samp{*@var{addr}}, where @var{addr} is the address of a memory cell
31185 @samp{*@var{addr}-@var{addr}} --- a memory address range (TBD)
31188 @samp{$@var{regname}} --- a CPU register name
31191 @cindex dynamic varobj
31192 A varobj's contents may be provided by a Python-based pretty-printer. In this
31193 case the varobj is known as a @dfn{dynamic varobj}. Dynamic varobjs
31194 have slightly different semantics in some cases. If the
31195 @code{-enable-pretty-printing} command is not sent, then @value{GDBN}
31196 will never create a dynamic varobj. This ensures backward
31197 compatibility for existing clients.
31199 @subsubheading Result
31201 This operation returns attributes of the newly-created varobj. These
31206 The name of the varobj.
31209 The number of children of the varobj. This number is not necessarily
31210 reliable for a dynamic varobj. Instead, you must examine the
31211 @samp{has_more} attribute.
31214 The varobj's scalar value. For a varobj whose type is some sort of
31215 aggregate (e.g., a @code{struct}), or for a dynamic varobj, this value
31216 will not be interesting.
31219 The varobj's type. This is a string representation of the type, as
31220 would be printed by the @value{GDBN} CLI. If @samp{print object}
31221 (@pxref{Print Settings, set print object}) is set to @code{on}, the
31222 @emph{actual} (derived) type of the object is shown rather than the
31223 @emph{declared} one.
31226 If a variable object is bound to a specific thread, then this is the
31227 thread's global identifier.
31230 For a dynamic varobj, this indicates whether there appear to be any
31231 children available. For a non-dynamic varobj, this will be 0.
31234 This attribute will be present and have the value @samp{1} if the
31235 varobj is a dynamic varobj. If the varobj is not a dynamic varobj,
31236 then this attribute will not be present.
31239 A dynamic varobj can supply a display hint to the front end. The
31240 value comes directly from the Python pretty-printer object's
31241 @code{display_hint} method. @xref{Pretty Printing API}.
31244 Typical output will look like this:
31247 name="@var{name}",numchild="@var{N}",type="@var{type}",thread-id="@var{M}",
31248 has_more="@var{has_more}"
31252 @subheading The @code{-var-delete} Command
31253 @findex -var-delete
31255 @subsubheading Synopsis
31258 -var-delete [ -c ] @var{name}
31261 Deletes a previously created variable object and all of its children.
31262 With the @samp{-c} option, just deletes the children.
31264 Returns an error if the object @var{name} is not found.
31267 @subheading The @code{-var-set-format} Command
31268 @findex -var-set-format
31270 @subsubheading Synopsis
31273 -var-set-format @var{name} @var{format-spec}
31276 Sets the output format for the value of the object @var{name} to be
31279 @anchor{-var-set-format}
31280 The syntax for the @var{format-spec} is as follows:
31283 @var{format-spec} @expansion{}
31284 @{binary | decimal | hexadecimal | octal | natural | zero-hexadecimal@}
31287 The natural format is the default format choosen automatically
31288 based on the variable type (like decimal for an @code{int}, hex
31289 for pointers, etc.).
31291 The zero-hexadecimal format has a representation similar to hexadecimal
31292 but with padding zeroes to the left of the value. For example, a 32-bit
31293 hexadecimal value of 0x1234 would be represented as 0x00001234 in the
31294 zero-hexadecimal format.
31296 For a variable with children, the format is set only on the
31297 variable itself, and the children are not affected.
31299 @subheading The @code{-var-show-format} Command
31300 @findex -var-show-format
31302 @subsubheading Synopsis
31305 -var-show-format @var{name}
31308 Returns the format used to display the value of the object @var{name}.
31311 @var{format} @expansion{}
31316 @subheading The @code{-var-info-num-children} Command
31317 @findex -var-info-num-children
31319 @subsubheading Synopsis
31322 -var-info-num-children @var{name}
31325 Returns the number of children of a variable object @var{name}:
31331 Note that this number is not completely reliable for a dynamic varobj.
31332 It will return the current number of children, but more children may
31336 @subheading The @code{-var-list-children} Command
31337 @findex -var-list-children
31339 @subsubheading Synopsis
31342 -var-list-children [@var{print-values}] @var{name} [@var{from} @var{to}]
31344 @anchor{-var-list-children}
31346 Return a list of the children of the specified variable object and
31347 create variable objects for them, if they do not already exist. With
31348 a single argument or if @var{print-values} has a value of 0 or
31349 @code{--no-values}, print only the names of the variables; if
31350 @var{print-values} is 1 or @code{--all-values}, also print their
31351 values; and if it is 2 or @code{--simple-values} print the name and
31352 value for simple data types and just the name for arrays, structures
31355 @var{from} and @var{to}, if specified, indicate the range of children
31356 to report. If @var{from} or @var{to} is less than zero, the range is
31357 reset and all children will be reported. Otherwise, children starting
31358 at @var{from} (zero-based) and up to and excluding @var{to} will be
31361 If a child range is requested, it will only affect the current call to
31362 @code{-var-list-children}, but not future calls to @code{-var-update}.
31363 For this, you must instead use @code{-var-set-update-range}. The
31364 intent of this approach is to enable a front end to implement any
31365 update approach it likes; for example, scrolling a view may cause the
31366 front end to request more children with @code{-var-list-children}, and
31367 then the front end could call @code{-var-set-update-range} with a
31368 different range to ensure that future updates are restricted to just
31371 For each child the following results are returned:
31376 Name of the variable object created for this child.
31379 The expression to be shown to the user by the front end to designate this child.
31380 For example this may be the name of a structure member.
31382 For a dynamic varobj, this value cannot be used to form an
31383 expression. There is no way to do this at all with a dynamic varobj.
31385 For C/C@t{++} structures there are several pseudo children returned to
31386 designate access qualifiers. For these pseudo children @var{exp} is
31387 @samp{public}, @samp{private}, or @samp{protected}. In this case the
31388 type and value are not present.
31390 A dynamic varobj will not report the access qualifying
31391 pseudo-children, regardless of the language. This information is not
31392 available at all with a dynamic varobj.
31395 Number of children this child has. For a dynamic varobj, this will be
31399 The type of the child. If @samp{print object}
31400 (@pxref{Print Settings, set print object}) is set to @code{on}, the
31401 @emph{actual} (derived) type of the object is shown rather than the
31402 @emph{declared} one.
31405 If values were requested, this is the value.
31408 If this variable object is associated with a thread, this is the
31409 thread's global thread id. Otherwise this result is not present.
31412 If the variable object is frozen, this variable will be present with a value of 1.
31415 A dynamic varobj can supply a display hint to the front end. The
31416 value comes directly from the Python pretty-printer object's
31417 @code{display_hint} method. @xref{Pretty Printing API}.
31420 This attribute will be present and have the value @samp{1} if the
31421 varobj is a dynamic varobj. If the varobj is not a dynamic varobj,
31422 then this attribute will not be present.
31426 The result may have its own attributes:
31430 A dynamic varobj can supply a display hint to the front end. The
31431 value comes directly from the Python pretty-printer object's
31432 @code{display_hint} method. @xref{Pretty Printing API}.
31435 This is an integer attribute which is nonzero if there are children
31436 remaining after the end of the selected range.
31439 @subsubheading Example
31443 -var-list-children n
31444 ^done,numchild=@var{n},children=[child=@{name=@var{name},exp=@var{exp},
31445 numchild=@var{n},type=@var{type}@},@r{(repeats N times)}]
31447 -var-list-children --all-values n
31448 ^done,numchild=@var{n},children=[child=@{name=@var{name},exp=@var{exp},
31449 numchild=@var{n},value=@var{value},type=@var{type}@},@r{(repeats N times)}]
31453 @subheading The @code{-var-info-type} Command
31454 @findex -var-info-type
31456 @subsubheading Synopsis
31459 -var-info-type @var{name}
31462 Returns the type of the specified variable @var{name}. The type is
31463 returned as a string in the same format as it is output by the
31467 type=@var{typename}
31471 @subheading The @code{-var-info-expression} Command
31472 @findex -var-info-expression
31474 @subsubheading Synopsis
31477 -var-info-expression @var{name}
31480 Returns a string that is suitable for presenting this
31481 variable object in user interface. The string is generally
31482 not valid expression in the current language, and cannot be evaluated.
31484 For example, if @code{a} is an array, and variable object
31485 @code{A} was created for @code{a}, then we'll get this output:
31488 (gdb) -var-info-expression A.1
31489 ^done,lang="C",exp="1"
31493 Here, the value of @code{lang} is the language name, which can be
31494 found in @ref{Supported Languages}.
31496 Note that the output of the @code{-var-list-children} command also
31497 includes those expressions, so the @code{-var-info-expression} command
31500 @subheading The @code{-var-info-path-expression} Command
31501 @findex -var-info-path-expression
31503 @subsubheading Synopsis
31506 -var-info-path-expression @var{name}
31509 Returns an expression that can be evaluated in the current
31510 context and will yield the same value that a variable object has.
31511 Compare this with the @code{-var-info-expression} command, which
31512 result can be used only for UI presentation. Typical use of
31513 the @code{-var-info-path-expression} command is creating a
31514 watchpoint from a variable object.
31516 This command is currently not valid for children of a dynamic varobj,
31517 and will give an error when invoked on one.
31519 For example, suppose @code{C} is a C@t{++} class, derived from class
31520 @code{Base}, and that the @code{Base} class has a member called
31521 @code{m_size}. Assume a variable @code{c} is has the type of
31522 @code{C} and a variable object @code{C} was created for variable
31523 @code{c}. Then, we'll get this output:
31525 (gdb) -var-info-path-expression C.Base.public.m_size
31526 ^done,path_expr=((Base)c).m_size)
31529 @subheading The @code{-var-show-attributes} Command
31530 @findex -var-show-attributes
31532 @subsubheading Synopsis
31535 -var-show-attributes @var{name}
31538 List attributes of the specified variable object @var{name}:
31541 status=@var{attr} [ ( ,@var{attr} )* ]
31545 where @var{attr} is @code{@{ @{ editable | noneditable @} | TBD @}}.
31547 @subheading The @code{-var-evaluate-expression} Command
31548 @findex -var-evaluate-expression
31550 @subsubheading Synopsis
31553 -var-evaluate-expression [-f @var{format-spec}] @var{name}
31556 Evaluates the expression that is represented by the specified variable
31557 object and returns its value as a string. The format of the string
31558 can be specified with the @samp{-f} option. The possible values of
31559 this option are the same as for @code{-var-set-format}
31560 (@pxref{-var-set-format}). If the @samp{-f} option is not specified,
31561 the current display format will be used. The current display format
31562 can be changed using the @code{-var-set-format} command.
31568 Note that one must invoke @code{-var-list-children} for a variable
31569 before the value of a child variable can be evaluated.
31571 @subheading The @code{-var-assign} Command
31572 @findex -var-assign
31574 @subsubheading Synopsis
31577 -var-assign @var{name} @var{expression}
31580 Assigns the value of @var{expression} to the variable object specified
31581 by @var{name}. The object must be @samp{editable}. If the variable's
31582 value is altered by the assign, the variable will show up in any
31583 subsequent @code{-var-update} list.
31585 @subsubheading Example
31593 ^done,changelist=[@{name="var1",in_scope="true",type_changed="false"@}]
31597 @subheading The @code{-var-update} Command
31598 @findex -var-update
31600 @subsubheading Synopsis
31603 -var-update [@var{print-values}] @{@var{name} | "*"@}
31606 Reevaluate the expressions corresponding to the variable object
31607 @var{name} and all its direct and indirect children, and return the
31608 list of variable objects whose values have changed; @var{name} must
31609 be a root variable object. Here, ``changed'' means that the result of
31610 @code{-var-evaluate-expression} before and after the
31611 @code{-var-update} is different. If @samp{*} is used as the variable
31612 object names, all existing variable objects are updated, except
31613 for frozen ones (@pxref{-var-set-frozen}). The option
31614 @var{print-values} determines whether both names and values, or just
31615 names are printed. The possible values of this option are the same
31616 as for @code{-var-list-children} (@pxref{-var-list-children}). It is
31617 recommended to use the @samp{--all-values} option, to reduce the
31618 number of MI commands needed on each program stop.
31620 With the @samp{*} parameter, if a variable object is bound to a
31621 currently running thread, it will not be updated, without any
31624 If @code{-var-set-update-range} was previously used on a varobj, then
31625 only the selected range of children will be reported.
31627 @code{-var-update} reports all the changed varobjs in a tuple named
31630 Each item in the change list is itself a tuple holding:
31634 The name of the varobj.
31637 If values were requested for this update, then this field will be
31638 present and will hold the value of the varobj.
31641 @anchor{-var-update}
31642 This field is a string which may take one of three values:
31646 The variable object's current value is valid.
31649 The variable object does not currently hold a valid value but it may
31650 hold one in the future if its associated expression comes back into
31654 The variable object no longer holds a valid value.
31655 This can occur when the executable file being debugged has changed,
31656 either through recompilation or by using the @value{GDBN} @code{file}
31657 command. The front end should normally choose to delete these variable
31661 In the future new values may be added to this list so the front should
31662 be prepared for this possibility. @xref{GDB/MI Development and Front Ends, ,@sc{GDB/MI} Development and Front Ends}.
31665 This is only present if the varobj is still valid. If the type
31666 changed, then this will be the string @samp{true}; otherwise it will
31669 When a varobj's type changes, its children are also likely to have
31670 become incorrect. Therefore, the varobj's children are automatically
31671 deleted when this attribute is @samp{true}. Also, the varobj's update
31672 range, when set using the @code{-var-set-update-range} command, is
31676 If the varobj's type changed, then this field will be present and will
31679 @item new_num_children
31680 For a dynamic varobj, if the number of children changed, or if the
31681 type changed, this will be the new number of children.
31683 The @samp{numchild} field in other varobj responses is generally not
31684 valid for a dynamic varobj -- it will show the number of children that
31685 @value{GDBN} knows about, but because dynamic varobjs lazily
31686 instantiate their children, this will not reflect the number of
31687 children which may be available.
31689 The @samp{new_num_children} attribute only reports changes to the
31690 number of children known by @value{GDBN}. This is the only way to
31691 detect whether an update has removed children (which necessarily can
31692 only happen at the end of the update range).
31695 The display hint, if any.
31698 This is an integer value, which will be 1 if there are more children
31699 available outside the varobj's update range.
31702 This attribute will be present and have the value @samp{1} if the
31703 varobj is a dynamic varobj. If the varobj is not a dynamic varobj,
31704 then this attribute will not be present.
31707 If new children were added to a dynamic varobj within the selected
31708 update range (as set by @code{-var-set-update-range}), then they will
31709 be listed in this attribute.
31712 @subsubheading Example
31719 -var-update --all-values var1
31720 ^done,changelist=[@{name="var1",value="3",in_scope="true",
31721 type_changed="false"@}]
31725 @subheading The @code{-var-set-frozen} Command
31726 @findex -var-set-frozen
31727 @anchor{-var-set-frozen}
31729 @subsubheading Synopsis
31732 -var-set-frozen @var{name} @var{flag}
31735 Set the frozenness flag on the variable object @var{name}. The
31736 @var{flag} parameter should be either @samp{1} to make the variable
31737 frozen or @samp{0} to make it unfrozen. If a variable object is
31738 frozen, then neither itself, nor any of its children, are
31739 implicitly updated by @code{-var-update} of
31740 a parent variable or by @code{-var-update *}. Only
31741 @code{-var-update} of the variable itself will update its value and
31742 values of its children. After a variable object is unfrozen, it is
31743 implicitly updated by all subsequent @code{-var-update} operations.
31744 Unfreezing a variable does not update it, only subsequent
31745 @code{-var-update} does.
31747 @subsubheading Example
31751 -var-set-frozen V 1
31756 @subheading The @code{-var-set-update-range} command
31757 @findex -var-set-update-range
31758 @anchor{-var-set-update-range}
31760 @subsubheading Synopsis
31763 -var-set-update-range @var{name} @var{from} @var{to}
31766 Set the range of children to be returned by future invocations of
31767 @code{-var-update}.
31769 @var{from} and @var{to} indicate the range of children to report. If
31770 @var{from} or @var{to} is less than zero, the range is reset and all
31771 children will be reported. Otherwise, children starting at @var{from}
31772 (zero-based) and up to and excluding @var{to} will be reported.
31774 @subsubheading Example
31778 -var-set-update-range V 1 2
31782 @subheading The @code{-var-set-visualizer} command
31783 @findex -var-set-visualizer
31784 @anchor{-var-set-visualizer}
31786 @subsubheading Synopsis
31789 -var-set-visualizer @var{name} @var{visualizer}
31792 Set a visualizer for the variable object @var{name}.
31794 @var{visualizer} is the visualizer to use. The special value
31795 @samp{None} means to disable any visualizer in use.
31797 If not @samp{None}, @var{visualizer} must be a Python expression.
31798 This expression must evaluate to a callable object which accepts a
31799 single argument. @value{GDBN} will call this object with the value of
31800 the varobj @var{name} as an argument (this is done so that the same
31801 Python pretty-printing code can be used for both the CLI and MI).
31802 When called, this object must return an object which conforms to the
31803 pretty-printing interface (@pxref{Pretty Printing API}).
31805 The pre-defined function @code{gdb.default_visualizer} may be used to
31806 select a visualizer by following the built-in process
31807 (@pxref{Selecting Pretty-Printers}). This is done automatically when
31808 a varobj is created, and so ordinarily is not needed.
31810 This feature is only available if Python support is enabled. The MI
31811 command @code{-list-features} (@pxref{GDB/MI Support Commands})
31812 can be used to check this.
31814 @subsubheading Example
31816 Resetting the visualizer:
31820 -var-set-visualizer V None
31824 Reselecting the default (type-based) visualizer:
31828 -var-set-visualizer V gdb.default_visualizer
31832 Suppose @code{SomeClass} is a visualizer class. A lambda expression
31833 can be used to instantiate this class for a varobj:
31837 -var-set-visualizer V "lambda val: SomeClass()"
31841 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
31842 @node GDB/MI Data Manipulation
31843 @section @sc{gdb/mi} Data Manipulation
31845 @cindex data manipulation, in @sc{gdb/mi}
31846 @cindex @sc{gdb/mi}, data manipulation
31847 This section describes the @sc{gdb/mi} commands that manipulate data:
31848 examine memory and registers, evaluate expressions, etc.
31850 For details about what an addressable memory unit is,
31851 @pxref{addressable memory unit}.
31853 @c REMOVED FROM THE INTERFACE.
31854 @c @subheading -data-assign
31855 @c Change the value of a program variable. Plenty of side effects.
31856 @c @subsubheading GDB Command
31858 @c @subsubheading Example
31861 @subheading The @code{-data-disassemble} Command
31862 @findex -data-disassemble
31864 @subsubheading Synopsis
31868 [ -s @var{start-addr} -e @var{end-addr} ]
31869 | [ -a @var{addr} ]
31870 | [ -f @var{filename} -l @var{linenum} [ -n @var{lines} ] ]
31878 @item @var{start-addr}
31879 is the beginning address (or @code{$pc})
31880 @item @var{end-addr}
31883 is an address anywhere within (or the name of) the function to
31884 disassemble. If an address is specified, the whole function
31885 surrounding that address will be disassembled. If a name is
31886 specified, the whole function with that name will be disassembled.
31887 @item @var{filename}
31888 is the name of the file to disassemble
31889 @item @var{linenum}
31890 is the line number to disassemble around
31892 is the number of disassembly lines to be produced. If it is -1,
31893 the whole function will be disassembled, in case no @var{end-addr} is
31894 specified. If @var{end-addr} is specified as a non-zero value, and
31895 @var{lines} is lower than the number of disassembly lines between
31896 @var{start-addr} and @var{end-addr}, only @var{lines} lines are
31897 displayed; if @var{lines} is higher than the number of lines between
31898 @var{start-addr} and @var{end-addr}, only the lines up to @var{end-addr}
31903 @item 0 disassembly only
31904 @item 1 mixed source and disassembly (deprecated)
31905 @item 2 disassembly with raw opcodes
31906 @item 3 mixed source and disassembly with raw opcodes (deprecated)
31907 @item 4 mixed source and disassembly
31908 @item 5 mixed source and disassembly with raw opcodes
31911 Modes 1 and 3 are deprecated. The output is ``source centric''
31912 which hasn't proved useful in practice.
31913 @xref{Machine Code}, for a discussion of the difference between
31914 @code{/m} and @code{/s} output of the @code{disassemble} command.
31917 @subsubheading Result
31919 The result of the @code{-data-disassemble} command will be a list named
31920 @samp{asm_insns}, the contents of this list depend on the @var{mode}
31921 used with the @code{-data-disassemble} command.
31923 For modes 0 and 2 the @samp{asm_insns} list contains tuples with the
31928 The address at which this instruction was disassembled.
31931 The name of the function this instruction is within.
31934 The decimal offset in bytes from the start of @samp{func-name}.
31937 The text disassembly for this @samp{address}.
31940 This field is only present for modes 2, 3 and 5. This contains the raw opcode
31941 bytes for the @samp{inst} field.
31945 For modes 1, 3, 4 and 5 the @samp{asm_insns} list contains tuples named
31946 @samp{src_and_asm_line}, each of which has the following fields:
31950 The line number within @samp{file}.
31953 The file name from the compilation unit. This might be an absolute
31954 file name or a relative file name depending on the compile command
31958 Absolute file name of @samp{file}. It is converted to a canonical form
31959 using the source file search path
31960 (@pxref{Source Path, ,Specifying Source Directories})
31961 and after resolving all the symbolic links.
31963 If the source file is not found this field will contain the path as
31964 present in the debug information.
31966 @item line_asm_insn
31967 This is a list of tuples containing the disassembly for @samp{line} in
31968 @samp{file}. The fields of each tuple are the same as for
31969 @code{-data-disassemble} in @var{mode} 0 and 2, so @samp{address},
31970 @samp{func-name}, @samp{offset}, @samp{inst}, and optionally
31975 Note that whatever included in the @samp{inst} field, is not
31976 manipulated directly by @sc{gdb/mi}, i.e., it is not possible to
31979 @subsubheading @value{GDBN} Command
31981 The corresponding @value{GDBN} command is @samp{disassemble}.
31983 @subsubheading Example
31985 Disassemble from the current value of @code{$pc} to @code{$pc + 20}:
31989 -data-disassemble -s $pc -e "$pc + 20" -- 0
31992 @{address="0x000107c0",func-name="main",offset="4",
31993 inst="mov 2, %o0"@},
31994 @{address="0x000107c4",func-name="main",offset="8",
31995 inst="sethi %hi(0x11800), %o2"@},
31996 @{address="0x000107c8",func-name="main",offset="12",
31997 inst="or %o2, 0x140, %o1\t! 0x11940 <_lib_version+8>"@},
31998 @{address="0x000107cc",func-name="main",offset="16",
31999 inst="sethi %hi(0x11800), %o2"@},
32000 @{address="0x000107d0",func-name="main",offset="20",
32001 inst="or %o2, 0x168, %o4\t! 0x11968 <_lib_version+48>"@}]
32005 Disassemble the whole @code{main} function. Line 32 is part of
32009 -data-disassemble -f basics.c -l 32 -- 0
32011 @{address="0x000107bc",func-name="main",offset="0",
32012 inst="save %sp, -112, %sp"@},
32013 @{address="0x000107c0",func-name="main",offset="4",
32014 inst="mov 2, %o0"@},
32015 @{address="0x000107c4",func-name="main",offset="8",
32016 inst="sethi %hi(0x11800), %o2"@},
32018 @{address="0x0001081c",func-name="main",offset="96",inst="ret "@},
32019 @{address="0x00010820",func-name="main",offset="100",inst="restore "@}]
32023 Disassemble 3 instructions from the start of @code{main}:
32027 -data-disassemble -f basics.c -l 32 -n 3 -- 0
32029 @{address="0x000107bc",func-name="main",offset="0",
32030 inst="save %sp, -112, %sp"@},
32031 @{address="0x000107c0",func-name="main",offset="4",
32032 inst="mov 2, %o0"@},
32033 @{address="0x000107c4",func-name="main",offset="8",
32034 inst="sethi %hi(0x11800), %o2"@}]
32038 Disassemble 3 instructions from the start of @code{main} in mixed mode:
32042 -data-disassemble -f basics.c -l 32 -n 3 -- 1
32044 src_and_asm_line=@{line="31",
32045 file="../../../src/gdb/testsuite/gdb.mi/basics.c",
32046 fullname="/absolute/path/to/src/gdb/testsuite/gdb.mi/basics.c",
32047 line_asm_insn=[@{address="0x000107bc",
32048 func-name="main",offset="0",inst="save %sp, -112, %sp"@}]@},
32049 src_and_asm_line=@{line="32",
32050 file="../../../src/gdb/testsuite/gdb.mi/basics.c",
32051 fullname="/absolute/path/to/src/gdb/testsuite/gdb.mi/basics.c",
32052 line_asm_insn=[@{address="0x000107c0",
32053 func-name="main",offset="4",inst="mov 2, %o0"@},
32054 @{address="0x000107c4",func-name="main",offset="8",
32055 inst="sethi %hi(0x11800), %o2"@}]@}]
32060 @subheading The @code{-data-evaluate-expression} Command
32061 @findex -data-evaluate-expression
32063 @subsubheading Synopsis
32066 -data-evaluate-expression @var{expr}
32069 Evaluate @var{expr} as an expression. The expression could contain an
32070 inferior function call. The function call will execute synchronously.
32071 If the expression contains spaces, it must be enclosed in double quotes.
32073 @subsubheading @value{GDBN} Command
32075 The corresponding @value{GDBN} commands are @samp{print}, @samp{output}, and
32076 @samp{call}. In @code{gdbtk} only, there's a corresponding
32077 @samp{gdb_eval} command.
32079 @subsubheading Example
32081 In the following example, the numbers that precede the commands are the
32082 @dfn{tokens} described in @ref{GDB/MI Command Syntax, ,@sc{gdb/mi}
32083 Command Syntax}. Notice how @sc{gdb/mi} returns the same tokens in its
32087 211-data-evaluate-expression A
32090 311-data-evaluate-expression &A
32091 311^done,value="0xefffeb7c"
32093 411-data-evaluate-expression A+3
32096 511-data-evaluate-expression "A + 3"
32102 @subheading The @code{-data-list-changed-registers} Command
32103 @findex -data-list-changed-registers
32105 @subsubheading Synopsis
32108 -data-list-changed-registers
32111 Display a list of the registers that have changed.
32113 @subsubheading @value{GDBN} Command
32115 @value{GDBN} doesn't have a direct analog for this command; @code{gdbtk}
32116 has the corresponding command @samp{gdb_changed_register_list}.
32118 @subsubheading Example
32120 On a PPC MBX board:
32128 *stopped,reason="breakpoint-hit",disp="keep",bkptno="1",frame=@{
32129 func="main",args=[],file="try.c",fullname="/home/foo/bar/try.c",
32130 line="5",arch="powerpc"@}
32132 -data-list-changed-registers
32133 ^done,changed-registers=["0","1","2","4","5","6","7","8","9",
32134 "10","11","13","14","15","16","17","18","19","20","21","22","23",
32135 "24","25","26","27","28","30","31","64","65","66","67","69"]
32140 @subheading The @code{-data-list-register-names} Command
32141 @findex -data-list-register-names
32143 @subsubheading Synopsis
32146 -data-list-register-names [ ( @var{regno} )+ ]
32149 Show a list of register names for the current target. If no arguments
32150 are given, it shows a list of the names of all the registers. If
32151 integer numbers are given as arguments, it will print a list of the
32152 names of the registers corresponding to the arguments. To ensure
32153 consistency between a register name and its number, the output list may
32154 include empty register names.
32156 @subsubheading @value{GDBN} Command
32158 @value{GDBN} does not have a command which corresponds to
32159 @samp{-data-list-register-names}. In @code{gdbtk} there is a
32160 corresponding command @samp{gdb_regnames}.
32162 @subsubheading Example
32164 For the PPC MBX board:
32167 -data-list-register-names
32168 ^done,register-names=["r0","r1","r2","r3","r4","r5","r6","r7",
32169 "r8","r9","r10","r11","r12","r13","r14","r15","r16","r17","r18",
32170 "r19","r20","r21","r22","r23","r24","r25","r26","r27","r28","r29",
32171 "r30","r31","f0","f1","f2","f3","f4","f5","f6","f7","f8","f9",
32172 "f10","f11","f12","f13","f14","f15","f16","f17","f18","f19","f20",
32173 "f21","f22","f23","f24","f25","f26","f27","f28","f29","f30","f31",
32174 "", "pc","ps","cr","lr","ctr","xer"]
32176 -data-list-register-names 1 2 3
32177 ^done,register-names=["r1","r2","r3"]
32181 @subheading The @code{-data-list-register-values} Command
32182 @findex -data-list-register-values
32184 @subsubheading Synopsis
32187 -data-list-register-values
32188 [ @code{--skip-unavailable} ] @var{fmt} [ ( @var{regno} )*]
32191 Display the registers' contents. The format according to which the
32192 registers' contents are to be returned is given by @var{fmt}, followed
32193 by an optional list of numbers specifying the registers to display. A
32194 missing list of numbers indicates that the contents of all the
32195 registers must be returned. The @code{--skip-unavailable} option
32196 indicates that only the available registers are to be returned.
32198 Allowed formats for @var{fmt} are:
32215 @subsubheading @value{GDBN} Command
32217 The corresponding @value{GDBN} commands are @samp{info reg}, @samp{info
32218 all-reg}, and (in @code{gdbtk}) @samp{gdb_fetch_registers}.
32220 @subsubheading Example
32222 For a PPC MBX board (note: line breaks are for readability only, they
32223 don't appear in the actual output):
32227 -data-list-register-values r 64 65
32228 ^done,register-values=[@{number="64",value="0xfe00a300"@},
32229 @{number="65",value="0x00029002"@}]
32231 -data-list-register-values x
32232 ^done,register-values=[@{number="0",value="0xfe0043c8"@},
32233 @{number="1",value="0x3fff88"@},@{number="2",value="0xfffffffe"@},
32234 @{number="3",value="0x0"@},@{number="4",value="0xa"@},
32235 @{number="5",value="0x3fff68"@},@{number="6",value="0x3fff58"@},
32236 @{number="7",value="0xfe011e98"@},@{number="8",value="0x2"@},
32237 @{number="9",value="0xfa202820"@},@{number="10",value="0xfa202808"@},
32238 @{number="11",value="0x1"@},@{number="12",value="0x0"@},
32239 @{number="13",value="0x4544"@},@{number="14",value="0xffdfffff"@},
32240 @{number="15",value="0xffffffff"@},@{number="16",value="0xfffffeff"@},
32241 @{number="17",value="0xefffffed"@},@{number="18",value="0xfffffffe"@},
32242 @{number="19",value="0xffffffff"@},@{number="20",value="0xffffffff"@},
32243 @{number="21",value="0xffffffff"@},@{number="22",value="0xfffffff7"@},
32244 @{number="23",value="0xffffffff"@},@{number="24",value="0xffffffff"@},
32245 @{number="25",value="0xffffffff"@},@{number="26",value="0xfffffffb"@},
32246 @{number="27",value="0xffffffff"@},@{number="28",value="0xf7bfffff"@},
32247 @{number="29",value="0x0"@},@{number="30",value="0xfe010000"@},
32248 @{number="31",value="0x0"@},@{number="32",value="0x0"@},
32249 @{number="33",value="0x0"@},@{number="34",value="0x0"@},
32250 @{number="35",value="0x0"@},@{number="36",value="0x0"@},
32251 @{number="37",value="0x0"@},@{number="38",value="0x0"@},
32252 @{number="39",value="0x0"@},@{number="40",value="0x0"@},
32253 @{number="41",value="0x0"@},@{number="42",value="0x0"@},
32254 @{number="43",value="0x0"@},@{number="44",value="0x0"@},
32255 @{number="45",value="0x0"@},@{number="46",value="0x0"@},
32256 @{number="47",value="0x0"@},@{number="48",value="0x0"@},
32257 @{number="49",value="0x0"@},@{number="50",value="0x0"@},
32258 @{number="51",value="0x0"@},@{number="52",value="0x0"@},
32259 @{number="53",value="0x0"@},@{number="54",value="0x0"@},
32260 @{number="55",value="0x0"@},@{number="56",value="0x0"@},
32261 @{number="57",value="0x0"@},@{number="58",value="0x0"@},
32262 @{number="59",value="0x0"@},@{number="60",value="0x0"@},
32263 @{number="61",value="0x0"@},@{number="62",value="0x0"@},
32264 @{number="63",value="0x0"@},@{number="64",value="0xfe00a300"@},
32265 @{number="65",value="0x29002"@},@{number="66",value="0x202f04b5"@},
32266 @{number="67",value="0xfe0043b0"@},@{number="68",value="0xfe00b3e4"@},
32267 @{number="69",value="0x20002b03"@}]
32272 @subheading The @code{-data-read-memory} Command
32273 @findex -data-read-memory
32275 This command is deprecated, use @code{-data-read-memory-bytes} instead.
32277 @subsubheading Synopsis
32280 -data-read-memory [ -o @var{byte-offset} ]
32281 @var{address} @var{word-format} @var{word-size}
32282 @var{nr-rows} @var{nr-cols} [ @var{aschar} ]
32289 @item @var{address}
32290 An expression specifying the address of the first memory word to be
32291 read. Complex expressions containing embedded white space should be
32292 quoted using the C convention.
32294 @item @var{word-format}
32295 The format to be used to print the memory words. The notation is the
32296 same as for @value{GDBN}'s @code{print} command (@pxref{Output Formats,
32299 @item @var{word-size}
32300 The size of each memory word in bytes.
32302 @item @var{nr-rows}
32303 The number of rows in the output table.
32305 @item @var{nr-cols}
32306 The number of columns in the output table.
32309 If present, indicates that each row should include an @sc{ascii} dump. The
32310 value of @var{aschar} is used as a padding character when a byte is not a
32311 member of the printable @sc{ascii} character set (printable @sc{ascii}
32312 characters are those whose code is between 32 and 126, inclusively).
32314 @item @var{byte-offset}
32315 An offset to add to the @var{address} before fetching memory.
32318 This command displays memory contents as a table of @var{nr-rows} by
32319 @var{nr-cols} words, each word being @var{word-size} bytes. In total,
32320 @code{@var{nr-rows} * @var{nr-cols} * @var{word-size}} bytes are read
32321 (returned as @samp{total-bytes}). Should less than the requested number
32322 of bytes be returned by the target, the missing words are identified
32323 using @samp{N/A}. The number of bytes read from the target is returned
32324 in @samp{nr-bytes} and the starting address used to read memory in
32327 The address of the next/previous row or page is available in
32328 @samp{next-row} and @samp{prev-row}, @samp{next-page} and
32331 @subsubheading @value{GDBN} Command
32333 The corresponding @value{GDBN} command is @samp{x}. @code{gdbtk} has
32334 @samp{gdb_get_mem} memory read command.
32336 @subsubheading Example
32338 Read six bytes of memory starting at @code{bytes+6} but then offset by
32339 @code{-6} bytes. Format as three rows of two columns. One byte per
32340 word. Display each word in hex.
32344 9-data-read-memory -o -6 -- bytes+6 x 1 3 2
32345 9^done,addr="0x00001390",nr-bytes="6",total-bytes="6",
32346 next-row="0x00001396",prev-row="0x0000138e",next-page="0x00001396",
32347 prev-page="0x0000138a",memory=[
32348 @{addr="0x00001390",data=["0x00","0x01"]@},
32349 @{addr="0x00001392",data=["0x02","0x03"]@},
32350 @{addr="0x00001394",data=["0x04","0x05"]@}]
32354 Read two bytes of memory starting at address @code{shorts + 64} and
32355 display as a single word formatted in decimal.
32359 5-data-read-memory shorts+64 d 2 1 1
32360 5^done,addr="0x00001510",nr-bytes="2",total-bytes="2",
32361 next-row="0x00001512",prev-row="0x0000150e",
32362 next-page="0x00001512",prev-page="0x0000150e",memory=[
32363 @{addr="0x00001510",data=["128"]@}]
32367 Read thirty two bytes of memory starting at @code{bytes+16} and format
32368 as eight rows of four columns. Include a string encoding with @samp{x}
32369 used as the non-printable character.
32373 4-data-read-memory bytes+16 x 1 8 4 x
32374 4^done,addr="0x000013a0",nr-bytes="32",total-bytes="32",
32375 next-row="0x000013c0",prev-row="0x0000139c",
32376 next-page="0x000013c0",prev-page="0x00001380",memory=[
32377 @{addr="0x000013a0",data=["0x10","0x11","0x12","0x13"],ascii="xxxx"@},
32378 @{addr="0x000013a4",data=["0x14","0x15","0x16","0x17"],ascii="xxxx"@},
32379 @{addr="0x000013a8",data=["0x18","0x19","0x1a","0x1b"],ascii="xxxx"@},
32380 @{addr="0x000013ac",data=["0x1c","0x1d","0x1e","0x1f"],ascii="xxxx"@},
32381 @{addr="0x000013b0",data=["0x20","0x21","0x22","0x23"],ascii=" !\"#"@},
32382 @{addr="0x000013b4",data=["0x24","0x25","0x26","0x27"],ascii="$%&'"@},
32383 @{addr="0x000013b8",data=["0x28","0x29","0x2a","0x2b"],ascii="()*+"@},
32384 @{addr="0x000013bc",data=["0x2c","0x2d","0x2e","0x2f"],ascii=",-./"@}]
32388 @subheading The @code{-data-read-memory-bytes} Command
32389 @findex -data-read-memory-bytes
32391 @subsubheading Synopsis
32394 -data-read-memory-bytes [ -o @var{offset} ]
32395 @var{address} @var{count}
32402 @item @var{address}
32403 An expression specifying the address of the first addressable memory unit
32404 to be read. Complex expressions containing embedded white space should be
32405 quoted using the C convention.
32408 The number of addressable memory units to read. This should be an integer
32412 The offset relative to @var{address} at which to start reading. This
32413 should be an integer literal. This option is provided so that a frontend
32414 is not required to first evaluate address and then perform address
32415 arithmetics itself.
32419 This command attempts to read all accessible memory regions in the
32420 specified range. First, all regions marked as unreadable in the memory
32421 map (if one is defined) will be skipped. @xref{Memory Region
32422 Attributes}. Second, @value{GDBN} will attempt to read the remaining
32423 regions. For each one, if reading full region results in an errors,
32424 @value{GDBN} will try to read a subset of the region.
32426 In general, every single memory unit in the region may be readable or not,
32427 and the only way to read every readable unit is to try a read at
32428 every address, which is not practical. Therefore, @value{GDBN} will
32429 attempt to read all accessible memory units at either beginning or the end
32430 of the region, using a binary division scheme. This heuristic works
32431 well for reading accross a memory map boundary. Note that if a region
32432 has a readable range that is neither at the beginning or the end,
32433 @value{GDBN} will not read it.
32435 The result record (@pxref{GDB/MI Result Records}) that is output of
32436 the command includes a field named @samp{memory} whose content is a
32437 list of tuples. Each tuple represent a successfully read memory block
32438 and has the following fields:
32442 The start address of the memory block, as hexadecimal literal.
32445 The end address of the memory block, as hexadecimal literal.
32448 The offset of the memory block, as hexadecimal literal, relative to
32449 the start address passed to @code{-data-read-memory-bytes}.
32452 The contents of the memory block, in hex.
32458 @subsubheading @value{GDBN} Command
32460 The corresponding @value{GDBN} command is @samp{x}.
32462 @subsubheading Example
32466 -data-read-memory-bytes &a 10
32467 ^done,memory=[@{begin="0xbffff154",offset="0x00000000",
32469 contents="01000000020000000300"@}]
32474 @subheading The @code{-data-write-memory-bytes} Command
32475 @findex -data-write-memory-bytes
32477 @subsubheading Synopsis
32480 -data-write-memory-bytes @var{address} @var{contents}
32481 -data-write-memory-bytes @var{address} @var{contents} @r{[}@var{count}@r{]}
32488 @item @var{address}
32489 An expression specifying the address of the first addressable memory unit
32490 to be written. Complex expressions containing embedded white space should
32491 be quoted using the C convention.
32493 @item @var{contents}
32494 The hex-encoded data to write. It is an error if @var{contents} does
32495 not represent an integral number of addressable memory units.
32498 Optional argument indicating the number of addressable memory units to be
32499 written. If @var{count} is greater than @var{contents}' length,
32500 @value{GDBN} will repeatedly write @var{contents} until it fills
32501 @var{count} memory units.
32505 @subsubheading @value{GDBN} Command
32507 There's no corresponding @value{GDBN} command.
32509 @subsubheading Example
32513 -data-write-memory-bytes &a "aabbccdd"
32520 -data-write-memory-bytes &a "aabbccdd" 16e
32525 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
32526 @node GDB/MI Tracepoint Commands
32527 @section @sc{gdb/mi} Tracepoint Commands
32529 The commands defined in this section implement MI support for
32530 tracepoints. For detailed introduction, see @ref{Tracepoints}.
32532 @subheading The @code{-trace-find} Command
32533 @findex -trace-find
32535 @subsubheading Synopsis
32538 -trace-find @var{mode} [@var{parameters}@dots{}]
32541 Find a trace frame using criteria defined by @var{mode} and
32542 @var{parameters}. The following table lists permissible
32543 modes and their parameters. For details of operation, see @ref{tfind}.
32548 No parameters are required. Stops examining trace frames.
32551 An integer is required as parameter. Selects tracepoint frame with
32554 @item tracepoint-number
32555 An integer is required as parameter. Finds next
32556 trace frame that corresponds to tracepoint with the specified number.
32559 An address is required as parameter. Finds
32560 next trace frame that corresponds to any tracepoint at the specified
32563 @item pc-inside-range
32564 Two addresses are required as parameters. Finds next trace
32565 frame that corresponds to a tracepoint at an address inside the
32566 specified range. Both bounds are considered to be inside the range.
32568 @item pc-outside-range
32569 Two addresses are required as parameters. Finds
32570 next trace frame that corresponds to a tracepoint at an address outside
32571 the specified range. Both bounds are considered to be inside the range.
32574 Line specification is required as parameter. @xref{Specify Location}.
32575 Finds next trace frame that corresponds to a tracepoint at
32576 the specified location.
32580 If @samp{none} was passed as @var{mode}, the response does not
32581 have fields. Otherwise, the response may have the following fields:
32585 This field has either @samp{0} or @samp{1} as the value, depending
32586 on whether a matching tracepoint was found.
32589 The index of the found traceframe. This field is present iff
32590 the @samp{found} field has value of @samp{1}.
32593 The index of the found tracepoint. This field is present iff
32594 the @samp{found} field has value of @samp{1}.
32597 The information about the frame corresponding to the found trace
32598 frame. This field is present only if a trace frame was found.
32599 @xref{GDB/MI Frame Information}, for description of this field.
32603 @subsubheading @value{GDBN} Command
32605 The corresponding @value{GDBN} command is @samp{tfind}.
32607 @subheading -trace-define-variable
32608 @findex -trace-define-variable
32610 @subsubheading Synopsis
32613 -trace-define-variable @var{name} [ @var{value} ]
32616 Create trace variable @var{name} if it does not exist. If
32617 @var{value} is specified, sets the initial value of the specified
32618 trace variable to that value. Note that the @var{name} should start
32619 with the @samp{$} character.
32621 @subsubheading @value{GDBN} Command
32623 The corresponding @value{GDBN} command is @samp{tvariable}.
32625 @subheading The @code{-trace-frame-collected} Command
32626 @findex -trace-frame-collected
32628 @subsubheading Synopsis
32631 -trace-frame-collected
32632 [--var-print-values @var{var_pval}]
32633 [--comp-print-values @var{comp_pval}]
32634 [--registers-format @var{regformat}]
32635 [--memory-contents]
32638 This command returns the set of collected objects, register names,
32639 trace state variable names, memory ranges and computed expressions
32640 that have been collected at a particular trace frame. The optional
32641 parameters to the command affect the output format in different ways.
32642 See the output description table below for more details.
32644 The reported names can be used in the normal manner to create
32645 varobjs and inspect the objects themselves. The items returned by
32646 this command are categorized so that it is clear which is a variable,
32647 which is a register, which is a trace state variable, which is a
32648 memory range and which is a computed expression.
32650 For instance, if the actions were
32652 collect myVar, myArray[myIndex], myObj.field, myPtr->field, myCount + 2
32653 collect *(int*)0xaf02bef0@@40
32657 the object collected in its entirety would be @code{myVar}. The
32658 object @code{myArray} would be partially collected, because only the
32659 element at index @code{myIndex} would be collected. The remaining
32660 objects would be computed expressions.
32662 An example output would be:
32666 -trace-frame-collected
32668 explicit-variables=[@{name="myVar",value="1"@}],
32669 computed-expressions=[@{name="myArray[myIndex]",value="0"@},
32670 @{name="myObj.field",value="0"@},
32671 @{name="myPtr->field",value="1"@},
32672 @{name="myCount + 2",value="3"@},
32673 @{name="$tvar1 + 1",value="43970027"@}],
32674 registers=[@{number="0",value="0x7fe2c6e79ec8"@},
32675 @{number="1",value="0x0"@},
32676 @{number="2",value="0x4"@},
32678 @{number="125",value="0x0"@}],
32679 tvars=[@{name="$tvar1",current="43970026"@}],
32680 memory=[@{address="0x0000000000602264",length="4"@},
32681 @{address="0x0000000000615bc0",length="4"@}]
32688 @item explicit-variables
32689 The set of objects that have been collected in their entirety (as
32690 opposed to collecting just a few elements of an array or a few struct
32691 members). For each object, its name and value are printed.
32692 The @code{--var-print-values} option affects how or whether the value
32693 field is output. If @var{var_pval} is 0, then print only the names;
32694 if it is 1, print also their values; and if it is 2, print the name,
32695 type and value for simple data types, and the name and type for
32696 arrays, structures and unions.
32698 @item computed-expressions
32699 The set of computed expressions that have been collected at the
32700 current trace frame. The @code{--comp-print-values} option affects
32701 this set like the @code{--var-print-values} option affects the
32702 @code{explicit-variables} set. See above.
32705 The registers that have been collected at the current trace frame.
32706 For each register collected, the name and current value are returned.
32707 The value is formatted according to the @code{--registers-format}
32708 option. See the @command{-data-list-register-values} command for a
32709 list of the allowed formats. The default is @samp{x}.
32712 The trace state variables that have been collected at the current
32713 trace frame. For each trace state variable collected, the name and
32714 current value are returned.
32717 The set of memory ranges that have been collected at the current trace
32718 frame. Its content is a list of tuples. Each tuple represents a
32719 collected memory range and has the following fields:
32723 The start address of the memory range, as hexadecimal literal.
32726 The length of the memory range, as decimal literal.
32729 The contents of the memory block, in hex. This field is only present
32730 if the @code{--memory-contents} option is specified.
32736 @subsubheading @value{GDBN} Command
32738 There is no corresponding @value{GDBN} command.
32740 @subsubheading Example
32742 @subheading -trace-list-variables
32743 @findex -trace-list-variables
32745 @subsubheading Synopsis
32748 -trace-list-variables
32751 Return a table of all defined trace variables. Each element of the
32752 table has the following fields:
32756 The name of the trace variable. This field is always present.
32759 The initial value. This is a 64-bit signed integer. This
32760 field is always present.
32763 The value the trace variable has at the moment. This is a 64-bit
32764 signed integer. This field is absent iff current value is
32765 not defined, for example if the trace was never run, or is
32770 @subsubheading @value{GDBN} Command
32772 The corresponding @value{GDBN} command is @samp{tvariables}.
32774 @subsubheading Example
32778 -trace-list-variables
32779 ^done,trace-variables=@{nr_rows="1",nr_cols="3",
32780 hdr=[@{width="15",alignment="-1",col_name="name",colhdr="Name"@},
32781 @{width="11",alignment="-1",col_name="initial",colhdr="Initial"@},
32782 @{width="11",alignment="-1",col_name="current",colhdr="Current"@}],
32783 body=[variable=@{name="$trace_timestamp",initial="0"@}
32784 variable=@{name="$foo",initial="10",current="15"@}]@}
32788 @subheading -trace-save
32789 @findex -trace-save
32791 @subsubheading Synopsis
32794 -trace-save [ -r ] [ -ctf ] @var{filename}
32797 Saves the collected trace data to @var{filename}. Without the
32798 @samp{-r} option, the data is downloaded from the target and saved
32799 in a local file. With the @samp{-r} option the target is asked
32800 to perform the save.
32802 By default, this command will save the trace in the tfile format. You can
32803 supply the optional @samp{-ctf} argument to save it the CTF format. See
32804 @ref{Trace Files} for more information about CTF.
32806 @subsubheading @value{GDBN} Command
32808 The corresponding @value{GDBN} command is @samp{tsave}.
32811 @subheading -trace-start
32812 @findex -trace-start
32814 @subsubheading Synopsis
32820 Starts a tracing experiment. The result of this command does not
32823 @subsubheading @value{GDBN} Command
32825 The corresponding @value{GDBN} command is @samp{tstart}.
32827 @subheading -trace-status
32828 @findex -trace-status
32830 @subsubheading Synopsis
32836 Obtains the status of a tracing experiment. The result may include
32837 the following fields:
32842 May have a value of either @samp{0}, when no tracing operations are
32843 supported, @samp{1}, when all tracing operations are supported, or
32844 @samp{file} when examining trace file. In the latter case, examining
32845 of trace frame is possible but new tracing experiement cannot be
32846 started. This field is always present.
32849 May have a value of either @samp{0} or @samp{1} depending on whether
32850 tracing experiement is in progress on target. This field is present
32851 if @samp{supported} field is not @samp{0}.
32854 Report the reason why the tracing was stopped last time. This field
32855 may be absent iff tracing was never stopped on target yet. The
32856 value of @samp{request} means the tracing was stopped as result of
32857 the @code{-trace-stop} command. The value of @samp{overflow} means
32858 the tracing buffer is full. The value of @samp{disconnection} means
32859 tracing was automatically stopped when @value{GDBN} has disconnected.
32860 The value of @samp{passcount} means tracing was stopped when a
32861 tracepoint was passed a maximal number of times for that tracepoint.
32862 This field is present if @samp{supported} field is not @samp{0}.
32864 @item stopping-tracepoint
32865 The number of tracepoint whose passcount as exceeded. This field is
32866 present iff the @samp{stop-reason} field has the value of
32870 @itemx frames-created
32871 The @samp{frames} field is a count of the total number of trace frames
32872 in the trace buffer, while @samp{frames-created} is the total created
32873 during the run, including ones that were discarded, such as when a
32874 circular trace buffer filled up. Both fields are optional.
32878 These fields tell the current size of the tracing buffer and the
32879 remaining space. These fields are optional.
32882 The value of the circular trace buffer flag. @code{1} means that the
32883 trace buffer is circular and old trace frames will be discarded if
32884 necessary to make room, @code{0} means that the trace buffer is linear
32888 The value of the disconnected tracing flag. @code{1} means that
32889 tracing will continue after @value{GDBN} disconnects, @code{0} means
32890 that the trace run will stop.
32893 The filename of the trace file being examined. This field is
32894 optional, and only present when examining a trace file.
32898 @subsubheading @value{GDBN} Command
32900 The corresponding @value{GDBN} command is @samp{tstatus}.
32902 @subheading -trace-stop
32903 @findex -trace-stop
32905 @subsubheading Synopsis
32911 Stops a tracing experiment. The result of this command has the same
32912 fields as @code{-trace-status}, except that the @samp{supported} and
32913 @samp{running} fields are not output.
32915 @subsubheading @value{GDBN} Command
32917 The corresponding @value{GDBN} command is @samp{tstop}.
32920 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
32921 @node GDB/MI Symbol Query
32922 @section @sc{gdb/mi} Symbol Query Commands
32926 @subheading The @code{-symbol-info-address} Command
32927 @findex -symbol-info-address
32929 @subsubheading Synopsis
32932 -symbol-info-address @var{symbol}
32935 Describe where @var{symbol} is stored.
32937 @subsubheading @value{GDBN} Command
32939 The corresponding @value{GDBN} command is @samp{info address}.
32941 @subsubheading Example
32945 @subheading The @code{-symbol-info-file} Command
32946 @findex -symbol-info-file
32948 @subsubheading Synopsis
32954 Show the file for the symbol.
32956 @subsubheading @value{GDBN} Command
32958 There's no equivalent @value{GDBN} command. @code{gdbtk} has
32959 @samp{gdb_find_file}.
32961 @subsubheading Example
32965 @subheading The @code{-symbol-info-function} Command
32966 @findex -symbol-info-function
32968 @subsubheading Synopsis
32971 -symbol-info-function
32974 Show which function the symbol lives in.
32976 @subsubheading @value{GDBN} Command
32978 @samp{gdb_get_function} in @code{gdbtk}.
32980 @subsubheading Example
32984 @subheading The @code{-symbol-info-line} Command
32985 @findex -symbol-info-line
32987 @subsubheading Synopsis
32993 Show the core addresses of the code for a source line.
32995 @subsubheading @value{GDBN} Command
32997 The corresponding @value{GDBN} command is @samp{info line}.
32998 @code{gdbtk} has the @samp{gdb_get_line} and @samp{gdb_get_file} commands.
33000 @subsubheading Example
33004 @subheading The @code{-symbol-info-symbol} Command
33005 @findex -symbol-info-symbol
33007 @subsubheading Synopsis
33010 -symbol-info-symbol @var{addr}
33013 Describe what symbol is at location @var{addr}.
33015 @subsubheading @value{GDBN} Command
33017 The corresponding @value{GDBN} command is @samp{info symbol}.
33019 @subsubheading Example
33023 @subheading The @code{-symbol-list-functions} Command
33024 @findex -symbol-list-functions
33026 @subsubheading Synopsis
33029 -symbol-list-functions
33032 List the functions in the executable.
33034 @subsubheading @value{GDBN} Command
33036 @samp{info functions} in @value{GDBN}, @samp{gdb_listfunc} and
33037 @samp{gdb_search} in @code{gdbtk}.
33039 @subsubheading Example
33044 @subheading The @code{-symbol-list-lines} Command
33045 @findex -symbol-list-lines
33047 @subsubheading Synopsis
33050 -symbol-list-lines @var{filename}
33053 Print the list of lines that contain code and their associated program
33054 addresses for the given source filename. The entries are sorted in
33055 ascending PC order.
33057 @subsubheading @value{GDBN} Command
33059 There is no corresponding @value{GDBN} command.
33061 @subsubheading Example
33064 -symbol-list-lines basics.c
33065 ^done,lines=[@{pc="0x08048554",line="7"@},@{pc="0x0804855a",line="8"@}]
33071 @subheading The @code{-symbol-list-types} Command
33072 @findex -symbol-list-types
33074 @subsubheading Synopsis
33080 List all the type names.
33082 @subsubheading @value{GDBN} Command
33084 The corresponding commands are @samp{info types} in @value{GDBN},
33085 @samp{gdb_search} in @code{gdbtk}.
33087 @subsubheading Example
33091 @subheading The @code{-symbol-list-variables} Command
33092 @findex -symbol-list-variables
33094 @subsubheading Synopsis
33097 -symbol-list-variables
33100 List all the global and static variable names.
33102 @subsubheading @value{GDBN} Command
33104 @samp{info variables} in @value{GDBN}, @samp{gdb_search} in @code{gdbtk}.
33106 @subsubheading Example
33110 @subheading The @code{-symbol-locate} Command
33111 @findex -symbol-locate
33113 @subsubheading Synopsis
33119 @subsubheading @value{GDBN} Command
33121 @samp{gdb_loc} in @code{gdbtk}.
33123 @subsubheading Example
33127 @subheading The @code{-symbol-type} Command
33128 @findex -symbol-type
33130 @subsubheading Synopsis
33133 -symbol-type @var{variable}
33136 Show type of @var{variable}.
33138 @subsubheading @value{GDBN} Command
33140 The corresponding @value{GDBN} command is @samp{ptype}, @code{gdbtk} has
33141 @samp{gdb_obj_variable}.
33143 @subsubheading Example
33148 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
33149 @node GDB/MI File Commands
33150 @section @sc{gdb/mi} File Commands
33152 This section describes the GDB/MI commands to specify executable file names
33153 and to read in and obtain symbol table information.
33155 @subheading The @code{-file-exec-and-symbols} Command
33156 @findex -file-exec-and-symbols
33158 @subsubheading Synopsis
33161 -file-exec-and-symbols @var{file}
33164 Specify the executable file to be debugged. This file is the one from
33165 which the symbol table is also read. If no file is specified, the
33166 command clears the executable and symbol information. If breakpoints
33167 are set when using this command with no arguments, @value{GDBN} will produce
33168 error messages. Otherwise, no output is produced, except a completion
33171 @subsubheading @value{GDBN} Command
33173 The corresponding @value{GDBN} command is @samp{file}.
33175 @subsubheading Example
33179 -file-exec-and-symbols /kwikemart/marge/ezannoni/TRUNK/mbx/hello.mbx
33185 @subheading The @code{-file-exec-file} Command
33186 @findex -file-exec-file
33188 @subsubheading Synopsis
33191 -file-exec-file @var{file}
33194 Specify the executable file to be debugged. Unlike
33195 @samp{-file-exec-and-symbols}, the symbol table is @emph{not} read
33196 from this file. If used without argument, @value{GDBN} clears the information
33197 about the executable file. No output is produced, except a completion
33200 @subsubheading @value{GDBN} Command
33202 The corresponding @value{GDBN} command is @samp{exec-file}.
33204 @subsubheading Example
33208 -file-exec-file /kwikemart/marge/ezannoni/TRUNK/mbx/hello.mbx
33215 @subheading The @code{-file-list-exec-sections} Command
33216 @findex -file-list-exec-sections
33218 @subsubheading Synopsis
33221 -file-list-exec-sections
33224 List the sections of the current executable file.
33226 @subsubheading @value{GDBN} Command
33228 The @value{GDBN} command @samp{info file} shows, among the rest, the same
33229 information as this command. @code{gdbtk} has a corresponding command
33230 @samp{gdb_load_info}.
33232 @subsubheading Example
33237 @subheading The @code{-file-list-exec-source-file} Command
33238 @findex -file-list-exec-source-file
33240 @subsubheading Synopsis
33243 -file-list-exec-source-file
33246 List the line number, the current source file, and the absolute path
33247 to the current source file for the current executable. The macro
33248 information field has a value of @samp{1} or @samp{0} depending on
33249 whether or not the file includes preprocessor macro information.
33251 @subsubheading @value{GDBN} Command
33253 The @value{GDBN} equivalent is @samp{info source}
33255 @subsubheading Example
33259 123-file-list-exec-source-file
33260 123^done,line="1",file="foo.c",fullname="/home/bar/foo.c,macro-info="1"
33265 @subheading The @code{-file-list-exec-source-files} Command
33266 @findex -file-list-exec-source-files
33268 @subsubheading Synopsis
33271 -file-list-exec-source-files
33274 List the source files for the current executable.
33276 It will always output both the filename and fullname (absolute file
33277 name) of a source file.
33279 @subsubheading @value{GDBN} Command
33281 The @value{GDBN} equivalent is @samp{info sources}.
33282 @code{gdbtk} has an analogous command @samp{gdb_listfiles}.
33284 @subsubheading Example
33287 -file-list-exec-source-files
33289 @{file=foo.c,fullname=/home/foo.c@},
33290 @{file=/home/bar.c,fullname=/home/bar.c@},
33291 @{file=gdb_could_not_find_fullpath.c@}]
33295 @subheading The @code{-file-list-shared-libraries} Command
33296 @findex -file-list-shared-libraries
33298 @subsubheading Synopsis
33301 -file-list-shared-libraries [ @var{regexp} ]
33304 List the shared libraries in the program.
33305 With a regular expression @var{regexp}, only those libraries whose
33306 names match @var{regexp} are listed.
33308 @subsubheading @value{GDBN} Command
33310 The corresponding @value{GDBN} command is @samp{info shared}. The fields
33311 have a similar meaning to the @code{=library-loaded} notification.
33312 The @code{ranges} field specifies the multiple segments belonging to this
33313 library. Each range has the following fields:
33317 The address defining the inclusive lower bound of the segment.
33319 The address defining the exclusive upper bound of the segment.
33322 @subsubheading Example
33325 -file-list-exec-source-files
33326 ^done,shared-libraries=[
33327 @{id="/lib/libfoo.so",target-name="/lib/libfoo.so",host-name="/lib/libfoo.so",symbols-loaded="1",thread-group="i1",ranges=[@{from="0x72815989",to="0x728162c0"@}]@},
33328 @{id="/lib/libbar.so",target-name="/lib/libbar.so",host-name="/lib/libbar.so",symbols-loaded="1",thread-group="i1",ranges=[@{from="0x76ee48c0",to="0x76ee9160"@}]@}]
33334 @subheading The @code{-file-list-symbol-files} Command
33335 @findex -file-list-symbol-files
33337 @subsubheading Synopsis
33340 -file-list-symbol-files
33345 @subsubheading @value{GDBN} Command
33347 The corresponding @value{GDBN} command is @samp{info file} (part of it).
33349 @subsubheading Example
33354 @subheading The @code{-file-symbol-file} Command
33355 @findex -file-symbol-file
33357 @subsubheading Synopsis
33360 -file-symbol-file @var{file}
33363 Read symbol table info from the specified @var{file} argument. When
33364 used without arguments, clears @value{GDBN}'s symbol table info. No output is
33365 produced, except for a completion notification.
33367 @subsubheading @value{GDBN} Command
33369 The corresponding @value{GDBN} command is @samp{symbol-file}.
33371 @subsubheading Example
33375 -file-symbol-file /kwikemart/marge/ezannoni/TRUNK/mbx/hello.mbx
33381 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
33382 @node GDB/MI Memory Overlay Commands
33383 @section @sc{gdb/mi} Memory Overlay Commands
33385 The memory overlay commands are not implemented.
33387 @c @subheading -overlay-auto
33389 @c @subheading -overlay-list-mapping-state
33391 @c @subheading -overlay-list-overlays
33393 @c @subheading -overlay-map
33395 @c @subheading -overlay-off
33397 @c @subheading -overlay-on
33399 @c @subheading -overlay-unmap
33401 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
33402 @node GDB/MI Signal Handling Commands
33403 @section @sc{gdb/mi} Signal Handling Commands
33405 Signal handling commands are not implemented.
33407 @c @subheading -signal-handle
33409 @c @subheading -signal-list-handle-actions
33411 @c @subheading -signal-list-signal-types
33415 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
33416 @node GDB/MI Target Manipulation
33417 @section @sc{gdb/mi} Target Manipulation Commands
33420 @subheading The @code{-target-attach} Command
33421 @findex -target-attach
33423 @subsubheading Synopsis
33426 -target-attach @var{pid} | @var{gid} | @var{file}
33429 Attach to a process @var{pid} or a file @var{file} outside of
33430 @value{GDBN}, or a thread group @var{gid}. If attaching to a thread
33431 group, the id previously returned by
33432 @samp{-list-thread-groups --available} must be used.
33434 @subsubheading @value{GDBN} Command
33436 The corresponding @value{GDBN} command is @samp{attach}.
33438 @subsubheading Example
33442 =thread-created,id="1"
33443 *stopped,thread-id="1",frame=@{addr="0xb7f7e410",func="bar",args=[]@}
33449 @subheading The @code{-target-compare-sections} Command
33450 @findex -target-compare-sections
33452 @subsubheading Synopsis
33455 -target-compare-sections [ @var{section} ]
33458 Compare data of section @var{section} on target to the exec file.
33459 Without the argument, all sections are compared.
33461 @subsubheading @value{GDBN} Command
33463 The @value{GDBN} equivalent is @samp{compare-sections}.
33465 @subsubheading Example
33470 @subheading The @code{-target-detach} Command
33471 @findex -target-detach
33473 @subsubheading Synopsis
33476 -target-detach [ @var{pid} | @var{gid} ]
33479 Detach from the remote target which normally resumes its execution.
33480 If either @var{pid} or @var{gid} is specified, detaches from either
33481 the specified process, or specified thread group. There's no output.
33483 @subsubheading @value{GDBN} Command
33485 The corresponding @value{GDBN} command is @samp{detach}.
33487 @subsubheading Example
33497 @subheading The @code{-target-disconnect} Command
33498 @findex -target-disconnect
33500 @subsubheading Synopsis
33506 Disconnect from the remote target. There's no output and the target is
33507 generally not resumed.
33509 @subsubheading @value{GDBN} Command
33511 The corresponding @value{GDBN} command is @samp{disconnect}.
33513 @subsubheading Example
33523 @subheading The @code{-target-download} Command
33524 @findex -target-download
33526 @subsubheading Synopsis
33532 Loads the executable onto the remote target.
33533 It prints out an update message every half second, which includes the fields:
33537 The name of the section.
33539 The size of what has been sent so far for that section.
33541 The size of the section.
33543 The total size of what was sent so far (the current and the previous sections).
33545 The size of the overall executable to download.
33549 Each message is sent as status record (@pxref{GDB/MI Output Syntax, ,
33550 @sc{gdb/mi} Output Syntax}).
33552 In addition, it prints the name and size of the sections, as they are
33553 downloaded. These messages include the following fields:
33557 The name of the section.
33559 The size of the section.
33561 The size of the overall executable to download.
33565 At the end, a summary is printed.
33567 @subsubheading @value{GDBN} Command
33569 The corresponding @value{GDBN} command is @samp{load}.
33571 @subsubheading Example
33573 Note: each status message appears on a single line. Here the messages
33574 have been broken down so that they can fit onto a page.
33579 +download,@{section=".text",section-size="6668",total-size="9880"@}
33580 +download,@{section=".text",section-sent="512",section-size="6668",
33581 total-sent="512",total-size="9880"@}
33582 +download,@{section=".text",section-sent="1024",section-size="6668",
33583 total-sent="1024",total-size="9880"@}
33584 +download,@{section=".text",section-sent="1536",section-size="6668",
33585 total-sent="1536",total-size="9880"@}
33586 +download,@{section=".text",section-sent="2048",section-size="6668",
33587 total-sent="2048",total-size="9880"@}
33588 +download,@{section=".text",section-sent="2560",section-size="6668",
33589 total-sent="2560",total-size="9880"@}
33590 +download,@{section=".text",section-sent="3072",section-size="6668",
33591 total-sent="3072",total-size="9880"@}
33592 +download,@{section=".text",section-sent="3584",section-size="6668",
33593 total-sent="3584",total-size="9880"@}
33594 +download,@{section=".text",section-sent="4096",section-size="6668",
33595 total-sent="4096",total-size="9880"@}
33596 +download,@{section=".text",section-sent="4608",section-size="6668",
33597 total-sent="4608",total-size="9880"@}
33598 +download,@{section=".text",section-sent="5120",section-size="6668",
33599 total-sent="5120",total-size="9880"@}
33600 +download,@{section=".text",section-sent="5632",section-size="6668",
33601 total-sent="5632",total-size="9880"@}
33602 +download,@{section=".text",section-sent="6144",section-size="6668",
33603 total-sent="6144",total-size="9880"@}
33604 +download,@{section=".text",section-sent="6656",section-size="6668",
33605 total-sent="6656",total-size="9880"@}
33606 +download,@{section=".init",section-size="28",total-size="9880"@}
33607 +download,@{section=".fini",section-size="28",total-size="9880"@}
33608 +download,@{section=".data",section-size="3156",total-size="9880"@}
33609 +download,@{section=".data",section-sent="512",section-size="3156",
33610 total-sent="7236",total-size="9880"@}
33611 +download,@{section=".data",section-sent="1024",section-size="3156",
33612 total-sent="7748",total-size="9880"@}
33613 +download,@{section=".data",section-sent="1536",section-size="3156",
33614 total-sent="8260",total-size="9880"@}
33615 +download,@{section=".data",section-sent="2048",section-size="3156",
33616 total-sent="8772",total-size="9880"@}
33617 +download,@{section=".data",section-sent="2560",section-size="3156",
33618 total-sent="9284",total-size="9880"@}
33619 +download,@{section=".data",section-sent="3072",section-size="3156",
33620 total-sent="9796",total-size="9880"@}
33621 ^done,address="0x10004",load-size="9880",transfer-rate="6586",
33628 @subheading The @code{-target-exec-status} Command
33629 @findex -target-exec-status
33631 @subsubheading Synopsis
33634 -target-exec-status
33637 Provide information on the state of the target (whether it is running or
33638 not, for instance).
33640 @subsubheading @value{GDBN} Command
33642 There's no equivalent @value{GDBN} command.
33644 @subsubheading Example
33648 @subheading The @code{-target-list-available-targets} Command
33649 @findex -target-list-available-targets
33651 @subsubheading Synopsis
33654 -target-list-available-targets
33657 List the possible targets to connect to.
33659 @subsubheading @value{GDBN} Command
33661 The corresponding @value{GDBN} command is @samp{help target}.
33663 @subsubheading Example
33667 @subheading The @code{-target-list-current-targets} Command
33668 @findex -target-list-current-targets
33670 @subsubheading Synopsis
33673 -target-list-current-targets
33676 Describe the current target.
33678 @subsubheading @value{GDBN} Command
33680 The corresponding information is printed by @samp{info file} (among
33683 @subsubheading Example
33687 @subheading The @code{-target-list-parameters} Command
33688 @findex -target-list-parameters
33690 @subsubheading Synopsis
33693 -target-list-parameters
33699 @subsubheading @value{GDBN} Command
33703 @subsubheading Example
33706 @subheading The @code{-target-flash-erase} Command
33707 @findex -target-flash-erase
33709 @subsubheading Synopsis
33712 -target-flash-erase
33715 Erases all known flash memory regions on the target.
33717 The corresponding @value{GDBN} command is @samp{flash-erase}.
33719 The output is a list of flash regions that have been erased, with starting
33720 addresses and memory region sizes.
33724 -target-flash-erase
33725 ^done,erased-regions=@{address="0x0",size="0x40000"@}
33729 @subheading The @code{-target-select} Command
33730 @findex -target-select
33732 @subsubheading Synopsis
33735 -target-select @var{type} @var{parameters @dots{}}
33738 Connect @value{GDBN} to the remote target. This command takes two args:
33742 The type of target, for instance @samp{remote}, etc.
33743 @item @var{parameters}
33744 Device names, host names and the like. @xref{Target Commands, ,
33745 Commands for Managing Targets}, for more details.
33748 The output is a connection notification, followed by the address at
33749 which the target program is, in the following form:
33752 ^connected,addr="@var{address}",func="@var{function name}",
33753 args=[@var{arg list}]
33756 @subsubheading @value{GDBN} Command
33758 The corresponding @value{GDBN} command is @samp{target}.
33760 @subsubheading Example
33764 -target-select remote /dev/ttya
33765 ^connected,addr="0xfe00a300",func="??",args=[]
33769 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
33770 @node GDB/MI File Transfer Commands
33771 @section @sc{gdb/mi} File Transfer Commands
33774 @subheading The @code{-target-file-put} Command
33775 @findex -target-file-put
33777 @subsubheading Synopsis
33780 -target-file-put @var{hostfile} @var{targetfile}
33783 Copy file @var{hostfile} from the host system (the machine running
33784 @value{GDBN}) to @var{targetfile} on the target system.
33786 @subsubheading @value{GDBN} Command
33788 The corresponding @value{GDBN} command is @samp{remote put}.
33790 @subsubheading Example
33794 -target-file-put localfile remotefile
33800 @subheading The @code{-target-file-get} Command
33801 @findex -target-file-get
33803 @subsubheading Synopsis
33806 -target-file-get @var{targetfile} @var{hostfile}
33809 Copy file @var{targetfile} from the target system to @var{hostfile}
33810 on the host system.
33812 @subsubheading @value{GDBN} Command
33814 The corresponding @value{GDBN} command is @samp{remote get}.
33816 @subsubheading Example
33820 -target-file-get remotefile localfile
33826 @subheading The @code{-target-file-delete} Command
33827 @findex -target-file-delete
33829 @subsubheading Synopsis
33832 -target-file-delete @var{targetfile}
33835 Delete @var{targetfile} from the target system.
33837 @subsubheading @value{GDBN} Command
33839 The corresponding @value{GDBN} command is @samp{remote delete}.
33841 @subsubheading Example
33845 -target-file-delete remotefile
33851 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
33852 @node GDB/MI Ada Exceptions Commands
33853 @section Ada Exceptions @sc{gdb/mi} Commands
33855 @subheading The @code{-info-ada-exceptions} Command
33856 @findex -info-ada-exceptions
33858 @subsubheading Synopsis
33861 -info-ada-exceptions [ @var{regexp}]
33864 List all Ada exceptions defined within the program being debugged.
33865 With a regular expression @var{regexp}, only those exceptions whose
33866 names match @var{regexp} are listed.
33868 @subsubheading @value{GDBN} Command
33870 The corresponding @value{GDBN} command is @samp{info exceptions}.
33872 @subsubheading Result
33874 The result is a table of Ada exceptions. The following columns are
33875 defined for each exception:
33879 The name of the exception.
33882 The address of the exception.
33886 @subsubheading Example
33889 -info-ada-exceptions aint
33890 ^done,ada-exceptions=@{nr_rows="2",nr_cols="2",
33891 hdr=[@{width="1",alignment="-1",col_name="name",colhdr="Name"@},
33892 @{width="1",alignment="-1",col_name="address",colhdr="Address"@}],
33893 body=[@{name="constraint_error",address="0x0000000000613da0"@},
33894 @{name="const.aint_global_e",address="0x0000000000613b00"@}]@}
33897 @subheading Catching Ada Exceptions
33899 The commands describing how to ask @value{GDBN} to stop when a program
33900 raises an exception are described at @ref{Ada Exception GDB/MI
33901 Catchpoint Commands}.
33904 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
33905 @node GDB/MI Support Commands
33906 @section @sc{gdb/mi} Support Commands
33908 Since new commands and features get regularly added to @sc{gdb/mi},
33909 some commands are available to help front-ends query the debugger
33910 about support for these capabilities. Similarly, it is also possible
33911 to query @value{GDBN} about target support of certain features.
33913 @subheading The @code{-info-gdb-mi-command} Command
33914 @cindex @code{-info-gdb-mi-command}
33915 @findex -info-gdb-mi-command
33917 @subsubheading Synopsis
33920 -info-gdb-mi-command @var{cmd_name}
33923 Query support for the @sc{gdb/mi} command named @var{cmd_name}.
33925 Note that the dash (@code{-}) starting all @sc{gdb/mi} commands
33926 is technically not part of the command name (@pxref{GDB/MI Input
33927 Syntax}), and thus should be omitted in @var{cmd_name}. However,
33928 for ease of use, this command also accepts the form with the leading
33931 @subsubheading @value{GDBN} Command
33933 There is no corresponding @value{GDBN} command.
33935 @subsubheading Result
33937 The result is a tuple. There is currently only one field:
33941 This field is equal to @code{"true"} if the @sc{gdb/mi} command exists,
33942 @code{"false"} otherwise.
33946 @subsubheading Example
33948 Here is an example where the @sc{gdb/mi} command does not exist:
33951 -info-gdb-mi-command unsupported-command
33952 ^done,command=@{exists="false"@}
33956 And here is an example where the @sc{gdb/mi} command is known
33960 -info-gdb-mi-command symbol-list-lines
33961 ^done,command=@{exists="true"@}
33964 @subheading The @code{-list-features} Command
33965 @findex -list-features
33966 @cindex supported @sc{gdb/mi} features, list
33968 Returns a list of particular features of the MI protocol that
33969 this version of gdb implements. A feature can be a command,
33970 or a new field in an output of some command, or even an
33971 important bugfix. While a frontend can sometimes detect presence
33972 of a feature at runtime, it is easier to perform detection at debugger
33975 The command returns a list of strings, with each string naming an
33976 available feature. Each returned string is just a name, it does not
33977 have any internal structure. The list of possible feature names
33983 (gdb) -list-features
33984 ^done,result=["feature1","feature2"]
33987 The current list of features is:
33990 @item frozen-varobjs
33991 Indicates support for the @code{-var-set-frozen} command, as well
33992 as possible presense of the @code{frozen} field in the output
33993 of @code{-varobj-create}.
33994 @item pending-breakpoints
33995 Indicates support for the @option{-f} option to the @code{-break-insert}
33998 Indicates Python scripting support, Python-based
33999 pretty-printing commands, and possible presence of the
34000 @samp{display_hint} field in the output of @code{-var-list-children}
34002 Indicates support for the @code{-thread-info} command.
34003 @item data-read-memory-bytes
34004 Indicates support for the @code{-data-read-memory-bytes} and the
34005 @code{-data-write-memory-bytes} commands.
34006 @item breakpoint-notifications
34007 Indicates that changes to breakpoints and breakpoints created via the
34008 CLI will be announced via async records.
34009 @item ada-task-info
34010 Indicates support for the @code{-ada-task-info} command.
34011 @item language-option
34012 Indicates that all @sc{gdb/mi} commands accept the @option{--language}
34013 option (@pxref{Context management}).
34014 @item info-gdb-mi-command
34015 Indicates support for the @code{-info-gdb-mi-command} command.
34016 @item undefined-command-error-code
34017 Indicates support for the "undefined-command" error code in error result
34018 records, produced when trying to execute an undefined @sc{gdb/mi} command
34019 (@pxref{GDB/MI Result Records}).
34020 @item exec-run-start-option
34021 Indicates that the @code{-exec-run} command supports the @option{--start}
34022 option (@pxref{GDB/MI Program Execution}).
34023 @item data-disassemble-a-option
34024 Indicates that the @code{-data-disassemble} command supports the @option{-a}
34025 option (@pxref{GDB/MI Data Manipulation}).
34028 @subheading The @code{-list-target-features} Command
34029 @findex -list-target-features
34031 Returns a list of particular features that are supported by the
34032 target. Those features affect the permitted MI commands, but
34033 unlike the features reported by the @code{-list-features} command, the
34034 features depend on which target GDB is using at the moment. Whenever
34035 a target can change, due to commands such as @code{-target-select},
34036 @code{-target-attach} or @code{-exec-run}, the list of target features
34037 may change, and the frontend should obtain it again.
34041 (gdb) -list-target-features
34042 ^done,result=["async"]
34045 The current list of features is:
34049 Indicates that the target is capable of asynchronous command
34050 execution, which means that @value{GDBN} will accept further commands
34051 while the target is running.
34054 Indicates that the target is capable of reverse execution.
34055 @xref{Reverse Execution}, for more information.
34059 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
34060 @node GDB/MI Miscellaneous Commands
34061 @section Miscellaneous @sc{gdb/mi} Commands
34063 @c @subheading -gdb-complete
34065 @subheading The @code{-gdb-exit} Command
34068 @subsubheading Synopsis
34074 Exit @value{GDBN} immediately.
34076 @subsubheading @value{GDBN} Command
34078 Approximately corresponds to @samp{quit}.
34080 @subsubheading Example
34090 @subheading The @code{-exec-abort} Command
34091 @findex -exec-abort
34093 @subsubheading Synopsis
34099 Kill the inferior running program.
34101 @subsubheading @value{GDBN} Command
34103 The corresponding @value{GDBN} command is @samp{kill}.
34105 @subsubheading Example
34110 @subheading The @code{-gdb-set} Command
34113 @subsubheading Synopsis
34119 Set an internal @value{GDBN} variable.
34120 @c IS THIS A DOLLAR VARIABLE? OR SOMETHING LIKE ANNOTATE ?????
34122 @subsubheading @value{GDBN} Command
34124 The corresponding @value{GDBN} command is @samp{set}.
34126 @subsubheading Example
34136 @subheading The @code{-gdb-show} Command
34139 @subsubheading Synopsis
34145 Show the current value of a @value{GDBN} variable.
34147 @subsubheading @value{GDBN} Command
34149 The corresponding @value{GDBN} command is @samp{show}.
34151 @subsubheading Example
34160 @c @subheading -gdb-source
34163 @subheading The @code{-gdb-version} Command
34164 @findex -gdb-version
34166 @subsubheading Synopsis
34172 Show version information for @value{GDBN}. Used mostly in testing.
34174 @subsubheading @value{GDBN} Command
34176 The @value{GDBN} equivalent is @samp{show version}. @value{GDBN} by
34177 default shows this information when you start an interactive session.
34179 @subsubheading Example
34181 @c This example modifies the actual output from GDB to avoid overfull
34187 ~Copyright 2000 Free Software Foundation, Inc.
34188 ~GDB is free software, covered by the GNU General Public License, and
34189 ~you are welcome to change it and/or distribute copies of it under
34190 ~ certain conditions.
34191 ~Type "show copying" to see the conditions.
34192 ~There is absolutely no warranty for GDB. Type "show warranty" for
34194 ~This GDB was configured as
34195 "--host=sparc-sun-solaris2.5.1 --target=ppc-eabi".
34200 @subheading The @code{-list-thread-groups} Command
34201 @findex -list-thread-groups
34203 @subheading Synopsis
34206 -list-thread-groups [ --available ] [ --recurse 1 ] [ @var{group} ... ]
34209 Lists thread groups (@pxref{Thread groups}). When a single thread
34210 group is passed as the argument, lists the children of that group.
34211 When several thread group are passed, lists information about those
34212 thread groups. Without any parameters, lists information about all
34213 top-level thread groups.
34215 Normally, thread groups that are being debugged are reported.
34216 With the @samp{--available} option, @value{GDBN} reports thread groups
34217 available on the target.
34219 The output of this command may have either a @samp{threads} result or
34220 a @samp{groups} result. The @samp{thread} result has a list of tuples
34221 as value, with each tuple describing a thread (@pxref{GDB/MI Thread
34222 Information}). The @samp{groups} result has a list of tuples as value,
34223 each tuple describing a thread group. If top-level groups are
34224 requested (that is, no parameter is passed), or when several groups
34225 are passed, the output always has a @samp{groups} result. The format
34226 of the @samp{group} result is described below.
34228 To reduce the number of roundtrips it's possible to list thread groups
34229 together with their children, by passing the @samp{--recurse} option
34230 and the recursion depth. Presently, only recursion depth of 1 is
34231 permitted. If this option is present, then every reported thread group
34232 will also include its children, either as @samp{group} or
34233 @samp{threads} field.
34235 In general, any combination of option and parameters is permitted, with
34236 the following caveats:
34240 When a single thread group is passed, the output will typically
34241 be the @samp{threads} result. Because threads may not contain
34242 anything, the @samp{recurse} option will be ignored.
34245 When the @samp{--available} option is passed, limited information may
34246 be available. In particular, the list of threads of a process might
34247 be inaccessible. Further, specifying specific thread groups might
34248 not give any performance advantage over listing all thread groups.
34249 The frontend should assume that @samp{-list-thread-groups --available}
34250 is always an expensive operation and cache the results.
34254 The @samp{groups} result is a list of tuples, where each tuple may
34255 have the following fields:
34259 Identifier of the thread group. This field is always present.
34260 The identifier is an opaque string; frontends should not try to
34261 convert it to an integer, even though it might look like one.
34264 The type of the thread group. At present, only @samp{process} is a
34268 The target-specific process identifier. This field is only present
34269 for thread groups of type @samp{process} and only if the process exists.
34272 The exit code of this group's last exited thread, formatted in octal.
34273 This field is only present for thread groups of type @samp{process} and
34274 only if the process is not running.
34277 The number of children this thread group has. This field may be
34278 absent for an available thread group.
34281 This field has a list of tuples as value, each tuple describing a
34282 thread. It may be present if the @samp{--recurse} option is
34283 specified, and it's actually possible to obtain the threads.
34286 This field is a list of integers, each identifying a core that one
34287 thread of the group is running on. This field may be absent if
34288 such information is not available.
34291 The name of the executable file that corresponds to this thread group.
34292 The field is only present for thread groups of type @samp{process},
34293 and only if there is a corresponding executable file.
34297 @subheading Example
34301 -list-thread-groups
34302 ^done,groups=[@{id="17",type="process",pid="yyy",num_children="2"@}]
34303 -list-thread-groups 17
34304 ^done,threads=[@{id="2",target-id="Thread 0xb7e14b90 (LWP 21257)",
34305 frame=@{level="0",addr="0xffffe410",func="__kernel_vsyscall",args=[]@},state="running"@},
34306 @{id="1",target-id="Thread 0xb7e156b0 (LWP 21254)",
34307 frame=@{level="0",addr="0x0804891f",func="foo",args=[@{name="i",value="10"@}],
34308 file="/tmp/a.c",fullname="/tmp/a.c",line="158",arch="i386:x86_64"@},state="running"@}]]
34309 -list-thread-groups --available
34310 ^done,groups=[@{id="17",type="process",pid="yyy",num_children="2",cores=[1,2]@}]
34311 -list-thread-groups --available --recurse 1
34312 ^done,groups=[@{id="17", types="process",pid="yyy",num_children="2",cores=[1,2],
34313 threads=[@{id="1",target-id="Thread 0xb7e14b90",cores=[1]@},
34314 @{id="2",target-id="Thread 0xb7e14b90",cores=[2]@}]@},..]
34315 -list-thread-groups --available --recurse 1 17 18
34316 ^done,groups=[@{id="17", types="process",pid="yyy",num_children="2",cores=[1,2],
34317 threads=[@{id="1",target-id="Thread 0xb7e14b90",cores=[1]@},
34318 @{id="2",target-id="Thread 0xb7e14b90",cores=[2]@}]@},...]
34321 @subheading The @code{-info-os} Command
34324 @subsubheading Synopsis
34327 -info-os [ @var{type} ]
34330 If no argument is supplied, the command returns a table of available
34331 operating-system-specific information types. If one of these types is
34332 supplied as an argument @var{type}, then the command returns a table
34333 of data of that type.
34335 The types of information available depend on the target operating
34338 @subsubheading @value{GDBN} Command
34340 The corresponding @value{GDBN} command is @samp{info os}.
34342 @subsubheading Example
34344 When run on a @sc{gnu}/Linux system, the output will look something
34350 ^done,OSDataTable=@{nr_rows="10",nr_cols="3",
34351 hdr=[@{width="10",alignment="-1",col_name="col0",colhdr="Type"@},
34352 @{width="10",alignment="-1",col_name="col1",colhdr="Description"@},
34353 @{width="10",alignment="-1",col_name="col2",colhdr="Title"@}],
34354 body=[item=@{col0="cpus",col1="Listing of all cpus/cores on the system",
34356 item=@{col0="files",col1="Listing of all file descriptors",
34357 col2="File descriptors"@},
34358 item=@{col0="modules",col1="Listing of all loaded kernel modules",
34359 col2="Kernel modules"@},
34360 item=@{col0="msg",col1="Listing of all message queues",
34361 col2="Message queues"@},
34362 item=@{col0="processes",col1="Listing of all processes",
34363 col2="Processes"@},
34364 item=@{col0="procgroups",col1="Listing of all process groups",
34365 col2="Process groups"@},
34366 item=@{col0="semaphores",col1="Listing of all semaphores",
34367 col2="Semaphores"@},
34368 item=@{col0="shm",col1="Listing of all shared-memory regions",
34369 col2="Shared-memory regions"@},
34370 item=@{col0="sockets",col1="Listing of all internet-domain sockets",
34372 item=@{col0="threads",col1="Listing of all threads",
34376 ^done,OSDataTable=@{nr_rows="190",nr_cols="4",
34377 hdr=[@{width="10",alignment="-1",col_name="col0",colhdr="pid"@},
34378 @{width="10",alignment="-1",col_name="col1",colhdr="user"@},
34379 @{width="10",alignment="-1",col_name="col2",colhdr="command"@},
34380 @{width="10",alignment="-1",col_name="col3",colhdr="cores"@}],
34381 body=[item=@{col0="1",col1="root",col2="/sbin/init",col3="0"@},
34382 item=@{col0="2",col1="root",col2="[kthreadd]",col3="1"@},
34383 item=@{col0="3",col1="root",col2="[ksoftirqd/0]",col3="0"@},
34385 item=@{col0="26446",col1="stan",col2="bash",col3="0"@},
34386 item=@{col0="28152",col1="stan",col2="bash",col3="1"@}]@}
34390 (Note that the MI output here includes a @code{"Title"} column that
34391 does not appear in command-line @code{info os}; this column is useful
34392 for MI clients that want to enumerate the types of data, such as in a
34393 popup menu, but is needless clutter on the command line, and
34394 @code{info os} omits it.)
34396 @subheading The @code{-add-inferior} Command
34397 @findex -add-inferior
34399 @subheading Synopsis
34405 Creates a new inferior (@pxref{Inferiors and Programs}). The created
34406 inferior is not associated with any executable. Such association may
34407 be established with the @samp{-file-exec-and-symbols} command
34408 (@pxref{GDB/MI File Commands}). The command response has a single
34409 field, @samp{inferior}, whose value is the identifier of the
34410 thread group corresponding to the new inferior.
34412 @subheading Example
34417 ^done,inferior="i3"
34420 @subheading The @code{-interpreter-exec} Command
34421 @findex -interpreter-exec
34423 @subheading Synopsis
34426 -interpreter-exec @var{interpreter} @var{command}
34428 @anchor{-interpreter-exec}
34430 Execute the specified @var{command} in the given @var{interpreter}.
34432 @subheading @value{GDBN} Command
34434 The corresponding @value{GDBN} command is @samp{interpreter-exec}.
34436 @subheading Example
34440 -interpreter-exec console "break main"
34441 &"During symbol reading, couldn't parse type; debugger out of date?.\n"
34442 &"During symbol reading, bad structure-type format.\n"
34443 ~"Breakpoint 1 at 0x8074fc6: file ../../src/gdb/main.c, line 743.\n"
34448 @subheading The @code{-inferior-tty-set} Command
34449 @findex -inferior-tty-set
34451 @subheading Synopsis
34454 -inferior-tty-set /dev/pts/1
34457 Set terminal for future runs of the program being debugged.
34459 @subheading @value{GDBN} Command
34461 The corresponding @value{GDBN} command is @samp{set inferior-tty} /dev/pts/1.
34463 @subheading Example
34467 -inferior-tty-set /dev/pts/1
34472 @subheading The @code{-inferior-tty-show} Command
34473 @findex -inferior-tty-show
34475 @subheading Synopsis
34481 Show terminal for future runs of program being debugged.
34483 @subheading @value{GDBN} Command
34485 The corresponding @value{GDBN} command is @samp{show inferior-tty}.
34487 @subheading Example
34491 -inferior-tty-set /dev/pts/1
34495 ^done,inferior_tty_terminal="/dev/pts/1"
34499 @subheading The @code{-enable-timings} Command
34500 @findex -enable-timings
34502 @subheading Synopsis
34505 -enable-timings [yes | no]
34508 Toggle the printing of the wallclock, user and system times for an MI
34509 command as a field in its output. This command is to help frontend
34510 developers optimize the performance of their code. No argument is
34511 equivalent to @samp{yes}.
34513 @subheading @value{GDBN} Command
34517 @subheading Example
34525 ^done,bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
34526 addr="0x080484ed",func="main",file="myprog.c",
34527 fullname="/home/nickrob/myprog.c",line="73",thread-groups=["i1"],
34529 time=@{wallclock="0.05185",user="0.00800",system="0.00000"@}
34537 *stopped,reason="breakpoint-hit",disp="keep",bkptno="1",thread-id="0",
34538 frame=@{addr="0x080484ed",func="main",args=[@{name="argc",value="1"@},
34539 @{name="argv",value="0xbfb60364"@}],file="myprog.c",
34540 fullname="/home/nickrob/myprog.c",line="73",arch="i386:x86_64"@}
34544 @subheading The @code{-complete} Command
34547 @subheading Synopsis
34550 -complete @var{command}
34553 Show a list of completions for partially typed CLI @var{command}.
34555 This command is intended for @sc{gdb/mi} frontends that cannot use two separate
34556 CLI and MI channels --- for example: because of lack of PTYs like on Windows or
34557 because @value{GDBN} is used remotely via a SSH connection.
34561 The result consists of two or three fields:
34565 This field contains the completed @var{command}. If @var{command}
34566 has no known completions, this field is omitted.
34569 This field contains a (possibly empty) array of matches. It is always present.
34571 @item max_completions_reached
34572 This field contains @code{1} if number of known completions is above
34573 @code{max-completions} limit (@pxref{Completion}), otherwise it contains
34574 @code{0}. It is always present.
34578 @subheading @value{GDBN} Command
34580 The corresponding @value{GDBN} command is @samp{complete}.
34582 @subheading Example
34587 ^done,completion="break",
34588 matches=["break","break-range"],
34589 max_completions_reached="0"
34592 ^done,completion="b ma",
34593 matches=["b madvise","b main"],max_completions_reached="0"
34595 -complete "b push_b"
34596 ^done,completion="b push_back(",
34598 "b A::push_back(void*)",
34599 "b std::string::push_back(char)",
34600 "b std::vector<int, std::allocator<int> >::push_back(int&&)"],
34601 max_completions_reached="0"
34603 -complete "nonexist"
34604 ^done,matches=[],max_completions_reached="0"
34610 @chapter @value{GDBN} Annotations
34612 This chapter describes annotations in @value{GDBN}. Annotations were
34613 designed to interface @value{GDBN} to graphical user interfaces or other
34614 similar programs which want to interact with @value{GDBN} at a
34615 relatively high level.
34617 The annotation mechanism has largely been superseded by @sc{gdb/mi}
34621 This is Edition @value{EDITION}, @value{DATE}.
34625 * Annotations Overview:: What annotations are; the general syntax.
34626 * Server Prefix:: Issuing a command without affecting user state.
34627 * Prompting:: Annotations marking @value{GDBN}'s need for input.
34628 * Errors:: Annotations for error messages.
34629 * Invalidation:: Some annotations describe things now invalid.
34630 * Annotations for Running::
34631 Whether the program is running, how it stopped, etc.
34632 * Source Annotations:: Annotations describing source code.
34635 @node Annotations Overview
34636 @section What is an Annotation?
34637 @cindex annotations
34639 Annotations start with a newline character, two @samp{control-z}
34640 characters, and the name of the annotation. If there is no additional
34641 information associated with this annotation, the name of the annotation
34642 is followed immediately by a newline. If there is additional
34643 information, the name of the annotation is followed by a space, the
34644 additional information, and a newline. The additional information
34645 cannot contain newline characters.
34647 Any output not beginning with a newline and two @samp{control-z}
34648 characters denotes literal output from @value{GDBN}. Currently there is
34649 no need for @value{GDBN} to output a newline followed by two
34650 @samp{control-z} characters, but if there was such a need, the
34651 annotations could be extended with an @samp{escape} annotation which
34652 means those three characters as output.
34654 The annotation @var{level}, which is specified using the
34655 @option{--annotate} command line option (@pxref{Mode Options}), controls
34656 how much information @value{GDBN} prints together with its prompt,
34657 values of expressions, source lines, and other types of output. Level 0
34658 is for no annotations, level 1 is for use when @value{GDBN} is run as a
34659 subprocess of @sc{gnu} Emacs, level 3 is the maximum annotation suitable
34660 for programs that control @value{GDBN}, and level 2 annotations have
34661 been made obsolete (@pxref{Limitations, , Limitations of the Annotation
34662 Interface, annotate, GDB's Obsolete Annotations}).
34665 @kindex set annotate
34666 @item set annotate @var{level}
34667 The @value{GDBN} command @code{set annotate} sets the level of
34668 annotations to the specified @var{level}.
34670 @item show annotate
34671 @kindex show annotate
34672 Show the current annotation level.
34675 This chapter describes level 3 annotations.
34677 A simple example of starting up @value{GDBN} with annotations is:
34680 $ @kbd{gdb --annotate=3}
34682 Copyright 2003 Free Software Foundation, Inc.
34683 GDB is free software, covered by the GNU General Public License,
34684 and you are welcome to change it and/or distribute copies of it
34685 under certain conditions.
34686 Type "show copying" to see the conditions.
34687 There is absolutely no warranty for GDB. Type "show warranty"
34689 This GDB was configured as "i386-pc-linux-gnu"
34700 Here @samp{quit} is input to @value{GDBN}; the rest is output from
34701 @value{GDBN}. The three lines beginning @samp{^Z^Z} (where @samp{^Z}
34702 denotes a @samp{control-z} character) are annotations; the rest is
34703 output from @value{GDBN}.
34705 @node Server Prefix
34706 @section The Server Prefix
34707 @cindex server prefix
34709 If you prefix a command with @samp{server } then it will not affect
34710 the command history, nor will it affect @value{GDBN}'s notion of which
34711 command to repeat if @key{RET} is pressed on a line by itself. This
34712 means that commands can be run behind a user's back by a front-end in
34713 a transparent manner.
34715 The @code{server } prefix does not affect the recording of values into
34716 the value history; to print a value without recording it into the
34717 value history, use the @code{output} command instead of the
34718 @code{print} command.
34720 Using this prefix also disables confirmation requests
34721 (@pxref{confirmation requests}).
34724 @section Annotation for @value{GDBN} Input
34726 @cindex annotations for prompts
34727 When @value{GDBN} prompts for input, it annotates this fact so it is possible
34728 to know when to send output, when the output from a given command is
34731 Different kinds of input each have a different @dfn{input type}. Each
34732 input type has three annotations: a @code{pre-} annotation, which
34733 denotes the beginning of any prompt which is being output, a plain
34734 annotation, which denotes the end of the prompt, and then a @code{post-}
34735 annotation which denotes the end of any echo which may (or may not) be
34736 associated with the input. For example, the @code{prompt} input type
34737 features the following annotations:
34745 The input types are
34748 @findex pre-prompt annotation
34749 @findex prompt annotation
34750 @findex post-prompt annotation
34752 When @value{GDBN} is prompting for a command (the main @value{GDBN} prompt).
34754 @findex pre-commands annotation
34755 @findex commands annotation
34756 @findex post-commands annotation
34758 When @value{GDBN} prompts for a set of commands, like in the @code{commands}
34759 command. The annotations are repeated for each command which is input.
34761 @findex pre-overload-choice annotation
34762 @findex overload-choice annotation
34763 @findex post-overload-choice annotation
34764 @item overload-choice
34765 When @value{GDBN} wants the user to select between various overloaded functions.
34767 @findex pre-query annotation
34768 @findex query annotation
34769 @findex post-query annotation
34771 When @value{GDBN} wants the user to confirm a potentially dangerous operation.
34773 @findex pre-prompt-for-continue annotation
34774 @findex prompt-for-continue annotation
34775 @findex post-prompt-for-continue annotation
34776 @item prompt-for-continue
34777 When @value{GDBN} is asking the user to press return to continue. Note: Don't
34778 expect this to work well; instead use @code{set height 0} to disable
34779 prompting. This is because the counting of lines is buggy in the
34780 presence of annotations.
34785 @cindex annotations for errors, warnings and interrupts
34787 @findex quit annotation
34792 This annotation occurs right before @value{GDBN} responds to an interrupt.
34794 @findex error annotation
34799 This annotation occurs right before @value{GDBN} responds to an error.
34801 Quit and error annotations indicate that any annotations which @value{GDBN} was
34802 in the middle of may end abruptly. For example, if a
34803 @code{value-history-begin} annotation is followed by a @code{error}, one
34804 cannot expect to receive the matching @code{value-history-end}. One
34805 cannot expect not to receive it either, however; an error annotation
34806 does not necessarily mean that @value{GDBN} is immediately returning all the way
34809 @findex error-begin annotation
34810 A quit or error annotation may be preceded by
34816 Any output between that and the quit or error annotation is the error
34819 Warning messages are not yet annotated.
34820 @c If we want to change that, need to fix warning(), type_error(),
34821 @c range_error(), and possibly other places.
34824 @section Invalidation Notices
34826 @cindex annotations for invalidation messages
34827 The following annotations say that certain pieces of state may have
34831 @findex frames-invalid annotation
34832 @item ^Z^Zframes-invalid
34834 The frames (for example, output from the @code{backtrace} command) may
34837 @findex breakpoints-invalid annotation
34838 @item ^Z^Zbreakpoints-invalid
34840 The breakpoints may have changed. For example, the user just added or
34841 deleted a breakpoint.
34844 @node Annotations for Running
34845 @section Running the Program
34846 @cindex annotations for running programs
34848 @findex starting annotation
34849 @findex stopping annotation
34850 When the program starts executing due to a @value{GDBN} command such as
34851 @code{step} or @code{continue},
34857 is output. When the program stops,
34863 is output. Before the @code{stopped} annotation, a variety of
34864 annotations describe how the program stopped.
34867 @findex exited annotation
34868 @item ^Z^Zexited @var{exit-status}
34869 The program exited, and @var{exit-status} is the exit status (zero for
34870 successful exit, otherwise nonzero).
34872 @findex signalled annotation
34873 @findex signal-name annotation
34874 @findex signal-name-end annotation
34875 @findex signal-string annotation
34876 @findex signal-string-end annotation
34877 @item ^Z^Zsignalled
34878 The program exited with a signal. After the @code{^Z^Zsignalled}, the
34879 annotation continues:
34885 ^Z^Zsignal-name-end
34889 ^Z^Zsignal-string-end
34894 where @var{name} is the name of the signal, such as @code{SIGILL} or
34895 @code{SIGSEGV}, and @var{string} is the explanation of the signal, such
34896 as @code{Illegal Instruction} or @code{Segmentation fault}. The arguments
34897 @var{intro-text}, @var{middle-text}, and @var{end-text} are for the
34898 user's benefit and have no particular format.
34900 @findex signal annotation
34902 The syntax of this annotation is just like @code{signalled}, but @value{GDBN} is
34903 just saying that the program received the signal, not that it was
34904 terminated with it.
34906 @findex breakpoint annotation
34907 @item ^Z^Zbreakpoint @var{number}
34908 The program hit breakpoint number @var{number}.
34910 @findex watchpoint annotation
34911 @item ^Z^Zwatchpoint @var{number}
34912 The program hit watchpoint number @var{number}.
34915 @node Source Annotations
34916 @section Displaying Source
34917 @cindex annotations for source display
34919 @findex source annotation
34920 The following annotation is used instead of displaying source code:
34923 ^Z^Zsource @var{filename}:@var{line}:@var{character}:@var{middle}:@var{addr}
34926 where @var{filename} is an absolute file name indicating which source
34927 file, @var{line} is the line number within that file (where 1 is the
34928 first line in the file), @var{character} is the character position
34929 within the file (where 0 is the first character in the file) (for most
34930 debug formats this will necessarily point to the beginning of a line),
34931 @var{middle} is @samp{middle} if @var{addr} is in the middle of the
34932 line, or @samp{beg} if @var{addr} is at the beginning of the line, and
34933 @var{addr} is the address in the target program associated with the
34934 source which is being displayed. The @var{addr} is in the form @samp{0x}
34935 followed by one or more lowercase hex digits (note that this does not
34936 depend on the language).
34938 @node JIT Interface
34939 @chapter JIT Compilation Interface
34940 @cindex just-in-time compilation
34941 @cindex JIT compilation interface
34943 This chapter documents @value{GDBN}'s @dfn{just-in-time} (JIT) compilation
34944 interface. A JIT compiler is a program or library that generates native
34945 executable code at runtime and executes it, usually in order to achieve good
34946 performance while maintaining platform independence.
34948 Programs that use JIT compilation are normally difficult to debug because
34949 portions of their code are generated at runtime, instead of being loaded from
34950 object files, which is where @value{GDBN} normally finds the program's symbols
34951 and debug information. In order to debug programs that use JIT compilation,
34952 @value{GDBN} has an interface that allows the program to register in-memory
34953 symbol files with @value{GDBN} at runtime.
34955 If you are using @value{GDBN} to debug a program that uses this interface, then
34956 it should work transparently so long as you have not stripped the binary. If
34957 you are developing a JIT compiler, then the interface is documented in the rest
34958 of this chapter. At this time, the only known client of this interface is the
34961 Broadly speaking, the JIT interface mirrors the dynamic loader interface. The
34962 JIT compiler communicates with @value{GDBN} by writing data into a global
34963 variable and calling a fuction at a well-known symbol. When @value{GDBN}
34964 attaches, it reads a linked list of symbol files from the global variable to
34965 find existing code, and puts a breakpoint in the function so that it can find
34966 out about additional code.
34969 * Declarations:: Relevant C struct declarations
34970 * Registering Code:: Steps to register code
34971 * Unregistering Code:: Steps to unregister code
34972 * Custom Debug Info:: Emit debug information in a custom format
34976 @section JIT Declarations
34978 These are the relevant struct declarations that a C program should include to
34979 implement the interface:
34989 struct jit_code_entry
34991 struct jit_code_entry *next_entry;
34992 struct jit_code_entry *prev_entry;
34993 const char *symfile_addr;
34994 uint64_t symfile_size;
34997 struct jit_descriptor
35000 /* This type should be jit_actions_t, but we use uint32_t
35001 to be explicit about the bitwidth. */
35002 uint32_t action_flag;
35003 struct jit_code_entry *relevant_entry;
35004 struct jit_code_entry *first_entry;
35007 /* GDB puts a breakpoint in this function. */
35008 void __attribute__((noinline)) __jit_debug_register_code() @{ @};
35010 /* Make sure to specify the version statically, because the
35011 debugger may check the version before we can set it. */
35012 struct jit_descriptor __jit_debug_descriptor = @{ 1, 0, 0, 0 @};
35015 If the JIT is multi-threaded, then it is important that the JIT synchronize any
35016 modifications to this global data properly, which can easily be done by putting
35017 a global mutex around modifications to these structures.
35019 @node Registering Code
35020 @section Registering Code
35022 To register code with @value{GDBN}, the JIT should follow this protocol:
35026 Generate an object file in memory with symbols and other desired debug
35027 information. The file must include the virtual addresses of the sections.
35030 Create a code entry for the file, which gives the start and size of the symbol
35034 Add it to the linked list in the JIT descriptor.
35037 Point the relevant_entry field of the descriptor at the entry.
35040 Set @code{action_flag} to @code{JIT_REGISTER} and call
35041 @code{__jit_debug_register_code}.
35044 When @value{GDBN} is attached and the breakpoint fires, @value{GDBN} uses the
35045 @code{relevant_entry} pointer so it doesn't have to walk the list looking for
35046 new code. However, the linked list must still be maintained in order to allow
35047 @value{GDBN} to attach to a running process and still find the symbol files.
35049 @node Unregistering Code
35050 @section Unregistering Code
35052 If code is freed, then the JIT should use the following protocol:
35056 Remove the code entry corresponding to the code from the linked list.
35059 Point the @code{relevant_entry} field of the descriptor at the code entry.
35062 Set @code{action_flag} to @code{JIT_UNREGISTER} and call
35063 @code{__jit_debug_register_code}.
35066 If the JIT frees or recompiles code without unregistering it, then @value{GDBN}
35067 and the JIT will leak the memory used for the associated symbol files.
35069 @node Custom Debug Info
35070 @section Custom Debug Info
35071 @cindex custom JIT debug info
35072 @cindex JIT debug info reader
35074 Generating debug information in platform-native file formats (like ELF
35075 or COFF) may be an overkill for JIT compilers; especially if all the
35076 debug info is used for is displaying a meaningful backtrace. The
35077 issue can be resolved by having the JIT writers decide on a debug info
35078 format and also provide a reader that parses the debug info generated
35079 by the JIT compiler. This section gives a brief overview on writing
35080 such a parser. More specific details can be found in the source file
35081 @file{gdb/jit-reader.in}, which is also installed as a header at
35082 @file{@var{includedir}/gdb/jit-reader.h} for easy inclusion.
35084 The reader is implemented as a shared object (so this functionality is
35085 not available on platforms which don't allow loading shared objects at
35086 runtime). Two @value{GDBN} commands, @code{jit-reader-load} and
35087 @code{jit-reader-unload} are provided, to be used to load and unload
35088 the readers from a preconfigured directory. Once loaded, the shared
35089 object is used the parse the debug information emitted by the JIT
35093 * Using JIT Debug Info Readers:: How to use supplied readers correctly
35094 * Writing JIT Debug Info Readers:: Creating a debug-info reader
35097 @node Using JIT Debug Info Readers
35098 @subsection Using JIT Debug Info Readers
35099 @kindex jit-reader-load
35100 @kindex jit-reader-unload
35102 Readers can be loaded and unloaded using the @code{jit-reader-load}
35103 and @code{jit-reader-unload} commands.
35106 @item jit-reader-load @var{reader}
35107 Load the JIT reader named @var{reader}, which is a shared
35108 object specified as either an absolute or a relative file name. In
35109 the latter case, @value{GDBN} will try to load the reader from a
35110 pre-configured directory, usually @file{@var{libdir}/gdb/} on a UNIX
35111 system (here @var{libdir} is the system library directory, often
35112 @file{/usr/local/lib}).
35114 Only one reader can be active at a time; trying to load a second
35115 reader when one is already loaded will result in @value{GDBN}
35116 reporting an error. A new JIT reader can be loaded by first unloading
35117 the current one using @code{jit-reader-unload} and then invoking
35118 @code{jit-reader-load}.
35120 @item jit-reader-unload
35121 Unload the currently loaded JIT reader.
35125 @node Writing JIT Debug Info Readers
35126 @subsection Writing JIT Debug Info Readers
35127 @cindex writing JIT debug info readers
35129 As mentioned, a reader is essentially a shared object conforming to a
35130 certain ABI. This ABI is described in @file{jit-reader.h}.
35132 @file{jit-reader.h} defines the structures, macros and functions
35133 required to write a reader. It is installed (along with
35134 @value{GDBN}), in @file{@var{includedir}/gdb} where @var{includedir} is
35135 the system include directory.
35137 Readers need to be released under a GPL compatible license. A reader
35138 can be declared as released under such a license by placing the macro
35139 @code{GDB_DECLARE_GPL_COMPATIBLE_READER} in a source file.
35141 The entry point for readers is the symbol @code{gdb_init_reader},
35142 which is expected to be a function with the prototype
35144 @findex gdb_init_reader
35146 extern struct gdb_reader_funcs *gdb_init_reader (void);
35149 @cindex @code{struct gdb_reader_funcs}
35151 @code{struct gdb_reader_funcs} contains a set of pointers to callback
35152 functions. These functions are executed to read the debug info
35153 generated by the JIT compiler (@code{read}), to unwind stack frames
35154 (@code{unwind}) and to create canonical frame IDs
35155 (@code{get_Frame_id}). It also has a callback that is called when the
35156 reader is being unloaded (@code{destroy}). The struct looks like this
35159 struct gdb_reader_funcs
35161 /* Must be set to GDB_READER_INTERFACE_VERSION. */
35162 int reader_version;
35164 /* For use by the reader. */
35167 gdb_read_debug_info *read;
35168 gdb_unwind_frame *unwind;
35169 gdb_get_frame_id *get_frame_id;
35170 gdb_destroy_reader *destroy;
35174 @cindex @code{struct gdb_symbol_callbacks}
35175 @cindex @code{struct gdb_unwind_callbacks}
35177 The callbacks are provided with another set of callbacks by
35178 @value{GDBN} to do their job. For @code{read}, these callbacks are
35179 passed in a @code{struct gdb_symbol_callbacks} and for @code{unwind}
35180 and @code{get_frame_id}, in a @code{struct gdb_unwind_callbacks}.
35181 @code{struct gdb_symbol_callbacks} has callbacks to create new object
35182 files and new symbol tables inside those object files. @code{struct
35183 gdb_unwind_callbacks} has callbacks to read registers off the current
35184 frame and to write out the values of the registers in the previous
35185 frame. Both have a callback (@code{target_read}) to read bytes off the
35186 target's address space.
35188 @node In-Process Agent
35189 @chapter In-Process Agent
35190 @cindex debugging agent
35191 The traditional debugging model is conceptually low-speed, but works fine,
35192 because most bugs can be reproduced in debugging-mode execution. However,
35193 as multi-core or many-core processors are becoming mainstream, and
35194 multi-threaded programs become more and more popular, there should be more
35195 and more bugs that only manifest themselves at normal-mode execution, for
35196 example, thread races, because debugger's interference with the program's
35197 timing may conceal the bugs. On the other hand, in some applications,
35198 it is not feasible for the debugger to interrupt the program's execution
35199 long enough for the developer to learn anything helpful about its behavior.
35200 If the program's correctness depends on its real-time behavior, delays
35201 introduced by a debugger might cause the program to fail, even when the
35202 code itself is correct. It is useful to be able to observe the program's
35203 behavior without interrupting it.
35205 Therefore, traditional debugging model is too intrusive to reproduce
35206 some bugs. In order to reduce the interference with the program, we can
35207 reduce the number of operations performed by debugger. The
35208 @dfn{In-Process Agent}, a shared library, is running within the same
35209 process with inferior, and is able to perform some debugging operations
35210 itself. As a result, debugger is only involved when necessary, and
35211 performance of debugging can be improved accordingly. Note that
35212 interference with program can be reduced but can't be removed completely,
35213 because the in-process agent will still stop or slow down the program.
35215 The in-process agent can interpret and execute Agent Expressions
35216 (@pxref{Agent Expressions}) during performing debugging operations. The
35217 agent expressions can be used for different purposes, such as collecting
35218 data in tracepoints, and condition evaluation in breakpoints.
35220 @anchor{Control Agent}
35221 You can control whether the in-process agent is used as an aid for
35222 debugging with the following commands:
35225 @kindex set agent on
35227 Causes the in-process agent to perform some operations on behalf of the
35228 debugger. Just which operations requested by the user will be done
35229 by the in-process agent depends on the its capabilities. For example,
35230 if you request to evaluate breakpoint conditions in the in-process agent,
35231 and the in-process agent has such capability as well, then breakpoint
35232 conditions will be evaluated in the in-process agent.
35234 @kindex set agent off
35235 @item set agent off
35236 Disables execution of debugging operations by the in-process agent. All
35237 of the operations will be performed by @value{GDBN}.
35241 Display the current setting of execution of debugging operations by
35242 the in-process agent.
35246 * In-Process Agent Protocol::
35249 @node In-Process Agent Protocol
35250 @section In-Process Agent Protocol
35251 @cindex in-process agent protocol
35253 The in-process agent is able to communicate with both @value{GDBN} and
35254 GDBserver (@pxref{In-Process Agent}). This section documents the protocol
35255 used for communications between @value{GDBN} or GDBserver and the IPA.
35256 In general, @value{GDBN} or GDBserver sends commands
35257 (@pxref{IPA Protocol Commands}) and data to in-process agent, and then
35258 in-process agent replies back with the return result of the command, or
35259 some other information. The data sent to in-process agent is composed
35260 of primitive data types, such as 4-byte or 8-byte type, and composite
35261 types, which are called objects (@pxref{IPA Protocol Objects}).
35264 * IPA Protocol Objects::
35265 * IPA Protocol Commands::
35268 @node IPA Protocol Objects
35269 @subsection IPA Protocol Objects
35270 @cindex ipa protocol objects
35272 The commands sent to and results received from agent may contain some
35273 complex data types called @dfn{objects}.
35275 The in-process agent is running on the same machine with @value{GDBN}
35276 or GDBserver, so it doesn't have to handle as much differences between
35277 two ends as remote protocol (@pxref{Remote Protocol}) tries to handle.
35278 However, there are still some differences of two ends in two processes:
35282 word size. On some 64-bit machines, @value{GDBN} or GDBserver can be
35283 compiled as a 64-bit executable, while in-process agent is a 32-bit one.
35285 ABI. Some machines may have multiple types of ABI, @value{GDBN} or
35286 GDBserver is compiled with one, and in-process agent is compiled with
35290 Here are the IPA Protocol Objects:
35294 agent expression object. It represents an agent expression
35295 (@pxref{Agent Expressions}).
35296 @anchor{agent expression object}
35298 tracepoint action object. It represents a tracepoint action
35299 (@pxref{Tracepoint Actions,,Tracepoint Action Lists}) to collect registers,
35300 memory, static trace data and to evaluate expression.
35301 @anchor{tracepoint action object}
35303 tracepoint object. It represents a tracepoint (@pxref{Tracepoints}).
35304 @anchor{tracepoint object}
35308 The following table describes important attributes of each IPA protocol
35311 @multitable @columnfractions .30 .20 .50
35312 @headitem Name @tab Size @tab Description
35313 @item @emph{agent expression object} @tab @tab
35314 @item length @tab 4 @tab length of bytes code
35315 @item byte code @tab @var{length} @tab contents of byte code
35316 @item @emph{tracepoint action for collecting memory} @tab @tab
35317 @item 'M' @tab 1 @tab type of tracepoint action
35318 @item addr @tab 8 @tab if @var{basereg} is @samp{-1}, @var{addr} is the
35319 address of the lowest byte to collect, otherwise @var{addr} is the offset
35320 of @var{basereg} for memory collecting.
35321 @item len @tab 8 @tab length of memory for collecting
35322 @item basereg @tab 4 @tab the register number containing the starting
35323 memory address for collecting.
35324 @item @emph{tracepoint action for collecting registers} @tab @tab
35325 @item 'R' @tab 1 @tab type of tracepoint action
35326 @item @emph{tracepoint action for collecting static trace data} @tab @tab
35327 @item 'L' @tab 1 @tab type of tracepoint action
35328 @item @emph{tracepoint action for expression evaluation} @tab @tab
35329 @item 'X' @tab 1 @tab type of tracepoint action
35330 @item agent expression @tab length of @tab @ref{agent expression object}
35331 @item @emph{tracepoint object} @tab @tab
35332 @item number @tab 4 @tab number of tracepoint
35333 @item address @tab 8 @tab address of tracepoint inserted on
35334 @item type @tab 4 @tab type of tracepoint
35335 @item enabled @tab 1 @tab enable or disable of tracepoint
35336 @item step_count @tab 8 @tab step
35337 @item pass_count @tab 8 @tab pass
35338 @item numactions @tab 4 @tab number of tracepoint actions
35339 @item hit count @tab 8 @tab hit count
35340 @item trace frame usage @tab 8 @tab trace frame usage
35341 @item compiled_cond @tab 8 @tab compiled condition
35342 @item orig_size @tab 8 @tab orig size
35343 @item condition @tab 4 if condition is NULL otherwise length of
35344 @ref{agent expression object}
35345 @tab zero if condition is NULL, otherwise is
35346 @ref{agent expression object}
35347 @item actions @tab variable
35348 @tab numactions number of @ref{tracepoint action object}
35351 @node IPA Protocol Commands
35352 @subsection IPA Protocol Commands
35353 @cindex ipa protocol commands
35355 The spaces in each command are delimiters to ease reading this commands
35356 specification. They don't exist in real commands.
35360 @item FastTrace:@var{tracepoint_object} @var{gdb_jump_pad_head}
35361 Installs a new fast tracepoint described by @var{tracepoint_object}
35362 (@pxref{tracepoint object}). The @var{gdb_jump_pad_head}, 8-byte long, is the
35363 head of @dfn{jumppad}, which is used to jump to data collection routine
35368 @item OK @var{target_address} @var{gdb_jump_pad_head} @var{fjump_size} @var{fjump}
35369 @var{target_address} is address of tracepoint in the inferior.
35370 The @var{gdb_jump_pad_head} is updated head of jumppad. Both of
35371 @var{target_address} and @var{gdb_jump_pad_head} are 8-byte long.
35372 The @var{fjump} contains a sequence of instructions jump to jumppad entry.
35373 The @var{fjump_size}, 4-byte long, is the size of @var{fjump}.
35380 Closes the in-process agent. This command is sent when @value{GDBN} or GDBserver
35381 is about to kill inferiors.
35389 @item probe_marker_at:@var{address}
35390 Asks in-process agent to probe the marker at @var{address}.
35397 @item unprobe_marker_at:@var{address}
35398 Asks in-process agent to unprobe the marker at @var{address}.
35402 @chapter Reporting Bugs in @value{GDBN}
35403 @cindex bugs in @value{GDBN}
35404 @cindex reporting bugs in @value{GDBN}
35406 Your bug reports play an essential role in making @value{GDBN} reliable.
35408 Reporting a bug may help you by bringing a solution to your problem, or it
35409 may not. But in any case the principal function of a bug report is to help
35410 the entire community by making the next version of @value{GDBN} work better. Bug
35411 reports are your contribution to the maintenance of @value{GDBN}.
35413 In order for a bug report to serve its purpose, you must include the
35414 information that enables us to fix the bug.
35417 * Bug Criteria:: Have you found a bug?
35418 * Bug Reporting:: How to report bugs
35422 @section Have You Found a Bug?
35423 @cindex bug criteria
35425 If you are not sure whether you have found a bug, here are some guidelines:
35428 @cindex fatal signal
35429 @cindex debugger crash
35430 @cindex crash of debugger
35432 If the debugger gets a fatal signal, for any input whatever, that is a
35433 @value{GDBN} bug. Reliable debuggers never crash.
35435 @cindex error on valid input
35437 If @value{GDBN} produces an error message for valid input, that is a
35438 bug. (Note that if you're cross debugging, the problem may also be
35439 somewhere in the connection to the target.)
35441 @cindex invalid input
35443 If @value{GDBN} does not produce an error message for invalid input,
35444 that is a bug. However, you should note that your idea of
35445 ``invalid input'' might be our idea of ``an extension'' or ``support
35446 for traditional practice''.
35449 If you are an experienced user of debugging tools, your suggestions
35450 for improvement of @value{GDBN} are welcome in any case.
35453 @node Bug Reporting
35454 @section How to Report Bugs
35455 @cindex bug reports
35456 @cindex @value{GDBN} bugs, reporting
35458 A number of companies and individuals offer support for @sc{gnu} products.
35459 If you obtained @value{GDBN} from a support organization, we recommend you
35460 contact that organization first.
35462 You can find contact information for many support companies and
35463 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
35465 @c should add a web page ref...
35468 @ifset BUGURL_DEFAULT
35469 In any event, we also recommend that you submit bug reports for
35470 @value{GDBN}. The preferred method is to submit them directly using
35471 @uref{http://www.gnu.org/software/gdb/bugs/, @value{GDBN}'s Bugs web
35472 page}. Alternatively, the @email{bug-gdb@@gnu.org, e-mail gateway} can
35475 @strong{Do not send bug reports to @samp{info-gdb}, or to
35476 @samp{help-gdb}, or to any newsgroups.} Most users of @value{GDBN} do
35477 not want to receive bug reports. Those that do have arranged to receive
35480 The mailing list @samp{bug-gdb} has a newsgroup @samp{gnu.gdb.bug} which
35481 serves as a repeater. The mailing list and the newsgroup carry exactly
35482 the same messages. Often people think of posting bug reports to the
35483 newsgroup instead of mailing them. This appears to work, but it has one
35484 problem which can be crucial: a newsgroup posting often lacks a mail
35485 path back to the sender. Thus, if we need to ask for more information,
35486 we may be unable to reach you. For this reason, it is better to send
35487 bug reports to the mailing list.
35489 @ifclear BUGURL_DEFAULT
35490 In any event, we also recommend that you submit bug reports for
35491 @value{GDBN} to @value{BUGURL}.
35495 The fundamental principle of reporting bugs usefully is this:
35496 @strong{report all the facts}. If you are not sure whether to state a
35497 fact or leave it out, state it!
35499 Often people omit facts because they think they know what causes the
35500 problem and assume that some details do not matter. Thus, you might
35501 assume that the name of the variable you use in an example does not matter.
35502 Well, probably it does not, but one cannot be sure. Perhaps the bug is a
35503 stray memory reference which happens to fetch from the location where that
35504 name is stored in memory; perhaps, if the name were different, the contents
35505 of that location would fool the debugger into doing the right thing despite
35506 the bug. Play it safe and give a specific, complete example. That is the
35507 easiest thing for you to do, and the most helpful.
35509 Keep in mind that the purpose of a bug report is to enable us to fix the
35510 bug. It may be that the bug has been reported previously, but neither
35511 you nor we can know that unless your bug report is complete and
35514 Sometimes people give a few sketchy facts and ask, ``Does this ring a
35515 bell?'' Those bug reports are useless, and we urge everyone to
35516 @emph{refuse to respond to them} except to chide the sender to report
35519 To enable us to fix the bug, you should include all these things:
35523 The version of @value{GDBN}. @value{GDBN} announces it if you start
35524 with no arguments; you can also print it at any time using @code{show
35527 Without this, we will not know whether there is any point in looking for
35528 the bug in the current version of @value{GDBN}.
35531 The type of machine you are using, and the operating system name and
35535 The details of the @value{GDBN} build-time configuration.
35536 @value{GDBN} shows these details if you invoke it with the
35537 @option{--configuration} command-line option, or if you type
35538 @code{show configuration} at @value{GDBN}'s prompt.
35541 What compiler (and its version) was used to compile @value{GDBN}---e.g.@:
35542 ``@value{GCC}--2.8.1''.
35545 What compiler (and its version) was used to compile the program you are
35546 debugging---e.g.@: ``@value{GCC}--2.8.1'', or ``HP92453-01 A.10.32.03 HP
35547 C Compiler''. For @value{NGCC}, you can say @kbd{@value{GCC} --version}
35548 to get this information; for other compilers, see the documentation for
35552 The command arguments you gave the compiler to compile your example and
35553 observe the bug. For example, did you use @samp{-O}? To guarantee
35554 you will not omit something important, list them all. A copy of the
35555 Makefile (or the output from make) is sufficient.
35557 If we were to try to guess the arguments, we would probably guess wrong
35558 and then we might not encounter the bug.
35561 A complete input script, and all necessary source files, that will
35565 A description of what behavior you observe that you believe is
35566 incorrect. For example, ``It gets a fatal signal.''
35568 Of course, if the bug is that @value{GDBN} gets a fatal signal, then we
35569 will certainly notice it. But if the bug is incorrect output, we might
35570 not notice unless it is glaringly wrong. You might as well not give us
35571 a chance to make a mistake.
35573 Even if the problem you experience is a fatal signal, you should still
35574 say so explicitly. Suppose something strange is going on, such as, your
35575 copy of @value{GDBN} is out of synch, or you have encountered a bug in
35576 the C library on your system. (This has happened!) Your copy might
35577 crash and ours would not. If you told us to expect a crash, then when
35578 ours fails to crash, we would know that the bug was not happening for
35579 us. If you had not told us to expect a crash, then we would not be able
35580 to draw any conclusion from our observations.
35583 @cindex recording a session script
35584 To collect all this information, you can use a session recording program
35585 such as @command{script}, which is available on many Unix systems.
35586 Just run your @value{GDBN} session inside @command{script} and then
35587 include the @file{typescript} file with your bug report.
35589 Another way to record a @value{GDBN} session is to run @value{GDBN}
35590 inside Emacs and then save the entire buffer to a file.
35593 If you wish to suggest changes to the @value{GDBN} source, send us context
35594 diffs. If you even discuss something in the @value{GDBN} source, refer to
35595 it by context, not by line number.
35597 The line numbers in our development sources will not match those in your
35598 sources. Your line numbers would convey no useful information to us.
35602 Here are some things that are not necessary:
35606 A description of the envelope of the bug.
35608 Often people who encounter a bug spend a lot of time investigating
35609 which changes to the input file will make the bug go away and which
35610 changes will not affect it.
35612 This is often time consuming and not very useful, because the way we
35613 will find the bug is by running a single example under the debugger
35614 with breakpoints, not by pure deduction from a series of examples.
35615 We recommend that you save your time for something else.
35617 Of course, if you can find a simpler example to report @emph{instead}
35618 of the original one, that is a convenience for us. Errors in the
35619 output will be easier to spot, running under the debugger will take
35620 less time, and so on.
35622 However, simplification is not vital; if you do not want to do this,
35623 report the bug anyway and send us the entire test case you used.
35626 A patch for the bug.
35628 A patch for the bug does help us if it is a good one. But do not omit
35629 the necessary information, such as the test case, on the assumption that
35630 a patch is all we need. We might see problems with your patch and decide
35631 to fix the problem another way, or we might not understand it at all.
35633 Sometimes with a program as complicated as @value{GDBN} it is very hard to
35634 construct an example that will make the program follow a certain path
35635 through the code. If you do not send us the example, we will not be able
35636 to construct one, so we will not be able to verify that the bug is fixed.
35638 And if we cannot understand what bug you are trying to fix, or why your
35639 patch should be an improvement, we will not install it. A test case will
35640 help us to understand.
35643 A guess about what the bug is or what it depends on.
35645 Such guesses are usually wrong. Even we cannot guess right about such
35646 things without first using the debugger to find the facts.
35649 @c The readline documentation is distributed with the readline code
35650 @c and consists of the two following files:
35653 @c Use -I with makeinfo to point to the appropriate directory,
35654 @c environment var TEXINPUTS with TeX.
35655 @ifclear SYSTEM_READLINE
35656 @include rluser.texi
35657 @include hsuser.texi
35661 @appendix In Memoriam
35663 The @value{GDBN} project mourns the loss of the following long-time
35668 Fred was a long-standing contributor to @value{GDBN} (1991-2006), and
35669 to Free Software in general. Outside of @value{GDBN}, he was known in
35670 the Amiga world for his series of Fish Disks, and the GeekGadget project.
35672 @item Michael Snyder
35673 Michael was one of the Global Maintainers of the @value{GDBN} project,
35674 with contributions recorded as early as 1996, until 2011. In addition
35675 to his day to day participation, he was a large driving force behind
35676 adding Reverse Debugging to @value{GDBN}.
35679 Beyond their technical contributions to the project, they were also
35680 enjoyable members of the Free Software Community. We will miss them.
35682 @node Formatting Documentation
35683 @appendix Formatting Documentation
35685 @cindex @value{GDBN} reference card
35686 @cindex reference card
35687 The @value{GDBN} 4 release includes an already-formatted reference card, ready
35688 for printing with PostScript or Ghostscript, in the @file{gdb}
35689 subdirectory of the main source directory@footnote{In
35690 @file{gdb-@value{GDBVN}/gdb/refcard.ps} of the version @value{GDBVN}
35691 release.}. If you can use PostScript or Ghostscript with your printer,
35692 you can print the reference card immediately with @file{refcard.ps}.
35694 The release also includes the source for the reference card. You
35695 can format it, using @TeX{}, by typing:
35701 The @value{GDBN} reference card is designed to print in @dfn{landscape}
35702 mode on US ``letter'' size paper;
35703 that is, on a sheet 11 inches wide by 8.5 inches
35704 high. You will need to specify this form of printing as an option to
35705 your @sc{dvi} output program.
35707 @cindex documentation
35709 All the documentation for @value{GDBN} comes as part of the machine-readable
35710 distribution. The documentation is written in Texinfo format, which is
35711 a documentation system that uses a single source file to produce both
35712 on-line information and a printed manual. You can use one of the Info
35713 formatting commands to create the on-line version of the documentation
35714 and @TeX{} (or @code{texi2roff}) to typeset the printed version.
35716 @value{GDBN} includes an already formatted copy of the on-line Info
35717 version of this manual in the @file{gdb} subdirectory. The main Info
35718 file is @file{gdb-@value{GDBVN}/gdb/gdb.info}, and it refers to
35719 subordinate files matching @samp{gdb.info*} in the same directory. If
35720 necessary, you can print out these files, or read them with any editor;
35721 but they are easier to read using the @code{info} subsystem in @sc{gnu}
35722 Emacs or the standalone @code{info} program, available as part of the
35723 @sc{gnu} Texinfo distribution.
35725 If you want to format these Info files yourself, you need one of the
35726 Info formatting programs, such as @code{texinfo-format-buffer} or
35729 If you have @code{makeinfo} installed, and are in the top level
35730 @value{GDBN} source directory (@file{gdb-@value{GDBVN}}, in the case of
35731 version @value{GDBVN}), you can make the Info file by typing:
35738 If you want to typeset and print copies of this manual, you need @TeX{},
35739 a program to print its @sc{dvi} output files, and @file{texinfo.tex}, the
35740 Texinfo definitions file.
35742 @TeX{} is a typesetting program; it does not print files directly, but
35743 produces output files called @sc{dvi} files. To print a typeset
35744 document, you need a program to print @sc{dvi} files. If your system
35745 has @TeX{} installed, chances are it has such a program. The precise
35746 command to use depends on your system; @kbd{lpr -d} is common; another
35747 (for PostScript devices) is @kbd{dvips}. The @sc{dvi} print command may
35748 require a file name without any extension or a @samp{.dvi} extension.
35750 @TeX{} also requires a macro definitions file called
35751 @file{texinfo.tex}. This file tells @TeX{} how to typeset a document
35752 written in Texinfo format. On its own, @TeX{} cannot either read or
35753 typeset a Texinfo file. @file{texinfo.tex} is distributed with GDB
35754 and is located in the @file{gdb-@var{version-number}/texinfo}
35757 If you have @TeX{} and a @sc{dvi} printer program installed, you can
35758 typeset and print this manual. First switch to the @file{gdb}
35759 subdirectory of the main source directory (for example, to
35760 @file{gdb-@value{GDBVN}/gdb}) and type:
35766 Then give @file{gdb.dvi} to your @sc{dvi} printing program.
35768 @node Installing GDB
35769 @appendix Installing @value{GDBN}
35770 @cindex installation
35773 * Requirements:: Requirements for building @value{GDBN}
35774 * Running Configure:: Invoking the @value{GDBN} @file{configure} script
35775 * Separate Objdir:: Compiling @value{GDBN} in another directory
35776 * Config Names:: Specifying names for hosts and targets
35777 * Configure Options:: Summary of options for configure
35778 * System-wide configuration:: Having a system-wide init file
35782 @section Requirements for Building @value{GDBN}
35783 @cindex building @value{GDBN}, requirements for
35785 Building @value{GDBN} requires various tools and packages to be available.
35786 Other packages will be used only if they are found.
35788 @heading Tools/Packages Necessary for Building @value{GDBN}
35790 @item C@t{++}11 compiler
35791 @value{GDBN} is written in C@t{++}11. It should be buildable with any
35792 recent C@t{++}11 compiler, e.g.@: GCC.
35795 @value{GDBN}'s build system relies on features only found in the GNU
35796 make program. Other variants of @code{make} will not work.
35799 @heading Tools/Packages Optional for Building @value{GDBN}
35803 @value{GDBN} can use the Expat XML parsing library. This library may be
35804 included with your operating system distribution; if it is not, you
35805 can get the latest version from @url{http://expat.sourceforge.net}.
35806 The @file{configure} script will search for this library in several
35807 standard locations; if it is installed in an unusual path, you can
35808 use the @option{--with-libexpat-prefix} option to specify its location.
35814 Remote protocol memory maps (@pxref{Memory Map Format})
35816 Target descriptions (@pxref{Target Descriptions})
35818 Remote shared library lists (@xref{Library List Format},
35819 or alternatively @pxref{Library List Format for SVR4 Targets})
35821 MS-Windows shared libraries (@pxref{Shared Libraries})
35823 Traceframe info (@pxref{Traceframe Info Format})
35825 Branch trace (@pxref{Branch Trace Format},
35826 @pxref{Branch Trace Configuration Format})
35830 @value{GDBN} can be scripted using GNU Guile. @xref{Guile}. By
35831 default, @value{GDBN} will be compiled if the Guile libraries are
35832 installed and are found by @file{configure}. You can use the
35833 @code{--with-guile} option to request Guile, and pass either the Guile
35834 version number or the file name of the relevant @code{pkg-config}
35835 program to choose a particular version of Guile.
35838 @value{GDBN}'s features related to character sets (@pxref{Character
35839 Sets}) require a functioning @code{iconv} implementation. If you are
35840 on a GNU system, then this is provided by the GNU C Library. Some
35841 other systems also provide a working @code{iconv}.
35843 If @value{GDBN} is using the @code{iconv} program which is installed
35844 in a non-standard place, you will need to tell @value{GDBN} where to
35845 find it. This is done with @option{--with-iconv-bin} which specifies
35846 the directory that contains the @code{iconv} program. This program is
35847 run in order to make a list of the available character sets.
35849 On systems without @code{iconv}, you can install GNU Libiconv. If
35850 Libiconv is installed in a standard place, @value{GDBN} will
35851 automatically use it if it is needed. If you have previously
35852 installed Libiconv in a non-standard place, you can use the
35853 @option{--with-libiconv-prefix} option to @file{configure}.
35855 @value{GDBN}'s top-level @file{configure} and @file{Makefile} will
35856 arrange to build Libiconv if a directory named @file{libiconv} appears
35857 in the top-most source directory. If Libiconv is built this way, and
35858 if the operating system does not provide a suitable @code{iconv}
35859 implementation, then the just-built library will automatically be used
35860 by @value{GDBN}. One easy way to set this up is to download GNU
35861 Libiconv, unpack it inside the top-level directory of the @value{GDBN}
35862 source tree, and then rename the directory holding the Libiconv source
35863 code to @samp{libiconv}.
35866 @value{GDBN} can support debugging sections that are compressed with
35867 the LZMA library. @xref{MiniDebugInfo}. If this library is not
35868 included with your operating system, you can find it in the xz package
35869 at @url{http://tukaani.org/xz/}. If the LZMA library is available in
35870 the usual place, then the @file{configure} script will use it
35871 automatically. If it is installed in an unusual path, you can use the
35872 @option{--with-lzma-prefix} option to specify its location.
35876 @value{GDBN} can use the GNU MPFR multiple-precision floating-point
35877 library. This library may be included with your operating system
35878 distribution; if it is not, you can get the latest version from
35879 @url{http://www.mpfr.org}. The @file{configure} script will search
35880 for this library in several standard locations; if it is installed
35881 in an unusual path, you can use the @option{--with-libmpfr-prefix}
35882 option to specify its location.
35884 GNU MPFR is used to emulate target floating-point arithmetic during
35885 expression evaluation when the target uses different floating-point
35886 formats than the host. If GNU MPFR it is not available, @value{GDBN}
35887 will fall back to using host floating-point arithmetic.
35890 @value{GDBN} can be scripted using Python language. @xref{Python}.
35891 By default, @value{GDBN} will be compiled if the Python libraries are
35892 installed and are found by @file{configure}. You can use the
35893 @code{--with-python} option to request Python, and pass either the
35894 file name of the relevant @code{python} executable, or the name of the
35895 directory in which Python is installed, to choose a particular
35896 installation of Python.
35899 @cindex compressed debug sections
35900 @value{GDBN} will use the @samp{zlib} library, if available, to read
35901 compressed debug sections. Some linkers, such as GNU gold, are capable
35902 of producing binaries with compressed debug sections. If @value{GDBN}
35903 is compiled with @samp{zlib}, it will be able to read the debug
35904 information in such binaries.
35906 The @samp{zlib} library is likely included with your operating system
35907 distribution; if it is not, you can get the latest version from
35908 @url{http://zlib.net}.
35911 @node Running Configure
35912 @section Invoking the @value{GDBN} @file{configure} Script
35913 @cindex configuring @value{GDBN}
35914 @value{GDBN} comes with a @file{configure} script that automates the process
35915 of preparing @value{GDBN} for installation; you can then use @code{make} to
35916 build the @code{gdb} program.
35918 @c irrelevant in info file; it's as current as the code it lives with.
35919 @footnote{If you have a more recent version of @value{GDBN} than @value{GDBVN},
35920 look at the @file{README} file in the sources; we may have improved the
35921 installation procedures since publishing this manual.}
35924 The @value{GDBN} distribution includes all the source code you need for
35925 @value{GDBN} in a single directory, whose name is usually composed by
35926 appending the version number to @samp{gdb}.
35928 For example, the @value{GDBN} version @value{GDBVN} distribution is in the
35929 @file{gdb-@value{GDBVN}} directory. That directory contains:
35932 @item gdb-@value{GDBVN}/configure @r{(and supporting files)}
35933 script for configuring @value{GDBN} and all its supporting libraries
35935 @item gdb-@value{GDBVN}/gdb
35936 the source specific to @value{GDBN} itself
35938 @item gdb-@value{GDBVN}/bfd
35939 source for the Binary File Descriptor library
35941 @item gdb-@value{GDBVN}/include
35942 @sc{gnu} include files
35944 @item gdb-@value{GDBVN}/libiberty
35945 source for the @samp{-liberty} free software library
35947 @item gdb-@value{GDBVN}/opcodes
35948 source for the library of opcode tables and disassemblers
35950 @item gdb-@value{GDBVN}/readline
35951 source for the @sc{gnu} command-line interface
35954 There may be other subdirectories as well.
35956 The simplest way to configure and build @value{GDBN} is to run @file{configure}
35957 from the @file{gdb-@var{version-number}} source directory, which in
35958 this example is the @file{gdb-@value{GDBVN}} directory.
35960 First switch to the @file{gdb-@var{version-number}} source directory
35961 if you are not already in it; then run @file{configure}. Pass the
35962 identifier for the platform on which @value{GDBN} will run as an
35968 cd gdb-@value{GDBVN}
35973 Running @samp{configure} and then running @code{make} builds the
35974 included supporting libraries, then @code{gdb} itself. The configured
35975 source files, and the binaries, are left in the corresponding source
35979 @file{configure} is a Bourne-shell (@code{/bin/sh}) script; if your
35980 system does not recognize this automatically when you run a different
35981 shell, you may need to run @code{sh} on it explicitly:
35987 You should run the @file{configure} script from the top directory in the
35988 source tree, the @file{gdb-@var{version-number}} directory. If you run
35989 @file{configure} from one of the subdirectories, you will configure only
35990 that subdirectory. That is usually not what you want. In particular,
35991 if you run the first @file{configure} from the @file{gdb} subdirectory
35992 of the @file{gdb-@var{version-number}} directory, you will omit the
35993 configuration of @file{bfd}, @file{readline}, and other sibling
35994 directories of the @file{gdb} subdirectory. This leads to build errors
35995 about missing include files such as @file{bfd/bfd.h}.
35997 You can install @code{@value{GDBN}} anywhere. The best way to do this
35998 is to pass the @code{--prefix} option to @code{configure}, and then
35999 install it with @code{make install}.
36001 @node Separate Objdir
36002 @section Compiling @value{GDBN} in Another Directory
36004 If you want to run @value{GDBN} versions for several host or target machines,
36005 you need a different @code{gdb} compiled for each combination of
36006 host and target. @file{configure} is designed to make this easy by
36007 allowing you to generate each configuration in a separate subdirectory,
36008 rather than in the source directory. If your @code{make} program
36009 handles the @samp{VPATH} feature (@sc{gnu} @code{make} does), running
36010 @code{make} in each of these directories builds the @code{gdb}
36011 program specified there.
36013 To build @code{gdb} in a separate directory, run @file{configure}
36014 with the @samp{--srcdir} option to specify where to find the source.
36015 (You also need to specify a path to find @file{configure}
36016 itself from your working directory. If the path to @file{configure}
36017 would be the same as the argument to @samp{--srcdir}, you can leave out
36018 the @samp{--srcdir} option; it is assumed.)
36020 For example, with version @value{GDBVN}, you can build @value{GDBN} in a
36021 separate directory for a Sun 4 like this:
36025 cd gdb-@value{GDBVN}
36028 ../gdb-@value{GDBVN}/configure
36033 When @file{configure} builds a configuration using a remote source
36034 directory, it creates a tree for the binaries with the same structure
36035 (and using the same names) as the tree under the source directory. In
36036 the example, you'd find the Sun 4 library @file{libiberty.a} in the
36037 directory @file{gdb-sun4/libiberty}, and @value{GDBN} itself in
36038 @file{gdb-sun4/gdb}.
36040 Make sure that your path to the @file{configure} script has just one
36041 instance of @file{gdb} in it. If your path to @file{configure} looks
36042 like @file{../gdb-@value{GDBVN}/gdb/configure}, you are configuring only
36043 one subdirectory of @value{GDBN}, not the whole package. This leads to
36044 build errors about missing include files such as @file{bfd/bfd.h}.
36046 One popular reason to build several @value{GDBN} configurations in separate
36047 directories is to configure @value{GDBN} for cross-compiling (where
36048 @value{GDBN} runs on one machine---the @dfn{host}---while debugging
36049 programs that run on another machine---the @dfn{target}).
36050 You specify a cross-debugging target by
36051 giving the @samp{--target=@var{target}} option to @file{configure}.
36053 When you run @code{make} to build a program or library, you must run
36054 it in a configured directory---whatever directory you were in when you
36055 called @file{configure} (or one of its subdirectories).
36057 The @code{Makefile} that @file{configure} generates in each source
36058 directory also runs recursively. If you type @code{make} in a source
36059 directory such as @file{gdb-@value{GDBVN}} (or in a separate configured
36060 directory configured with @samp{--srcdir=@var{dirname}/gdb-@value{GDBVN}}), you
36061 will build all the required libraries, and then build GDB.
36063 When you have multiple hosts or targets configured in separate
36064 directories, you can run @code{make} on them in parallel (for example,
36065 if they are NFS-mounted on each of the hosts); they will not interfere
36069 @section Specifying Names for Hosts and Targets
36071 The specifications used for hosts and targets in the @file{configure}
36072 script are based on a three-part naming scheme, but some short predefined
36073 aliases are also supported. The full naming scheme encodes three pieces
36074 of information in the following pattern:
36077 @var{architecture}-@var{vendor}-@var{os}
36080 For example, you can use the alias @code{sun4} as a @var{host} argument,
36081 or as the value for @var{target} in a @code{--target=@var{target}}
36082 option. The equivalent full name is @samp{sparc-sun-sunos4}.
36084 The @file{configure} script accompanying @value{GDBN} does not provide
36085 any query facility to list all supported host and target names or
36086 aliases. @file{configure} calls the Bourne shell script
36087 @code{config.sub} to map abbreviations to full names; you can read the
36088 script, if you wish, or you can use it to test your guesses on
36089 abbreviations---for example:
36092 % sh config.sub i386-linux
36094 % sh config.sub alpha-linux
36095 alpha-unknown-linux-gnu
36096 % sh config.sub hp9k700
36098 % sh config.sub sun4
36099 sparc-sun-sunos4.1.1
36100 % sh config.sub sun3
36101 m68k-sun-sunos4.1.1
36102 % sh config.sub i986v
36103 Invalid configuration `i986v': machine `i986v' not recognized
36107 @code{config.sub} is also distributed in the @value{GDBN} source
36108 directory (@file{gdb-@value{GDBVN}}, for version @value{GDBVN}).
36110 @node Configure Options
36111 @section @file{configure} Options
36113 Here is a summary of the @file{configure} options and arguments that
36114 are most often useful for building @value{GDBN}. @file{configure}
36115 also has several other options not listed here. @inforef{Running
36116 configure scripts,,autoconf.info}, for a full
36117 explanation of @file{configure}.
36120 configure @r{[}--help@r{]}
36121 @r{[}--prefix=@var{dir}@r{]}
36122 @r{[}--exec-prefix=@var{dir}@r{]}
36123 @r{[}--srcdir=@var{dirname}@r{]}
36124 @r{[}--target=@var{target}@r{]}
36128 You may introduce options with a single @samp{-} rather than
36129 @samp{--} if you prefer; but you may abbreviate option names if you use
36134 Display a quick summary of how to invoke @file{configure}.
36136 @item --prefix=@var{dir}
36137 Configure the source to install programs and files under directory
36140 @item --exec-prefix=@var{dir}
36141 Configure the source to install programs under directory
36144 @c avoid splitting the warning from the explanation:
36146 @item --srcdir=@var{dirname}
36147 Use this option to make configurations in directories separate from the
36148 @value{GDBN} source directories. Among other things, you can use this to
36149 build (or maintain) several configurations simultaneously, in separate
36150 directories. @file{configure} writes configuration-specific files in
36151 the current directory, but arranges for them to use the source in the
36152 directory @var{dirname}. @file{configure} creates directories under
36153 the working directory in parallel to the source directories below
36156 @item --target=@var{target}
36157 Configure @value{GDBN} for cross-debugging programs running on the specified
36158 @var{target}. Without this option, @value{GDBN} is configured to debug
36159 programs that run on the same machine (@var{host}) as @value{GDBN} itself.
36161 There is no convenient way to generate a list of all available
36162 targets. Also see the @code{--enable-targets} option, below.
36165 There are many other options that are specific to @value{GDBN}. This
36166 lists just the most common ones; there are some very specialized
36167 options not described here.
36170 @item --enable-targets=@r{[}@var{target}@r{]}@dots{}
36171 @itemx --enable-targets=all
36172 Configure @value{GDBN} for cross-debugging programs running on the
36173 specified list of targets. The special value @samp{all} configures
36174 @value{GDBN} for debugging programs running on any target it supports.
36176 @item --with-gdb-datadir=@var{path}
36177 Set the @value{GDBN}-specific data directory. @value{GDBN} will look
36178 here for certain supporting files or scripts. This defaults to the
36179 @file{gdb} subdirectory of @samp{datadi} (which can be set using
36182 @item --with-relocated-sources=@var{dir}
36183 Sets up the default source path substitution rule so that directory
36184 names recorded in debug information will be automatically adjusted for
36185 any directory under @var{dir}. @var{dir} should be a subdirectory of
36186 @value{GDBN}'s configured prefix, the one mentioned in the
36187 @code{--prefix} or @code{--exec-prefix} options to configure. This
36188 option is useful if GDB is supposed to be moved to a different place
36191 @item --enable-64-bit-bfd
36192 Enable 64-bit support in BFD on 32-bit hosts.
36194 @item --disable-gdbmi
36195 Build @value{GDBN} without the GDB/MI machine interface
36199 Build @value{GDBN} with the text-mode full-screen user interface
36200 (TUI). Requires a curses library (ncurses and cursesX are also
36203 @item --with-curses
36204 Use the curses library instead of the termcap library, for text-mode
36205 terminal operations.
36207 @item --with-libunwind-ia64
36208 Use the libunwind library for unwinding function call stack on ia64
36209 target platforms. See http://www.nongnu.org/libunwind/index.html for
36212 @item --with-system-readline
36213 Use the readline library installed on the host, rather than the
36214 library supplied as part of @value{GDBN}.
36216 @item --with-system-zlib
36217 Use the zlib library installed on the host, rather than the library
36218 supplied as part of @value{GDBN}.
36221 Build @value{GDBN} with Expat, a library for XML parsing. (Done by
36222 default if libexpat is installed and found at configure time.) This
36223 library is used to read XML files supplied with @value{GDBN}. If it
36224 is unavailable, some features, such as remote protocol memory maps,
36225 target descriptions, and shared library lists, that are based on XML
36226 files, will not be available in @value{GDBN}. If your host does not
36227 have libexpat installed, you can get the latest version from
36228 `http://expat.sourceforge.net'.
36230 @item --with-libiconv-prefix@r{[}=@var{dir}@r{]}
36232 Build @value{GDBN} with GNU libiconv, a character set encoding
36233 conversion library. This is not done by default, as on GNU systems
36234 the @code{iconv} that is built in to the C library is sufficient. If
36235 your host does not have a working @code{iconv}, you can get the latest
36236 version of GNU iconv from `https://www.gnu.org/software/libiconv/'.
36238 @value{GDBN}'s build system also supports building GNU libiconv as
36239 part of the overall build. @xref{Requirements}.
36242 Build @value{GDBN} with LZMA, a compression library. (Done by default
36243 if liblzma is installed and found at configure time.) LZMA is used by
36244 @value{GDBN}'s "mini debuginfo" feature, which is only useful on
36245 platforms using the ELF object file format. If your host does not
36246 have liblzma installed, you can get the latest version from
36247 `https://tukaani.org/xz/'.
36250 Build @value{GDBN} with GNU MPFR, a library for multiple-precision
36251 floating-point computation with correct rounding. (Done by default if
36252 GNU MPFR is installed and found at configure time.) This library is
36253 used to emulate target floating-point arithmetic during expression
36254 evaluation when the target uses different floating-point formats than
36255 the host. If GNU MPFR is not available, @value{GDBN} will fall back
36256 to using host floating-point arithmetic. If your host does not have
36257 GNU MPFR installed, you can get the latest version from
36258 `http://www.mpfr.org'.
36260 @item --with-python@r{[}=@var{python}@r{]}
36261 Build @value{GDBN} with Python scripting support. (Done by default if
36262 libpython is present and found at configure time.) Python makes
36263 @value{GDBN} scripting much more powerful than the restricted CLI
36264 scripting language. If your host does not have Python installed, you
36265 can find it on `http://www.python.org/download/'. The oldest version
36266 of Python supported by GDB is 2.6. The optional argument @var{python}
36267 is used to find the Python headers and libraries. It can be either
36268 the name of a Python executable, or the name of the directory in which
36269 Python is installed.
36271 @item --with-guile[=GUILE]'
36272 Build @value{GDBN} with GNU Guile scripting support. (Done by default
36273 if libguile is present and found at configure time.) If your host
36274 does not have Guile installed, you can find it at
36275 `https://www.gnu.org/software/guile/'. The optional argument GUILE
36276 can be a version number, which will cause @code{configure} to try to
36277 use that version of Guile; or the file name of a @code{pkg-config}
36278 executable, which will be queried to find the information needed to
36279 compile and link against Guile.
36281 @item --without-included-regex
36282 Don't use the regex library included with @value{GDBN} (as part of the
36283 libiberty library). This is the default on hosts with version 2 of
36286 @item --with-sysroot=@var{dir}
36287 Use @var{dir} as the default system root directory for libraries whose
36288 file names begin with @file{/lib}' or @file{/usr/lib'}. (The value of
36289 @var{dir} can be modified at run time by using the @command{set
36290 sysroot} command.) If @var{dir} is under the @value{GDBN} configured
36291 prefix (set with @code{--prefix} or @code{--exec-prefix options}, the
36292 default system root will be automatically adjusted if and when
36293 @value{GDBN} is moved to a different location.
36295 @item --with-system-gdbinit=@var{file}
36296 Configure @value{GDBN} to automatically load a system-wide init file.
36297 @var{file} should be an absolute file name. If @var{file} is in a
36298 directory under the configured prefix, and @value{GDBN} is moved to
36299 another location after being built, the location of the system-wide
36300 init file will be adjusted accordingly.
36302 @item --enable-build-warnings
36303 When building the @value{GDBN} sources, ask the compiler to warn about
36304 any code which looks even vaguely suspicious. It passes many
36305 different warning flags, depending on the exact version of the
36306 compiler you are using.
36308 @item --enable-werror
36309 Treat compiler warnings as werrors. It adds the @code{-Werror} flag
36310 to the compiler, which will fail the compilation if the compiler
36311 outputs any warning messages.
36313 @item --enable-ubsan
36314 Enable the GCC undefined behavior sanitizer. This is disabled by
36315 default, but passing @code{--enable-ubsan=yes} or
36316 @code{--enable-ubsan=auto} to @code{configure} will enable it. The
36317 undefined behavior sanitizer checks for C@t{++} undefined behavior.
36318 It has a performance cost, so if you are looking at @value{GDBN}'s
36319 performance, you should disable it. The undefined behavior sanitizer
36320 was first introduced in GCC 4.9.
36323 @node System-wide configuration
36324 @section System-wide configuration and settings
36325 @cindex system-wide init file
36327 @value{GDBN} can be configured to have a system-wide init file;
36328 this file will be read and executed at startup (@pxref{Startup, , What
36329 @value{GDBN} does during startup}).
36331 Here is the corresponding configure option:
36334 @item --with-system-gdbinit=@var{file}
36335 Specify that the default location of the system-wide init file is
36339 If @value{GDBN} has been configured with the option @option{--prefix=$prefix},
36340 it may be subject to relocation. Two possible cases:
36344 If the default location of this init file contains @file{$prefix},
36345 it will be subject to relocation. Suppose that the configure options
36346 are @option{--prefix=$prefix --with-system-gdbinit=$prefix/etc/gdbinit};
36347 if @value{GDBN} is moved from @file{$prefix} to @file{$install}, the system
36348 init file is looked for as @file{$install/etc/gdbinit} instead of
36349 @file{$prefix/etc/gdbinit}.
36352 By contrast, if the default location does not contain the prefix,
36353 it will not be relocated. E.g.@: if @value{GDBN} has been configured with
36354 @option{--prefix=/usr/local --with-system-gdbinit=/usr/share/gdb/gdbinit},
36355 then @value{GDBN} will always look for @file{/usr/share/gdb/gdbinit},
36356 wherever @value{GDBN} is installed.
36359 If the configured location of the system-wide init file (as given by the
36360 @option{--with-system-gdbinit} option at configure time) is in the
36361 data-directory (as specified by @option{--with-gdb-datadir} at configure
36362 time) or in one of its subdirectories, then @value{GDBN} will look for the
36363 system-wide init file in the directory specified by the
36364 @option{--data-directory} command-line option.
36365 Note that the system-wide init file is only read once, during @value{GDBN}
36366 initialization. If the data-directory is changed after @value{GDBN} has
36367 started with the @code{set data-directory} command, the file will not be
36371 * System-wide Configuration Scripts:: Installed System-wide Configuration Scripts
36374 @node System-wide Configuration Scripts
36375 @subsection Installed System-wide Configuration Scripts
36376 @cindex system-wide configuration scripts
36378 The @file{system-gdbinit} directory, located inside the data-directory
36379 (as specified by @option{--with-gdb-datadir} at configure time) contains
36380 a number of scripts which can be used as system-wide init files. To
36381 automatically source those scripts at startup, @value{GDBN} should be
36382 configured with @option{--with-system-gdbinit}. Otherwise, any user
36383 should be able to source them by hand as needed.
36385 The following scripts are currently available:
36388 @item @file{elinos.py}
36390 @cindex ELinOS system-wide configuration script
36391 This script is useful when debugging a program on an ELinOS target.
36392 It takes advantage of the environment variables defined in a standard
36393 ELinOS environment in order to determine the location of the system
36394 shared libraries, and then sets the @samp{solib-absolute-prefix}
36395 and @samp{solib-search-path} variables appropriately.
36397 @item @file{wrs-linux.py}
36398 @pindex wrs-linux.py
36399 @cindex Wind River Linux system-wide configuration script
36400 This script is useful when debugging a program on a target running
36401 Wind River Linux. It expects the @env{ENV_PREFIX} to be set to
36402 the host-side sysroot used by the target system.
36406 @node Maintenance Commands
36407 @appendix Maintenance Commands
36408 @cindex maintenance commands
36409 @cindex internal commands
36411 In addition to commands intended for @value{GDBN} users, @value{GDBN}
36412 includes a number of commands intended for @value{GDBN} developers,
36413 that are not documented elsewhere in this manual. These commands are
36414 provided here for reference. (For commands that turn on debugging
36415 messages, see @ref{Debugging Output}.)
36418 @kindex maint agent
36419 @kindex maint agent-eval
36420 @item maint agent @r{[}-at @var{location}@r{,}@r{]} @var{expression}
36421 @itemx maint agent-eval @r{[}-at @var{location}@r{,}@r{]} @var{expression}
36422 Translate the given @var{expression} into remote agent bytecodes.
36423 This command is useful for debugging the Agent Expression mechanism
36424 (@pxref{Agent Expressions}). The @samp{agent} version produces an
36425 expression useful for data collection, such as by tracepoints, while
36426 @samp{maint agent-eval} produces an expression that evaluates directly
36427 to a result. For instance, a collection expression for @code{globa +
36428 globb} will include bytecodes to record four bytes of memory at each
36429 of the addresses of @code{globa} and @code{globb}, while discarding
36430 the result of the addition, while an evaluation expression will do the
36431 addition and return the sum.
36432 If @code{-at} is given, generate remote agent bytecode for @var{location}.
36433 If not, generate remote agent bytecode for current frame PC address.
36435 @kindex maint agent-printf
36436 @item maint agent-printf @var{format},@var{expr},...
36437 Translate the given format string and list of argument expressions
36438 into remote agent bytecodes and display them as a disassembled list.
36439 This command is useful for debugging the agent version of dynamic
36440 printf (@pxref{Dynamic Printf}).
36442 @kindex maint info breakpoints
36443 @item @anchor{maint info breakpoints}maint info breakpoints
36444 Using the same format as @samp{info breakpoints}, display both the
36445 breakpoints you've set explicitly, and those @value{GDBN} is using for
36446 internal purposes. Internal breakpoints are shown with negative
36447 breakpoint numbers. The type column identifies what kind of breakpoint
36452 Normal, explicitly set breakpoint.
36455 Normal, explicitly set watchpoint.
36458 Internal breakpoint, used to handle correctly stepping through
36459 @code{longjmp} calls.
36461 @item longjmp resume
36462 Internal breakpoint at the target of a @code{longjmp}.
36465 Temporary internal breakpoint used by the @value{GDBN} @code{until} command.
36468 Temporary internal breakpoint used by the @value{GDBN} @code{finish} command.
36471 Shared library events.
36475 @kindex maint info btrace
36476 @item maint info btrace
36477 Pint information about raw branch tracing data.
36479 @kindex maint btrace packet-history
36480 @item maint btrace packet-history
36481 Print the raw branch trace packets that are used to compute the
36482 execution history for the @samp{record btrace} command. Both the
36483 information and the format in which it is printed depend on the btrace
36488 For the BTS recording format, print a list of blocks of sequential
36489 code. For each block, the following information is printed:
36493 Newer blocks have higher numbers. The oldest block has number zero.
36494 @item Lowest @samp{PC}
36495 @item Highest @samp{PC}
36499 For the Intel Processor Trace recording format, print a list of
36500 Intel Processor Trace packets. For each packet, the following
36501 information is printed:
36504 @item Packet number
36505 Newer packets have higher numbers. The oldest packet has number zero.
36507 The packet's offset in the trace stream.
36508 @item Packet opcode and payload
36512 @kindex maint btrace clear-packet-history
36513 @item maint btrace clear-packet-history
36514 Discards the cached packet history printed by the @samp{maint btrace
36515 packet-history} command. The history will be computed again when
36518 @kindex maint btrace clear
36519 @item maint btrace clear
36520 Discard the branch trace data. The data will be fetched anew and the
36521 branch trace will be recomputed when needed.
36523 This implicitly truncates the branch trace to a single branch trace
36524 buffer. When updating branch trace incrementally, the branch trace
36525 available to @value{GDBN} may be bigger than a single branch trace
36528 @kindex maint set btrace pt skip-pad
36529 @item maint set btrace pt skip-pad
36530 @kindex maint show btrace pt skip-pad
36531 @item maint show btrace pt skip-pad
36532 Control whether @value{GDBN} will skip PAD packets when computing the
36535 @kindex set displaced-stepping
36536 @kindex show displaced-stepping
36537 @cindex displaced stepping support
36538 @cindex out-of-line single-stepping
36539 @item set displaced-stepping
36540 @itemx show displaced-stepping
36541 Control whether or not @value{GDBN} will do @dfn{displaced stepping}
36542 if the target supports it. Displaced stepping is a way to single-step
36543 over breakpoints without removing them from the inferior, by executing
36544 an out-of-line copy of the instruction that was originally at the
36545 breakpoint location. It is also known as out-of-line single-stepping.
36548 @item set displaced-stepping on
36549 If the target architecture supports it, @value{GDBN} will use
36550 displaced stepping to step over breakpoints.
36552 @item set displaced-stepping off
36553 @value{GDBN} will not use displaced stepping to step over breakpoints,
36554 even if such is supported by the target architecture.
36556 @cindex non-stop mode, and @samp{set displaced-stepping}
36557 @item set displaced-stepping auto
36558 This is the default mode. @value{GDBN} will use displaced stepping
36559 only if non-stop mode is active (@pxref{Non-Stop Mode}) and the target
36560 architecture supports displaced stepping.
36563 @kindex maint check-psymtabs
36564 @item maint check-psymtabs
36565 Check the consistency of currently expanded psymtabs versus symtabs.
36566 Use this to check, for example, whether a symbol is in one but not the other.
36568 @kindex maint check-symtabs
36569 @item maint check-symtabs
36570 Check the consistency of currently expanded symtabs.
36572 @kindex maint expand-symtabs
36573 @item maint expand-symtabs [@var{regexp}]
36574 Expand symbol tables.
36575 If @var{regexp} is specified, only expand symbol tables for file
36576 names matching @var{regexp}.
36578 @kindex maint set catch-demangler-crashes
36579 @kindex maint show catch-demangler-crashes
36580 @cindex demangler crashes
36581 @item maint set catch-demangler-crashes [on|off]
36582 @itemx maint show catch-demangler-crashes
36583 Control whether @value{GDBN} should attempt to catch crashes in the
36584 symbol name demangler. The default is to attempt to catch crashes.
36585 If enabled, the first time a crash is caught, a core file is created,
36586 the offending symbol is displayed and the user is presented with the
36587 option to terminate the current session.
36589 @kindex maint cplus first_component
36590 @item maint cplus first_component @var{name}
36591 Print the first C@t{++} class/namespace component of @var{name}.
36593 @kindex maint cplus namespace
36594 @item maint cplus namespace
36595 Print the list of possible C@t{++} namespaces.
36597 @kindex maint deprecate
36598 @kindex maint undeprecate
36599 @cindex deprecated commands
36600 @item maint deprecate @var{command} @r{[}@var{replacement}@r{]}
36601 @itemx maint undeprecate @var{command}
36602 Deprecate or undeprecate the named @var{command}. Deprecated commands
36603 cause @value{GDBN} to issue a warning when you use them. The optional
36604 argument @var{replacement} says which newer command should be used in
36605 favor of the deprecated one; if it is given, @value{GDBN} will mention
36606 the replacement as part of the warning.
36608 @kindex maint dump-me
36609 @item maint dump-me
36610 @cindex @code{SIGQUIT} signal, dump core of @value{GDBN}
36611 Cause a fatal signal in the debugger and force it to dump its core.
36612 This is supported only on systems which support aborting a program
36613 with the @code{SIGQUIT} signal.
36615 @kindex maint internal-error
36616 @kindex maint internal-warning
36617 @kindex maint demangler-warning
36618 @cindex demangler crashes
36619 @item maint internal-error @r{[}@var{message-text}@r{]}
36620 @itemx maint internal-warning @r{[}@var{message-text}@r{]}
36621 @itemx maint demangler-warning @r{[}@var{message-text}@r{]}
36623 Cause @value{GDBN} to call the internal function @code{internal_error},
36624 @code{internal_warning} or @code{demangler_warning} and hence behave
36625 as though an internal problem has been detected. In addition to
36626 reporting the internal problem, these functions give the user the
36627 opportunity to either quit @value{GDBN} or (for @code{internal_error}
36628 and @code{internal_warning}) create a core file of the current
36629 @value{GDBN} session.
36631 These commands take an optional parameter @var{message-text} that is
36632 used as the text of the error or warning message.
36634 Here's an example of using @code{internal-error}:
36637 (@value{GDBP}) @kbd{maint internal-error testing, 1, 2}
36638 @dots{}/maint.c:121: internal-error: testing, 1, 2
36639 A problem internal to GDB has been detected. Further
36640 debugging may prove unreliable.
36641 Quit this debugging session? (y or n) @kbd{n}
36642 Create a core file? (y or n) @kbd{n}
36646 @cindex @value{GDBN} internal error
36647 @cindex internal errors, control of @value{GDBN} behavior
36648 @cindex demangler crashes
36650 @kindex maint set internal-error
36651 @kindex maint show internal-error
36652 @kindex maint set internal-warning
36653 @kindex maint show internal-warning
36654 @kindex maint set demangler-warning
36655 @kindex maint show demangler-warning
36656 @item maint set internal-error @var{action} [ask|yes|no]
36657 @itemx maint show internal-error @var{action}
36658 @itemx maint set internal-warning @var{action} [ask|yes|no]
36659 @itemx maint show internal-warning @var{action}
36660 @itemx maint set demangler-warning @var{action} [ask|yes|no]
36661 @itemx maint show demangler-warning @var{action}
36662 When @value{GDBN} reports an internal problem (error or warning) it
36663 gives the user the opportunity to both quit @value{GDBN} and create a
36664 core file of the current @value{GDBN} session. These commands let you
36665 override the default behaviour for each particular @var{action},
36666 described in the table below.
36670 You can specify that @value{GDBN} should always (yes) or never (no)
36671 quit. The default is to ask the user what to do.
36674 You can specify that @value{GDBN} should always (yes) or never (no)
36675 create a core file. The default is to ask the user what to do. Note
36676 that there is no @code{corefile} option for @code{demangler-warning}:
36677 demangler warnings always create a core file and this cannot be
36681 @kindex maint packet
36682 @item maint packet @var{text}
36683 If @value{GDBN} is talking to an inferior via the serial protocol,
36684 then this command sends the string @var{text} to the inferior, and
36685 displays the response packet. @value{GDBN} supplies the initial
36686 @samp{$} character, the terminating @samp{#} character, and the
36689 @kindex maint print architecture
36690 @item maint print architecture @r{[}@var{file}@r{]}
36691 Print the entire architecture configuration. The optional argument
36692 @var{file} names the file where the output goes.
36694 @kindex maint print c-tdesc @r{[}@var{file}@r{]}
36695 @item maint print c-tdesc
36696 Print the target description (@pxref{Target Descriptions}) as
36697 a C source file. By default, the target description is for the current
36698 target, but if the optional argument @var{file} is provided, that file
36699 is used to produce the description. The @var{file} should be an XML
36700 document, of the form described in @ref{Target Description Format}.
36701 The created source file is built into @value{GDBN} when @value{GDBN} is
36702 built again. This command is used by developers after they add or
36703 modify XML target descriptions.
36705 @kindex maint check xml-descriptions
36706 @item maint check xml-descriptions @var{dir}
36707 Check that the target descriptions dynamically created by @value{GDBN}
36708 equal the descriptions created from XML files found in @var{dir}.
36710 @anchor{maint check libthread-db}
36711 @kindex maint check libthread-db
36712 @item maint check libthread-db
36713 Run integrity checks on the current inferior's thread debugging
36714 library. This exercises all @code{libthread_db} functionality used by
36715 @value{GDBN} on GNU/Linux systems, and by extension also exercises the
36716 @code{proc_service} functions provided by @value{GDBN} that
36717 @code{libthread_db} uses. Note that parts of the test may be skipped
36718 on some platforms when debugging core files.
36720 @kindex maint print dummy-frames
36721 @item maint print dummy-frames
36722 Prints the contents of @value{GDBN}'s internal dummy-frame stack.
36725 (@value{GDBP}) @kbd{b add}
36727 (@value{GDBP}) @kbd{print add(2,3)}
36728 Breakpoint 2, add (a=2, b=3) at @dots{}
36730 The program being debugged stopped while in a function called from GDB.
36732 (@value{GDBP}) @kbd{maint print dummy-frames}
36733 0xa8206d8: id=@{stack=0xbfffe734,code=0xbfffe73f,!special@}, ptid=process 9353
36737 Takes an optional file parameter.
36739 @kindex maint print registers
36740 @kindex maint print raw-registers
36741 @kindex maint print cooked-registers
36742 @kindex maint print register-groups
36743 @kindex maint print remote-registers
36744 @item maint print registers @r{[}@var{file}@r{]}
36745 @itemx maint print raw-registers @r{[}@var{file}@r{]}
36746 @itemx maint print cooked-registers @r{[}@var{file}@r{]}
36747 @itemx maint print register-groups @r{[}@var{file}@r{]}
36748 @itemx maint print remote-registers @r{[}@var{file}@r{]}
36749 Print @value{GDBN}'s internal register data structures.
36751 The command @code{maint print raw-registers} includes the contents of
36752 the raw register cache; the command @code{maint print
36753 cooked-registers} includes the (cooked) value of all registers,
36754 including registers which aren't available on the target nor visible
36755 to user; the command @code{maint print register-groups} includes the
36756 groups that each register is a member of; and the command @code{maint
36757 print remote-registers} includes the remote target's register numbers
36758 and offsets in the `G' packets.
36760 These commands take an optional parameter, a file name to which to
36761 write the information.
36763 @kindex maint print reggroups
36764 @item maint print reggroups @r{[}@var{file}@r{]}
36765 Print @value{GDBN}'s internal register group data structures. The
36766 optional argument @var{file} tells to what file to write the
36769 The register groups info looks like this:
36772 (@value{GDBP}) @kbd{maint print reggroups}
36785 This command forces @value{GDBN} to flush its internal register cache.
36787 @kindex maint print objfiles
36788 @cindex info for known object files
36789 @item maint print objfiles @r{[}@var{regexp}@r{]}
36790 Print a dump of all known object files.
36791 If @var{regexp} is specified, only print object files whose names
36792 match @var{regexp}. For each object file, this command prints its name,
36793 address in memory, and all of its psymtabs and symtabs.
36795 @kindex maint print user-registers
36796 @cindex user registers
36797 @item maint print user-registers
36798 List all currently available @dfn{user registers}. User registers
36799 typically provide alternate names for actual hardware registers. They
36800 include the four ``standard'' registers @code{$fp}, @code{$pc},
36801 @code{$sp}, and @code{$ps}. @xref{standard registers}. User
36802 registers can be used in expressions in the same way as the canonical
36803 register names, but only the latter are listed by the @code{info
36804 registers} and @code{maint print registers} commands.
36806 @kindex maint print section-scripts
36807 @cindex info for known .debug_gdb_scripts-loaded scripts
36808 @item maint print section-scripts [@var{regexp}]
36809 Print a dump of scripts specified in the @code{.debug_gdb_section} section.
36810 If @var{regexp} is specified, only print scripts loaded by object files
36811 matching @var{regexp}.
36812 For each script, this command prints its name as specified in the objfile,
36813 and the full path if known.
36814 @xref{dotdebug_gdb_scripts section}.
36816 @kindex maint print statistics
36817 @cindex bcache statistics
36818 @item maint print statistics
36819 This command prints, for each object file in the program, various data
36820 about that object file followed by the byte cache (@dfn{bcache})
36821 statistics for the object file. The objfile data includes the number
36822 of minimal, partial, full, and stabs symbols, the number of types
36823 defined by the objfile, the number of as yet unexpanded psym tables,
36824 the number of line tables and string tables, and the amount of memory
36825 used by the various tables. The bcache statistics include the counts,
36826 sizes, and counts of duplicates of all and unique objects, max,
36827 average, and median entry size, total memory used and its overhead and
36828 savings, and various measures of the hash table size and chain
36831 @kindex maint print target-stack
36832 @cindex target stack description
36833 @item maint print target-stack
36834 A @dfn{target} is an interface between the debugger and a particular
36835 kind of file or process. Targets can be stacked in @dfn{strata},
36836 so that more than one target can potentially respond to a request.
36837 In particular, memory accesses will walk down the stack of targets
36838 until they find a target that is interested in handling that particular
36841 This command prints a short description of each layer that was pushed on
36842 the @dfn{target stack}, starting from the top layer down to the bottom one.
36844 @kindex maint print type
36845 @cindex type chain of a data type
36846 @item maint print type @var{expr}
36847 Print the type chain for a type specified by @var{expr}. The argument
36848 can be either a type name or a symbol. If it is a symbol, the type of
36849 that symbol is described. The type chain produced by this command is
36850 a recursive definition of the data type as stored in @value{GDBN}'s
36851 data structures, including its flags and contained types.
36853 @kindex maint selftest
36855 @item maint selftest @r{[}@var{filter}@r{]}
36856 Run any self tests that were compiled in to @value{GDBN}. This will
36857 print a message showing how many tests were run, and how many failed.
36858 If a @var{filter} is passed, only the tests with @var{filter} in their
36861 @kindex maint info selftests
36863 @item maint info selftests
36864 List the selftests compiled in to @value{GDBN}.
36866 @kindex maint set dwarf always-disassemble
36867 @kindex maint show dwarf always-disassemble
36868 @item maint set dwarf always-disassemble
36869 @item maint show dwarf always-disassemble
36870 Control the behavior of @code{info address} when using DWARF debugging
36873 The default is @code{off}, which means that @value{GDBN} should try to
36874 describe a variable's location in an easily readable format. When
36875 @code{on}, @value{GDBN} will instead display the DWARF location
36876 expression in an assembly-like format. Note that some locations are
36877 too complex for @value{GDBN} to describe simply; in this case you will
36878 always see the disassembly form.
36880 Here is an example of the resulting disassembly:
36883 (gdb) info addr argc
36884 Symbol "argc" is a complex DWARF expression:
36888 For more information on these expressions, see
36889 @uref{http://www.dwarfstd.org/, the DWARF standard}.
36891 @kindex maint set dwarf max-cache-age
36892 @kindex maint show dwarf max-cache-age
36893 @item maint set dwarf max-cache-age
36894 @itemx maint show dwarf max-cache-age
36895 Control the DWARF compilation unit cache.
36897 @cindex DWARF compilation units cache
36898 In object files with inter-compilation-unit references, such as those
36899 produced by the GCC option @samp{-feliminate-dwarf2-dups}, the DWARF
36900 reader needs to frequently refer to previously read compilation units.
36901 This setting controls how long a compilation unit will remain in the
36902 cache if it is not referenced. A higher limit means that cached
36903 compilation units will be stored in memory longer, and more total
36904 memory will be used. Setting it to zero disables caching, which will
36905 slow down @value{GDBN} startup, but reduce memory consumption.
36907 @kindex maint set dwarf unwinders
36908 @kindex maint show dwarf unwinders
36909 @item maint set dwarf unwinders
36910 @itemx maint show dwarf unwinders
36911 Control use of the DWARF frame unwinders.
36913 @cindex DWARF frame unwinders
36914 Many targets that support DWARF debugging use @value{GDBN}'s DWARF
36915 frame unwinders to build the backtrace. Many of these targets will
36916 also have a second mechanism for building the backtrace for use in
36917 cases where DWARF information is not available, this second mechanism
36918 is often an analysis of a function's prologue.
36920 In order to extend testing coverage of the second level stack
36921 unwinding mechanisms it is helpful to be able to disable the DWARF
36922 stack unwinders, this can be done with this switch.
36924 In normal use of @value{GDBN} disabling the DWARF unwinders is not
36925 advisable, there are cases that are better handled through DWARF than
36926 prologue analysis, and the debug experience is likely to be better
36927 with the DWARF frame unwinders enabled.
36929 If DWARF frame unwinders are not supported for a particular target
36930 architecture, then enabling this flag does not cause them to be used.
36931 @kindex maint set profile
36932 @kindex maint show profile
36933 @cindex profiling GDB
36934 @item maint set profile
36935 @itemx maint show profile
36936 Control profiling of @value{GDBN}.
36938 Profiling will be disabled until you use the @samp{maint set profile}
36939 command to enable it. When you enable profiling, the system will begin
36940 collecting timing and execution count data; when you disable profiling or
36941 exit @value{GDBN}, the results will be written to a log file. Remember that
36942 if you use profiling, @value{GDBN} will overwrite the profiling log file
36943 (often called @file{gmon.out}). If you have a record of important profiling
36944 data in a @file{gmon.out} file, be sure to move it to a safe location.
36946 Configuring with @samp{--enable-profiling} arranges for @value{GDBN} to be
36947 compiled with the @samp{-pg} compiler option.
36949 @kindex maint set show-debug-regs
36950 @kindex maint show show-debug-regs
36951 @cindex hardware debug registers
36952 @item maint set show-debug-regs
36953 @itemx maint show show-debug-regs
36954 Control whether to show variables that mirror the hardware debug
36955 registers. Use @code{on} to enable, @code{off} to disable. If
36956 enabled, the debug registers values are shown when @value{GDBN} inserts or
36957 removes a hardware breakpoint or watchpoint, and when the inferior
36958 triggers a hardware-assisted breakpoint or watchpoint.
36960 @kindex maint set show-all-tib
36961 @kindex maint show show-all-tib
36962 @item maint set show-all-tib
36963 @itemx maint show show-all-tib
36964 Control whether to show all non zero areas within a 1k block starting
36965 at thread local base, when using the @samp{info w32 thread-information-block}
36968 @kindex maint set target-async
36969 @kindex maint show target-async
36970 @item maint set target-async
36971 @itemx maint show target-async
36972 This controls whether @value{GDBN} targets operate in synchronous or
36973 asynchronous mode (@pxref{Background Execution}). Normally the
36974 default is asynchronous, if it is available; but this can be changed
36975 to more easily debug problems occurring only in synchronous mode.
36977 @kindex maint set target-non-stop @var{mode} [on|off|auto]
36978 @kindex maint show target-non-stop
36979 @item maint set target-non-stop
36980 @itemx maint show target-non-stop
36982 This controls whether @value{GDBN} targets always operate in non-stop
36983 mode even if @code{set non-stop} is @code{off} (@pxref{Non-Stop
36984 Mode}). The default is @code{auto}, meaning non-stop mode is enabled
36985 if supported by the target.
36988 @item maint set target-non-stop auto
36989 This is the default mode. @value{GDBN} controls the target in
36990 non-stop mode if the target supports it.
36992 @item maint set target-non-stop on
36993 @value{GDBN} controls the target in non-stop mode even if the target
36994 does not indicate support.
36996 @item maint set target-non-stop off
36997 @value{GDBN} does not control the target in non-stop mode even if the
36998 target supports it.
37001 @kindex maint set per-command
37002 @kindex maint show per-command
37003 @item maint set per-command
37004 @itemx maint show per-command
37005 @cindex resources used by commands
37007 @value{GDBN} can display the resources used by each command.
37008 This is useful in debugging performance problems.
37011 @item maint set per-command space [on|off]
37012 @itemx maint show per-command space
37013 Enable or disable the printing of the memory used by GDB for each command.
37014 If enabled, @value{GDBN} will display how much memory each command
37015 took, following the command's own output.
37016 This can also be requested by invoking @value{GDBN} with the
37017 @option{--statistics} command-line switch (@pxref{Mode Options}).
37019 @item maint set per-command time [on|off]
37020 @itemx maint show per-command time
37021 Enable or disable the printing of the execution time of @value{GDBN}
37023 If enabled, @value{GDBN} will display how much time it
37024 took to execute each command, following the command's own output.
37025 Both CPU time and wallclock time are printed.
37026 Printing both is useful when trying to determine whether the cost is
37027 CPU or, e.g., disk/network latency.
37028 Note that the CPU time printed is for @value{GDBN} only, it does not include
37029 the execution time of the inferior because there's no mechanism currently
37030 to compute how much time was spent by @value{GDBN} and how much time was
37031 spent by the program been debugged.
37032 This can also be requested by invoking @value{GDBN} with the
37033 @option{--statistics} command-line switch (@pxref{Mode Options}).
37035 @item maint set per-command symtab [on|off]
37036 @itemx maint show per-command symtab
37037 Enable or disable the printing of basic symbol table statistics
37039 If enabled, @value{GDBN} will display the following information:
37043 number of symbol tables
37045 number of primary symbol tables
37047 number of blocks in the blockvector
37051 @kindex maint set check-libthread-db
37052 @kindex maint show check-libthread-db
37053 @item maint set check-libthread-db [on|off]
37054 @itemx maint show check-libthread-db
37055 Control whether @value{GDBN} should run integrity checks on inferior
37056 specific thread debugging libraries as they are loaded. The default
37057 is not to perform such checks. If any check fails @value{GDBN} will
37058 unload the library and continue searching for a suitable candidate as
37059 described in @ref{set libthread-db-search-path}. For more information
37060 about the tests, see @ref{maint check libthread-db}.
37062 @kindex maint space
37063 @cindex memory used by commands
37064 @item maint space @var{value}
37065 An alias for @code{maint set per-command space}.
37066 A non-zero value enables it, zero disables it.
37069 @cindex time of command execution
37070 @item maint time @var{value}
37071 An alias for @code{maint set per-command time}.
37072 A non-zero value enables it, zero disables it.
37074 @kindex maint translate-address
37075 @item maint translate-address @r{[}@var{section}@r{]} @var{addr}
37076 Find the symbol stored at the location specified by the address
37077 @var{addr} and an optional section name @var{section}. If found,
37078 @value{GDBN} prints the name of the closest symbol and an offset from
37079 the symbol's location to the specified address. This is similar to
37080 the @code{info address} command (@pxref{Symbols}), except that this
37081 command also allows to find symbols in other sections.
37083 If section was not specified, the section in which the symbol was found
37084 is also printed. For dynamically linked executables, the name of
37085 executable or shared library containing the symbol is printed as well.
37089 The following command is useful for non-interactive invocations of
37090 @value{GDBN}, such as in the test suite.
37093 @item set watchdog @var{nsec}
37094 @kindex set watchdog
37095 @cindex watchdog timer
37096 @cindex timeout for commands
37097 Set the maximum number of seconds @value{GDBN} will wait for the
37098 target operation to finish. If this time expires, @value{GDBN}
37099 reports and error and the command is aborted.
37101 @item show watchdog
37102 Show the current setting of the target wait timeout.
37105 @node Remote Protocol
37106 @appendix @value{GDBN} Remote Serial Protocol
37111 * Stop Reply Packets::
37112 * General Query Packets::
37113 * Architecture-Specific Protocol Details::
37114 * Tracepoint Packets::
37115 * Host I/O Packets::
37117 * Notification Packets::
37118 * Remote Non-Stop::
37119 * Packet Acknowledgment::
37121 * File-I/O Remote Protocol Extension::
37122 * Library List Format::
37123 * Library List Format for SVR4 Targets::
37124 * Memory Map Format::
37125 * Thread List Format::
37126 * Traceframe Info Format::
37127 * Branch Trace Format::
37128 * Branch Trace Configuration Format::
37134 There may be occasions when you need to know something about the
37135 protocol---for example, if there is only one serial port to your target
37136 machine, you might want your program to do something special if it
37137 recognizes a packet meant for @value{GDBN}.
37139 In the examples below, @samp{->} and @samp{<-} are used to indicate
37140 transmitted and received data, respectively.
37142 @cindex protocol, @value{GDBN} remote serial
37143 @cindex serial protocol, @value{GDBN} remote
37144 @cindex remote serial protocol
37145 All @value{GDBN} commands and responses (other than acknowledgments
37146 and notifications, see @ref{Notification Packets}) are sent as a
37147 @var{packet}. A @var{packet} is introduced with the character
37148 @samp{$}, the actual @var{packet-data}, and the terminating character
37149 @samp{#} followed by a two-digit @var{checksum}:
37152 @code{$}@var{packet-data}@code{#}@var{checksum}
37156 @cindex checksum, for @value{GDBN} remote
37158 The two-digit @var{checksum} is computed as the modulo 256 sum of all
37159 characters between the leading @samp{$} and the trailing @samp{#} (an
37160 eight bit unsigned checksum).
37162 Implementors should note that prior to @value{GDBN} 5.0 the protocol
37163 specification also included an optional two-digit @var{sequence-id}:
37166 @code{$}@var{sequence-id}@code{:}@var{packet-data}@code{#}@var{checksum}
37169 @cindex sequence-id, for @value{GDBN} remote
37171 That @var{sequence-id} was appended to the acknowledgment. @value{GDBN}
37172 has never output @var{sequence-id}s. Stubs that handle packets added
37173 since @value{GDBN} 5.0 must not accept @var{sequence-id}.
37175 When either the host or the target machine receives a packet, the first
37176 response expected is an acknowledgment: either @samp{+} (to indicate
37177 the package was received correctly) or @samp{-} (to request
37181 -> @code{$}@var{packet-data}@code{#}@var{checksum}
37186 The @samp{+}/@samp{-} acknowledgments can be disabled
37187 once a connection is established.
37188 @xref{Packet Acknowledgment}, for details.
37190 The host (@value{GDBN}) sends @var{command}s, and the target (the
37191 debugging stub incorporated in your program) sends a @var{response}. In
37192 the case of step and continue @var{command}s, the response is only sent
37193 when the operation has completed, and the target has again stopped all
37194 threads in all attached processes. This is the default all-stop mode
37195 behavior, but the remote protocol also supports @value{GDBN}'s non-stop
37196 execution mode; see @ref{Remote Non-Stop}, for details.
37198 @var{packet-data} consists of a sequence of characters with the
37199 exception of @samp{#} and @samp{$} (see @samp{X} packet for additional
37202 @cindex remote protocol, field separator
37203 Fields within the packet should be separated using @samp{,} @samp{;} or
37204 @samp{:}. Except where otherwise noted all numbers are represented in
37205 @sc{hex} with leading zeros suppressed.
37207 Implementors should note that prior to @value{GDBN} 5.0, the character
37208 @samp{:} could not appear as the third character in a packet (as it
37209 would potentially conflict with the @var{sequence-id}).
37211 @cindex remote protocol, binary data
37212 @anchor{Binary Data}
37213 Binary data in most packets is encoded either as two hexadecimal
37214 digits per byte of binary data. This allowed the traditional remote
37215 protocol to work over connections which were only seven-bit clean.
37216 Some packets designed more recently assume an eight-bit clean
37217 connection, and use a more efficient encoding to send and receive
37220 The binary data representation uses @code{7d} (@sc{ascii} @samp{@}})
37221 as an escape character. Any escaped byte is transmitted as the escape
37222 character followed by the original character XORed with @code{0x20}.
37223 For example, the byte @code{0x7d} would be transmitted as the two
37224 bytes @code{0x7d 0x5d}. The bytes @code{0x23} (@sc{ascii} @samp{#}),
37225 @code{0x24} (@sc{ascii} @samp{$}), and @code{0x7d} (@sc{ascii}
37226 @samp{@}}) must always be escaped. Responses sent by the stub
37227 must also escape @code{0x2a} (@sc{ascii} @samp{*}), so that it
37228 is not interpreted as the start of a run-length encoded sequence
37231 Response @var{data} can be run-length encoded to save space.
37232 Run-length encoding replaces runs of identical characters with one
37233 instance of the repeated character, followed by a @samp{*} and a
37234 repeat count. The repeat count is itself sent encoded, to avoid
37235 binary characters in @var{data}: a value of @var{n} is sent as
37236 @code{@var{n}+29}. For a repeat count greater or equal to 3, this
37237 produces a printable @sc{ascii} character, e.g.@: a space (@sc{ascii}
37238 code 32) for a repeat count of 3. (This is because run-length
37239 encoding starts to win for counts 3 or more.) Thus, for example,
37240 @samp{0* } is a run-length encoding of ``0000'': the space character
37241 after @samp{*} means repeat the leading @code{0} @w{@code{32 - 29 =
37244 The printable characters @samp{#} and @samp{$} or with a numeric value
37245 greater than 126 must not be used. Runs of six repeats (@samp{#}) or
37246 seven repeats (@samp{$}) can be expanded using a repeat count of only
37247 five (@samp{"}). For example, @samp{00000000} can be encoded as
37250 The error response returned for some packets includes a two character
37251 error number. That number is not well defined.
37253 @cindex empty response, for unsupported packets
37254 For any @var{command} not supported by the stub, an empty response
37255 (@samp{$#00}) should be returned. That way it is possible to extend the
37256 protocol. A newer @value{GDBN} can tell if a packet is supported based
37259 At a minimum, a stub is required to support the @samp{g} and @samp{G}
37260 commands for register access, and the @samp{m} and @samp{M} commands
37261 for memory access. Stubs that only control single-threaded targets
37262 can implement run control with the @samp{c} (continue), and @samp{s}
37263 (step) commands. Stubs that support multi-threading targets should
37264 support the @samp{vCont} command. All other commands are optional.
37269 The following table provides a complete list of all currently defined
37270 @var{command}s and their corresponding response @var{data}.
37271 @xref{File-I/O Remote Protocol Extension}, for details about the File
37272 I/O extension of the remote protocol.
37274 Each packet's description has a template showing the packet's overall
37275 syntax, followed by an explanation of the packet's meaning. We
37276 include spaces in some of the templates for clarity; these are not
37277 part of the packet's syntax. No @value{GDBN} packet uses spaces to
37278 separate its components. For example, a template like @samp{foo
37279 @var{bar} @var{baz}} describes a packet beginning with the three ASCII
37280 bytes @samp{foo}, followed by a @var{bar}, followed directly by a
37281 @var{baz}. @value{GDBN} does not transmit a space character between the
37282 @samp{foo} and the @var{bar}, or between the @var{bar} and the
37285 @cindex @var{thread-id}, in remote protocol
37286 @anchor{thread-id syntax}
37287 Several packets and replies include a @var{thread-id} field to identify
37288 a thread. Normally these are positive numbers with a target-specific
37289 interpretation, formatted as big-endian hex strings. A @var{thread-id}
37290 can also be a literal @samp{-1} to indicate all threads, or @samp{0} to
37293 In addition, the remote protocol supports a multiprocess feature in
37294 which the @var{thread-id} syntax is extended to optionally include both
37295 process and thread ID fields, as @samp{p@var{pid}.@var{tid}}.
37296 The @var{pid} (process) and @var{tid} (thread) components each have the
37297 format described above: a positive number with target-specific
37298 interpretation formatted as a big-endian hex string, literal @samp{-1}
37299 to indicate all processes or threads (respectively), or @samp{0} to
37300 indicate an arbitrary process or thread. Specifying just a process, as
37301 @samp{p@var{pid}}, is equivalent to @samp{p@var{pid}.-1}. It is an
37302 error to specify all processes but a specific thread, such as
37303 @samp{p-1.@var{tid}}. Note that the @samp{p} prefix is @emph{not} used
37304 for those packets and replies explicitly documented to include a process
37305 ID, rather than a @var{thread-id}.
37307 The multiprocess @var{thread-id} syntax extensions are only used if both
37308 @value{GDBN} and the stub report support for the @samp{multiprocess}
37309 feature using @samp{qSupported}. @xref{multiprocess extensions}, for
37312 Note that all packet forms beginning with an upper- or lower-case
37313 letter, other than those described here, are reserved for future use.
37315 Here are the packet descriptions.
37320 @cindex @samp{!} packet
37321 @anchor{extended mode}
37322 Enable extended mode. In extended mode, the remote server is made
37323 persistent. The @samp{R} packet is used to restart the program being
37329 The remote target both supports and has enabled extended mode.
37333 @cindex @samp{?} packet
37335 Indicate the reason the target halted. The reply is the same as for
37336 step and continue. This packet has a special interpretation when the
37337 target is in non-stop mode; see @ref{Remote Non-Stop}.
37340 @xref{Stop Reply Packets}, for the reply specifications.
37342 @item A @var{arglen},@var{argnum},@var{arg},@dots{}
37343 @cindex @samp{A} packet
37344 Initialized @code{argv[]} array passed into program. @var{arglen}
37345 specifies the number of bytes in the hex encoded byte stream
37346 @var{arg}. See @code{gdbserver} for more details.
37351 The arguments were set.
37357 @cindex @samp{b} packet
37358 (Don't use this packet; its behavior is not well-defined.)
37359 Change the serial line speed to @var{baud}.
37361 JTC: @emph{When does the transport layer state change? When it's
37362 received, or after the ACK is transmitted. In either case, there are
37363 problems if the command or the acknowledgment packet is dropped.}
37365 Stan: @emph{If people really wanted to add something like this, and get
37366 it working for the first time, they ought to modify ser-unix.c to send
37367 some kind of out-of-band message to a specially-setup stub and have the
37368 switch happen "in between" packets, so that from remote protocol's point
37369 of view, nothing actually happened.}
37371 @item B @var{addr},@var{mode}
37372 @cindex @samp{B} packet
37373 Set (@var{mode} is @samp{S}) or clear (@var{mode} is @samp{C}) a
37374 breakpoint at @var{addr}.
37376 Don't use this packet. Use the @samp{Z} and @samp{z} packets instead
37377 (@pxref{insert breakpoint or watchpoint packet}).
37379 @cindex @samp{bc} packet
37382 Backward continue. Execute the target system in reverse. No parameter.
37383 @xref{Reverse Execution}, for more information.
37386 @xref{Stop Reply Packets}, for the reply specifications.
37388 @cindex @samp{bs} packet
37391 Backward single step. Execute one instruction in reverse. No parameter.
37392 @xref{Reverse Execution}, for more information.
37395 @xref{Stop Reply Packets}, for the reply specifications.
37397 @item c @r{[}@var{addr}@r{]}
37398 @cindex @samp{c} packet
37399 Continue at @var{addr}, which is the address to resume. If @var{addr}
37400 is omitted, resume at current address.
37402 This packet is deprecated for multi-threading support. @xref{vCont
37406 @xref{Stop Reply Packets}, for the reply specifications.
37408 @item C @var{sig}@r{[};@var{addr}@r{]}
37409 @cindex @samp{C} packet
37410 Continue with signal @var{sig} (hex signal number). If
37411 @samp{;@var{addr}} is omitted, resume at same address.
37413 This packet is deprecated for multi-threading support. @xref{vCont
37417 @xref{Stop Reply Packets}, for the reply specifications.
37420 @cindex @samp{d} packet
37423 Don't use this packet; instead, define a general set packet
37424 (@pxref{General Query Packets}).
37428 @cindex @samp{D} packet
37429 The first form of the packet is used to detach @value{GDBN} from the
37430 remote system. It is sent to the remote target
37431 before @value{GDBN} disconnects via the @code{detach} command.
37433 The second form, including a process ID, is used when multiprocess
37434 protocol extensions are enabled (@pxref{multiprocess extensions}), to
37435 detach only a specific process. The @var{pid} is specified as a
37436 big-endian hex string.
37446 @item F @var{RC},@var{EE},@var{CF};@var{XX}
37447 @cindex @samp{F} packet
37448 A reply from @value{GDBN} to an @samp{F} packet sent by the target.
37449 This is part of the File-I/O protocol extension. @xref{File-I/O
37450 Remote Protocol Extension}, for the specification.
37453 @anchor{read registers packet}
37454 @cindex @samp{g} packet
37455 Read general registers.
37459 @item @var{XX@dots{}}
37460 Each byte of register data is described by two hex digits. The bytes
37461 with the register are transmitted in target byte order. The size of
37462 each register and their position within the @samp{g} packet are
37463 determined by the @value{GDBN} internal gdbarch functions
37464 @code{DEPRECATED_REGISTER_RAW_SIZE} and @code{gdbarch_register_name}.
37466 When reading registers from a trace frame (@pxref{Analyze Collected
37467 Data,,Using the Collected Data}), the stub may also return a string of
37468 literal @samp{x}'s in place of the register data digits, to indicate
37469 that the corresponding register has not been collected, thus its value
37470 is unavailable. For example, for an architecture with 4 registers of
37471 4 bytes each, the following reply indicates to @value{GDBN} that
37472 registers 0 and 2 have not been collected, while registers 1 and 3
37473 have been collected, and both have zero value:
37477 <- @code{xxxxxxxx00000000xxxxxxxx00000000}
37484 @item G @var{XX@dots{}}
37485 @cindex @samp{G} packet
37486 Write general registers. @xref{read registers packet}, for a
37487 description of the @var{XX@dots{}} data.
37497 @item H @var{op} @var{thread-id}
37498 @cindex @samp{H} packet
37499 Set thread for subsequent operations (@samp{m}, @samp{M}, @samp{g},
37500 @samp{G}, et.al.). Depending on the operation to be performed, @var{op}
37501 should be @samp{c} for step and continue operations (note that this
37502 is deprecated, supporting the @samp{vCont} command is a better
37503 option), and @samp{g} for other operations. The thread designator
37504 @var{thread-id} has the format and interpretation described in
37505 @ref{thread-id syntax}.
37516 @c 'H': How restrictive (or permissive) is the thread model. If a
37517 @c thread is selected and stopped, are other threads allowed
37518 @c to continue to execute? As I mentioned above, I think the
37519 @c semantics of each command when a thread is selected must be
37520 @c described. For example:
37522 @c 'g': If the stub supports threads and a specific thread is
37523 @c selected, returns the register block from that thread;
37524 @c otherwise returns current registers.
37526 @c 'G' If the stub supports threads and a specific thread is
37527 @c selected, sets the registers of the register block of
37528 @c that thread; otherwise sets current registers.
37530 @item i @r{[}@var{addr}@r{[},@var{nnn}@r{]]}
37531 @anchor{cycle step packet}
37532 @cindex @samp{i} packet
37533 Step the remote target by a single clock cycle. If @samp{,@var{nnn}} is
37534 present, cycle step @var{nnn} cycles. If @var{addr} is present, cycle
37535 step starting at that address.
37538 @cindex @samp{I} packet
37539 Signal, then cycle step. @xref{step with signal packet}. @xref{cycle
37543 @cindex @samp{k} packet
37546 The exact effect of this packet is not specified.
37548 For a bare-metal target, it may power cycle or reset the target
37549 system. For that reason, the @samp{k} packet has no reply.
37551 For a single-process target, it may kill that process if possible.
37553 A multiple-process target may choose to kill just one process, or all
37554 that are under @value{GDBN}'s control. For more precise control, use
37555 the vKill packet (@pxref{vKill packet}).
37557 If the target system immediately closes the connection in response to
37558 @samp{k}, @value{GDBN} does not consider the lack of packet
37559 acknowledgment to be an error, and assumes the kill was successful.
37561 If connected using @kbd{target extended-remote}, and the target does
37562 not close the connection in response to a kill request, @value{GDBN}
37563 probes the target state as if a new connection was opened
37564 (@pxref{? packet}).
37566 @item m @var{addr},@var{length}
37567 @cindex @samp{m} packet
37568 Read @var{length} addressable memory units starting at address @var{addr}
37569 (@pxref{addressable memory unit}). Note that @var{addr} may not be aligned to
37570 any particular boundary.
37572 The stub need not use any particular size or alignment when gathering
37573 data from memory for the response; even if @var{addr} is word-aligned
37574 and @var{length} is a multiple of the word size, the stub is free to
37575 use byte accesses, or not. For this reason, this packet may not be
37576 suitable for accessing memory-mapped I/O devices.
37577 @cindex alignment of remote memory accesses
37578 @cindex size of remote memory accesses
37579 @cindex memory, alignment and size of remote accesses
37583 @item @var{XX@dots{}}
37584 Memory contents; each byte is transmitted as a two-digit hexadecimal number.
37585 The reply may contain fewer addressable memory units than requested if the
37586 server was able to read only part of the region of memory.
37591 @item M @var{addr},@var{length}:@var{XX@dots{}}
37592 @cindex @samp{M} packet
37593 Write @var{length} addressable memory units starting at address @var{addr}
37594 (@pxref{addressable memory unit}). The data is given by @var{XX@dots{}}; each
37595 byte is transmitted as a two-digit hexadecimal number.
37602 for an error (this includes the case where only part of the data was
37607 @cindex @samp{p} packet
37608 Read the value of register @var{n}; @var{n} is in hex.
37609 @xref{read registers packet}, for a description of how the returned
37610 register value is encoded.
37614 @item @var{XX@dots{}}
37615 the register's value
37619 Indicating an unrecognized @var{query}.
37622 @item P @var{n@dots{}}=@var{r@dots{}}
37623 @anchor{write register packet}
37624 @cindex @samp{P} packet
37625 Write register @var{n@dots{}} with value @var{r@dots{}}. The register
37626 number @var{n} is in hexadecimal, and @var{r@dots{}} contains two hex
37627 digits for each byte in the register (target byte order).
37637 @item q @var{name} @var{params}@dots{}
37638 @itemx Q @var{name} @var{params}@dots{}
37639 @cindex @samp{q} packet
37640 @cindex @samp{Q} packet
37641 General query (@samp{q}) and set (@samp{Q}). These packets are
37642 described fully in @ref{General Query Packets}.
37645 @cindex @samp{r} packet
37646 Reset the entire system.
37648 Don't use this packet; use the @samp{R} packet instead.
37651 @cindex @samp{R} packet
37652 Restart the program being debugged. The @var{XX}, while needed, is ignored.
37653 This packet is only available in extended mode (@pxref{extended mode}).
37655 The @samp{R} packet has no reply.
37657 @item s @r{[}@var{addr}@r{]}
37658 @cindex @samp{s} packet
37659 Single step, resuming at @var{addr}. If
37660 @var{addr} is omitted, resume at same address.
37662 This packet is deprecated for multi-threading support. @xref{vCont
37666 @xref{Stop Reply Packets}, for the reply specifications.
37668 @item S @var{sig}@r{[};@var{addr}@r{]}
37669 @anchor{step with signal packet}
37670 @cindex @samp{S} packet
37671 Step with signal. This is analogous to the @samp{C} packet, but
37672 requests a single-step, rather than a normal resumption of execution.
37674 This packet is deprecated for multi-threading support. @xref{vCont
37678 @xref{Stop Reply Packets}, for the reply specifications.
37680 @item t @var{addr}:@var{PP},@var{MM}
37681 @cindex @samp{t} packet
37682 Search backwards starting at address @var{addr} for a match with pattern
37683 @var{PP} and mask @var{MM}, both of which are are 4 byte long.
37684 There must be at least 3 digits in @var{addr}.
37686 @item T @var{thread-id}
37687 @cindex @samp{T} packet
37688 Find out if the thread @var{thread-id} is alive. @xref{thread-id syntax}.
37693 thread is still alive
37699 Packets starting with @samp{v} are identified by a multi-letter name,
37700 up to the first @samp{;} or @samp{?} (or the end of the packet).
37702 @item vAttach;@var{pid}
37703 @cindex @samp{vAttach} packet
37704 Attach to a new process with the specified process ID @var{pid}.
37705 The process ID is a
37706 hexadecimal integer identifying the process. In all-stop mode, all
37707 threads in the attached process are stopped; in non-stop mode, it may be
37708 attached without being stopped if that is supported by the target.
37710 @c In non-stop mode, on a successful vAttach, the stub should set the
37711 @c current thread to a thread of the newly-attached process. After
37712 @c attaching, GDB queries for the attached process's thread ID with qC.
37713 @c Also note that, from a user perspective, whether or not the
37714 @c target is stopped on attach in non-stop mode depends on whether you
37715 @c use the foreground or background version of the attach command, not
37716 @c on what vAttach does; GDB does the right thing with respect to either
37717 @c stopping or restarting threads.
37719 This packet is only available in extended mode (@pxref{extended mode}).
37725 @item @r{Any stop packet}
37726 for success in all-stop mode (@pxref{Stop Reply Packets})
37728 for success in non-stop mode (@pxref{Remote Non-Stop})
37731 @item vCont@r{[};@var{action}@r{[}:@var{thread-id}@r{]]}@dots{}
37732 @cindex @samp{vCont} packet
37733 @anchor{vCont packet}
37734 Resume the inferior, specifying different actions for each thread.
37736 For each inferior thread, the leftmost action with a matching
37737 @var{thread-id} is applied. Threads that don't match any action
37738 remain in their current state. Thread IDs are specified using the
37739 syntax described in @ref{thread-id syntax}. If multiprocess
37740 extensions (@pxref{multiprocess extensions}) are supported, actions
37741 can be specified to match all threads in a process by using the
37742 @samp{p@var{pid}.-1} form of the @var{thread-id}. An action with no
37743 @var{thread-id} matches all threads. Specifying no actions is an
37746 Currently supported actions are:
37752 Continue with signal @var{sig}. The signal @var{sig} should be two hex digits.
37756 Step with signal @var{sig}. The signal @var{sig} should be two hex digits.
37759 @item r @var{start},@var{end}
37760 Step once, and then keep stepping as long as the thread stops at
37761 addresses between @var{start} (inclusive) and @var{end} (exclusive).
37762 The remote stub reports a stop reply when either the thread goes out
37763 of the range or is stopped due to an unrelated reason, such as hitting
37764 a breakpoint. @xref{range stepping}.
37766 If the range is empty (@var{start} == @var{end}), then the action
37767 becomes equivalent to the @samp{s} action. In other words,
37768 single-step once, and report the stop (even if the stepped instruction
37769 jumps to @var{start}).
37771 (A stop reply may be sent at any point even if the PC is still within
37772 the stepping range; for example, it is valid to implement this packet
37773 in a degenerate way as a single instruction step operation.)
37777 The optional argument @var{addr} normally associated with the
37778 @samp{c}, @samp{C}, @samp{s}, and @samp{S} packets is
37779 not supported in @samp{vCont}.
37781 The @samp{t} action is only relevant in non-stop mode
37782 (@pxref{Remote Non-Stop}) and may be ignored by the stub otherwise.
37783 A stop reply should be generated for any affected thread not already stopped.
37784 When a thread is stopped by means of a @samp{t} action,
37785 the corresponding stop reply should indicate that the thread has stopped with
37786 signal @samp{0}, regardless of whether the target uses some other signal
37787 as an implementation detail.
37789 The server must ignore @samp{c}, @samp{C}, @samp{s}, @samp{S}, and
37790 @samp{r} actions for threads that are already running. Conversely,
37791 the server must ignore @samp{t} actions for threads that are already
37794 @emph{Note:} In non-stop mode, a thread is considered running until
37795 @value{GDBN} acknowleges an asynchronous stop notification for it with
37796 the @samp{vStopped} packet (@pxref{Remote Non-Stop}).
37798 The stub must support @samp{vCont} if it reports support for
37799 multiprocess extensions (@pxref{multiprocess extensions}).
37802 @xref{Stop Reply Packets}, for the reply specifications.
37805 @cindex @samp{vCont?} packet
37806 Request a list of actions supported by the @samp{vCont} packet.
37810 @item vCont@r{[};@var{action}@dots{}@r{]}
37811 The @samp{vCont} packet is supported. Each @var{action} is a supported
37812 command in the @samp{vCont} packet.
37814 The @samp{vCont} packet is not supported.
37817 @anchor{vCtrlC packet}
37819 @cindex @samp{vCtrlC} packet
37820 Interrupt remote target as if a control-C was pressed on the remote
37821 terminal. This is the equivalent to reacting to the @code{^C}
37822 (@samp{\003}, the control-C character) character in all-stop mode
37823 while the target is running, except this works in non-stop mode.
37824 @xref{interrupting remote targets}, for more info on the all-stop
37835 @item vFile:@var{operation}:@var{parameter}@dots{}
37836 @cindex @samp{vFile} packet
37837 Perform a file operation on the target system. For details,
37838 see @ref{Host I/O Packets}.
37840 @item vFlashErase:@var{addr},@var{length}
37841 @cindex @samp{vFlashErase} packet
37842 Direct the stub to erase @var{length} bytes of flash starting at
37843 @var{addr}. The region may enclose any number of flash blocks, but
37844 its start and end must fall on block boundaries, as indicated by the
37845 flash block size appearing in the memory map (@pxref{Memory Map
37846 Format}). @value{GDBN} groups flash memory programming operations
37847 together, and sends a @samp{vFlashDone} request after each group; the
37848 stub is allowed to delay erase operation until the @samp{vFlashDone}
37849 packet is received.
37859 @item vFlashWrite:@var{addr}:@var{XX@dots{}}
37860 @cindex @samp{vFlashWrite} packet
37861 Direct the stub to write data to flash address @var{addr}. The data
37862 is passed in binary form using the same encoding as for the @samp{X}
37863 packet (@pxref{Binary Data}). The memory ranges specified by
37864 @samp{vFlashWrite} packets preceding a @samp{vFlashDone} packet must
37865 not overlap, and must appear in order of increasing addresses
37866 (although @samp{vFlashErase} packets for higher addresses may already
37867 have been received; the ordering is guaranteed only between
37868 @samp{vFlashWrite} packets). If a packet writes to an address that was
37869 neither erased by a preceding @samp{vFlashErase} packet nor by some other
37870 target-specific method, the results are unpredictable.
37878 for vFlashWrite addressing non-flash memory
37884 @cindex @samp{vFlashDone} packet
37885 Indicate to the stub that flash programming operation is finished.
37886 The stub is permitted to delay or batch the effects of a group of
37887 @samp{vFlashErase} and @samp{vFlashWrite} packets until a
37888 @samp{vFlashDone} packet is received. The contents of the affected
37889 regions of flash memory are unpredictable until the @samp{vFlashDone}
37890 request is completed.
37892 @item vKill;@var{pid}
37893 @cindex @samp{vKill} packet
37894 @anchor{vKill packet}
37895 Kill the process with the specified process ID @var{pid}, which is a
37896 hexadecimal integer identifying the process. This packet is used in
37897 preference to @samp{k} when multiprocess protocol extensions are
37898 supported; see @ref{multiprocess extensions}.
37908 @item vMustReplyEmpty
37909 @cindex @samp{vMustReplyEmpty} packet
37910 The correct reply to an unknown @samp{v} packet is to return the empty
37911 string, however, some older versions of @command{gdbserver} would
37912 incorrectly return @samp{OK} for unknown @samp{v} packets.
37914 The @samp{vMustReplyEmpty} is used as a feature test to check how
37915 @command{gdbserver} handles unknown packets, it is important that this
37916 packet be handled in the same way as other unknown @samp{v} packets.
37917 If this packet is handled differently to other unknown @samp{v}
37918 packets then it is possile that @value{GDBN} may run into problems in
37919 other areas, specifically around use of @samp{vFile:setfs:}.
37921 @item vRun;@var{filename}@r{[};@var{argument}@r{]}@dots{}
37922 @cindex @samp{vRun} packet
37923 Run the program @var{filename}, passing it each @var{argument} on its
37924 command line. The file and arguments are hex-encoded strings. If
37925 @var{filename} is an empty string, the stub may use a default program
37926 (e.g.@: the last program run). The program is created in the stopped
37929 @c FIXME: What about non-stop mode?
37931 This packet is only available in extended mode (@pxref{extended mode}).
37937 @item @r{Any stop packet}
37938 for success (@pxref{Stop Reply Packets})
37942 @cindex @samp{vStopped} packet
37943 @xref{Notification Packets}.
37945 @item X @var{addr},@var{length}:@var{XX@dots{}}
37947 @cindex @samp{X} packet
37948 Write data to memory, where the data is transmitted in binary.
37949 Memory is specified by its address @var{addr} and number of addressable memory
37950 units @var{length} (@pxref{addressable memory unit});
37951 @samp{@var{XX}@dots{}} is binary data (@pxref{Binary Data}).
37961 @item z @var{type},@var{addr},@var{kind}
37962 @itemx Z @var{type},@var{addr},@var{kind}
37963 @anchor{insert breakpoint or watchpoint packet}
37964 @cindex @samp{z} packet
37965 @cindex @samp{Z} packets
37966 Insert (@samp{Z}) or remove (@samp{z}) a @var{type} breakpoint or
37967 watchpoint starting at address @var{address} of kind @var{kind}.
37969 Each breakpoint and watchpoint packet @var{type} is documented
37972 @emph{Implementation notes: A remote target shall return an empty string
37973 for an unrecognized breakpoint or watchpoint packet @var{type}. A
37974 remote target shall support either both or neither of a given
37975 @samp{Z@var{type}@dots{}} and @samp{z@var{type}@dots{}} packet pair. To
37976 avoid potential problems with duplicate packets, the operations should
37977 be implemented in an idempotent way.}
37979 @item z0,@var{addr},@var{kind}
37980 @itemx Z0,@var{addr},@var{kind}@r{[};@var{cond_list}@dots{}@r{]}@r{[};cmds:@var{persist},@var{cmd_list}@dots{}@r{]}
37981 @cindex @samp{z0} packet
37982 @cindex @samp{Z0} packet
37983 Insert (@samp{Z0}) or remove (@samp{z0}) a software breakpoint at address
37984 @var{addr} of type @var{kind}.
37986 A software breakpoint is implemented by replacing the instruction at
37987 @var{addr} with a software breakpoint or trap instruction. The
37988 @var{kind} is target-specific and typically indicates the size of the
37989 breakpoint in bytes that should be inserted. E.g., the @sc{arm} and
37990 @sc{mips} can insert either a 2 or 4 byte breakpoint. Some
37991 architectures have additional meanings for @var{kind}
37992 (@pxref{Architecture-Specific Protocol Details}); if no
37993 architecture-specific value is being used, it should be @samp{0}.
37994 @var{kind} is hex-encoded. @var{cond_list} is an optional list of
37995 conditional expressions in bytecode form that should be evaluated on
37996 the target's side. These are the conditions that should be taken into
37997 consideration when deciding if the breakpoint trigger should be
37998 reported back to @value{GDBN}.
38000 See also the @samp{swbreak} stop reason (@pxref{swbreak stop reason})
38001 for how to best report a software breakpoint event to @value{GDBN}.
38003 The @var{cond_list} parameter is comprised of a series of expressions,
38004 concatenated without separators. Each expression has the following form:
38008 @item X @var{len},@var{expr}
38009 @var{len} is the length of the bytecode expression and @var{expr} is the
38010 actual conditional expression in bytecode form.
38014 The optional @var{cmd_list} parameter introduces commands that may be
38015 run on the target, rather than being reported back to @value{GDBN}.
38016 The parameter starts with a numeric flag @var{persist}; if the flag is
38017 nonzero, then the breakpoint may remain active and the commands
38018 continue to be run even when @value{GDBN} disconnects from the target.
38019 Following this flag is a series of expressions concatenated with no
38020 separators. Each expression has the following form:
38024 @item X @var{len},@var{expr}
38025 @var{len} is the length of the bytecode expression and @var{expr} is the
38026 actual commands expression in bytecode form.
38030 @emph{Implementation note: It is possible for a target to copy or move
38031 code that contains software breakpoints (e.g., when implementing
38032 overlays). The behavior of this packet, in the presence of such a
38033 target, is not defined.}
38045 @item z1,@var{addr},@var{kind}
38046 @itemx Z1,@var{addr},@var{kind}@r{[};@var{cond_list}@dots{}@r{]}@r{[};cmds:@var{persist},@var{cmd_list}@dots{}@r{]}
38047 @cindex @samp{z1} packet
38048 @cindex @samp{Z1} packet
38049 Insert (@samp{Z1}) or remove (@samp{z1}) a hardware breakpoint at
38050 address @var{addr}.
38052 A hardware breakpoint is implemented using a mechanism that is not
38053 dependent on being able to modify the target's memory. The
38054 @var{kind}, @var{cond_list}, and @var{cmd_list} arguments have the
38055 same meaning as in @samp{Z0} packets.
38057 @emph{Implementation note: A hardware breakpoint is not affected by code
38070 @item z2,@var{addr},@var{kind}
38071 @itemx Z2,@var{addr},@var{kind}
38072 @cindex @samp{z2} packet
38073 @cindex @samp{Z2} packet
38074 Insert (@samp{Z2}) or remove (@samp{z2}) a write watchpoint at @var{addr}.
38075 The number of bytes to watch is specified by @var{kind}.
38087 @item z3,@var{addr},@var{kind}
38088 @itemx Z3,@var{addr},@var{kind}
38089 @cindex @samp{z3} packet
38090 @cindex @samp{Z3} packet
38091 Insert (@samp{Z3}) or remove (@samp{z3}) a read watchpoint at @var{addr}.
38092 The number of bytes to watch is specified by @var{kind}.
38104 @item z4,@var{addr},@var{kind}
38105 @itemx Z4,@var{addr},@var{kind}
38106 @cindex @samp{z4} packet
38107 @cindex @samp{Z4} packet
38108 Insert (@samp{Z4}) or remove (@samp{z4}) an access watchpoint at @var{addr}.
38109 The number of bytes to watch is specified by @var{kind}.
38123 @node Stop Reply Packets
38124 @section Stop Reply Packets
38125 @cindex stop reply packets
38127 The @samp{C}, @samp{c}, @samp{S}, @samp{s}, @samp{vCont},
38128 @samp{vAttach}, @samp{vRun}, @samp{vStopped}, and @samp{?} packets can
38129 receive any of the below as a reply. Except for @samp{?}
38130 and @samp{vStopped}, that reply is only returned
38131 when the target halts. In the below the exact meaning of @dfn{signal
38132 number} is defined by the header @file{include/gdb/signals.h} in the
38133 @value{GDBN} source code.
38135 In non-stop mode, the server will simply reply @samp{OK} to commands
38136 such as @samp{vCont}; any stop will be the subject of a future
38137 notification. @xref{Remote Non-Stop}.
38139 As in the description of request packets, we include spaces in the
38140 reply templates for clarity; these are not part of the reply packet's
38141 syntax. No @value{GDBN} stop reply packet uses spaces to separate its
38147 The program received signal number @var{AA} (a two-digit hexadecimal
38148 number). This is equivalent to a @samp{T} response with no
38149 @var{n}:@var{r} pairs.
38151 @item T @var{AA} @var{n1}:@var{r1};@var{n2}:@var{r2};@dots{}
38152 @cindex @samp{T} packet reply
38153 The program received signal number @var{AA} (a two-digit hexadecimal
38154 number). This is equivalent to an @samp{S} response, except that the
38155 @samp{@var{n}:@var{r}} pairs can carry values of important registers
38156 and other information directly in the stop reply packet, reducing
38157 round-trip latency. Single-step and breakpoint traps are reported
38158 this way. Each @samp{@var{n}:@var{r}} pair is interpreted as follows:
38162 If @var{n} is a hexadecimal number, it is a register number, and the
38163 corresponding @var{r} gives that register's value. The data @var{r} is a
38164 series of bytes in target byte order, with each byte given by a
38165 two-digit hex number.
38168 If @var{n} is @samp{thread}, then @var{r} is the @var{thread-id} of
38169 the stopped thread, as specified in @ref{thread-id syntax}.
38172 If @var{n} is @samp{core}, then @var{r} is the hexadecimal number of
38173 the core on which the stop event was detected.
38176 If @var{n} is a recognized @dfn{stop reason}, it describes a more
38177 specific event that stopped the target. The currently defined stop
38178 reasons are listed below. The @var{aa} should be @samp{05}, the trap
38179 signal. At most one stop reason should be present.
38182 Otherwise, @value{GDBN} should ignore this @samp{@var{n}:@var{r}} pair
38183 and go on to the next; this allows us to extend the protocol in the
38187 The currently defined stop reasons are:
38193 The packet indicates a watchpoint hit, and @var{r} is the data address, in
38196 @item syscall_entry
38197 @itemx syscall_return
38198 The packet indicates a syscall entry or return, and @var{r} is the
38199 syscall number, in hex.
38201 @cindex shared library events, remote reply
38203 The packet indicates that the loaded libraries have changed.
38204 @value{GDBN} should use @samp{qXfer:libraries:read} to fetch a new
38205 list of loaded libraries. The @var{r} part is ignored.
38207 @cindex replay log events, remote reply
38209 The packet indicates that the target cannot continue replaying
38210 logged execution events, because it has reached the end (or the
38211 beginning when executing backward) of the log. The value of @var{r}
38212 will be either @samp{begin} or @samp{end}. @xref{Reverse Execution},
38213 for more information.
38216 @anchor{swbreak stop reason}
38217 The packet indicates a software breakpoint instruction was executed,
38218 irrespective of whether it was @value{GDBN} that planted the
38219 breakpoint or the breakpoint is hardcoded in the program. The @var{r}
38220 part must be left empty.
38222 On some architectures, such as x86, at the architecture level, when a
38223 breakpoint instruction executes the program counter points at the
38224 breakpoint address plus an offset. On such targets, the stub is
38225 responsible for adjusting the PC to point back at the breakpoint
38228 This packet should not be sent by default; older @value{GDBN} versions
38229 did not support it. @value{GDBN} requests it, by supplying an
38230 appropriate @samp{qSupported} feature (@pxref{qSupported}). The
38231 remote stub must also supply the appropriate @samp{qSupported} feature
38232 indicating support.
38234 This packet is required for correct non-stop mode operation.
38237 The packet indicates the target stopped for a hardware breakpoint.
38238 The @var{r} part must be left empty.
38240 The same remarks about @samp{qSupported} and non-stop mode above
38243 @cindex fork events, remote reply
38245 The packet indicates that @code{fork} was called, and @var{r}
38246 is the thread ID of the new child process. Refer to
38247 @ref{thread-id syntax} for the format of the @var{thread-id}
38248 field. This packet is only applicable to targets that support
38251 This packet should not be sent by default; older @value{GDBN} versions
38252 did not support it. @value{GDBN} requests it, by supplying an
38253 appropriate @samp{qSupported} feature (@pxref{qSupported}). The
38254 remote stub must also supply the appropriate @samp{qSupported} feature
38255 indicating support.
38257 @cindex vfork events, remote reply
38259 The packet indicates that @code{vfork} was called, and @var{r}
38260 is the thread ID of the new child process. Refer to
38261 @ref{thread-id syntax} for the format of the @var{thread-id}
38262 field. This packet is only applicable to targets that support
38265 This packet should not be sent by default; older @value{GDBN} versions
38266 did not support it. @value{GDBN} requests it, by supplying an
38267 appropriate @samp{qSupported} feature (@pxref{qSupported}). The
38268 remote stub must also supply the appropriate @samp{qSupported} feature
38269 indicating support.
38271 @cindex vforkdone events, remote reply
38273 The packet indicates that a child process created by a vfork
38274 has either called @code{exec} or terminated, so that the
38275 address spaces of the parent and child process are no longer
38276 shared. The @var{r} part is ignored. This packet is only
38277 applicable to targets that support vforkdone events.
38279 This packet should not be sent by default; older @value{GDBN} versions
38280 did not support it. @value{GDBN} requests it, by supplying an
38281 appropriate @samp{qSupported} feature (@pxref{qSupported}). The
38282 remote stub must also supply the appropriate @samp{qSupported} feature
38283 indicating support.
38285 @cindex exec events, remote reply
38287 The packet indicates that @code{execve} was called, and @var{r}
38288 is the absolute pathname of the file that was executed, in hex.
38289 This packet is only applicable to targets that support exec events.
38291 This packet should not be sent by default; older @value{GDBN} versions
38292 did not support it. @value{GDBN} requests it, by supplying an
38293 appropriate @samp{qSupported} feature (@pxref{qSupported}). The
38294 remote stub must also supply the appropriate @samp{qSupported} feature
38295 indicating support.
38297 @cindex thread create event, remote reply
38298 @anchor{thread create event}
38300 The packet indicates that the thread was just created. The new thread
38301 is stopped until @value{GDBN} sets it running with a resumption packet
38302 (@pxref{vCont packet}). This packet should not be sent by default;
38303 @value{GDBN} requests it with the @ref{QThreadEvents} packet. See
38304 also the @samp{w} (@pxref{thread exit event}) remote reply below. The
38305 @var{r} part is ignored.
38310 @itemx W @var{AA} ; process:@var{pid}
38311 The process exited, and @var{AA} is the exit status. This is only
38312 applicable to certain targets.
38314 The second form of the response, including the process ID of the
38315 exited process, can be used only when @value{GDBN} has reported
38316 support for multiprocess protocol extensions; see @ref{multiprocess
38317 extensions}. Both @var{AA} and @var{pid} are formatted as big-endian
38321 @itemx X @var{AA} ; process:@var{pid}
38322 The process terminated with signal @var{AA}.
38324 The second form of the response, including the process ID of the
38325 terminated process, can be used only when @value{GDBN} has reported
38326 support for multiprocess protocol extensions; see @ref{multiprocess
38327 extensions}. Both @var{AA} and @var{pid} are formatted as big-endian
38330 @anchor{thread exit event}
38331 @cindex thread exit event, remote reply
38332 @item w @var{AA} ; @var{tid}
38334 The thread exited, and @var{AA} is the exit status. This response
38335 should not be sent by default; @value{GDBN} requests it with the
38336 @ref{QThreadEvents} packet. See also @ref{thread create event} above.
38337 @var{AA} is formatted as a big-endian hex string.
38340 There are no resumed threads left in the target. In other words, even
38341 though the process is alive, the last resumed thread has exited. For
38342 example, say the target process has two threads: thread 1 and thread
38343 2. The client leaves thread 1 stopped, and resumes thread 2, which
38344 subsequently exits. At this point, even though the process is still
38345 alive, and thus no @samp{W} stop reply is sent, no thread is actually
38346 executing either. The @samp{N} stop reply thus informs the client
38347 that it can stop waiting for stop replies. This packet should not be
38348 sent by default; older @value{GDBN} versions did not support it.
38349 @value{GDBN} requests it, by supplying an appropriate
38350 @samp{qSupported} feature (@pxref{qSupported}). The remote stub must
38351 also supply the appropriate @samp{qSupported} feature indicating
38354 @item O @var{XX}@dots{}
38355 @samp{@var{XX}@dots{}} is hex encoding of @sc{ascii} data, to be
38356 written as the program's console output. This can happen at any time
38357 while the program is running and the debugger should continue to wait
38358 for @samp{W}, @samp{T}, etc. This reply is not permitted in non-stop mode.
38360 @item F @var{call-id},@var{parameter}@dots{}
38361 @var{call-id} is the identifier which says which host system call should
38362 be called. This is just the name of the function. Translation into the
38363 correct system call is only applicable as it's defined in @value{GDBN}.
38364 @xref{File-I/O Remote Protocol Extension}, for a list of implemented
38367 @samp{@var{parameter}@dots{}} is a list of parameters as defined for
38368 this very system call.
38370 The target replies with this packet when it expects @value{GDBN} to
38371 call a host system call on behalf of the target. @value{GDBN} replies
38372 with an appropriate @samp{F} packet and keeps up waiting for the next
38373 reply packet from the target. The latest @samp{C}, @samp{c}, @samp{S}
38374 or @samp{s} action is expected to be continued. @xref{File-I/O Remote
38375 Protocol Extension}, for more details.
38379 @node General Query Packets
38380 @section General Query Packets
38381 @cindex remote query requests
38383 Packets starting with @samp{q} are @dfn{general query packets};
38384 packets starting with @samp{Q} are @dfn{general set packets}. General
38385 query and set packets are a semi-unified form for retrieving and
38386 sending information to and from the stub.
38388 The initial letter of a query or set packet is followed by a name
38389 indicating what sort of thing the packet applies to. For example,
38390 @value{GDBN} may use a @samp{qSymbol} packet to exchange symbol
38391 definitions with the stub. These packet names follow some
38396 The name must not contain commas, colons or semicolons.
38398 Most @value{GDBN} query and set packets have a leading upper case
38401 The names of custom vendor packets should use a company prefix, in
38402 lower case, followed by a period. For example, packets designed at
38403 the Acme Corporation might begin with @samp{qacme.foo} (for querying
38404 foos) or @samp{Qacme.bar} (for setting bars).
38407 The name of a query or set packet should be separated from any
38408 parameters by a @samp{:}; the parameters themselves should be
38409 separated by @samp{,} or @samp{;}. Stubs must be careful to match the
38410 full packet name, and check for a separator or the end of the packet,
38411 in case two packet names share a common prefix. New packets should not begin
38412 with @samp{qC}, @samp{qP}, or @samp{qL}@footnote{The @samp{qP} and @samp{qL}
38413 packets predate these conventions, and have arguments without any terminator
38414 for the packet name; we suspect they are in widespread use in places that
38415 are difficult to upgrade. The @samp{qC} packet has no arguments, but some
38416 existing stubs (e.g.@: RedBoot) are known to not check for the end of the
38419 Like the descriptions of the other packets, each description here
38420 has a template showing the packet's overall syntax, followed by an
38421 explanation of the packet's meaning. We include spaces in some of the
38422 templates for clarity; these are not part of the packet's syntax. No
38423 @value{GDBN} packet uses spaces to separate its components.
38425 Here are the currently defined query and set packets:
38431 Turn on or off the agent as a helper to perform some debugging operations
38432 delegated from @value{GDBN} (@pxref{Control Agent}).
38434 @item QAllow:@var{op}:@var{val}@dots{}
38435 @cindex @samp{QAllow} packet
38436 Specify which operations @value{GDBN} expects to request of the
38437 target, as a semicolon-separated list of operation name and value
38438 pairs. Possible values for @var{op} include @samp{WriteReg},
38439 @samp{WriteMem}, @samp{InsertBreak}, @samp{InsertTrace},
38440 @samp{InsertFastTrace}, and @samp{Stop}. @var{val} is either 0,
38441 indicating that @value{GDBN} will not request the operation, or 1,
38442 indicating that it may. (The target can then use this to set up its
38443 own internals optimally, for instance if the debugger never expects to
38444 insert breakpoints, it may not need to install its own trap handler.)
38447 @cindex current thread, remote request
38448 @cindex @samp{qC} packet
38449 Return the current thread ID.
38453 @item QC @var{thread-id}
38454 Where @var{thread-id} is a thread ID as documented in
38455 @ref{thread-id syntax}.
38456 @item @r{(anything else)}
38457 Any other reply implies the old thread ID.
38460 @item qCRC:@var{addr},@var{length}
38461 @cindex CRC of memory block, remote request
38462 @cindex @samp{qCRC} packet
38463 @anchor{qCRC packet}
38464 Compute the CRC checksum of a block of memory using CRC-32 defined in
38465 IEEE 802.3. The CRC is computed byte at a time, taking the most
38466 significant bit of each byte first. The initial pattern code
38467 @code{0xffffffff} is used to ensure leading zeros affect the CRC.
38469 @emph{Note:} This is the same CRC used in validating separate debug
38470 files (@pxref{Separate Debug Files, , Debugging Information in Separate
38471 Files}). However the algorithm is slightly different. When validating
38472 separate debug files, the CRC is computed taking the @emph{least}
38473 significant bit of each byte first, and the final result is inverted to
38474 detect trailing zeros.
38479 An error (such as memory fault)
38480 @item C @var{crc32}
38481 The specified memory region's checksum is @var{crc32}.
38484 @item QDisableRandomization:@var{value}
38485 @cindex disable address space randomization, remote request
38486 @cindex @samp{QDisableRandomization} packet
38487 Some target operating systems will randomize the virtual address space
38488 of the inferior process as a security feature, but provide a feature
38489 to disable such randomization, e.g.@: to allow for a more deterministic
38490 debugging experience. On such systems, this packet with a @var{value}
38491 of 1 directs the target to disable address space randomization for
38492 processes subsequently started via @samp{vRun} packets, while a packet
38493 with a @var{value} of 0 tells the target to enable address space
38496 This packet is only available in extended mode (@pxref{extended mode}).
38501 The request succeeded.
38504 An error occurred. The error number @var{nn} is given as hex digits.
38507 An empty reply indicates that @samp{QDisableRandomization} is not supported
38511 This packet is not probed by default; the remote stub must request it,
38512 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
38513 This should only be done on targets that actually support disabling
38514 address space randomization.
38516 @item QStartupWithShell:@var{value}
38517 @cindex startup with shell, remote request
38518 @cindex @samp{QStartupWithShell} packet
38519 On UNIX-like targets, it is possible to start the inferior using a
38520 shell program. This is the default behavior on both @value{GDBN} and
38521 @command{gdbserver} (@pxref{set startup-with-shell}). This packet is
38522 used to inform @command{gdbserver} whether it should start the
38523 inferior using a shell or not.
38525 If @var{value} is @samp{0}, @command{gdbserver} will not use a shell
38526 to start the inferior. If @var{value} is @samp{1},
38527 @command{gdbserver} will use a shell to start the inferior. All other
38528 values are considered an error.
38530 This packet is only available in extended mode (@pxref{extended
38536 The request succeeded.
38539 An error occurred. The error number @var{nn} is given as hex digits.
38542 This packet is not probed by default; the remote stub must request it,
38543 by supplying an appropriate @samp{qSupported} response
38544 (@pxref{qSupported}). This should only be done on targets that
38545 actually support starting the inferior using a shell.
38547 Use of this packet is controlled by the @code{set startup-with-shell}
38548 command; @pxref{set startup-with-shell}.
38550 @item QEnvironmentHexEncoded:@var{hex-value}
38551 @anchor{QEnvironmentHexEncoded}
38552 @cindex set environment variable, remote request
38553 @cindex @samp{QEnvironmentHexEncoded} packet
38554 On UNIX-like targets, it is possible to set environment variables that
38555 will be passed to the inferior during the startup process. This
38556 packet is used to inform @command{gdbserver} of an environment
38557 variable that has been defined by the user on @value{GDBN} (@pxref{set
38560 The packet is composed by @var{hex-value}, an hex encoded
38561 representation of the @var{name=value} format representing an
38562 environment variable. The name of the environment variable is
38563 represented by @var{name}, and the value to be assigned to the
38564 environment variable is represented by @var{value}. If the variable
38565 has no value (i.e., the value is @code{null}), then @var{value} will
38568 This packet is only available in extended mode (@pxref{extended
38574 The request succeeded.
38577 This packet is not probed by default; the remote stub must request it,
38578 by supplying an appropriate @samp{qSupported} response
38579 (@pxref{qSupported}). This should only be done on targets that
38580 actually support passing environment variables to the starting
38583 This packet is related to the @code{set environment} command;
38584 @pxref{set environment}.
38586 @item QEnvironmentUnset:@var{hex-value}
38587 @anchor{QEnvironmentUnset}
38588 @cindex unset environment variable, remote request
38589 @cindex @samp{QEnvironmentUnset} packet
38590 On UNIX-like targets, it is possible to unset environment variables
38591 before starting the inferior in the remote target. This packet is
38592 used to inform @command{gdbserver} of an environment variable that has
38593 been unset by the user on @value{GDBN} (@pxref{unset environment}).
38595 The packet is composed by @var{hex-value}, an hex encoded
38596 representation of the name of the environment variable to be unset.
38598 This packet is only available in extended mode (@pxref{extended
38604 The request succeeded.
38607 This packet is not probed by default; the remote stub must request it,
38608 by supplying an appropriate @samp{qSupported} response
38609 (@pxref{qSupported}). This should only be done on targets that
38610 actually support passing environment variables to the starting
38613 This packet is related to the @code{unset environment} command;
38614 @pxref{unset environment}.
38616 @item QEnvironmentReset
38617 @anchor{QEnvironmentReset}
38618 @cindex reset environment, remote request
38619 @cindex @samp{QEnvironmentReset} packet
38620 On UNIX-like targets, this packet is used to reset the state of
38621 environment variables in the remote target before starting the
38622 inferior. In this context, reset means unsetting all environment
38623 variables that were previously set by the user (i.e., were not
38624 initially present in the environment). It is sent to
38625 @command{gdbserver} before the @samp{QEnvironmentHexEncoded}
38626 (@pxref{QEnvironmentHexEncoded}) and the @samp{QEnvironmentUnset}
38627 (@pxref{QEnvironmentUnset}) packets.
38629 This packet is only available in extended mode (@pxref{extended
38635 The request succeeded.
38638 This packet is not probed by default; the remote stub must request it,
38639 by supplying an appropriate @samp{qSupported} response
38640 (@pxref{qSupported}). This should only be done on targets that
38641 actually support passing environment variables to the starting
38644 @item QSetWorkingDir:@r{[}@var{directory}@r{]}
38645 @anchor{QSetWorkingDir packet}
38646 @cindex set working directory, remote request
38647 @cindex @samp{QSetWorkingDir} packet
38648 This packet is used to inform the remote server of the intended
38649 current working directory for programs that are going to be executed.
38651 The packet is composed by @var{directory}, an hex encoded
38652 representation of the directory that the remote inferior will use as
38653 its current working directory. If @var{directory} is an empty string,
38654 the remote server should reset the inferior's current working
38655 directory to its original, empty value.
38657 This packet is only available in extended mode (@pxref{extended
38663 The request succeeded.
38667 @itemx qsThreadInfo
38668 @cindex list active threads, remote request
38669 @cindex @samp{qfThreadInfo} packet
38670 @cindex @samp{qsThreadInfo} packet
38671 Obtain a list of all active thread IDs from the target (OS). Since there
38672 may be too many active threads to fit into one reply packet, this query
38673 works iteratively: it may require more than one query/reply sequence to
38674 obtain the entire list of threads. The first query of the sequence will
38675 be the @samp{qfThreadInfo} query; subsequent queries in the
38676 sequence will be the @samp{qsThreadInfo} query.
38678 NOTE: This packet replaces the @samp{qL} query (see below).
38682 @item m @var{thread-id}
38684 @item m @var{thread-id},@var{thread-id}@dots{}
38685 a comma-separated list of thread IDs
38687 (lower case letter @samp{L}) denotes end of list.
38690 In response to each query, the target will reply with a list of one or
38691 more thread IDs, separated by commas.
38692 @value{GDBN} will respond to each reply with a request for more thread
38693 ids (using the @samp{qs} form of the query), until the target responds
38694 with @samp{l} (lower-case ell, for @dfn{last}).
38695 Refer to @ref{thread-id syntax}, for the format of the @var{thread-id}
38698 @emph{Note: @value{GDBN} will send the @code{qfThreadInfo} query during the
38699 initial connection with the remote target, and the very first thread ID
38700 mentioned in the reply will be stopped by @value{GDBN} in a subsequent
38701 message. Therefore, the stub should ensure that the first thread ID in
38702 the @code{qfThreadInfo} reply is suitable for being stopped by @value{GDBN}.}
38704 @item qGetTLSAddr:@var{thread-id},@var{offset},@var{lm}
38705 @cindex get thread-local storage address, remote request
38706 @cindex @samp{qGetTLSAddr} packet
38707 Fetch the address associated with thread local storage specified
38708 by @var{thread-id}, @var{offset}, and @var{lm}.
38710 @var{thread-id} is the thread ID associated with the
38711 thread for which to fetch the TLS address. @xref{thread-id syntax}.
38713 @var{offset} is the (big endian, hex encoded) offset associated with the
38714 thread local variable. (This offset is obtained from the debug
38715 information associated with the variable.)
38717 @var{lm} is the (big endian, hex encoded) OS/ABI-specific encoding of the
38718 load module associated with the thread local storage. For example,
38719 a @sc{gnu}/Linux system will pass the link map address of the shared
38720 object associated with the thread local storage under consideration.
38721 Other operating environments may choose to represent the load module
38722 differently, so the precise meaning of this parameter will vary.
38726 @item @var{XX}@dots{}
38727 Hex encoded (big endian) bytes representing the address of the thread
38728 local storage requested.
38731 An error occurred. The error number @var{nn} is given as hex digits.
38734 An empty reply indicates that @samp{qGetTLSAddr} is not supported by the stub.
38737 @item qGetTIBAddr:@var{thread-id}
38738 @cindex get thread information block address
38739 @cindex @samp{qGetTIBAddr} packet
38740 Fetch address of the Windows OS specific Thread Information Block.
38742 @var{thread-id} is the thread ID associated with the thread.
38746 @item @var{XX}@dots{}
38747 Hex encoded (big endian) bytes representing the linear address of the
38748 thread information block.
38751 An error occured. This means that either the thread was not found, or the
38752 address could not be retrieved.
38755 An empty reply indicates that @samp{qGetTIBAddr} is not supported by the stub.
38758 @item qL @var{startflag} @var{threadcount} @var{nextthread}
38759 Obtain thread information from RTOS. Where: @var{startflag} (one hex
38760 digit) is one to indicate the first query and zero to indicate a
38761 subsequent query; @var{threadcount} (two hex digits) is the maximum
38762 number of threads the response packet can contain; and @var{nextthread}
38763 (eight hex digits), for subsequent queries (@var{startflag} is zero), is
38764 returned in the response as @var{argthread}.
38766 Don't use this packet; use the @samp{qfThreadInfo} query instead (see above).
38770 @item qM @var{count} @var{done} @var{argthread} @var{thread}@dots{}
38771 Where: @var{count} (two hex digits) is the number of threads being
38772 returned; @var{done} (one hex digit) is zero to indicate more threads
38773 and one indicates no further threads; @var{argthreadid} (eight hex
38774 digits) is @var{nextthread} from the request packet; @var{thread}@dots{}
38775 is a sequence of thread IDs, @var{threadid} (eight hex
38776 digits), from the target. See @code{remote.c:parse_threadlist_response()}.
38780 @cindex section offsets, remote request
38781 @cindex @samp{qOffsets} packet
38782 Get section offsets that the target used when relocating the downloaded
38787 @item Text=@var{xxx};Data=@var{yyy}@r{[};Bss=@var{zzz}@r{]}
38788 Relocate the @code{Text} section by @var{xxx} from its original address.
38789 Relocate the @code{Data} section by @var{yyy} from its original address.
38790 If the object file format provides segment information (e.g.@: @sc{elf}
38791 @samp{PT_LOAD} program headers), @value{GDBN} will relocate entire
38792 segments by the supplied offsets.
38794 @emph{Note: while a @code{Bss} offset may be included in the response,
38795 @value{GDBN} ignores this and instead applies the @code{Data} offset
38796 to the @code{Bss} section.}
38798 @item TextSeg=@var{xxx}@r{[};DataSeg=@var{yyy}@r{]}
38799 Relocate the first segment of the object file, which conventionally
38800 contains program code, to a starting address of @var{xxx}. If
38801 @samp{DataSeg} is specified, relocate the second segment, which
38802 conventionally contains modifiable data, to a starting address of
38803 @var{yyy}. @value{GDBN} will report an error if the object file
38804 does not contain segment information, or does not contain at least
38805 as many segments as mentioned in the reply. Extra segments are
38806 kept at fixed offsets relative to the last relocated segment.
38809 @item qP @var{mode} @var{thread-id}
38810 @cindex thread information, remote request
38811 @cindex @samp{qP} packet
38812 Returns information on @var{thread-id}. Where: @var{mode} is a hex
38813 encoded 32 bit mode; @var{thread-id} is a thread ID
38814 (@pxref{thread-id syntax}).
38816 Don't use this packet; use the @samp{qThreadExtraInfo} query instead
38819 Reply: see @code{remote.c:remote_unpack_thread_info_response()}.
38823 @cindex non-stop mode, remote request
38824 @cindex @samp{QNonStop} packet
38826 Enter non-stop (@samp{QNonStop:1}) or all-stop (@samp{QNonStop:0}) mode.
38827 @xref{Remote Non-Stop}, for more information.
38832 The request succeeded.
38835 An error occurred. The error number @var{nn} is given as hex digits.
38838 An empty reply indicates that @samp{QNonStop} is not supported by
38842 This packet is not probed by default; the remote stub must request it,
38843 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
38844 Use of this packet is controlled by the @code{set non-stop} command;
38845 @pxref{Non-Stop Mode}.
38847 @item QCatchSyscalls:1 @r{[};@var{sysno}@r{]}@dots{}
38848 @itemx QCatchSyscalls:0
38849 @cindex catch syscalls from inferior, remote request
38850 @cindex @samp{QCatchSyscalls} packet
38851 @anchor{QCatchSyscalls}
38852 Enable (@samp{QCatchSyscalls:1}) or disable (@samp{QCatchSyscalls:0})
38853 catching syscalls from the inferior process.
38855 For @samp{QCatchSyscalls:1}, each listed syscall @var{sysno} (encoded
38856 in hex) should be reported to @value{GDBN}. If no syscall @var{sysno}
38857 is listed, every system call should be reported.
38859 Note that if a syscall not in the list is reported, @value{GDBN} will
38860 still filter the event according to its own list from all corresponding
38861 @code{catch syscall} commands. However, it is more efficient to only
38862 report the requested syscalls.
38864 Multiple @samp{QCatchSyscalls:1} packets do not combine; any earlier
38865 @samp{QCatchSyscalls:1} list is completely replaced by the new list.
38867 If the inferior process execs, the state of @samp{QCatchSyscalls} is
38868 kept for the new process too. On targets where exec may affect syscall
38869 numbers, for example with exec between 32 and 64-bit processes, the
38870 client should send a new packet with the new syscall list.
38875 The request succeeded.
38878 An error occurred. @var{nn} are hex digits.
38881 An empty reply indicates that @samp{QCatchSyscalls} is not supported by
38885 Use of this packet is controlled by the @code{set remote catch-syscalls}
38886 command (@pxref{Remote Configuration, set remote catch-syscalls}).
38887 This packet is not probed by default; the remote stub must request it,
38888 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
38890 @item QPassSignals: @var{signal} @r{[};@var{signal}@r{]}@dots{}
38891 @cindex pass signals to inferior, remote request
38892 @cindex @samp{QPassSignals} packet
38893 @anchor{QPassSignals}
38894 Each listed @var{signal} should be passed directly to the inferior process.
38895 Signals are numbered identically to continue packets and stop replies
38896 (@pxref{Stop Reply Packets}). Each @var{signal} list item should be
38897 strictly greater than the previous item. These signals do not need to stop
38898 the inferior, or be reported to @value{GDBN}. All other signals should be
38899 reported to @value{GDBN}. Multiple @samp{QPassSignals} packets do not
38900 combine; any earlier @samp{QPassSignals} list is completely replaced by the
38901 new list. This packet improves performance when using @samp{handle
38902 @var{signal} nostop noprint pass}.
38907 The request succeeded.
38910 An error occurred. The error number @var{nn} is given as hex digits.
38913 An empty reply indicates that @samp{QPassSignals} is not supported by
38917 Use of this packet is controlled by the @code{set remote pass-signals}
38918 command (@pxref{Remote Configuration, set remote pass-signals}).
38919 This packet is not probed by default; the remote stub must request it,
38920 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
38922 @item QProgramSignals: @var{signal} @r{[};@var{signal}@r{]}@dots{}
38923 @cindex signals the inferior may see, remote request
38924 @cindex @samp{QProgramSignals} packet
38925 @anchor{QProgramSignals}
38926 Each listed @var{signal} may be delivered to the inferior process.
38927 Others should be silently discarded.
38929 In some cases, the remote stub may need to decide whether to deliver a
38930 signal to the program or not without @value{GDBN} involvement. One
38931 example of that is while detaching --- the program's threads may have
38932 stopped for signals that haven't yet had a chance of being reported to
38933 @value{GDBN}, and so the remote stub can use the signal list specified
38934 by this packet to know whether to deliver or ignore those pending
38937 This does not influence whether to deliver a signal as requested by a
38938 resumption packet (@pxref{vCont packet}).
38940 Signals are numbered identically to continue packets and stop replies
38941 (@pxref{Stop Reply Packets}). Each @var{signal} list item should be
38942 strictly greater than the previous item. Multiple
38943 @samp{QProgramSignals} packets do not combine; any earlier
38944 @samp{QProgramSignals} list is completely replaced by the new list.
38949 The request succeeded.
38952 An error occurred. The error number @var{nn} is given as hex digits.
38955 An empty reply indicates that @samp{QProgramSignals} is not supported
38959 Use of this packet is controlled by the @code{set remote program-signals}
38960 command (@pxref{Remote Configuration, set remote program-signals}).
38961 This packet is not probed by default; the remote stub must request it,
38962 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
38964 @anchor{QThreadEvents}
38965 @item QThreadEvents:1
38966 @itemx QThreadEvents:0
38967 @cindex thread create/exit events, remote request
38968 @cindex @samp{QThreadEvents} packet
38970 Enable (@samp{QThreadEvents:1}) or disable (@samp{QThreadEvents:0})
38971 reporting of thread create and exit events. @xref{thread create
38972 event}, for the reply specifications. For example, this is used in
38973 non-stop mode when @value{GDBN} stops a set of threads and
38974 synchronously waits for the their corresponding stop replies. Without
38975 exit events, if one of the threads exits, @value{GDBN} would hang
38976 forever not knowing that it should no longer expect a stop for that
38977 same thread. @value{GDBN} does not enable this feature unless the
38978 stub reports that it supports it by including @samp{QThreadEvents+} in
38979 its @samp{qSupported} reply.
38984 The request succeeded.
38987 An error occurred. The error number @var{nn} is given as hex digits.
38990 An empty reply indicates that @samp{QThreadEvents} is not supported by
38994 Use of this packet is controlled by the @code{set remote thread-events}
38995 command (@pxref{Remote Configuration, set remote thread-events}).
38997 @item qRcmd,@var{command}
38998 @cindex execute remote command, remote request
38999 @cindex @samp{qRcmd} packet
39000 @var{command} (hex encoded) is passed to the local interpreter for
39001 execution. Invalid commands should be reported using the output
39002 string. Before the final result packet, the target may also respond
39003 with a number of intermediate @samp{O@var{output}} console output
39004 packets. @emph{Implementors should note that providing access to a
39005 stubs's interpreter may have security implications}.
39010 A command response with no output.
39012 A command response with the hex encoded output string @var{OUTPUT}.
39014 Indicate a badly formed request.
39016 An empty reply indicates that @samp{qRcmd} is not recognized.
39019 (Note that the @code{qRcmd} packet's name is separated from the
39020 command by a @samp{,}, not a @samp{:}, contrary to the naming
39021 conventions above. Please don't use this packet as a model for new
39024 @item qSearch:memory:@var{address};@var{length};@var{search-pattern}
39025 @cindex searching memory, in remote debugging
39027 @cindex @samp{qSearch:memory} packet
39029 @cindex @samp{qSearch memory} packet
39030 @anchor{qSearch memory}
39031 Search @var{length} bytes at @var{address} for @var{search-pattern}.
39032 Both @var{address} and @var{length} are encoded in hex;
39033 @var{search-pattern} is a sequence of bytes, also hex encoded.
39038 The pattern was not found.
39040 The pattern was found at @var{address}.
39042 A badly formed request or an error was encountered while searching memory.
39044 An empty reply indicates that @samp{qSearch:memory} is not recognized.
39047 @item QStartNoAckMode
39048 @cindex @samp{QStartNoAckMode} packet
39049 @anchor{QStartNoAckMode}
39050 Request that the remote stub disable the normal @samp{+}/@samp{-}
39051 protocol acknowledgments (@pxref{Packet Acknowledgment}).
39056 The stub has switched to no-acknowledgment mode.
39057 @value{GDBN} acknowledges this reponse,
39058 but neither the stub nor @value{GDBN} shall send or expect further
39059 @samp{+}/@samp{-} acknowledgments in the current connection.
39061 An empty reply indicates that the stub does not support no-acknowledgment mode.
39064 @item qSupported @r{[}:@var{gdbfeature} @r{[};@var{gdbfeature}@r{]}@dots{} @r{]}
39065 @cindex supported packets, remote query
39066 @cindex features of the remote protocol
39067 @cindex @samp{qSupported} packet
39068 @anchor{qSupported}
39069 Tell the remote stub about features supported by @value{GDBN}, and
39070 query the stub for features it supports. This packet allows
39071 @value{GDBN} and the remote stub to take advantage of each others'
39072 features. @samp{qSupported} also consolidates multiple feature probes
39073 at startup, to improve @value{GDBN} performance---a single larger
39074 packet performs better than multiple smaller probe packets on
39075 high-latency links. Some features may enable behavior which must not
39076 be on by default, e.g.@: because it would confuse older clients or
39077 stubs. Other features may describe packets which could be
39078 automatically probed for, but are not. These features must be
39079 reported before @value{GDBN} will use them. This ``default
39080 unsupported'' behavior is not appropriate for all packets, but it
39081 helps to keep the initial connection time under control with new
39082 versions of @value{GDBN} which support increasing numbers of packets.
39086 @item @var{stubfeature} @r{[};@var{stubfeature}@r{]}@dots{}
39087 The stub supports or does not support each returned @var{stubfeature},
39088 depending on the form of each @var{stubfeature} (see below for the
39091 An empty reply indicates that @samp{qSupported} is not recognized,
39092 or that no features needed to be reported to @value{GDBN}.
39095 The allowed forms for each feature (either a @var{gdbfeature} in the
39096 @samp{qSupported} packet, or a @var{stubfeature} in the response)
39100 @item @var{name}=@var{value}
39101 The remote protocol feature @var{name} is supported, and associated
39102 with the specified @var{value}. The format of @var{value} depends
39103 on the feature, but it must not include a semicolon.
39105 The remote protocol feature @var{name} is supported, and does not
39106 need an associated value.
39108 The remote protocol feature @var{name} is not supported.
39110 The remote protocol feature @var{name} may be supported, and
39111 @value{GDBN} should auto-detect support in some other way when it is
39112 needed. This form will not be used for @var{gdbfeature} notifications,
39113 but may be used for @var{stubfeature} responses.
39116 Whenever the stub receives a @samp{qSupported} request, the
39117 supplied set of @value{GDBN} features should override any previous
39118 request. This allows @value{GDBN} to put the stub in a known
39119 state, even if the stub had previously been communicating with
39120 a different version of @value{GDBN}.
39122 The following values of @var{gdbfeature} (for the packet sent by @value{GDBN})
39127 This feature indicates whether @value{GDBN} supports multiprocess
39128 extensions to the remote protocol. @value{GDBN} does not use such
39129 extensions unless the stub also reports that it supports them by
39130 including @samp{multiprocess+} in its @samp{qSupported} reply.
39131 @xref{multiprocess extensions}, for details.
39134 This feature indicates that @value{GDBN} supports the XML target
39135 description. If the stub sees @samp{xmlRegisters=} with target
39136 specific strings separated by a comma, it will report register
39140 This feature indicates whether @value{GDBN} supports the
39141 @samp{qRelocInsn} packet (@pxref{Tracepoint Packets,,Relocate
39142 instruction reply packet}).
39145 This feature indicates whether @value{GDBN} supports the swbreak stop
39146 reason in stop replies. @xref{swbreak stop reason}, for details.
39149 This feature indicates whether @value{GDBN} supports the hwbreak stop
39150 reason in stop replies. @xref{swbreak stop reason}, for details.
39153 This feature indicates whether @value{GDBN} supports fork event
39154 extensions to the remote protocol. @value{GDBN} does not use such
39155 extensions unless the stub also reports that it supports them by
39156 including @samp{fork-events+} in its @samp{qSupported} reply.
39159 This feature indicates whether @value{GDBN} supports vfork event
39160 extensions to the remote protocol. @value{GDBN} does not use such
39161 extensions unless the stub also reports that it supports them by
39162 including @samp{vfork-events+} in its @samp{qSupported} reply.
39165 This feature indicates whether @value{GDBN} supports exec event
39166 extensions to the remote protocol. @value{GDBN} does not use such
39167 extensions unless the stub also reports that it supports them by
39168 including @samp{exec-events+} in its @samp{qSupported} reply.
39170 @item vContSupported
39171 This feature indicates whether @value{GDBN} wants to know the
39172 supported actions in the reply to @samp{vCont?} packet.
39175 Stubs should ignore any unknown values for
39176 @var{gdbfeature}. Any @value{GDBN} which sends a @samp{qSupported}
39177 packet supports receiving packets of unlimited length (earlier
39178 versions of @value{GDBN} may reject overly long responses). Additional values
39179 for @var{gdbfeature} may be defined in the future to let the stub take
39180 advantage of new features in @value{GDBN}, e.g.@: incompatible
39181 improvements in the remote protocol---the @samp{multiprocess} feature is
39182 an example of such a feature. The stub's reply should be independent
39183 of the @var{gdbfeature} entries sent by @value{GDBN}; first @value{GDBN}
39184 describes all the features it supports, and then the stub replies with
39185 all the features it supports.
39187 Similarly, @value{GDBN} will silently ignore unrecognized stub feature
39188 responses, as long as each response uses one of the standard forms.
39190 Some features are flags. A stub which supports a flag feature
39191 should respond with a @samp{+} form response. Other features
39192 require values, and the stub should respond with an @samp{=}
39195 Each feature has a default value, which @value{GDBN} will use if
39196 @samp{qSupported} is not available or if the feature is not mentioned
39197 in the @samp{qSupported} response. The default values are fixed; a
39198 stub is free to omit any feature responses that match the defaults.
39200 Not all features can be probed, but for those which can, the probing
39201 mechanism is useful: in some cases, a stub's internal
39202 architecture may not allow the protocol layer to know some information
39203 about the underlying target in advance. This is especially common in
39204 stubs which may be configured for multiple targets.
39206 These are the currently defined stub features and their properties:
39208 @multitable @columnfractions 0.35 0.2 0.12 0.2
39209 @c NOTE: The first row should be @headitem, but we do not yet require
39210 @c a new enough version of Texinfo (4.7) to use @headitem.
39212 @tab Value Required
39216 @item @samp{PacketSize}
39221 @item @samp{qXfer:auxv:read}
39226 @item @samp{qXfer:btrace:read}
39231 @item @samp{qXfer:btrace-conf:read}
39236 @item @samp{qXfer:exec-file:read}
39241 @item @samp{qXfer:features:read}
39246 @item @samp{qXfer:libraries:read}
39251 @item @samp{qXfer:libraries-svr4:read}
39256 @item @samp{augmented-libraries-svr4-read}
39261 @item @samp{qXfer:memory-map:read}
39266 @item @samp{qXfer:sdata:read}
39271 @item @samp{qXfer:spu:read}
39276 @item @samp{qXfer:spu:write}
39281 @item @samp{qXfer:siginfo:read}
39286 @item @samp{qXfer:siginfo:write}
39291 @item @samp{qXfer:threads:read}
39296 @item @samp{qXfer:traceframe-info:read}
39301 @item @samp{qXfer:uib:read}
39306 @item @samp{qXfer:fdpic:read}
39311 @item @samp{Qbtrace:off}
39316 @item @samp{Qbtrace:bts}
39321 @item @samp{Qbtrace:pt}
39326 @item @samp{Qbtrace-conf:bts:size}
39331 @item @samp{Qbtrace-conf:pt:size}
39336 @item @samp{QNonStop}
39341 @item @samp{QCatchSyscalls}
39346 @item @samp{QPassSignals}
39351 @item @samp{QStartNoAckMode}
39356 @item @samp{multiprocess}
39361 @item @samp{ConditionalBreakpoints}
39366 @item @samp{ConditionalTracepoints}
39371 @item @samp{ReverseContinue}
39376 @item @samp{ReverseStep}
39381 @item @samp{TracepointSource}
39386 @item @samp{QAgent}
39391 @item @samp{QAllow}
39396 @item @samp{QDisableRandomization}
39401 @item @samp{EnableDisableTracepoints}
39406 @item @samp{QTBuffer:size}
39411 @item @samp{tracenz}
39416 @item @samp{BreakpointCommands}
39421 @item @samp{swbreak}
39426 @item @samp{hwbreak}
39431 @item @samp{fork-events}
39436 @item @samp{vfork-events}
39441 @item @samp{exec-events}
39446 @item @samp{QThreadEvents}
39451 @item @samp{no-resumed}
39458 These are the currently defined stub features, in more detail:
39461 @cindex packet size, remote protocol
39462 @item PacketSize=@var{bytes}
39463 The remote stub can accept packets up to at least @var{bytes} in
39464 length. @value{GDBN} will send packets up to this size for bulk
39465 transfers, and will never send larger packets. This is a limit on the
39466 data characters in the packet, including the frame and checksum.
39467 There is no trailing NUL byte in a remote protocol packet; if the stub
39468 stores packets in a NUL-terminated format, it should allow an extra
39469 byte in its buffer for the NUL. If this stub feature is not supported,
39470 @value{GDBN} guesses based on the size of the @samp{g} packet response.
39472 @item qXfer:auxv:read
39473 The remote stub understands the @samp{qXfer:auxv:read} packet
39474 (@pxref{qXfer auxiliary vector read}).
39476 @item qXfer:btrace:read
39477 The remote stub understands the @samp{qXfer:btrace:read}
39478 packet (@pxref{qXfer btrace read}).
39480 @item qXfer:btrace-conf:read
39481 The remote stub understands the @samp{qXfer:btrace-conf:read}
39482 packet (@pxref{qXfer btrace-conf read}).
39484 @item qXfer:exec-file:read
39485 The remote stub understands the @samp{qXfer:exec-file:read} packet
39486 (@pxref{qXfer executable filename read}).
39488 @item qXfer:features:read
39489 The remote stub understands the @samp{qXfer:features:read} packet
39490 (@pxref{qXfer target description read}).
39492 @item qXfer:libraries:read
39493 The remote stub understands the @samp{qXfer:libraries:read} packet
39494 (@pxref{qXfer library list read}).
39496 @item qXfer:libraries-svr4:read
39497 The remote stub understands the @samp{qXfer:libraries-svr4:read} packet
39498 (@pxref{qXfer svr4 library list read}).
39500 @item augmented-libraries-svr4-read
39501 The remote stub understands the augmented form of the
39502 @samp{qXfer:libraries-svr4:read} packet
39503 (@pxref{qXfer svr4 library list read}).
39505 @item qXfer:memory-map:read
39506 The remote stub understands the @samp{qXfer:memory-map:read} packet
39507 (@pxref{qXfer memory map read}).
39509 @item qXfer:sdata:read
39510 The remote stub understands the @samp{qXfer:sdata:read} packet
39511 (@pxref{qXfer sdata read}).
39513 @item qXfer:spu:read
39514 The remote stub understands the @samp{qXfer:spu:read} packet
39515 (@pxref{qXfer spu read}).
39517 @item qXfer:spu:write
39518 The remote stub understands the @samp{qXfer:spu:write} packet
39519 (@pxref{qXfer spu write}).
39521 @item qXfer:siginfo:read
39522 The remote stub understands the @samp{qXfer:siginfo:read} packet
39523 (@pxref{qXfer siginfo read}).
39525 @item qXfer:siginfo:write
39526 The remote stub understands the @samp{qXfer:siginfo:write} packet
39527 (@pxref{qXfer siginfo write}).
39529 @item qXfer:threads:read
39530 The remote stub understands the @samp{qXfer:threads:read} packet
39531 (@pxref{qXfer threads read}).
39533 @item qXfer:traceframe-info:read
39534 The remote stub understands the @samp{qXfer:traceframe-info:read}
39535 packet (@pxref{qXfer traceframe info read}).
39537 @item qXfer:uib:read
39538 The remote stub understands the @samp{qXfer:uib:read}
39539 packet (@pxref{qXfer unwind info block}).
39541 @item qXfer:fdpic:read
39542 The remote stub understands the @samp{qXfer:fdpic:read}
39543 packet (@pxref{qXfer fdpic loadmap read}).
39546 The remote stub understands the @samp{QNonStop} packet
39547 (@pxref{QNonStop}).
39549 @item QCatchSyscalls
39550 The remote stub understands the @samp{QCatchSyscalls} packet
39551 (@pxref{QCatchSyscalls}).
39554 The remote stub understands the @samp{QPassSignals} packet
39555 (@pxref{QPassSignals}).
39557 @item QStartNoAckMode
39558 The remote stub understands the @samp{QStartNoAckMode} packet and
39559 prefers to operate in no-acknowledgment mode. @xref{Packet Acknowledgment}.
39562 @anchor{multiprocess extensions}
39563 @cindex multiprocess extensions, in remote protocol
39564 The remote stub understands the multiprocess extensions to the remote
39565 protocol syntax. The multiprocess extensions affect the syntax of
39566 thread IDs in both packets and replies (@pxref{thread-id syntax}), and
39567 add process IDs to the @samp{D} packet and @samp{W} and @samp{X}
39568 replies. Note that reporting this feature indicates support for the
39569 syntactic extensions only, not that the stub necessarily supports
39570 debugging of more than one process at a time. The stub must not use
39571 multiprocess extensions in packet replies unless @value{GDBN} has also
39572 indicated it supports them in its @samp{qSupported} request.
39574 @item qXfer:osdata:read
39575 The remote stub understands the @samp{qXfer:osdata:read} packet
39576 ((@pxref{qXfer osdata read}).
39578 @item ConditionalBreakpoints
39579 The target accepts and implements evaluation of conditional expressions
39580 defined for breakpoints. The target will only report breakpoint triggers
39581 when such conditions are true (@pxref{Conditions, ,Break Conditions}).
39583 @item ConditionalTracepoints
39584 The remote stub accepts and implements conditional expressions defined
39585 for tracepoints (@pxref{Tracepoint Conditions}).
39587 @item ReverseContinue
39588 The remote stub accepts and implements the reverse continue packet
39592 The remote stub accepts and implements the reverse step packet
39595 @item TracepointSource
39596 The remote stub understands the @samp{QTDPsrc} packet that supplies
39597 the source form of tracepoint definitions.
39600 The remote stub understands the @samp{QAgent} packet.
39603 The remote stub understands the @samp{QAllow} packet.
39605 @item QDisableRandomization
39606 The remote stub understands the @samp{QDisableRandomization} packet.
39608 @item StaticTracepoint
39609 @cindex static tracepoints, in remote protocol
39610 The remote stub supports static tracepoints.
39612 @item InstallInTrace
39613 @anchor{install tracepoint in tracing}
39614 The remote stub supports installing tracepoint in tracing.
39616 @item EnableDisableTracepoints
39617 The remote stub supports the @samp{QTEnable} (@pxref{QTEnable}) and
39618 @samp{QTDisable} (@pxref{QTDisable}) packets that allow tracepoints
39619 to be enabled and disabled while a trace experiment is running.
39621 @item QTBuffer:size
39622 The remote stub supports the @samp{QTBuffer:size} (@pxref{QTBuffer-size})
39623 packet that allows to change the size of the trace buffer.
39626 @cindex string tracing, in remote protocol
39627 The remote stub supports the @samp{tracenz} bytecode for collecting strings.
39628 See @ref{Bytecode Descriptions} for details about the bytecode.
39630 @item BreakpointCommands
39631 @cindex breakpoint commands, in remote protocol
39632 The remote stub supports running a breakpoint's command list itself,
39633 rather than reporting the hit to @value{GDBN}.
39636 The remote stub understands the @samp{Qbtrace:off} packet.
39639 The remote stub understands the @samp{Qbtrace:bts} packet.
39642 The remote stub understands the @samp{Qbtrace:pt} packet.
39644 @item Qbtrace-conf:bts:size
39645 The remote stub understands the @samp{Qbtrace-conf:bts:size} packet.
39647 @item Qbtrace-conf:pt:size
39648 The remote stub understands the @samp{Qbtrace-conf:pt:size} packet.
39651 The remote stub reports the @samp{swbreak} stop reason for memory
39655 The remote stub reports the @samp{hwbreak} stop reason for hardware
39659 The remote stub reports the @samp{fork} stop reason for fork events.
39662 The remote stub reports the @samp{vfork} stop reason for vfork events
39663 and vforkdone events.
39666 The remote stub reports the @samp{exec} stop reason for exec events.
39668 @item vContSupported
39669 The remote stub reports the supported actions in the reply to
39670 @samp{vCont?} packet.
39672 @item QThreadEvents
39673 The remote stub understands the @samp{QThreadEvents} packet.
39676 The remote stub reports the @samp{N} stop reply.
39681 @cindex symbol lookup, remote request
39682 @cindex @samp{qSymbol} packet
39683 Notify the target that @value{GDBN} is prepared to serve symbol lookup
39684 requests. Accept requests from the target for the values of symbols.
39689 The target does not need to look up any (more) symbols.
39690 @item qSymbol:@var{sym_name}
39691 The target requests the value of symbol @var{sym_name} (hex encoded).
39692 @value{GDBN} may provide the value by using the
39693 @samp{qSymbol:@var{sym_value}:@var{sym_name}} message, described
39697 @item qSymbol:@var{sym_value}:@var{sym_name}
39698 Set the value of @var{sym_name} to @var{sym_value}.
39700 @var{sym_name} (hex encoded) is the name of a symbol whose value the
39701 target has previously requested.
39703 @var{sym_value} (hex) is the value for symbol @var{sym_name}. If
39704 @value{GDBN} cannot supply a value for @var{sym_name}, then this field
39710 The target does not need to look up any (more) symbols.
39711 @item qSymbol:@var{sym_name}
39712 The target requests the value of a new symbol @var{sym_name} (hex
39713 encoded). @value{GDBN} will continue to supply the values of symbols
39714 (if available), until the target ceases to request them.
39719 @itemx QTDisconnected
39726 @itemx qTMinFTPILen
39728 @xref{Tracepoint Packets}.
39730 @item qThreadExtraInfo,@var{thread-id}
39731 @cindex thread attributes info, remote request
39732 @cindex @samp{qThreadExtraInfo} packet
39733 Obtain from the target OS a printable string description of thread
39734 attributes for the thread @var{thread-id}; see @ref{thread-id syntax},
39735 for the forms of @var{thread-id}. This
39736 string may contain anything that the target OS thinks is interesting
39737 for @value{GDBN} to tell the user about the thread. The string is
39738 displayed in @value{GDBN}'s @code{info threads} display. Some
39739 examples of possible thread extra info strings are @samp{Runnable}, or
39740 @samp{Blocked on Mutex}.
39744 @item @var{XX}@dots{}
39745 Where @samp{@var{XX}@dots{}} is a hex encoding of @sc{ascii} data,
39746 comprising the printable string containing the extra information about
39747 the thread's attributes.
39750 (Note that the @code{qThreadExtraInfo} packet's name is separated from
39751 the command by a @samp{,}, not a @samp{:}, contrary to the naming
39752 conventions above. Please don't use this packet as a model for new
39771 @xref{Tracepoint Packets}.
39773 @item qXfer:@var{object}:read:@var{annex}:@var{offset},@var{length}
39774 @cindex read special object, remote request
39775 @cindex @samp{qXfer} packet
39776 @anchor{qXfer read}
39777 Read uninterpreted bytes from the target's special data area
39778 identified by the keyword @var{object}. Request @var{length} bytes
39779 starting at @var{offset} bytes into the data. The content and
39780 encoding of @var{annex} is specific to @var{object}; it can supply
39781 additional details about what data to access.
39786 Data @var{data} (@pxref{Binary Data}) has been read from the
39787 target. There may be more data at a higher address (although
39788 it is permitted to return @samp{m} even for the last valid
39789 block of data, as long as at least one byte of data was read).
39790 It is possible for @var{data} to have fewer bytes than the @var{length} in the
39794 Data @var{data} (@pxref{Binary Data}) has been read from the target.
39795 There is no more data to be read. It is possible for @var{data} to
39796 have fewer bytes than the @var{length} in the request.
39799 The @var{offset} in the request is at the end of the data.
39800 There is no more data to be read.
39803 The request was malformed, or @var{annex} was invalid.
39806 The offset was invalid, or there was an error encountered reading the data.
39807 The @var{nn} part is a hex-encoded @code{errno} value.
39810 An empty reply indicates the @var{object} string was not recognized by
39811 the stub, or that the object does not support reading.
39814 Here are the specific requests of this form defined so far. All the
39815 @samp{qXfer:@var{object}:read:@dots{}} requests use the same reply
39816 formats, listed above.
39819 @item qXfer:auxv:read::@var{offset},@var{length}
39820 @anchor{qXfer auxiliary vector read}
39821 Access the target's @dfn{auxiliary vector}. @xref{OS Information,
39822 auxiliary vector}. Note @var{annex} must be empty.
39824 This packet is not probed by default; the remote stub must request it,
39825 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
39827 @item qXfer:btrace:read:@var{annex}:@var{offset},@var{length}
39828 @anchor{qXfer btrace read}
39830 Return a description of the current branch trace.
39831 @xref{Branch Trace Format}. The annex part of the generic @samp{qXfer}
39832 packet may have one of the following values:
39836 Returns all available branch trace.
39839 Returns all available branch trace if the branch trace changed since
39840 the last read request.
39843 Returns the new branch trace since the last read request. Adds a new
39844 block to the end of the trace that begins at zero and ends at the source
39845 location of the first branch in the trace buffer. This extra block is
39846 used to stitch traces together.
39848 If the trace buffer overflowed, returns an error indicating the overflow.
39851 This packet is not probed by default; the remote stub must request it
39852 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
39854 @item qXfer:btrace-conf:read::@var{offset},@var{length}
39855 @anchor{qXfer btrace-conf read}
39857 Return a description of the current branch trace configuration.
39858 @xref{Branch Trace Configuration Format}.
39860 This packet is not probed by default; the remote stub must request it
39861 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
39863 @item qXfer:exec-file:read:@var{annex}:@var{offset},@var{length}
39864 @anchor{qXfer executable filename read}
39865 Return the full absolute name of the file that was executed to create
39866 a process running on the remote system. The annex specifies the
39867 numeric process ID of the process to query, encoded as a hexadecimal
39868 number. If the annex part is empty the remote stub should return the
39869 filename corresponding to the currently executing process.
39871 This packet is not probed by default; the remote stub must request it,
39872 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
39874 @item qXfer:features:read:@var{annex}:@var{offset},@var{length}
39875 @anchor{qXfer target description read}
39876 Access the @dfn{target description}. @xref{Target Descriptions}. The
39877 annex specifies which XML document to access. The main description is
39878 always loaded from the @samp{target.xml} annex.
39880 This packet is not probed by default; the remote stub must request it,
39881 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
39883 @item qXfer:libraries:read:@var{annex}:@var{offset},@var{length}
39884 @anchor{qXfer library list read}
39885 Access the target's list of loaded libraries. @xref{Library List Format}.
39886 The annex part of the generic @samp{qXfer} packet must be empty
39887 (@pxref{qXfer read}).
39889 Targets which maintain a list of libraries in the program's memory do
39890 not need to implement this packet; it is designed for platforms where
39891 the operating system manages the list of loaded libraries.
39893 This packet is not probed by default; the remote stub must request it,
39894 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
39896 @item qXfer:libraries-svr4:read:@var{annex}:@var{offset},@var{length}
39897 @anchor{qXfer svr4 library list read}
39898 Access the target's list of loaded libraries when the target is an SVR4
39899 platform. @xref{Library List Format for SVR4 Targets}. The annex part
39900 of the generic @samp{qXfer} packet must be empty unless the remote
39901 stub indicated it supports the augmented form of this packet
39902 by supplying an appropriate @samp{qSupported} response
39903 (@pxref{qXfer read}, @ref{qSupported}).
39905 This packet is optional for better performance on SVR4 targets.
39906 @value{GDBN} uses memory read packets to read the SVR4 library list otherwise.
39908 This packet is not probed by default; the remote stub must request it,
39909 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
39911 If the remote stub indicates it supports the augmented form of this
39912 packet then the annex part of the generic @samp{qXfer} packet may
39913 contain a semicolon-separated list of @samp{@var{name}=@var{value}}
39914 arguments. The currently supported arguments are:
39917 @item start=@var{address}
39918 A hexadecimal number specifying the address of the @samp{struct
39919 link_map} to start reading the library list from. If unset or zero
39920 then the first @samp{struct link_map} in the library list will be
39921 chosen as the starting point.
39923 @item prev=@var{address}
39924 A hexadecimal number specifying the address of the @samp{struct
39925 link_map} immediately preceding the @samp{struct link_map}
39926 specified by the @samp{start} argument. If unset or zero then
39927 the remote stub will expect that no @samp{struct link_map}
39928 exists prior to the starting point.
39932 Arguments that are not understood by the remote stub will be silently
39935 @item qXfer:memory-map:read::@var{offset},@var{length}
39936 @anchor{qXfer memory map read}
39937 Access the target's @dfn{memory-map}. @xref{Memory Map Format}. The
39938 annex part of the generic @samp{qXfer} packet must be empty
39939 (@pxref{qXfer read}).
39941 This packet is not probed by default; the remote stub must request it,
39942 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
39944 @item qXfer:sdata:read::@var{offset},@var{length}
39945 @anchor{qXfer sdata read}
39947 Read contents of the extra collected static tracepoint marker
39948 information. The annex part of the generic @samp{qXfer} packet must
39949 be empty (@pxref{qXfer read}). @xref{Tracepoint Actions,,Tracepoint
39952 This packet is not probed by default; the remote stub must request it,
39953 by supplying an appropriate @samp{qSupported} response
39954 (@pxref{qSupported}).
39956 @item qXfer:siginfo:read::@var{offset},@var{length}
39957 @anchor{qXfer siginfo read}
39958 Read contents of the extra signal information on the target
39959 system. The annex part of the generic @samp{qXfer} packet must be
39960 empty (@pxref{qXfer read}).
39962 This packet is not probed by default; the remote stub must request it,
39963 by supplying an appropriate @samp{qSupported} response
39964 (@pxref{qSupported}).
39966 @item qXfer:spu:read:@var{annex}:@var{offset},@var{length}
39967 @anchor{qXfer spu read}
39968 Read contents of an @code{spufs} file on the target system. The
39969 annex specifies which file to read; it must be of the form
39970 @file{@var{id}/@var{name}}, where @var{id} specifies an SPU context ID
39971 in the target process, and @var{name} identifes the @code{spufs} file
39972 in that context to be accessed.
39974 This packet is not probed by default; the remote stub must request it,
39975 by supplying an appropriate @samp{qSupported} response
39976 (@pxref{qSupported}).
39978 @item qXfer:threads:read::@var{offset},@var{length}
39979 @anchor{qXfer threads read}
39980 Access the list of threads on target. @xref{Thread List Format}. The
39981 annex part of the generic @samp{qXfer} packet must be empty
39982 (@pxref{qXfer read}).
39984 This packet is not probed by default; the remote stub must request it,
39985 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
39987 @item qXfer:traceframe-info:read::@var{offset},@var{length}
39988 @anchor{qXfer traceframe info read}
39990 Return a description of the current traceframe's contents.
39991 @xref{Traceframe Info Format}. The annex part of the generic
39992 @samp{qXfer} packet must be empty (@pxref{qXfer read}).
39994 This packet is not probed by default; the remote stub must request it,
39995 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
39997 @item qXfer:uib:read:@var{pc}:@var{offset},@var{length}
39998 @anchor{qXfer unwind info block}
40000 Return the unwind information block for @var{pc}. This packet is used
40001 on OpenVMS/ia64 to ask the kernel unwind information.
40003 This packet is not probed by default.
40005 @item qXfer:fdpic:read:@var{annex}:@var{offset},@var{length}
40006 @anchor{qXfer fdpic loadmap read}
40007 Read contents of @code{loadmap}s on the target system. The
40008 annex, either @samp{exec} or @samp{interp}, specifies which @code{loadmap},
40009 executable @code{loadmap} or interpreter @code{loadmap} to read.
40011 This packet is not probed by default; the remote stub must request it,
40012 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
40014 @item qXfer:osdata:read::@var{offset},@var{length}
40015 @anchor{qXfer osdata read}
40016 Access the target's @dfn{operating system information}.
40017 @xref{Operating System Information}.
40021 @item qXfer:@var{object}:write:@var{annex}:@var{offset}:@var{data}@dots{}
40022 @cindex write data into object, remote request
40023 @anchor{qXfer write}
40024 Write uninterpreted bytes into the target's special data area
40025 identified by the keyword @var{object}, starting at @var{offset} bytes
40026 into the data. The binary-encoded data (@pxref{Binary Data}) to be
40027 written is given by @var{data}@dots{}. The content and encoding of @var{annex}
40028 is specific to @var{object}; it can supply additional details about what data
40034 @var{nn} (hex encoded) is the number of bytes written.
40035 This may be fewer bytes than supplied in the request.
40038 The request was malformed, or @var{annex} was invalid.
40041 The offset was invalid, or there was an error encountered writing the data.
40042 The @var{nn} part is a hex-encoded @code{errno} value.
40045 An empty reply indicates the @var{object} string was not
40046 recognized by the stub, or that the object does not support writing.
40049 Here are the specific requests of this form defined so far. All the
40050 @samp{qXfer:@var{object}:write:@dots{}} requests use the same reply
40051 formats, listed above.
40054 @item qXfer:siginfo:write::@var{offset}:@var{data}@dots{}
40055 @anchor{qXfer siginfo write}
40056 Write @var{data} to the extra signal information on the target system.
40057 The annex part of the generic @samp{qXfer} packet must be
40058 empty (@pxref{qXfer write}).
40060 This packet is not probed by default; the remote stub must request it,
40061 by supplying an appropriate @samp{qSupported} response
40062 (@pxref{qSupported}).
40064 @item qXfer:spu:write:@var{annex}:@var{offset}:@var{data}@dots{}
40065 @anchor{qXfer spu write}
40066 Write @var{data} to an @code{spufs} file on the target system. The
40067 annex specifies which file to write; it must be of the form
40068 @file{@var{id}/@var{name}}, where @var{id} specifies an SPU context ID
40069 in the target process, and @var{name} identifes the @code{spufs} file
40070 in that context to be accessed.
40072 This packet is not probed by default; the remote stub must request it,
40073 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
40076 @item qXfer:@var{object}:@var{operation}:@dots{}
40077 Requests of this form may be added in the future. When a stub does
40078 not recognize the @var{object} keyword, or its support for
40079 @var{object} does not recognize the @var{operation} keyword, the stub
40080 must respond with an empty packet.
40082 @item qAttached:@var{pid}
40083 @cindex query attached, remote request
40084 @cindex @samp{qAttached} packet
40085 Return an indication of whether the remote server attached to an
40086 existing process or created a new process. When the multiprocess
40087 protocol extensions are supported (@pxref{multiprocess extensions}),
40088 @var{pid} is an integer in hexadecimal format identifying the target
40089 process. Otherwise, @value{GDBN} will omit the @var{pid} field and
40090 the query packet will be simplified as @samp{qAttached}.
40092 This query is used, for example, to know whether the remote process
40093 should be detached or killed when a @value{GDBN} session is ended with
40094 the @code{quit} command.
40099 The remote server attached to an existing process.
40101 The remote server created a new process.
40103 A badly formed request or an error was encountered.
40107 Enable branch tracing for the current thread using Branch Trace Store.
40112 Branch tracing has been enabled.
40114 A badly formed request or an error was encountered.
40118 Enable branch tracing for the current thread using Intel Processor Trace.
40123 Branch tracing has been enabled.
40125 A badly formed request or an error was encountered.
40129 Disable branch tracing for the current thread.
40134 Branch tracing has been disabled.
40136 A badly formed request or an error was encountered.
40139 @item Qbtrace-conf:bts:size=@var{value}
40140 Set the requested ring buffer size for new threads that use the
40141 btrace recording method in bts format.
40146 The ring buffer size has been set.
40148 A badly formed request or an error was encountered.
40151 @item Qbtrace-conf:pt:size=@var{value}
40152 Set the requested ring buffer size for new threads that use the
40153 btrace recording method in pt format.
40158 The ring buffer size has been set.
40160 A badly formed request or an error was encountered.
40165 @node Architecture-Specific Protocol Details
40166 @section Architecture-Specific Protocol Details
40168 This section describes how the remote protocol is applied to specific
40169 target architectures. Also see @ref{Standard Target Features}, for
40170 details of XML target descriptions for each architecture.
40173 * ARM-Specific Protocol Details::
40174 * MIPS-Specific Protocol Details::
40177 @node ARM-Specific Protocol Details
40178 @subsection @acronym{ARM}-specific Protocol Details
40181 * ARM Breakpoint Kinds::
40184 @node ARM Breakpoint Kinds
40185 @subsubsection @acronym{ARM} Breakpoint Kinds
40186 @cindex breakpoint kinds, @acronym{ARM}
40188 These breakpoint kinds are defined for the @samp{Z0} and @samp{Z1} packets.
40193 16-bit Thumb mode breakpoint.
40196 32-bit Thumb mode (Thumb-2) breakpoint.
40199 32-bit @acronym{ARM} mode breakpoint.
40203 @node MIPS-Specific Protocol Details
40204 @subsection @acronym{MIPS}-specific Protocol Details
40207 * MIPS Register packet Format::
40208 * MIPS Breakpoint Kinds::
40211 @node MIPS Register packet Format
40212 @subsubsection @acronym{MIPS} Register Packet Format
40213 @cindex register packet format, @acronym{MIPS}
40215 The following @code{g}/@code{G} packets have previously been defined.
40216 In the below, some thirty-two bit registers are transferred as
40217 sixty-four bits. Those registers should be zero/sign extended (which?)
40218 to fill the space allocated. Register bytes are transferred in target
40219 byte order. The two nibbles within a register byte are transferred
40220 most-significant -- least-significant.
40225 All registers are transferred as thirty-two bit quantities in the order:
40226 32 general-purpose; sr; lo; hi; bad; cause; pc; 32 floating-point
40227 registers; fsr; fir; fp.
40230 All registers are transferred as sixty-four bit quantities (including
40231 thirty-two bit registers such as @code{sr}). The ordering is the same
40236 @node MIPS Breakpoint Kinds
40237 @subsubsection @acronym{MIPS} Breakpoint Kinds
40238 @cindex breakpoint kinds, @acronym{MIPS}
40240 These breakpoint kinds are defined for the @samp{Z0} and @samp{Z1} packets.
40245 16-bit @acronym{MIPS16} mode breakpoint.
40248 16-bit @acronym{microMIPS} mode breakpoint.
40251 32-bit standard @acronym{MIPS} mode breakpoint.
40254 32-bit @acronym{microMIPS} mode breakpoint.
40258 @node Tracepoint Packets
40259 @section Tracepoint Packets
40260 @cindex tracepoint packets
40261 @cindex packets, tracepoint
40263 Here we describe the packets @value{GDBN} uses to implement
40264 tracepoints (@pxref{Tracepoints}).
40268 @item QTDP:@var{n}:@var{addr}:@var{ena}:@var{step}:@var{pass}[:F@var{flen}][:X@var{len},@var{bytes}]@r{[}-@r{]}
40269 @cindex @samp{QTDP} packet
40270 Create a new tracepoint, number @var{n}, at @var{addr}. If @var{ena}
40271 is @samp{E}, then the tracepoint is enabled; if it is @samp{D}, then
40272 the tracepoint is disabled. The @var{step} gives the tracepoint's step
40273 count, and @var{pass} gives its pass count. If an @samp{F} is present,
40274 then the tracepoint is to be a fast tracepoint, and the @var{flen} is
40275 the number of bytes that the target should copy elsewhere to make room
40276 for the tracepoint. If an @samp{X} is present, it introduces a
40277 tracepoint condition, which consists of a hexadecimal length, followed
40278 by a comma and hex-encoded bytes, in a manner similar to action
40279 encodings as described below. If the trailing @samp{-} is present,
40280 further @samp{QTDP} packets will follow to specify this tracepoint's
40286 The packet was understood and carried out.
40288 @xref{Tracepoint Packets,,Relocate instruction reply packet}.
40290 The packet was not recognized.
40293 @item QTDP:-@var{n}:@var{addr}:@r{[}S@r{]}@var{action}@dots{}@r{[}-@r{]}
40294 Define actions to be taken when a tracepoint is hit. The @var{n} and
40295 @var{addr} must be the same as in the initial @samp{QTDP} packet for
40296 this tracepoint. This packet may only be sent immediately after
40297 another @samp{QTDP} packet that ended with a @samp{-}. If the
40298 trailing @samp{-} is present, further @samp{QTDP} packets will follow,
40299 specifying more actions for this tracepoint.
40301 In the series of action packets for a given tracepoint, at most one
40302 can have an @samp{S} before its first @var{action}. If such a packet
40303 is sent, it and the following packets define ``while-stepping''
40304 actions. Any prior packets define ordinary actions --- that is, those
40305 taken when the tracepoint is first hit. If no action packet has an
40306 @samp{S}, then all the packets in the series specify ordinary
40307 tracepoint actions.
40309 The @samp{@var{action}@dots{}} portion of the packet is a series of
40310 actions, concatenated without separators. Each action has one of the
40316 Collect the registers whose bits are set in @var{mask},
40317 a hexadecimal number whose @var{i}'th bit is set if register number
40318 @var{i} should be collected. (The least significant bit is numbered
40319 zero.) Note that @var{mask} may be any number of digits long; it may
40320 not fit in a 32-bit word.
40322 @item M @var{basereg},@var{offset},@var{len}
40323 Collect @var{len} bytes of memory starting at the address in register
40324 number @var{basereg}, plus @var{offset}. If @var{basereg} is
40325 @samp{-1}, then the range has a fixed address: @var{offset} is the
40326 address of the lowest byte to collect. The @var{basereg},
40327 @var{offset}, and @var{len} parameters are all unsigned hexadecimal
40328 values (the @samp{-1} value for @var{basereg} is a special case).
40330 @item X @var{len},@var{expr}
40331 Evaluate @var{expr}, whose length is @var{len}, and collect memory as
40332 it directs. The agent expression @var{expr} is as described in
40333 @ref{Agent Expressions}. Each byte of the expression is encoded as a
40334 two-digit hex number in the packet; @var{len} is the number of bytes
40335 in the expression (and thus one-half the number of hex digits in the
40340 Any number of actions may be packed together in a single @samp{QTDP}
40341 packet, as long as the packet does not exceed the maximum packet
40342 length (400 bytes, for many stubs). There may be only one @samp{R}
40343 action per tracepoint, and it must precede any @samp{M} or @samp{X}
40344 actions. Any registers referred to by @samp{M} and @samp{X} actions
40345 must be collected by a preceding @samp{R} action. (The
40346 ``while-stepping'' actions are treated as if they were attached to a
40347 separate tracepoint, as far as these restrictions are concerned.)
40352 The packet was understood and carried out.
40354 @xref{Tracepoint Packets,,Relocate instruction reply packet}.
40356 The packet was not recognized.
40359 @item QTDPsrc:@var{n}:@var{addr}:@var{type}:@var{start}:@var{slen}:@var{bytes}
40360 @cindex @samp{QTDPsrc} packet
40361 Specify a source string of tracepoint @var{n} at address @var{addr}.
40362 This is useful to get accurate reproduction of the tracepoints
40363 originally downloaded at the beginning of the trace run. The @var{type}
40364 is the name of the tracepoint part, such as @samp{cond} for the
40365 tracepoint's conditional expression (see below for a list of types), while
40366 @var{bytes} is the string, encoded in hexadecimal.
40368 @var{start} is the offset of the @var{bytes} within the overall source
40369 string, while @var{slen} is the total length of the source string.
40370 This is intended for handling source strings that are longer than will
40371 fit in a single packet.
40372 @c Add detailed example when this info is moved into a dedicated
40373 @c tracepoint descriptions section.
40375 The available string types are @samp{at} for the location,
40376 @samp{cond} for the conditional, and @samp{cmd} for an action command.
40377 @value{GDBN} sends a separate packet for each command in the action
40378 list, in the same order in which the commands are stored in the list.
40380 The target does not need to do anything with source strings except
40381 report them back as part of the replies to the @samp{qTfP}/@samp{qTsP}
40384 Although this packet is optional, and @value{GDBN} will only send it
40385 if the target replies with @samp{TracepointSource} @xref{General
40386 Query Packets}, it makes both disconnected tracing and trace files
40387 much easier to use. Otherwise the user must be careful that the
40388 tracepoints in effect while looking at trace frames are identical to
40389 the ones in effect during the trace run; even a small discrepancy
40390 could cause @samp{tdump} not to work, or a particular trace frame not
40393 @item QTDV:@var{n}:@var{value}:@var{builtin}:@var{name}
40394 @cindex define trace state variable, remote request
40395 @cindex @samp{QTDV} packet
40396 Create a new trace state variable, number @var{n}, with an initial
40397 value of @var{value}, which is a 64-bit signed integer. Both @var{n}
40398 and @var{value} are encoded as hexadecimal values. @value{GDBN} has
40399 the option of not using this packet for initial values of zero; the
40400 target should simply create the trace state variables as they are
40401 mentioned in expressions. The value @var{builtin} should be 1 (one)
40402 if the trace state variable is builtin and 0 (zero) if it is not builtin.
40403 @value{GDBN} only sets @var{builtin} to 1 if a previous @samp{qTfV} or
40404 @samp{qTsV} packet had it set. The contents of @var{name} is the
40405 hex-encoded name (without the leading @samp{$}) of the trace state
40408 @item QTFrame:@var{n}
40409 @cindex @samp{QTFrame} packet
40410 Select the @var{n}'th tracepoint frame from the buffer, and use the
40411 register and memory contents recorded there to answer subsequent
40412 request packets from @value{GDBN}.
40414 A successful reply from the stub indicates that the stub has found the
40415 requested frame. The response is a series of parts, concatenated
40416 without separators, describing the frame we selected. Each part has
40417 one of the following forms:
40421 The selected frame is number @var{n} in the trace frame buffer;
40422 @var{f} is a hexadecimal number. If @var{f} is @samp{-1}, then there
40423 was no frame matching the criteria in the request packet.
40426 The selected trace frame records a hit of tracepoint number @var{t};
40427 @var{t} is a hexadecimal number.
40431 @item QTFrame:pc:@var{addr}
40432 Like @samp{QTFrame:@var{n}}, but select the first tracepoint frame after the
40433 currently selected frame whose PC is @var{addr};
40434 @var{addr} is a hexadecimal number.
40436 @item QTFrame:tdp:@var{t}
40437 Like @samp{QTFrame:@var{n}}, but select the first tracepoint frame after the
40438 currently selected frame that is a hit of tracepoint @var{t}; @var{t}
40439 is a hexadecimal number.
40441 @item QTFrame:range:@var{start}:@var{end}
40442 Like @samp{QTFrame:@var{n}}, but select the first tracepoint frame after the
40443 currently selected frame whose PC is between @var{start} (inclusive)
40444 and @var{end} (inclusive); @var{start} and @var{end} are hexadecimal
40447 @item QTFrame:outside:@var{start}:@var{end}
40448 Like @samp{QTFrame:range:@var{start}:@var{end}}, but select the first
40449 frame @emph{outside} the given range of addresses (exclusive).
40452 @cindex @samp{qTMinFTPILen} packet
40453 This packet requests the minimum length of instruction at which a fast
40454 tracepoint (@pxref{Set Tracepoints}) may be placed. For instance, on
40455 the 32-bit x86 architecture, it is possible to use a 4-byte jump, but
40456 it depends on the target system being able to create trampolines in
40457 the first 64K of memory, which might or might not be possible for that
40458 system. So the reply to this packet will be 4 if it is able to
40465 The minimum instruction length is currently unknown.
40467 The minimum instruction length is @var{length}, where @var{length}
40468 is a hexadecimal number greater or equal to 1. A reply
40469 of 1 means that a fast tracepoint may be placed on any instruction
40470 regardless of size.
40472 An error has occurred.
40474 An empty reply indicates that the request is not supported by the stub.
40478 @cindex @samp{QTStart} packet
40479 Begin the tracepoint experiment. Begin collecting data from
40480 tracepoint hits in the trace frame buffer. This packet supports the
40481 @samp{qRelocInsn} reply (@pxref{Tracepoint Packets,,Relocate
40482 instruction reply packet}).
40485 @cindex @samp{QTStop} packet
40486 End the tracepoint experiment. Stop collecting trace frames.
40488 @item QTEnable:@var{n}:@var{addr}
40490 @cindex @samp{QTEnable} packet
40491 Enable tracepoint @var{n} at address @var{addr} in a started tracepoint
40492 experiment. If the tracepoint was previously disabled, then collection
40493 of data from it will resume.
40495 @item QTDisable:@var{n}:@var{addr}
40497 @cindex @samp{QTDisable} packet
40498 Disable tracepoint @var{n} at address @var{addr} in a started tracepoint
40499 experiment. No more data will be collected from the tracepoint unless
40500 @samp{QTEnable:@var{n}:@var{addr}} is subsequently issued.
40503 @cindex @samp{QTinit} packet
40504 Clear the table of tracepoints, and empty the trace frame buffer.
40506 @item QTro:@var{start1},@var{end1}:@var{start2},@var{end2}:@dots{}
40507 @cindex @samp{QTro} packet
40508 Establish the given ranges of memory as ``transparent''. The stub
40509 will answer requests for these ranges from memory's current contents,
40510 if they were not collected as part of the tracepoint hit.
40512 @value{GDBN} uses this to mark read-only regions of memory, like those
40513 containing program code. Since these areas never change, they should
40514 still have the same contents they did when the tracepoint was hit, so
40515 there's no reason for the stub to refuse to provide their contents.
40517 @item QTDisconnected:@var{value}
40518 @cindex @samp{QTDisconnected} packet
40519 Set the choice to what to do with the tracing run when @value{GDBN}
40520 disconnects from the target. A @var{value} of 1 directs the target to
40521 continue the tracing run, while 0 tells the target to stop tracing if
40522 @value{GDBN} is no longer in the picture.
40525 @cindex @samp{qTStatus} packet
40526 Ask the stub if there is a trace experiment running right now.
40528 The reply has the form:
40532 @item T@var{running}@r{[};@var{field}@r{]}@dots{}
40533 @var{running} is a single digit @code{1} if the trace is presently
40534 running, or @code{0} if not. It is followed by semicolon-separated
40535 optional fields that an agent may use to report additional status.
40539 If the trace is not running, the agent may report any of several
40540 explanations as one of the optional fields:
40545 No trace has been run yet.
40547 @item tstop[:@var{text}]:0
40548 The trace was stopped by a user-originated stop command. The optional
40549 @var{text} field is a user-supplied string supplied as part of the
40550 stop command (for instance, an explanation of why the trace was
40551 stopped manually). It is hex-encoded.
40554 The trace stopped because the trace buffer filled up.
40556 @item tdisconnected:0
40557 The trace stopped because @value{GDBN} disconnected from the target.
40559 @item tpasscount:@var{tpnum}
40560 The trace stopped because tracepoint @var{tpnum} exceeded its pass count.
40562 @item terror:@var{text}:@var{tpnum}
40563 The trace stopped because tracepoint @var{tpnum} had an error. The
40564 string @var{text} is available to describe the nature of the error
40565 (for instance, a divide by zero in the condition expression); it
40569 The trace stopped for some other reason.
40573 Additional optional fields supply statistical and other information.
40574 Although not required, they are extremely useful for users monitoring
40575 the progress of a trace run. If a trace has stopped, and these
40576 numbers are reported, they must reflect the state of the just-stopped
40581 @item tframes:@var{n}
40582 The number of trace frames in the buffer.
40584 @item tcreated:@var{n}
40585 The total number of trace frames created during the run. This may
40586 be larger than the trace frame count, if the buffer is circular.
40588 @item tsize:@var{n}
40589 The total size of the trace buffer, in bytes.
40591 @item tfree:@var{n}
40592 The number of bytes still unused in the buffer.
40594 @item circular:@var{n}
40595 The value of the circular trace buffer flag. @code{1} means that the
40596 trace buffer is circular and old trace frames will be discarded if
40597 necessary to make room, @code{0} means that the trace buffer is linear
40600 @item disconn:@var{n}
40601 The value of the disconnected tracing flag. @code{1} means that
40602 tracing will continue after @value{GDBN} disconnects, @code{0} means
40603 that the trace run will stop.
40607 @item qTP:@var{tp}:@var{addr}
40608 @cindex tracepoint status, remote request
40609 @cindex @samp{qTP} packet
40610 Ask the stub for the current state of tracepoint number @var{tp} at
40611 address @var{addr}.
40615 @item V@var{hits}:@var{usage}
40616 The tracepoint has been hit @var{hits} times so far during the trace
40617 run, and accounts for @var{usage} in the trace buffer. Note that
40618 @code{while-stepping} steps are not counted as separate hits, but the
40619 steps' space consumption is added into the usage number.
40623 @item qTV:@var{var}
40624 @cindex trace state variable value, remote request
40625 @cindex @samp{qTV} packet
40626 Ask the stub for the value of the trace state variable number @var{var}.
40631 The value of the variable is @var{value}. This will be the current
40632 value of the variable if the user is examining a running target, or a
40633 saved value if the variable was collected in the trace frame that the
40634 user is looking at. Note that multiple requests may result in
40635 different reply values, such as when requesting values while the
40636 program is running.
40639 The value of the variable is unknown. This would occur, for example,
40640 if the user is examining a trace frame in which the requested variable
40645 @cindex @samp{qTfP} packet
40647 @cindex @samp{qTsP} packet
40648 These packets request data about tracepoints that are being used by
40649 the target. @value{GDBN} sends @code{qTfP} to get the first piece
40650 of data, and multiple @code{qTsP} to get additional pieces. Replies
40651 to these packets generally take the form of the @code{QTDP} packets
40652 that define tracepoints. (FIXME add detailed syntax)
40655 @cindex @samp{qTfV} packet
40657 @cindex @samp{qTsV} packet
40658 These packets request data about trace state variables that are on the
40659 target. @value{GDBN} sends @code{qTfV} to get the first vari of data,
40660 and multiple @code{qTsV} to get additional variables. Replies to
40661 these packets follow the syntax of the @code{QTDV} packets that define
40662 trace state variables.
40668 @cindex @samp{qTfSTM} packet
40669 @cindex @samp{qTsSTM} packet
40670 These packets request data about static tracepoint markers that exist
40671 in the target program. @value{GDBN} sends @code{qTfSTM} to get the
40672 first piece of data, and multiple @code{qTsSTM} to get additional
40673 pieces. Replies to these packets take the following form:
40677 @item m @var{address}:@var{id}:@var{extra}
40679 @item m @var{address}:@var{id}:@var{extra},@var{address}:@var{id}:@var{extra}@dots{}
40680 a comma-separated list of markers
40682 (lower case letter @samp{L}) denotes end of list.
40684 An error occurred. The error number @var{nn} is given as hex digits.
40686 An empty reply indicates that the request is not supported by the
40690 The @var{address} is encoded in hex;
40691 @var{id} and @var{extra} are strings encoded in hex.
40693 In response to each query, the target will reply with a list of one or
40694 more markers, separated by commas. @value{GDBN} will respond to each
40695 reply with a request for more markers (using the @samp{qs} form of the
40696 query), until the target responds with @samp{l} (lower-case ell, for
40699 @item qTSTMat:@var{address}
40701 @cindex @samp{qTSTMat} packet
40702 This packets requests data about static tracepoint markers in the
40703 target program at @var{address}. Replies to this packet follow the
40704 syntax of the @samp{qTfSTM} and @code{qTsSTM} packets that list static
40705 tracepoint markers.
40707 @item QTSave:@var{filename}
40708 @cindex @samp{QTSave} packet
40709 This packet directs the target to save trace data to the file name
40710 @var{filename} in the target's filesystem. The @var{filename} is encoded
40711 as a hex string; the interpretation of the file name (relative vs
40712 absolute, wild cards, etc) is up to the target.
40714 @item qTBuffer:@var{offset},@var{len}
40715 @cindex @samp{qTBuffer} packet
40716 Return up to @var{len} bytes of the current contents of trace buffer,
40717 starting at @var{offset}. The trace buffer is treated as if it were
40718 a contiguous collection of traceframes, as per the trace file format.
40719 The reply consists as many hex-encoded bytes as the target can deliver
40720 in a packet; it is not an error to return fewer than were asked for.
40721 A reply consisting of just @code{l} indicates that no bytes are
40724 @item QTBuffer:circular:@var{value}
40725 This packet directs the target to use a circular trace buffer if
40726 @var{value} is 1, or a linear buffer if the value is 0.
40728 @item QTBuffer:size:@var{size}
40729 @anchor{QTBuffer-size}
40730 @cindex @samp{QTBuffer size} packet
40731 This packet directs the target to make the trace buffer be of size
40732 @var{size} if possible. A value of @code{-1} tells the target to
40733 use whatever size it prefers.
40735 @item QTNotes:@r{[}@var{type}:@var{text}@r{]}@r{[};@var{type}:@var{text}@r{]}@dots{}
40736 @cindex @samp{QTNotes} packet
40737 This packet adds optional textual notes to the trace run. Allowable
40738 types include @code{user}, @code{notes}, and @code{tstop}, the
40739 @var{text} fields are arbitrary strings, hex-encoded.
40743 @subsection Relocate instruction reply packet
40744 When installing fast tracepoints in memory, the target may need to
40745 relocate the instruction currently at the tracepoint address to a
40746 different address in memory. For most instructions, a simple copy is
40747 enough, but, for example, call instructions that implicitly push the
40748 return address on the stack, and relative branches or other
40749 PC-relative instructions require offset adjustment, so that the effect
40750 of executing the instruction at a different address is the same as if
40751 it had executed in the original location.
40753 In response to several of the tracepoint packets, the target may also
40754 respond with a number of intermediate @samp{qRelocInsn} request
40755 packets before the final result packet, to have @value{GDBN} handle
40756 this relocation operation. If a packet supports this mechanism, its
40757 documentation will explicitly say so. See for example the above
40758 descriptions for the @samp{QTStart} and @samp{QTDP} packets. The
40759 format of the request is:
40762 @item qRelocInsn:@var{from};@var{to}
40764 This requests @value{GDBN} to copy instruction at address @var{from}
40765 to address @var{to}, possibly adjusted so that executing the
40766 instruction at @var{to} has the same effect as executing it at
40767 @var{from}. @value{GDBN} writes the adjusted instruction to target
40768 memory starting at @var{to}.
40773 @item qRelocInsn:@var{adjusted_size}
40774 Informs the stub the relocation is complete. The @var{adjusted_size} is
40775 the length in bytes of resulting relocated instruction sequence.
40777 A badly formed request was detected, or an error was encountered while
40778 relocating the instruction.
40781 @node Host I/O Packets
40782 @section Host I/O Packets
40783 @cindex Host I/O, remote protocol
40784 @cindex file transfer, remote protocol
40786 The @dfn{Host I/O} packets allow @value{GDBN} to perform I/O
40787 operations on the far side of a remote link. For example, Host I/O is
40788 used to upload and download files to a remote target with its own
40789 filesystem. Host I/O uses the same constant values and data structure
40790 layout as the target-initiated File-I/O protocol. However, the
40791 Host I/O packets are structured differently. The target-initiated
40792 protocol relies on target memory to store parameters and buffers.
40793 Host I/O requests are initiated by @value{GDBN}, and the
40794 target's memory is not involved. @xref{File-I/O Remote Protocol
40795 Extension}, for more details on the target-initiated protocol.
40797 The Host I/O request packets all encode a single operation along with
40798 its arguments. They have this format:
40802 @item vFile:@var{operation}: @var{parameter}@dots{}
40803 @var{operation} is the name of the particular request; the target
40804 should compare the entire packet name up to the second colon when checking
40805 for a supported operation. The format of @var{parameter} depends on
40806 the operation. Numbers are always passed in hexadecimal. Negative
40807 numbers have an explicit minus sign (i.e.@: two's complement is not
40808 used). Strings (e.g.@: filenames) are encoded as a series of
40809 hexadecimal bytes. The last argument to a system call may be a
40810 buffer of escaped binary data (@pxref{Binary Data}).
40814 The valid responses to Host I/O packets are:
40818 @item F @var{result} [, @var{errno}] [; @var{attachment}]
40819 @var{result} is the integer value returned by this operation, usually
40820 non-negative for success and -1 for errors. If an error has occured,
40821 @var{errno} will be included in the result specifying a
40822 value defined by the File-I/O protocol (@pxref{Errno Values}). For
40823 operations which return data, @var{attachment} supplies the data as a
40824 binary buffer. Binary buffers in response packets are escaped in the
40825 normal way (@pxref{Binary Data}). See the individual packet
40826 documentation for the interpretation of @var{result} and
40830 An empty response indicates that this operation is not recognized.
40834 These are the supported Host I/O operations:
40837 @item vFile:open: @var{filename}, @var{flags}, @var{mode}
40838 Open a file at @var{filename} and return a file descriptor for it, or
40839 return -1 if an error occurs. The @var{filename} is a string,
40840 @var{flags} is an integer indicating a mask of open flags
40841 (@pxref{Open Flags}), and @var{mode} is an integer indicating a mask
40842 of mode bits to use if the file is created (@pxref{mode_t Values}).
40843 @xref{open}, for details of the open flags and mode values.
40845 @item vFile:close: @var{fd}
40846 Close the open file corresponding to @var{fd} and return 0, or
40847 -1 if an error occurs.
40849 @item vFile:pread: @var{fd}, @var{count}, @var{offset}
40850 Read data from the open file corresponding to @var{fd}. Up to
40851 @var{count} bytes will be read from the file, starting at @var{offset}
40852 relative to the start of the file. The target may read fewer bytes;
40853 common reasons include packet size limits and an end-of-file
40854 condition. The number of bytes read is returned. Zero should only be
40855 returned for a successful read at the end of the file, or if
40856 @var{count} was zero.
40858 The data read should be returned as a binary attachment on success.
40859 If zero bytes were read, the response should include an empty binary
40860 attachment (i.e.@: a trailing semicolon). The return value is the
40861 number of target bytes read; the binary attachment may be longer if
40862 some characters were escaped.
40864 @item vFile:pwrite: @var{fd}, @var{offset}, @var{data}
40865 Write @var{data} (a binary buffer) to the open file corresponding
40866 to @var{fd}. Start the write at @var{offset} from the start of the
40867 file. Unlike many @code{write} system calls, there is no
40868 separate @var{count} argument; the length of @var{data} in the
40869 packet is used. @samp{vFile:write} returns the number of bytes written,
40870 which may be shorter than the length of @var{data}, or -1 if an
40873 @item vFile:fstat: @var{fd}
40874 Get information about the open file corresponding to @var{fd}.
40875 On success the information is returned as a binary attachment
40876 and the return value is the size of this attachment in bytes.
40877 If an error occurs the return value is -1. The format of the
40878 returned binary attachment is as described in @ref{struct stat}.
40880 @item vFile:unlink: @var{filename}
40881 Delete the file at @var{filename} on the target. Return 0,
40882 or -1 if an error occurs. The @var{filename} is a string.
40884 @item vFile:readlink: @var{filename}
40885 Read value of symbolic link @var{filename} on the target. Return
40886 the number of bytes read, or -1 if an error occurs.
40888 The data read should be returned as a binary attachment on success.
40889 If zero bytes were read, the response should include an empty binary
40890 attachment (i.e.@: a trailing semicolon). The return value is the
40891 number of target bytes read; the binary attachment may be longer if
40892 some characters were escaped.
40894 @item vFile:setfs: @var{pid}
40895 Select the filesystem on which @code{vFile} operations with
40896 @var{filename} arguments will operate. This is required for
40897 @value{GDBN} to be able to access files on remote targets where
40898 the remote stub does not share a common filesystem with the
40901 If @var{pid} is nonzero, select the filesystem as seen by process
40902 @var{pid}. If @var{pid} is zero, select the filesystem as seen by
40903 the remote stub. Return 0 on success, or -1 if an error occurs.
40904 If @code{vFile:setfs:} indicates success, the selected filesystem
40905 remains selected until the next successful @code{vFile:setfs:}
40911 @section Interrupts
40912 @cindex interrupts (remote protocol)
40913 @anchor{interrupting remote targets}
40915 In all-stop mode, when a program on the remote target is running,
40916 @value{GDBN} may attempt to interrupt it by sending a @samp{Ctrl-C},
40917 @code{BREAK} or a @code{BREAK} followed by @code{g}, control of which
40918 is specified via @value{GDBN}'s @samp{interrupt-sequence}.
40920 The precise meaning of @code{BREAK} is defined by the transport
40921 mechanism and may, in fact, be undefined. @value{GDBN} does not
40922 currently define a @code{BREAK} mechanism for any of the network
40923 interfaces except for TCP, in which case @value{GDBN} sends the
40924 @code{telnet} BREAK sequence.
40926 @samp{Ctrl-C}, on the other hand, is defined and implemented for all
40927 transport mechanisms. It is represented by sending the single byte
40928 @code{0x03} without any of the usual packet overhead described in
40929 the Overview section (@pxref{Overview}). When a @code{0x03} byte is
40930 transmitted as part of a packet, it is considered to be packet data
40931 and does @emph{not} represent an interrupt. E.g., an @samp{X} packet
40932 (@pxref{X packet}), used for binary downloads, may include an unescaped
40933 @code{0x03} as part of its packet.
40935 @code{BREAK} followed by @code{g} is also known as Magic SysRq g.
40936 When Linux kernel receives this sequence from serial port,
40937 it stops execution and connects to gdb.
40939 In non-stop mode, because packet resumptions are asynchronous
40940 (@pxref{vCont packet}), @value{GDBN} is always free to send a remote
40941 command to the remote stub, even when the target is running. For that
40942 reason, @value{GDBN} instead sends a regular packet (@pxref{vCtrlC
40943 packet}) with the usual packet framing instead of the single byte
40946 Stubs are not required to recognize these interrupt mechanisms and the
40947 precise meaning associated with receipt of the interrupt is
40948 implementation defined. If the target supports debugging of multiple
40949 threads and/or processes, it should attempt to interrupt all
40950 currently-executing threads and processes.
40951 If the stub is successful at interrupting the
40952 running program, it should send one of the stop
40953 reply packets (@pxref{Stop Reply Packets}) to @value{GDBN} as a result
40954 of successfully stopping the program in all-stop mode, and a stop reply
40955 for each stopped thread in non-stop mode.
40956 Interrupts received while the
40957 program is stopped are queued and the program will be interrupted when
40958 it is resumed next time.
40960 @node Notification Packets
40961 @section Notification Packets
40962 @cindex notification packets
40963 @cindex packets, notification
40965 The @value{GDBN} remote serial protocol includes @dfn{notifications},
40966 packets that require no acknowledgment. Both the GDB and the stub
40967 may send notifications (although the only notifications defined at
40968 present are sent by the stub). Notifications carry information
40969 without incurring the round-trip latency of an acknowledgment, and so
40970 are useful for low-impact communications where occasional packet loss
40973 A notification packet has the form @samp{% @var{data} #
40974 @var{checksum}}, where @var{data} is the content of the notification,
40975 and @var{checksum} is a checksum of @var{data}, computed and formatted
40976 as for ordinary @value{GDBN} packets. A notification's @var{data}
40977 never contains @samp{$}, @samp{%} or @samp{#} characters. Upon
40978 receiving a notification, the recipient sends no @samp{+} or @samp{-}
40979 to acknowledge the notification's receipt or to report its corruption.
40981 Every notification's @var{data} begins with a name, which contains no
40982 colon characters, followed by a colon character.
40984 Recipients should silently ignore corrupted notifications and
40985 notifications they do not understand. Recipients should restart
40986 timeout periods on receipt of a well-formed notification, whether or
40987 not they understand it.
40989 Senders should only send the notifications described here when this
40990 protocol description specifies that they are permitted. In the
40991 future, we may extend the protocol to permit existing notifications in
40992 new contexts; this rule helps older senders avoid confusing newer
40995 (Older versions of @value{GDBN} ignore bytes received until they see
40996 the @samp{$} byte that begins an ordinary packet, so new stubs may
40997 transmit notifications without fear of confusing older clients. There
40998 are no notifications defined for @value{GDBN} to send at the moment, but we
40999 assume that most older stubs would ignore them, as well.)
41001 Each notification is comprised of three parts:
41003 @item @var{name}:@var{event}
41004 The notification packet is sent by the side that initiates the
41005 exchange (currently, only the stub does that), with @var{event}
41006 carrying the specific information about the notification, and
41007 @var{name} specifying the name of the notification.
41009 The acknowledge sent by the other side, usually @value{GDBN}, to
41010 acknowledge the exchange and request the event.
41013 The purpose of an asynchronous notification mechanism is to report to
41014 @value{GDBN} that something interesting happened in the remote stub.
41016 The remote stub may send notification @var{name}:@var{event}
41017 at any time, but @value{GDBN} acknowledges the notification when
41018 appropriate. The notification event is pending before @value{GDBN}
41019 acknowledges. Only one notification at a time may be pending; if
41020 additional events occur before @value{GDBN} has acknowledged the
41021 previous notification, they must be queued by the stub for later
41022 synchronous transmission in response to @var{ack} packets from
41023 @value{GDBN}. Because the notification mechanism is unreliable,
41024 the stub is permitted to resend a notification if it believes
41025 @value{GDBN} may not have received it.
41027 Specifically, notifications may appear when @value{GDBN} is not
41028 otherwise reading input from the stub, or when @value{GDBN} is
41029 expecting to read a normal synchronous response or a
41030 @samp{+}/@samp{-} acknowledgment to a packet it has sent.
41031 Notification packets are distinct from any other communication from
41032 the stub so there is no ambiguity.
41034 After receiving a notification, @value{GDBN} shall acknowledge it by
41035 sending a @var{ack} packet as a regular, synchronous request to the
41036 stub. Such acknowledgment is not required to happen immediately, as
41037 @value{GDBN} is permitted to send other, unrelated packets to the
41038 stub first, which the stub should process normally.
41040 Upon receiving a @var{ack} packet, if the stub has other queued
41041 events to report to @value{GDBN}, it shall respond by sending a
41042 normal @var{event}. @value{GDBN} shall then send another @var{ack}
41043 packet to solicit further responses; again, it is permitted to send
41044 other, unrelated packets as well which the stub should process
41047 If the stub receives a @var{ack} packet and there are no additional
41048 @var{event} to report, the stub shall return an @samp{OK} response.
41049 At this point, @value{GDBN} has finished processing a notification
41050 and the stub has completed sending any queued events. @value{GDBN}
41051 won't accept any new notifications until the final @samp{OK} is
41052 received . If further notification events occur, the stub shall send
41053 a new notification, @value{GDBN} shall accept the notification, and
41054 the process shall be repeated.
41056 The process of asynchronous notification can be illustrated by the
41059 <- @code{%Stop:T0505:98e7ffbf;04:4ce6ffbf;08:b1b6e54c;thread:p7526.7526;core:0;}
41062 <- @code{T0505:68f37db7;04:40f37db7;08:63850408;thread:p7526.7528;core:0;}
41064 <- @code{T0505:68e3fdb6;04:40e3fdb6;08:63850408;thread:p7526.7529;core:0;}
41069 The following notifications are defined:
41070 @multitable @columnfractions 0.12 0.12 0.38 0.38
41079 @tab @var{reply}. The @var{reply} has the form of a stop reply, as
41080 described in @ref{Stop Reply Packets}. Refer to @ref{Remote Non-Stop},
41081 for information on how these notifications are acknowledged by
41083 @tab Report an asynchronous stop event in non-stop mode.
41087 @node Remote Non-Stop
41088 @section Remote Protocol Support for Non-Stop Mode
41090 @value{GDBN}'s remote protocol supports non-stop debugging of
41091 multi-threaded programs, as described in @ref{Non-Stop Mode}. If the stub
41092 supports non-stop mode, it should report that to @value{GDBN} by including
41093 @samp{QNonStop+} in its @samp{qSupported} response (@pxref{qSupported}).
41095 @value{GDBN} typically sends a @samp{QNonStop} packet only when
41096 establishing a new connection with the stub. Entering non-stop mode
41097 does not alter the state of any currently-running threads, but targets
41098 must stop all threads in any already-attached processes when entering
41099 all-stop mode. @value{GDBN} uses the @samp{?} packet as necessary to
41100 probe the target state after a mode change.
41102 In non-stop mode, when an attached process encounters an event that
41103 would otherwise be reported with a stop reply, it uses the
41104 asynchronous notification mechanism (@pxref{Notification Packets}) to
41105 inform @value{GDBN}. In contrast to all-stop mode, where all threads
41106 in all processes are stopped when a stop reply is sent, in non-stop
41107 mode only the thread reporting the stop event is stopped. That is,
41108 when reporting a @samp{S} or @samp{T} response to indicate completion
41109 of a step operation, hitting a breakpoint, or a fault, only the
41110 affected thread is stopped; any other still-running threads continue
41111 to run. When reporting a @samp{W} or @samp{X} response, all running
41112 threads belonging to other attached processes continue to run.
41114 In non-stop mode, the target shall respond to the @samp{?} packet as
41115 follows. First, any incomplete stop reply notification/@samp{vStopped}
41116 sequence in progress is abandoned. The target must begin a new
41117 sequence reporting stop events for all stopped threads, whether or not
41118 it has previously reported those events to @value{GDBN}. The first
41119 stop reply is sent as a synchronous reply to the @samp{?} packet, and
41120 subsequent stop replies are sent as responses to @samp{vStopped} packets
41121 using the mechanism described above. The target must not send
41122 asynchronous stop reply notifications until the sequence is complete.
41123 If all threads are running when the target receives the @samp{?} packet,
41124 or if the target is not attached to any process, it shall respond
41127 If the stub supports non-stop mode, it should also support the
41128 @samp{swbreak} stop reason if software breakpoints are supported, and
41129 the @samp{hwbreak} stop reason if hardware breakpoints are supported
41130 (@pxref{swbreak stop reason}). This is because given the asynchronous
41131 nature of non-stop mode, between the time a thread hits a breakpoint
41132 and the time the event is finally processed by @value{GDBN}, the
41133 breakpoint may have already been removed from the target. Due to
41134 this, @value{GDBN} needs to be able to tell whether a trap stop was
41135 caused by a delayed breakpoint event, which should be ignored, as
41136 opposed to a random trap signal, which should be reported to the user.
41137 Note the @samp{swbreak} feature implies that the target is responsible
41138 for adjusting the PC when a software breakpoint triggers, if
41139 necessary, such as on the x86 architecture.
41141 @node Packet Acknowledgment
41142 @section Packet Acknowledgment
41144 @cindex acknowledgment, for @value{GDBN} remote
41145 @cindex packet acknowledgment, for @value{GDBN} remote
41146 By default, when either the host or the target machine receives a packet,
41147 the first response expected is an acknowledgment: either @samp{+} (to indicate
41148 the package was received correctly) or @samp{-} (to request retransmission).
41149 This mechanism allows the @value{GDBN} remote protocol to operate over
41150 unreliable transport mechanisms, such as a serial line.
41152 In cases where the transport mechanism is itself reliable (such as a pipe or
41153 TCP connection), the @samp{+}/@samp{-} acknowledgments are redundant.
41154 It may be desirable to disable them in that case to reduce communication
41155 overhead, or for other reasons. This can be accomplished by means of the
41156 @samp{QStartNoAckMode} packet; @pxref{QStartNoAckMode}.
41158 When in no-acknowledgment mode, neither the stub nor @value{GDBN} shall send or
41159 expect @samp{+}/@samp{-} protocol acknowledgments. The packet
41160 and response format still includes the normal checksum, as described in
41161 @ref{Overview}, but the checksum may be ignored by the receiver.
41163 If the stub supports @samp{QStartNoAckMode} and prefers to operate in
41164 no-acknowledgment mode, it should report that to @value{GDBN}
41165 by including @samp{QStartNoAckMode+} in its response to @samp{qSupported};
41166 @pxref{qSupported}.
41167 If @value{GDBN} also supports @samp{QStartNoAckMode} and it has not been
41168 disabled via the @code{set remote noack-packet off} command
41169 (@pxref{Remote Configuration}),
41170 @value{GDBN} may then send a @samp{QStartNoAckMode} packet to the stub.
41171 Only then may the stub actually turn off packet acknowledgments.
41172 @value{GDBN} sends a final @samp{+} acknowledgment of the stub's @samp{OK}
41173 response, which can be safely ignored by the stub.
41175 Note that @code{set remote noack-packet} command only affects negotiation
41176 between @value{GDBN} and the stub when subsequent connections are made;
41177 it does not affect the protocol acknowledgment state for any current
41179 Since @samp{+}/@samp{-} acknowledgments are enabled by default when a
41180 new connection is established,
41181 there is also no protocol request to re-enable the acknowledgments
41182 for the current connection, once disabled.
41187 Example sequence of a target being re-started. Notice how the restart
41188 does not get any direct output:
41193 @emph{target restarts}
41196 <- @code{T001:1234123412341234}
41200 Example sequence of a target being stepped by a single instruction:
41203 -> @code{G1445@dots{}}
41208 <- @code{T001:1234123412341234}
41212 <- @code{1455@dots{}}
41216 @node File-I/O Remote Protocol Extension
41217 @section File-I/O Remote Protocol Extension
41218 @cindex File-I/O remote protocol extension
41221 * File-I/O Overview::
41222 * Protocol Basics::
41223 * The F Request Packet::
41224 * The F Reply Packet::
41225 * The Ctrl-C Message::
41227 * List of Supported Calls::
41228 * Protocol-specific Representation of Datatypes::
41230 * File-I/O Examples::
41233 @node File-I/O Overview
41234 @subsection File-I/O Overview
41235 @cindex file-i/o overview
41237 The @dfn{File I/O remote protocol extension} (short: File-I/O) allows the
41238 target to use the host's file system and console I/O to perform various
41239 system calls. System calls on the target system are translated into a
41240 remote protocol packet to the host system, which then performs the needed
41241 actions and returns a response packet to the target system.
41242 This simulates file system operations even on targets that lack file systems.
41244 The protocol is defined to be independent of both the host and target systems.
41245 It uses its own internal representation of datatypes and values. Both
41246 @value{GDBN} and the target's @value{GDBN} stub are responsible for
41247 translating the system-dependent value representations into the internal
41248 protocol representations when data is transmitted.
41250 The communication is synchronous. A system call is possible only when
41251 @value{GDBN} is waiting for a response from the @samp{C}, @samp{c}, @samp{S}
41252 or @samp{s} packets. While @value{GDBN} handles the request for a system call,
41253 the target is stopped to allow deterministic access to the target's
41254 memory. Therefore File-I/O is not interruptible by target signals. On
41255 the other hand, it is possible to interrupt File-I/O by a user interrupt
41256 (@samp{Ctrl-C}) within @value{GDBN}.
41258 The target's request to perform a host system call does not finish
41259 the latest @samp{C}, @samp{c}, @samp{S} or @samp{s} action. That means,
41260 after finishing the system call, the target returns to continuing the
41261 previous activity (continue, step). No additional continue or step
41262 request from @value{GDBN} is required.
41265 (@value{GDBP}) continue
41266 <- target requests 'system call X'
41267 target is stopped, @value{GDBN} executes system call
41268 -> @value{GDBN} returns result
41269 ... target continues, @value{GDBN} returns to wait for the target
41270 <- target hits breakpoint and sends a Txx packet
41273 The protocol only supports I/O on the console and to regular files on
41274 the host file system. Character or block special devices, pipes,
41275 named pipes, sockets or any other communication method on the host
41276 system are not supported by this protocol.
41278 File I/O is not supported in non-stop mode.
41280 @node Protocol Basics
41281 @subsection Protocol Basics
41282 @cindex protocol basics, file-i/o
41284 The File-I/O protocol uses the @code{F} packet as the request as well
41285 as reply packet. Since a File-I/O system call can only occur when
41286 @value{GDBN} is waiting for a response from the continuing or stepping target,
41287 the File-I/O request is a reply that @value{GDBN} has to expect as a result
41288 of a previous @samp{C}, @samp{c}, @samp{S} or @samp{s} packet.
41289 This @code{F} packet contains all information needed to allow @value{GDBN}
41290 to call the appropriate host system call:
41294 A unique identifier for the requested system call.
41297 All parameters to the system call. Pointers are given as addresses
41298 in the target memory address space. Pointers to strings are given as
41299 pointer/length pair. Numerical values are given as they are.
41300 Numerical control flags are given in a protocol-specific representation.
41304 At this point, @value{GDBN} has to perform the following actions.
41308 If the parameters include pointer values to data needed as input to a
41309 system call, @value{GDBN} requests this data from the target with a
41310 standard @code{m} packet request. This additional communication has to be
41311 expected by the target implementation and is handled as any other @code{m}
41315 @value{GDBN} translates all value from protocol representation to host
41316 representation as needed. Datatypes are coerced into the host types.
41319 @value{GDBN} calls the system call.
41322 It then coerces datatypes back to protocol representation.
41325 If the system call is expected to return data in buffer space specified
41326 by pointer parameters to the call, the data is transmitted to the
41327 target using a @code{M} or @code{X} packet. This packet has to be expected
41328 by the target implementation and is handled as any other @code{M} or @code{X}
41333 Eventually @value{GDBN} replies with another @code{F} packet which contains all
41334 necessary information for the target to continue. This at least contains
41341 @code{errno}, if has been changed by the system call.
41348 After having done the needed type and value coercion, the target continues
41349 the latest continue or step action.
41351 @node The F Request Packet
41352 @subsection The @code{F} Request Packet
41353 @cindex file-i/o request packet
41354 @cindex @code{F} request packet
41356 The @code{F} request packet has the following format:
41359 @item F@var{call-id},@var{parameter@dots{}}
41361 @var{call-id} is the identifier to indicate the host system call to be called.
41362 This is just the name of the function.
41364 @var{parameter@dots{}} are the parameters to the system call.
41365 Parameters are hexadecimal integer values, either the actual values in case
41366 of scalar datatypes, pointers to target buffer space in case of compound
41367 datatypes and unspecified memory areas, or pointer/length pairs in case
41368 of string parameters. These are appended to the @var{call-id} as a
41369 comma-delimited list. All values are transmitted in ASCII
41370 string representation, pointer/length pairs separated by a slash.
41376 @node The F Reply Packet
41377 @subsection The @code{F} Reply Packet
41378 @cindex file-i/o reply packet
41379 @cindex @code{F} reply packet
41381 The @code{F} reply packet has the following format:
41385 @item F@var{retcode},@var{errno},@var{Ctrl-C flag};@var{call-specific attachment}
41387 @var{retcode} is the return code of the system call as hexadecimal value.
41389 @var{errno} is the @code{errno} set by the call, in protocol-specific
41391 This parameter can be omitted if the call was successful.
41393 @var{Ctrl-C flag} is only sent if the user requested a break. In this
41394 case, @var{errno} must be sent as well, even if the call was successful.
41395 The @var{Ctrl-C flag} itself consists of the character @samp{C}:
41402 or, if the call was interrupted before the host call has been performed:
41409 assuming 4 is the protocol-specific representation of @code{EINTR}.
41414 @node The Ctrl-C Message
41415 @subsection The @samp{Ctrl-C} Message
41416 @cindex ctrl-c message, in file-i/o protocol
41418 If the @samp{Ctrl-C} flag is set in the @value{GDBN}
41419 reply packet (@pxref{The F Reply Packet}),
41420 the target should behave as if it had
41421 gotten a break message. The meaning for the target is ``system call
41422 interrupted by @code{SIGINT}''. Consequentially, the target should actually stop
41423 (as with a break message) and return to @value{GDBN} with a @code{T02}
41426 It's important for the target to know in which
41427 state the system call was interrupted. There are two possible cases:
41431 The system call hasn't been performed on the host yet.
41434 The system call on the host has been finished.
41438 These two states can be distinguished by the target by the value of the
41439 returned @code{errno}. If it's the protocol representation of @code{EINTR}, the system
41440 call hasn't been performed. This is equivalent to the @code{EINTR} handling
41441 on POSIX systems. In any other case, the target may presume that the
41442 system call has been finished --- successfully or not --- and should behave
41443 as if the break message arrived right after the system call.
41445 @value{GDBN} must behave reliably. If the system call has not been called
41446 yet, @value{GDBN} may send the @code{F} reply immediately, setting @code{EINTR} as
41447 @code{errno} in the packet. If the system call on the host has been finished
41448 before the user requests a break, the full action must be finished by
41449 @value{GDBN}. This requires sending @code{M} or @code{X} packets as necessary.
41450 The @code{F} packet may only be sent when either nothing has happened
41451 or the full action has been completed.
41454 @subsection Console I/O
41455 @cindex console i/o as part of file-i/o
41457 By default and if not explicitly closed by the target system, the file
41458 descriptors 0, 1 and 2 are connected to the @value{GDBN} console. Output
41459 on the @value{GDBN} console is handled as any other file output operation
41460 (@code{write(1, @dots{})} or @code{write(2, @dots{})}). Console input is handled
41461 by @value{GDBN} so that after the target read request from file descriptor
41462 0 all following typing is buffered until either one of the following
41467 The user types @kbd{Ctrl-c}. The behaviour is as explained above, and the
41469 system call is treated as finished.
41472 The user presses @key{RET}. This is treated as end of input with a trailing
41476 The user types @kbd{Ctrl-d}. This is treated as end of input. No trailing
41477 character (neither newline nor @samp{Ctrl-D}) is appended to the input.
41481 If the user has typed more characters than fit in the buffer given to
41482 the @code{read} call, the trailing characters are buffered in @value{GDBN} until
41483 either another @code{read(0, @dots{})} is requested by the target, or debugging
41484 is stopped at the user's request.
41487 @node List of Supported Calls
41488 @subsection List of Supported Calls
41489 @cindex list of supported file-i/o calls
41506 @unnumberedsubsubsec open
41507 @cindex open, file-i/o system call
41512 int open(const char *pathname, int flags);
41513 int open(const char *pathname, int flags, mode_t mode);
41517 @samp{Fopen,@var{pathptr}/@var{len},@var{flags},@var{mode}}
41520 @var{flags} is the bitwise @code{OR} of the following values:
41524 If the file does not exist it will be created. The host
41525 rules apply as far as file ownership and time stamps
41529 When used with @code{O_CREAT}, if the file already exists it is
41530 an error and open() fails.
41533 If the file already exists and the open mode allows
41534 writing (@code{O_RDWR} or @code{O_WRONLY} is given) it will be
41535 truncated to zero length.
41538 The file is opened in append mode.
41541 The file is opened for reading only.
41544 The file is opened for writing only.
41547 The file is opened for reading and writing.
41551 Other bits are silently ignored.
41555 @var{mode} is the bitwise @code{OR} of the following values:
41559 User has read permission.
41562 User has write permission.
41565 Group has read permission.
41568 Group has write permission.
41571 Others have read permission.
41574 Others have write permission.
41578 Other bits are silently ignored.
41581 @item Return value:
41582 @code{open} returns the new file descriptor or -1 if an error
41589 @var{pathname} already exists and @code{O_CREAT} and @code{O_EXCL} were used.
41592 @var{pathname} refers to a directory.
41595 The requested access is not allowed.
41598 @var{pathname} was too long.
41601 A directory component in @var{pathname} does not exist.
41604 @var{pathname} refers to a device, pipe, named pipe or socket.
41607 @var{pathname} refers to a file on a read-only filesystem and
41608 write access was requested.
41611 @var{pathname} is an invalid pointer value.
41614 No space on device to create the file.
41617 The process already has the maximum number of files open.
41620 The limit on the total number of files open on the system
41624 The call was interrupted by the user.
41630 @unnumberedsubsubsec close
41631 @cindex close, file-i/o system call
41640 @samp{Fclose,@var{fd}}
41642 @item Return value:
41643 @code{close} returns zero on success, or -1 if an error occurred.
41649 @var{fd} isn't a valid open file descriptor.
41652 The call was interrupted by the user.
41658 @unnumberedsubsubsec read
41659 @cindex read, file-i/o system call
41664 int read(int fd, void *buf, unsigned int count);
41668 @samp{Fread,@var{fd},@var{bufptr},@var{count}}
41670 @item Return value:
41671 On success, the number of bytes read is returned.
41672 Zero indicates end of file. If count is zero, read
41673 returns zero as well. On error, -1 is returned.
41679 @var{fd} is not a valid file descriptor or is not open for
41683 @var{bufptr} is an invalid pointer value.
41686 The call was interrupted by the user.
41692 @unnumberedsubsubsec write
41693 @cindex write, file-i/o system call
41698 int write(int fd, const void *buf, unsigned int count);
41702 @samp{Fwrite,@var{fd},@var{bufptr},@var{count}}
41704 @item Return value:
41705 On success, the number of bytes written are returned.
41706 Zero indicates nothing was written. On error, -1
41713 @var{fd} is not a valid file descriptor or is not open for
41717 @var{bufptr} is an invalid pointer value.
41720 An attempt was made to write a file that exceeds the
41721 host-specific maximum file size allowed.
41724 No space on device to write the data.
41727 The call was interrupted by the user.
41733 @unnumberedsubsubsec lseek
41734 @cindex lseek, file-i/o system call
41739 long lseek (int fd, long offset, int flag);
41743 @samp{Flseek,@var{fd},@var{offset},@var{flag}}
41745 @var{flag} is one of:
41749 The offset is set to @var{offset} bytes.
41752 The offset is set to its current location plus @var{offset}
41756 The offset is set to the size of the file plus @var{offset}
41760 @item Return value:
41761 On success, the resulting unsigned offset in bytes from
41762 the beginning of the file is returned. Otherwise, a
41763 value of -1 is returned.
41769 @var{fd} is not a valid open file descriptor.
41772 @var{fd} is associated with the @value{GDBN} console.
41775 @var{flag} is not a proper value.
41778 The call was interrupted by the user.
41784 @unnumberedsubsubsec rename
41785 @cindex rename, file-i/o system call
41790 int rename(const char *oldpath, const char *newpath);
41794 @samp{Frename,@var{oldpathptr}/@var{len},@var{newpathptr}/@var{len}}
41796 @item Return value:
41797 On success, zero is returned. On error, -1 is returned.
41803 @var{newpath} is an existing directory, but @var{oldpath} is not a
41807 @var{newpath} is a non-empty directory.
41810 @var{oldpath} or @var{newpath} is a directory that is in use by some
41814 An attempt was made to make a directory a subdirectory
41818 A component used as a directory in @var{oldpath} or new
41819 path is not a directory. Or @var{oldpath} is a directory
41820 and @var{newpath} exists but is not a directory.
41823 @var{oldpathptr} or @var{newpathptr} are invalid pointer values.
41826 No access to the file or the path of the file.
41830 @var{oldpath} or @var{newpath} was too long.
41833 A directory component in @var{oldpath} or @var{newpath} does not exist.
41836 The file is on a read-only filesystem.
41839 The device containing the file has no room for the new
41843 The call was interrupted by the user.
41849 @unnumberedsubsubsec unlink
41850 @cindex unlink, file-i/o system call
41855 int unlink(const char *pathname);
41859 @samp{Funlink,@var{pathnameptr}/@var{len}}
41861 @item Return value:
41862 On success, zero is returned. On error, -1 is returned.
41868 No access to the file or the path of the file.
41871 The system does not allow unlinking of directories.
41874 The file @var{pathname} cannot be unlinked because it's
41875 being used by another process.
41878 @var{pathnameptr} is an invalid pointer value.
41881 @var{pathname} was too long.
41884 A directory component in @var{pathname} does not exist.
41887 A component of the path is not a directory.
41890 The file is on a read-only filesystem.
41893 The call was interrupted by the user.
41899 @unnumberedsubsubsec stat/fstat
41900 @cindex fstat, file-i/o system call
41901 @cindex stat, file-i/o system call
41906 int stat(const char *pathname, struct stat *buf);
41907 int fstat(int fd, struct stat *buf);
41911 @samp{Fstat,@var{pathnameptr}/@var{len},@var{bufptr}}@*
41912 @samp{Ffstat,@var{fd},@var{bufptr}}
41914 @item Return value:
41915 On success, zero is returned. On error, -1 is returned.
41921 @var{fd} is not a valid open file.
41924 A directory component in @var{pathname} does not exist or the
41925 path is an empty string.
41928 A component of the path is not a directory.
41931 @var{pathnameptr} is an invalid pointer value.
41934 No access to the file or the path of the file.
41937 @var{pathname} was too long.
41940 The call was interrupted by the user.
41946 @unnumberedsubsubsec gettimeofday
41947 @cindex gettimeofday, file-i/o system call
41952 int gettimeofday(struct timeval *tv, void *tz);
41956 @samp{Fgettimeofday,@var{tvptr},@var{tzptr}}
41958 @item Return value:
41959 On success, 0 is returned, -1 otherwise.
41965 @var{tz} is a non-NULL pointer.
41968 @var{tvptr} and/or @var{tzptr} is an invalid pointer value.
41974 @unnumberedsubsubsec isatty
41975 @cindex isatty, file-i/o system call
41980 int isatty(int fd);
41984 @samp{Fisatty,@var{fd}}
41986 @item Return value:
41987 Returns 1 if @var{fd} refers to the @value{GDBN} console, 0 otherwise.
41993 The call was interrupted by the user.
41998 Note that the @code{isatty} call is treated as a special case: it returns
41999 1 to the target if the file descriptor is attached
42000 to the @value{GDBN} console, 0 otherwise. Implementing through system calls
42001 would require implementing @code{ioctl} and would be more complex than
42006 @unnumberedsubsubsec system
42007 @cindex system, file-i/o system call
42012 int system(const char *command);
42016 @samp{Fsystem,@var{commandptr}/@var{len}}
42018 @item Return value:
42019 If @var{len} is zero, the return value indicates whether a shell is
42020 available. A zero return value indicates a shell is not available.
42021 For non-zero @var{len}, the value returned is -1 on error and the
42022 return status of the command otherwise. Only the exit status of the
42023 command is returned, which is extracted from the host's @code{system}
42024 return value by calling @code{WEXITSTATUS(retval)}. In case
42025 @file{/bin/sh} could not be executed, 127 is returned.
42031 The call was interrupted by the user.
42036 @value{GDBN} takes over the full task of calling the necessary host calls
42037 to perform the @code{system} call. The return value of @code{system} on
42038 the host is simplified before it's returned
42039 to the target. Any termination signal information from the child process
42040 is discarded, and the return value consists
42041 entirely of the exit status of the called command.
42043 Due to security concerns, the @code{system} call is by default refused
42044 by @value{GDBN}. The user has to allow this call explicitly with the
42045 @code{set remote system-call-allowed 1} command.
42048 @item set remote system-call-allowed
42049 @kindex set remote system-call-allowed
42050 Control whether to allow the @code{system} calls in the File I/O
42051 protocol for the remote target. The default is zero (disabled).
42053 @item show remote system-call-allowed
42054 @kindex show remote system-call-allowed
42055 Show whether the @code{system} calls are allowed in the File I/O
42059 @node Protocol-specific Representation of Datatypes
42060 @subsection Protocol-specific Representation of Datatypes
42061 @cindex protocol-specific representation of datatypes, in file-i/o protocol
42064 * Integral Datatypes::
42066 * Memory Transfer::
42071 @node Integral Datatypes
42072 @unnumberedsubsubsec Integral Datatypes
42073 @cindex integral datatypes, in file-i/o protocol
42075 The integral datatypes used in the system calls are @code{int},
42076 @code{unsigned int}, @code{long}, @code{unsigned long},
42077 @code{mode_t}, and @code{time_t}.
42079 @code{int}, @code{unsigned int}, @code{mode_t} and @code{time_t} are
42080 implemented as 32 bit values in this protocol.
42082 @code{long} and @code{unsigned long} are implemented as 64 bit types.
42084 @xref{Limits}, for corresponding MIN and MAX values (similar to those
42085 in @file{limits.h}) to allow range checking on host and target.
42087 @code{time_t} datatypes are defined as seconds since the Epoch.
42089 All integral datatypes transferred as part of a memory read or write of a
42090 structured datatype e.g.@: a @code{struct stat} have to be given in big endian
42093 @node Pointer Values
42094 @unnumberedsubsubsec Pointer Values
42095 @cindex pointer values, in file-i/o protocol
42097 Pointers to target data are transmitted as they are. An exception
42098 is made for pointers to buffers for which the length isn't
42099 transmitted as part of the function call, namely strings. Strings
42100 are transmitted as a pointer/length pair, both as hex values, e.g.@:
42107 which is a pointer to data of length 18 bytes at position 0x1aaf.
42108 The length is defined as the full string length in bytes, including
42109 the trailing null byte. For example, the string @code{"hello world"}
42110 at address 0x123456 is transmitted as
42116 @node Memory Transfer
42117 @unnumberedsubsubsec Memory Transfer
42118 @cindex memory transfer, in file-i/o protocol
42120 Structured data which is transferred using a memory read or write (for
42121 example, a @code{struct stat}) is expected to be in a protocol-specific format
42122 with all scalar multibyte datatypes being big endian. Translation to
42123 this representation needs to be done both by the target before the @code{F}
42124 packet is sent, and by @value{GDBN} before
42125 it transfers memory to the target. Transferred pointers to structured
42126 data should point to the already-coerced data at any time.
42130 @unnumberedsubsubsec struct stat
42131 @cindex struct stat, in file-i/o protocol
42133 The buffer of type @code{struct stat} used by the target and @value{GDBN}
42134 is defined as follows:
42138 unsigned int st_dev; /* device */
42139 unsigned int st_ino; /* inode */
42140 mode_t st_mode; /* protection */
42141 unsigned int st_nlink; /* number of hard links */
42142 unsigned int st_uid; /* user ID of owner */
42143 unsigned int st_gid; /* group ID of owner */
42144 unsigned int st_rdev; /* device type (if inode device) */
42145 unsigned long st_size; /* total size, in bytes */
42146 unsigned long st_blksize; /* blocksize for filesystem I/O */
42147 unsigned long st_blocks; /* number of blocks allocated */
42148 time_t st_atime; /* time of last access */
42149 time_t st_mtime; /* time of last modification */
42150 time_t st_ctime; /* time of last change */
42154 The integral datatypes conform to the definitions given in the
42155 appropriate section (see @ref{Integral Datatypes}, for details) so this
42156 structure is of size 64 bytes.
42158 The values of several fields have a restricted meaning and/or
42164 A value of 0 represents a file, 1 the console.
42167 No valid meaning for the target. Transmitted unchanged.
42170 Valid mode bits are described in @ref{Constants}. Any other
42171 bits have currently no meaning for the target.
42176 No valid meaning for the target. Transmitted unchanged.
42181 These values have a host and file system dependent
42182 accuracy. Especially on Windows hosts, the file system may not
42183 support exact timing values.
42186 The target gets a @code{struct stat} of the above representation and is
42187 responsible for coercing it to the target representation before
42190 Note that due to size differences between the host, target, and protocol
42191 representations of @code{struct stat} members, these members could eventually
42192 get truncated on the target.
42194 @node struct timeval
42195 @unnumberedsubsubsec struct timeval
42196 @cindex struct timeval, in file-i/o protocol
42198 The buffer of type @code{struct timeval} used by the File-I/O protocol
42199 is defined as follows:
42203 time_t tv_sec; /* second */
42204 long tv_usec; /* microsecond */
42208 The integral datatypes conform to the definitions given in the
42209 appropriate section (see @ref{Integral Datatypes}, for details) so this
42210 structure is of size 8 bytes.
42213 @subsection Constants
42214 @cindex constants, in file-i/o protocol
42216 The following values are used for the constants inside of the
42217 protocol. @value{GDBN} and target are responsible for translating these
42218 values before and after the call as needed.
42229 @unnumberedsubsubsec Open Flags
42230 @cindex open flags, in file-i/o protocol
42232 All values are given in hexadecimal representation.
42244 @node mode_t Values
42245 @unnumberedsubsubsec mode_t Values
42246 @cindex mode_t values, in file-i/o protocol
42248 All values are given in octal representation.
42265 @unnumberedsubsubsec Errno Values
42266 @cindex errno values, in file-i/o protocol
42268 All values are given in decimal representation.
42293 @code{EUNKNOWN} is used as a fallback error value if a host system returns
42294 any error value not in the list of supported error numbers.
42297 @unnumberedsubsubsec Lseek Flags
42298 @cindex lseek flags, in file-i/o protocol
42307 @unnumberedsubsubsec Limits
42308 @cindex limits, in file-i/o protocol
42310 All values are given in decimal representation.
42313 INT_MIN -2147483648
42315 UINT_MAX 4294967295
42316 LONG_MIN -9223372036854775808
42317 LONG_MAX 9223372036854775807
42318 ULONG_MAX 18446744073709551615
42321 @node File-I/O Examples
42322 @subsection File-I/O Examples
42323 @cindex file-i/o examples
42325 Example sequence of a write call, file descriptor 3, buffer is at target
42326 address 0x1234, 6 bytes should be written:
42329 <- @code{Fwrite,3,1234,6}
42330 @emph{request memory read from target}
42333 @emph{return "6 bytes written"}
42337 Example sequence of a read call, file descriptor 3, buffer is at target
42338 address 0x1234, 6 bytes should be read:
42341 <- @code{Fread,3,1234,6}
42342 @emph{request memory write to target}
42343 -> @code{X1234,6:XXXXXX}
42344 @emph{return "6 bytes read"}
42348 Example sequence of a read call, call fails on the host due to invalid
42349 file descriptor (@code{EBADF}):
42352 <- @code{Fread,3,1234,6}
42356 Example sequence of a read call, user presses @kbd{Ctrl-c} before syscall on
42360 <- @code{Fread,3,1234,6}
42365 Example sequence of a read call, user presses @kbd{Ctrl-c} after syscall on
42369 <- @code{Fread,3,1234,6}
42370 -> @code{X1234,6:XXXXXX}
42374 @node Library List Format
42375 @section Library List Format
42376 @cindex library list format, remote protocol
42378 On some platforms, a dynamic loader (e.g.@: @file{ld.so}) runs in the
42379 same process as your application to manage libraries. In this case,
42380 @value{GDBN} can use the loader's symbol table and normal memory
42381 operations to maintain a list of shared libraries. On other
42382 platforms, the operating system manages loaded libraries.
42383 @value{GDBN} can not retrieve the list of currently loaded libraries
42384 through memory operations, so it uses the @samp{qXfer:libraries:read}
42385 packet (@pxref{qXfer library list read}) instead. The remote stub
42386 queries the target's operating system and reports which libraries
42389 The @samp{qXfer:libraries:read} packet returns an XML document which
42390 lists loaded libraries and their offsets. Each library has an
42391 associated name and one or more segment or section base addresses,
42392 which report where the library was loaded in memory.
42394 For the common case of libraries that are fully linked binaries, the
42395 library should have a list of segments. If the target supports
42396 dynamic linking of a relocatable object file, its library XML element
42397 should instead include a list of allocated sections. The segment or
42398 section bases are start addresses, not relocation offsets; they do not
42399 depend on the library's link-time base addresses.
42401 @value{GDBN} must be linked with the Expat library to support XML
42402 library lists. @xref{Expat}.
42404 A simple memory map, with one loaded library relocated by a single
42405 offset, looks like this:
42409 <library name="/lib/libc.so.6">
42410 <segment address="0x10000000"/>
42415 Another simple memory map, with one loaded library with three
42416 allocated sections (.text, .data, .bss), looks like this:
42420 <library name="sharedlib.o">
42421 <section address="0x10000000"/>
42422 <section address="0x20000000"/>
42423 <section address="0x30000000"/>
42428 The format of a library list is described by this DTD:
42431 <!-- library-list: Root element with versioning -->
42432 <!ELEMENT library-list (library)*>
42433 <!ATTLIST library-list version CDATA #FIXED "1.0">
42434 <!ELEMENT library (segment*, section*)>
42435 <!ATTLIST library name CDATA #REQUIRED>
42436 <!ELEMENT segment EMPTY>
42437 <!ATTLIST segment address CDATA #REQUIRED>
42438 <!ELEMENT section EMPTY>
42439 <!ATTLIST section address CDATA #REQUIRED>
42442 In addition, segments and section descriptors cannot be mixed within a
42443 single library element, and you must supply at least one segment or
42444 section for each library.
42446 @node Library List Format for SVR4 Targets
42447 @section Library List Format for SVR4 Targets
42448 @cindex library list format, remote protocol
42450 On SVR4 platforms @value{GDBN} can use the symbol table of a dynamic loader
42451 (e.g.@: @file{ld.so}) and normal memory operations to maintain a list of
42452 shared libraries. Still a special library list provided by this packet is
42453 more efficient for the @value{GDBN} remote protocol.
42455 The @samp{qXfer:libraries-svr4:read} packet returns an XML document which lists
42456 loaded libraries and their SVR4 linker parameters. For each library on SVR4
42457 target, the following parameters are reported:
42461 @code{name}, the absolute file name from the @code{l_name} field of
42462 @code{struct link_map}.
42464 @code{lm} with address of @code{struct link_map} used for TLS
42465 (Thread Local Storage) access.
42467 @code{l_addr}, the displacement as read from the field @code{l_addr} of
42468 @code{struct link_map}. For prelinked libraries this is not an absolute
42469 memory address. It is a displacement of absolute memory address against
42470 address the file was prelinked to during the library load.
42472 @code{l_ld}, which is memory address of the @code{PT_DYNAMIC} segment
42475 Additionally the single @code{main-lm} attribute specifies address of
42476 @code{struct link_map} used for the main executable. This parameter is used
42477 for TLS access and its presence is optional.
42479 @value{GDBN} must be linked with the Expat library to support XML
42480 SVR4 library lists. @xref{Expat}.
42482 A simple memory map, with two loaded libraries (which do not use prelink),
42486 <library-list-svr4 version="1.0" main-lm="0xe4f8f8">
42487 <library name="/lib/ld-linux.so.2" lm="0xe4f51c" l_addr="0xe2d000"
42489 <library name="/lib/libc.so.6" lm="0xe4fbe8" l_addr="0x154000"
42491 </library-list-svr>
42494 The format of an SVR4 library list is described by this DTD:
42497 <!-- library-list-svr4: Root element with versioning -->
42498 <!ELEMENT library-list-svr4 (library)*>
42499 <!ATTLIST library-list-svr4 version CDATA #FIXED "1.0">
42500 <!ATTLIST library-list-svr4 main-lm CDATA #IMPLIED>
42501 <!ELEMENT library EMPTY>
42502 <!ATTLIST library name CDATA #REQUIRED>
42503 <!ATTLIST library lm CDATA #REQUIRED>
42504 <!ATTLIST library l_addr CDATA #REQUIRED>
42505 <!ATTLIST library l_ld CDATA #REQUIRED>
42508 @node Memory Map Format
42509 @section Memory Map Format
42510 @cindex memory map format
42512 To be able to write into flash memory, @value{GDBN} needs to obtain a
42513 memory map from the target. This section describes the format of the
42516 The memory map is obtained using the @samp{qXfer:memory-map:read}
42517 (@pxref{qXfer memory map read}) packet and is an XML document that
42518 lists memory regions.
42520 @value{GDBN} must be linked with the Expat library to support XML
42521 memory maps. @xref{Expat}.
42523 The top-level structure of the document is shown below:
42526 <?xml version="1.0"?>
42527 <!DOCTYPE memory-map
42528 PUBLIC "+//IDN gnu.org//DTD GDB Memory Map V1.0//EN"
42529 "http://sourceware.org/gdb/gdb-memory-map.dtd">
42535 Each region can be either:
42540 A region of RAM starting at @var{addr} and extending for @var{length}
42544 <memory type="ram" start="@var{addr}" length="@var{length}"/>
42549 A region of read-only memory:
42552 <memory type="rom" start="@var{addr}" length="@var{length}"/>
42557 A region of flash memory, with erasure blocks @var{blocksize}
42561 <memory type="flash" start="@var{addr}" length="@var{length}">
42562 <property name="blocksize">@var{blocksize}</property>
42568 Regions must not overlap. @value{GDBN} assumes that areas of memory not covered
42569 by the memory map are RAM, and uses the ordinary @samp{M} and @samp{X}
42570 packets to write to addresses in such ranges.
42572 The formal DTD for memory map format is given below:
42575 <!-- ................................................... -->
42576 <!-- Memory Map XML DTD ................................ -->
42577 <!-- File: memory-map.dtd .............................. -->
42578 <!-- .................................... .............. -->
42579 <!-- memory-map.dtd -->
42580 <!-- memory-map: Root element with versioning -->
42581 <!ELEMENT memory-map (memory)*>
42582 <!ATTLIST memory-map version CDATA #FIXED "1.0.0">
42583 <!ELEMENT memory (property)*>
42584 <!-- memory: Specifies a memory region,
42585 and its type, or device. -->
42586 <!ATTLIST memory type (ram|rom|flash) #REQUIRED
42587 start CDATA #REQUIRED
42588 length CDATA #REQUIRED>
42589 <!-- property: Generic attribute tag -->
42590 <!ELEMENT property (#PCDATA | property)*>
42591 <!ATTLIST property name (blocksize) #REQUIRED>
42594 @node Thread List Format
42595 @section Thread List Format
42596 @cindex thread list format
42598 To efficiently update the list of threads and their attributes,
42599 @value{GDBN} issues the @samp{qXfer:threads:read} packet
42600 (@pxref{qXfer threads read}) and obtains the XML document with
42601 the following structure:
42604 <?xml version="1.0"?>
42606 <thread id="id" core="0" name="name">
42607 ... description ...
42612 Each @samp{thread} element must have the @samp{id} attribute that
42613 identifies the thread (@pxref{thread-id syntax}). The
42614 @samp{core} attribute, if present, specifies which processor core
42615 the thread was last executing on. The @samp{name} attribute, if
42616 present, specifies the human-readable name of the thread. The content
42617 of the of @samp{thread} element is interpreted as human-readable
42618 auxiliary information. The @samp{handle} attribute, if present,
42619 is a hex encoded representation of the thread handle.
42622 @node Traceframe Info Format
42623 @section Traceframe Info Format
42624 @cindex traceframe info format
42626 To be able to know which objects in the inferior can be examined when
42627 inspecting a tracepoint hit, @value{GDBN} needs to obtain the list of
42628 memory ranges, registers and trace state variables that have been
42629 collected in a traceframe.
42631 This list is obtained using the @samp{qXfer:traceframe-info:read}
42632 (@pxref{qXfer traceframe info read}) packet and is an XML document.
42634 @value{GDBN} must be linked with the Expat library to support XML
42635 traceframe info discovery. @xref{Expat}.
42637 The top-level structure of the document is shown below:
42640 <?xml version="1.0"?>
42641 <!DOCTYPE traceframe-info
42642 PUBLIC "+//IDN gnu.org//DTD GDB Memory Map V1.0//EN"
42643 "http://sourceware.org/gdb/gdb-traceframe-info.dtd">
42649 Each traceframe block can be either:
42654 A region of collected memory starting at @var{addr} and extending for
42655 @var{length} bytes from there:
42658 <memory start="@var{addr}" length="@var{length}"/>
42662 A block indicating trace state variable numbered @var{number} has been
42666 <tvar id="@var{number}"/>
42671 The formal DTD for the traceframe info format is given below:
42674 <!ELEMENT traceframe-info (memory | tvar)* >
42675 <!ATTLIST traceframe-info version CDATA #FIXED "1.0">
42677 <!ELEMENT memory EMPTY>
42678 <!ATTLIST memory start CDATA #REQUIRED
42679 length CDATA #REQUIRED>
42681 <!ATTLIST tvar id CDATA #REQUIRED>
42684 @node Branch Trace Format
42685 @section Branch Trace Format
42686 @cindex branch trace format
42688 In order to display the branch trace of an inferior thread,
42689 @value{GDBN} needs to obtain the list of branches. This list is
42690 represented as list of sequential code blocks that are connected via
42691 branches. The code in each block has been executed sequentially.
42693 This list is obtained using the @samp{qXfer:btrace:read}
42694 (@pxref{qXfer btrace read}) packet and is an XML document.
42696 @value{GDBN} must be linked with the Expat library to support XML
42697 traceframe info discovery. @xref{Expat}.
42699 The top-level structure of the document is shown below:
42702 <?xml version="1.0"?>
42704 PUBLIC "+//IDN gnu.org//DTD GDB Branch Trace V1.0//EN"
42705 "http://sourceware.org/gdb/gdb-btrace.dtd">
42714 A block of sequentially executed instructions starting at @var{begin}
42715 and ending at @var{end}:
42718 <block begin="@var{begin}" end="@var{end}"/>
42723 The formal DTD for the branch trace format is given below:
42726 <!ELEMENT btrace (block* | pt) >
42727 <!ATTLIST btrace version CDATA #FIXED "1.0">
42729 <!ELEMENT block EMPTY>
42730 <!ATTLIST block begin CDATA #REQUIRED
42731 end CDATA #REQUIRED>
42733 <!ELEMENT pt (pt-config?, raw?)>
42735 <!ELEMENT pt-config (cpu?)>
42737 <!ELEMENT cpu EMPTY>
42738 <!ATTLIST cpu vendor CDATA #REQUIRED
42739 family CDATA #REQUIRED
42740 model CDATA #REQUIRED
42741 stepping CDATA #REQUIRED>
42743 <!ELEMENT raw (#PCDATA)>
42746 @node Branch Trace Configuration Format
42747 @section Branch Trace Configuration Format
42748 @cindex branch trace configuration format
42750 For each inferior thread, @value{GDBN} can obtain the branch trace
42751 configuration using the @samp{qXfer:btrace-conf:read}
42752 (@pxref{qXfer btrace-conf read}) packet.
42754 The configuration describes the branch trace format and configuration
42755 settings for that format. The following information is described:
42759 This thread uses the @dfn{Branch Trace Store} (@acronym{BTS}) format.
42762 The size of the @acronym{BTS} ring buffer in bytes.
42765 This thread uses the @dfn{Intel Processor Trace} (@acronym{Intel
42769 The size of the @acronym{Intel PT} ring buffer in bytes.
42773 @value{GDBN} must be linked with the Expat library to support XML
42774 branch trace configuration discovery. @xref{Expat}.
42776 The formal DTD for the branch trace configuration format is given below:
42779 <!ELEMENT btrace-conf (bts?, pt?)>
42780 <!ATTLIST btrace-conf version CDATA #FIXED "1.0">
42782 <!ELEMENT bts EMPTY>
42783 <!ATTLIST bts size CDATA #IMPLIED>
42785 <!ELEMENT pt EMPTY>
42786 <!ATTLIST pt size CDATA #IMPLIED>
42789 @include agentexpr.texi
42791 @node Target Descriptions
42792 @appendix Target Descriptions
42793 @cindex target descriptions
42795 One of the challenges of using @value{GDBN} to debug embedded systems
42796 is that there are so many minor variants of each processor
42797 architecture in use. It is common practice for vendors to start with
42798 a standard processor core --- ARM, PowerPC, or @acronym{MIPS}, for example ---
42799 and then make changes to adapt it to a particular market niche. Some
42800 architectures have hundreds of variants, available from dozens of
42801 vendors. This leads to a number of problems:
42805 With so many different customized processors, it is difficult for
42806 the @value{GDBN} maintainers to keep up with the changes.
42808 Since individual variants may have short lifetimes or limited
42809 audiences, it may not be worthwhile to carry information about every
42810 variant in the @value{GDBN} source tree.
42812 When @value{GDBN} does support the architecture of the embedded system
42813 at hand, the task of finding the correct architecture name to give the
42814 @command{set architecture} command can be error-prone.
42817 To address these problems, the @value{GDBN} remote protocol allows a
42818 target system to not only identify itself to @value{GDBN}, but to
42819 actually describe its own features. This lets @value{GDBN} support
42820 processor variants it has never seen before --- to the extent that the
42821 descriptions are accurate, and that @value{GDBN} understands them.
42823 @value{GDBN} must be linked with the Expat library to support XML
42824 target descriptions. @xref{Expat}.
42827 * Retrieving Descriptions:: How descriptions are fetched from a target.
42828 * Target Description Format:: The contents of a target description.
42829 * Predefined Target Types:: Standard types available for target
42831 * Enum Target Types:: How to define enum target types.
42832 * Standard Target Features:: Features @value{GDBN} knows about.
42835 @node Retrieving Descriptions
42836 @section Retrieving Descriptions
42838 Target descriptions can be read from the target automatically, or
42839 specified by the user manually. The default behavior is to read the
42840 description from the target. @value{GDBN} retrieves it via the remote
42841 protocol using @samp{qXfer} requests (@pxref{General Query Packets,
42842 qXfer}). The @var{annex} in the @samp{qXfer} packet will be
42843 @samp{target.xml}. The contents of the @samp{target.xml} annex are an
42844 XML document, of the form described in @ref{Target Description
42847 Alternatively, you can specify a file to read for the target description.
42848 If a file is set, the target will not be queried. The commands to
42849 specify a file are:
42852 @cindex set tdesc filename
42853 @item set tdesc filename @var{path}
42854 Read the target description from @var{path}.
42856 @cindex unset tdesc filename
42857 @item unset tdesc filename
42858 Do not read the XML target description from a file. @value{GDBN}
42859 will use the description supplied by the current target.
42861 @cindex show tdesc filename
42862 @item show tdesc filename
42863 Show the filename to read for a target description, if any.
42867 @node Target Description Format
42868 @section Target Description Format
42869 @cindex target descriptions, XML format
42871 A target description annex is an @uref{http://www.w3.org/XML/, XML}
42872 document which complies with the Document Type Definition provided in
42873 the @value{GDBN} sources in @file{gdb/features/gdb-target.dtd}. This
42874 means you can use generally available tools like @command{xmllint} to
42875 check that your feature descriptions are well-formed and valid.
42876 However, to help people unfamiliar with XML write descriptions for
42877 their targets, we also describe the grammar here.
42879 Target descriptions can identify the architecture of the remote target
42880 and (for some architectures) provide information about custom register
42881 sets. They can also identify the OS ABI of the remote target.
42882 @value{GDBN} can use this information to autoconfigure for your
42883 target, or to warn you if you connect to an unsupported target.
42885 Here is a simple target description:
42888 <target version="1.0">
42889 <architecture>i386:x86-64</architecture>
42894 This minimal description only says that the target uses
42895 the x86-64 architecture.
42897 A target description has the following overall form, with [ ] marking
42898 optional elements and @dots{} marking repeatable elements. The elements
42899 are explained further below.
42902 <?xml version="1.0"?>
42903 <!DOCTYPE target SYSTEM "gdb-target.dtd">
42904 <target version="1.0">
42905 @r{[}@var{architecture}@r{]}
42906 @r{[}@var{osabi}@r{]}
42907 @r{[}@var{compatible}@r{]}
42908 @r{[}@var{feature}@dots{}@r{]}
42913 The description is generally insensitive to whitespace and line
42914 breaks, under the usual common-sense rules. The XML version
42915 declaration and document type declaration can generally be omitted
42916 (@value{GDBN} does not require them), but specifying them may be
42917 useful for XML validation tools. The @samp{version} attribute for
42918 @samp{<target>} may also be omitted, but we recommend
42919 including it; if future versions of @value{GDBN} use an incompatible
42920 revision of @file{gdb-target.dtd}, they will detect and report
42921 the version mismatch.
42923 @subsection Inclusion
42924 @cindex target descriptions, inclusion
42927 @cindex <xi:include>
42930 It can sometimes be valuable to split a target description up into
42931 several different annexes, either for organizational purposes, or to
42932 share files between different possible target descriptions. You can
42933 divide a description into multiple files by replacing any element of
42934 the target description with an inclusion directive of the form:
42937 <xi:include href="@var{document}"/>
42941 When @value{GDBN} encounters an element of this form, it will retrieve
42942 the named XML @var{document}, and replace the inclusion directive with
42943 the contents of that document. If the current description was read
42944 using @samp{qXfer}, then so will be the included document;
42945 @var{document} will be interpreted as the name of an annex. If the
42946 current description was read from a file, @value{GDBN} will look for
42947 @var{document} as a file in the same directory where it found the
42948 original description.
42950 @subsection Architecture
42951 @cindex <architecture>
42953 An @samp{<architecture>} element has this form:
42956 <architecture>@var{arch}</architecture>
42959 @var{arch} is one of the architectures from the set accepted by
42960 @code{set architecture} (@pxref{Targets, ,Specifying a Debugging Target}).
42963 @cindex @code{<osabi>}
42965 This optional field was introduced in @value{GDBN} version 7.0.
42966 Previous versions of @value{GDBN} ignore it.
42968 An @samp{<osabi>} element has this form:
42971 <osabi>@var{abi-name}</osabi>
42974 @var{abi-name} is an OS ABI name from the same selection accepted by
42975 @w{@code{set osabi}} (@pxref{ABI, ,Configuring the Current ABI}).
42977 @subsection Compatible Architecture
42978 @cindex @code{<compatible>}
42980 This optional field was introduced in @value{GDBN} version 7.0.
42981 Previous versions of @value{GDBN} ignore it.
42983 A @samp{<compatible>} element has this form:
42986 <compatible>@var{arch}</compatible>
42989 @var{arch} is one of the architectures from the set accepted by
42990 @code{set architecture} (@pxref{Targets, ,Specifying a Debugging Target}).
42992 A @samp{<compatible>} element is used to specify that the target
42993 is able to run binaries in some other than the main target architecture
42994 given by the @samp{<architecture>} element. For example, on the
42995 Cell Broadband Engine, the main architecture is @code{powerpc:common}
42996 or @code{powerpc:common64}, but the system is able to run binaries
42997 in the @code{spu} architecture as well. The way to describe this
42998 capability with @samp{<compatible>} is as follows:
43001 <architecture>powerpc:common</architecture>
43002 <compatible>spu</compatible>
43005 @subsection Features
43008 Each @samp{<feature>} describes some logical portion of the target
43009 system. Features are currently used to describe available CPU
43010 registers and the types of their contents. A @samp{<feature>} element
43014 <feature name="@var{name}">
43015 @r{[}@var{type}@dots{}@r{]}
43021 Each feature's name should be unique within the description. The name
43022 of a feature does not matter unless @value{GDBN} has some special
43023 knowledge of the contents of that feature; if it does, the feature
43024 should have its standard name. @xref{Standard Target Features}.
43028 Any register's value is a collection of bits which @value{GDBN} must
43029 interpret. The default interpretation is a two's complement integer,
43030 but other types can be requested by name in the register description.
43031 Some predefined types are provided by @value{GDBN} (@pxref{Predefined
43032 Target Types}), and the description can define additional composite
43035 Each type element must have an @samp{id} attribute, which gives
43036 a unique (within the containing @samp{<feature>}) name to the type.
43037 Types must be defined before they are used.
43040 Some targets offer vector registers, which can be treated as arrays
43041 of scalar elements. These types are written as @samp{<vector>} elements,
43042 specifying the array element type, @var{type}, and the number of elements,
43046 <vector id="@var{id}" type="@var{type}" count="@var{count}"/>
43050 If a register's value is usefully viewed in multiple ways, define it
43051 with a union type containing the useful representations. The
43052 @samp{<union>} element contains one or more @samp{<field>} elements,
43053 each of which has a @var{name} and a @var{type}:
43056 <union id="@var{id}">
43057 <field name="@var{name}" type="@var{type}"/>
43064 If a register's value is composed from several separate values, define
43065 it with either a structure type or a flags type.
43066 A flags type may only contain bitfields.
43067 A structure type may either contain only bitfields or contain no bitfields.
43068 If the value contains only bitfields, its total size in bytes must be
43071 Non-bitfield values have a @var{name} and @var{type}.
43074 <struct id="@var{id}">
43075 <field name="@var{name}" type="@var{type}"/>
43080 Both @var{name} and @var{type} values are required.
43081 No implicit padding is added.
43083 Bitfield values have a @var{name}, @var{start}, @var{end} and @var{type}.
43086 <struct id="@var{id}" size="@var{size}">
43087 <field name="@var{name}" start="@var{start}" end="@var{end}" type="@var{type}"/>
43093 <flags id="@var{id}" size="@var{size}">
43094 <field name="@var{name}" start="@var{start}" end="@var{end}" type="@var{type}"/>
43099 The @var{name} value is required.
43100 Bitfield values may be named with the empty string, @samp{""},
43101 in which case the field is ``filler'' and its value is not printed.
43102 Not all bits need to be specified, so ``filler'' fields are optional.
43104 The @var{start} and @var{end} values are required, and @var{type}
43106 The field's @var{start} must be less than or equal to its @var{end},
43107 and zero represents the least significant bit.
43109 The default value of @var{type} is @code{bool} for single bit fields,
43110 and an unsigned integer otherwise.
43112 Which to choose? Structures or flags?
43114 Registers defined with @samp{flags} have these advantages over
43115 defining them with @samp{struct}:
43119 Arithmetic may be performed on them as if they were integers.
43121 They are printed in a more readable fashion.
43124 Registers defined with @samp{struct} have one advantage over
43125 defining them with @samp{flags}:
43129 One can fetch individual fields like in @samp{C}.
43132 (gdb) print $my_struct_reg.field3
43138 @subsection Registers
43141 Each register is represented as an element with this form:
43144 <reg name="@var{name}"
43145 bitsize="@var{size}"
43146 @r{[}regnum="@var{num}"@r{]}
43147 @r{[}save-restore="@var{save-restore}"@r{]}
43148 @r{[}type="@var{type}"@r{]}
43149 @r{[}group="@var{group}"@r{]}/>
43153 The components are as follows:
43158 The register's name; it must be unique within the target description.
43161 The register's size, in bits.
43164 The register's number. If omitted, a register's number is one greater
43165 than that of the previous register (either in the current feature or in
43166 a preceding feature); the first register in the target description
43167 defaults to zero. This register number is used to read or write
43168 the register; e.g.@: it is used in the remote @code{p} and @code{P}
43169 packets, and registers appear in the @code{g} and @code{G} packets
43170 in order of increasing register number.
43173 Whether the register should be preserved across inferior function
43174 calls; this must be either @code{yes} or @code{no}. The default is
43175 @code{yes}, which is appropriate for most registers except for
43176 some system control registers; this is not related to the target's
43180 The type of the register. It may be a predefined type, a type
43181 defined in the current feature, or one of the special types @code{int}
43182 and @code{float}. @code{int} is an integer type of the correct size
43183 for @var{bitsize}, and @code{float} is a floating point type (in the
43184 architecture's normal floating point format) of the correct size for
43185 @var{bitsize}. The default is @code{int}.
43188 The register group to which this register belongs. It can be one of the
43189 standard register groups @code{general}, @code{float}, @code{vector} or an
43190 arbitrary string. Group names should be limited to alphanumeric characters.
43191 If a group name is made up of multiple words the words may be separated by
43192 hyphens; e.g.@: @code{special-group} or @code{ultra-special-group}. If no
43193 @var{group} is specified, @value{GDBN} will not display the register in
43194 @code{info registers}.
43198 @node Predefined Target Types
43199 @section Predefined Target Types
43200 @cindex target descriptions, predefined types
43202 Type definitions in the self-description can build up composite types
43203 from basic building blocks, but can not define fundamental types. Instead,
43204 standard identifiers are provided by @value{GDBN} for the fundamental
43205 types. The currently supported types are:
43210 Boolean type, occupying a single bit.
43218 Signed integer types holding the specified number of bits.
43226 Unsigned integer types holding the specified number of bits.
43230 Pointers to unspecified code and data. The program counter and
43231 any dedicated return address register may be marked as code
43232 pointers; printing a code pointer converts it into a symbolic
43233 address. The stack pointer and any dedicated address registers
43234 may be marked as data pointers.
43237 Single precision IEEE floating point.
43240 Double precision IEEE floating point.
43243 The 12-byte extended precision format used by ARM FPA registers.
43246 The 10-byte extended precision format used by x87 registers.
43249 32bit @sc{eflags} register used by x86.
43252 32bit @sc{mxcsr} register used by x86.
43256 @node Enum Target Types
43257 @section Enum Target Types
43258 @cindex target descriptions, enum types
43260 Enum target types are useful in @samp{struct} and @samp{flags}
43261 register descriptions. @xref{Target Description Format}.
43263 Enum types have a name, size and a list of name/value pairs.
43266 <enum id="@var{id}" size="@var{size}">
43267 <evalue name="@var{name}" value="@var{value}"/>
43272 Enums must be defined before they are used.
43275 <enum id="levels_type" size="4">
43276 <evalue name="low" value="0"/>
43277 <evalue name="high" value="1"/>
43279 <flags id="flags_type" size="4">
43280 <field name="X" start="0"/>
43281 <field name="LEVEL" start="1" end="1" type="levels_type"/>
43283 <reg name="flags" bitsize="32" type="flags_type"/>
43286 Given that description, a value of 3 for the @samp{flags} register
43287 would be printed as:
43290 (gdb) info register flags
43291 flags 0x3 [ X LEVEL=high ]
43294 @node Standard Target Features
43295 @section Standard Target Features
43296 @cindex target descriptions, standard features
43298 A target description must contain either no registers or all the
43299 target's registers. If the description contains no registers, then
43300 @value{GDBN} will assume a default register layout, selected based on
43301 the architecture. If the description contains any registers, the
43302 default layout will not be used; the standard registers must be
43303 described in the target description, in such a way that @value{GDBN}
43304 can recognize them.
43306 This is accomplished by giving specific names to feature elements
43307 which contain standard registers. @value{GDBN} will look for features
43308 with those names and verify that they contain the expected registers;
43309 if any known feature is missing required registers, or if any required
43310 feature is missing, @value{GDBN} will reject the target
43311 description. You can add additional registers to any of the
43312 standard features --- @value{GDBN} will display them just as if
43313 they were added to an unrecognized feature.
43315 This section lists the known features and their expected contents.
43316 Sample XML documents for these features are included in the
43317 @value{GDBN} source tree, in the directory @file{gdb/features}.
43319 Names recognized by @value{GDBN} should include the name of the
43320 company or organization which selected the name, and the overall
43321 architecture to which the feature applies; so e.g.@: the feature
43322 containing ARM core registers is named @samp{org.gnu.gdb.arm.core}.
43324 The names of registers are not case sensitive for the purpose
43325 of recognizing standard features, but @value{GDBN} will only display
43326 registers using the capitalization used in the description.
43329 * AArch64 Features::
43333 * MicroBlaze Features::
43337 * Nios II Features::
43338 * OpenRISC 1000 Features::
43339 * PowerPC Features::
43340 * RISC-V Features::
43341 * S/390 and System z Features::
43347 @node AArch64 Features
43348 @subsection AArch64 Features
43349 @cindex target descriptions, AArch64 features
43351 The @samp{org.gnu.gdb.aarch64.core} feature is required for AArch64
43352 targets. It should contain registers @samp{x0} through @samp{x30},
43353 @samp{sp}, @samp{pc}, and @samp{cpsr}.
43355 The @samp{org.gnu.gdb.aarch64.fpu} feature is optional. If present,
43356 it should contain registers @samp{v0} through @samp{v31}, @samp{fpsr},
43359 The @samp{org.gnu.gdb.aarch64.sve} feature is optional. If present,
43360 it should contain registers @samp{z0} through @samp{z31}, @samp{p0}
43361 through @samp{p15}, @samp{ffr} and @samp{vg}.
43363 The @samp{org.gnu.gdb.aarch64.pauth} feature is optional. If present,
43364 it should contain registers @samp{pauth_dmask} and @samp{pauth_cmask}.
43367 @subsection ARC Features
43368 @cindex target descriptions, ARC Features
43370 ARC processors are highly configurable, so even core registers and their number
43371 are not completely predetermined. In addition flags and PC registers which are
43372 important to @value{GDBN} are not ``core'' registers in ARC. It is required
43373 that one of the core registers features is present.
43374 @samp{org.gnu.gdb.arc.aux-minimal} feature is mandatory.
43376 The @samp{org.gnu.gdb.arc.core.v2} feature is required for ARC EM and ARC HS
43377 targets with a normal register file. It should contain registers @samp{r0}
43378 through @samp{r25}, @samp{gp}, @samp{fp}, @samp{sp}, @samp{r30}, @samp{blink},
43379 @samp{lp_count} and @samp{pcl}. This feature may contain register @samp{ilink}
43380 and any of extension core registers @samp{r32} through @samp{r59/acch}.
43381 @samp{ilink} and extension core registers are not available to read/write, when
43382 debugging GNU/Linux applications, thus @samp{ilink} is made optional.
43384 The @samp{org.gnu.gdb.arc.core-reduced.v2} feature is required for ARC EM and
43385 ARC HS targets with a reduced register file. It should contain registers
43386 @samp{r0} through @samp{r3}, @samp{r10} through @samp{r15}, @samp{gp},
43387 @samp{fp}, @samp{sp}, @samp{r30}, @samp{blink}, @samp{lp_count} and @samp{pcl}.
43388 This feature may contain register @samp{ilink} and any of extension core
43389 registers @samp{r32} through @samp{r59/acch}.
43391 The @samp{org.gnu.gdb.arc.core.arcompact} feature is required for ARCompact
43392 targets with a normal register file. It should contain registers @samp{r0}
43393 through @samp{r25}, @samp{gp}, @samp{fp}, @samp{sp}, @samp{r30}, @samp{blink},
43394 @samp{lp_count} and @samp{pcl}. This feature may contain registers
43395 @samp{ilink1}, @samp{ilink2} and any of extension core registers @samp{r32}
43396 through @samp{r59/acch}. @samp{ilink1} and @samp{ilink2} and extension core
43397 registers are not available when debugging GNU/Linux applications. The only
43398 difference with @samp{org.gnu.gdb.arc.core.v2} feature is in the names of
43399 @samp{ilink1} and @samp{ilink2} registers and that @samp{r30} is mandatory in
43400 ARC v2, but @samp{ilink2} is optional on ARCompact.
43402 The @samp{org.gnu.gdb.arc.aux-minimal} feature is required for all ARC
43403 targets. It should contain registers @samp{pc} and @samp{status32}.
43406 @subsection ARM Features
43407 @cindex target descriptions, ARM features
43409 The @samp{org.gnu.gdb.arm.core} feature is required for non-M-profile
43411 It should contain registers @samp{r0} through @samp{r13}, @samp{sp},
43412 @samp{lr}, @samp{pc}, and @samp{cpsr}.
43414 For M-profile targets (e.g. Cortex-M3), the @samp{org.gnu.gdb.arm.core}
43415 feature is replaced by @samp{org.gnu.gdb.arm.m-profile}. It should contain
43416 registers @samp{r0} through @samp{r13}, @samp{sp}, @samp{lr}, @samp{pc},
43419 The @samp{org.gnu.gdb.arm.fpa} feature is optional. If present, it
43420 should contain registers @samp{f0} through @samp{f7} and @samp{fps}.
43422 The @samp{org.gnu.gdb.xscale.iwmmxt} feature is optional. If present,
43423 it should contain at least registers @samp{wR0} through @samp{wR15} and
43424 @samp{wCGR0} through @samp{wCGR3}. The @samp{wCID}, @samp{wCon},
43425 @samp{wCSSF}, and @samp{wCASF} registers are optional.
43427 The @samp{org.gnu.gdb.arm.vfp} feature is optional. If present, it
43428 should contain at least registers @samp{d0} through @samp{d15}. If
43429 they are present, @samp{d16} through @samp{d31} should also be included.
43430 @value{GDBN} will synthesize the single-precision registers from
43431 halves of the double-precision registers.
43433 The @samp{org.gnu.gdb.arm.neon} feature is optional. It does not
43434 need to contain registers; it instructs @value{GDBN} to display the
43435 VFP double-precision registers as vectors and to synthesize the
43436 quad-precision registers from pairs of double-precision registers.
43437 If this feature is present, @samp{org.gnu.gdb.arm.vfp} must also
43438 be present and include 32 double-precision registers.
43440 @node i386 Features
43441 @subsection i386 Features
43442 @cindex target descriptions, i386 features
43444 The @samp{org.gnu.gdb.i386.core} feature is required for i386/amd64
43445 targets. It should describe the following registers:
43449 @samp{eax} through @samp{edi} plus @samp{eip} for i386
43451 @samp{rax} through @samp{r15} plus @samp{rip} for amd64
43453 @samp{eflags}, @samp{cs}, @samp{ss}, @samp{ds}, @samp{es},
43454 @samp{fs}, @samp{gs}
43456 @samp{st0} through @samp{st7}
43458 @samp{fctrl}, @samp{fstat}, @samp{ftag}, @samp{fiseg}, @samp{fioff},
43459 @samp{foseg}, @samp{fooff} and @samp{fop}
43462 The register sets may be different, depending on the target.
43464 The @samp{org.gnu.gdb.i386.sse} feature is optional. It should
43465 describe registers:
43469 @samp{xmm0} through @samp{xmm7} for i386
43471 @samp{xmm0} through @samp{xmm15} for amd64
43476 The @samp{org.gnu.gdb.i386.avx} feature is optional and requires the
43477 @samp{org.gnu.gdb.i386.sse} feature. It should
43478 describe the upper 128 bits of @sc{ymm} registers:
43482 @samp{ymm0h} through @samp{ymm7h} for i386
43484 @samp{ymm0h} through @samp{ymm15h} for amd64
43487 The @samp{org.gnu.gdb.i386.mpx} is an optional feature representing Intel
43488 Memory Protection Extension (MPX). It should describe the following registers:
43492 @samp{bnd0raw} through @samp{bnd3raw} for i386 and amd64.
43494 @samp{bndcfgu} and @samp{bndstatus} for i386 and amd64.
43497 The @samp{org.gnu.gdb.i386.linux} feature is optional. It should
43498 describe a single register, @samp{orig_eax}.
43500 The @samp{org.gnu.gdb.i386.segments} feature is optional. It should
43501 describe two system registers: @samp{fs_base} and @samp{gs_base}.
43503 The @samp{org.gnu.gdb.i386.avx512} feature is optional and requires the
43504 @samp{org.gnu.gdb.i386.avx} feature. It should
43505 describe additional @sc{xmm} registers:
43509 @samp{xmm16h} through @samp{xmm31h}, only valid for amd64.
43512 It should describe the upper 128 bits of additional @sc{ymm} registers:
43516 @samp{ymm16h} through @samp{ymm31h}, only valid for amd64.
43520 describe the upper 256 bits of @sc{zmm} registers:
43524 @samp{zmm0h} through @samp{zmm7h} for i386.
43526 @samp{zmm0h} through @samp{zmm15h} for amd64.
43530 describe the additional @sc{zmm} registers:
43534 @samp{zmm16h} through @samp{zmm31h}, only valid for amd64.
43537 The @samp{org.gnu.gdb.i386.pkeys} feature is optional. It should
43538 describe a single register, @samp{pkru}. It is a 32-bit register
43539 valid for i386 and amd64.
43541 @node MicroBlaze Features
43542 @subsection MicroBlaze Features
43543 @cindex target descriptions, MicroBlaze features
43545 The @samp{org.gnu.gdb.microblaze.core} feature is required for MicroBlaze
43546 targets. It should contain registers @samp{r0} through @samp{r31},
43547 @samp{rpc}, @samp{rmsr}, @samp{rear}, @samp{resr}, @samp{rfsr}, @samp{rbtr},
43548 @samp{rpvr}, @samp{rpvr1} through @samp{rpvr11}, @samp{redr}, @samp{rpid},
43549 @samp{rzpr}, @samp{rtlbx}, @samp{rtlbsx}, @samp{rtlblo}, and @samp{rtlbhi}.
43551 The @samp{org.gnu.gdb.microblaze.stack-protect} feature is optional.
43552 If present, it should contain registers @samp{rshr} and @samp{rslr}
43554 @node MIPS Features
43555 @subsection @acronym{MIPS} Features
43556 @cindex target descriptions, @acronym{MIPS} features
43558 The @samp{org.gnu.gdb.mips.cpu} feature is required for @acronym{MIPS} targets.
43559 It should contain registers @samp{r0} through @samp{r31}, @samp{lo},
43560 @samp{hi}, and @samp{pc}. They may be 32-bit or 64-bit depending
43563 The @samp{org.gnu.gdb.mips.cp0} feature is also required. It should
43564 contain at least the @samp{status}, @samp{badvaddr}, and @samp{cause}
43565 registers. They may be 32-bit or 64-bit depending on the target.
43567 The @samp{org.gnu.gdb.mips.fpu} feature is currently required, though
43568 it may be optional in a future version of @value{GDBN}. It should
43569 contain registers @samp{f0} through @samp{f31}, @samp{fcsr}, and
43570 @samp{fir}. They may be 32-bit or 64-bit depending on the target.
43572 The @samp{org.gnu.gdb.mips.dsp} feature is optional. It should
43573 contain registers @samp{hi1} through @samp{hi3}, @samp{lo1} through
43574 @samp{lo3}, and @samp{dspctl}. The @samp{dspctl} register should
43575 be 32-bit and the rest may be 32-bit or 64-bit depending on the target.
43577 The @samp{org.gnu.gdb.mips.linux} feature is optional. It should
43578 contain a single register, @samp{restart}, which is used by the
43579 Linux kernel to control restartable syscalls.
43581 @node M68K Features
43582 @subsection M68K Features
43583 @cindex target descriptions, M68K features
43586 @item @samp{org.gnu.gdb.m68k.core}
43587 @itemx @samp{org.gnu.gdb.coldfire.core}
43588 @itemx @samp{org.gnu.gdb.fido.core}
43589 One of those features must be always present.
43590 The feature that is present determines which flavor of m68k is
43591 used. The feature that is present should contain registers
43592 @samp{d0} through @samp{d7}, @samp{a0} through @samp{a5}, @samp{fp},
43593 @samp{sp}, @samp{ps} and @samp{pc}.
43595 @item @samp{org.gnu.gdb.coldfire.fp}
43596 This feature is optional. If present, it should contain registers
43597 @samp{fp0} through @samp{fp7}, @samp{fpcontrol}, @samp{fpstatus} and
43601 @node NDS32 Features
43602 @subsection NDS32 Features
43603 @cindex target descriptions, NDS32 features
43605 The @samp{org.gnu.gdb.nds32.core} feature is required for NDS32
43606 targets. It should contain at least registers @samp{r0} through
43607 @samp{r10}, @samp{r15}, @samp{fp}, @samp{gp}, @samp{lp}, @samp{sp},
43610 The @samp{org.gnu.gdb.nds32.fpu} feature is optional. If present,
43611 it should contain 64-bit double-precision floating-point registers
43612 @samp{fd0} through @emph{fdN}, which should be @samp{fd3}, @samp{fd7},
43613 @samp{fd15}, or @samp{fd31} based on the FPU configuration implemented.
43615 @emph{Note:} The first sixteen 64-bit double-precision floating-point
43616 registers are overlapped with the thirty-two 32-bit single-precision
43617 floating-point registers. The 32-bit single-precision registers, if
43618 not being listed explicitly, will be synthesized from halves of the
43619 overlapping 64-bit double-precision registers. Listing 32-bit
43620 single-precision registers explicitly is deprecated, and the
43621 support to it could be totally removed some day.
43623 @node Nios II Features
43624 @subsection Nios II Features
43625 @cindex target descriptions, Nios II features
43627 The @samp{org.gnu.gdb.nios2.cpu} feature is required for Nios II
43628 targets. It should contain the 32 core registers (@samp{zero},
43629 @samp{at}, @samp{r2} through @samp{r23}, @samp{et} through @samp{ra}),
43630 @samp{pc}, and the 16 control registers (@samp{status} through
43633 @node OpenRISC 1000 Features
43634 @subsection Openrisc 1000 Features
43635 @cindex target descriptions, OpenRISC 1000 features
43637 The @samp{org.gnu.gdb.or1k.group0} feature is required for OpenRISC 1000
43638 targets. It should contain the 32 general purpose registers (@samp{r0}
43639 through @samp{r31}), @samp{ppc}, @samp{npc} and @samp{sr}.
43641 @node PowerPC Features
43642 @subsection PowerPC Features
43643 @cindex target descriptions, PowerPC features
43645 The @samp{org.gnu.gdb.power.core} feature is required for PowerPC
43646 targets. It should contain registers @samp{r0} through @samp{r31},
43647 @samp{pc}, @samp{msr}, @samp{cr}, @samp{lr}, @samp{ctr}, and
43648 @samp{xer}. They may be 32-bit or 64-bit depending on the target.
43650 The @samp{org.gnu.gdb.power.fpu} feature is optional. It should
43651 contain registers @samp{f0} through @samp{f31} and @samp{fpscr}.
43653 The @samp{org.gnu.gdb.power.altivec} feature is optional. It should
43654 contain registers @samp{vr0} through @samp{vr31}, @samp{vscr}, and
43655 @samp{vrsave}. @value{GDBN} will define pseudo-registers @samp{v0}
43656 through @samp{v31} as aliases for the corresponding @samp{vrX}
43659 The @samp{org.gnu.gdb.power.vsx} feature is optional. It should
43660 contain registers @samp{vs0h} through @samp{vs31h}. @value{GDBN} will
43661 combine these registers with the floating point registers (@samp{f0}
43662 through @samp{f31}) and the altivec registers (@samp{vr0} through
43663 @samp{vr31}) to present the 128-bit wide registers @samp{vs0} through
43664 @samp{vs63}, the set of vector-scalar registers for POWER7.
43665 Therefore, this feature requires both @samp{org.gnu.gdb.power.fpu} and
43666 @samp{org.gnu.gdb.power.altivec}.
43668 The @samp{org.gnu.gdb.power.spe} feature is optional. It should
43669 contain registers @samp{ev0h} through @samp{ev31h}, @samp{acc}, and
43670 @samp{spefscr}. SPE targets should provide 32-bit registers in
43671 @samp{org.gnu.gdb.power.core} and provide the upper halves in
43672 @samp{ev0h} through @samp{ev31h}. @value{GDBN} will combine
43673 these to present registers @samp{ev0} through @samp{ev31} to the
43676 The @samp{org.gnu.gdb.power.ppr} feature is optional. It should
43677 contain the 64-bit register @samp{ppr}.
43679 The @samp{org.gnu.gdb.power.dscr} feature is optional. It should
43680 contain the 64-bit register @samp{dscr}.
43682 The @samp{org.gnu.gdb.power.tar} feature is optional. It should
43683 contain the 64-bit register @samp{tar}.
43685 The @samp{org.gnu.gdb.power.ebb} feature is optional. It should
43686 contain registers @samp{bescr}, @samp{ebbhr} and @samp{ebbrr}, all
43689 The @samp{org.gnu.gdb.power.linux.pmu} feature is optional. It should
43690 contain registers @samp{mmcr0}, @samp{mmcr2}, @samp{siar}, @samp{sdar}
43691 and @samp{sier}, all 64-bit wide. This is the subset of the isa 2.07
43692 server PMU registers provided by @sc{gnu}/Linux.
43694 The @samp{org.gnu.gdb.power.htm.spr} feature is optional. It should
43695 contain registers @samp{tfhar}, @samp{texasr} and @samp{tfiar}, all
43698 The @samp{org.gnu.gdb.power.htm.core} feature is optional. It should
43699 contain the checkpointed general-purpose registers @samp{cr0} through
43700 @samp{cr31}, as well as the checkpointed registers @samp{clr} and
43701 @samp{cctr}. These registers may all be either 32-bit or 64-bit
43702 depending on the target. It should also contain the checkpointed
43703 registers @samp{ccr} and @samp{cxer}, which should both be 32-bit
43706 The @samp{org.gnu.gdb.power.htm.fpu} feature is optional. It should
43707 contain the checkpointed 64-bit floating-point registers @samp{cf0}
43708 through @samp{cf31}, as well as the checkpointed 64-bit register
43711 The @samp{org.gnu.gdb.power.htm.altivec} feature is optional. It
43712 should contain the checkpointed altivec registers @samp{cvr0} through
43713 @samp{cvr31}, all 128-bit wide. It should also contain the
43714 checkpointed registers @samp{cvscr} and @samp{cvrsave}, both 32-bit
43717 The @samp{org.gnu.gdb.power.htm.vsx} feature is optional. It should
43718 contain registers @samp{cvs0h} through @samp{cvs31h}. @value{GDBN}
43719 will combine these registers with the checkpointed floating point
43720 registers (@samp{cf0} through @samp{cf31}) and the checkpointed
43721 altivec registers (@samp{cvr0} through @samp{cvr31}) to present the
43722 128-bit wide checkpointed vector-scalar registers @samp{cvs0} through
43723 @samp{cvs63}. Therefore, this feature requires both
43724 @samp{org.gnu.gdb.power.htm.altivec} and
43725 @samp{org.gnu.gdb.power.htm.fpu}.
43727 The @samp{org.gnu.gdb.power.htm.ppr} feature is optional. It should
43728 contain the 64-bit checkpointed register @samp{cppr}.
43730 The @samp{org.gnu.gdb.power.htm.dscr} feature is optional. It should
43731 contain the 64-bit checkpointed register @samp{cdscr}.
43733 The @samp{org.gnu.gdb.power.htm.tar} feature is optional. It should
43734 contain the 64-bit checkpointed register @samp{ctar}.
43737 @node RISC-V Features
43738 @subsection RISC-V Features
43739 @cindex target descriptions, RISC-V Features
43741 The @samp{org.gnu.gdb.riscv.cpu} feature is required for RISC-V
43742 targets. It should contain the registers @samp{x0} through
43743 @samp{x31}, and @samp{pc}. Either the architectural names (@samp{x0},
43744 @samp{x1}, etc) can be used, or the ABI names (@samp{zero}, @samp{ra},
43747 The @samp{org.gnu.gdb.riscv.fpu} feature is optional. If present, it
43748 should contain registers @samp{f0} through @samp{f31}, @samp{fflags},
43749 @samp{frm}, and @samp{fcsr}. As with the cpu feature, either the
43750 architectural register names, or the ABI names can be used.
43752 The @samp{org.gnu.gdb.riscv.virtual} feature is optional. If present,
43753 it should contain registers that are not backed by real registers on
43754 the target, but are instead virtual, where the register value is
43755 derived from other target state. In many ways these are like
43756 @value{GDBN}s pseudo-registers, except implemented by the target.
43757 Currently the only register expected in this set is the one byte
43758 @samp{priv} register that contains the target's privilege level in the
43759 least significant two bits.
43761 The @samp{org.gnu.gdb.riscv.csr} feature is optional. If present, it
43762 should contain all of the target's standard CSRs. Standard CSRs are
43763 those defined in the RISC-V specification documents. There is some
43764 overlap between this feature and the fpu feature; the @samp{fflags},
43765 @samp{frm}, and @samp{fcsr} registers could be in either feature. The
43766 expectation is that these registers will be in the fpu feature if the
43767 target has floating point hardware, but can be moved into the csr
43768 feature if the target has the floating point control registers, but no
43769 other floating point hardware.
43771 @node S/390 and System z Features
43772 @subsection S/390 and System z Features
43773 @cindex target descriptions, S/390 features
43774 @cindex target descriptions, System z features
43776 The @samp{org.gnu.gdb.s390.core} feature is required for S/390 and
43777 System z targets. It should contain the PSW and the 16 general
43778 registers. In particular, System z targets should provide the 64-bit
43779 registers @samp{pswm}, @samp{pswa}, and @samp{r0} through @samp{r15}.
43780 S/390 targets should provide the 32-bit versions of these registers.
43781 A System z target that runs in 31-bit addressing mode should provide
43782 32-bit versions of @samp{pswm} and @samp{pswa}, as well as the general
43783 register's upper halves @samp{r0h} through @samp{r15h}, and their
43784 lower halves @samp{r0l} through @samp{r15l}.
43786 The @samp{org.gnu.gdb.s390.fpr} feature is required. It should
43787 contain the 64-bit registers @samp{f0} through @samp{f15}, and
43790 The @samp{org.gnu.gdb.s390.acr} feature is required. It should
43791 contain the 32-bit registers @samp{acr0} through @samp{acr15}.
43793 The @samp{org.gnu.gdb.s390.linux} feature is optional. It should
43794 contain the register @samp{orig_r2}, which is 64-bit wide on System z
43795 targets and 32-bit otherwise. In addition, the feature may contain
43796 the @samp{last_break} register, whose width depends on the addressing
43797 mode, as well as the @samp{system_call} register, which is always
43800 The @samp{org.gnu.gdb.s390.tdb} feature is optional. It should
43801 contain the 64-bit registers @samp{tdb0}, @samp{tac}, @samp{tct},
43802 @samp{atia}, and @samp{tr0} through @samp{tr15}.
43804 The @samp{org.gnu.gdb.s390.vx} feature is optional. It should contain
43805 64-bit wide registers @samp{v0l} through @samp{v15l}, which will be
43806 combined by @value{GDBN} with the floating point registers @samp{f0}
43807 through @samp{f15} to present the 128-bit wide vector registers
43808 @samp{v0} through @samp{v15}. In addition, this feature should
43809 contain the 128-bit wide vector registers @samp{v16} through
43812 The @samp{org.gnu.gdb.s390.gs} feature is optional. It should contain
43813 the 64-bit wide guarded-storage-control registers @samp{gsd},
43814 @samp{gssm}, and @samp{gsepla}.
43816 The @samp{org.gnu.gdb.s390.gsbc} feature is optional. It should contain
43817 the 64-bit wide guarded-storage broadcast control registers
43818 @samp{bc_gsd}, @samp{bc_gssm}, and @samp{bc_gsepla}.
43820 @node Sparc Features
43821 @subsection Sparc Features
43822 @cindex target descriptions, sparc32 features
43823 @cindex target descriptions, sparc64 features
43824 The @samp{org.gnu.gdb.sparc.cpu} feature is required for sparc32/sparc64
43825 targets. It should describe the following registers:
43829 @samp{g0} through @samp{g7}
43831 @samp{o0} through @samp{o7}
43833 @samp{l0} through @samp{l7}
43835 @samp{i0} through @samp{i7}
43838 They may be 32-bit or 64-bit depending on the target.
43840 Also the @samp{org.gnu.gdb.sparc.fpu} feature is required for sparc32/sparc64
43841 targets. It should describe the following registers:
43845 @samp{f0} through @samp{f31}
43847 @samp{f32} through @samp{f62} for sparc64
43850 The @samp{org.gnu.gdb.sparc.cp0} feature is required for sparc32/sparc64
43851 targets. It should describe the following registers:
43855 @samp{y}, @samp{psr}, @samp{wim}, @samp{tbr}, @samp{pc}, @samp{npc},
43856 @samp{fsr}, and @samp{csr} for sparc32
43858 @samp{pc}, @samp{npc}, @samp{state}, @samp{fsr}, @samp{fprs}, and @samp{y}
43862 @node TIC6x Features
43863 @subsection TMS320C6x Features
43864 @cindex target descriptions, TIC6x features
43865 @cindex target descriptions, TMS320C6x features
43866 The @samp{org.gnu.gdb.tic6x.core} feature is required for TMS320C6x
43867 targets. It should contain registers @samp{A0} through @samp{A15},
43868 registers @samp{B0} through @samp{B15}, @samp{CSR} and @samp{PC}.
43870 The @samp{org.gnu.gdb.tic6x.gp} feature is optional. It should
43871 contain registers @samp{A16} through @samp{A31} and @samp{B16}
43872 through @samp{B31}.
43874 The @samp{org.gnu.gdb.tic6x.c6xp} feature is optional. It should
43875 contain registers @samp{TSR}, @samp{ILC} and @samp{RILC}.
43877 @node Operating System Information
43878 @appendix Operating System Information
43879 @cindex operating system information
43885 Users of @value{GDBN} often wish to obtain information about the state of
43886 the operating system running on the target---for example the list of
43887 processes, or the list of open files. This section describes the
43888 mechanism that makes it possible. This mechanism is similar to the
43889 target features mechanism (@pxref{Target Descriptions}), but focuses
43890 on a different aspect of target.
43892 Operating system information is retrived from the target via the
43893 remote protocol, using @samp{qXfer} requests (@pxref{qXfer osdata
43894 read}). The object name in the request should be @samp{osdata}, and
43895 the @var{annex} identifies the data to be fetched.
43898 @appendixsection Process list
43899 @cindex operating system information, process list
43901 When requesting the process list, the @var{annex} field in the
43902 @samp{qXfer} request should be @samp{processes}. The returned data is
43903 an XML document. The formal syntax of this document is defined in
43904 @file{gdb/features/osdata.dtd}.
43906 An example document is:
43909 <?xml version="1.0"?>
43910 <!DOCTYPE target SYSTEM "osdata.dtd">
43911 <osdata type="processes">
43913 <column name="pid">1</column>
43914 <column name="user">root</column>
43915 <column name="command">/sbin/init</column>
43916 <column name="cores">1,2,3</column>
43921 Each item should include a column whose name is @samp{pid}. The value
43922 of that column should identify the process on the target. The
43923 @samp{user} and @samp{command} columns are optional, and will be
43924 displayed by @value{GDBN}. The @samp{cores} column, if present,
43925 should contain a comma-separated list of cores that this process
43926 is running on. Target may provide additional columns,
43927 which @value{GDBN} currently ignores.
43929 @node Trace File Format
43930 @appendix Trace File Format
43931 @cindex trace file format
43933 The trace file comes in three parts: a header, a textual description
43934 section, and a trace frame section with binary data.
43936 The header has the form @code{\x7fTRACE0\n}. The first byte is
43937 @code{0x7f} so as to indicate that the file contains binary data,
43938 while the @code{0} is a version number that may have different values
43941 The description section consists of multiple lines of @sc{ascii} text
43942 separated by newline characters (@code{0xa}). The lines may include a
43943 variety of optional descriptive or context-setting information, such
43944 as tracepoint definitions or register set size. @value{GDBN} will
43945 ignore any line that it does not recognize. An empty line marks the end
43950 Specifies the size of a register block in bytes. This is equal to the
43951 size of a @code{g} packet payload in the remote protocol. @var{size}
43952 is an ascii decimal number. There should be only one such line in
43953 a single trace file.
43955 @item status @var{status}
43956 Trace status. @var{status} has the same format as a @code{qTStatus}
43957 remote packet reply. There should be only one such line in a single trace
43960 @item tp @var{payload}
43961 Tracepoint definition. The @var{payload} has the same format as
43962 @code{qTfP}/@code{qTsP} remote packet reply payload. A single tracepoint
43963 may take multiple lines of definition, corresponding to the multiple
43966 @item tsv @var{payload}
43967 Trace state variable definition. The @var{payload} has the same format as
43968 @code{qTfV}/@code{qTsV} remote packet reply payload. A single variable
43969 may take multiple lines of definition, corresponding to the multiple
43972 @item tdesc @var{payload}
43973 Target description in XML format. The @var{payload} is a single line of
43974 the XML file. All such lines should be concatenated together to get
43975 the original XML file. This file is in the same format as @code{qXfer}
43976 @code{features} payload, and corresponds to the main @code{target.xml}
43977 file. Includes are not allowed.
43981 The trace frame section consists of a number of consecutive frames.
43982 Each frame begins with a two-byte tracepoint number, followed by a
43983 four-byte size giving the amount of data in the frame. The data in
43984 the frame consists of a number of blocks, each introduced by a
43985 character indicating its type (at least register, memory, and trace
43986 state variable). The data in this section is raw binary, not a
43987 hexadecimal or other encoding; its endianness matches the target's
43990 @c FIXME bi-arch may require endianness/arch info in description section
43993 @item R @var{bytes}
43994 Register block. The number and ordering of bytes matches that of a
43995 @code{g} packet in the remote protocol. Note that these are the
43996 actual bytes, in target order, not a hexadecimal encoding.
43998 @item M @var{address} @var{length} @var{bytes}...
43999 Memory block. This is a contiguous block of memory, at the 8-byte
44000 address @var{address}, with a 2-byte length @var{length}, followed by
44001 @var{length} bytes.
44003 @item V @var{number} @var{value}
44004 Trace state variable block. This records the 8-byte signed value
44005 @var{value} of trace state variable numbered @var{number}.
44009 Future enhancements of the trace file format may include additional types
44012 @node Index Section Format
44013 @appendix @code{.gdb_index} section format
44014 @cindex .gdb_index section format
44015 @cindex index section format
44017 This section documents the index section that is created by @code{save
44018 gdb-index} (@pxref{Index Files}). The index section is
44019 DWARF-specific; some knowledge of DWARF is assumed in this
44022 The mapped index file format is designed to be directly
44023 @code{mmap}able on any architecture. In most cases, a datum is
44024 represented using a little-endian 32-bit integer value, called an
44025 @code{offset_type}. Big endian machines must byte-swap the values
44026 before using them. Exceptions to this rule are noted. The data is
44027 laid out such that alignment is always respected.
44029 A mapped index consists of several areas, laid out in order.
44033 The file header. This is a sequence of values, of @code{offset_type}
44034 unless otherwise noted:
44038 The version number, currently 8. Versions 1, 2 and 3 are obsolete.
44039 Version 4 uses a different hashing function from versions 5 and 6.
44040 Version 6 includes symbols for inlined functions, whereas versions 4
44041 and 5 do not. Version 7 adds attributes to the CU indices in the
44042 symbol table. Version 8 specifies that symbols from DWARF type units
44043 (@samp{DW_TAG_type_unit}) refer to the type unit's symbol table and not the
44044 compilation unit (@samp{DW_TAG_comp_unit}) using the type.
44046 @value{GDBN} will only read version 4, 5, or 6 indices
44047 by specifying @code{set use-deprecated-index-sections on}.
44048 GDB has a workaround for potentially broken version 7 indices so it is
44049 currently not flagged as deprecated.
44052 The offset, from the start of the file, of the CU list.
44055 The offset, from the start of the file, of the types CU list. Note
44056 that this area can be empty, in which case this offset will be equal
44057 to the next offset.
44060 The offset, from the start of the file, of the address area.
44063 The offset, from the start of the file, of the symbol table.
44066 The offset, from the start of the file, of the constant pool.
44070 The CU list. This is a sequence of pairs of 64-bit little-endian
44071 values, sorted by the CU offset. The first element in each pair is
44072 the offset of a CU in the @code{.debug_info} section. The second
44073 element in each pair is the length of that CU. References to a CU
44074 elsewhere in the map are done using a CU index, which is just the
44075 0-based index into this table. Note that if there are type CUs, then
44076 conceptually CUs and type CUs form a single list for the purposes of
44080 The types CU list. This is a sequence of triplets of 64-bit
44081 little-endian values. In a triplet, the first value is the CU offset,
44082 the second value is the type offset in the CU, and the third value is
44083 the type signature. The types CU list is not sorted.
44086 The address area. The address area consists of a sequence of address
44087 entries. Each address entry has three elements:
44091 The low address. This is a 64-bit little-endian value.
44094 The high address. This is a 64-bit little-endian value. Like
44095 @code{DW_AT_high_pc}, the value is one byte beyond the end.
44098 The CU index. This is an @code{offset_type} value.
44102 The symbol table. This is an open-addressed hash table. The size of
44103 the hash table is always a power of 2.
44105 Each slot in the hash table consists of a pair of @code{offset_type}
44106 values. The first value is the offset of the symbol's name in the
44107 constant pool. The second value is the offset of the CU vector in the
44110 If both values are 0, then this slot in the hash table is empty. This
44111 is ok because while 0 is a valid constant pool index, it cannot be a
44112 valid index for both a string and a CU vector.
44114 The hash value for a table entry is computed by applying an
44115 iterative hash function to the symbol's name. Starting with an
44116 initial value of @code{r = 0}, each (unsigned) character @samp{c} in
44117 the string is incorporated into the hash using the formula depending on the
44122 The formula is @code{r = r * 67 + c - 113}.
44124 @item Versions 5 to 7
44125 The formula is @code{r = r * 67 + tolower (c) - 113}.
44128 The terminating @samp{\0} is not incorporated into the hash.
44130 The step size used in the hash table is computed via
44131 @code{((hash * 17) & (size - 1)) | 1}, where @samp{hash} is the hash
44132 value, and @samp{size} is the size of the hash table. The step size
44133 is used to find the next candidate slot when handling a hash
44136 The names of C@t{++} symbols in the hash table are canonicalized. We
44137 don't currently have a simple description of the canonicalization
44138 algorithm; if you intend to create new index sections, you must read
44142 The constant pool. This is simply a bunch of bytes. It is organized
44143 so that alignment is correct: CU vectors are stored first, followed by
44146 A CU vector in the constant pool is a sequence of @code{offset_type}
44147 values. The first value is the number of CU indices in the vector.
44148 Each subsequent value is the index and symbol attributes of a CU in
44149 the CU list. This element in the hash table is used to indicate which
44150 CUs define the symbol and how the symbol is used.
44151 See below for the format of each CU index+attributes entry.
44153 A string in the constant pool is zero-terminated.
44156 Attributes were added to CU index values in @code{.gdb_index} version 7.
44157 If a symbol has multiple uses within a CU then there is one
44158 CU index+attributes value for each use.
44160 The format of each CU index+attributes entry is as follows
44166 This is the index of the CU in the CU list.
44168 These bits are reserved for future purposes and must be zero.
44170 The kind of the symbol in the CU.
44174 This value is reserved and should not be used.
44175 By reserving zero the full @code{offset_type} value is backwards compatible
44176 with previous versions of the index.
44178 The symbol is a type.
44180 The symbol is a variable or an enum value.
44182 The symbol is a function.
44184 Any other kind of symbol.
44186 These values are reserved.
44190 This bit is zero if the value is global and one if it is static.
44192 The determination of whether a symbol is global or static is complicated.
44193 The authorative reference is the file @file{dwarf2read.c} in
44194 @value{GDBN} sources.
44198 This pseudo-code describes the computation of a symbol's kind and
44199 global/static attributes in the index.
44202 is_external = get_attribute (die, DW_AT_external);
44203 language = get_attribute (cu_die, DW_AT_language);
44206 case DW_TAG_typedef:
44207 case DW_TAG_base_type:
44208 case DW_TAG_subrange_type:
44212 case DW_TAG_enumerator:
44214 is_static = language != CPLUS;
44216 case DW_TAG_subprogram:
44218 is_static = ! (is_external || language == ADA);
44220 case DW_TAG_constant:
44222 is_static = ! is_external;
44224 case DW_TAG_variable:
44226 is_static = ! is_external;
44228 case DW_TAG_namespace:
44232 case DW_TAG_class_type:
44233 case DW_TAG_interface_type:
44234 case DW_TAG_structure_type:
44235 case DW_TAG_union_type:
44236 case DW_TAG_enumeration_type:
44238 is_static = language != CPLUS;
44246 @appendix Manual pages
44250 * gdb man:: The GNU Debugger man page
44251 * gdbserver man:: Remote Server for the GNU Debugger man page
44252 * gcore man:: Generate a core file of a running program
44253 * gdbinit man:: gdbinit scripts
44254 * gdb-add-index man:: Add index files to speed up GDB
44260 @c man title gdb The GNU Debugger
44262 @c man begin SYNOPSIS gdb
44263 gdb [@option{-help}] [@option{-nh}] [@option{-nx}] [@option{-q}]
44264 [@option{-batch}] [@option{-cd=}@var{dir}] [@option{-f}]
44265 [@option{-b}@w{ }@var{bps}]
44266 [@option{-tty=}@var{dev}] [@option{-s} @var{symfile}]
44267 [@option{-e}@w{ }@var{prog}] [@option{-se}@w{ }@var{prog}]
44268 [@option{-c}@w{ }@var{core}] [@option{-p}@w{ }@var{procID}]
44269 [@option{-x}@w{ }@var{cmds}] [@option{-d}@w{ }@var{dir}]
44270 [@var{prog}|@var{prog} @var{procID}|@var{prog} @var{core}]
44273 @c man begin DESCRIPTION gdb
44274 The purpose of a debugger such as @value{GDBN} is to allow you to see what is
44275 going on ``inside'' another program while it executes -- or what another
44276 program was doing at the moment it crashed.
44278 @value{GDBN} can do four main kinds of things (plus other things in support of
44279 these) to help you catch bugs in the act:
44283 Start your program, specifying anything that might affect its behavior.
44286 Make your program stop on specified conditions.
44289 Examine what has happened, when your program has stopped.
44292 Change things in your program, so you can experiment with correcting the
44293 effects of one bug and go on to learn about another.
44296 You can use @value{GDBN} to debug programs written in C, C@t{++}, Fortran and
44299 @value{GDBN} is invoked with the shell command @code{gdb}. Once started, it reads
44300 commands from the terminal until you tell it to exit with the @value{GDBN}
44301 command @code{quit}. You can get online help from @value{GDBN} itself
44302 by using the command @code{help}.
44304 You can run @code{gdb} with no arguments or options; but the most
44305 usual way to start @value{GDBN} is with one argument or two, specifying an
44306 executable program as the argument:
44312 You can also start with both an executable program and a core file specified:
44318 You can, instead, specify a process ID as a second argument, if you want
44319 to debug a running process:
44327 would attach @value{GDBN} to process @code{1234} (unless you also have a file
44328 named @file{1234}; @value{GDBN} does check for a core file first).
44329 With option @option{-p} you can omit the @var{program} filename.
44331 Here are some of the most frequently needed @value{GDBN} commands:
44333 @c pod2man highlights the right hand side of the @item lines.
44335 @item break [@var{file}:]@var{function}
44336 Set a breakpoint at @var{function} (in @var{file}).
44338 @item run [@var{arglist}]
44339 Start your program (with @var{arglist}, if specified).
44342 Backtrace: display the program stack.
44344 @item print @var{expr}
44345 Display the value of an expression.
44348 Continue running your program (after stopping, e.g. at a breakpoint).
44351 Execute next program line (after stopping); step @emph{over} any
44352 function calls in the line.
44354 @item edit [@var{file}:]@var{function}
44355 look at the program line where it is presently stopped.
44357 @item list [@var{file}:]@var{function}
44358 type the text of the program in the vicinity of where it is presently stopped.
44361 Execute next program line (after stopping); step @emph{into} any
44362 function calls in the line.
44364 @item help [@var{name}]
44365 Show information about @value{GDBN} command @var{name}, or general information
44366 about using @value{GDBN}.
44369 Exit from @value{GDBN}.
44373 For full details on @value{GDBN},
44374 see @cite{Using GDB: A Guide to the GNU Source-Level Debugger},
44375 by Richard M. Stallman and Roland H. Pesch. The same text is available online
44376 as the @code{gdb} entry in the @code{info} program.
44380 @c man begin OPTIONS gdb
44381 Any arguments other than options specify an executable
44382 file and core file (or process ID); that is, the first argument
44383 encountered with no
44384 associated option flag is equivalent to a @option{-se} option, and the second,
44385 if any, is equivalent to a @option{-c} option if it's the name of a file.
44387 both long and short forms; both are shown here. The long forms are also
44388 recognized if you truncate them, so long as enough of the option is
44389 present to be unambiguous. (If you prefer, you can flag option
44390 arguments with @option{+} rather than @option{-}, though we illustrate the
44391 more usual convention.)
44393 All the options and command line arguments you give are processed
44394 in sequential order. The order makes a difference when the @option{-x}
44400 List all options, with brief explanations.
44402 @item -symbols=@var{file}
44403 @itemx -s @var{file}
44404 Read symbol table from file @var{file}.
44407 Enable writing into executable and core files.
44409 @item -exec=@var{file}
44410 @itemx -e @var{file}
44411 Use file @var{file} as the executable file to execute when
44412 appropriate, and for examining pure data in conjunction with a core
44415 @item -se=@var{file}
44416 Read symbol table from file @var{file} and use it as the executable
44419 @item -core=@var{file}
44420 @itemx -c @var{file}
44421 Use file @var{file} as a core dump to examine.
44423 @item -command=@var{file}
44424 @itemx -x @var{file}
44425 Execute @value{GDBN} commands from file @var{file}.
44427 @item -ex @var{command}
44428 Execute given @value{GDBN} @var{command}.
44430 @item -directory=@var{directory}
44431 @itemx -d @var{directory}
44432 Add @var{directory} to the path to search for source files.
44435 Do not execute commands from @file{~/.gdbinit}.
44439 Do not execute commands from any @file{.gdbinit} initialization files.
44443 ``Quiet''. Do not print the introductory and copyright messages. These
44444 messages are also suppressed in batch mode.
44447 Run in batch mode. Exit with status @code{0} after processing all the command
44448 files specified with @option{-x} (and @file{.gdbinit}, if not inhibited).
44449 Exit with nonzero status if an error occurs in executing the @value{GDBN}
44450 commands in the command files.
44452 Batch mode may be useful for running @value{GDBN} as a filter, for example to
44453 download and run a program on another computer; in order to make this
44454 more useful, the message
44457 Program exited normally.
44461 (which is ordinarily issued whenever a program running under @value{GDBN} control
44462 terminates) is not issued when running in batch mode.
44464 @item -cd=@var{directory}
44465 Run @value{GDBN} using @var{directory} as its working directory,
44466 instead of the current directory.
44470 Emacs sets this option when it runs @value{GDBN} as a subprocess. It tells
44471 @value{GDBN} to output the full file name and line number in a standard,
44472 recognizable fashion each time a stack frame is displayed (which
44473 includes each time the program stops). This recognizable format looks
44474 like two @samp{\032} characters, followed by the file name, line number
44475 and character position separated by colons, and a newline. The
44476 Emacs-to-@value{GDBN} interface program uses the two @samp{\032}
44477 characters as a signal to display the source code for the frame.
44480 Set the line speed (baud rate or bits per second) of any serial
44481 interface used by @value{GDBN} for remote debugging.
44483 @item -tty=@var{device}
44484 Run using @var{device} for your program's standard input and output.
44488 @c man begin SEEALSO gdb
44490 The full documentation for @value{GDBN} is maintained as a Texinfo manual.
44491 If the @code{info} and @code{gdb} programs and @value{GDBN}'s Texinfo
44492 documentation are properly installed at your site, the command
44499 should give you access to the complete manual.
44501 @cite{Using GDB: A Guide to the GNU Source-Level Debugger},
44502 Richard M. Stallman and Roland H. Pesch, July 1991.
44506 @node gdbserver man
44507 @heading gdbserver man
44509 @c man title gdbserver Remote Server for the GNU Debugger
44511 @c man begin SYNOPSIS gdbserver
44512 gdbserver @var{comm} @var{prog} [@var{args}@dots{}]
44514 gdbserver --attach @var{comm} @var{pid}
44516 gdbserver --multi @var{comm}
44520 @c man begin DESCRIPTION gdbserver
44521 @command{gdbserver} is a program that allows you to run @value{GDBN} on a different machine
44522 than the one which is running the program being debugged.
44525 @subheading Usage (server (target) side)
44528 Usage (server (target) side):
44531 First, you need to have a copy of the program you want to debug put onto
44532 the target system. The program can be stripped to save space if needed, as
44533 @command{gdbserver} doesn't care about symbols. All symbol handling is taken care of by
44534 the @value{GDBN} running on the host system.
44536 To use the server, you log on to the target system, and run the @command{gdbserver}
44537 program. You must tell it (a) how to communicate with @value{GDBN}, (b) the name of
44538 your program, and (c) its arguments. The general syntax is:
44541 target> gdbserver @var{comm} @var{program} [@var{args} ...]
44544 For example, using a serial port, you might say:
44548 @c @file would wrap it as F</dev/com1>.
44549 target> gdbserver /dev/com1 emacs foo.txt
44552 target> gdbserver @file{/dev/com1} emacs foo.txt
44556 This tells @command{gdbserver} to debug emacs with an argument of foo.txt, and
44557 to communicate with @value{GDBN} via @file{/dev/com1}. @command{gdbserver} now
44558 waits patiently for the host @value{GDBN} to communicate with it.
44560 To use a TCP connection, you could say:
44563 target> gdbserver host:2345 emacs foo.txt
44566 This says pretty much the same thing as the last example, except that we are
44567 going to communicate with the @code{host} @value{GDBN} via TCP. The @code{host:2345} argument means
44568 that we are expecting to see a TCP connection from @code{host} to local TCP port
44569 2345. (Currently, the @code{host} part is ignored.) You can choose any number you
44570 want for the port number as long as it does not conflict with any existing TCP
44571 ports on the target system. This same port number must be used in the host
44572 @value{GDBN}s @code{target remote} command, which will be described shortly. Note that if
44573 you chose a port number that conflicts with another service, @command{gdbserver} will
44574 print an error message and exit.
44576 @command{gdbserver} can also attach to running programs.
44577 This is accomplished via the @option{--attach} argument. The syntax is:
44580 target> gdbserver --attach @var{comm} @var{pid}
44583 @var{pid} is the process ID of a currently running process. It isn't
44584 necessary to point @command{gdbserver} at a binary for the running process.
44586 To start @code{gdbserver} without supplying an initial command to run
44587 or process ID to attach, use the @option{--multi} command line option.
44588 In such case you should connect using @kbd{target extended-remote} to start
44589 the program you want to debug.
44592 target> gdbserver --multi @var{comm}
44596 @subheading Usage (host side)
44602 You need an unstripped copy of the target program on your host system, since
44603 @value{GDBN} needs to examine its symbol tables and such. Start up @value{GDBN} as you normally
44604 would, with the target program as the first argument. (You may need to use the
44605 @option{--baud} option if the serial line is running at anything except 9600 baud.)
44606 That is @code{gdb TARGET-PROG}, or @code{gdb --baud BAUD TARGET-PROG}. After that, the only
44607 new command you need to know about is @code{target remote}
44608 (or @code{target extended-remote}). Its argument is either
44609 a device name (usually a serial device, like @file{/dev/ttyb}), or a @code{HOST:PORT}
44610 descriptor. For example:
44614 @c @file would wrap it as F</dev/ttyb>.
44615 (gdb) target remote /dev/ttyb
44618 (gdb) target remote @file{/dev/ttyb}
44623 communicates with the server via serial line @file{/dev/ttyb}, and:
44626 (gdb) target remote the-target:2345
44630 communicates via a TCP connection to port 2345 on host `the-target', where
44631 you previously started up @command{gdbserver} with the same port number. Note that for
44632 TCP connections, you must start up @command{gdbserver} prior to using the `target remote'
44633 command, otherwise you may get an error that looks something like
44634 `Connection refused'.
44636 @command{gdbserver} can also debug multiple inferiors at once,
44639 the @value{GDBN} manual in node @code{Inferiors and Programs}
44640 -- shell command @code{info -f gdb -n 'Inferiors and Programs'}.
44643 @ref{Inferiors and Programs}.
44645 In such case use the @code{extended-remote} @value{GDBN} command variant:
44648 (gdb) target extended-remote the-target:2345
44651 The @command{gdbserver} option @option{--multi} may or may not be used in such
44655 @c man begin OPTIONS gdbserver
44656 There are three different modes for invoking @command{gdbserver}:
44661 Debug a specific program specified by its program name:
44664 gdbserver @var{comm} @var{prog} [@var{args}@dots{}]
44667 The @var{comm} parameter specifies how should the server communicate
44668 with @value{GDBN}; it is either a device name (to use a serial line),
44669 a TCP port number (@code{:1234}), or @code{-} or @code{stdio} to use
44670 stdin/stdout of @code{gdbserver}. Specify the name of the program to
44671 debug in @var{prog}. Any remaining arguments will be passed to the
44672 program verbatim. When the program exits, @value{GDBN} will close the
44673 connection, and @code{gdbserver} will exit.
44676 Debug a specific program by specifying the process ID of a running
44680 gdbserver --attach @var{comm} @var{pid}
44683 The @var{comm} parameter is as described above. Supply the process ID
44684 of a running program in @var{pid}; @value{GDBN} will do everything
44685 else. Like with the previous mode, when the process @var{pid} exits,
44686 @value{GDBN} will close the connection, and @code{gdbserver} will exit.
44689 Multi-process mode -- debug more than one program/process:
44692 gdbserver --multi @var{comm}
44695 In this mode, @value{GDBN} can instruct @command{gdbserver} which
44696 command(s) to run. Unlike the other 2 modes, @value{GDBN} will not
44697 close the connection when a process being debugged exits, so you can
44698 debug several processes in the same session.
44701 In each of the modes you may specify these options:
44706 List all options, with brief explanations.
44709 This option causes @command{gdbserver} to print its version number and exit.
44712 @command{gdbserver} will attach to a running program. The syntax is:
44715 target> gdbserver --attach @var{comm} @var{pid}
44718 @var{pid} is the process ID of a currently running process. It isn't
44719 necessary to point @command{gdbserver} at a binary for the running process.
44722 To start @code{gdbserver} without supplying an initial command to run
44723 or process ID to attach, use this command line option.
44724 Then you can connect using @kbd{target extended-remote} and start
44725 the program you want to debug. The syntax is:
44728 target> gdbserver --multi @var{comm}
44732 Instruct @code{gdbserver} to display extra status information about the debugging
44734 This option is intended for @code{gdbserver} development and for bug reports to
44737 @item --remote-debug
44738 Instruct @code{gdbserver} to display remote protocol debug output.
44739 This option is intended for @code{gdbserver} development and for bug reports to
44742 @item --debug-file=@var{filename}
44743 Instruct @code{gdbserver} to send any debug output to the given @var{filename}.
44744 This option is intended for @code{gdbserver} development and for bug reports to
44747 @item --debug-format=option1@r{[},option2,...@r{]}
44748 Instruct @code{gdbserver} to include extra information in each line
44749 of debugging output.
44750 @xref{Other Command-Line Arguments for gdbserver}.
44753 Specify a wrapper to launch programs
44754 for debugging. The option should be followed by the name of the
44755 wrapper, then any command-line arguments to pass to the wrapper, then
44756 @kbd{--} indicating the end of the wrapper arguments.
44759 By default, @command{gdbserver} keeps the listening TCP port open, so that
44760 additional connections are possible. However, if you start @code{gdbserver}
44761 with the @option{--once} option, it will stop listening for any further
44762 connection attempts after connecting to the first @value{GDBN} session.
44764 @c --disable-packet is not documented for users.
44766 @c --disable-randomization and --no-disable-randomization are superseded by
44767 @c QDisableRandomization.
44772 @c man begin SEEALSO gdbserver
44774 The full documentation for @value{GDBN} is maintained as a Texinfo manual.
44775 If the @code{info} and @code{gdb} programs and @value{GDBN}'s Texinfo
44776 documentation are properly installed at your site, the command
44782 should give you access to the complete manual.
44784 @cite{Using GDB: A Guide to the GNU Source-Level Debugger},
44785 Richard M. Stallman and Roland H. Pesch, July 1991.
44792 @c man title gcore Generate a core file of a running program
44795 @c man begin SYNOPSIS gcore
44796 gcore [-a] [-o @var{prefix}] @var{pid1} [@var{pid2}...@var{pidN}]
44800 @c man begin DESCRIPTION gcore
44801 Generate core dumps of one or more running programs with process IDs
44802 @var{pid1}, @var{pid2}, etc. A core file produced by @command{gcore}
44803 is equivalent to one produced by the kernel when the process crashes
44804 (and when @kbd{ulimit -c} was used to set up an appropriate core dump
44805 limit). However, unlike after a crash, after @command{gcore} finishes
44806 its job the program remains running without any change.
44809 @c man begin OPTIONS gcore
44812 Dump all memory mappings. The actual effect of this option depends on
44813 the Operating System. On @sc{gnu}/Linux, it will disable
44814 @code{use-coredump-filter} (@pxref{set use-coredump-filter}) and
44815 enable @code{dump-excluded-mappings} (@pxref{set
44816 dump-excluded-mappings}).
44818 @item -o @var{prefix}
44819 The optional argument @var{prefix} specifies the prefix to be used
44820 when composing the file names of the core dumps. The file name is
44821 composed as @file{@var{prefix}.@var{pid}}, where @var{pid} is the
44822 process ID of the running program being analyzed by @command{gcore}.
44823 If not specified, @var{prefix} defaults to @var{gcore}.
44827 @c man begin SEEALSO gcore
44829 The full documentation for @value{GDBN} is maintained as a Texinfo manual.
44830 If the @code{info} and @code{gdb} programs and @value{GDBN}'s Texinfo
44831 documentation are properly installed at your site, the command
44838 should give you access to the complete manual.
44840 @cite{Using GDB: A Guide to the GNU Source-Level Debugger},
44841 Richard M. Stallman and Roland H. Pesch, July 1991.
44848 @c man title gdbinit GDB initialization scripts
44851 @c man begin SYNOPSIS gdbinit
44852 @ifset SYSTEM_GDBINIT
44853 @value{SYSTEM_GDBINIT}
44862 @c man begin DESCRIPTION gdbinit
44863 These files contain @value{GDBN} commands to automatically execute during
44864 @value{GDBN} startup. The lines of contents are canned sequences of commands,
44867 the @value{GDBN} manual in node @code{Sequences}
44868 -- shell command @code{info -f gdb -n Sequences}.
44874 Please read more in
44876 the @value{GDBN} manual in node @code{Startup}
44877 -- shell command @code{info -f gdb -n Startup}.
44884 @ifset SYSTEM_GDBINIT
44885 @item @value{SYSTEM_GDBINIT}
44887 @ifclear SYSTEM_GDBINIT
44888 @item (not enabled with @code{--with-system-gdbinit} during compilation)
44890 System-wide initialization file. It is executed unless user specified
44891 @value{GDBN} option @code{-nx} or @code{-n}.
44894 the @value{GDBN} manual in node @code{System-wide configuration}
44895 -- shell command @code{info -f gdb -n 'System-wide configuration'}.
44898 @ref{System-wide configuration}.
44902 User initialization file. It is executed unless user specified
44903 @value{GDBN} options @code{-nx}, @code{-n} or @code{-nh}.
44906 Initialization file for current directory. It may need to be enabled with
44907 @value{GDBN} security command @code{set auto-load local-gdbinit}.
44910 the @value{GDBN} manual in node @code{Init File in the Current Directory}
44911 -- shell command @code{info -f gdb -n 'Init File in the Current Directory'}.
44914 @ref{Init File in the Current Directory}.
44919 @c man begin SEEALSO gdbinit
44921 gdb(1), @code{info -f gdb -n Startup}
44923 The full documentation for @value{GDBN} is maintained as a Texinfo manual.
44924 If the @code{info} and @code{gdb} programs and @value{GDBN}'s Texinfo
44925 documentation are properly installed at your site, the command
44931 should give you access to the complete manual.
44933 @cite{Using GDB: A Guide to the GNU Source-Level Debugger},
44934 Richard M. Stallman and Roland H. Pesch, July 1991.
44938 @node gdb-add-index man
44939 @heading gdb-add-index
44940 @pindex gdb-add-index
44941 @anchor{gdb-add-index}
44943 @c man title gdb-add-index Add index files to speed up GDB
44945 @c man begin SYNOPSIS gdb-add-index
44946 gdb-add-index @var{filename}
44949 @c man begin DESCRIPTION gdb-add-index
44950 When @value{GDBN} finds a symbol file, it scans the symbols in the
44951 file in order to construct an internal symbol table. This lets most
44952 @value{GDBN} operations work quickly--at the cost of a delay early on.
44953 For large programs, this delay can be quite lengthy, so @value{GDBN}
44954 provides a way to build an index, which speeds up startup.
44956 To determine whether a file contains such an index, use the command
44957 @kbd{readelf -S filename}: the index is stored in a section named
44958 @code{.gdb_index}. The index file can only be produced on systems
44959 which use ELF binaries and DWARF debug information (i.e., sections
44960 named @code{.debug_*}).
44962 @command{gdb-add-index} uses @value{GDBN} and @command{objdump} found
44963 in the @env{PATH} environment variable. If you want to use different
44964 versions of these programs, you can specify them through the
44965 @env{GDB} and @env{OBJDUMP} environment variables.
44969 the @value{GDBN} manual in node @code{Index Files}
44970 -- shell command @kbd{info -f gdb -n "Index Files"}.
44977 @c man begin SEEALSO gdb-add-index
44979 The full documentation for @value{GDBN} is maintained as a Texinfo manual.
44980 If the @code{info} and @code{gdb} programs and @value{GDBN}'s Texinfo
44981 documentation are properly installed at your site, the command
44987 should give you access to the complete manual.
44989 @cite{Using GDB: A Guide to the GNU Source-Level Debugger},
44990 Richard M. Stallman and Roland H. Pesch, July 1991.
44996 @node GNU Free Documentation License
44997 @appendix GNU Free Documentation License
45000 @node Concept Index
45001 @unnumbered Concept Index
45005 @node Command and Variable Index
45006 @unnumbered Command, Variable, and Function Index
45011 % I think something like @@colophon should be in texinfo. In the
45013 \long\def\colophon{\hbox to0pt{}\vfill
45014 \centerline{The body of this manual is set in}
45015 \centerline{\fontname\tenrm,}
45016 \centerline{with headings in {\bf\fontname\tenbf}}
45017 \centerline{and examples in {\tt\fontname\tentt}.}
45018 \centerline{{\it\fontname\tenit\/},}
45019 \centerline{{\bf\fontname\tenbf}, and}
45020 \centerline{{\sl\fontname\tensl\/}}
45021 \centerline{are used for emphasis.}\vfill}
45023 % Blame: doc@@cygnus.com, 1991.