1 \input texinfo @c -*-texinfo-*-
2 @c Copyright (C) 1988, 1989, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1998,
3 @c 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
4 @c Free Software Foundation, Inc.
7 @c makeinfo ignores cmds prev to setfilename, so its arg cannot make use
8 @c of @set vars. However, you can override filename with makeinfo -o.
13 @settitle Debugging with @value{GDBN}
14 @setchapternewpage odd
26 @c readline appendices use @vindex, @findex and @ftable,
27 @c annotate.texi and gdbmi use @findex.
31 @c !!set GDB manual's edition---not the same as GDB version!
32 @c This is updated by GNU Press.
35 @c !!set GDB edit command default editor
38 @c THIS MANUAL REQUIRES TEXINFO 4.0 OR LATER.
40 @c This is a dir.info fragment to support semi-automated addition of
41 @c manuals to an info tree.
42 @dircategory Software development
44 * Gdb: (gdb). The GNU debugger.
48 Copyright @copyright{} 1988, 1989, 1990, 1991, 1992, 1993, 1994, 1995, 1996,
49 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
50 Free Software Foundation, Inc.
52 Permission is granted to copy, distribute and/or modify this document
53 under the terms of the GNU Free Documentation License, Version 1.1 or
54 any later version published by the Free Software Foundation; with the
55 Invariant Sections being ``Free Software'' and ``Free Software Needs
56 Free Documentation'', with the Front-Cover Texts being ``A GNU Manual,''
57 and with the Back-Cover Texts as in (a) below.
59 (a) The FSF's Back-Cover Text is: ``You are free to copy and modify
60 this GNU Manual. Buying copies from GNU Press supports the FSF in
61 developing GNU and promoting software freedom.''
65 This file documents the @sc{gnu} debugger @value{GDBN}.
67 This is the @value{EDITION} Edition, of @cite{Debugging with
68 @value{GDBN}: the @sc{gnu} Source-Level Debugger} for @value{GDBN}
69 @ifset VERSION_PACKAGE
70 @value{VERSION_PACKAGE}
72 Version @value{GDBVN}.
78 @title Debugging with @value{GDBN}
79 @subtitle The @sc{gnu} Source-Level Debugger
81 @subtitle @value{EDITION} Edition, for @value{GDBN} version @value{GDBVN}
82 @ifset VERSION_PACKAGE
84 @subtitle @value{VERSION_PACKAGE}
86 @author Richard Stallman, Roland Pesch, Stan Shebs, et al.
90 \hfill (Send bugs and comments on @value{GDBN} to @value{BUGURL}.)\par
91 \hfill {\it Debugging with @value{GDBN}}\par
92 \hfill \TeX{}info \texinfoversion\par
96 @vskip 0pt plus 1filll
97 Published by the Free Software Foundation @*
98 51 Franklin Street, Fifth Floor,
99 Boston, MA 02110-1301, USA@*
100 ISBN 1-882114-77-9 @*
104 This edition of the GDB manual is dedicated to the memory of Fred
105 Fish. Fred was a long-standing contributor to GDB and to Free
106 software in general. We will miss him.
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) 1988-2010 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.
164 * GDB Bugs:: Reporting bugs in @value{GDBN}
166 * Command Line Editing:: Command Line Editing
167 * Using History Interactively:: Using History Interactively
168 * Formatting Documentation:: How to format and print @value{GDBN} documentation
169 * Installing GDB:: Installing GDB
170 * Maintenance Commands:: Maintenance Commands
171 * Remote Protocol:: GDB Remote Serial Protocol
172 * Agent Expressions:: The GDB Agent Expression Mechanism
173 * Target Descriptions:: How targets can describe themselves to
175 * Operating System Information:: Getting additional information from
177 * Trace File Format:: GDB trace file format
178 * Copying:: GNU General Public License says
179 how you can copy and share GDB
180 * GNU Free Documentation License:: The license for this documentation
189 @unnumbered Summary of @value{GDBN}
191 The purpose of a debugger such as @value{GDBN} is to allow you to see what is
192 going on ``inside'' another program while it executes---or what another
193 program was doing at the moment it crashed.
195 @value{GDBN} can do four main kinds of things (plus other things in support of
196 these) to help you catch bugs in the act:
200 Start your program, specifying anything that might affect its behavior.
203 Make your program stop on specified conditions.
206 Examine what has happened, when your program has stopped.
209 Change things in your program, so you can experiment with correcting the
210 effects of one bug and go on to learn about another.
213 You can use @value{GDBN} to debug programs written in C and C@t{++}.
214 For more information, see @ref{Supported Languages,,Supported Languages}.
215 For more information, see @ref{C,,C and C++}.
217 Support for D is partial. For information on D, see
221 Support for Modula-2 is partial. For information on Modula-2, see
222 @ref{Modula-2,,Modula-2}.
225 Debugging Pascal programs which use sets, subranges, file variables, or
226 nested functions does not currently work. @value{GDBN} does not support
227 entering expressions, printing values, or similar features using Pascal
231 @value{GDBN} can be used to debug programs written in Fortran, although
232 it may be necessary to refer to some variables with a trailing
235 @value{GDBN} can be used to debug programs written in Objective-C,
236 using either the Apple/NeXT or the GNU Objective-C runtime.
239 * Free Software:: Freely redistributable software
240 * Contributors:: Contributors to GDB
244 @unnumberedsec Free Software
246 @value{GDBN} is @dfn{free software}, protected by the @sc{gnu}
247 General Public License
248 (GPL). The GPL gives you the freedom to copy or adapt a licensed
249 program---but every person getting a copy also gets with it the
250 freedom to modify that copy (which means that they must get access to
251 the source code), and the freedom to distribute further copies.
252 Typical software companies use copyrights to limit your freedoms; the
253 Free Software Foundation uses the GPL to preserve these freedoms.
255 Fundamentally, the General Public License is a license which says that
256 you have these freedoms and that you cannot take these freedoms away
259 @unnumberedsec Free Software Needs Free Documentation
261 The biggest deficiency in the free software community today is not in
262 the software---it is the lack of good free documentation that we can
263 include with the free software. Many of our most important
264 programs do not come with free reference manuals and free introductory
265 texts. Documentation is an essential part of any software package;
266 when an important free software package does not come with a free
267 manual and a free tutorial, that is a major gap. We have many such
270 Consider Perl, for instance. The tutorial manuals that people
271 normally use are non-free. How did this come about? Because the
272 authors of those manuals published them with restrictive terms---no
273 copying, no modification, source files not available---which exclude
274 them from the free software world.
276 That wasn't the first time this sort of thing happened, and it was far
277 from the last. Many times we have heard a GNU user eagerly describe a
278 manual that he is writing, his intended contribution to the community,
279 only to learn that he had ruined everything by signing a publication
280 contract to make it non-free.
282 Free documentation, like free software, is a matter of freedom, not
283 price. The problem with the non-free manual is not that publishers
284 charge a price for printed copies---that in itself is fine. (The Free
285 Software Foundation sells printed copies of manuals, too.) The
286 problem is the restrictions on the use of the manual. Free manuals
287 are available in source code form, and give you permission to copy and
288 modify. Non-free manuals do not allow this.
290 The criteria of freedom for a free manual are roughly the same as for
291 free software. Redistribution (including the normal kinds of
292 commercial redistribution) must be permitted, so that the manual can
293 accompany every copy of the program, both on-line and on paper.
295 Permission for modification of the technical content is crucial too.
296 When people modify the software, adding or changing features, if they
297 are conscientious they will change the manual too---so they can
298 provide accurate and clear documentation for the modified program. A
299 manual that leaves you no choice but to write a new manual to document
300 a changed version of the program is not really available to our
303 Some kinds of limits on the way modification is handled are
304 acceptable. For example, requirements to preserve the original
305 author's copyright notice, the distribution terms, or the list of
306 authors, are ok. It is also no problem to require modified versions
307 to include notice that they were modified. Even entire sections that
308 may not be deleted or changed are acceptable, as long as they deal
309 with nontechnical topics (like this one). These kinds of restrictions
310 are acceptable because they don't obstruct the community's normal use
313 However, it must be possible to modify all the @emph{technical}
314 content of the manual, and then distribute the result in all the usual
315 media, through all the usual channels. Otherwise, the restrictions
316 obstruct the use of the manual, it is not free, and we need another
317 manual to replace it.
319 Please spread the word about this issue. Our community continues to
320 lose manuals to proprietary publishing. If we spread the word that
321 free software needs free reference manuals and free tutorials, perhaps
322 the next person who wants to contribute by writing documentation will
323 realize, before it is too late, that only free manuals contribute to
324 the free software community.
326 If you are writing documentation, please insist on publishing it under
327 the GNU Free Documentation License or another free documentation
328 license. Remember that this decision requires your approval---you
329 don't have to let the publisher decide. Some commercial publishers
330 will use a free license if you insist, but they will not propose the
331 option; it is up to you to raise the issue and say firmly that this is
332 what you want. If the publisher you are dealing with refuses, please
333 try other publishers. If you're not sure whether a proposed license
334 is free, write to @email{licensing@@gnu.org}.
336 You can encourage commercial publishers to sell more free, copylefted
337 manuals and tutorials by buying them, and particularly by buying
338 copies from the publishers that paid for their writing or for major
339 improvements. Meanwhile, try to avoid buying non-free documentation
340 at all. Check the distribution terms of a manual before you buy it,
341 and insist that whoever seeks your business must respect your freedom.
342 Check the history of the book, and try to reward the publishers that
343 have paid or pay the authors to work on it.
345 The Free Software Foundation maintains a list of free documentation
346 published by other publishers, at
347 @url{http://www.fsf.org/doc/other-free-books.html}.
350 @unnumberedsec Contributors to @value{GDBN}
352 Richard Stallman was the original author of @value{GDBN}, and of many
353 other @sc{gnu} programs. Many others have contributed to its
354 development. This section attempts to credit major contributors. One
355 of the virtues of free software is that everyone is free to contribute
356 to it; with regret, we cannot actually acknowledge everyone here. The
357 file @file{ChangeLog} in the @value{GDBN} distribution approximates a
358 blow-by-blow account.
360 Changes much prior to version 2.0 are lost in the mists of time.
363 @emph{Plea:} Additions to this section are particularly welcome. If you
364 or your friends (or enemies, to be evenhanded) have been unfairly
365 omitted from this list, we would like to add your names!
368 So that they may not regard their many labors as thankless, we
369 particularly thank those who shepherded @value{GDBN} through major
371 Andrew Cagney (releases 6.3, 6.2, 6.1, 6.0, 5.3, 5.2, 5.1 and 5.0);
372 Jim Blandy (release 4.18);
373 Jason Molenda (release 4.17);
374 Stan Shebs (release 4.14);
375 Fred Fish (releases 4.16, 4.15, 4.13, 4.12, 4.11, 4.10, and 4.9);
376 Stu Grossman and John Gilmore (releases 4.8, 4.7, 4.6, 4.5, and 4.4);
377 John Gilmore (releases 4.3, 4.2, 4.1, 4.0, and 3.9);
378 Jim Kingdon (releases 3.5, 3.4, and 3.3);
379 and Randy Smith (releases 3.2, 3.1, and 3.0).
381 Richard Stallman, assisted at various times by Peter TerMaat, Chris
382 Hanson, and Richard Mlynarik, handled releases through 2.8.
384 Michael Tiemann is the author of most of the @sc{gnu} C@t{++} support
385 in @value{GDBN}, with significant additional contributions from Per
386 Bothner and Daniel Berlin. James Clark wrote the @sc{gnu} C@t{++}
387 demangler. Early work on C@t{++} was by Peter TerMaat (who also did
388 much general update work leading to release 3.0).
390 @value{GDBN} uses the BFD subroutine library to examine multiple
391 object-file formats; BFD was a joint project of David V.
392 Henkel-Wallace, Rich Pixley, Steve Chamberlain, and John Gilmore.
394 David Johnson wrote the original COFF support; Pace Willison did
395 the original support for encapsulated COFF.
397 Brent Benson of Harris Computer Systems contributed DWARF 2 support.
399 Adam de Boor and Bradley Davis contributed the ISI Optimum V support.
400 Per Bothner, Noboyuki Hikichi, and Alessandro Forin contributed MIPS
402 Jean-Daniel Fekete contributed Sun 386i support.
403 Chris Hanson improved the HP9000 support.
404 Noboyuki Hikichi and Tomoyuki Hasei contributed Sony/News OS 3 support.
405 David Johnson contributed Encore Umax support.
406 Jyrki Kuoppala contributed Altos 3068 support.
407 Jeff Law contributed HP PA and SOM support.
408 Keith Packard contributed NS32K support.
409 Doug Rabson contributed Acorn Risc Machine support.
410 Bob Rusk contributed Harris Nighthawk CX-UX support.
411 Chris Smith contributed Convex support (and Fortran debugging).
412 Jonathan Stone contributed Pyramid support.
413 Michael Tiemann contributed SPARC support.
414 Tim Tucker contributed support for the Gould NP1 and Gould Powernode.
415 Pace Willison contributed Intel 386 support.
416 Jay Vosburgh contributed Symmetry support.
417 Marko Mlinar contributed OpenRISC 1000 support.
419 Andreas Schwab contributed M68K @sc{gnu}/Linux support.
421 Rich Schaefer and Peter Schauer helped with support of SunOS shared
424 Jay Fenlason and Roland McGrath ensured that @value{GDBN} and GAS agree
425 about several machine instruction sets.
427 Patrick Duval, Ted Goldstein, Vikram Koka and Glenn Engel helped develop
428 remote debugging. Intel Corporation, Wind River Systems, AMD, and ARM
429 contributed remote debugging modules for the i960, VxWorks, A29K UDI,
430 and RDI targets, respectively.
432 Brian Fox is the author of the readline libraries providing
433 command-line editing and command history.
435 Andrew Beers of SUNY Buffalo wrote the language-switching code, the
436 Modula-2 support, and contributed the Languages chapter of this manual.
438 Fred Fish wrote most of the support for Unix System Vr4.
439 He also enhanced the command-completion support to cover C@t{++} overloaded
442 Hitachi America (now Renesas America), Ltd. sponsored the support for
443 H8/300, H8/500, and Super-H processors.
445 NEC sponsored the support for the v850, Vr4xxx, and Vr5xxx processors.
447 Mitsubishi (now Renesas) sponsored the support for D10V, D30V, and M32R/D
450 Toshiba sponsored the support for the TX39 Mips processor.
452 Matsushita sponsored the support for the MN10200 and MN10300 processors.
454 Fujitsu sponsored the support for SPARClite and FR30 processors.
456 Kung Hsu, Jeff Law, and Rick Sladkey added support for hardware
459 Michael Snyder added support for tracepoints.
461 Stu Grossman wrote gdbserver.
463 Jim Kingdon, Peter Schauer, Ian Taylor, and Stu Grossman made
464 nearly innumerable bug fixes and cleanups throughout @value{GDBN}.
466 The following people at the Hewlett-Packard Company contributed
467 support for the PA-RISC 2.0 architecture, HP-UX 10.20, 10.30, and 11.0
468 (narrow mode), HP's implementation of kernel threads, HP's aC@t{++}
469 compiler, and the Text User Interface (nee Terminal User Interface):
470 Ben Krepp, Richard Title, John Bishop, Susan Macchia, Kathy Mann,
471 Satish Pai, India Paul, Steve Rehrauer, and Elena Zannoni. Kim Haase
472 provided HP-specific information in this manual.
474 DJ Delorie ported @value{GDBN} to MS-DOS, for the DJGPP project.
475 Robert Hoehne made significant contributions to the DJGPP port.
477 Cygnus Solutions has sponsored @value{GDBN} maintenance and much of its
478 development since 1991. Cygnus engineers who have worked on @value{GDBN}
479 fulltime include Mark Alexander, Jim Blandy, Per Bothner, Kevin
480 Buettner, Edith Epstein, Chris Faylor, Fred Fish, Martin Hunt, Jim
481 Ingham, John Gilmore, Stu Grossman, Kung Hsu, Jim Kingdon, John Metzler,
482 Fernando Nasser, Geoffrey Noer, Dawn Perchik, Rich Pixley, Zdenek
483 Radouch, Keith Seitz, Stan Shebs, David Taylor, and Elena Zannoni. In
484 addition, Dave Brolley, Ian Carmichael, Steve Chamberlain, Nick Clifton,
485 JT Conklin, Stan Cox, DJ Delorie, Ulrich Drepper, Frank Eigler, Doug
486 Evans, Sean Fagan, David Henkel-Wallace, Richard Henderson, Jeff
487 Holcomb, Jeff Law, Jim Lemke, Tom Lord, Bob Manson, Michael Meissner,
488 Jason Merrill, Catherine Moore, Drew Moseley, Ken Raeburn, Gavin
489 Romig-Koch, Rob Savoye, Jamie Smith, Mike Stump, Ian Taylor, Angela
490 Thomas, Michael Tiemann, Tom Tromey, Ron Unrau, Jim Wilson, and David
491 Zuhn have made contributions both large and small.
493 Andrew Cagney, Fernando Nasser, and Elena Zannoni, while working for
494 Cygnus Solutions, implemented the original @sc{gdb/mi} interface.
496 Jim Blandy added support for preprocessor macros, while working for Red
499 Andrew Cagney designed @value{GDBN}'s architecture vector. Many
500 people including Andrew Cagney, Stephane Carrez, Randolph Chung, Nick
501 Duffek, Richard Henderson, Mark Kettenis, Grace Sainsbury, Kei
502 Sakamoto, Yoshinori Sato, Michael Snyder, Andreas Schwab, Jason
503 Thorpe, Corinna Vinschen, Ulrich Weigand, and Elena Zannoni, helped
504 with the migration of old architectures to this new framework.
506 Andrew Cagney completely re-designed and re-implemented @value{GDBN}'s
507 unwinder framework, this consisting of a fresh new design featuring
508 frame IDs, independent frame sniffers, and the sentinel frame. Mark
509 Kettenis implemented the @sc{dwarf 2} unwinder, Jeff Johnston the
510 libunwind unwinder, and Andrew Cagney the dummy, sentinel, tramp, and
511 trad unwinders. The architecture-specific changes, each involving a
512 complete rewrite of the architecture's frame code, were carried out by
513 Jim Blandy, Joel Brobecker, Kevin Buettner, Andrew Cagney, Stephane
514 Carrez, Randolph Chung, Orjan Friberg, Richard Henderson, Daniel
515 Jacobowitz, Jeff Johnston, Mark Kettenis, Theodore A. Roth, Kei
516 Sakamoto, Yoshinori Sato, Michael Snyder, Corinna Vinschen, and Ulrich
519 Christian Zankel, Ross Morley, Bob Wilson, and Maxim Grigoriev from
520 Tensilica, Inc.@: contributed support for Xtensa processors. Others
521 who have worked on the Xtensa port of @value{GDBN} in the past include
522 Steve Tjiang, John Newlin, and Scott Foehner.
524 Michael Eager and staff of Xilinx, Inc., contributed support for the
525 Xilinx MicroBlaze architecture.
528 @chapter A Sample @value{GDBN} Session
530 You can use this manual at your leisure to read all about @value{GDBN}.
531 However, a handful of commands are enough to get started using the
532 debugger. This chapter illustrates those commands.
535 In this sample session, we emphasize user input like this: @b{input},
536 to make it easier to pick out from the surrounding output.
539 @c FIXME: this example may not be appropriate for some configs, where
540 @c FIXME...primary interest is in remote use.
542 One of the preliminary versions of @sc{gnu} @code{m4} (a generic macro
543 processor) exhibits the following bug: sometimes, when we change its
544 quote strings from the default, the commands used to capture one macro
545 definition within another stop working. In the following short @code{m4}
546 session, we define a macro @code{foo} which expands to @code{0000}; we
547 then use the @code{m4} built-in @code{defn} to define @code{bar} as the
548 same thing. However, when we change the open quote string to
549 @code{<QUOTE>} and the close quote string to @code{<UNQUOTE>}, the same
550 procedure fails to define a new synonym @code{baz}:
559 @b{define(bar,defn(`foo'))}
563 @b{changequote(<QUOTE>,<UNQUOTE>)}
565 @b{define(baz,defn(<QUOTE>foo<UNQUOTE>))}
568 m4: End of input: 0: fatal error: EOF in string
572 Let us use @value{GDBN} to try to see what is going on.
575 $ @b{@value{GDBP} m4}
576 @c FIXME: this falsifies the exact text played out, to permit smallbook
577 @c FIXME... format to come out better.
578 @value{GDBN} is free software and you are welcome to distribute copies
579 of it under certain conditions; type "show copying" to see
581 There is absolutely no warranty for @value{GDBN}; type "show warranty"
584 @value{GDBN} @value{GDBVN}, Copyright 1999 Free Software Foundation, Inc...
589 @value{GDBN} reads only enough symbol data to know where to find the
590 rest when needed; as a result, the first prompt comes up very quickly.
591 We now tell @value{GDBN} to use a narrower display width than usual, so
592 that examples fit in this manual.
595 (@value{GDBP}) @b{set width 70}
599 We need to see how the @code{m4} built-in @code{changequote} works.
600 Having looked at the source, we know the relevant subroutine is
601 @code{m4_changequote}, so we set a breakpoint there with the @value{GDBN}
602 @code{break} command.
605 (@value{GDBP}) @b{break m4_changequote}
606 Breakpoint 1 at 0x62f4: file builtin.c, line 879.
610 Using the @code{run} command, we start @code{m4} running under @value{GDBN}
611 control; as long as control does not reach the @code{m4_changequote}
612 subroutine, the program runs as usual:
615 (@value{GDBP}) @b{run}
616 Starting program: /work/Editorial/gdb/gnu/m4/m4
624 To trigger the breakpoint, we call @code{changequote}. @value{GDBN}
625 suspends execution of @code{m4}, displaying information about the
626 context where it stops.
629 @b{changequote(<QUOTE>,<UNQUOTE>)}
631 Breakpoint 1, m4_changequote (argc=3, argv=0x33c70)
633 879 if (bad_argc(TOKEN_DATA_TEXT(argv[0]),argc,1,3))
637 Now we use the command @code{n} (@code{next}) to advance execution to
638 the next line of the current function.
642 882 set_quotes((argc >= 2) ? TOKEN_DATA_TEXT(argv[1])\
647 @code{set_quotes} looks like a promising subroutine. We can go into it
648 by using the command @code{s} (@code{step}) instead of @code{next}.
649 @code{step} goes to the next line to be executed in @emph{any}
650 subroutine, so it steps into @code{set_quotes}.
654 set_quotes (lq=0x34c78 "<QUOTE>", rq=0x34c88 "<UNQUOTE>")
656 530 if (lquote != def_lquote)
660 The display that shows the subroutine where @code{m4} is now
661 suspended (and its arguments) is called a stack frame display. It
662 shows a summary of the stack. We can use the @code{backtrace}
663 command (which can also be spelled @code{bt}), to see where we are
664 in the stack as a whole: the @code{backtrace} command displays a
665 stack frame for each active subroutine.
668 (@value{GDBP}) @b{bt}
669 #0 set_quotes (lq=0x34c78 "<QUOTE>", rq=0x34c88 "<UNQUOTE>")
671 #1 0x6344 in m4_changequote (argc=3, argv=0x33c70)
673 #2 0x8174 in expand_macro (sym=0x33320) at macro.c:242
674 #3 0x7a88 in expand_token (obs=0x0, t=209696, td=0xf7fffa30)
676 #4 0x79dc in expand_input () at macro.c:40
677 #5 0x2930 in main (argc=0, argv=0xf7fffb20) at m4.c:195
681 We step through a few more lines to see what happens. The first two
682 times, we can use @samp{s}; the next two times we use @code{n} to avoid
683 falling into the @code{xstrdup} subroutine.
687 0x3b5c 532 if (rquote != def_rquote)
689 0x3b80 535 lquote = (lq == nil || *lq == '\0') ? \
690 def_lquote : xstrdup(lq);
692 536 rquote = (rq == nil || *rq == '\0') ? def_rquote\
695 538 len_lquote = strlen(rquote);
699 The last line displayed looks a little odd; we can examine the variables
700 @code{lquote} and @code{rquote} to see if they are in fact the new left
701 and right quotes we specified. We use the command @code{p}
702 (@code{print}) to see their values.
705 (@value{GDBP}) @b{p lquote}
706 $1 = 0x35d40 "<QUOTE>"
707 (@value{GDBP}) @b{p rquote}
708 $2 = 0x35d50 "<UNQUOTE>"
712 @code{lquote} and @code{rquote} are indeed the new left and right quotes.
713 To look at some context, we can display ten lines of source
714 surrounding the current line with the @code{l} (@code{list}) command.
720 535 lquote = (lq == nil || *lq == '\0') ? def_lquote\
722 536 rquote = (rq == nil || *rq == '\0') ? def_rquote\
725 538 len_lquote = strlen(rquote);
726 539 len_rquote = strlen(lquote);
733 Let us step past the two lines that set @code{len_lquote} and
734 @code{len_rquote}, and then examine the values of those variables.
738 539 len_rquote = strlen(lquote);
741 (@value{GDBP}) @b{p len_lquote}
743 (@value{GDBP}) @b{p len_rquote}
748 That certainly looks wrong, assuming @code{len_lquote} and
749 @code{len_rquote} are meant to be the lengths of @code{lquote} and
750 @code{rquote} respectively. We can set them to better values using
751 the @code{p} command, since it can print the value of
752 any expression---and that expression can include subroutine calls and
756 (@value{GDBP}) @b{p len_lquote=strlen(lquote)}
758 (@value{GDBP}) @b{p len_rquote=strlen(rquote)}
763 Is that enough to fix the problem of using the new quotes with the
764 @code{m4} built-in @code{defn}? We can allow @code{m4} to continue
765 executing with the @code{c} (@code{continue}) command, and then try the
766 example that caused trouble initially:
772 @b{define(baz,defn(<QUOTE>foo<UNQUOTE>))}
779 Success! The new quotes now work just as well as the default ones. The
780 problem seems to have been just the two typos defining the wrong
781 lengths. We allow @code{m4} exit by giving it an EOF as input:
785 Program exited normally.
789 The message @samp{Program exited normally.} is from @value{GDBN}; it
790 indicates @code{m4} has finished executing. We can end our @value{GDBN}
791 session with the @value{GDBN} @code{quit} command.
794 (@value{GDBP}) @b{quit}
798 @chapter Getting In and Out of @value{GDBN}
800 This chapter discusses how to start @value{GDBN}, and how to get out of it.
804 type @samp{@value{GDBP}} to start @value{GDBN}.
806 type @kbd{quit} or @kbd{Ctrl-d} to exit.
810 * Invoking GDB:: How to start @value{GDBN}
811 * Quitting GDB:: How to quit @value{GDBN}
812 * Shell Commands:: How to use shell commands inside @value{GDBN}
813 * Logging Output:: How to log @value{GDBN}'s output to a file
817 @section Invoking @value{GDBN}
819 Invoke @value{GDBN} by running the program @code{@value{GDBP}}. Once started,
820 @value{GDBN} reads commands from the terminal until you tell it to exit.
822 You can also run @code{@value{GDBP}} with a variety of arguments and options,
823 to specify more of your debugging environment at the outset.
825 The command-line options described here are designed
826 to cover a variety of situations; in some environments, some of these
827 options may effectively be unavailable.
829 The most usual way to start @value{GDBN} is with one argument,
830 specifying an executable program:
833 @value{GDBP} @var{program}
837 You can also start with both an executable program and a core file
841 @value{GDBP} @var{program} @var{core}
844 You can, instead, specify a process ID as a second argument, if you want
845 to debug a running process:
848 @value{GDBP} @var{program} 1234
852 would attach @value{GDBN} to process @code{1234} (unless you also have a file
853 named @file{1234}; @value{GDBN} does check for a core file first).
855 Taking advantage of the second command-line argument requires a fairly
856 complete operating system; when you use @value{GDBN} as a remote
857 debugger attached to a bare board, there may not be any notion of
858 ``process'', and there is often no way to get a core dump. @value{GDBN}
859 will warn you if it is unable to attach or to read core dumps.
861 You can optionally have @code{@value{GDBP}} pass any arguments after the
862 executable file to the inferior using @code{--args}. This option stops
865 @value{GDBP} --args gcc -O2 -c foo.c
867 This will cause @code{@value{GDBP}} to debug @code{gcc}, and to set
868 @code{gcc}'s command-line arguments (@pxref{Arguments}) to @samp{-O2 -c foo.c}.
870 You can run @code{@value{GDBP}} without printing the front material, which describes
871 @value{GDBN}'s non-warranty, by specifying @code{-silent}:
878 You can further control how @value{GDBN} starts up by using command-line
879 options. @value{GDBN} itself can remind you of the options available.
889 to display all available options and briefly describe their use
890 (@samp{@value{GDBP} -h} is a shorter equivalent).
892 All options and command line arguments you give are processed
893 in sequential order. The order makes a difference when the
894 @samp{-x} option is used.
898 * File Options:: Choosing files
899 * Mode Options:: Choosing modes
900 * Startup:: What @value{GDBN} does during startup
904 @subsection Choosing Files
906 When @value{GDBN} starts, it reads any arguments other than options as
907 specifying an executable file and core file (or process ID). This is
908 the same as if the arguments were specified by the @samp{-se} and
909 @samp{-c} (or @samp{-p}) options respectively. (@value{GDBN} reads the
910 first argument that does not have an associated option flag as
911 equivalent to the @samp{-se} option followed by that argument; and the
912 second argument that does not have an associated option flag, if any, as
913 equivalent to the @samp{-c}/@samp{-p} option followed by that argument.)
914 If the second argument begins with a decimal digit, @value{GDBN} will
915 first attempt to attach to it as a process, and if that fails, attempt
916 to open it as a corefile. If you have a corefile whose name begins with
917 a digit, you can prevent @value{GDBN} from treating it as a pid by
918 prefixing it with @file{./}, e.g.@: @file{./12345}.
920 If @value{GDBN} has not been configured to included core file support,
921 such as for most embedded targets, then it will complain about a second
922 argument and ignore it.
924 Many options have both long and short forms; both are shown in the
925 following list. @value{GDBN} also recognizes the long forms if you truncate
926 them, so long as enough of the option is present to be unambiguous.
927 (If you prefer, you can flag option arguments with @samp{--} rather
928 than @samp{-}, though we illustrate the more usual convention.)
930 @c NOTE: the @cindex entries here use double dashes ON PURPOSE. This
931 @c way, both those who look for -foo and --foo in the index, will find
935 @item -symbols @var{file}
937 @cindex @code{--symbols}
939 Read symbol table from file @var{file}.
941 @item -exec @var{file}
943 @cindex @code{--exec}
945 Use file @var{file} as the executable file to execute when appropriate,
946 and for examining pure data in conjunction with a core dump.
950 Read symbol table from file @var{file} and use it as the executable
953 @item -core @var{file}
955 @cindex @code{--core}
957 Use file @var{file} as a core dump to examine.
959 @item -pid @var{number}
960 @itemx -p @var{number}
963 Connect to process ID @var{number}, as with the @code{attach} command.
965 @item -command @var{file}
967 @cindex @code{--command}
969 Execute commands from file @var{file}. The contents of this file is
970 evaluated exactly as the @code{source} command would.
971 @xref{Command Files,, Command files}.
973 @item -eval-command @var{command}
974 @itemx -ex @var{command}
975 @cindex @code{--eval-command}
977 Execute a single @value{GDBN} command.
979 This option may be used multiple times to call multiple commands. It may
980 also be interleaved with @samp{-command} as required.
983 @value{GDBP} -ex 'target sim' -ex 'load' \
984 -x setbreakpoints -ex 'run' a.out
987 @item -directory @var{directory}
988 @itemx -d @var{directory}
989 @cindex @code{--directory}
991 Add @var{directory} to the path to search for source and script files.
995 @cindex @code{--readnow}
997 Read each symbol file's entire symbol table immediately, rather than
998 the default, which is to read it incrementally as it is needed.
999 This makes startup slower, but makes future operations faster.
1004 @subsection Choosing Modes
1006 You can run @value{GDBN} in various alternative modes---for example, in
1007 batch mode or quiet mode.
1014 Do not execute commands found in any initialization files. Normally,
1015 @value{GDBN} executes the commands in these files after all the command
1016 options and arguments have been processed. @xref{Command Files,,Command
1022 @cindex @code{--quiet}
1023 @cindex @code{--silent}
1025 ``Quiet''. Do not print the introductory and copyright messages. These
1026 messages are also suppressed in batch mode.
1029 @cindex @code{--batch}
1030 Run in batch mode. Exit with status @code{0} after processing all the
1031 command files specified with @samp{-x} (and all commands from
1032 initialization files, if not inhibited with @samp{-n}). Exit with
1033 nonzero status if an error occurs in executing the @value{GDBN} commands
1034 in the command files. Batch mode also disables pagination;
1035 @pxref{Screen Size} and acts as if @kbd{set confirm off} were in
1036 effect (@pxref{Messages/Warnings}).
1038 Batch mode may be useful for running @value{GDBN} as a filter, for
1039 example to download and run a program on another computer; in order to
1040 make this more useful, the message
1043 Program exited normally.
1047 (which is ordinarily issued whenever a program running under
1048 @value{GDBN} control terminates) is not issued when running in batch
1052 @cindex @code{--batch-silent}
1053 Run in batch mode exactly like @samp{-batch}, but totally silently. All
1054 @value{GDBN} output to @code{stdout} is prevented (@code{stderr} is
1055 unaffected). This is much quieter than @samp{-silent} and would be useless
1056 for an interactive session.
1058 This is particularly useful when using targets that give @samp{Loading section}
1059 messages, for example.
1061 Note that targets that give their output via @value{GDBN}, as opposed to
1062 writing directly to @code{stdout}, will also be made silent.
1064 @item -return-child-result
1065 @cindex @code{--return-child-result}
1066 The return code from @value{GDBN} will be the return code from the child
1067 process (the process being debugged), with the following exceptions:
1071 @value{GDBN} exits abnormally. E.g., due to an incorrect argument or an
1072 internal error. In this case the exit code is the same as it would have been
1073 without @samp{-return-child-result}.
1075 The user quits with an explicit value. E.g., @samp{quit 1}.
1077 The child process never runs, or is not allowed to terminate, in which case
1078 the exit code will be -1.
1081 This option is useful in conjunction with @samp{-batch} or @samp{-batch-silent},
1082 when @value{GDBN} is being used as a remote program loader or simulator
1087 @cindex @code{--nowindows}
1089 ``No windows''. If @value{GDBN} comes with a graphical user interface
1090 (GUI) built in, then this option tells @value{GDBN} to only use the command-line
1091 interface. If no GUI is available, this option has no effect.
1095 @cindex @code{--windows}
1097 If @value{GDBN} includes a GUI, then this option requires it to be
1100 @item -cd @var{directory}
1102 Run @value{GDBN} using @var{directory} as its working directory,
1103 instead of the current directory.
1107 @cindex @code{--fullname}
1109 @sc{gnu} Emacs sets this option when it runs @value{GDBN} as a
1110 subprocess. It tells @value{GDBN} to output the full file name and line
1111 number in a standard, recognizable fashion each time a stack frame is
1112 displayed (which includes each time your program stops). This
1113 recognizable format looks like two @samp{\032} characters, followed by
1114 the file name, line number and character position separated by colons,
1115 and a newline. The Emacs-to-@value{GDBN} interface program uses the two
1116 @samp{\032} characters as a signal to display the source code for the
1120 @cindex @code{--epoch}
1121 The Epoch Emacs-@value{GDBN} interface sets this option when it runs
1122 @value{GDBN} as a subprocess. It tells @value{GDBN} to modify its print
1123 routines so as to allow Epoch to display values of expressions in a
1126 @item -annotate @var{level}
1127 @cindex @code{--annotate}
1128 This option sets the @dfn{annotation level} inside @value{GDBN}. Its
1129 effect is identical to using @samp{set annotate @var{level}}
1130 (@pxref{Annotations}). The annotation @var{level} controls how much
1131 information @value{GDBN} prints together with its prompt, values of
1132 expressions, source lines, and other types of output. Level 0 is the
1133 normal, level 1 is for use when @value{GDBN} is run as a subprocess of
1134 @sc{gnu} Emacs, level 3 is the maximum annotation suitable for programs
1135 that control @value{GDBN}, and level 2 has been deprecated.
1137 The annotation mechanism has largely been superseded by @sc{gdb/mi}
1141 @cindex @code{--args}
1142 Change interpretation of command line so that arguments following the
1143 executable file are passed as command line arguments to the inferior.
1144 This option stops option processing.
1146 @item -baud @var{bps}
1148 @cindex @code{--baud}
1150 Set the line speed (baud rate or bits per second) of any serial
1151 interface used by @value{GDBN} for remote debugging.
1153 @item -l @var{timeout}
1155 Set the timeout (in seconds) of any communication used by @value{GDBN}
1156 for remote debugging.
1158 @item -tty @var{device}
1159 @itemx -t @var{device}
1160 @cindex @code{--tty}
1162 Run using @var{device} for your program's standard input and output.
1163 @c FIXME: kingdon thinks there is more to -tty. Investigate.
1165 @c resolve the situation of these eventually
1167 @cindex @code{--tui}
1168 Activate the @dfn{Text User Interface} when starting. The Text User
1169 Interface manages several text windows on the terminal, showing
1170 source, assembly, registers and @value{GDBN} command outputs
1171 (@pxref{TUI, ,@value{GDBN} Text User Interface}). Alternatively, the
1172 Text User Interface can be enabled by invoking the program
1173 @samp{@value{GDBTUI}}. Do not use this option if you run @value{GDBN} from
1174 Emacs (@pxref{Emacs, ,Using @value{GDBN} under @sc{gnu} Emacs}).
1177 @c @cindex @code{--xdb}
1178 @c Run in XDB compatibility mode, allowing the use of certain XDB commands.
1179 @c For information, see the file @file{xdb_trans.html}, which is usually
1180 @c installed in the directory @code{/opt/langtools/wdb/doc} on HP-UX
1183 @item -interpreter @var{interp}
1184 @cindex @code{--interpreter}
1185 Use the interpreter @var{interp} for interface with the controlling
1186 program or device. This option is meant to be set by programs which
1187 communicate with @value{GDBN} using it as a back end.
1188 @xref{Interpreters, , Command Interpreters}.
1190 @samp{--interpreter=mi} (or @samp{--interpreter=mi2}) causes
1191 @value{GDBN} to use the @dfn{@sc{gdb/mi} interface} (@pxref{GDB/MI, ,
1192 The @sc{gdb/mi} Interface}) included since @value{GDBN} version 6.0. The
1193 previous @sc{gdb/mi} interface, included in @value{GDBN} version 5.3 and
1194 selected with @samp{--interpreter=mi1}, is deprecated. Earlier
1195 @sc{gdb/mi} interfaces are no longer supported.
1198 @cindex @code{--write}
1199 Open the executable and core files for both reading and writing. This
1200 is equivalent to the @samp{set write on} command inside @value{GDBN}
1204 @cindex @code{--statistics}
1205 This option causes @value{GDBN} to print statistics about time and
1206 memory usage after it completes each command and returns to the prompt.
1209 @cindex @code{--version}
1210 This option causes @value{GDBN} to print its version number and
1211 no-warranty blurb, and exit.
1216 @subsection What @value{GDBN} Does During Startup
1217 @cindex @value{GDBN} startup
1219 Here's the description of what @value{GDBN} does during session startup:
1223 Sets up the command interpreter as specified by the command line
1224 (@pxref{Mode Options, interpreter}).
1228 Reads the system-wide @dfn{init file} (if @option{--with-system-gdbinit} was
1229 used when building @value{GDBN}; @pxref{System-wide configuration,
1230 ,System-wide configuration and settings}) and executes all the commands in
1234 Reads the init file (if any) in your home directory@footnote{On
1235 DOS/Windows systems, the home directory is the one pointed to by the
1236 @code{HOME} environment variable.} and executes all the commands in
1240 Processes command line options and operands.
1243 Reads and executes the commands from init file (if any) in the current
1244 working directory. This is only done if the current directory is
1245 different from your home directory. Thus, you can have more than one
1246 init file, one generic in your home directory, and another, specific
1247 to the program you are debugging, in the directory where you invoke
1251 Reads command files specified by the @samp{-x} option. @xref{Command
1252 Files}, for more details about @value{GDBN} command files.
1255 Reads the command history recorded in the @dfn{history file}.
1256 @xref{Command History}, for more details about the command history and the
1257 files where @value{GDBN} records it.
1260 Init files use the same syntax as @dfn{command files} (@pxref{Command
1261 Files}) and are processed by @value{GDBN} in the same way. The init
1262 file in your home directory can set options (such as @samp{set
1263 complaints}) that affect subsequent processing of command line options
1264 and operands. Init files are not executed if you use the @samp{-nx}
1265 option (@pxref{Mode Options, ,Choosing Modes}).
1267 To display the list of init files loaded by gdb at startup, you
1268 can use @kbd{gdb --help}.
1270 @cindex init file name
1271 @cindex @file{.gdbinit}
1272 @cindex @file{gdb.ini}
1273 The @value{GDBN} init files are normally called @file{.gdbinit}.
1274 The DJGPP port of @value{GDBN} uses the name @file{gdb.ini}, due to
1275 the limitations of file names imposed by DOS filesystems. The Windows
1276 ports of @value{GDBN} use the standard name, but if they find a
1277 @file{gdb.ini} file, they warn you about that and suggest to rename
1278 the file to the standard name.
1282 @section Quitting @value{GDBN}
1283 @cindex exiting @value{GDBN}
1284 @cindex leaving @value{GDBN}
1287 @kindex quit @r{[}@var{expression}@r{]}
1288 @kindex q @r{(@code{quit})}
1289 @item quit @r{[}@var{expression}@r{]}
1291 To exit @value{GDBN}, use the @code{quit} command (abbreviated
1292 @code{q}), or type an end-of-file character (usually @kbd{Ctrl-d}). If you
1293 do not supply @var{expression}, @value{GDBN} will terminate normally;
1294 otherwise it will terminate using the result of @var{expression} as the
1299 An interrupt (often @kbd{Ctrl-c}) does not exit from @value{GDBN}, but rather
1300 terminates the action of any @value{GDBN} command that is in progress and
1301 returns to @value{GDBN} command level. It is safe to type the interrupt
1302 character at any time because @value{GDBN} does not allow it to take effect
1303 until a time when it is safe.
1305 If you have been using @value{GDBN} to control an attached process or
1306 device, you can release it with the @code{detach} command
1307 (@pxref{Attach, ,Debugging an Already-running Process}).
1309 @node Shell Commands
1310 @section Shell Commands
1312 If you need to execute occasional shell commands during your
1313 debugging session, there is no need to leave or suspend @value{GDBN}; you can
1314 just use the @code{shell} command.
1318 @cindex shell escape
1319 @item shell @var{command string}
1320 Invoke a standard shell to execute @var{command string}.
1321 If it exists, the environment variable @code{SHELL} determines which
1322 shell to run. Otherwise @value{GDBN} uses the default shell
1323 (@file{/bin/sh} on Unix systems, @file{COMMAND.COM} on MS-DOS, etc.).
1326 The utility @code{make} is often needed in development environments.
1327 You do not have to use the @code{shell} command for this purpose in
1332 @cindex calling make
1333 @item make @var{make-args}
1334 Execute the @code{make} program with the specified
1335 arguments. This is equivalent to @samp{shell make @var{make-args}}.
1338 @node Logging Output
1339 @section Logging Output
1340 @cindex logging @value{GDBN} output
1341 @cindex save @value{GDBN} output to a file
1343 You may want to save the output of @value{GDBN} commands to a file.
1344 There are several commands to control @value{GDBN}'s logging.
1348 @item set logging on
1350 @item set logging off
1352 @cindex logging file name
1353 @item set logging file @var{file}
1354 Change the name of the current logfile. The default logfile is @file{gdb.txt}.
1355 @item set logging overwrite [on|off]
1356 By default, @value{GDBN} will append to the logfile. Set @code{overwrite} if
1357 you want @code{set logging on} to overwrite the logfile instead.
1358 @item set logging redirect [on|off]
1359 By default, @value{GDBN} output will go to both the terminal and the logfile.
1360 Set @code{redirect} if you want output to go only to the log file.
1361 @kindex show logging
1363 Show the current values of the logging settings.
1367 @chapter @value{GDBN} Commands
1369 You can abbreviate a @value{GDBN} command to the first few letters of the command
1370 name, if that abbreviation is unambiguous; and you can repeat certain
1371 @value{GDBN} commands by typing just @key{RET}. You can also use the @key{TAB}
1372 key to get @value{GDBN} to fill out the rest of a word in a command (or to
1373 show you the alternatives available, if there is more than one possibility).
1376 * Command Syntax:: How to give commands to @value{GDBN}
1377 * Completion:: Command completion
1378 * Help:: How to ask @value{GDBN} for help
1381 @node Command Syntax
1382 @section Command Syntax
1384 A @value{GDBN} command is a single line of input. There is no limit on
1385 how long it can be. It starts with a command name, which is followed by
1386 arguments whose meaning depends on the command name. For example, the
1387 command @code{step} accepts an argument which is the number of times to
1388 step, as in @samp{step 5}. You can also use the @code{step} command
1389 with no arguments. Some commands do not allow any arguments.
1391 @cindex abbreviation
1392 @value{GDBN} command names may always be truncated if that abbreviation is
1393 unambiguous. Other possible command abbreviations are listed in the
1394 documentation for individual commands. In some cases, even ambiguous
1395 abbreviations are allowed; for example, @code{s} is specially defined as
1396 equivalent to @code{step} even though there are other commands whose
1397 names start with @code{s}. You can test abbreviations by using them as
1398 arguments to the @code{help} command.
1400 @cindex repeating commands
1401 @kindex RET @r{(repeat last command)}
1402 A blank line as input to @value{GDBN} (typing just @key{RET}) means to
1403 repeat the previous command. Certain commands (for example, @code{run})
1404 will not repeat this way; these are commands whose unintentional
1405 repetition might cause trouble and which you are unlikely to want to
1406 repeat. User-defined commands can disable this feature; see
1407 @ref{Define, dont-repeat}.
1409 The @code{list} and @code{x} commands, when you repeat them with
1410 @key{RET}, construct new arguments rather than repeating
1411 exactly as typed. This permits easy scanning of source or memory.
1413 @value{GDBN} can also use @key{RET} in another way: to partition lengthy
1414 output, in a way similar to the common utility @code{more}
1415 (@pxref{Screen Size,,Screen Size}). Since it is easy to press one
1416 @key{RET} too many in this situation, @value{GDBN} disables command
1417 repetition after any command that generates this sort of display.
1419 @kindex # @r{(a comment)}
1421 Any text from a @kbd{#} to the end of the line is a comment; it does
1422 nothing. This is useful mainly in command files (@pxref{Command
1423 Files,,Command Files}).
1425 @cindex repeating command sequences
1426 @kindex Ctrl-o @r{(operate-and-get-next)}
1427 The @kbd{Ctrl-o} binding is useful for repeating a complex sequence of
1428 commands. This command accepts the current line, like @key{RET}, and
1429 then fetches the next line relative to the current line from the history
1433 @section Command Completion
1436 @cindex word completion
1437 @value{GDBN} can fill in the rest of a word in a command for you, if there is
1438 only one possibility; it can also show you what the valid possibilities
1439 are for the next word in a command, at any time. This works for @value{GDBN}
1440 commands, @value{GDBN} subcommands, and the names of symbols in your program.
1442 Press the @key{TAB} key whenever you want @value{GDBN} to fill out the rest
1443 of a word. If there is only one possibility, @value{GDBN} fills in the
1444 word, and waits for you to finish the command (or press @key{RET} to
1445 enter it). For example, if you type
1447 @c FIXME "@key" does not distinguish its argument sufficiently to permit
1448 @c complete accuracy in these examples; space introduced for clarity.
1449 @c If texinfo enhancements make it unnecessary, it would be nice to
1450 @c replace " @key" by "@key" in the following...
1452 (@value{GDBP}) info bre @key{TAB}
1456 @value{GDBN} fills in the rest of the word @samp{breakpoints}, since that is
1457 the only @code{info} subcommand beginning with @samp{bre}:
1460 (@value{GDBP}) info breakpoints
1464 You can either press @key{RET} at this point, to run the @code{info
1465 breakpoints} command, or backspace and enter something else, if
1466 @samp{breakpoints} does not look like the command you expected. (If you
1467 were sure you wanted @code{info breakpoints} in the first place, you
1468 might as well just type @key{RET} immediately after @samp{info bre},
1469 to exploit command abbreviations rather than command completion).
1471 If there is more than one possibility for the next word when you press
1472 @key{TAB}, @value{GDBN} sounds a bell. You can either supply more
1473 characters and try again, or just press @key{TAB} a second time;
1474 @value{GDBN} displays all the possible completions for that word. For
1475 example, you might want to set a breakpoint on a subroutine whose name
1476 begins with @samp{make_}, but when you type @kbd{b make_@key{TAB}} @value{GDBN}
1477 just sounds the bell. Typing @key{TAB} again displays all the
1478 function names in your program that begin with those characters, for
1482 (@value{GDBP}) b make_ @key{TAB}
1483 @exdent @value{GDBN} sounds bell; press @key{TAB} again, to see:
1484 make_a_section_from_file make_environ
1485 make_abs_section make_function_type
1486 make_blockvector make_pointer_type
1487 make_cleanup make_reference_type
1488 make_command make_symbol_completion_list
1489 (@value{GDBP}) b make_
1493 After displaying the available possibilities, @value{GDBN} copies your
1494 partial input (@samp{b make_} in the example) so you can finish the
1497 If you just want to see the list of alternatives in the first place, you
1498 can press @kbd{M-?} rather than pressing @key{TAB} twice. @kbd{M-?}
1499 means @kbd{@key{META} ?}. You can type this either by holding down a
1500 key designated as the @key{META} shift on your keyboard (if there is
1501 one) while typing @kbd{?}, or as @key{ESC} followed by @kbd{?}.
1503 @cindex quotes in commands
1504 @cindex completion of quoted strings
1505 Sometimes the string you need, while logically a ``word'', may contain
1506 parentheses or other characters that @value{GDBN} normally excludes from
1507 its notion of a word. To permit word completion to work in this
1508 situation, you may enclose words in @code{'} (single quote marks) in
1509 @value{GDBN} commands.
1511 The most likely situation where you might need this is in typing the
1512 name of a C@t{++} function. This is because C@t{++} allows function
1513 overloading (multiple definitions of the same function, distinguished
1514 by argument type). For example, when you want to set a breakpoint you
1515 may need to distinguish whether you mean the version of @code{name}
1516 that takes an @code{int} parameter, @code{name(int)}, or the version
1517 that takes a @code{float} parameter, @code{name(float)}. To use the
1518 word-completion facilities in this situation, type a single quote
1519 @code{'} at the beginning of the function name. This alerts
1520 @value{GDBN} that it may need to consider more information than usual
1521 when you press @key{TAB} or @kbd{M-?} to request word completion:
1524 (@value{GDBP}) b 'bubble( @kbd{M-?}
1525 bubble(double,double) bubble(int,int)
1526 (@value{GDBP}) b 'bubble(
1529 In some cases, @value{GDBN} can tell that completing a name requires using
1530 quotes. When this happens, @value{GDBN} inserts the quote for you (while
1531 completing as much as it can) if you do not type the quote in the first
1535 (@value{GDBP}) b bub @key{TAB}
1536 @exdent @value{GDBN} alters your input line to the following, and rings a bell:
1537 (@value{GDBP}) b 'bubble(
1541 In general, @value{GDBN} can tell that a quote is needed (and inserts it) if
1542 you have not yet started typing the argument list when you ask for
1543 completion on an overloaded symbol.
1545 For more information about overloaded functions, see @ref{C Plus Plus
1546 Expressions, ,C@t{++} Expressions}. You can use the command @code{set
1547 overload-resolution off} to disable overload resolution;
1548 see @ref{Debugging C Plus Plus, ,@value{GDBN} Features for C@t{++}}.
1550 @cindex completion of structure field names
1551 @cindex structure field name completion
1552 @cindex completion of union field names
1553 @cindex union field name completion
1554 When completing in an expression which looks up a field in a
1555 structure, @value{GDBN} also tries@footnote{The completer can be
1556 confused by certain kinds of invalid expressions. Also, it only
1557 examines the static type of the expression, not the dynamic type.} to
1558 limit completions to the field names available in the type of the
1562 (@value{GDBP}) p gdb_stdout.@kbd{M-?}
1563 magic to_delete to_fputs to_put to_rewind
1564 to_data to_flush to_isatty to_read to_write
1568 This is because the @code{gdb_stdout} is a variable of the type
1569 @code{struct ui_file} that is defined in @value{GDBN} sources as
1576 ui_file_flush_ftype *to_flush;
1577 ui_file_write_ftype *to_write;
1578 ui_file_fputs_ftype *to_fputs;
1579 ui_file_read_ftype *to_read;
1580 ui_file_delete_ftype *to_delete;
1581 ui_file_isatty_ftype *to_isatty;
1582 ui_file_rewind_ftype *to_rewind;
1583 ui_file_put_ftype *to_put;
1590 @section Getting Help
1591 @cindex online documentation
1594 You can always ask @value{GDBN} itself for information on its commands,
1595 using the command @code{help}.
1598 @kindex h @r{(@code{help})}
1601 You can use @code{help} (abbreviated @code{h}) with no arguments to
1602 display a short list of named classes of commands:
1606 List of classes of commands:
1608 aliases -- Aliases of other commands
1609 breakpoints -- Making program stop at certain points
1610 data -- Examining data
1611 files -- Specifying and examining files
1612 internals -- Maintenance commands
1613 obscure -- Obscure features
1614 running -- Running the program
1615 stack -- Examining the stack
1616 status -- Status inquiries
1617 support -- Support facilities
1618 tracepoints -- Tracing of program execution without
1619 stopping the program
1620 user-defined -- User-defined commands
1622 Type "help" followed by a class name for a list of
1623 commands in that class.
1624 Type "help" followed by command name for full
1626 Command name abbreviations are allowed if unambiguous.
1629 @c the above line break eliminates huge line overfull...
1631 @item help @var{class}
1632 Using one of the general help classes as an argument, you can get a
1633 list of the individual commands in that class. For example, here is the
1634 help display for the class @code{status}:
1637 (@value{GDBP}) help status
1642 @c Line break in "show" line falsifies real output, but needed
1643 @c to fit in smallbook page size.
1644 info -- Generic command for showing things
1645 about the program being debugged
1646 show -- Generic command for showing things
1649 Type "help" followed by command name for full
1651 Command name abbreviations are allowed if unambiguous.
1655 @item help @var{command}
1656 With a command name as @code{help} argument, @value{GDBN} displays a
1657 short paragraph on how to use that command.
1660 @item apropos @var{args}
1661 The @code{apropos} command searches through all of the @value{GDBN}
1662 commands, and their documentation, for the regular expression specified in
1663 @var{args}. It prints out all matches found. For example:
1674 set symbol-reloading -- Set dynamic symbol table reloading
1675 multiple times in one run
1676 show symbol-reloading -- Show dynamic symbol table reloading
1677 multiple times in one run
1682 @item complete @var{args}
1683 The @code{complete @var{args}} command lists all the possible completions
1684 for the beginning of a command. Use @var{args} to specify the beginning of the
1685 command you want completed. For example:
1691 @noindent results in:
1702 @noindent This is intended for use by @sc{gnu} Emacs.
1705 In addition to @code{help}, you can use the @value{GDBN} commands @code{info}
1706 and @code{show} to inquire about the state of your program, or the state
1707 of @value{GDBN} itself. Each command supports many topics of inquiry; this
1708 manual introduces each of them in the appropriate context. The listings
1709 under @code{info} and under @code{show} in the Index point to
1710 all the sub-commands. @xref{Index}.
1715 @kindex i @r{(@code{info})}
1717 This command (abbreviated @code{i}) is for describing the state of your
1718 program. For example, you can show the arguments passed to a function
1719 with @code{info args}, list the registers currently in use with @code{info
1720 registers}, or list the breakpoints you have set with @code{info breakpoints}.
1721 You can get a complete list of the @code{info} sub-commands with
1722 @w{@code{help info}}.
1726 You can assign the result of an expression to an environment variable with
1727 @code{set}. For example, you can set the @value{GDBN} prompt to a $-sign with
1728 @code{set prompt $}.
1732 In contrast to @code{info}, @code{show} is for describing the state of
1733 @value{GDBN} itself.
1734 You can change most of the things you can @code{show}, by using the
1735 related command @code{set}; for example, you can control what number
1736 system is used for displays with @code{set radix}, or simply inquire
1737 which is currently in use with @code{show radix}.
1740 To display all the settable parameters and their current
1741 values, you can use @code{show} with no arguments; you may also use
1742 @code{info set}. Both commands produce the same display.
1743 @c FIXME: "info set" violates the rule that "info" is for state of
1744 @c FIXME...program. Ck w/ GNU: "info set" to be called something else,
1745 @c FIXME...or change desc of rule---eg "state of prog and debugging session"?
1749 Here are three miscellaneous @code{show} subcommands, all of which are
1750 exceptional in lacking corresponding @code{set} commands:
1753 @kindex show version
1754 @cindex @value{GDBN} version number
1756 Show what version of @value{GDBN} is running. You should include this
1757 information in @value{GDBN} bug-reports. If multiple versions of
1758 @value{GDBN} are in use at your site, you may need to determine which
1759 version of @value{GDBN} you are running; as @value{GDBN} evolves, new
1760 commands are introduced, and old ones may wither away. Also, many
1761 system vendors ship variant versions of @value{GDBN}, and there are
1762 variant versions of @value{GDBN} in @sc{gnu}/Linux distributions as well.
1763 The version number is the same as the one announced when you start
1766 @kindex show copying
1767 @kindex info copying
1768 @cindex display @value{GDBN} copyright
1771 Display information about permission for copying @value{GDBN}.
1773 @kindex show warranty
1774 @kindex info warranty
1776 @itemx info warranty
1777 Display the @sc{gnu} ``NO WARRANTY'' statement, or a warranty,
1778 if your version of @value{GDBN} comes with one.
1783 @chapter Running Programs Under @value{GDBN}
1785 When you run a program under @value{GDBN}, you must first generate
1786 debugging information when you compile it.
1788 You may start @value{GDBN} with its arguments, if any, in an environment
1789 of your choice. If you are doing native debugging, you may redirect
1790 your program's input and output, debug an already running process, or
1791 kill a child process.
1794 * Compilation:: Compiling for debugging
1795 * Starting:: Starting your program
1796 * Arguments:: Your program's arguments
1797 * Environment:: Your program's environment
1799 * Working Directory:: Your program's working directory
1800 * Input/Output:: Your program's input and output
1801 * Attach:: Debugging an already-running process
1802 * Kill Process:: Killing the child process
1804 * Inferiors and Programs:: Debugging multiple inferiors and programs
1805 * Threads:: Debugging programs with multiple threads
1806 * Forks:: Debugging forks
1807 * Checkpoint/Restart:: Setting a @emph{bookmark} to return to later
1811 @section Compiling for Debugging
1813 In order to debug a program effectively, you need to generate
1814 debugging information when you compile it. This debugging information
1815 is stored in the object file; it describes the data type of each
1816 variable or function and the correspondence between source line numbers
1817 and addresses in the executable code.
1819 To request debugging information, specify the @samp{-g} option when you run
1822 Programs that are to be shipped to your customers are compiled with
1823 optimizations, using the @samp{-O} compiler option. However, some
1824 compilers are unable to handle the @samp{-g} and @samp{-O} options
1825 together. Using those compilers, you cannot generate optimized
1826 executables containing debugging information.
1828 @value{NGCC}, the @sc{gnu} C/C@t{++} compiler, supports @samp{-g} with or
1829 without @samp{-O}, making it possible to debug optimized code. We
1830 recommend that you @emph{always} use @samp{-g} whenever you compile a
1831 program. You may think your program is correct, but there is no sense
1832 in pushing your luck. For more information, see @ref{Optimized Code}.
1834 Older versions of the @sc{gnu} C compiler permitted a variant option
1835 @w{@samp{-gg}} for debugging information. @value{GDBN} no longer supports this
1836 format; if your @sc{gnu} C compiler has this option, do not use it.
1838 @value{GDBN} knows about preprocessor macros and can show you their
1839 expansion (@pxref{Macros}). Most compilers do not include information
1840 about preprocessor macros in the debugging information if you specify
1841 the @option{-g} flag alone, because this information is rather large.
1842 Version 3.1 and later of @value{NGCC}, the @sc{gnu} C compiler,
1843 provides macro information if you specify the options
1844 @option{-gdwarf-2} and @option{-g3}; the former option requests
1845 debugging information in the Dwarf 2 format, and the latter requests
1846 ``extra information''. In the future, we hope to find more compact
1847 ways to represent macro information, so that it can be included with
1852 @section Starting your Program
1858 @kindex r @r{(@code{run})}
1861 Use the @code{run} command to start your program under @value{GDBN}.
1862 You must first specify the program name (except on VxWorks) with an
1863 argument to @value{GDBN} (@pxref{Invocation, ,Getting In and Out of
1864 @value{GDBN}}), or by using the @code{file} or @code{exec-file} command
1865 (@pxref{Files, ,Commands to Specify Files}).
1869 If you are running your program in an execution environment that
1870 supports processes, @code{run} creates an inferior process and makes
1871 that process run your program. In some environments without processes,
1872 @code{run} jumps to the start of your program. Other targets,
1873 like @samp{remote}, are always running. If you get an error
1874 message like this one:
1877 The "remote" target does not support "run".
1878 Try "help target" or "continue".
1882 then use @code{continue} to run your program. You may need @code{load}
1883 first (@pxref{load}).
1885 The execution of a program is affected by certain information it
1886 receives from its superior. @value{GDBN} provides ways to specify this
1887 information, which you must do @emph{before} starting your program. (You
1888 can change it after starting your program, but such changes only affect
1889 your program the next time you start it.) This information may be
1890 divided into four categories:
1893 @item The @emph{arguments.}
1894 Specify the arguments to give your program as the arguments of the
1895 @code{run} command. If a shell is available on your target, the shell
1896 is used to pass the arguments, so that you may use normal conventions
1897 (such as wildcard expansion or variable substitution) in describing
1899 In Unix systems, you can control which shell is used with the
1900 @code{SHELL} environment variable.
1901 @xref{Arguments, ,Your Program's Arguments}.
1903 @item The @emph{environment.}
1904 Your program normally inherits its environment from @value{GDBN}, but you can
1905 use the @value{GDBN} commands @code{set environment} and @code{unset
1906 environment} to change parts of the environment that affect
1907 your program. @xref{Environment, ,Your Program's Environment}.
1909 @item The @emph{working directory.}
1910 Your program inherits its working directory from @value{GDBN}. You can set
1911 the @value{GDBN} working directory with the @code{cd} command in @value{GDBN}.
1912 @xref{Working Directory, ,Your Program's Working Directory}.
1914 @item The @emph{standard input and output.}
1915 Your program normally uses the same device for standard input and
1916 standard output as @value{GDBN} is using. You can redirect input and output
1917 in the @code{run} command line, or you can use the @code{tty} command to
1918 set a different device for your program.
1919 @xref{Input/Output, ,Your Program's Input and Output}.
1922 @emph{Warning:} While input and output redirection work, you cannot use
1923 pipes to pass the output of the program you are debugging to another
1924 program; if you attempt this, @value{GDBN} is likely to wind up debugging the
1928 When you issue the @code{run} command, your program begins to execute
1929 immediately. @xref{Stopping, ,Stopping and Continuing}, for discussion
1930 of how to arrange for your program to stop. Once your program has
1931 stopped, you may call functions in your program, using the @code{print}
1932 or @code{call} commands. @xref{Data, ,Examining Data}.
1934 If the modification time of your symbol file has changed since the last
1935 time @value{GDBN} read its symbols, @value{GDBN} discards its symbol
1936 table, and reads it again. When it does this, @value{GDBN} tries to retain
1937 your current breakpoints.
1942 @cindex run to main procedure
1943 The name of the main procedure can vary from language to language.
1944 With C or C@t{++}, the main procedure name is always @code{main}, but
1945 other languages such as Ada do not require a specific name for their
1946 main procedure. The debugger provides a convenient way to start the
1947 execution of the program and to stop at the beginning of the main
1948 procedure, depending on the language used.
1950 The @samp{start} command does the equivalent of setting a temporary
1951 breakpoint at the beginning of the main procedure and then invoking
1952 the @samp{run} command.
1954 @cindex elaboration phase
1955 Some programs contain an @dfn{elaboration} phase where some startup code is
1956 executed before the main procedure is called. This depends on the
1957 languages used to write your program. In C@t{++}, for instance,
1958 constructors for static and global objects are executed before
1959 @code{main} is called. It is therefore possible that the debugger stops
1960 before reaching the main procedure. However, the temporary breakpoint
1961 will remain to halt execution.
1963 Specify the arguments to give to your program as arguments to the
1964 @samp{start} command. These arguments will be given verbatim to the
1965 underlying @samp{run} command. Note that the same arguments will be
1966 reused if no argument is provided during subsequent calls to
1967 @samp{start} or @samp{run}.
1969 It is sometimes necessary to debug the program during elaboration. In
1970 these cases, using the @code{start} command would stop the execution of
1971 your program too late, as the program would have already completed the
1972 elaboration phase. Under these circumstances, insert breakpoints in your
1973 elaboration code before running your program.
1975 @kindex set exec-wrapper
1976 @item set exec-wrapper @var{wrapper}
1977 @itemx show exec-wrapper
1978 @itemx unset exec-wrapper
1979 When @samp{exec-wrapper} is set, the specified wrapper is used to
1980 launch programs for debugging. @value{GDBN} starts your program
1981 with a shell command of the form @kbd{exec @var{wrapper}
1982 @var{program}}. Quoting is added to @var{program} and its
1983 arguments, but not to @var{wrapper}, so you should add quotes if
1984 appropriate for your shell. The wrapper runs until it executes
1985 your program, and then @value{GDBN} takes control.
1987 You can use any program that eventually calls @code{execve} with
1988 its arguments as a wrapper. Several standard Unix utilities do
1989 this, e.g.@: @code{env} and @code{nohup}. Any Unix shell script ending
1990 with @code{exec "$@@"} will also work.
1992 For example, you can use @code{env} to pass an environment variable to
1993 the debugged program, without setting the variable in your shell's
1997 (@value{GDBP}) set exec-wrapper env 'LD_PRELOAD=libtest.so'
2001 This command is available when debugging locally on most targets, excluding
2002 @sc{djgpp}, Cygwin, MS Windows, and QNX Neutrino.
2004 @kindex set disable-randomization
2005 @item set disable-randomization
2006 @itemx set disable-randomization on
2007 This option (enabled by default in @value{GDBN}) will turn off the native
2008 randomization of the virtual address space of the started program. This option
2009 is useful for multiple debugging sessions to make the execution better
2010 reproducible and memory addresses reusable across debugging sessions.
2012 This feature is implemented only on @sc{gnu}/Linux. You can get the same
2016 (@value{GDBP}) set exec-wrapper setarch `uname -m` -R
2019 @item set disable-randomization off
2020 Leave the behavior of the started executable unchanged. Some bugs rear their
2021 ugly heads only when the program is loaded at certain addresses. If your bug
2022 disappears when you run the program under @value{GDBN}, that might be because
2023 @value{GDBN} by default disables the address randomization on platforms, such
2024 as @sc{gnu}/Linux, which do that for stand-alone programs. Use @kbd{set
2025 disable-randomization off} to try to reproduce such elusive bugs.
2027 The virtual address space randomization is implemented only on @sc{gnu}/Linux.
2028 It protects the programs against some kinds of security attacks. In these
2029 cases the attacker needs to know the exact location of a concrete executable
2030 code. Randomizing its location makes it impossible to inject jumps misusing
2031 a code at its expected addresses.
2033 Prelinking shared libraries provides a startup performance advantage but it
2034 makes addresses in these libraries predictable for privileged processes by
2035 having just unprivileged access at the target system. Reading the shared
2036 library binary gives enough information for assembling the malicious code
2037 misusing it. Still even a prelinked shared library can get loaded at a new
2038 random address just requiring the regular relocation process during the
2039 startup. Shared libraries not already prelinked are always loaded at
2040 a randomly chosen address.
2042 Position independent executables (PIE) contain position independent code
2043 similar to the shared libraries and therefore such executables get loaded at
2044 a randomly chosen address upon startup. PIE executables always load even
2045 already prelinked shared libraries at a random address. You can build such
2046 executable using @command{gcc -fPIE -pie}.
2048 Heap (malloc storage), stack and custom mmap areas are always placed randomly
2049 (as long as the randomization is enabled).
2051 @item show disable-randomization
2052 Show the current setting of the explicit disable of the native randomization of
2053 the virtual address space of the started program.
2058 @section Your Program's Arguments
2060 @cindex arguments (to your program)
2061 The arguments to your program can be specified by the arguments of the
2063 They are passed to a shell, which expands wildcard characters and
2064 performs redirection of I/O, and thence to your program. Your
2065 @code{SHELL} environment variable (if it exists) specifies what shell
2066 @value{GDBN} uses. If you do not define @code{SHELL}, @value{GDBN} uses
2067 the default shell (@file{/bin/sh} on Unix).
2069 On non-Unix systems, the program is usually invoked directly by
2070 @value{GDBN}, which emulates I/O redirection via the appropriate system
2071 calls, and the wildcard characters are expanded by the startup code of
2072 the program, not by the shell.
2074 @code{run} with no arguments uses the same arguments used by the previous
2075 @code{run}, or those set by the @code{set args} command.
2080 Specify the arguments to be used the next time your program is run. If
2081 @code{set args} has no arguments, @code{run} executes your program
2082 with no arguments. Once you have run your program with arguments,
2083 using @code{set args} before the next @code{run} is the only way to run
2084 it again without arguments.
2088 Show the arguments to give your program when it is started.
2092 @section Your Program's Environment
2094 @cindex environment (of your program)
2095 The @dfn{environment} consists of a set of environment variables and
2096 their values. Environment variables conventionally record such things as
2097 your user name, your home directory, your terminal type, and your search
2098 path for programs to run. Usually you set up environment variables with
2099 the shell and they are inherited by all the other programs you run. When
2100 debugging, it can be useful to try running your program with a modified
2101 environment without having to start @value{GDBN} over again.
2105 @item path @var{directory}
2106 Add @var{directory} to the front of the @code{PATH} environment variable
2107 (the search path for executables) that will be passed to your program.
2108 The value of @code{PATH} used by @value{GDBN} does not change.
2109 You may specify several directory names, separated by whitespace or by a
2110 system-dependent separator character (@samp{:} on Unix, @samp{;} on
2111 MS-DOS and MS-Windows). If @var{directory} is already in the path, it
2112 is moved to the front, so it is searched sooner.
2114 You can use the string @samp{$cwd} to refer to whatever is the current
2115 working directory at the time @value{GDBN} searches the path. If you
2116 use @samp{.} instead, it refers to the directory where you executed the
2117 @code{path} command. @value{GDBN} replaces @samp{.} in the
2118 @var{directory} argument (with the current path) before adding
2119 @var{directory} to the search path.
2120 @c 'path' is explicitly nonrepeatable, but RMS points out it is silly to
2121 @c document that, since repeating it would be a no-op.
2125 Display the list of search paths for executables (the @code{PATH}
2126 environment variable).
2128 @kindex show environment
2129 @item show environment @r{[}@var{varname}@r{]}
2130 Print the value of environment variable @var{varname} to be given to
2131 your program when it starts. If you do not supply @var{varname},
2132 print the names and values of all environment variables to be given to
2133 your program. You can abbreviate @code{environment} as @code{env}.
2135 @kindex set environment
2136 @item set environment @var{varname} @r{[}=@var{value}@r{]}
2137 Set environment variable @var{varname} to @var{value}. The value
2138 changes for your program only, not for @value{GDBN} itself. @var{value} may
2139 be any string; the values of environment variables are just strings, and
2140 any interpretation is supplied by your program itself. The @var{value}
2141 parameter is optional; if it is eliminated, the variable is set to a
2143 @c "any string" here does not include leading, trailing
2144 @c blanks. Gnu asks: does anyone care?
2146 For example, this command:
2153 tells the debugged program, when subsequently run, that its user is named
2154 @samp{foo}. (The spaces around @samp{=} are used for clarity here; they
2155 are not actually required.)
2157 @kindex unset environment
2158 @item unset environment @var{varname}
2159 Remove variable @var{varname} from the environment to be passed to your
2160 program. This is different from @samp{set env @var{varname} =};
2161 @code{unset environment} removes the variable from the environment,
2162 rather than assigning it an empty value.
2165 @emph{Warning:} On Unix systems, @value{GDBN} runs your program using
2167 by your @code{SHELL} environment variable if it exists (or
2168 @code{/bin/sh} if not). If your @code{SHELL} variable names a shell
2169 that runs an initialization file---such as @file{.cshrc} for C-shell, or
2170 @file{.bashrc} for BASH---any variables you set in that file affect
2171 your program. You may wish to move setting of environment variables to
2172 files that are only run when you sign on, such as @file{.login} or
2175 @node Working Directory
2176 @section Your Program's Working Directory
2178 @cindex working directory (of your program)
2179 Each time you start your program with @code{run}, it inherits its
2180 working directory from the current working directory of @value{GDBN}.
2181 The @value{GDBN} working directory is initially whatever it inherited
2182 from its parent process (typically the shell), but you can specify a new
2183 working directory in @value{GDBN} with the @code{cd} command.
2185 The @value{GDBN} working directory also serves as a default for the commands
2186 that specify files for @value{GDBN} to operate on. @xref{Files, ,Commands to
2191 @cindex change working directory
2192 @item cd @var{directory}
2193 Set the @value{GDBN} working directory to @var{directory}.
2197 Print the @value{GDBN} working directory.
2200 It is generally impossible to find the current working directory of
2201 the process being debugged (since a program can change its directory
2202 during its run). If you work on a system where @value{GDBN} is
2203 configured with the @file{/proc} support, you can use the @code{info
2204 proc} command (@pxref{SVR4 Process Information}) to find out the
2205 current working directory of the debuggee.
2208 @section Your Program's Input and Output
2213 By default, the program you run under @value{GDBN} does input and output to
2214 the same terminal that @value{GDBN} uses. @value{GDBN} switches the terminal
2215 to its own terminal modes to interact with you, but it records the terminal
2216 modes your program was using and switches back to them when you continue
2217 running your program.
2220 @kindex info terminal
2222 Displays information recorded by @value{GDBN} about the terminal modes your
2226 You can redirect your program's input and/or output using shell
2227 redirection with the @code{run} command. For example,
2234 starts your program, diverting its output to the file @file{outfile}.
2237 @cindex controlling terminal
2238 Another way to specify where your program should do input and output is
2239 with the @code{tty} command. This command accepts a file name as
2240 argument, and causes this file to be the default for future @code{run}
2241 commands. It also resets the controlling terminal for the child
2242 process, for future @code{run} commands. For example,
2249 directs that processes started with subsequent @code{run} commands
2250 default to do input and output on the terminal @file{/dev/ttyb} and have
2251 that as their controlling terminal.
2253 An explicit redirection in @code{run} overrides the @code{tty} command's
2254 effect on the input/output device, but not its effect on the controlling
2257 When you use the @code{tty} command or redirect input in the @code{run}
2258 command, only the input @emph{for your program} is affected. The input
2259 for @value{GDBN} still comes from your terminal. @code{tty} is an alias
2260 for @code{set inferior-tty}.
2262 @cindex inferior tty
2263 @cindex set inferior controlling terminal
2264 You can use the @code{show inferior-tty} command to tell @value{GDBN} to
2265 display the name of the terminal that will be used for future runs of your
2269 @item set inferior-tty /dev/ttyb
2270 @kindex set inferior-tty
2271 Set the tty for the program being debugged to /dev/ttyb.
2273 @item show inferior-tty
2274 @kindex show inferior-tty
2275 Show the current tty for the program being debugged.
2279 @section Debugging an Already-running Process
2284 @item attach @var{process-id}
2285 This command attaches to a running process---one that was started
2286 outside @value{GDBN}. (@code{info files} shows your active
2287 targets.) The command takes as argument a process ID. The usual way to
2288 find out the @var{process-id} of a Unix process is with the @code{ps} utility,
2289 or with the @samp{jobs -l} shell command.
2291 @code{attach} does not repeat if you press @key{RET} a second time after
2292 executing the command.
2295 To use @code{attach}, your program must be running in an environment
2296 which supports processes; for example, @code{attach} does not work for
2297 programs on bare-board targets that lack an operating system. You must
2298 also have permission to send the process a signal.
2300 When you use @code{attach}, the debugger finds the program running in
2301 the process first by looking in the current working directory, then (if
2302 the program is not found) by using the source file search path
2303 (@pxref{Source Path, ,Specifying Source Directories}). You can also use
2304 the @code{file} command to load the program. @xref{Files, ,Commands to
2307 The first thing @value{GDBN} does after arranging to debug the specified
2308 process is to stop it. You can examine and modify an attached process
2309 with all the @value{GDBN} commands that are ordinarily available when
2310 you start processes with @code{run}. You can insert breakpoints; you
2311 can step and continue; you can modify storage. If you would rather the
2312 process continue running, you may use the @code{continue} command after
2313 attaching @value{GDBN} to the process.
2318 When you have finished debugging the attached process, you can use the
2319 @code{detach} command to release it from @value{GDBN} control. Detaching
2320 the process continues its execution. After the @code{detach} command,
2321 that process and @value{GDBN} become completely independent once more, and you
2322 are ready to @code{attach} another process or start one with @code{run}.
2323 @code{detach} does not repeat if you press @key{RET} again after
2324 executing the command.
2327 If you exit @value{GDBN} while you have an attached process, you detach
2328 that process. If you use the @code{run} command, you kill that process.
2329 By default, @value{GDBN} asks for confirmation if you try to do either of these
2330 things; you can control whether or not you need to confirm by using the
2331 @code{set confirm} command (@pxref{Messages/Warnings, ,Optional Warnings and
2335 @section Killing the Child Process
2340 Kill the child process in which your program is running under @value{GDBN}.
2343 This command is useful if you wish to debug a core dump instead of a
2344 running process. @value{GDBN} ignores any core dump file while your program
2347 On some operating systems, a program cannot be executed outside @value{GDBN}
2348 while you have breakpoints set on it inside @value{GDBN}. You can use the
2349 @code{kill} command in this situation to permit running your program
2350 outside the debugger.
2352 The @code{kill} command is also useful if you wish to recompile and
2353 relink your program, since on many systems it is impossible to modify an
2354 executable file while it is running in a process. In this case, when you
2355 next type @code{run}, @value{GDBN} notices that the file has changed, and
2356 reads the symbol table again (while trying to preserve your current
2357 breakpoint settings).
2359 @node Inferiors and Programs
2360 @section Debugging Multiple Inferiors and Programs
2362 @value{GDBN} lets you run and debug multiple programs in a single
2363 session. In addition, @value{GDBN} on some systems may let you run
2364 several programs simultaneously (otherwise you have to exit from one
2365 before starting another). In the most general case, you can have
2366 multiple threads of execution in each of multiple processes, launched
2367 from multiple executables.
2370 @value{GDBN} represents the state of each program execution with an
2371 object called an @dfn{inferior}. An inferior typically corresponds to
2372 a process, but is more general and applies also to targets that do not
2373 have processes. Inferiors may be created before a process runs, and
2374 may be retained after a process exits. Inferiors have unique
2375 identifiers that are different from process ids. Usually each
2376 inferior will also have its own distinct address space, although some
2377 embedded targets may have several inferiors running in different parts
2378 of a single address space. Each inferior may in turn have multiple
2379 threads running in it.
2381 To find out what inferiors exist at any moment, use @w{@code{info
2385 @kindex info inferiors
2386 @item info inferiors
2387 Print a list of all inferiors currently being managed by @value{GDBN}.
2389 @value{GDBN} displays for each inferior (in this order):
2393 the inferior number assigned by @value{GDBN}
2396 the target system's inferior identifier
2399 the name of the executable the inferior is running.
2404 An asterisk @samp{*} preceding the @value{GDBN} inferior number
2405 indicates the current inferior.
2409 @c end table here to get a little more width for example
2412 (@value{GDBP}) info inferiors
2413 Num Description Executable
2414 2 process 2307 hello
2415 * 1 process 3401 goodbye
2418 To switch focus between inferiors, use the @code{inferior} command:
2421 @kindex inferior @var{infno}
2422 @item inferior @var{infno}
2423 Make inferior number @var{infno} the current inferior. The argument
2424 @var{infno} is the inferior number assigned by @value{GDBN}, as shown
2425 in the first field of the @samp{info inferiors} display.
2429 You can get multiple executables into a debugging session via the
2430 @code{add-inferior} and @w{@code{clone-inferior}} commands. On some
2431 systems @value{GDBN} can add inferiors to the debug session
2432 automatically by following calls to @code{fork} and @code{exec}. To
2433 remove inferiors from the debugging session use the
2434 @w{@code{remove-inferior}} command.
2437 @kindex add-inferior
2438 @item add-inferior [ -copies @var{n} ] [ -exec @var{executable} ]
2439 Adds @var{n} inferiors to be run using @var{executable} as the
2440 executable. @var{n} defaults to 1. If no executable is specified,
2441 the inferiors begins empty, with no program. You can still assign or
2442 change the program assigned to the inferior at any time by using the
2443 @code{file} command with the executable name as its argument.
2445 @kindex clone-inferior
2446 @item clone-inferior [ -copies @var{n} ] [ @var{infno} ]
2447 Adds @var{n} inferiors ready to execute the same program as inferior
2448 @var{infno}. @var{n} defaults to 1. @var{infno} defaults to the
2449 number of the current inferior. This is a convenient command when you
2450 want to run another instance of the inferior you are debugging.
2453 (@value{GDBP}) info inferiors
2454 Num Description Executable
2455 * 1 process 29964 helloworld
2456 (@value{GDBP}) clone-inferior
2459 (@value{GDBP}) info inferiors
2460 Num Description Executable
2462 * 1 process 29964 helloworld
2465 You can now simply switch focus to inferior 2 and run it.
2467 @kindex remove-inferior
2468 @item remove-inferior @var{infno}
2469 Removes the inferior @var{infno}. It is not possible to remove an
2470 inferior that is running with this command. For those, use the
2471 @code{kill} or @code{detach} command first.
2475 To quit debugging one of the running inferiors that is not the current
2476 inferior, you can either detach from it by using the @w{@code{detach
2477 inferior}} command (allowing it to run independently), or kill it
2478 using the @w{@code{kill inferior}} command:
2481 @kindex detach inferior @var{infno}
2482 @item detach inferior @var{infno}
2483 Detach from the inferior identified by @value{GDBN} inferior number
2484 @var{infno}, and remove it from the inferior list.
2486 @kindex kill inferior @var{infno}
2487 @item kill inferior @var{infno}
2488 Kill the inferior identified by @value{GDBN} inferior number
2489 @var{infno}, and remove it from the inferior list.
2492 After the successful completion of a command such as @code{detach},
2493 @code{detach inferior}, @code{kill} or @code{kill inferior}, or after
2494 a normal process exit, the inferior is still valid and listed with
2495 @code{info inferiors}, ready to be restarted.
2498 To be notified when inferiors are started or exit under @value{GDBN}'s
2499 control use @w{@code{set print inferior-events}}:
2502 @kindex set print inferior-events
2503 @cindex print messages on inferior start and exit
2504 @item set print inferior-events
2505 @itemx set print inferior-events on
2506 @itemx set print inferior-events off
2507 The @code{set print inferior-events} command allows you to enable or
2508 disable printing of messages when @value{GDBN} notices that new
2509 inferiors have started or that inferiors have exited or have been
2510 detached. By default, these messages will not be printed.
2512 @kindex show print inferior-events
2513 @item show print inferior-events
2514 Show whether messages will be printed when @value{GDBN} detects that
2515 inferiors have started, exited or have been detached.
2518 Many commands will work the same with multiple programs as with a
2519 single program: e.g., @code{print myglobal} will simply display the
2520 value of @code{myglobal} in the current inferior.
2523 Occasionaly, when debugging @value{GDBN} itself, it may be useful to
2524 get more info about the relationship of inferiors, programs, address
2525 spaces in a debug session. You can do that with the @w{@code{maint
2526 info program-spaces}} command.
2529 @kindex maint info program-spaces
2530 @item maint info program-spaces
2531 Print a list of all program spaces currently being managed by
2534 @value{GDBN} displays for each program space (in this order):
2538 the program space number assigned by @value{GDBN}
2541 the name of the executable loaded into the program space, with e.g.,
2542 the @code{file} command.
2547 An asterisk @samp{*} preceding the @value{GDBN} program space number
2548 indicates the current program space.
2550 In addition, below each program space line, @value{GDBN} prints extra
2551 information that isn't suitable to display in tabular form. For
2552 example, the list of inferiors bound to the program space.
2555 (@value{GDBP}) maint info program-spaces
2558 Bound inferiors: ID 1 (process 21561)
2562 Here we can see that no inferior is running the program @code{hello},
2563 while @code{process 21561} is running the program @code{goodbye}. On
2564 some targets, it is possible that multiple inferiors are bound to the
2565 same program space. The most common example is that of debugging both
2566 the parent and child processes of a @code{vfork} call. For example,
2569 (@value{GDBP}) maint info program-spaces
2572 Bound inferiors: ID 2 (process 18050), ID 1 (process 18045)
2575 Here, both inferior 2 and inferior 1 are running in the same program
2576 space as a result of inferior 1 having executed a @code{vfork} call.
2580 @section Debugging Programs with Multiple Threads
2582 @cindex threads of execution
2583 @cindex multiple threads
2584 @cindex switching threads
2585 In some operating systems, such as HP-UX and Solaris, a single program
2586 may have more than one @dfn{thread} of execution. The precise semantics
2587 of threads differ from one operating system to another, but in general
2588 the threads of a single program are akin to multiple processes---except
2589 that they share one address space (that is, they can all examine and
2590 modify the same variables). On the other hand, each thread has its own
2591 registers and execution stack, and perhaps private memory.
2593 @value{GDBN} provides these facilities for debugging multi-thread
2597 @item automatic notification of new threads
2598 @item @samp{thread @var{threadno}}, a command to switch among threads
2599 @item @samp{info threads}, a command to inquire about existing threads
2600 @item @samp{thread apply [@var{threadno}] [@var{all}] @var{args}},
2601 a command to apply a command to a list of threads
2602 @item thread-specific breakpoints
2603 @item @samp{set print thread-events}, which controls printing of
2604 messages on thread start and exit.
2605 @item @samp{set libthread-db-search-path @var{path}}, which lets
2606 the user specify which @code{libthread_db} to use if the default choice
2607 isn't compatible with the program.
2611 @emph{Warning:} These facilities are not yet available on every
2612 @value{GDBN} configuration where the operating system supports threads.
2613 If your @value{GDBN} does not support threads, these commands have no
2614 effect. For example, a system without thread support shows no output
2615 from @samp{info threads}, and always rejects the @code{thread} command,
2619 (@value{GDBP}) info threads
2620 (@value{GDBP}) thread 1
2621 Thread ID 1 not known. Use the "info threads" command to
2622 see the IDs of currently known threads.
2624 @c FIXME to implementors: how hard would it be to say "sorry, this GDB
2625 @c doesn't support threads"?
2628 @cindex focus of debugging
2629 @cindex current thread
2630 The @value{GDBN} thread debugging facility allows you to observe all
2631 threads while your program runs---but whenever @value{GDBN} takes
2632 control, one thread in particular is always the focus of debugging.
2633 This thread is called the @dfn{current thread}. Debugging commands show
2634 program information from the perspective of the current thread.
2636 @cindex @code{New} @var{systag} message
2637 @cindex thread identifier (system)
2638 @c FIXME-implementors!! It would be more helpful if the [New...] message
2639 @c included GDB's numeric thread handle, so you could just go to that
2640 @c thread without first checking `info threads'.
2641 Whenever @value{GDBN} detects a new thread in your program, it displays
2642 the target system's identification for the thread with a message in the
2643 form @samp{[New @var{systag}]}. @var{systag} is a thread identifier
2644 whose form varies depending on the particular system. For example, on
2645 @sc{gnu}/Linux, you might see
2648 [New Thread 46912507313328 (LWP 25582)]
2652 when @value{GDBN} notices a new thread. In contrast, on an SGI system,
2653 the @var{systag} is simply something like @samp{process 368}, with no
2656 @c FIXME!! (1) Does the [New...] message appear even for the very first
2657 @c thread of a program, or does it only appear for the
2658 @c second---i.e.@: when it becomes obvious we have a multithread
2660 @c (2) *Is* there necessarily a first thread always? Or do some
2661 @c multithread systems permit starting a program with multiple
2662 @c threads ab initio?
2664 @cindex thread number
2665 @cindex thread identifier (GDB)
2666 For debugging purposes, @value{GDBN} associates its own thread
2667 number---always a single integer---with each thread in your program.
2670 @kindex info threads
2672 Display a summary of all threads currently in your
2673 program. @value{GDBN} displays for each thread (in this order):
2677 the thread number assigned by @value{GDBN}
2680 the target system's thread identifier (@var{systag})
2683 the current stack frame summary for that thread
2687 An asterisk @samp{*} to the left of the @value{GDBN} thread number
2688 indicates the current thread.
2692 @c end table here to get a little more width for example
2695 (@value{GDBP}) info threads
2696 3 process 35 thread 27 0x34e5 in sigpause ()
2697 2 process 35 thread 23 0x34e5 in sigpause ()
2698 * 1 process 35 thread 13 main (argc=1, argv=0x7ffffff8)
2704 @cindex debugging multithreaded programs (on HP-UX)
2705 @cindex thread identifier (GDB), on HP-UX
2706 For debugging purposes, @value{GDBN} associates its own thread
2707 number---a small integer assigned in thread-creation order---with each
2708 thread in your program.
2710 @cindex @code{New} @var{systag} message, on HP-UX
2711 @cindex thread identifier (system), on HP-UX
2712 @c FIXME-implementors!! It would be more helpful if the [New...] message
2713 @c included GDB's numeric thread handle, so you could just go to that
2714 @c thread without first checking `info threads'.
2715 Whenever @value{GDBN} detects a new thread in your program, it displays
2716 both @value{GDBN}'s thread number and the target system's identification for the thread with a message in the
2717 form @samp{[New @var{systag}]}. @var{systag} is a thread identifier
2718 whose form varies depending on the particular system. For example, on
2722 [New thread 2 (system thread 26594)]
2726 when @value{GDBN} notices a new thread.
2729 @kindex info threads (HP-UX)
2731 Display a summary of all threads currently in your
2732 program. @value{GDBN} displays for each thread (in this order):
2735 @item the thread number assigned by @value{GDBN}
2737 @item the target system's thread identifier (@var{systag})
2739 @item the current stack frame summary for that thread
2743 An asterisk @samp{*} to the left of the @value{GDBN} thread number
2744 indicates the current thread.
2748 @c end table here to get a little more width for example
2751 (@value{GDBP}) info threads
2752 * 3 system thread 26607 worker (wptr=0x7b09c318 "@@") \@*
2754 2 system thread 26606 0x7b0030d8 in __ksleep () \@*
2755 from /usr/lib/libc.2
2756 1 system thread 27905 0x7b003498 in _brk () \@*
2757 from /usr/lib/libc.2
2760 On Solaris, you can display more information about user threads with a
2761 Solaris-specific command:
2764 @item maint info sol-threads
2765 @kindex maint info sol-threads
2766 @cindex thread info (Solaris)
2767 Display info on Solaris user threads.
2771 @kindex thread @var{threadno}
2772 @item thread @var{threadno}
2773 Make thread number @var{threadno} the current thread. The command
2774 argument @var{threadno} is the internal @value{GDBN} thread number, as
2775 shown in the first field of the @samp{info threads} display.
2776 @value{GDBN} responds by displaying the system identifier of the thread
2777 you selected, and its current stack frame summary:
2780 @c FIXME!! This example made up; find a @value{GDBN} w/threads and get real one
2781 (@value{GDBP}) thread 2
2782 [Switching to process 35 thread 23]
2783 0x34e5 in sigpause ()
2787 As with the @samp{[New @dots{}]} message, the form of the text after
2788 @samp{Switching to} depends on your system's conventions for identifying
2791 @kindex thread apply
2792 @cindex apply command to several threads
2793 @item thread apply [@var{threadno}] [@var{all}] @var{command}
2794 The @code{thread apply} command allows you to apply the named
2795 @var{command} to one or more threads. Specify the numbers of the
2796 threads that you want affected with the command argument
2797 @var{threadno}. It can be a single thread number, one of the numbers
2798 shown in the first field of the @samp{info threads} display; or it
2799 could be a range of thread numbers, as in @code{2-4}. To apply a
2800 command to all threads, type @kbd{thread apply all @var{command}}.
2802 @kindex set print thread-events
2803 @cindex print messages on thread start and exit
2804 @item set print thread-events
2805 @itemx set print thread-events on
2806 @itemx set print thread-events off
2807 The @code{set print thread-events} command allows you to enable or
2808 disable printing of messages when @value{GDBN} notices that new threads have
2809 started or that threads have exited. By default, these messages will
2810 be printed if detection of these events is supported by the target.
2811 Note that these messages cannot be disabled on all targets.
2813 @kindex show print thread-events
2814 @item show print thread-events
2815 Show whether messages will be printed when @value{GDBN} detects that threads
2816 have started and exited.
2819 @xref{Thread Stops,,Stopping and Starting Multi-thread Programs}, for
2820 more information about how @value{GDBN} behaves when you stop and start
2821 programs with multiple threads.
2823 @xref{Set Watchpoints,,Setting Watchpoints}, for information about
2824 watchpoints in programs with multiple threads.
2827 @kindex set libthread-db-search-path
2828 @cindex search path for @code{libthread_db}
2829 @item set libthread-db-search-path @r{[}@var{path}@r{]}
2830 If this variable is set, @var{path} is a colon-separated list of
2831 directories @value{GDBN} will use to search for @code{libthread_db}.
2832 If you omit @var{path}, @samp{libthread-db-search-path} will be reset to
2835 On @sc{gnu}/Linux and Solaris systems, @value{GDBN} uses a ``helper''
2836 @code{libthread_db} library to obtain information about threads in the
2837 inferior process. @value{GDBN} will use @samp{libthread-db-search-path}
2838 to find @code{libthread_db}. If that fails, @value{GDBN} will continue
2839 with default system shared library directories, and finally the directory
2840 from which @code{libpthread} was loaded in the inferior process.
2842 For any @code{libthread_db} library @value{GDBN} finds in above directories,
2843 @value{GDBN} attempts to initialize it with the current inferior process.
2844 If this initialization fails (which could happen because of a version
2845 mismatch between @code{libthread_db} and @code{libpthread}), @value{GDBN}
2846 will unload @code{libthread_db}, and continue with the next directory.
2847 If none of @code{libthread_db} libraries initialize successfully,
2848 @value{GDBN} will issue a warning and thread debugging will be disabled.
2850 Setting @code{libthread-db-search-path} is currently implemented
2851 only on some platforms.
2853 @kindex show libthread-db-search-path
2854 @item show libthread-db-search-path
2855 Display current libthread_db search path.
2859 @section Debugging Forks
2861 @cindex fork, debugging programs which call
2862 @cindex multiple processes
2863 @cindex processes, multiple
2864 On most systems, @value{GDBN} has no special support for debugging
2865 programs which create additional processes using the @code{fork}
2866 function. When a program forks, @value{GDBN} will continue to debug the
2867 parent process and the child process will run unimpeded. If you have
2868 set a breakpoint in any code which the child then executes, the child
2869 will get a @code{SIGTRAP} signal which (unless it catches the signal)
2870 will cause it to terminate.
2872 However, if you want to debug the child process there is a workaround
2873 which isn't too painful. Put a call to @code{sleep} in the code which
2874 the child process executes after the fork. It may be useful to sleep
2875 only if a certain environment variable is set, or a certain file exists,
2876 so that the delay need not occur when you don't want to run @value{GDBN}
2877 on the child. While the child is sleeping, use the @code{ps} program to
2878 get its process ID. Then tell @value{GDBN} (a new invocation of
2879 @value{GDBN} if you are also debugging the parent process) to attach to
2880 the child process (@pxref{Attach}). From that point on you can debug
2881 the child process just like any other process which you attached to.
2883 On some systems, @value{GDBN} provides support for debugging programs that
2884 create additional processes using the @code{fork} or @code{vfork} functions.
2885 Currently, the only platforms with this feature are HP-UX (11.x and later
2886 only?) and @sc{gnu}/Linux (kernel version 2.5.60 and later).
2888 By default, when a program forks, @value{GDBN} will continue to debug
2889 the parent process and the child process will run unimpeded.
2891 If you want to follow the child process instead of the parent process,
2892 use the command @w{@code{set follow-fork-mode}}.
2895 @kindex set follow-fork-mode
2896 @item set follow-fork-mode @var{mode}
2897 Set the debugger response to a program call of @code{fork} or
2898 @code{vfork}. A call to @code{fork} or @code{vfork} creates a new
2899 process. The @var{mode} argument can be:
2903 The original process is debugged after a fork. The child process runs
2904 unimpeded. This is the default.
2907 The new process is debugged after a fork. The parent process runs
2912 @kindex show follow-fork-mode
2913 @item show follow-fork-mode
2914 Display the current debugger response to a @code{fork} or @code{vfork} call.
2917 @cindex debugging multiple processes
2918 On Linux, if you want to debug both the parent and child processes, use the
2919 command @w{@code{set detach-on-fork}}.
2922 @kindex set detach-on-fork
2923 @item set detach-on-fork @var{mode}
2924 Tells gdb whether to detach one of the processes after a fork, or
2925 retain debugger control over them both.
2929 The child process (or parent process, depending on the value of
2930 @code{follow-fork-mode}) will be detached and allowed to run
2931 independently. This is the default.
2934 Both processes will be held under the control of @value{GDBN}.
2935 One process (child or parent, depending on the value of
2936 @code{follow-fork-mode}) is debugged as usual, while the other
2941 @kindex show detach-on-fork
2942 @item show detach-on-fork
2943 Show whether detach-on-fork mode is on/off.
2946 If you choose to set @samp{detach-on-fork} mode off, then @value{GDBN}
2947 will retain control of all forked processes (including nested forks).
2948 You can list the forked processes under the control of @value{GDBN} by
2949 using the @w{@code{info inferiors}} command, and switch from one fork
2950 to another by using the @code{inferior} command (@pxref{Inferiors and
2951 Programs, ,Debugging Multiple Inferiors and Programs}).
2953 To quit debugging one of the forked processes, you can either detach
2954 from it by using the @w{@code{detach inferior}} command (allowing it
2955 to run independently), or kill it using the @w{@code{kill inferior}}
2956 command. @xref{Inferiors and Programs, ,Debugging Multiple Inferiors
2959 If you ask to debug a child process and a @code{vfork} is followed by an
2960 @code{exec}, @value{GDBN} executes the new target up to the first
2961 breakpoint in the new target. If you have a breakpoint set on
2962 @code{main} in your original program, the breakpoint will also be set on
2963 the child process's @code{main}.
2965 On some systems, when a child process is spawned by @code{vfork}, you
2966 cannot debug the child or parent until an @code{exec} call completes.
2968 If you issue a @code{run} command to @value{GDBN} after an @code{exec}
2969 call executes, the new target restarts. To restart the parent
2970 process, use the @code{file} command with the parent executable name
2971 as its argument. By default, after an @code{exec} call executes,
2972 @value{GDBN} discards the symbols of the previous executable image.
2973 You can change this behaviour with the @w{@code{set follow-exec-mode}}
2977 @kindex set follow-exec-mode
2978 @item set follow-exec-mode @var{mode}
2980 Set debugger response to a program call of @code{exec}. An
2981 @code{exec} call replaces the program image of a process.
2983 @code{follow-exec-mode} can be:
2987 @value{GDBN} creates a new inferior and rebinds the process to this
2988 new inferior. The program the process was running before the
2989 @code{exec} call can be restarted afterwards by restarting the
2995 (@value{GDBP}) info inferiors
2997 Id Description Executable
3000 process 12020 is executing new program: prog2
3001 Program exited normally.
3002 (@value{GDBP}) info inferiors
3003 Id Description Executable
3009 @value{GDBN} keeps the process bound to the same inferior. The new
3010 executable image replaces the previous executable loaded in the
3011 inferior. Restarting the inferior after the @code{exec} call, with
3012 e.g., the @code{run} command, restarts the executable the process was
3013 running after the @code{exec} call. This is the default mode.
3018 (@value{GDBP}) info inferiors
3019 Id Description Executable
3022 process 12020 is executing new program: prog2
3023 Program exited normally.
3024 (@value{GDBP}) info inferiors
3025 Id Description Executable
3032 You can use the @code{catch} command to make @value{GDBN} stop whenever
3033 a @code{fork}, @code{vfork}, or @code{exec} call is made. @xref{Set
3034 Catchpoints, ,Setting Catchpoints}.
3036 @node Checkpoint/Restart
3037 @section Setting a @emph{Bookmark} to Return to Later
3042 @cindex snapshot of a process
3043 @cindex rewind program state
3045 On certain operating systems@footnote{Currently, only
3046 @sc{gnu}/Linux.}, @value{GDBN} is able to save a @dfn{snapshot} of a
3047 program's state, called a @dfn{checkpoint}, and come back to it
3050 Returning to a checkpoint effectively undoes everything that has
3051 happened in the program since the @code{checkpoint} was saved. This
3052 includes changes in memory, registers, and even (within some limits)
3053 system state. Effectively, it is like going back in time to the
3054 moment when the checkpoint was saved.
3056 Thus, if you're stepping thru a program and you think you're
3057 getting close to the point where things go wrong, you can save
3058 a checkpoint. Then, if you accidentally go too far and miss
3059 the critical statement, instead of having to restart your program
3060 from the beginning, you can just go back to the checkpoint and
3061 start again from there.
3063 This can be especially useful if it takes a lot of time or
3064 steps to reach the point where you think the bug occurs.
3066 To use the @code{checkpoint}/@code{restart} method of debugging:
3071 Save a snapshot of the debugged program's current execution state.
3072 The @code{checkpoint} command takes no arguments, but each checkpoint
3073 is assigned a small integer id, similar to a breakpoint id.
3075 @kindex info checkpoints
3076 @item info checkpoints
3077 List the checkpoints that have been saved in the current debugging
3078 session. For each checkpoint, the following information will be
3085 @item Source line, or label
3088 @kindex restart @var{checkpoint-id}
3089 @item restart @var{checkpoint-id}
3090 Restore the program state that was saved as checkpoint number
3091 @var{checkpoint-id}. All program variables, registers, stack frames
3092 etc.@: will be returned to the values that they had when the checkpoint
3093 was saved. In essence, gdb will ``wind back the clock'' to the point
3094 in time when the checkpoint was saved.
3096 Note that breakpoints, @value{GDBN} variables, command history etc.
3097 are not affected by restoring a checkpoint. In general, a checkpoint
3098 only restores things that reside in the program being debugged, not in
3101 @kindex delete checkpoint @var{checkpoint-id}
3102 @item delete checkpoint @var{checkpoint-id}
3103 Delete the previously-saved checkpoint identified by @var{checkpoint-id}.
3107 Returning to a previously saved checkpoint will restore the user state
3108 of the program being debugged, plus a significant subset of the system
3109 (OS) state, including file pointers. It won't ``un-write'' data from
3110 a file, but it will rewind the file pointer to the previous location,
3111 so that the previously written data can be overwritten. For files
3112 opened in read mode, the pointer will also be restored so that the
3113 previously read data can be read again.
3115 Of course, characters that have been sent to a printer (or other
3116 external device) cannot be ``snatched back'', and characters received
3117 from eg.@: a serial device can be removed from internal program buffers,
3118 but they cannot be ``pushed back'' into the serial pipeline, ready to
3119 be received again. Similarly, the actual contents of files that have
3120 been changed cannot be restored (at this time).
3122 However, within those constraints, you actually can ``rewind'' your
3123 program to a previously saved point in time, and begin debugging it
3124 again --- and you can change the course of events so as to debug a
3125 different execution path this time.
3127 @cindex checkpoints and process id
3128 Finally, there is one bit of internal program state that will be
3129 different when you return to a checkpoint --- the program's process
3130 id. Each checkpoint will have a unique process id (or @var{pid}),
3131 and each will be different from the program's original @var{pid}.
3132 If your program has saved a local copy of its process id, this could
3133 potentially pose a problem.
3135 @subsection A Non-obvious Benefit of Using Checkpoints
3137 On some systems such as @sc{gnu}/Linux, address space randomization
3138 is performed on new processes for security reasons. This makes it
3139 difficult or impossible to set a breakpoint, or watchpoint, on an
3140 absolute address if you have to restart the program, since the
3141 absolute location of a symbol will change from one execution to the
3144 A checkpoint, however, is an @emph{identical} copy of a process.
3145 Therefore if you create a checkpoint at (eg.@:) the start of main,
3146 and simply return to that checkpoint instead of restarting the
3147 process, you can avoid the effects of address randomization and
3148 your symbols will all stay in the same place.
3151 @chapter Stopping and Continuing
3153 The principal purposes of using a debugger are so that you can stop your
3154 program before it terminates; or so that, if your program runs into
3155 trouble, you can investigate and find out why.
3157 Inside @value{GDBN}, your program may stop for any of several reasons,
3158 such as a signal, a breakpoint, or reaching a new line after a
3159 @value{GDBN} command such as @code{step}. You may then examine and
3160 change variables, set new breakpoints or remove old ones, and then
3161 continue execution. Usually, the messages shown by @value{GDBN} provide
3162 ample explanation of the status of your program---but you can also
3163 explicitly request this information at any time.
3166 @kindex info program
3168 Display information about the status of your program: whether it is
3169 running or not, what process it is, and why it stopped.
3173 * Breakpoints:: Breakpoints, watchpoints, and catchpoints
3174 * Continuing and Stepping:: Resuming execution
3176 * Thread Stops:: Stopping and starting multi-thread programs
3180 @section Breakpoints, Watchpoints, and Catchpoints
3183 A @dfn{breakpoint} makes your program stop whenever a certain point in
3184 the program is reached. For each breakpoint, you can add conditions to
3185 control in finer detail whether your program stops. You can set
3186 breakpoints with the @code{break} command and its variants (@pxref{Set
3187 Breaks, ,Setting Breakpoints}), to specify the place where your program
3188 should stop by line number, function name or exact address in the
3191 On some systems, you can set breakpoints in shared libraries before
3192 the executable is run. There is a minor limitation on HP-UX systems:
3193 you must wait until the executable is run in order to set breakpoints
3194 in shared library routines that are not called directly by the program
3195 (for example, routines that are arguments in a @code{pthread_create}
3199 @cindex data breakpoints
3200 @cindex memory tracing
3201 @cindex breakpoint on memory address
3202 @cindex breakpoint on variable modification
3203 A @dfn{watchpoint} is a special breakpoint that stops your program
3204 when the value of an expression changes. The expression may be a value
3205 of a variable, or it could involve values of one or more variables
3206 combined by operators, such as @samp{a + b}. This is sometimes called
3207 @dfn{data breakpoints}. You must use a different command to set
3208 watchpoints (@pxref{Set Watchpoints, ,Setting Watchpoints}), but aside
3209 from that, you can manage a watchpoint like any other breakpoint: you
3210 enable, disable, and delete both breakpoints and watchpoints using the
3213 You can arrange to have values from your program displayed automatically
3214 whenever @value{GDBN} stops at a breakpoint. @xref{Auto Display,,
3218 @cindex breakpoint on events
3219 A @dfn{catchpoint} is another special breakpoint that stops your program
3220 when a certain kind of event occurs, such as the throwing of a C@t{++}
3221 exception or the loading of a library. As with watchpoints, you use a
3222 different command to set a catchpoint (@pxref{Set Catchpoints, ,Setting
3223 Catchpoints}), but aside from that, you can manage a catchpoint like any
3224 other breakpoint. (To stop when your program receives a signal, use the
3225 @code{handle} command; see @ref{Signals, ,Signals}.)
3227 @cindex breakpoint numbers
3228 @cindex numbers for breakpoints
3229 @value{GDBN} assigns a number to each breakpoint, watchpoint, or
3230 catchpoint when you create it; these numbers are successive integers
3231 starting with one. In many of the commands for controlling various
3232 features of breakpoints you use the breakpoint number to say which
3233 breakpoint you want to change. Each breakpoint may be @dfn{enabled} or
3234 @dfn{disabled}; if disabled, it has no effect on your program until you
3237 @cindex breakpoint ranges
3238 @cindex ranges of breakpoints
3239 Some @value{GDBN} commands accept a range of breakpoints on which to
3240 operate. A breakpoint range is either a single breakpoint number, like
3241 @samp{5}, or two such numbers, in increasing order, separated by a
3242 hyphen, like @samp{5-7}. When a breakpoint range is given to a command,
3243 all breakpoints in that range are operated on.
3246 * Set Breaks:: Setting breakpoints
3247 * Set Watchpoints:: Setting watchpoints
3248 * Set Catchpoints:: Setting catchpoints
3249 * Delete Breaks:: Deleting breakpoints
3250 * Disabling:: Disabling breakpoints
3251 * Conditions:: Break conditions
3252 * Break Commands:: Breakpoint command lists
3253 * Save Breakpoints:: How to save breakpoints in a file
3254 * Error in Breakpoints:: ``Cannot insert breakpoints''
3255 * Breakpoint-related Warnings:: ``Breakpoint address adjusted...''
3259 @subsection Setting Breakpoints
3261 @c FIXME LMB what does GDB do if no code on line of breakpt?
3262 @c consider in particular declaration with/without initialization.
3264 @c FIXME 2 is there stuff on this already? break at fun start, already init?
3267 @kindex b @r{(@code{break})}
3268 @vindex $bpnum@r{, convenience variable}
3269 @cindex latest breakpoint
3270 Breakpoints are set with the @code{break} command (abbreviated
3271 @code{b}). The debugger convenience variable @samp{$bpnum} records the
3272 number of the breakpoint you've set most recently; see @ref{Convenience
3273 Vars,, Convenience Variables}, for a discussion of what you can do with
3274 convenience variables.
3277 @item break @var{location}
3278 Set a breakpoint at the given @var{location}, which can specify a
3279 function name, a line number, or an address of an instruction.
3280 (@xref{Specify Location}, for a list of all the possible ways to
3281 specify a @var{location}.) The breakpoint will stop your program just
3282 before it executes any of the code in the specified @var{location}.
3284 When using source languages that permit overloading of symbols, such as
3285 C@t{++}, a function name may refer to more than one possible place to break.
3286 @xref{Ambiguous Expressions,,Ambiguous Expressions}, for a discussion of
3289 It is also possible to insert a breakpoint that will stop the program
3290 only if a specific thread (@pxref{Thread-Specific Breakpoints})
3291 or a specific task (@pxref{Ada Tasks}) hits that breakpoint.
3294 When called without any arguments, @code{break} sets a breakpoint at
3295 the next instruction to be executed in the selected stack frame
3296 (@pxref{Stack, ,Examining the Stack}). In any selected frame but the
3297 innermost, this makes your program stop as soon as control
3298 returns to that frame. This is similar to the effect of a
3299 @code{finish} command in the frame inside the selected frame---except
3300 that @code{finish} does not leave an active breakpoint. If you use
3301 @code{break} without an argument in the innermost frame, @value{GDBN} stops
3302 the next time it reaches the current location; this may be useful
3305 @value{GDBN} normally ignores breakpoints when it resumes execution, until at
3306 least one instruction has been executed. If it did not do this, you
3307 would be unable to proceed past a breakpoint without first disabling the
3308 breakpoint. This rule applies whether or not the breakpoint already
3309 existed when your program stopped.
3311 @item break @dots{} if @var{cond}
3312 Set a breakpoint with condition @var{cond}; evaluate the expression
3313 @var{cond} each time the breakpoint is reached, and stop only if the
3314 value is nonzero---that is, if @var{cond} evaluates as true.
3315 @samp{@dots{}} stands for one of the possible arguments described
3316 above (or no argument) specifying where to break. @xref{Conditions,
3317 ,Break Conditions}, for more information on breakpoint conditions.
3320 @item tbreak @var{args}
3321 Set a breakpoint enabled only for one stop. @var{args} are the
3322 same as for the @code{break} command, and the breakpoint is set in the same
3323 way, but the breakpoint is automatically deleted after the first time your
3324 program stops there. @xref{Disabling, ,Disabling Breakpoints}.
3327 @cindex hardware breakpoints
3328 @item hbreak @var{args}
3329 Set a hardware-assisted breakpoint. @var{args} are the same as for the
3330 @code{break} command and the breakpoint is set in the same way, but the
3331 breakpoint requires hardware support and some target hardware may not
3332 have this support. The main purpose of this is EPROM/ROM code
3333 debugging, so you can set a breakpoint at an instruction without
3334 changing the instruction. This can be used with the new trap-generation
3335 provided by SPARClite DSU and most x86-based targets. These targets
3336 will generate traps when a program accesses some data or instruction
3337 address that is assigned to the debug registers. However the hardware
3338 breakpoint registers can take a limited number of breakpoints. For
3339 example, on the DSU, only two data breakpoints can be set at a time, and
3340 @value{GDBN} will reject this command if more than two are used. Delete
3341 or disable unused hardware breakpoints before setting new ones
3342 (@pxref{Disabling, ,Disabling Breakpoints}).
3343 @xref{Conditions, ,Break Conditions}.
3344 For remote targets, you can restrict the number of hardware
3345 breakpoints @value{GDBN} will use, see @ref{set remote
3346 hardware-breakpoint-limit}.
3349 @item thbreak @var{args}
3350 Set a hardware-assisted breakpoint enabled only for one stop. @var{args}
3351 are the same as for the @code{hbreak} command and the breakpoint is set in
3352 the same way. However, like the @code{tbreak} command,
3353 the breakpoint is automatically deleted after the
3354 first time your program stops there. Also, like the @code{hbreak}
3355 command, the breakpoint requires hardware support and some target hardware
3356 may not have this support. @xref{Disabling, ,Disabling Breakpoints}.
3357 See also @ref{Conditions, ,Break Conditions}.
3360 @cindex regular expression
3361 @cindex breakpoints at functions matching a regexp
3362 @cindex set breakpoints in many functions
3363 @item rbreak @var{regex}
3364 Set breakpoints on all functions matching the regular expression
3365 @var{regex}. This command sets an unconditional breakpoint on all
3366 matches, printing a list of all breakpoints it set. Once these
3367 breakpoints are set, they are treated just like the breakpoints set with
3368 the @code{break} command. You can delete them, disable them, or make
3369 them conditional the same way as any other breakpoint.
3371 The syntax of the regular expression is the standard one used with tools
3372 like @file{grep}. Note that this is different from the syntax used by
3373 shells, so for instance @code{foo*} matches all functions that include
3374 an @code{fo} followed by zero or more @code{o}s. There is an implicit
3375 @code{.*} leading and trailing the regular expression you supply, so to
3376 match only functions that begin with @code{foo}, use @code{^foo}.
3378 @cindex non-member C@t{++} functions, set breakpoint in
3379 When debugging C@t{++} programs, @code{rbreak} is useful for setting
3380 breakpoints on overloaded functions that are not members of any special
3383 @cindex set breakpoints on all functions
3384 The @code{rbreak} command can be used to set breakpoints in
3385 @strong{all} the functions in a program, like this:
3388 (@value{GDBP}) rbreak .
3391 @item rbreak @var{file}:@var{regex}
3392 If @code{rbreak} is called with a filename qualification, it limits
3393 the search for functions matching the given regular expression to the
3394 specified @var{file}. This can be used, for example, to set breakpoints on
3395 every function in a given file:
3398 (@value{GDBP}) rbreak file.c:.
3401 The colon separating the filename qualifier from the regex may
3402 optionally be surrounded by spaces.
3404 @kindex info breakpoints
3405 @cindex @code{$_} and @code{info breakpoints}
3406 @item info breakpoints @r{[}@var{n}@r{]}
3407 @itemx info break @r{[}@var{n}@r{]}
3408 Print a table of all breakpoints, watchpoints, and catchpoints set and
3409 not deleted. Optional argument @var{n} means print information only
3410 about the specified breakpoint (or watchpoint or catchpoint). For
3411 each breakpoint, following columns are printed:
3414 @item Breakpoint Numbers
3416 Breakpoint, watchpoint, or catchpoint.
3418 Whether the breakpoint is marked to be disabled or deleted when hit.
3419 @item Enabled or Disabled
3420 Enabled breakpoints are marked with @samp{y}. @samp{n} marks breakpoints
3421 that are not enabled.
3423 Where the breakpoint is in your program, as a memory address. For a
3424 pending breakpoint whose address is not yet known, this field will
3425 contain @samp{<PENDING>}. Such breakpoint won't fire until a shared
3426 library that has the symbol or line referred by breakpoint is loaded.
3427 See below for details. A breakpoint with several locations will
3428 have @samp{<MULTIPLE>} in this field---see below for details.
3430 Where the breakpoint is in the source for your program, as a file and
3431 line number. For a pending breakpoint, the original string passed to
3432 the breakpoint command will be listed as it cannot be resolved until
3433 the appropriate shared library is loaded in the future.
3437 If a breakpoint is conditional, @code{info break} shows the condition on
3438 the line following the affected breakpoint; breakpoint commands, if any,
3439 are listed after that. A pending breakpoint is allowed to have a condition
3440 specified for it. The condition is not parsed for validity until a shared
3441 library is loaded that allows the pending breakpoint to resolve to a
3445 @code{info break} with a breakpoint
3446 number @var{n} as argument lists only that breakpoint. The
3447 convenience variable @code{$_} and the default examining-address for
3448 the @code{x} command are set to the address of the last breakpoint
3449 listed (@pxref{Memory, ,Examining Memory}).
3452 @code{info break} displays a count of the number of times the breakpoint
3453 has been hit. This is especially useful in conjunction with the
3454 @code{ignore} command. You can ignore a large number of breakpoint
3455 hits, look at the breakpoint info to see how many times the breakpoint
3456 was hit, and then run again, ignoring one less than that number. This
3457 will get you quickly to the last hit of that breakpoint.
3460 @value{GDBN} allows you to set any number of breakpoints at the same place in
3461 your program. There is nothing silly or meaningless about this. When
3462 the breakpoints are conditional, this is even useful
3463 (@pxref{Conditions, ,Break Conditions}).
3465 @cindex multiple locations, breakpoints
3466 @cindex breakpoints, multiple locations
3467 It is possible that a breakpoint corresponds to several locations
3468 in your program. Examples of this situation are:
3472 For a C@t{++} constructor, the @value{NGCC} compiler generates several
3473 instances of the function body, used in different cases.
3476 For a C@t{++} template function, a given line in the function can
3477 correspond to any number of instantiations.
3480 For an inlined function, a given source line can correspond to
3481 several places where that function is inlined.
3484 In all those cases, @value{GDBN} will insert a breakpoint at all
3485 the relevant locations@footnote{
3486 As of this writing, multiple-location breakpoints work only if there's
3487 line number information for all the locations. This means that they
3488 will generally not work in system libraries, unless you have debug
3489 info with line numbers for them.}.
3491 A breakpoint with multiple locations is displayed in the breakpoint
3492 table using several rows---one header row, followed by one row for
3493 each breakpoint location. The header row has @samp{<MULTIPLE>} in the
3494 address column. The rows for individual locations contain the actual
3495 addresses for locations, and show the functions to which those
3496 locations belong. The number column for a location is of the form
3497 @var{breakpoint-number}.@var{location-number}.
3502 Num Type Disp Enb Address What
3503 1 breakpoint keep y <MULTIPLE>
3505 breakpoint already hit 1 time
3506 1.1 y 0x080486a2 in void foo<int>() at t.cc:8
3507 1.2 y 0x080486ca in void foo<double>() at t.cc:8
3510 Each location can be individually enabled or disabled by passing
3511 @var{breakpoint-number}.@var{location-number} as argument to the
3512 @code{enable} and @code{disable} commands. Note that you cannot
3513 delete the individual locations from the list, you can only delete the
3514 entire list of locations that belong to their parent breakpoint (with
3515 the @kbd{delete @var{num}} command, where @var{num} is the number of
3516 the parent breakpoint, 1 in the above example). Disabling or enabling
3517 the parent breakpoint (@pxref{Disabling}) affects all of the locations
3518 that belong to that breakpoint.
3520 @cindex pending breakpoints
3521 It's quite common to have a breakpoint inside a shared library.
3522 Shared libraries can be loaded and unloaded explicitly,
3523 and possibly repeatedly, as the program is executed. To support
3524 this use case, @value{GDBN} updates breakpoint locations whenever
3525 any shared library is loaded or unloaded. Typically, you would
3526 set a breakpoint in a shared library at the beginning of your
3527 debugging session, when the library is not loaded, and when the
3528 symbols from the library are not available. When you try to set
3529 breakpoint, @value{GDBN} will ask you if you want to set
3530 a so called @dfn{pending breakpoint}---breakpoint whose address
3531 is not yet resolved.
3533 After the program is run, whenever a new shared library is loaded,
3534 @value{GDBN} reevaluates all the breakpoints. When a newly loaded
3535 shared library contains the symbol or line referred to by some
3536 pending breakpoint, that breakpoint is resolved and becomes an
3537 ordinary breakpoint. When a library is unloaded, all breakpoints
3538 that refer to its symbols or source lines become pending again.
3540 This logic works for breakpoints with multiple locations, too. For
3541 example, if you have a breakpoint in a C@t{++} template function, and
3542 a newly loaded shared library has an instantiation of that template,
3543 a new location is added to the list of locations for the breakpoint.
3545 Except for having unresolved address, pending breakpoints do not
3546 differ from regular breakpoints. You can set conditions or commands,
3547 enable and disable them and perform other breakpoint operations.
3549 @value{GDBN} provides some additional commands for controlling what
3550 happens when the @samp{break} command cannot resolve breakpoint
3551 address specification to an address:
3553 @kindex set breakpoint pending
3554 @kindex show breakpoint pending
3556 @item set breakpoint pending auto
3557 This is the default behavior. When @value{GDBN} cannot find the breakpoint
3558 location, it queries you whether a pending breakpoint should be created.
3560 @item set breakpoint pending on
3561 This indicates that an unrecognized breakpoint location should automatically
3562 result in a pending breakpoint being created.
3564 @item set breakpoint pending off
3565 This indicates that pending breakpoints are not to be created. Any
3566 unrecognized breakpoint location results in an error. This setting does
3567 not affect any pending breakpoints previously created.
3569 @item show breakpoint pending
3570 Show the current behavior setting for creating pending breakpoints.
3573 The settings above only affect the @code{break} command and its
3574 variants. Once breakpoint is set, it will be automatically updated
3575 as shared libraries are loaded and unloaded.
3577 @cindex automatic hardware breakpoints
3578 For some targets, @value{GDBN} can automatically decide if hardware or
3579 software breakpoints should be used, depending on whether the
3580 breakpoint address is read-only or read-write. This applies to
3581 breakpoints set with the @code{break} command as well as to internal
3582 breakpoints set by commands like @code{next} and @code{finish}. For
3583 breakpoints set with @code{hbreak}, @value{GDBN} will always use hardware
3586 You can control this automatic behaviour with the following commands::
3588 @kindex set breakpoint auto-hw
3589 @kindex show breakpoint auto-hw
3591 @item set breakpoint auto-hw on
3592 This is the default behavior. When @value{GDBN} sets a breakpoint, it
3593 will try to use the target memory map to decide if software or hardware
3594 breakpoint must be used.
3596 @item set breakpoint auto-hw off
3597 This indicates @value{GDBN} should not automatically select breakpoint
3598 type. If the target provides a memory map, @value{GDBN} will warn when
3599 trying to set software breakpoint at a read-only address.
3602 @value{GDBN} normally implements breakpoints by replacing the program code
3603 at the breakpoint address with a special instruction, which, when
3604 executed, given control to the debugger. By default, the program
3605 code is so modified only when the program is resumed. As soon as
3606 the program stops, @value{GDBN} restores the original instructions. This
3607 behaviour guards against leaving breakpoints inserted in the
3608 target should gdb abrubptly disconnect. However, with slow remote
3609 targets, inserting and removing breakpoint can reduce the performance.
3610 This behavior can be controlled with the following commands::
3612 @kindex set breakpoint always-inserted
3613 @kindex show breakpoint always-inserted
3615 @item set breakpoint always-inserted off
3616 All breakpoints, including newly added by the user, are inserted in
3617 the target only when the target is resumed. All breakpoints are
3618 removed from the target when it stops.
3620 @item set breakpoint always-inserted on
3621 Causes all breakpoints to be inserted in the target at all times. If
3622 the user adds a new breakpoint, or changes an existing breakpoint, the
3623 breakpoints in the target are updated immediately. A breakpoint is
3624 removed from the target only when breakpoint itself is removed.
3626 @cindex non-stop mode, and @code{breakpoint always-inserted}
3627 @item set breakpoint always-inserted auto
3628 This is the default mode. If @value{GDBN} is controlling the inferior
3629 in non-stop mode (@pxref{Non-Stop Mode}), gdb behaves as if
3630 @code{breakpoint always-inserted} mode is on. If @value{GDBN} is
3631 controlling the inferior in all-stop mode, @value{GDBN} behaves as if
3632 @code{breakpoint always-inserted} mode is off.
3635 @cindex negative breakpoint numbers
3636 @cindex internal @value{GDBN} breakpoints
3637 @value{GDBN} itself sometimes sets breakpoints in your program for
3638 special purposes, such as proper handling of @code{longjmp} (in C
3639 programs). These internal breakpoints are assigned negative numbers,
3640 starting with @code{-1}; @samp{info breakpoints} does not display them.
3641 You can see these breakpoints with the @value{GDBN} maintenance command
3642 @samp{maint info breakpoints} (@pxref{maint info breakpoints}).
3645 @node Set Watchpoints
3646 @subsection Setting Watchpoints
3648 @cindex setting watchpoints
3649 You can use a watchpoint to stop execution whenever the value of an
3650 expression changes, without having to predict a particular place where
3651 this may happen. (This is sometimes called a @dfn{data breakpoint}.)
3652 The expression may be as simple as the value of a single variable, or
3653 as complex as many variables combined by operators. Examples include:
3657 A reference to the value of a single variable.
3660 An address cast to an appropriate data type. For example,
3661 @samp{*(int *)0x12345678} will watch a 4-byte region at the specified
3662 address (assuming an @code{int} occupies 4 bytes).
3665 An arbitrarily complex expression, such as @samp{a*b + c/d}. The
3666 expression can use any operators valid in the program's native
3667 language (@pxref{Languages}).
3670 You can set a watchpoint on an expression even if the expression can
3671 not be evaluated yet. For instance, you can set a watchpoint on
3672 @samp{*global_ptr} before @samp{global_ptr} is initialized.
3673 @value{GDBN} will stop when your program sets @samp{global_ptr} and
3674 the expression produces a valid value. If the expression becomes
3675 valid in some other way than changing a variable (e.g.@: if the memory
3676 pointed to by @samp{*global_ptr} becomes readable as the result of a
3677 @code{malloc} call), @value{GDBN} may not stop until the next time
3678 the expression changes.
3680 @cindex software watchpoints
3681 @cindex hardware watchpoints
3682 Depending on your system, watchpoints may be implemented in software or
3683 hardware. @value{GDBN} does software watchpointing by single-stepping your
3684 program and testing the variable's value each time, which is hundreds of
3685 times slower than normal execution. (But this may still be worth it, to
3686 catch errors where you have no clue what part of your program is the
3689 On some systems, such as HP-UX, PowerPC, @sc{gnu}/Linux and most other
3690 x86-based targets, @value{GDBN} includes support for hardware
3691 watchpoints, which do not slow down the running of your program.
3695 @item watch @var{expr} @r{[}thread @var{threadnum}@r{]}
3696 Set a watchpoint for an expression. @value{GDBN} will break when the
3697 expression @var{expr} is written into by the program and its value
3698 changes. The simplest (and the most popular) use of this command is
3699 to watch the value of a single variable:
3702 (@value{GDBP}) watch foo
3705 If the command includes a @code{@r{[}thread @var{threadnum}@r{]}}
3706 clause, @value{GDBN} breaks only when the thread identified by
3707 @var{threadnum} changes the value of @var{expr}. If any other threads
3708 change the value of @var{expr}, @value{GDBN} will not break. Note
3709 that watchpoints restricted to a single thread in this way only work
3710 with Hardware Watchpoints.
3713 @item rwatch @var{expr} @r{[}thread @var{threadnum}@r{]}
3714 Set a watchpoint that will break when the value of @var{expr} is read
3718 @item awatch @var{expr} @r{[}thread @var{threadnum}@r{]}
3719 Set a watchpoint that will break when @var{expr} is either read from
3720 or written into by the program.
3722 @kindex info watchpoints @r{[}@var{n}@r{]}
3723 @item info watchpoints
3724 This command prints a list of watchpoints, using the same format as
3725 @code{info break} (@pxref{Set Breaks}).
3728 If you watch for a change in a numerically entered address you need to
3729 dereference it, as the address itself is just a constant number which will
3730 never change. @value{GDBN} refuses to create a watchpoint that watches
3731 a never-changing value:
3734 (@value{GDBP}) watch 0x600850
3735 Cannot watch constant value 0x600850.
3736 (@value{GDBP}) watch *(int *) 0x600850
3737 Watchpoint 1: *(int *) 6293584
3740 @value{GDBN} sets a @dfn{hardware watchpoint} if possible. Hardware
3741 watchpoints execute very quickly, and the debugger reports a change in
3742 value at the exact instruction where the change occurs. If @value{GDBN}
3743 cannot set a hardware watchpoint, it sets a software watchpoint, which
3744 executes more slowly and reports the change in value at the next
3745 @emph{statement}, not the instruction, after the change occurs.
3747 @cindex use only software watchpoints
3748 You can force @value{GDBN} to use only software watchpoints with the
3749 @kbd{set can-use-hw-watchpoints 0} command. With this variable set to
3750 zero, @value{GDBN} will never try to use hardware watchpoints, even if
3751 the underlying system supports them. (Note that hardware-assisted
3752 watchpoints that were set @emph{before} setting
3753 @code{can-use-hw-watchpoints} to zero will still use the hardware
3754 mechanism of watching expression values.)
3757 @item set can-use-hw-watchpoints
3758 @kindex set can-use-hw-watchpoints
3759 Set whether or not to use hardware watchpoints.
3761 @item show can-use-hw-watchpoints
3762 @kindex show can-use-hw-watchpoints
3763 Show the current mode of using hardware watchpoints.
3766 For remote targets, you can restrict the number of hardware
3767 watchpoints @value{GDBN} will use, see @ref{set remote
3768 hardware-breakpoint-limit}.
3770 When you issue the @code{watch} command, @value{GDBN} reports
3773 Hardware watchpoint @var{num}: @var{expr}
3777 if it was able to set a hardware watchpoint.
3779 Currently, the @code{awatch} and @code{rwatch} commands can only set
3780 hardware watchpoints, because accesses to data that don't change the
3781 value of the watched expression cannot be detected without examining
3782 every instruction as it is being executed, and @value{GDBN} does not do
3783 that currently. If @value{GDBN} finds that it is unable to set a
3784 hardware breakpoint with the @code{awatch} or @code{rwatch} command, it
3785 will print a message like this:
3788 Expression cannot be implemented with read/access watchpoint.
3791 Sometimes, @value{GDBN} cannot set a hardware watchpoint because the
3792 data type of the watched expression is wider than what a hardware
3793 watchpoint on the target machine can handle. For example, some systems
3794 can only watch regions that are up to 4 bytes wide; on such systems you
3795 cannot set hardware watchpoints for an expression that yields a
3796 double-precision floating-point number (which is typically 8 bytes
3797 wide). As a work-around, it might be possible to break the large region
3798 into a series of smaller ones and watch them with separate watchpoints.
3800 If you set too many hardware watchpoints, @value{GDBN} might be unable
3801 to insert all of them when you resume the execution of your program.
3802 Since the precise number of active watchpoints is unknown until such
3803 time as the program is about to be resumed, @value{GDBN} might not be
3804 able to warn you about this when you set the watchpoints, and the
3805 warning will be printed only when the program is resumed:
3808 Hardware watchpoint @var{num}: Could not insert watchpoint
3812 If this happens, delete or disable some of the watchpoints.
3814 Watching complex expressions that reference many variables can also
3815 exhaust the resources available for hardware-assisted watchpoints.
3816 That's because @value{GDBN} needs to watch every variable in the
3817 expression with separately allocated resources.
3819 If you call a function interactively using @code{print} or @code{call},
3820 any watchpoints you have set will be inactive until @value{GDBN} reaches another
3821 kind of breakpoint or the call completes.
3823 @value{GDBN} automatically deletes watchpoints that watch local
3824 (automatic) variables, or expressions that involve such variables, when
3825 they go out of scope, that is, when the execution leaves the block in
3826 which these variables were defined. In particular, when the program
3827 being debugged terminates, @emph{all} local variables go out of scope,
3828 and so only watchpoints that watch global variables remain set. If you
3829 rerun the program, you will need to set all such watchpoints again. One
3830 way of doing that would be to set a code breakpoint at the entry to the
3831 @code{main} function and when it breaks, set all the watchpoints.
3833 @cindex watchpoints and threads
3834 @cindex threads and watchpoints
3835 In multi-threaded programs, watchpoints will detect changes to the
3836 watched expression from every thread.
3839 @emph{Warning:} In multi-threaded programs, software watchpoints
3840 have only limited usefulness. If @value{GDBN} creates a software
3841 watchpoint, it can only watch the value of an expression @emph{in a
3842 single thread}. If you are confident that the expression can only
3843 change due to the current thread's activity (and if you are also
3844 confident that no other thread can become current), then you can use
3845 software watchpoints as usual. However, @value{GDBN} may not notice
3846 when a non-current thread's activity changes the expression. (Hardware
3847 watchpoints, in contrast, watch an expression in all threads.)
3850 @xref{set remote hardware-watchpoint-limit}.
3852 @node Set Catchpoints
3853 @subsection Setting Catchpoints
3854 @cindex catchpoints, setting
3855 @cindex exception handlers
3856 @cindex event handling
3858 You can use @dfn{catchpoints} to cause the debugger to stop for certain
3859 kinds of program events, such as C@t{++} exceptions or the loading of a
3860 shared library. Use the @code{catch} command to set a catchpoint.
3864 @item catch @var{event}
3865 Stop when @var{event} occurs. @var{event} can be any of the following:
3868 @cindex stop on C@t{++} exceptions
3869 The throwing of a C@t{++} exception.
3872 The catching of a C@t{++} exception.
3875 @cindex Ada exception catching
3876 @cindex catch Ada exceptions
3877 An Ada exception being raised. If an exception name is specified
3878 at the end of the command (eg @code{catch exception Program_Error}),
3879 the debugger will stop only when this specific exception is raised.
3880 Otherwise, the debugger stops execution when any Ada exception is raised.
3882 When inserting an exception catchpoint on a user-defined exception whose
3883 name is identical to one of the exceptions defined by the language, the
3884 fully qualified name must be used as the exception name. Otherwise,
3885 @value{GDBN} will assume that it should stop on the pre-defined exception
3886 rather than the user-defined one. For instance, assuming an exception
3887 called @code{Constraint_Error} is defined in package @code{Pck}, then
3888 the command to use to catch such exceptions is @kbd{catch exception
3889 Pck.Constraint_Error}.
3891 @item exception unhandled
3892 An exception that was raised but is not handled by the program.
3895 A failed Ada assertion.
3898 @cindex break on fork/exec
3899 A call to @code{exec}. This is currently only available for HP-UX
3903 @itemx syscall @r{[}@var{name} @r{|} @var{number}@r{]} @dots{}
3904 @cindex break on a system call.
3905 A call to or return from a system call, a.k.a.@: @dfn{syscall}. A
3906 syscall is a mechanism for application programs to request a service
3907 from the operating system (OS) or one of the OS system services.
3908 @value{GDBN} can catch some or all of the syscalls issued by the
3909 debuggee, and show the related information for each syscall. If no
3910 argument is specified, calls to and returns from all system calls
3913 @var{name} can be any system call name that is valid for the
3914 underlying OS. Just what syscalls are valid depends on the OS. On
3915 GNU and Unix systems, you can find the full list of valid syscall
3916 names on @file{/usr/include/asm/unistd.h}.
3918 @c For MS-Windows, the syscall names and the corresponding numbers
3919 @c can be found, e.g., on this URL:
3920 @c http://www.metasploit.com/users/opcode/syscalls.html
3921 @c but we don't support Windows syscalls yet.
3923 Normally, @value{GDBN} knows in advance which syscalls are valid for
3924 each OS, so you can use the @value{GDBN} command-line completion
3925 facilities (@pxref{Completion,, command completion}) to list the
3928 You may also specify the system call numerically. A syscall's
3929 number is the value passed to the OS's syscall dispatcher to
3930 identify the requested service. When you specify the syscall by its
3931 name, @value{GDBN} uses its database of syscalls to convert the name
3932 into the corresponding numeric code, but using the number directly
3933 may be useful if @value{GDBN}'s database does not have the complete
3934 list of syscalls on your system (e.g., because @value{GDBN} lags
3935 behind the OS upgrades).
3937 The example below illustrates how this command works if you don't provide
3941 (@value{GDBP}) catch syscall
3942 Catchpoint 1 (syscall)
3944 Starting program: /tmp/catch-syscall
3946 Catchpoint 1 (call to syscall 'close'), \
3947 0xffffe424 in __kernel_vsyscall ()
3951 Catchpoint 1 (returned from syscall 'close'), \
3952 0xffffe424 in __kernel_vsyscall ()
3956 Here is an example of catching a system call by name:
3959 (@value{GDBP}) catch syscall chroot
3960 Catchpoint 1 (syscall 'chroot' [61])
3962 Starting program: /tmp/catch-syscall
3964 Catchpoint 1 (call to syscall 'chroot'), \
3965 0xffffe424 in __kernel_vsyscall ()
3969 Catchpoint 1 (returned from syscall 'chroot'), \
3970 0xffffe424 in __kernel_vsyscall ()
3974 An example of specifying a system call numerically. In the case
3975 below, the syscall number has a corresponding entry in the XML
3976 file, so @value{GDBN} finds its name and prints it:
3979 (@value{GDBP}) catch syscall 252
3980 Catchpoint 1 (syscall(s) 'exit_group')
3982 Starting program: /tmp/catch-syscall
3984 Catchpoint 1 (call to syscall 'exit_group'), \
3985 0xffffe424 in __kernel_vsyscall ()
3989 Program exited normally.
3993 However, there can be situations when there is no corresponding name
3994 in XML file for that syscall number. In this case, @value{GDBN} prints
3995 a warning message saying that it was not able to find the syscall name,
3996 but the catchpoint will be set anyway. See the example below:
3999 (@value{GDBP}) catch syscall 764
4000 warning: The number '764' does not represent a known syscall.
4001 Catchpoint 2 (syscall 764)
4005 If you configure @value{GDBN} using the @samp{--without-expat} option,
4006 it will not be able to display syscall names. Also, if your
4007 architecture does not have an XML file describing its system calls,
4008 you will not be able to see the syscall names. It is important to
4009 notice that these two features are used for accessing the syscall
4010 name database. In either case, you will see a warning like this:
4013 (@value{GDBP}) catch syscall
4014 warning: Could not open "syscalls/i386-linux.xml"
4015 warning: Could not load the syscall XML file 'syscalls/i386-linux.xml'.
4016 GDB will not be able to display syscall names.
4017 Catchpoint 1 (syscall)
4021 Of course, the file name will change depending on your architecture and system.
4023 Still using the example above, you can also try to catch a syscall by its
4024 number. In this case, you would see something like:
4027 (@value{GDBP}) catch syscall 252
4028 Catchpoint 1 (syscall(s) 252)
4031 Again, in this case @value{GDBN} would not be able to display syscall's names.
4034 A call to @code{fork}. This is currently only available for HP-UX
4038 A call to @code{vfork}. This is currently only available for HP-UX
4043 @item tcatch @var{event}
4044 Set a catchpoint that is enabled only for one stop. The catchpoint is
4045 automatically deleted after the first time the event is caught.
4049 Use the @code{info break} command to list the current catchpoints.
4051 There are currently some limitations to C@t{++} exception handling
4052 (@code{catch throw} and @code{catch catch}) in @value{GDBN}:
4056 If you call a function interactively, @value{GDBN} normally returns
4057 control to you when the function has finished executing. If the call
4058 raises an exception, however, the call may bypass the mechanism that
4059 returns control to you and cause your program either to abort or to
4060 simply continue running until it hits a breakpoint, catches a signal
4061 that @value{GDBN} is listening for, or exits. This is the case even if
4062 you set a catchpoint for the exception; catchpoints on exceptions are
4063 disabled within interactive calls.
4066 You cannot raise an exception interactively.
4069 You cannot install an exception handler interactively.
4072 @cindex raise exceptions
4073 Sometimes @code{catch} is not the best way to debug exception handling:
4074 if you need to know exactly where an exception is raised, it is better to
4075 stop @emph{before} the exception handler is called, since that way you
4076 can see the stack before any unwinding takes place. If you set a
4077 breakpoint in an exception handler instead, it may not be easy to find
4078 out where the exception was raised.
4080 To stop just before an exception handler is called, you need some
4081 knowledge of the implementation. In the case of @sc{gnu} C@t{++}, exceptions are
4082 raised by calling a library function named @code{__raise_exception}
4083 which has the following ANSI C interface:
4086 /* @var{addr} is where the exception identifier is stored.
4087 @var{id} is the exception identifier. */
4088 void __raise_exception (void **addr, void *id);
4092 To make the debugger catch all exceptions before any stack
4093 unwinding takes place, set a breakpoint on @code{__raise_exception}
4094 (@pxref{Breakpoints, ,Breakpoints; Watchpoints; and Exceptions}).
4096 With a conditional breakpoint (@pxref{Conditions, ,Break Conditions})
4097 that depends on the value of @var{id}, you can stop your program when
4098 a specific exception is raised. You can use multiple conditional
4099 breakpoints to stop your program when any of a number of exceptions are
4104 @subsection Deleting Breakpoints
4106 @cindex clearing breakpoints, watchpoints, catchpoints
4107 @cindex deleting breakpoints, watchpoints, catchpoints
4108 It is often necessary to eliminate a breakpoint, watchpoint, or
4109 catchpoint once it has done its job and you no longer want your program
4110 to stop there. This is called @dfn{deleting} the breakpoint. A
4111 breakpoint that has been deleted no longer exists; it is forgotten.
4113 With the @code{clear} command you can delete breakpoints according to
4114 where they are in your program. With the @code{delete} command you can
4115 delete individual breakpoints, watchpoints, or catchpoints by specifying
4116 their breakpoint numbers.
4118 It is not necessary to delete a breakpoint to proceed past it. @value{GDBN}
4119 automatically ignores breakpoints on the first instruction to be executed
4120 when you continue execution without changing the execution address.
4125 Delete any breakpoints at the next instruction to be executed in the
4126 selected stack frame (@pxref{Selection, ,Selecting a Frame}). When
4127 the innermost frame is selected, this is a good way to delete a
4128 breakpoint where your program just stopped.
4130 @item clear @var{location}
4131 Delete any breakpoints set at the specified @var{location}.
4132 @xref{Specify Location}, for the various forms of @var{location}; the
4133 most useful ones are listed below:
4136 @item clear @var{function}
4137 @itemx clear @var{filename}:@var{function}
4138 Delete any breakpoints set at entry to the named @var{function}.
4140 @item clear @var{linenum}
4141 @itemx clear @var{filename}:@var{linenum}
4142 Delete any breakpoints set at or within the code of the specified
4143 @var{linenum} of the specified @var{filename}.
4146 @cindex delete breakpoints
4148 @kindex d @r{(@code{delete})}
4149 @item delete @r{[}breakpoints@r{]} @r{[}@var{range}@dots{}@r{]}
4150 Delete the breakpoints, watchpoints, or catchpoints of the breakpoint
4151 ranges specified as arguments. If no argument is specified, delete all
4152 breakpoints (@value{GDBN} asks confirmation, unless you have @code{set
4153 confirm off}). You can abbreviate this command as @code{d}.
4157 @subsection Disabling Breakpoints
4159 @cindex enable/disable a breakpoint
4160 Rather than deleting a breakpoint, watchpoint, or catchpoint, you might
4161 prefer to @dfn{disable} it. This makes the breakpoint inoperative as if
4162 it had been deleted, but remembers the information on the breakpoint so
4163 that you can @dfn{enable} it again later.
4165 You disable and enable breakpoints, watchpoints, and catchpoints with
4166 the @code{enable} and @code{disable} commands, optionally specifying
4167 one or more breakpoint numbers as arguments. Use @code{info break} to
4168 print a list of all breakpoints, watchpoints, and catchpoints if you
4169 do not know which numbers to use.
4171 Disabling and enabling a breakpoint that has multiple locations
4172 affects all of its locations.
4174 A breakpoint, watchpoint, or catchpoint can have any of four different
4175 states of enablement:
4179 Enabled. The breakpoint stops your program. A breakpoint set
4180 with the @code{break} command starts out in this state.
4182 Disabled. The breakpoint has no effect on your program.
4184 Enabled once. The breakpoint stops your program, but then becomes
4187 Enabled for deletion. The breakpoint stops your program, but
4188 immediately after it does so it is deleted permanently. A breakpoint
4189 set with the @code{tbreak} command starts out in this state.
4192 You can use the following commands to enable or disable breakpoints,
4193 watchpoints, and catchpoints:
4197 @kindex dis @r{(@code{disable})}
4198 @item disable @r{[}breakpoints@r{]} @r{[}@var{range}@dots{}@r{]}
4199 Disable the specified breakpoints---or all breakpoints, if none are
4200 listed. A disabled breakpoint has no effect but is not forgotten. All
4201 options such as ignore-counts, conditions and commands are remembered in
4202 case the breakpoint is enabled again later. You may abbreviate
4203 @code{disable} as @code{dis}.
4206 @item enable @r{[}breakpoints@r{]} @r{[}@var{range}@dots{}@r{]}
4207 Enable the specified breakpoints (or all defined breakpoints). They
4208 become effective once again in stopping your program.
4210 @item enable @r{[}breakpoints@r{]} once @var{range}@dots{}
4211 Enable the specified breakpoints temporarily. @value{GDBN} disables any
4212 of these breakpoints immediately after stopping your program.
4214 @item enable @r{[}breakpoints@r{]} delete @var{range}@dots{}
4215 Enable the specified breakpoints to work once, then die. @value{GDBN}
4216 deletes any of these breakpoints as soon as your program stops there.
4217 Breakpoints set by the @code{tbreak} command start out in this state.
4220 @c FIXME: I think the following ``Except for [...] @code{tbreak}'' is
4221 @c confusing: tbreak is also initially enabled.
4222 Except for a breakpoint set with @code{tbreak} (@pxref{Set Breaks,
4223 ,Setting Breakpoints}), breakpoints that you set are initially enabled;
4224 subsequently, they become disabled or enabled only when you use one of
4225 the commands above. (The command @code{until} can set and delete a
4226 breakpoint of its own, but it does not change the state of your other
4227 breakpoints; see @ref{Continuing and Stepping, ,Continuing and
4231 @subsection Break Conditions
4232 @cindex conditional breakpoints
4233 @cindex breakpoint conditions
4235 @c FIXME what is scope of break condition expr? Context where wanted?
4236 @c in particular for a watchpoint?
4237 The simplest sort of breakpoint breaks every time your program reaches a
4238 specified place. You can also specify a @dfn{condition} for a
4239 breakpoint. A condition is just a Boolean expression in your
4240 programming language (@pxref{Expressions, ,Expressions}). A breakpoint with
4241 a condition evaluates the expression each time your program reaches it,
4242 and your program stops only if the condition is @emph{true}.
4244 This is the converse of using assertions for program validation; in that
4245 situation, you want to stop when the assertion is violated---that is,
4246 when the condition is false. In C, if you want to test an assertion expressed
4247 by the condition @var{assert}, you should set the condition
4248 @samp{! @var{assert}} on the appropriate breakpoint.
4250 Conditions are also accepted for watchpoints; you may not need them,
4251 since a watchpoint is inspecting the value of an expression anyhow---but
4252 it might be simpler, say, to just set a watchpoint on a variable name,
4253 and specify a condition that tests whether the new value is an interesting
4256 Break conditions can have side effects, and may even call functions in
4257 your program. This can be useful, for example, to activate functions
4258 that log program progress, or to use your own print functions to
4259 format special data structures. The effects are completely predictable
4260 unless there is another enabled breakpoint at the same address. (In
4261 that case, @value{GDBN} might see the other breakpoint first and stop your
4262 program without checking the condition of this one.) Note that
4263 breakpoint commands are usually more convenient and flexible than break
4265 purpose of performing side effects when a breakpoint is reached
4266 (@pxref{Break Commands, ,Breakpoint Command Lists}).
4268 Break conditions can be specified when a breakpoint is set, by using
4269 @samp{if} in the arguments to the @code{break} command. @xref{Set
4270 Breaks, ,Setting Breakpoints}. They can also be changed at any time
4271 with the @code{condition} command.
4273 You can also use the @code{if} keyword with the @code{watch} command.
4274 The @code{catch} command does not recognize the @code{if} keyword;
4275 @code{condition} is the only way to impose a further condition on a
4280 @item condition @var{bnum} @var{expression}
4281 Specify @var{expression} as the break condition for breakpoint,
4282 watchpoint, or catchpoint number @var{bnum}. After you set a condition,
4283 breakpoint @var{bnum} stops your program only if the value of
4284 @var{expression} is true (nonzero, in C). When you use
4285 @code{condition}, @value{GDBN} checks @var{expression} immediately for
4286 syntactic correctness, and to determine whether symbols in it have
4287 referents in the context of your breakpoint. If @var{expression} uses
4288 symbols not referenced in the context of the breakpoint, @value{GDBN}
4289 prints an error message:
4292 No symbol "foo" in current context.
4297 not actually evaluate @var{expression} at the time the @code{condition}
4298 command (or a command that sets a breakpoint with a condition, like
4299 @code{break if @dots{}}) is given, however. @xref{Expressions, ,Expressions}.
4301 @item condition @var{bnum}
4302 Remove the condition from breakpoint number @var{bnum}. It becomes
4303 an ordinary unconditional breakpoint.
4306 @cindex ignore count (of breakpoint)
4307 A special case of a breakpoint condition is to stop only when the
4308 breakpoint has been reached a certain number of times. This is so
4309 useful that there is a special way to do it, using the @dfn{ignore
4310 count} of the breakpoint. Every breakpoint has an ignore count, which
4311 is an integer. Most of the time, the ignore count is zero, and
4312 therefore has no effect. But if your program reaches a breakpoint whose
4313 ignore count is positive, then instead of stopping, it just decrements
4314 the ignore count by one and continues. As a result, if the ignore count
4315 value is @var{n}, the breakpoint does not stop the next @var{n} times
4316 your program reaches it.
4320 @item ignore @var{bnum} @var{count}
4321 Set the ignore count of breakpoint number @var{bnum} to @var{count}.
4322 The next @var{count} times the breakpoint is reached, your program's
4323 execution does not stop; other than to decrement the ignore count, @value{GDBN}
4326 To make the breakpoint stop the next time it is reached, specify
4329 When you use @code{continue} to resume execution of your program from a
4330 breakpoint, you can specify an ignore count directly as an argument to
4331 @code{continue}, rather than using @code{ignore}. @xref{Continuing and
4332 Stepping,,Continuing and Stepping}.
4334 If a breakpoint has a positive ignore count and a condition, the
4335 condition is not checked. Once the ignore count reaches zero,
4336 @value{GDBN} resumes checking the condition.
4338 You could achieve the effect of the ignore count with a condition such
4339 as @w{@samp{$foo-- <= 0}} using a debugger convenience variable that
4340 is decremented each time. @xref{Convenience Vars, ,Convenience
4344 Ignore counts apply to breakpoints, watchpoints, and catchpoints.
4347 @node Break Commands
4348 @subsection Breakpoint Command Lists
4350 @cindex breakpoint commands
4351 You can give any breakpoint (or watchpoint or catchpoint) a series of
4352 commands to execute when your program stops due to that breakpoint. For
4353 example, you might want to print the values of certain expressions, or
4354 enable other breakpoints.
4358 @kindex end@r{ (breakpoint commands)}
4359 @item commands @r{[}@var{range}@dots{}@r{]}
4360 @itemx @dots{} @var{command-list} @dots{}
4362 Specify a list of commands for the given breakpoints. The commands
4363 themselves appear on the following lines. Type a line containing just
4364 @code{end} to terminate the commands.
4366 To remove all commands from a breakpoint, type @code{commands} and
4367 follow it immediately with @code{end}; that is, give no commands.
4369 With no argument, @code{commands} refers to the last breakpoint,
4370 watchpoint, or catchpoint set (not to the breakpoint most recently
4371 encountered). If the most recent breakpoints were set with a single
4372 command, then the @code{commands} will apply to all the breakpoints
4373 set by that command. This applies to breakpoints set by
4374 @code{rbreak}, and also applies when a single @code{break} command
4375 creates multiple breakpoints (@pxref{Ambiguous Expressions,,Ambiguous
4379 Pressing @key{RET} as a means of repeating the last @value{GDBN} command is
4380 disabled within a @var{command-list}.
4382 You can use breakpoint commands to start your program up again. Simply
4383 use the @code{continue} command, or @code{step}, or any other command
4384 that resumes execution.
4386 Any other commands in the command list, after a command that resumes
4387 execution, are ignored. This is because any time you resume execution
4388 (even with a simple @code{next} or @code{step}), you may encounter
4389 another breakpoint---which could have its own command list, leading to
4390 ambiguities about which list to execute.
4393 If the first command you specify in a command list is @code{silent}, the
4394 usual message about stopping at a breakpoint is not printed. This may
4395 be desirable for breakpoints that are to print a specific message and
4396 then continue. If none of the remaining commands print anything, you
4397 see no sign that the breakpoint was reached. @code{silent} is
4398 meaningful only at the beginning of a breakpoint command list.
4400 The commands @code{echo}, @code{output}, and @code{printf} allow you to
4401 print precisely controlled output, and are often useful in silent
4402 breakpoints. @xref{Output, ,Commands for Controlled Output}.
4404 For example, here is how you could use breakpoint commands to print the
4405 value of @code{x} at entry to @code{foo} whenever @code{x} is positive.
4411 printf "x is %d\n",x
4416 One application for breakpoint commands is to compensate for one bug so
4417 you can test for another. Put a breakpoint just after the erroneous line
4418 of code, give it a condition to detect the case in which something
4419 erroneous has been done, and give it commands to assign correct values
4420 to any variables that need them. End with the @code{continue} command
4421 so that your program does not stop, and start with the @code{silent}
4422 command so that no output is produced. Here is an example:
4433 @node Save Breakpoints
4434 @subsection How to save breakpoints to a file
4436 To save breakpoint definitions to a file use the @w{@code{save
4437 breakpoints}} command.
4440 @kindex save breakpoints
4441 @cindex save breakpoints to a file for future sessions
4442 @item save breakpoints [@var{filename}]
4443 This command saves all current breakpoint definitions together with
4444 their commands and ignore counts, into a file @file{@var{filename}}
4445 suitable for use in a later debugging session. This includes all
4446 types of breakpoints (breakpoints, watchpoints, catchpoints,
4447 tracepoints). To read the saved breakpoint definitions, use the
4448 @code{source} command (@pxref{Command Files}). Note that watchpoints
4449 with expressions involving local variables may fail to be recreated
4450 because it may not be possible to access the context where the
4451 watchpoint is valid anymore. Because the saved breakpoint definitions
4452 are simply a sequence of @value{GDBN} commands that recreate the
4453 breakpoints, you can edit the file in your favorite editing program,
4454 and remove the breakpoint definitions you're not interested in, or
4455 that can no longer be recreated.
4458 @c @ifclear BARETARGET
4459 @node Error in Breakpoints
4460 @subsection ``Cannot insert breakpoints''
4462 If you request too many active hardware-assisted breakpoints and
4463 watchpoints, you will see this error message:
4465 @c FIXME: the precise wording of this message may change; the relevant
4466 @c source change is not committed yet (Sep 3, 1999).
4468 Stopped; cannot insert breakpoints.
4469 You may have requested too many hardware breakpoints and watchpoints.
4473 This message is printed when you attempt to resume the program, since
4474 only then @value{GDBN} knows exactly how many hardware breakpoints and
4475 watchpoints it needs to insert.
4477 When this message is printed, you need to disable or remove some of the
4478 hardware-assisted breakpoints and watchpoints, and then continue.
4480 @node Breakpoint-related Warnings
4481 @subsection ``Breakpoint address adjusted...''
4482 @cindex breakpoint address adjusted
4484 Some processor architectures place constraints on the addresses at
4485 which breakpoints may be placed. For architectures thus constrained,
4486 @value{GDBN} will attempt to adjust the breakpoint's address to comply
4487 with the constraints dictated by the architecture.
4489 One example of such an architecture is the Fujitsu FR-V. The FR-V is
4490 a VLIW architecture in which a number of RISC-like instructions may be
4491 bundled together for parallel execution. The FR-V architecture
4492 constrains the location of a breakpoint instruction within such a
4493 bundle to the instruction with the lowest address. @value{GDBN}
4494 honors this constraint by adjusting a breakpoint's address to the
4495 first in the bundle.
4497 It is not uncommon for optimized code to have bundles which contain
4498 instructions from different source statements, thus it may happen that
4499 a breakpoint's address will be adjusted from one source statement to
4500 another. Since this adjustment may significantly alter @value{GDBN}'s
4501 breakpoint related behavior from what the user expects, a warning is
4502 printed when the breakpoint is first set and also when the breakpoint
4505 A warning like the one below is printed when setting a breakpoint
4506 that's been subject to address adjustment:
4509 warning: Breakpoint address adjusted from 0x00010414 to 0x00010410.
4512 Such warnings are printed both for user settable and @value{GDBN}'s
4513 internal breakpoints. If you see one of these warnings, you should
4514 verify that a breakpoint set at the adjusted address will have the
4515 desired affect. If not, the breakpoint in question may be removed and
4516 other breakpoints may be set which will have the desired behavior.
4517 E.g., it may be sufficient to place the breakpoint at a later
4518 instruction. A conditional breakpoint may also be useful in some
4519 cases to prevent the breakpoint from triggering too often.
4521 @value{GDBN} will also issue a warning when stopping at one of these
4522 adjusted breakpoints:
4525 warning: Breakpoint 1 address previously adjusted from 0x00010414
4529 When this warning is encountered, it may be too late to take remedial
4530 action except in cases where the breakpoint is hit earlier or more
4531 frequently than expected.
4533 @node Continuing and Stepping
4534 @section Continuing and Stepping
4538 @cindex resuming execution
4539 @dfn{Continuing} means resuming program execution until your program
4540 completes normally. In contrast, @dfn{stepping} means executing just
4541 one more ``step'' of your program, where ``step'' may mean either one
4542 line of source code, or one machine instruction (depending on what
4543 particular command you use). Either when continuing or when stepping,
4544 your program may stop even sooner, due to a breakpoint or a signal. (If
4545 it stops due to a signal, you may want to use @code{handle}, or use
4546 @samp{signal 0} to resume execution. @xref{Signals, ,Signals}.)
4550 @kindex c @r{(@code{continue})}
4551 @kindex fg @r{(resume foreground execution)}
4552 @item continue @r{[}@var{ignore-count}@r{]}
4553 @itemx c @r{[}@var{ignore-count}@r{]}
4554 @itemx fg @r{[}@var{ignore-count}@r{]}
4555 Resume program execution, at the address where your program last stopped;
4556 any breakpoints set at that address are bypassed. The optional argument
4557 @var{ignore-count} allows you to specify a further number of times to
4558 ignore a breakpoint at this location; its effect is like that of
4559 @code{ignore} (@pxref{Conditions, ,Break Conditions}).
4561 The argument @var{ignore-count} is meaningful only when your program
4562 stopped due to a breakpoint. At other times, the argument to
4563 @code{continue} is ignored.
4565 The synonyms @code{c} and @code{fg} (for @dfn{foreground}, as the
4566 debugged program is deemed to be the foreground program) are provided
4567 purely for convenience, and have exactly the same behavior as
4571 To resume execution at a different place, you can use @code{return}
4572 (@pxref{Returning, ,Returning from a Function}) to go back to the
4573 calling function; or @code{jump} (@pxref{Jumping, ,Continuing at a
4574 Different Address}) to go to an arbitrary location in your program.
4576 A typical technique for using stepping is to set a breakpoint
4577 (@pxref{Breakpoints, ,Breakpoints; Watchpoints; and Catchpoints}) at the
4578 beginning of the function or the section of your program where a problem
4579 is believed to lie, run your program until it stops at that breakpoint,
4580 and then step through the suspect area, examining the variables that are
4581 interesting, until you see the problem happen.
4585 @kindex s @r{(@code{step})}
4587 Continue running your program until control reaches a different source
4588 line, then stop it and return control to @value{GDBN}. This command is
4589 abbreviated @code{s}.
4592 @c "without debugging information" is imprecise; actually "without line
4593 @c numbers in the debugging information". (gcc -g1 has debugging info but
4594 @c not line numbers). But it seems complex to try to make that
4595 @c distinction here.
4596 @emph{Warning:} If you use the @code{step} command while control is
4597 within a function that was compiled without debugging information,
4598 execution proceeds until control reaches a function that does have
4599 debugging information. Likewise, it will not step into a function which
4600 is compiled without debugging information. To step through functions
4601 without debugging information, use the @code{stepi} command, described
4605 The @code{step} command only stops at the first instruction of a source
4606 line. This prevents the multiple stops that could otherwise occur in
4607 @code{switch} statements, @code{for} loops, etc. @code{step} continues
4608 to stop if a function that has debugging information is called within
4609 the line. In other words, @code{step} @emph{steps inside} any functions
4610 called within the line.
4612 Also, the @code{step} command only enters a function if there is line
4613 number information for the function. Otherwise it acts like the
4614 @code{next} command. This avoids problems when using @code{cc -gl}
4615 on MIPS machines. Previously, @code{step} entered subroutines if there
4616 was any debugging information about the routine.
4618 @item step @var{count}
4619 Continue running as in @code{step}, but do so @var{count} times. If a
4620 breakpoint is reached, or a signal not related to stepping occurs before
4621 @var{count} steps, stepping stops right away.
4624 @kindex n @r{(@code{next})}
4625 @item next @r{[}@var{count}@r{]}
4626 Continue to the next source line in the current (innermost) stack frame.
4627 This is similar to @code{step}, but function calls that appear within
4628 the line of code are executed without stopping. Execution stops when
4629 control reaches a different line of code at the original stack level
4630 that was executing when you gave the @code{next} command. This command
4631 is abbreviated @code{n}.
4633 An argument @var{count} is a repeat count, as for @code{step}.
4636 @c FIX ME!! Do we delete this, or is there a way it fits in with
4637 @c the following paragraph? --- Vctoria
4639 @c @code{next} within a function that lacks debugging information acts like
4640 @c @code{step}, but any function calls appearing within the code of the
4641 @c function are executed without stopping.
4643 The @code{next} command only stops at the first instruction of a
4644 source line. This prevents multiple stops that could otherwise occur in
4645 @code{switch} statements, @code{for} loops, etc.
4647 @kindex set step-mode
4649 @cindex functions without line info, and stepping
4650 @cindex stepping into functions with no line info
4651 @itemx set step-mode on
4652 The @code{set step-mode on} command causes the @code{step} command to
4653 stop at the first instruction of a function which contains no debug line
4654 information rather than stepping over it.
4656 This is useful in cases where you may be interested in inspecting the
4657 machine instructions of a function which has no symbolic info and do not
4658 want @value{GDBN} to automatically skip over this function.
4660 @item set step-mode off
4661 Causes the @code{step} command to step over any functions which contains no
4662 debug information. This is the default.
4664 @item show step-mode
4665 Show whether @value{GDBN} will stop in or step over functions without
4666 source line debug information.
4669 @kindex fin @r{(@code{finish})}
4671 Continue running until just after function in the selected stack frame
4672 returns. Print the returned value (if any). This command can be
4673 abbreviated as @code{fin}.
4675 Contrast this with the @code{return} command (@pxref{Returning,
4676 ,Returning from a Function}).
4679 @kindex u @r{(@code{until})}
4680 @cindex run until specified location
4683 Continue running until a source line past the current line, in the
4684 current stack frame, is reached. This command is used to avoid single
4685 stepping through a loop more than once. It is like the @code{next}
4686 command, except that when @code{until} encounters a jump, it
4687 automatically continues execution until the program counter is greater
4688 than the address of the jump.
4690 This means that when you reach the end of a loop after single stepping
4691 though it, @code{until} makes your program continue execution until it
4692 exits the loop. In contrast, a @code{next} command at the end of a loop
4693 simply steps back to the beginning of the loop, which forces you to step
4694 through the next iteration.
4696 @code{until} always stops your program if it attempts to exit the current
4699 @code{until} may produce somewhat counterintuitive results if the order
4700 of machine code does not match the order of the source lines. For
4701 example, in the following excerpt from a debugging session, the @code{f}
4702 (@code{frame}) command shows that execution is stopped at line
4703 @code{206}; yet when we use @code{until}, we get to line @code{195}:
4707 #0 main (argc=4, argv=0xf7fffae8) at m4.c:206
4709 (@value{GDBP}) until
4710 195 for ( ; argc > 0; NEXTARG) @{
4713 This happened because, for execution efficiency, the compiler had
4714 generated code for the loop closure test at the end, rather than the
4715 start, of the loop---even though the test in a C @code{for}-loop is
4716 written before the body of the loop. The @code{until} command appeared
4717 to step back to the beginning of the loop when it advanced to this
4718 expression; however, it has not really gone to an earlier
4719 statement---not in terms of the actual machine code.
4721 @code{until} with no argument works by means of single
4722 instruction stepping, and hence is slower than @code{until} with an
4725 @item until @var{location}
4726 @itemx u @var{location}
4727 Continue running your program until either the specified location is
4728 reached, or the current stack frame returns. @var{location} is any of
4729 the forms described in @ref{Specify Location}.
4730 This form of the command uses temporary breakpoints, and
4731 hence is quicker than @code{until} without an argument. The specified
4732 location is actually reached only if it is in the current frame. This
4733 implies that @code{until} can be used to skip over recursive function
4734 invocations. For instance in the code below, if the current location is
4735 line @code{96}, issuing @code{until 99} will execute the program up to
4736 line @code{99} in the same invocation of factorial, i.e., after the inner
4737 invocations have returned.
4740 94 int factorial (int value)
4742 96 if (value > 1) @{
4743 97 value *= factorial (value - 1);
4750 @kindex advance @var{location}
4751 @itemx advance @var{location}
4752 Continue running the program up to the given @var{location}. An argument is
4753 required, which should be of one of the forms described in
4754 @ref{Specify Location}.
4755 Execution will also stop upon exit from the current stack
4756 frame. This command is similar to @code{until}, but @code{advance} will
4757 not skip over recursive function calls, and the target location doesn't
4758 have to be in the same frame as the current one.
4762 @kindex si @r{(@code{stepi})}
4764 @itemx stepi @var{arg}
4766 Execute one machine instruction, then stop and return to the debugger.
4768 It is often useful to do @samp{display/i $pc} when stepping by machine
4769 instructions. This makes @value{GDBN} automatically display the next
4770 instruction to be executed, each time your program stops. @xref{Auto
4771 Display,, Automatic Display}.
4773 An argument is a repeat count, as in @code{step}.
4777 @kindex ni @r{(@code{nexti})}
4779 @itemx nexti @var{arg}
4781 Execute one machine instruction, but if it is a function call,
4782 proceed until the function returns.
4784 An argument is a repeat count, as in @code{next}.
4791 A signal is an asynchronous event that can happen in a program. The
4792 operating system defines the possible kinds of signals, and gives each
4793 kind a name and a number. For example, in Unix @code{SIGINT} is the
4794 signal a program gets when you type an interrupt character (often @kbd{Ctrl-c});
4795 @code{SIGSEGV} is the signal a program gets from referencing a place in
4796 memory far away from all the areas in use; @code{SIGALRM} occurs when
4797 the alarm clock timer goes off (which happens only if your program has
4798 requested an alarm).
4800 @cindex fatal signals
4801 Some signals, including @code{SIGALRM}, are a normal part of the
4802 functioning of your program. Others, such as @code{SIGSEGV}, indicate
4803 errors; these signals are @dfn{fatal} (they kill your program immediately) if the
4804 program has not specified in advance some other way to handle the signal.
4805 @code{SIGINT} does not indicate an error in your program, but it is normally
4806 fatal so it can carry out the purpose of the interrupt: to kill the program.
4808 @value{GDBN} has the ability to detect any occurrence of a signal in your
4809 program. You can tell @value{GDBN} in advance what to do for each kind of
4812 @cindex handling signals
4813 Normally, @value{GDBN} is set up to let the non-erroneous signals like
4814 @code{SIGALRM} be silently passed to your program
4815 (so as not to interfere with their role in the program's functioning)
4816 but to stop your program immediately whenever an error signal happens.
4817 You can change these settings with the @code{handle} command.
4820 @kindex info signals
4824 Print a table of all the kinds of signals and how @value{GDBN} has been told to
4825 handle each one. You can use this to see the signal numbers of all
4826 the defined types of signals.
4828 @item info signals @var{sig}
4829 Similar, but print information only about the specified signal number.
4831 @code{info handle} is an alias for @code{info signals}.
4834 @item handle @var{signal} @r{[}@var{keywords}@dots{}@r{]}
4835 Change the way @value{GDBN} handles signal @var{signal}. @var{signal}
4836 can be the number of a signal or its name (with or without the
4837 @samp{SIG} at the beginning); a list of signal numbers of the form
4838 @samp{@var{low}-@var{high}}; or the word @samp{all}, meaning all the
4839 known signals. Optional arguments @var{keywords}, described below,
4840 say what change to make.
4844 The keywords allowed by the @code{handle} command can be abbreviated.
4845 Their full names are:
4849 @value{GDBN} should not stop your program when this signal happens. It may
4850 still print a message telling you that the signal has come in.
4853 @value{GDBN} should stop your program when this signal happens. This implies
4854 the @code{print} keyword as well.
4857 @value{GDBN} should print a message when this signal happens.
4860 @value{GDBN} should not mention the occurrence of the signal at all. This
4861 implies the @code{nostop} keyword as well.
4865 @value{GDBN} should allow your program to see this signal; your program
4866 can handle the signal, or else it may terminate if the signal is fatal
4867 and not handled. @code{pass} and @code{noignore} are synonyms.
4871 @value{GDBN} should not allow your program to see this signal.
4872 @code{nopass} and @code{ignore} are synonyms.
4876 When a signal stops your program, the signal is not visible to the
4878 continue. Your program sees the signal then, if @code{pass} is in
4879 effect for the signal in question @emph{at that time}. In other words,
4880 after @value{GDBN} reports a signal, you can use the @code{handle}
4881 command with @code{pass} or @code{nopass} to control whether your
4882 program sees that signal when you continue.
4884 The default is set to @code{nostop}, @code{noprint}, @code{pass} for
4885 non-erroneous signals such as @code{SIGALRM}, @code{SIGWINCH} and
4886 @code{SIGCHLD}, and to @code{stop}, @code{print}, @code{pass} for the
4889 You can also use the @code{signal} command to prevent your program from
4890 seeing a signal, or cause it to see a signal it normally would not see,
4891 or to give it any signal at any time. For example, if your program stopped
4892 due to some sort of memory reference error, you might store correct
4893 values into the erroneous variables and continue, hoping to see more
4894 execution; but your program would probably terminate immediately as
4895 a result of the fatal signal once it saw the signal. To prevent this,
4896 you can continue with @samp{signal 0}. @xref{Signaling, ,Giving your
4899 @cindex extra signal information
4900 @anchor{extra signal information}
4902 On some targets, @value{GDBN} can inspect extra signal information
4903 associated with the intercepted signal, before it is actually
4904 delivered to the program being debugged. This information is exported
4905 by the convenience variable @code{$_siginfo}, and consists of data
4906 that is passed by the kernel to the signal handler at the time of the
4907 receipt of a signal. The data type of the information itself is
4908 target dependent. You can see the data type using the @code{ptype
4909 $_siginfo} command. On Unix systems, it typically corresponds to the
4910 standard @code{siginfo_t} type, as defined in the @file{signal.h}
4913 Here's an example, on a @sc{gnu}/Linux system, printing the stray
4914 referenced address that raised a segmentation fault.
4918 (@value{GDBP}) continue
4919 Program received signal SIGSEGV, Segmentation fault.
4920 0x0000000000400766 in main ()
4922 (@value{GDBP}) ptype $_siginfo
4929 struct @{...@} _kill;
4930 struct @{...@} _timer;
4932 struct @{...@} _sigchld;
4933 struct @{...@} _sigfault;
4934 struct @{...@} _sigpoll;
4937 (@value{GDBP}) ptype $_siginfo._sifields._sigfault
4941 (@value{GDBP}) p $_siginfo._sifields._sigfault.si_addr
4942 $1 = (void *) 0x7ffff7ff7000
4946 Depending on target support, @code{$_siginfo} may also be writable.
4949 @section Stopping and Starting Multi-thread Programs
4951 @cindex stopped threads
4952 @cindex threads, stopped
4954 @cindex continuing threads
4955 @cindex threads, continuing
4957 @value{GDBN} supports debugging programs with multiple threads
4958 (@pxref{Threads,, Debugging Programs with Multiple Threads}). There
4959 are two modes of controlling execution of your program within the
4960 debugger. In the default mode, referred to as @dfn{all-stop mode},
4961 when any thread in your program stops (for example, at a breakpoint
4962 or while being stepped), all other threads in the program are also stopped by
4963 @value{GDBN}. On some targets, @value{GDBN} also supports
4964 @dfn{non-stop mode}, in which other threads can continue to run freely while
4965 you examine the stopped thread in the debugger.
4968 * All-Stop Mode:: All threads stop when GDB takes control
4969 * Non-Stop Mode:: Other threads continue to execute
4970 * Background Execution:: Running your program asynchronously
4971 * Thread-Specific Breakpoints:: Controlling breakpoints
4972 * Interrupted System Calls:: GDB may interfere with system calls
4973 * Observer Mode:: GDB does not alter program behavior
4977 @subsection All-Stop Mode
4979 @cindex all-stop mode
4981 In all-stop mode, whenever your program stops under @value{GDBN} for any reason,
4982 @emph{all} threads of execution stop, not just the current thread. This
4983 allows you to examine the overall state of the program, including
4984 switching between threads, without worrying that things may change
4987 Conversely, whenever you restart the program, @emph{all} threads start
4988 executing. @emph{This is true even when single-stepping} with commands
4989 like @code{step} or @code{next}.
4991 In particular, @value{GDBN} cannot single-step all threads in lockstep.
4992 Since thread scheduling is up to your debugging target's operating
4993 system (not controlled by @value{GDBN}), other threads may
4994 execute more than one statement while the current thread completes a
4995 single step. Moreover, in general other threads stop in the middle of a
4996 statement, rather than at a clean statement boundary, when the program
4999 You might even find your program stopped in another thread after
5000 continuing or even single-stepping. This happens whenever some other
5001 thread runs into a breakpoint, a signal, or an exception before the
5002 first thread completes whatever you requested.
5004 @cindex automatic thread selection
5005 @cindex switching threads automatically
5006 @cindex threads, automatic switching
5007 Whenever @value{GDBN} stops your program, due to a breakpoint or a
5008 signal, it automatically selects the thread where that breakpoint or
5009 signal happened. @value{GDBN} alerts you to the context switch with a
5010 message such as @samp{[Switching to Thread @var{n}]} to identify the
5013 On some OSes, you can modify @value{GDBN}'s default behavior by
5014 locking the OS scheduler to allow only a single thread to run.
5017 @item set scheduler-locking @var{mode}
5018 @cindex scheduler locking mode
5019 @cindex lock scheduler
5020 Set the scheduler locking mode. If it is @code{off}, then there is no
5021 locking and any thread may run at any time. If @code{on}, then only the
5022 current thread may run when the inferior is resumed. The @code{step}
5023 mode optimizes for single-stepping; it prevents other threads
5024 from preempting the current thread while you are stepping, so that
5025 the focus of debugging does not change unexpectedly.
5026 Other threads only rarely (or never) get a chance to run
5027 when you step. They are more likely to run when you @samp{next} over a
5028 function call, and they are completely free to run when you use commands
5029 like @samp{continue}, @samp{until}, or @samp{finish}. However, unless another
5030 thread hits a breakpoint during its timeslice, @value{GDBN} does not change
5031 the current thread away from the thread that you are debugging.
5033 @item show scheduler-locking
5034 Display the current scheduler locking mode.
5037 @cindex resume threads of multiple processes simultaneously
5038 By default, when you issue one of the execution commands such as
5039 @code{continue}, @code{next} or @code{step}, @value{GDBN} allows only
5040 threads of the current inferior to run. For example, if @value{GDBN}
5041 is attached to two inferiors, each with two threads, the
5042 @code{continue} command resumes only the two threads of the current
5043 inferior. This is useful, for example, when you debug a program that
5044 forks and you want to hold the parent stopped (so that, for instance,
5045 it doesn't run to exit), while you debug the child. In other
5046 situations, you may not be interested in inspecting the current state
5047 of any of the processes @value{GDBN} is attached to, and you may want
5048 to resume them all until some breakpoint is hit. In the latter case,
5049 you can instruct @value{GDBN} to allow all threads of all the
5050 inferiors to run with the @w{@code{set schedule-multiple}} command.
5053 @kindex set schedule-multiple
5054 @item set schedule-multiple
5055 Set the mode for allowing threads of multiple processes to be resumed
5056 when an execution command is issued. When @code{on}, all threads of
5057 all processes are allowed to run. When @code{off}, only the threads
5058 of the current process are resumed. The default is @code{off}. The
5059 @code{scheduler-locking} mode takes precedence when set to @code{on},
5060 or while you are stepping and set to @code{step}.
5062 @item show schedule-multiple
5063 Display the current mode for resuming the execution of threads of
5068 @subsection Non-Stop Mode
5070 @cindex non-stop mode
5072 @c This section is really only a place-holder, and needs to be expanded
5073 @c with more details.
5075 For some multi-threaded targets, @value{GDBN} supports an optional
5076 mode of operation in which you can examine stopped program threads in
5077 the debugger while other threads continue to execute freely. This
5078 minimizes intrusion when debugging live systems, such as programs
5079 where some threads have real-time constraints or must continue to
5080 respond to external events. This is referred to as @dfn{non-stop} mode.
5082 In non-stop mode, when a thread stops to report a debugging event,
5083 @emph{only} that thread is stopped; @value{GDBN} does not stop other
5084 threads as well, in contrast to the all-stop mode behavior. Additionally,
5085 execution commands such as @code{continue} and @code{step} apply by default
5086 only to the current thread in non-stop mode, rather than all threads as
5087 in all-stop mode. This allows you to control threads explicitly in
5088 ways that are not possible in all-stop mode --- for example, stepping
5089 one thread while allowing others to run freely, stepping
5090 one thread while holding all others stopped, or stepping several threads
5091 independently and simultaneously.
5093 To enter non-stop mode, use this sequence of commands before you run
5094 or attach to your program:
5097 # Enable the async interface.
5100 # If using the CLI, pagination breaks non-stop.
5103 # Finally, turn it on!
5107 You can use these commands to manipulate the non-stop mode setting:
5110 @kindex set non-stop
5111 @item set non-stop on
5112 Enable selection of non-stop mode.
5113 @item set non-stop off
5114 Disable selection of non-stop mode.
5115 @kindex show non-stop
5117 Show the current non-stop enablement setting.
5120 Note these commands only reflect whether non-stop mode is enabled,
5121 not whether the currently-executing program is being run in non-stop mode.
5122 In particular, the @code{set non-stop} preference is only consulted when
5123 @value{GDBN} starts or connects to the target program, and it is generally
5124 not possible to switch modes once debugging has started. Furthermore,
5125 since not all targets support non-stop mode, even when you have enabled
5126 non-stop mode, @value{GDBN} may still fall back to all-stop operation by
5129 In non-stop mode, all execution commands apply only to the current thread
5130 by default. That is, @code{continue} only continues one thread.
5131 To continue all threads, issue @code{continue -a} or @code{c -a}.
5133 You can use @value{GDBN}'s background execution commands
5134 (@pxref{Background Execution}) to run some threads in the background
5135 while you continue to examine or step others from @value{GDBN}.
5136 The MI execution commands (@pxref{GDB/MI Program Execution}) are
5137 always executed asynchronously in non-stop mode.
5139 Suspending execution is done with the @code{interrupt} command when
5140 running in the background, or @kbd{Ctrl-c} during foreground execution.
5141 In all-stop mode, this stops the whole process;
5142 but in non-stop mode the interrupt applies only to the current thread.
5143 To stop the whole program, use @code{interrupt -a}.
5145 Other execution commands do not currently support the @code{-a} option.
5147 In non-stop mode, when a thread stops, @value{GDBN} doesn't automatically make
5148 that thread current, as it does in all-stop mode. This is because the
5149 thread stop notifications are asynchronous with respect to @value{GDBN}'s
5150 command interpreter, and it would be confusing if @value{GDBN} unexpectedly
5151 changed to a different thread just as you entered a command to operate on the
5152 previously current thread.
5154 @node Background Execution
5155 @subsection Background Execution
5157 @cindex foreground execution
5158 @cindex background execution
5159 @cindex asynchronous execution
5160 @cindex execution, foreground, background and asynchronous
5162 @value{GDBN}'s execution commands have two variants: the normal
5163 foreground (synchronous) behavior, and a background
5164 (asynchronous) behavior. In foreground execution, @value{GDBN} waits for
5165 the program to report that some thread has stopped before prompting for
5166 another command. In background execution, @value{GDBN} immediately gives
5167 a command prompt so that you can issue other commands while your program runs.
5169 You need to explicitly enable asynchronous mode before you can use
5170 background execution commands. You can use these commands to
5171 manipulate the asynchronous mode setting:
5174 @kindex set target-async
5175 @item set target-async on
5176 Enable asynchronous mode.
5177 @item set target-async off
5178 Disable asynchronous mode.
5179 @kindex show target-async
5180 @item show target-async
5181 Show the current target-async setting.
5184 If the target doesn't support async mode, @value{GDBN} issues an error
5185 message if you attempt to use the background execution commands.
5187 To specify background execution, add a @code{&} to the command. For example,
5188 the background form of the @code{continue} command is @code{continue&}, or
5189 just @code{c&}. The execution commands that accept background execution
5195 @xref{Starting, , Starting your Program}.
5199 @xref{Attach, , Debugging an Already-running Process}.
5203 @xref{Continuing and Stepping, step}.
5207 @xref{Continuing and Stepping, stepi}.
5211 @xref{Continuing and Stepping, next}.
5215 @xref{Continuing and Stepping, nexti}.
5219 @xref{Continuing and Stepping, continue}.
5223 @xref{Continuing and Stepping, finish}.
5227 @xref{Continuing and Stepping, until}.
5231 Background execution is especially useful in conjunction with non-stop
5232 mode for debugging programs with multiple threads; see @ref{Non-Stop Mode}.
5233 However, you can also use these commands in the normal all-stop mode with
5234 the restriction that you cannot issue another execution command until the
5235 previous one finishes. Examples of commands that are valid in all-stop
5236 mode while the program is running include @code{help} and @code{info break}.
5238 You can interrupt your program while it is running in the background by
5239 using the @code{interrupt} command.
5246 Suspend execution of the running program. In all-stop mode,
5247 @code{interrupt} stops the whole process, but in non-stop mode, it stops
5248 only the current thread. To stop the whole program in non-stop mode,
5249 use @code{interrupt -a}.
5252 @node Thread-Specific Breakpoints
5253 @subsection Thread-Specific Breakpoints
5255 When your program has multiple threads (@pxref{Threads,, Debugging
5256 Programs with Multiple Threads}), you can choose whether to set
5257 breakpoints on all threads, or on a particular thread.
5260 @cindex breakpoints and threads
5261 @cindex thread breakpoints
5262 @kindex break @dots{} thread @var{threadno}
5263 @item break @var{linespec} thread @var{threadno}
5264 @itemx break @var{linespec} thread @var{threadno} if @dots{}
5265 @var{linespec} specifies source lines; there are several ways of
5266 writing them (@pxref{Specify Location}), but the effect is always to
5267 specify some source line.
5269 Use the qualifier @samp{thread @var{threadno}} with a breakpoint command
5270 to specify that you only want @value{GDBN} to stop the program when a
5271 particular thread reaches this breakpoint. @var{threadno} is one of the
5272 numeric thread identifiers assigned by @value{GDBN}, shown in the first
5273 column of the @samp{info threads} display.
5275 If you do not specify @samp{thread @var{threadno}} when you set a
5276 breakpoint, the breakpoint applies to @emph{all} threads of your
5279 You can use the @code{thread} qualifier on conditional breakpoints as
5280 well; in this case, place @samp{thread @var{threadno}} before or
5281 after the breakpoint condition, like this:
5284 (@value{GDBP}) break frik.c:13 thread 28 if bartab > lim
5289 @node Interrupted System Calls
5290 @subsection Interrupted System Calls
5292 @cindex thread breakpoints and system calls
5293 @cindex system calls and thread breakpoints
5294 @cindex premature return from system calls
5295 There is an unfortunate side effect when using @value{GDBN} to debug
5296 multi-threaded programs. If one thread stops for a
5297 breakpoint, or for some other reason, and another thread is blocked in a
5298 system call, then the system call may return prematurely. This is a
5299 consequence of the interaction between multiple threads and the signals
5300 that @value{GDBN} uses to implement breakpoints and other events that
5303 To handle this problem, your program should check the return value of
5304 each system call and react appropriately. This is good programming
5307 For example, do not write code like this:
5313 The call to @code{sleep} will return early if a different thread stops
5314 at a breakpoint or for some other reason.
5316 Instead, write this:
5321 unslept = sleep (unslept);
5324 A system call is allowed to return early, so the system is still
5325 conforming to its specification. But @value{GDBN} does cause your
5326 multi-threaded program to behave differently than it would without
5329 Also, @value{GDBN} uses internal breakpoints in the thread library to
5330 monitor certain events such as thread creation and thread destruction.
5331 When such an event happens, a system call in another thread may return
5332 prematurely, even though your program does not appear to stop.
5335 @subsection Observer Mode
5337 If you want to build on non-stop mode and observe program behavior
5338 without any chance of disruption by @value{GDBN}, you can set
5339 variables to disable all of the debugger's attempts to modify state,
5340 whether by writing memory, inserting breakpoints, etc. These operate
5341 at a low level, intercepting operations from all commands.
5343 When all of these are set to @code{off}, then @value{GDBN} is said to
5344 be @dfn{observer mode}. As a convenience, the variable
5345 @code{observer} can be set to disable these, plus enable non-stop
5348 Note that @value{GDBN} will not prevent you from making nonsensical
5349 combinations of these settings. For instance, if you have enabled
5350 @code{may-insert-breakpoints} but disabled @code{may-write-memory},
5351 then breakpoints that work by writing trap instructions into the code
5352 stream will still not be able to be placed.
5357 @item set observer on
5358 @itemx set observer off
5359 When set to @code{on}, this disables all the permission variables
5360 below (except for @code{insert-fast-tracepoints}), plus enables
5361 non-stop debugging. Setting this to @code{off} switches back to
5362 normal debugging, though remaining in non-stop mode.
5365 Show whether observer mode is on or off.
5367 @kindex may-write-registers
5368 @item set may-write-registers on
5369 @itemx set may-write-registers off
5370 This controls whether @value{GDBN} will attempt to alter the values of
5371 registers, such as with assignment expressions in @code{print}, or the
5372 @code{jump} command. It defaults to @code{on}.
5374 @item show may-write-registers
5375 Show the current permission to write registers.
5377 @kindex may-write-memory
5378 @item set may-write-memory on
5379 @itemx set may-write-memory off
5380 This controls whether @value{GDBN} will attempt to alter the contents
5381 of memory, such as with assignment expressions in @code{print}. It
5382 defaults to @code{on}.
5384 @item show may-write-memory
5385 Show the current permission to write memory.
5387 @kindex may-insert-breakpoints
5388 @item set may-insert-breakpoints on
5389 @itemx set may-insert-breakpoints off
5390 This controls whether @value{GDBN} will attempt to insert breakpoints.
5391 This affects all breakpoints, including internal breakpoints defined
5392 by @value{GDBN}. It defaults to @code{on}.
5394 @item show may-insert-breakpoints
5395 Show the current permission to insert breakpoints.
5397 @kindex may-insert-tracepoints
5398 @item set may-insert-tracepoints on
5399 @itemx set may-insert-tracepoints off
5400 This controls whether @value{GDBN} will attempt to insert (regular)
5401 tracepoints at the beginning of a tracing experiment. It affects only
5402 non-fast tracepoints, fast tracepoints being under the control of
5403 @code{may-insert-fast-tracepoints}. It defaults to @code{on}.
5405 @item show may-insert-tracepoints
5406 Show the current permission to insert tracepoints.
5408 @kindex may-insert-fast-tracepoints
5409 @item set may-insert-fast-tracepoints on
5410 @itemx set may-insert-fast-tracepoints off
5411 This controls whether @value{GDBN} will attempt to insert fast
5412 tracepoints at the beginning of a tracing experiment. It affects only
5413 fast tracepoints, regular (non-fast) tracepoints being under the
5414 control of @code{may-insert-tracepoints}. It defaults to @code{on}.
5416 @item show may-insert-fast-tracepoints
5417 Show the current permission to insert fast tracepoints.
5419 @kindex may-interrupt
5420 @item set may-interrupt on
5421 @itemx set may-interrupt off
5422 This controls whether @value{GDBN} will attempt to interrupt or stop
5423 program execution. When this variable is @code{off}, the
5424 @code{interrupt} command will have no effect, nor will
5425 @kbd{Ctrl-c}. It defaults to @code{on}.
5427 @item show may-interrupt
5428 Show the current permission to interrupt or stop the program.
5432 @node Reverse Execution
5433 @chapter Running programs backward
5434 @cindex reverse execution
5435 @cindex running programs backward
5437 When you are debugging a program, it is not unusual to realize that
5438 you have gone too far, and some event of interest has already happened.
5439 If the target environment supports it, @value{GDBN} can allow you to
5440 ``rewind'' the program by running it backward.
5442 A target environment that supports reverse execution should be able
5443 to ``undo'' the changes in machine state that have taken place as the
5444 program was executing normally. Variables, registers etc.@: should
5445 revert to their previous values. Obviously this requires a great
5446 deal of sophistication on the part of the target environment; not
5447 all target environments can support reverse execution.
5449 When a program is executed in reverse, the instructions that
5450 have most recently been executed are ``un-executed'', in reverse
5451 order. The program counter runs backward, following the previous
5452 thread of execution in reverse. As each instruction is ``un-executed'',
5453 the values of memory and/or registers that were changed by that
5454 instruction are reverted to their previous states. After executing
5455 a piece of source code in reverse, all side effects of that code
5456 should be ``undone'', and all variables should be returned to their
5457 prior values@footnote{
5458 Note that some side effects are easier to undo than others. For instance,
5459 memory and registers are relatively easy, but device I/O is hard. Some
5460 targets may be able undo things like device I/O, and some may not.
5462 The contract between @value{GDBN} and the reverse executing target
5463 requires only that the target do something reasonable when
5464 @value{GDBN} tells it to execute backwards, and then report the
5465 results back to @value{GDBN}. Whatever the target reports back to
5466 @value{GDBN}, @value{GDBN} will report back to the user. @value{GDBN}
5467 assumes that the memory and registers that the target reports are in a
5468 consistant state, but @value{GDBN} accepts whatever it is given.
5471 If you are debugging in a target environment that supports
5472 reverse execution, @value{GDBN} provides the following commands.
5475 @kindex reverse-continue
5476 @kindex rc @r{(@code{reverse-continue})}
5477 @item reverse-continue @r{[}@var{ignore-count}@r{]}
5478 @itemx rc @r{[}@var{ignore-count}@r{]}
5479 Beginning at the point where your program last stopped, start executing
5480 in reverse. Reverse execution will stop for breakpoints and synchronous
5481 exceptions (signals), just like normal execution. Behavior of
5482 asynchronous signals depends on the target environment.
5484 @kindex reverse-step
5485 @kindex rs @r{(@code{step})}
5486 @item reverse-step @r{[}@var{count}@r{]}
5487 Run the program backward until control reaches the start of a
5488 different source line; then stop it, and return control to @value{GDBN}.
5490 Like the @code{step} command, @code{reverse-step} will only stop
5491 at the beginning of a source line. It ``un-executes'' the previously
5492 executed source line. If the previous source line included calls to
5493 debuggable functions, @code{reverse-step} will step (backward) into
5494 the called function, stopping at the beginning of the @emph{last}
5495 statement in the called function (typically a return statement).
5497 Also, as with the @code{step} command, if non-debuggable functions are
5498 called, @code{reverse-step} will run thru them backward without stopping.
5500 @kindex reverse-stepi
5501 @kindex rsi @r{(@code{reverse-stepi})}
5502 @item reverse-stepi @r{[}@var{count}@r{]}
5503 Reverse-execute one machine instruction. Note that the instruction
5504 to be reverse-executed is @emph{not} the one pointed to by the program
5505 counter, but the instruction executed prior to that one. For instance,
5506 if the last instruction was a jump, @code{reverse-stepi} will take you
5507 back from the destination of the jump to the jump instruction itself.
5509 @kindex reverse-next
5510 @kindex rn @r{(@code{reverse-next})}
5511 @item reverse-next @r{[}@var{count}@r{]}
5512 Run backward to the beginning of the previous line executed in
5513 the current (innermost) stack frame. If the line contains function
5514 calls, they will be ``un-executed'' without stopping. Starting from
5515 the first line of a function, @code{reverse-next} will take you back
5516 to the caller of that function, @emph{before} the function was called,
5517 just as the normal @code{next} command would take you from the last
5518 line of a function back to its return to its caller
5519 @footnote{Unless the code is too heavily optimized.}.
5521 @kindex reverse-nexti
5522 @kindex rni @r{(@code{reverse-nexti})}
5523 @item reverse-nexti @r{[}@var{count}@r{]}
5524 Like @code{nexti}, @code{reverse-nexti} executes a single instruction
5525 in reverse, except that called functions are ``un-executed'' atomically.
5526 That is, if the previously executed instruction was a return from
5527 another function, @code{reverse-nexti} will continue to execute
5528 in reverse until the call to that function (from the current stack
5531 @kindex reverse-finish
5532 @item reverse-finish
5533 Just as the @code{finish} command takes you to the point where the
5534 current function returns, @code{reverse-finish} takes you to the point
5535 where it was called. Instead of ending up at the end of the current
5536 function invocation, you end up at the beginning.
5538 @kindex set exec-direction
5539 @item set exec-direction
5540 Set the direction of target execution.
5541 @itemx set exec-direction reverse
5542 @cindex execute forward or backward in time
5543 @value{GDBN} will perform all execution commands in reverse, until the
5544 exec-direction mode is changed to ``forward''. Affected commands include
5545 @code{step, stepi, next, nexti, continue, and finish}. The @code{return}
5546 command cannot be used in reverse mode.
5547 @item set exec-direction forward
5548 @value{GDBN} will perform all execution commands in the normal fashion.
5549 This is the default.
5553 @node Process Record and Replay
5554 @chapter Recording Inferior's Execution and Replaying It
5555 @cindex process record and replay
5556 @cindex recording inferior's execution and replaying it
5558 On some platforms, @value{GDBN} provides a special @dfn{process record
5559 and replay} target that can record a log of the process execution, and
5560 replay it later with both forward and reverse execution commands.
5563 When this target is in use, if the execution log includes the record
5564 for the next instruction, @value{GDBN} will debug in @dfn{replay
5565 mode}. In the replay mode, the inferior does not really execute code
5566 instructions. Instead, all the events that normally happen during
5567 code execution are taken from the execution log. While code is not
5568 really executed in replay mode, the values of registers (including the
5569 program counter register) and the memory of the inferior are still
5570 changed as they normally would. Their contents are taken from the
5574 If the record for the next instruction is not in the execution log,
5575 @value{GDBN} will debug in @dfn{record mode}. In this mode, the
5576 inferior executes normally, and @value{GDBN} records the execution log
5579 The process record and replay target supports reverse execution
5580 (@pxref{Reverse Execution}), even if the platform on which the
5581 inferior runs does not. However, the reverse execution is limited in
5582 this case by the range of the instructions recorded in the execution
5583 log. In other words, reverse execution on platforms that don't
5584 support it directly can only be done in the replay mode.
5586 When debugging in the reverse direction, @value{GDBN} will work in
5587 replay mode as long as the execution log includes the record for the
5588 previous instruction; otherwise, it will work in record mode, if the
5589 platform supports reverse execution, or stop if not.
5591 For architecture environments that support process record and replay,
5592 @value{GDBN} provides the following commands:
5595 @kindex target record
5599 This command starts the process record and replay target. The process
5600 record and replay target can only debug a process that is already
5601 running. Therefore, you need first to start the process with the
5602 @kbd{run} or @kbd{start} commands, and then start the recording with
5603 the @kbd{target record} command.
5605 Both @code{record} and @code{rec} are aliases of @code{target record}.
5607 @cindex displaced stepping, and process record and replay
5608 Displaced stepping (@pxref{Maintenance Commands,, displaced stepping})
5609 will be automatically disabled when process record and replay target
5610 is started. That's because the process record and replay target
5611 doesn't support displaced stepping.
5613 @cindex non-stop mode, and process record and replay
5614 @cindex asynchronous execution, and process record and replay
5615 If the inferior is in the non-stop mode (@pxref{Non-Stop Mode}) or in
5616 the asynchronous execution mode (@pxref{Background Execution}), the
5617 process record and replay target cannot be started because it doesn't
5618 support these two modes.
5623 Stop the process record and replay target. When process record and
5624 replay target stops, the entire execution log will be deleted and the
5625 inferior will either be terminated, or will remain in its final state.
5627 When you stop the process record and replay target in record mode (at
5628 the end of the execution log), the inferior will be stopped at the
5629 next instruction that would have been recorded. In other words, if
5630 you record for a while and then stop recording, the inferior process
5631 will be left in the same state as if the recording never happened.
5633 On the other hand, if the process record and replay target is stopped
5634 while in replay mode (that is, not at the end of the execution log,
5635 but at some earlier point), the inferior process will become ``live''
5636 at that earlier state, and it will then be possible to continue the
5637 usual ``live'' debugging of the process from that state.
5639 When the inferior process exits, or @value{GDBN} detaches from it,
5640 process record and replay target will automatically stop itself.
5642 @kindex set record insn-number-max
5643 @item set record insn-number-max @var{limit}
5644 Set the limit of instructions to be recorded. Default value is 200000.
5646 If @var{limit} is a positive number, then @value{GDBN} will start
5647 deleting instructions from the log once the number of the record
5648 instructions becomes greater than @var{limit}. For every new recorded
5649 instruction, @value{GDBN} will delete the earliest recorded
5650 instruction to keep the number of recorded instructions at the limit.
5651 (Since deleting recorded instructions loses information, @value{GDBN}
5652 lets you control what happens when the limit is reached, by means of
5653 the @code{stop-at-limit} option, described below.)
5655 If @var{limit} is zero, @value{GDBN} will never delete recorded
5656 instructions from the execution log. The number of recorded
5657 instructions is unlimited in this case.
5659 @kindex show record insn-number-max
5660 @item show record insn-number-max
5661 Show the limit of instructions to be recorded.
5663 @kindex set record stop-at-limit
5664 @item set record stop-at-limit
5665 Control the behavior when the number of recorded instructions reaches
5666 the limit. If ON (the default), @value{GDBN} will stop when the limit
5667 is reached for the first time and ask you whether you want to stop the
5668 inferior or continue running it and recording the execution log. If
5669 you decide to continue recording, each new recorded instruction will
5670 cause the oldest one to be deleted.
5672 If this option is OFF, @value{GDBN} will automatically delete the
5673 oldest record to make room for each new one, without asking.
5675 @kindex show record stop-at-limit
5676 @item show record stop-at-limit
5677 Show the current setting of @code{stop-at-limit}.
5681 Show various statistics about the state of process record and its
5682 in-memory execution log buffer, including:
5686 Whether in record mode or replay mode.
5688 Lowest recorded instruction number (counting from when the current execution log started recording instructions).
5690 Highest recorded instruction number.
5692 Current instruction about to be replayed (if in replay mode).
5694 Number of instructions contained in the execution log.
5696 Maximum number of instructions that may be contained in the execution log.
5699 @kindex record delete
5702 When record target runs in replay mode (``in the past''), delete the
5703 subsequent execution log and begin to record a new execution log starting
5704 from the current address. This means you will abandon the previously
5705 recorded ``future'' and begin recording a new ``future''.
5710 @chapter Examining the Stack
5712 When your program has stopped, the first thing you need to know is where it
5713 stopped and how it got there.
5716 Each time your program performs a function call, information about the call
5718 That information includes the location of the call in your program,
5719 the arguments of the call,
5720 and the local variables of the function being called.
5721 The information is saved in a block of data called a @dfn{stack frame}.
5722 The stack frames are allocated in a region of memory called the @dfn{call
5725 When your program stops, the @value{GDBN} commands for examining the
5726 stack allow you to see all of this information.
5728 @cindex selected frame
5729 One of the stack frames is @dfn{selected} by @value{GDBN} and many
5730 @value{GDBN} commands refer implicitly to the selected frame. In
5731 particular, whenever you ask @value{GDBN} for the value of a variable in
5732 your program, the value is found in the selected frame. There are
5733 special @value{GDBN} commands to select whichever frame you are
5734 interested in. @xref{Selection, ,Selecting a Frame}.
5736 When your program stops, @value{GDBN} automatically selects the
5737 currently executing frame and describes it briefly, similar to the
5738 @code{frame} command (@pxref{Frame Info, ,Information about a Frame}).
5741 * Frames:: Stack frames
5742 * Backtrace:: Backtraces
5743 * Selection:: Selecting a frame
5744 * Frame Info:: Information on a frame
5749 @section Stack Frames
5751 @cindex frame, definition
5753 The call stack is divided up into contiguous pieces called @dfn{stack
5754 frames}, or @dfn{frames} for short; each frame is the data associated
5755 with one call to one function. The frame contains the arguments given
5756 to the function, the function's local variables, and the address at
5757 which the function is executing.
5759 @cindex initial frame
5760 @cindex outermost frame
5761 @cindex innermost frame
5762 When your program is started, the stack has only one frame, that of the
5763 function @code{main}. This is called the @dfn{initial} frame or the
5764 @dfn{outermost} frame. Each time a function is called, a new frame is
5765 made. Each time a function returns, the frame for that function invocation
5766 is eliminated. If a function is recursive, there can be many frames for
5767 the same function. The frame for the function in which execution is
5768 actually occurring is called the @dfn{innermost} frame. This is the most
5769 recently created of all the stack frames that still exist.
5771 @cindex frame pointer
5772 Inside your program, stack frames are identified by their addresses. A
5773 stack frame consists of many bytes, each of which has its own address; each
5774 kind of computer has a convention for choosing one byte whose
5775 address serves as the address of the frame. Usually this address is kept
5776 in a register called the @dfn{frame pointer register}
5777 (@pxref{Registers, $fp}) while execution is going on in that frame.
5779 @cindex frame number
5780 @value{GDBN} assigns numbers to all existing stack frames, starting with
5781 zero for the innermost frame, one for the frame that called it,
5782 and so on upward. These numbers do not really exist in your program;
5783 they are assigned by @value{GDBN} to give you a way of designating stack
5784 frames in @value{GDBN} commands.
5786 @c The -fomit-frame-pointer below perennially causes hbox overflow
5787 @c underflow problems.
5788 @cindex frameless execution
5789 Some compilers provide a way to compile functions so that they operate
5790 without stack frames. (For example, the @value{NGCC} option
5792 @samp{-fomit-frame-pointer}
5794 generates functions without a frame.)
5795 This is occasionally done with heavily used library functions to save
5796 the frame setup time. @value{GDBN} has limited facilities for dealing
5797 with these function invocations. If the innermost function invocation
5798 has no stack frame, @value{GDBN} nevertheless regards it as though
5799 it had a separate frame, which is numbered zero as usual, allowing
5800 correct tracing of the function call chain. However, @value{GDBN} has
5801 no provision for frameless functions elsewhere in the stack.
5804 @kindex frame@r{, command}
5805 @cindex current stack frame
5806 @item frame @var{args}
5807 The @code{frame} command allows you to move from one stack frame to another,
5808 and to print the stack frame you select. @var{args} may be either the
5809 address of the frame or the stack frame number. Without an argument,
5810 @code{frame} prints the current stack frame.
5812 @kindex select-frame
5813 @cindex selecting frame silently
5815 The @code{select-frame} command allows you to move from one stack frame
5816 to another without printing the frame. This is the silent version of
5824 @cindex call stack traces
5825 A backtrace is a summary of how your program got where it is. It shows one
5826 line per frame, for many frames, starting with the currently executing
5827 frame (frame zero), followed by its caller (frame one), and on up the
5832 @kindex bt @r{(@code{backtrace})}
5835 Print a backtrace of the entire stack: one line per frame for all
5836 frames in the stack.
5838 You can stop the backtrace at any time by typing the system interrupt
5839 character, normally @kbd{Ctrl-c}.
5841 @item backtrace @var{n}
5843 Similar, but print only the innermost @var{n} frames.
5845 @item backtrace -@var{n}
5847 Similar, but print only the outermost @var{n} frames.
5849 @item backtrace full
5851 @itemx bt full @var{n}
5852 @itemx bt full -@var{n}
5853 Print the values of the local variables also. @var{n} specifies the
5854 number of frames to print, as described above.
5859 The names @code{where} and @code{info stack} (abbreviated @code{info s})
5860 are additional aliases for @code{backtrace}.
5862 @cindex multiple threads, backtrace
5863 In a multi-threaded program, @value{GDBN} by default shows the
5864 backtrace only for the current thread. To display the backtrace for
5865 several or all of the threads, use the command @code{thread apply}
5866 (@pxref{Threads, thread apply}). For example, if you type @kbd{thread
5867 apply all backtrace}, @value{GDBN} will display the backtrace for all
5868 the threads; this is handy when you debug a core dump of a
5869 multi-threaded program.
5871 Each line in the backtrace shows the frame number and the function name.
5872 The program counter value is also shown---unless you use @code{set
5873 print address off}. The backtrace also shows the source file name and
5874 line number, as well as the arguments to the function. The program
5875 counter value is omitted if it is at the beginning of the code for that
5878 Here is an example of a backtrace. It was made with the command
5879 @samp{bt 3}, so it shows the innermost three frames.
5883 #0 m4_traceon (obs=0x24eb0, argc=1, argv=0x2b8c8)
5885 #1 0x6e38 in expand_macro (sym=0x2b600, data=...) at macro.c:242
5886 #2 0x6840 in expand_token (obs=0x0, t=177664, td=0xf7fffb08)
5888 (More stack frames follow...)
5893 The display for frame zero does not begin with a program counter
5894 value, indicating that your program has stopped at the beginning of the
5895 code for line @code{993} of @code{builtin.c}.
5898 The value of parameter @code{data} in frame 1 has been replaced by
5899 @code{@dots{}}. By default, @value{GDBN} prints the value of a parameter
5900 only if it is a scalar (integer, pointer, enumeration, etc). See command
5901 @kbd{set print frame-arguments} in @ref{Print Settings} for more details
5902 on how to configure the way function parameter values are printed.
5904 @cindex value optimized out, in backtrace
5905 @cindex function call arguments, optimized out
5906 If your program was compiled with optimizations, some compilers will
5907 optimize away arguments passed to functions if those arguments are
5908 never used after the call. Such optimizations generate code that
5909 passes arguments through registers, but doesn't store those arguments
5910 in the stack frame. @value{GDBN} has no way of displaying such
5911 arguments in stack frames other than the innermost one. Here's what
5912 such a backtrace might look like:
5916 #0 m4_traceon (obs=0x24eb0, argc=1, argv=0x2b8c8)
5918 #1 0x6e38 in expand_macro (sym=<value optimized out>) at macro.c:242
5919 #2 0x6840 in expand_token (obs=0x0, t=<value optimized out>, td=0xf7fffb08)
5921 (More stack frames follow...)
5926 The values of arguments that were not saved in their stack frames are
5927 shown as @samp{<value optimized out>}.
5929 If you need to display the values of such optimized-out arguments,
5930 either deduce that from other variables whose values depend on the one
5931 you are interested in, or recompile without optimizations.
5933 @cindex backtrace beyond @code{main} function
5934 @cindex program entry point
5935 @cindex startup code, and backtrace
5936 Most programs have a standard user entry point---a place where system
5937 libraries and startup code transition into user code. For C this is
5938 @code{main}@footnote{
5939 Note that embedded programs (the so-called ``free-standing''
5940 environment) are not required to have a @code{main} function as the
5941 entry point. They could even have multiple entry points.}.
5942 When @value{GDBN} finds the entry function in a backtrace
5943 it will terminate the backtrace, to avoid tracing into highly
5944 system-specific (and generally uninteresting) code.
5946 If you need to examine the startup code, or limit the number of levels
5947 in a backtrace, you can change this behavior:
5950 @item set backtrace past-main
5951 @itemx set backtrace past-main on
5952 @kindex set backtrace
5953 Backtraces will continue past the user entry point.
5955 @item set backtrace past-main off
5956 Backtraces will stop when they encounter the user entry point. This is the
5959 @item show backtrace past-main
5960 @kindex show backtrace
5961 Display the current user entry point backtrace policy.
5963 @item set backtrace past-entry
5964 @itemx set backtrace past-entry on
5965 Backtraces will continue past the internal entry point of an application.
5966 This entry point is encoded by the linker when the application is built,
5967 and is likely before the user entry point @code{main} (or equivalent) is called.
5969 @item set backtrace past-entry off
5970 Backtraces will stop when they encounter the internal entry point of an
5971 application. This is the default.
5973 @item show backtrace past-entry
5974 Display the current internal entry point backtrace policy.
5976 @item set backtrace limit @var{n}
5977 @itemx set backtrace limit 0
5978 @cindex backtrace limit
5979 Limit the backtrace to @var{n} levels. A value of zero means
5982 @item show backtrace limit
5983 Display the current limit on backtrace levels.
5987 @section Selecting a Frame
5989 Most commands for examining the stack and other data in your program work on
5990 whichever stack frame is selected at the moment. Here are the commands for
5991 selecting a stack frame; all of them finish by printing a brief description
5992 of the stack frame just selected.
5995 @kindex frame@r{, selecting}
5996 @kindex f @r{(@code{frame})}
5999 Select frame number @var{n}. Recall that frame zero is the innermost
6000 (currently executing) frame, frame one is the frame that called the
6001 innermost one, and so on. The highest-numbered frame is the one for
6004 @item frame @var{addr}
6006 Select the frame at address @var{addr}. This is useful mainly if the
6007 chaining of stack frames has been damaged by a bug, making it
6008 impossible for @value{GDBN} to assign numbers properly to all frames. In
6009 addition, this can be useful when your program has multiple stacks and
6010 switches between them.
6012 On the SPARC architecture, @code{frame} needs two addresses to
6013 select an arbitrary frame: a frame pointer and a stack pointer.
6015 On the MIPS and Alpha architecture, it needs two addresses: a stack
6016 pointer and a program counter.
6018 On the 29k architecture, it needs three addresses: a register stack
6019 pointer, a program counter, and a memory stack pointer.
6023 Move @var{n} frames up the stack. For positive numbers @var{n}, this
6024 advances toward the outermost frame, to higher frame numbers, to frames
6025 that have existed longer. @var{n} defaults to one.
6028 @kindex do @r{(@code{down})}
6030 Move @var{n} frames down the stack. For positive numbers @var{n}, this
6031 advances toward the innermost frame, to lower frame numbers, to frames
6032 that were created more recently. @var{n} defaults to one. You may
6033 abbreviate @code{down} as @code{do}.
6036 All of these commands end by printing two lines of output describing the
6037 frame. The first line shows the frame number, the function name, the
6038 arguments, and the source file and line number of execution in that
6039 frame. The second line shows the text of that source line.
6047 #1 0x22f0 in main (argc=1, argv=0xf7fffbf4, env=0xf7fffbfc)
6049 10 read_input_file (argv[i]);
6053 After such a printout, the @code{list} command with no arguments
6054 prints ten lines centered on the point of execution in the frame.
6055 You can also edit the program at the point of execution with your favorite
6056 editing program by typing @code{edit}.
6057 @xref{List, ,Printing Source Lines},
6061 @kindex down-silently
6063 @item up-silently @var{n}
6064 @itemx down-silently @var{n}
6065 These two commands are variants of @code{up} and @code{down},
6066 respectively; they differ in that they do their work silently, without
6067 causing display of the new frame. They are intended primarily for use
6068 in @value{GDBN} command scripts, where the output might be unnecessary and
6073 @section Information About a Frame
6075 There are several other commands to print information about the selected
6081 When used without any argument, this command does not change which
6082 frame is selected, but prints a brief description of the currently
6083 selected stack frame. It can be abbreviated @code{f}. With an
6084 argument, this command is used to select a stack frame.
6085 @xref{Selection, ,Selecting a Frame}.
6088 @kindex info f @r{(@code{info frame})}
6091 This command prints a verbose description of the selected stack frame,
6096 the address of the frame
6098 the address of the next frame down (called by this frame)
6100 the address of the next frame up (caller of this frame)
6102 the language in which the source code corresponding to this frame is written
6104 the address of the frame's arguments
6106 the address of the frame's local variables
6108 the program counter saved in it (the address of execution in the caller frame)
6110 which registers were saved in the frame
6113 @noindent The verbose description is useful when
6114 something has gone wrong that has made the stack format fail to fit
6115 the usual conventions.
6117 @item info frame @var{addr}
6118 @itemx info f @var{addr}
6119 Print a verbose description of the frame at address @var{addr}, without
6120 selecting that frame. The selected frame remains unchanged by this
6121 command. This requires the same kind of address (more than one for some
6122 architectures) that you specify in the @code{frame} command.
6123 @xref{Selection, ,Selecting a Frame}.
6127 Print the arguments of the selected frame, each on a separate line.
6131 Print the local variables of the selected frame, each on a separate
6132 line. These are all variables (declared either static or automatic)
6133 accessible at the point of execution of the selected frame.
6136 @cindex catch exceptions, list active handlers
6137 @cindex exception handlers, how to list
6139 Print a list of all the exception handlers that are active in the
6140 current stack frame at the current point of execution. To see other
6141 exception handlers, visit the associated frame (using the @code{up},
6142 @code{down}, or @code{frame} commands); then type @code{info catch}.
6143 @xref{Set Catchpoints, , Setting Catchpoints}.
6149 @chapter Examining Source Files
6151 @value{GDBN} can print parts of your program's source, since the debugging
6152 information recorded in the program tells @value{GDBN} what source files were
6153 used to build it. When your program stops, @value{GDBN} spontaneously prints
6154 the line where it stopped. Likewise, when you select a stack frame
6155 (@pxref{Selection, ,Selecting a Frame}), @value{GDBN} prints the line where
6156 execution in that frame has stopped. You can print other portions of
6157 source files by explicit command.
6159 If you use @value{GDBN} through its @sc{gnu} Emacs interface, you may
6160 prefer to use Emacs facilities to view source; see @ref{Emacs, ,Using
6161 @value{GDBN} under @sc{gnu} Emacs}.
6164 * List:: Printing source lines
6165 * Specify Location:: How to specify code locations
6166 * Edit:: Editing source files
6167 * Search:: Searching source files
6168 * Source Path:: Specifying source directories
6169 * Machine Code:: Source and machine code
6173 @section Printing Source Lines
6176 @kindex l @r{(@code{list})}
6177 To print lines from a source file, use the @code{list} command
6178 (abbreviated @code{l}). By default, ten lines are printed.
6179 There are several ways to specify what part of the file you want to
6180 print; see @ref{Specify Location}, for the full list.
6182 Here are the forms of the @code{list} command most commonly used:
6185 @item list @var{linenum}
6186 Print lines centered around line number @var{linenum} in the
6187 current source file.
6189 @item list @var{function}
6190 Print lines centered around the beginning of function
6194 Print more lines. If the last lines printed were printed with a
6195 @code{list} command, this prints lines following the last lines
6196 printed; however, if the last line printed was a solitary line printed
6197 as part of displaying a stack frame (@pxref{Stack, ,Examining the
6198 Stack}), this prints lines centered around that line.
6201 Print lines just before the lines last printed.
6204 @cindex @code{list}, how many lines to display
6205 By default, @value{GDBN} prints ten source lines with any of these forms of
6206 the @code{list} command. You can change this using @code{set listsize}:
6209 @kindex set listsize
6210 @item set listsize @var{count}
6211 Make the @code{list} command display @var{count} source lines (unless
6212 the @code{list} argument explicitly specifies some other number).
6214 @kindex show listsize
6216 Display the number of lines that @code{list} prints.
6219 Repeating a @code{list} command with @key{RET} discards the argument,
6220 so it is equivalent to typing just @code{list}. This is more useful
6221 than listing the same lines again. An exception is made for an
6222 argument of @samp{-}; that argument is preserved in repetition so that
6223 each repetition moves up in the source file.
6225 In general, the @code{list} command expects you to supply zero, one or two
6226 @dfn{linespecs}. Linespecs specify source lines; there are several ways
6227 of writing them (@pxref{Specify Location}), but the effect is always
6228 to specify some source line.
6230 Here is a complete description of the possible arguments for @code{list}:
6233 @item list @var{linespec}
6234 Print lines centered around the line specified by @var{linespec}.
6236 @item list @var{first},@var{last}
6237 Print lines from @var{first} to @var{last}. Both arguments are
6238 linespecs. When a @code{list} command has two linespecs, and the
6239 source file of the second linespec is omitted, this refers to
6240 the same source file as the first linespec.
6242 @item list ,@var{last}
6243 Print lines ending with @var{last}.
6245 @item list @var{first},
6246 Print lines starting with @var{first}.
6249 Print lines just after the lines last printed.
6252 Print lines just before the lines last printed.
6255 As described in the preceding table.
6258 @node Specify Location
6259 @section Specifying a Location
6260 @cindex specifying location
6263 Several @value{GDBN} commands accept arguments that specify a location
6264 of your program's code. Since @value{GDBN} is a source-level
6265 debugger, a location usually specifies some line in the source code;
6266 for that reason, locations are also known as @dfn{linespecs}.
6268 Here are all the different ways of specifying a code location that
6269 @value{GDBN} understands:
6273 Specifies the line number @var{linenum} of the current source file.
6276 @itemx +@var{offset}
6277 Specifies the line @var{offset} lines before or after the @dfn{current
6278 line}. For the @code{list} command, the current line is the last one
6279 printed; for the breakpoint commands, this is the line at which
6280 execution stopped in the currently selected @dfn{stack frame}
6281 (@pxref{Frames, ,Frames}, for a description of stack frames.) When
6282 used as the second of the two linespecs in a @code{list} command,
6283 this specifies the line @var{offset} lines up or down from the first
6286 @item @var{filename}:@var{linenum}
6287 Specifies the line @var{linenum} in the source file @var{filename}.
6289 @item @var{function}
6290 Specifies the line that begins the body of the function @var{function}.
6291 For example, in C, this is the line with the open brace.
6293 @item @var{filename}:@var{function}
6294 Specifies the line that begins the body of the function @var{function}
6295 in the file @var{filename}. You only need the file name with a
6296 function name to avoid ambiguity when there are identically named
6297 functions in different source files.
6299 @item *@var{address}
6300 Specifies the program address @var{address}. For line-oriented
6301 commands, such as @code{list} and @code{edit}, this specifies a source
6302 line that contains @var{address}. For @code{break} and other
6303 breakpoint oriented commands, this can be used to set breakpoints in
6304 parts of your program which do not have debugging information or
6307 Here @var{address} may be any expression valid in the current working
6308 language (@pxref{Languages, working language}) that specifies a code
6309 address. In addition, as a convenience, @value{GDBN} extends the
6310 semantics of expressions used in locations to cover the situations
6311 that frequently happen during debugging. Here are the various forms
6315 @item @var{expression}
6316 Any expression valid in the current working language.
6318 @item @var{funcaddr}
6319 An address of a function or procedure derived from its name. In C,
6320 C@t{++}, Java, Objective-C, Fortran, minimal, and assembly, this is
6321 simply the function's name @var{function} (and actually a special case
6322 of a valid expression). In Pascal and Modula-2, this is
6323 @code{&@var{function}}. In Ada, this is @code{@var{function}'Address}
6324 (although the Pascal form also works).
6326 This form specifies the address of the function's first instruction,
6327 before the stack frame and arguments have been set up.
6329 @item '@var{filename}'::@var{funcaddr}
6330 Like @var{funcaddr} above, but also specifies the name of the source
6331 file explicitly. This is useful if the name of the function does not
6332 specify the function unambiguously, e.g., if there are several
6333 functions with identical names in different source files.
6340 @section Editing Source Files
6341 @cindex editing source files
6344 @kindex e @r{(@code{edit})}
6345 To edit the lines in a source file, use the @code{edit} command.
6346 The editing program of your choice
6347 is invoked with the current line set to
6348 the active line in the program.
6349 Alternatively, there are several ways to specify what part of the file you
6350 want to print if you want to see other parts of the program:
6353 @item edit @var{location}
6354 Edit the source file specified by @code{location}. Editing starts at
6355 that @var{location}, e.g., at the specified source line of the
6356 specified file. @xref{Specify Location}, for all the possible forms
6357 of the @var{location} argument; here are the forms of the @code{edit}
6358 command most commonly used:
6361 @item edit @var{number}
6362 Edit the current source file with @var{number} as the active line number.
6364 @item edit @var{function}
6365 Edit the file containing @var{function} at the beginning of its definition.
6370 @subsection Choosing your Editor
6371 You can customize @value{GDBN} to use any editor you want
6373 The only restriction is that your editor (say @code{ex}), recognizes the
6374 following command-line syntax:
6376 ex +@var{number} file
6378 The optional numeric value +@var{number} specifies the number of the line in
6379 the file where to start editing.}.
6380 By default, it is @file{@value{EDITOR}}, but you can change this
6381 by setting the environment variable @code{EDITOR} before using
6382 @value{GDBN}. For example, to configure @value{GDBN} to use the
6383 @code{vi} editor, you could use these commands with the @code{sh} shell:
6389 or in the @code{csh} shell,
6391 setenv EDITOR /usr/bin/vi
6396 @section Searching Source Files
6397 @cindex searching source files
6399 There are two commands for searching through the current source file for a
6404 @kindex forward-search
6405 @item forward-search @var{regexp}
6406 @itemx search @var{regexp}
6407 The command @samp{forward-search @var{regexp}} checks each line,
6408 starting with the one following the last line listed, for a match for
6409 @var{regexp}. It lists the line that is found. You can use the
6410 synonym @samp{search @var{regexp}} or abbreviate the command name as
6413 @kindex reverse-search
6414 @item reverse-search @var{regexp}
6415 The command @samp{reverse-search @var{regexp}} checks each line, starting
6416 with the one before the last line listed and going backward, for a match
6417 for @var{regexp}. It lists the line that is found. You can abbreviate
6418 this command as @code{rev}.
6422 @section Specifying Source Directories
6425 @cindex directories for source files
6426 Executable programs sometimes do not record the directories of the source
6427 files from which they were compiled, just the names. Even when they do,
6428 the directories could be moved between the compilation and your debugging
6429 session. @value{GDBN} has a list of directories to search for source files;
6430 this is called the @dfn{source path}. Each time @value{GDBN} wants a source file,
6431 it tries all the directories in the list, in the order they are present
6432 in the list, until it finds a file with the desired name.
6434 For example, suppose an executable references the file
6435 @file{/usr/src/foo-1.0/lib/foo.c}, and our source path is
6436 @file{/mnt/cross}. The file is first looked up literally; if this
6437 fails, @file{/mnt/cross/usr/src/foo-1.0/lib/foo.c} is tried; if this
6438 fails, @file{/mnt/cross/foo.c} is opened; if this fails, an error
6439 message is printed. @value{GDBN} does not look up the parts of the
6440 source file name, such as @file{/mnt/cross/src/foo-1.0/lib/foo.c}.
6441 Likewise, the subdirectories of the source path are not searched: if
6442 the source path is @file{/mnt/cross}, and the binary refers to
6443 @file{foo.c}, @value{GDBN} would not find it under
6444 @file{/mnt/cross/usr/src/foo-1.0/lib}.
6446 Plain file names, relative file names with leading directories, file
6447 names containing dots, etc.@: are all treated as described above; for
6448 instance, if the source path is @file{/mnt/cross}, and the source file
6449 is recorded as @file{../lib/foo.c}, @value{GDBN} would first try
6450 @file{../lib/foo.c}, then @file{/mnt/cross/../lib/foo.c}, and after
6451 that---@file{/mnt/cross/foo.c}.
6453 Note that the executable search path is @emph{not} used to locate the
6456 Whenever you reset or rearrange the source path, @value{GDBN} clears out
6457 any information it has cached about where source files are found and where
6458 each line is in the file.
6462 When you start @value{GDBN}, its source path includes only @samp{cdir}
6463 and @samp{cwd}, in that order.
6464 To add other directories, use the @code{directory} command.
6466 The search path is used to find both program source files and @value{GDBN}
6467 script files (read using the @samp{-command} option and @samp{source} command).
6469 In addition to the source path, @value{GDBN} provides a set of commands
6470 that manage a list of source path substitution rules. A @dfn{substitution
6471 rule} specifies how to rewrite source directories stored in the program's
6472 debug information in case the sources were moved to a different
6473 directory between compilation and debugging. A rule is made of
6474 two strings, the first specifying what needs to be rewritten in
6475 the path, and the second specifying how it should be rewritten.
6476 In @ref{set substitute-path}, we name these two parts @var{from} and
6477 @var{to} respectively. @value{GDBN} does a simple string replacement
6478 of @var{from} with @var{to} at the start of the directory part of the
6479 source file name, and uses that result instead of the original file
6480 name to look up the sources.
6482 Using the previous example, suppose the @file{foo-1.0} tree has been
6483 moved from @file{/usr/src} to @file{/mnt/cross}, then you can tell
6484 @value{GDBN} to replace @file{/usr/src} in all source path names with
6485 @file{/mnt/cross}. The first lookup will then be
6486 @file{/mnt/cross/foo-1.0/lib/foo.c} in place of the original location
6487 of @file{/usr/src/foo-1.0/lib/foo.c}. To define a source path
6488 substitution rule, use the @code{set substitute-path} command
6489 (@pxref{set substitute-path}).
6491 To avoid unexpected substitution results, a rule is applied only if the
6492 @var{from} part of the directory name ends at a directory separator.
6493 For instance, a rule substituting @file{/usr/source} into
6494 @file{/mnt/cross} will be applied to @file{/usr/source/foo-1.0} but
6495 not to @file{/usr/sourceware/foo-2.0}. And because the substitution
6496 is applied only at the beginning of the directory name, this rule will
6497 not be applied to @file{/root/usr/source/baz.c} either.
6499 In many cases, you can achieve the same result using the @code{directory}
6500 command. However, @code{set substitute-path} can be more efficient in
6501 the case where the sources are organized in a complex tree with multiple
6502 subdirectories. With the @code{directory} command, you need to add each
6503 subdirectory of your project. If you moved the entire tree while
6504 preserving its internal organization, then @code{set substitute-path}
6505 allows you to direct the debugger to all the sources with one single
6508 @code{set substitute-path} is also more than just a shortcut command.
6509 The source path is only used if the file at the original location no
6510 longer exists. On the other hand, @code{set substitute-path} modifies
6511 the debugger behavior to look at the rewritten location instead. So, if
6512 for any reason a source file that is not relevant to your executable is
6513 located at the original location, a substitution rule is the only
6514 method available to point @value{GDBN} at the new location.
6516 @cindex @samp{--with-relocated-sources}
6517 @cindex default source path substitution
6518 You can configure a default source path substitution rule by
6519 configuring @value{GDBN} with the
6520 @samp{--with-relocated-sources=@var{dir}} option. The @var{dir}
6521 should be the name of a directory under @value{GDBN}'s configured
6522 prefix (set with @samp{--prefix} or @samp{--exec-prefix}), and
6523 directory names in debug information under @var{dir} will be adjusted
6524 automatically if the installed @value{GDBN} is moved to a new
6525 location. This is useful if @value{GDBN}, libraries or executables
6526 with debug information and corresponding source code are being moved
6530 @item directory @var{dirname} @dots{}
6531 @item dir @var{dirname} @dots{}
6532 Add directory @var{dirname} to the front of the source path. Several
6533 directory names may be given to this command, separated by @samp{:}
6534 (@samp{;} on MS-DOS and MS-Windows, where @samp{:} usually appears as
6535 part of absolute file names) or
6536 whitespace. You may specify a directory that is already in the source
6537 path; this moves it forward, so @value{GDBN} searches it sooner.
6541 @vindex $cdir@r{, convenience variable}
6542 @vindex $cwd@r{, convenience variable}
6543 @cindex compilation directory
6544 @cindex current directory
6545 @cindex working directory
6546 @cindex directory, current
6547 @cindex directory, compilation
6548 You can use the string @samp{$cdir} to refer to the compilation
6549 directory (if one is recorded), and @samp{$cwd} to refer to the current
6550 working directory. @samp{$cwd} is not the same as @samp{.}---the former
6551 tracks the current working directory as it changes during your @value{GDBN}
6552 session, while the latter is immediately expanded to the current
6553 directory at the time you add an entry to the source path.
6556 Reset the source path to its default value (@samp{$cdir:$cwd} on Unix systems). This requires confirmation.
6558 @c RET-repeat for @code{directory} is explicitly disabled, but since
6559 @c repeating it would be a no-op we do not say that. (thanks to RMS)
6561 @item show directories
6562 @kindex show directories
6563 Print the source path: show which directories it contains.
6565 @anchor{set substitute-path}
6566 @item set substitute-path @var{from} @var{to}
6567 @kindex set substitute-path
6568 Define a source path substitution rule, and add it at the end of the
6569 current list of existing substitution rules. If a rule with the same
6570 @var{from} was already defined, then the old rule is also deleted.
6572 For example, if the file @file{/foo/bar/baz.c} was moved to
6573 @file{/mnt/cross/baz.c}, then the command
6576 (@value{GDBP}) set substitute-path /usr/src /mnt/cross
6580 will tell @value{GDBN} to replace @samp{/usr/src} with
6581 @samp{/mnt/cross}, which will allow @value{GDBN} to find the file
6582 @file{baz.c} even though it was moved.
6584 In the case when more than one substitution rule have been defined,
6585 the rules are evaluated one by one in the order where they have been
6586 defined. The first one matching, if any, is selected to perform
6589 For instance, if we had entered the following commands:
6592 (@value{GDBP}) set substitute-path /usr/src/include /mnt/include
6593 (@value{GDBP}) set substitute-path /usr/src /mnt/src
6597 @value{GDBN} would then rewrite @file{/usr/src/include/defs.h} into
6598 @file{/mnt/include/defs.h} by using the first rule. However, it would
6599 use the second rule to rewrite @file{/usr/src/lib/foo.c} into
6600 @file{/mnt/src/lib/foo.c}.
6603 @item unset substitute-path [path]
6604 @kindex unset substitute-path
6605 If a path is specified, search the current list of substitution rules
6606 for a rule that would rewrite that path. Delete that rule if found.
6607 A warning is emitted by the debugger if no rule could be found.
6609 If no path is specified, then all substitution rules are deleted.
6611 @item show substitute-path [path]
6612 @kindex show substitute-path
6613 If a path is specified, then print the source path substitution rule
6614 which would rewrite that path, if any.
6616 If no path is specified, then print all existing source path substitution
6621 If your source path is cluttered with directories that are no longer of
6622 interest, @value{GDBN} may sometimes cause confusion by finding the wrong
6623 versions of source. You can correct the situation as follows:
6627 Use @code{directory} with no argument to reset the source path to its default value.
6630 Use @code{directory} with suitable arguments to reinstall the
6631 directories you want in the source path. You can add all the
6632 directories in one command.
6636 @section Source and Machine Code
6637 @cindex source line and its code address
6639 You can use the command @code{info line} to map source lines to program
6640 addresses (and vice versa), and the command @code{disassemble} to display
6641 a range of addresses as machine instructions. You can use the command
6642 @code{set disassemble-next-line} to set whether to disassemble next
6643 source line when execution stops. When run under @sc{gnu} Emacs
6644 mode, the @code{info line} command causes the arrow to point to the
6645 line specified. Also, @code{info line} prints addresses in symbolic form as
6650 @item info line @var{linespec}
6651 Print the starting and ending addresses of the compiled code for
6652 source line @var{linespec}. You can specify source lines in any of
6653 the ways documented in @ref{Specify Location}.
6656 For example, we can use @code{info line} to discover the location of
6657 the object code for the first line of function
6658 @code{m4_changequote}:
6660 @c FIXME: I think this example should also show the addresses in
6661 @c symbolic form, as they usually would be displayed.
6663 (@value{GDBP}) info line m4_changequote
6664 Line 895 of "builtin.c" starts at pc 0x634c and ends at 0x6350.
6668 @cindex code address and its source line
6669 We can also inquire (using @code{*@var{addr}} as the form for
6670 @var{linespec}) what source line covers a particular address:
6672 (@value{GDBP}) info line *0x63ff
6673 Line 926 of "builtin.c" starts at pc 0x63e4 and ends at 0x6404.
6676 @cindex @code{$_} and @code{info line}
6677 @cindex @code{x} command, default address
6678 @kindex x@r{(examine), and} info line
6679 After @code{info line}, the default address for the @code{x} command
6680 is changed to the starting address of the line, so that @samp{x/i} is
6681 sufficient to begin examining the machine code (@pxref{Memory,
6682 ,Examining Memory}). Also, this address is saved as the value of the
6683 convenience variable @code{$_} (@pxref{Convenience Vars, ,Convenience
6688 @cindex assembly instructions
6689 @cindex instructions, assembly
6690 @cindex machine instructions
6691 @cindex listing machine instructions
6693 @itemx disassemble /m
6694 @itemx disassemble /r
6695 This specialized command dumps a range of memory as machine
6696 instructions. It can also print mixed source+disassembly by specifying
6697 the @code{/m} modifier and print the raw instructions in hex as well as
6698 in symbolic form by specifying the @code{/r}.
6699 The default memory range is the function surrounding the
6700 program counter of the selected frame. A single argument to this
6701 command is a program counter value; @value{GDBN} dumps the function
6702 surrounding this value. When two arguments are given, they should
6703 be separated by a comma, possibly surrounded by whitespace. The
6704 arguments specify a range of addresses (first inclusive, second exclusive)
6705 to dump. In that case, the name of the function is also printed (since
6706 there could be several functions in the given range).
6708 The argument(s) can be any expression yielding a numeric value, such as
6709 @samp{0x32c4}, @samp{&main+10} or @samp{$pc - 8}.
6711 If the range of memory being disassembled contains current program counter,
6712 the instruction at that location is shown with a @code{=>} marker.
6715 The following example shows the disassembly of a range of addresses of
6716 HP PA-RISC 2.0 code:
6719 (@value{GDBP}) disas 0x32c4, 0x32e4
6720 Dump of assembler code from 0x32c4 to 0x32e4:
6721 0x32c4 <main+204>: addil 0,dp
6722 0x32c8 <main+208>: ldw 0x22c(sr0,r1),r26
6723 0x32cc <main+212>: ldil 0x3000,r31
6724 0x32d0 <main+216>: ble 0x3f8(sr4,r31)
6725 0x32d4 <main+220>: ldo 0(r31),rp
6726 0x32d8 <main+224>: addil -0x800,dp
6727 0x32dc <main+228>: ldo 0x588(r1),r26
6728 0x32e0 <main+232>: ldil 0x3000,r31
6729 End of assembler dump.
6732 Here is an example showing mixed source+assembly for Intel x86, when the
6733 program is stopped just after function prologue:
6736 (@value{GDBP}) disas /m main
6737 Dump of assembler code for function main:
6739 0x08048330 <+0>: push %ebp
6740 0x08048331 <+1>: mov %esp,%ebp
6741 0x08048333 <+3>: sub $0x8,%esp
6742 0x08048336 <+6>: and $0xfffffff0,%esp
6743 0x08048339 <+9>: sub $0x10,%esp
6745 6 printf ("Hello.\n");
6746 => 0x0804833c <+12>: movl $0x8048440,(%esp)
6747 0x08048343 <+19>: call 0x8048284 <puts@@plt>
6751 0x08048348 <+24>: mov $0x0,%eax
6752 0x0804834d <+29>: leave
6753 0x0804834e <+30>: ret
6755 End of assembler dump.
6758 Some architectures have more than one commonly-used set of instruction
6759 mnemonics or other syntax.
6761 For programs that were dynamically linked and use shared libraries,
6762 instructions that call functions or branch to locations in the shared
6763 libraries might show a seemingly bogus location---it's actually a
6764 location of the relocation table. On some architectures, @value{GDBN}
6765 might be able to resolve these to actual function names.
6768 @kindex set disassembly-flavor
6769 @cindex Intel disassembly flavor
6770 @cindex AT&T disassembly flavor
6771 @item set disassembly-flavor @var{instruction-set}
6772 Select the instruction set to use when disassembling the
6773 program via the @code{disassemble} or @code{x/i} commands.
6775 Currently this command is only defined for the Intel x86 family. You
6776 can set @var{instruction-set} to either @code{intel} or @code{att}.
6777 The default is @code{att}, the AT&T flavor used by default by Unix
6778 assemblers for x86-based targets.
6780 @kindex show disassembly-flavor
6781 @item show disassembly-flavor
6782 Show the current setting of the disassembly flavor.
6786 @kindex set disassemble-next-line
6787 @kindex show disassemble-next-line
6788 @item set disassemble-next-line
6789 @itemx show disassemble-next-line
6790 Control whether or not @value{GDBN} will disassemble the next source
6791 line or instruction when execution stops. If ON, @value{GDBN} will
6792 display disassembly of the next source line when execution of the
6793 program being debugged stops. This is @emph{in addition} to
6794 displaying the source line itself, which @value{GDBN} always does if
6795 possible. If the next source line cannot be displayed for some reason
6796 (e.g., if @value{GDBN} cannot find the source file, or there's no line
6797 info in the debug info), @value{GDBN} will display disassembly of the
6798 next @emph{instruction} instead of showing the next source line. If
6799 AUTO, @value{GDBN} will display disassembly of next instruction only
6800 if the source line cannot be displayed. This setting causes
6801 @value{GDBN} to display some feedback when you step through a function
6802 with no line info or whose source file is unavailable. The default is
6803 OFF, which means never display the disassembly of the next line or
6809 @chapter Examining Data
6811 @cindex printing data
6812 @cindex examining data
6815 @c "inspect" is not quite a synonym if you are using Epoch, which we do not
6816 @c document because it is nonstandard... Under Epoch it displays in a
6817 @c different window or something like that.
6818 The usual way to examine data in your program is with the @code{print}
6819 command (abbreviated @code{p}), or its synonym @code{inspect}. It
6820 evaluates and prints the value of an expression of the language your
6821 program is written in (@pxref{Languages, ,Using @value{GDBN} with
6822 Different Languages}). It may also print the expression using a
6823 Python-based pretty-printer (@pxref{Pretty Printing}).
6826 @item print @var{expr}
6827 @itemx print /@var{f} @var{expr}
6828 @var{expr} is an expression (in the source language). By default the
6829 value of @var{expr} is printed in a format appropriate to its data type;
6830 you can choose a different format by specifying @samp{/@var{f}}, where
6831 @var{f} is a letter specifying the format; see @ref{Output Formats,,Output
6835 @itemx print /@var{f}
6836 @cindex reprint the last value
6837 If you omit @var{expr}, @value{GDBN} displays the last value again (from the
6838 @dfn{value history}; @pxref{Value History, ,Value History}). This allows you to
6839 conveniently inspect the same value in an alternative format.
6842 A more low-level way of examining data is with the @code{x} command.
6843 It examines data in memory at a specified address and prints it in a
6844 specified format. @xref{Memory, ,Examining Memory}.
6846 If you are interested in information about types, or about how the
6847 fields of a struct or a class are declared, use the @code{ptype @var{exp}}
6848 command rather than @code{print}. @xref{Symbols, ,Examining the Symbol
6852 * Expressions:: Expressions
6853 * Ambiguous Expressions:: Ambiguous Expressions
6854 * Variables:: Program variables
6855 * Arrays:: Artificial arrays
6856 * Output Formats:: Output formats
6857 * Memory:: Examining memory
6858 * Auto Display:: Automatic display
6859 * Print Settings:: Print settings
6860 * Pretty Printing:: Python pretty printing
6861 * Value History:: Value history
6862 * Convenience Vars:: Convenience variables
6863 * Registers:: Registers
6864 * Floating Point Hardware:: Floating point hardware
6865 * Vector Unit:: Vector Unit
6866 * OS Information:: Auxiliary data provided by operating system
6867 * Memory Region Attributes:: Memory region attributes
6868 * Dump/Restore Files:: Copy between memory and a file
6869 * Core File Generation:: Cause a program dump its core
6870 * Character Sets:: Debugging programs that use a different
6871 character set than GDB does
6872 * Caching Remote Data:: Data caching for remote targets
6873 * Searching Memory:: Searching memory for a sequence of bytes
6877 @section Expressions
6880 @code{print} and many other @value{GDBN} commands accept an expression and
6881 compute its value. Any kind of constant, variable or operator defined
6882 by the programming language you are using is valid in an expression in
6883 @value{GDBN}. This includes conditional expressions, function calls,
6884 casts, and string constants. It also includes preprocessor macros, if
6885 you compiled your program to include this information; see
6888 @cindex arrays in expressions
6889 @value{GDBN} supports array constants in expressions input by
6890 the user. The syntax is @{@var{element}, @var{element}@dots{}@}. For example,
6891 you can use the command @code{print @{1, 2, 3@}} to create an array
6892 of three integers. If you pass an array to a function or assign it
6893 to a program variable, @value{GDBN} copies the array to memory that
6894 is @code{malloc}ed in the target program.
6896 Because C is so widespread, most of the expressions shown in examples in
6897 this manual are in C. @xref{Languages, , Using @value{GDBN} with Different
6898 Languages}, for information on how to use expressions in other
6901 In this section, we discuss operators that you can use in @value{GDBN}
6902 expressions regardless of your programming language.
6904 @cindex casts, in expressions
6905 Casts are supported in all languages, not just in C, because it is so
6906 useful to cast a number into a pointer in order to examine a structure
6907 at that address in memory.
6908 @c FIXME: casts supported---Mod2 true?
6910 @value{GDBN} supports these operators, in addition to those common
6911 to programming languages:
6915 @samp{@@} is a binary operator for treating parts of memory as arrays.
6916 @xref{Arrays, ,Artificial Arrays}, for more information.
6919 @samp{::} allows you to specify a variable in terms of the file or
6920 function where it is defined. @xref{Variables, ,Program Variables}.
6922 @cindex @{@var{type}@}
6923 @cindex type casting memory
6924 @cindex memory, viewing as typed object
6925 @cindex casts, to view memory
6926 @item @{@var{type}@} @var{addr}
6927 Refers to an object of type @var{type} stored at address @var{addr} in
6928 memory. @var{addr} may be any expression whose value is an integer or
6929 pointer (but parentheses are required around binary operators, just as in
6930 a cast). This construct is allowed regardless of what kind of data is
6931 normally supposed to reside at @var{addr}.
6934 @node Ambiguous Expressions
6935 @section Ambiguous Expressions
6936 @cindex ambiguous expressions
6938 Expressions can sometimes contain some ambiguous elements. For instance,
6939 some programming languages (notably Ada, C@t{++} and Objective-C) permit
6940 a single function name to be defined several times, for application in
6941 different contexts. This is called @dfn{overloading}. Another example
6942 involving Ada is generics. A @dfn{generic package} is similar to C@t{++}
6943 templates and is typically instantiated several times, resulting in
6944 the same function name being defined in different contexts.
6946 In some cases and depending on the language, it is possible to adjust
6947 the expression to remove the ambiguity. For instance in C@t{++}, you
6948 can specify the signature of the function you want to break on, as in
6949 @kbd{break @var{function}(@var{types})}. In Ada, using the fully
6950 qualified name of your function often makes the expression unambiguous
6953 When an ambiguity that needs to be resolved is detected, the debugger
6954 has the capability to display a menu of numbered choices for each
6955 possibility, and then waits for the selection with the prompt @samp{>}.
6956 The first option is always @samp{[0] cancel}, and typing @kbd{0 @key{RET}}
6957 aborts the current command. If the command in which the expression was
6958 used allows more than one choice to be selected, the next option in the
6959 menu is @samp{[1] all}, and typing @kbd{1 @key{RET}} selects all possible
6962 For example, the following session excerpt shows an attempt to set a
6963 breakpoint at the overloaded symbol @code{String::after}.
6964 We choose three particular definitions of that function name:
6966 @c FIXME! This is likely to change to show arg type lists, at least
6969 (@value{GDBP}) b String::after
6972 [2] file:String.cc; line number:867
6973 [3] file:String.cc; line number:860
6974 [4] file:String.cc; line number:875
6975 [5] file:String.cc; line number:853
6976 [6] file:String.cc; line number:846
6977 [7] file:String.cc; line number:735
6979 Breakpoint 1 at 0xb26c: file String.cc, line 867.
6980 Breakpoint 2 at 0xb344: file String.cc, line 875.
6981 Breakpoint 3 at 0xafcc: file String.cc, line 846.
6982 Multiple breakpoints were set.
6983 Use the "delete" command to delete unwanted
6990 @kindex set multiple-symbols
6991 @item set multiple-symbols @var{mode}
6992 @cindex multiple-symbols menu
6994 This option allows you to adjust the debugger behavior when an expression
6997 By default, @var{mode} is set to @code{all}. If the command with which
6998 the expression is used allows more than one choice, then @value{GDBN}
6999 automatically selects all possible choices. For instance, inserting
7000 a breakpoint on a function using an ambiguous name results in a breakpoint
7001 inserted on each possible match. However, if a unique choice must be made,
7002 then @value{GDBN} uses the menu to help you disambiguate the expression.
7003 For instance, printing the address of an overloaded function will result
7004 in the use of the menu.
7006 When @var{mode} is set to @code{ask}, the debugger always uses the menu
7007 when an ambiguity is detected.
7009 Finally, when @var{mode} is set to @code{cancel}, the debugger reports
7010 an error due to the ambiguity and the command is aborted.
7012 @kindex show multiple-symbols
7013 @item show multiple-symbols
7014 Show the current value of the @code{multiple-symbols} setting.
7018 @section Program Variables
7020 The most common kind of expression to use is the name of a variable
7023 Variables in expressions are understood in the selected stack frame
7024 (@pxref{Selection, ,Selecting a Frame}); they must be either:
7028 global (or file-static)
7035 visible according to the scope rules of the
7036 programming language from the point of execution in that frame
7039 @noindent This means that in the function
7054 you can examine and use the variable @code{a} whenever your program is
7055 executing within the function @code{foo}, but you can only use or
7056 examine the variable @code{b} while your program is executing inside
7057 the block where @code{b} is declared.
7059 @cindex variable name conflict
7060 There is an exception: you can refer to a variable or function whose
7061 scope is a single source file even if the current execution point is not
7062 in this file. But it is possible to have more than one such variable or
7063 function with the same name (in different source files). If that
7064 happens, referring to that name has unpredictable effects. If you wish,
7065 you can specify a static variable in a particular function or file,
7066 using the colon-colon (@code{::}) notation:
7068 @cindex colon-colon, context for variables/functions
7070 @c info cannot cope with a :: index entry, but why deprive hard copy readers?
7071 @cindex @code{::}, context for variables/functions
7074 @var{file}::@var{variable}
7075 @var{function}::@var{variable}
7079 Here @var{file} or @var{function} is the name of the context for the
7080 static @var{variable}. In the case of file names, you can use quotes to
7081 make sure @value{GDBN} parses the file name as a single word---for example,
7082 to print a global value of @code{x} defined in @file{f2.c}:
7085 (@value{GDBP}) p 'f2.c'::x
7088 @cindex C@t{++} scope resolution
7089 This use of @samp{::} is very rarely in conflict with the very similar
7090 use of the same notation in C@t{++}. @value{GDBN} also supports use of the C@t{++}
7091 scope resolution operator in @value{GDBN} expressions.
7092 @c FIXME: Um, so what happens in one of those rare cases where it's in
7095 @cindex wrong values
7096 @cindex variable values, wrong
7097 @cindex function entry/exit, wrong values of variables
7098 @cindex optimized code, wrong values of variables
7100 @emph{Warning:} Occasionally, a local variable may appear to have the
7101 wrong value at certain points in a function---just after entry to a new
7102 scope, and just before exit.
7104 You may see this problem when you are stepping by machine instructions.
7105 This is because, on most machines, it takes more than one instruction to
7106 set up a stack frame (including local variable definitions); if you are
7107 stepping by machine instructions, variables may appear to have the wrong
7108 values until the stack frame is completely built. On exit, it usually
7109 also takes more than one machine instruction to destroy a stack frame;
7110 after you begin stepping through that group of instructions, local
7111 variable definitions may be gone.
7113 This may also happen when the compiler does significant optimizations.
7114 To be sure of always seeing accurate values, turn off all optimization
7117 @cindex ``No symbol "foo" in current context''
7118 Another possible effect of compiler optimizations is to optimize
7119 unused variables out of existence, or assign variables to registers (as
7120 opposed to memory addresses). Depending on the support for such cases
7121 offered by the debug info format used by the compiler, @value{GDBN}
7122 might not be able to display values for such local variables. If that
7123 happens, @value{GDBN} will print a message like this:
7126 No symbol "foo" in current context.
7129 To solve such problems, either recompile without optimizations, or use a
7130 different debug info format, if the compiler supports several such
7131 formats. For example, @value{NGCC}, the @sc{gnu} C/C@t{++} compiler,
7132 usually supports the @option{-gstabs+} option. @option{-gstabs+}
7133 produces debug info in a format that is superior to formats such as
7134 COFF. You may be able to use DWARF 2 (@option{-gdwarf-2}), which is also
7135 an effective form for debug info. @xref{Debugging Options,,Options
7136 for Debugging Your Program or GCC, gcc.info, Using the @sc{gnu}
7137 Compiler Collection (GCC)}.
7138 @xref{C, ,C and C@t{++}}, for more information about debug info formats
7139 that are best suited to C@t{++} programs.
7141 If you ask to print an object whose contents are unknown to
7142 @value{GDBN}, e.g., because its data type is not completely specified
7143 by the debug information, @value{GDBN} will say @samp{<incomplete
7144 type>}. @xref{Symbols, incomplete type}, for more about this.
7146 Strings are identified as arrays of @code{char} values without specified
7147 signedness. Arrays of either @code{signed char} or @code{unsigned char} get
7148 printed as arrays of 1 byte sized integers. @code{-fsigned-char} or
7149 @code{-funsigned-char} @value{NGCC} options have no effect as @value{GDBN}
7150 defines literal string type @code{"char"} as @code{char} without a sign.
7155 signed char var1[] = "A";
7158 You get during debugging
7163 $2 = @{65 'A', 0 '\0'@}
7167 @section Artificial Arrays
7169 @cindex artificial array
7171 @kindex @@@r{, referencing memory as an array}
7172 It is often useful to print out several successive objects of the
7173 same type in memory; a section of an array, or an array of
7174 dynamically determined size for which only a pointer exists in the
7177 You can do this by referring to a contiguous span of memory as an
7178 @dfn{artificial array}, using the binary operator @samp{@@}. The left
7179 operand of @samp{@@} should be the first element of the desired array
7180 and be an individual object. The right operand should be the desired length
7181 of the array. The result is an array value whose elements are all of
7182 the type of the left argument. The first element is actually the left
7183 argument; the second element comes from bytes of memory immediately
7184 following those that hold the first element, and so on. Here is an
7185 example. If a program says
7188 int *array = (int *) malloc (len * sizeof (int));
7192 you can print the contents of @code{array} with
7198 The left operand of @samp{@@} must reside in memory. Array values made
7199 with @samp{@@} in this way behave just like other arrays in terms of
7200 subscripting, and are coerced to pointers when used in expressions.
7201 Artificial arrays most often appear in expressions via the value history
7202 (@pxref{Value History, ,Value History}), after printing one out.
7204 Another way to create an artificial array is to use a cast.
7205 This re-interprets a value as if it were an array.
7206 The value need not be in memory:
7208 (@value{GDBP}) p/x (short[2])0x12345678
7209 $1 = @{0x1234, 0x5678@}
7212 As a convenience, if you leave the array length out (as in
7213 @samp{(@var{type}[])@var{value}}) @value{GDBN} calculates the size to fill
7214 the value (as @samp{sizeof(@var{value})/sizeof(@var{type})}:
7216 (@value{GDBP}) p/x (short[])0x12345678
7217 $2 = @{0x1234, 0x5678@}
7220 Sometimes the artificial array mechanism is not quite enough; in
7221 moderately complex data structures, the elements of interest may not
7222 actually be adjacent---for example, if you are interested in the values
7223 of pointers in an array. One useful work-around in this situation is
7224 to use a convenience variable (@pxref{Convenience Vars, ,Convenience
7225 Variables}) as a counter in an expression that prints the first
7226 interesting value, and then repeat that expression via @key{RET}. For
7227 instance, suppose you have an array @code{dtab} of pointers to
7228 structures, and you are interested in the values of a field @code{fv}
7229 in each structure. Here is an example of what you might type:
7239 @node Output Formats
7240 @section Output Formats
7242 @cindex formatted output
7243 @cindex output formats
7244 By default, @value{GDBN} prints a value according to its data type. Sometimes
7245 this is not what you want. For example, you might want to print a number
7246 in hex, or a pointer in decimal. Or you might want to view data in memory
7247 at a certain address as a character string or as an instruction. To do
7248 these things, specify an @dfn{output format} when you print a value.
7250 The simplest use of output formats is to say how to print a value
7251 already computed. This is done by starting the arguments of the
7252 @code{print} command with a slash and a format letter. The format
7253 letters supported are:
7257 Regard the bits of the value as an integer, and print the integer in
7261 Print as integer in signed decimal.
7264 Print as integer in unsigned decimal.
7267 Print as integer in octal.
7270 Print as integer in binary. The letter @samp{t} stands for ``two''.
7271 @footnote{@samp{b} cannot be used because these format letters are also
7272 used with the @code{x} command, where @samp{b} stands for ``byte'';
7273 see @ref{Memory,,Examining Memory}.}
7276 @cindex unknown address, locating
7277 @cindex locate address
7278 Print as an address, both absolute in hexadecimal and as an offset from
7279 the nearest preceding symbol. You can use this format used to discover
7280 where (in what function) an unknown address is located:
7283 (@value{GDBP}) p/a 0x54320
7284 $3 = 0x54320 <_initialize_vx+396>
7288 The command @code{info symbol 0x54320} yields similar results.
7289 @xref{Symbols, info symbol}.
7292 Regard as an integer and print it as a character constant. This
7293 prints both the numerical value and its character representation. The
7294 character representation is replaced with the octal escape @samp{\nnn}
7295 for characters outside the 7-bit @sc{ascii} range.
7297 Without this format, @value{GDBN} displays @code{char},
7298 @w{@code{unsigned char}}, and @w{@code{signed char}} data as character
7299 constants. Single-byte members of vectors are displayed as integer
7303 Regard the bits of the value as a floating point number and print
7304 using typical floating point syntax.
7307 @cindex printing strings
7308 @cindex printing byte arrays
7309 Regard as a string, if possible. With this format, pointers to single-byte
7310 data are displayed as null-terminated strings and arrays of single-byte data
7311 are displayed as fixed-length strings. Other values are displayed in their
7314 Without this format, @value{GDBN} displays pointers to and arrays of
7315 @code{char}, @w{@code{unsigned char}}, and @w{@code{signed char}} as
7316 strings. Single-byte members of a vector are displayed as an integer
7320 @cindex raw printing
7321 Print using the @samp{raw} formatting. By default, @value{GDBN} will
7322 use a Python-based pretty-printer, if one is available (@pxref{Pretty
7323 Printing}). This typically results in a higher-level display of the
7324 value's contents. The @samp{r} format bypasses any Python
7325 pretty-printer which might exist.
7328 For example, to print the program counter in hex (@pxref{Registers}), type
7335 Note that no space is required before the slash; this is because command
7336 names in @value{GDBN} cannot contain a slash.
7338 To reprint the last value in the value history with a different format,
7339 you can use the @code{print} command with just a format and no
7340 expression. For example, @samp{p/x} reprints the last value in hex.
7343 @section Examining Memory
7345 You can use the command @code{x} (for ``examine'') to examine memory in
7346 any of several formats, independently of your program's data types.
7348 @cindex examining memory
7350 @kindex x @r{(examine memory)}
7351 @item x/@var{nfu} @var{addr}
7354 Use the @code{x} command to examine memory.
7357 @var{n}, @var{f}, and @var{u} are all optional parameters that specify how
7358 much memory to display and how to format it; @var{addr} is an
7359 expression giving the address where you want to start displaying memory.
7360 If you use defaults for @var{nfu}, you need not type the slash @samp{/}.
7361 Several commands set convenient defaults for @var{addr}.
7364 @item @var{n}, the repeat count
7365 The repeat count is a decimal integer; the default is 1. It specifies
7366 how much memory (counting by units @var{u}) to display.
7367 @c This really is **decimal**; unaffected by 'set radix' as of GDB
7370 @item @var{f}, the display format
7371 The display format is one of the formats used by @code{print}
7372 (@samp{x}, @samp{d}, @samp{u}, @samp{o}, @samp{t}, @samp{a}, @samp{c},
7373 @samp{f}, @samp{s}), and in addition @samp{i} (for machine instructions).
7374 The default is @samp{x} (hexadecimal) initially. The default changes
7375 each time you use either @code{x} or @code{print}.
7377 @item @var{u}, the unit size
7378 The unit size is any of
7384 Halfwords (two bytes).
7386 Words (four bytes). This is the initial default.
7388 Giant words (eight bytes).
7391 Each time you specify a unit size with @code{x}, that size becomes the
7392 default unit the next time you use @code{x}. For the @samp{i} format,
7393 the unit size is ignored and is normally not written. For the @samp{s} format,
7394 the unit size defaults to @samp{b}, unless it is explicitly given.
7395 Use @kbd{x /hs} to display 16-bit char strings and @kbd{x /ws} to display
7396 32-bit strings. The next use of @kbd{x /s} will again display 8-bit strings.
7397 Note that the results depend on the programming language of the
7398 current compilation unit. If the language is C, the @samp{s}
7399 modifier will use the UTF-16 encoding while @samp{w} will use
7400 UTF-32. The encoding is set by the programming language and cannot
7403 @item @var{addr}, starting display address
7404 @var{addr} is the address where you want @value{GDBN} to begin displaying
7405 memory. The expression need not have a pointer value (though it may);
7406 it is always interpreted as an integer address of a byte of memory.
7407 @xref{Expressions, ,Expressions}, for more information on expressions. The default for
7408 @var{addr} is usually just after the last address examined---but several
7409 other commands also set the default address: @code{info breakpoints} (to
7410 the address of the last breakpoint listed), @code{info line} (to the
7411 starting address of a line), and @code{print} (if you use it to display
7412 a value from memory).
7415 For example, @samp{x/3uh 0x54320} is a request to display three halfwords
7416 (@code{h}) of memory, formatted as unsigned decimal integers (@samp{u}),
7417 starting at address @code{0x54320}. @samp{x/4xw $sp} prints the four
7418 words (@samp{w}) of memory above the stack pointer (here, @samp{$sp};
7419 @pxref{Registers, ,Registers}) in hexadecimal (@samp{x}).
7421 Since the letters indicating unit sizes are all distinct from the
7422 letters specifying output formats, you do not have to remember whether
7423 unit size or format comes first; either order works. The output
7424 specifications @samp{4xw} and @samp{4wx} mean exactly the same thing.
7425 (However, the count @var{n} must come first; @samp{wx4} does not work.)
7427 Even though the unit size @var{u} is ignored for the formats @samp{s}
7428 and @samp{i}, you might still want to use a count @var{n}; for example,
7429 @samp{3i} specifies that you want to see three machine instructions,
7430 including any operands. For convenience, especially when used with
7431 the @code{display} command, the @samp{i} format also prints branch delay
7432 slot instructions, if any, beyond the count specified, which immediately
7433 follow the last instruction that is within the count. The command
7434 @code{disassemble} gives an alternative way of inspecting machine
7435 instructions; see @ref{Machine Code,,Source and Machine Code}.
7437 All the defaults for the arguments to @code{x} are designed to make it
7438 easy to continue scanning memory with minimal specifications each time
7439 you use @code{x}. For example, after you have inspected three machine
7440 instructions with @samp{x/3i @var{addr}}, you can inspect the next seven
7441 with just @samp{x/7}. If you use @key{RET} to repeat the @code{x} command,
7442 the repeat count @var{n} is used again; the other arguments default as
7443 for successive uses of @code{x}.
7445 When examining machine instructions, the instruction at current program
7446 counter is shown with a @code{=>} marker. For example:
7449 (@value{GDBP}) x/5i $pc-6
7450 0x804837f <main+11>: mov %esp,%ebp
7451 0x8048381 <main+13>: push %ecx
7452 0x8048382 <main+14>: sub $0x4,%esp
7453 => 0x8048385 <main+17>: movl $0x8048460,(%esp)
7454 0x804838c <main+24>: call 0x80482d4 <puts@@plt>
7457 @cindex @code{$_}, @code{$__}, and value history
7458 The addresses and contents printed by the @code{x} command are not saved
7459 in the value history because there is often too much of them and they
7460 would get in the way. Instead, @value{GDBN} makes these values available for
7461 subsequent use in expressions as values of the convenience variables
7462 @code{$_} and @code{$__}. After an @code{x} command, the last address
7463 examined is available for use in expressions in the convenience variable
7464 @code{$_}. The contents of that address, as examined, are available in
7465 the convenience variable @code{$__}.
7467 If the @code{x} command has a repeat count, the address and contents saved
7468 are from the last memory unit printed; this is not the same as the last
7469 address printed if several units were printed on the last line of output.
7471 @cindex remote memory comparison
7472 @cindex verify remote memory image
7473 When you are debugging a program running on a remote target machine
7474 (@pxref{Remote Debugging}), you may wish to verify the program's image in the
7475 remote machine's memory against the executable file you downloaded to
7476 the target. The @code{compare-sections} command is provided for such
7480 @kindex compare-sections
7481 @item compare-sections @r{[}@var{section-name}@r{]}
7482 Compare the data of a loadable section @var{section-name} in the
7483 executable file of the program being debugged with the same section in
7484 the remote machine's memory, and report any mismatches. With no
7485 arguments, compares all loadable sections. This command's
7486 availability depends on the target's support for the @code{"qCRC"}
7491 @section Automatic Display
7492 @cindex automatic display
7493 @cindex display of expressions
7495 If you find that you want to print the value of an expression frequently
7496 (to see how it changes), you might want to add it to the @dfn{automatic
7497 display list} so that @value{GDBN} prints its value each time your program stops.
7498 Each expression added to the list is given a number to identify it;
7499 to remove an expression from the list, you specify that number.
7500 The automatic display looks like this:
7504 3: bar[5] = (struct hack *) 0x3804
7508 This display shows item numbers, expressions and their current values. As with
7509 displays you request manually using @code{x} or @code{print}, you can
7510 specify the output format you prefer; in fact, @code{display} decides
7511 whether to use @code{print} or @code{x} depending your format
7512 specification---it uses @code{x} if you specify either the @samp{i}
7513 or @samp{s} format, or a unit size; otherwise it uses @code{print}.
7517 @item display @var{expr}
7518 Add the expression @var{expr} to the list of expressions to display
7519 each time your program stops. @xref{Expressions, ,Expressions}.
7521 @code{display} does not repeat if you press @key{RET} again after using it.
7523 @item display/@var{fmt} @var{expr}
7524 For @var{fmt} specifying only a display format and not a size or
7525 count, add the expression @var{expr} to the auto-display list but
7526 arrange to display it each time in the specified format @var{fmt}.
7527 @xref{Output Formats,,Output Formats}.
7529 @item display/@var{fmt} @var{addr}
7530 For @var{fmt} @samp{i} or @samp{s}, or including a unit-size or a
7531 number of units, add the expression @var{addr} as a memory address to
7532 be examined each time your program stops. Examining means in effect
7533 doing @samp{x/@var{fmt} @var{addr}}. @xref{Memory, ,Examining Memory}.
7536 For example, @samp{display/i $pc} can be helpful, to see the machine
7537 instruction about to be executed each time execution stops (@samp{$pc}
7538 is a common name for the program counter; @pxref{Registers, ,Registers}).
7541 @kindex delete display
7543 @item undisplay @var{dnums}@dots{}
7544 @itemx delete display @var{dnums}@dots{}
7545 Remove item numbers @var{dnums} from the list of expressions to display.
7547 @code{undisplay} does not repeat if you press @key{RET} after using it.
7548 (Otherwise you would just get the error @samp{No display number @dots{}}.)
7550 @kindex disable display
7551 @item disable display @var{dnums}@dots{}
7552 Disable the display of item numbers @var{dnums}. A disabled display
7553 item is not printed automatically, but is not forgotten. It may be
7554 enabled again later.
7556 @kindex enable display
7557 @item enable display @var{dnums}@dots{}
7558 Enable display of item numbers @var{dnums}. It becomes effective once
7559 again in auto display of its expression, until you specify otherwise.
7562 Display the current values of the expressions on the list, just as is
7563 done when your program stops.
7565 @kindex info display
7567 Print the list of expressions previously set up to display
7568 automatically, each one with its item number, but without showing the
7569 values. This includes disabled expressions, which are marked as such.
7570 It also includes expressions which would not be displayed right now
7571 because they refer to automatic variables not currently available.
7574 @cindex display disabled out of scope
7575 If a display expression refers to local variables, then it does not make
7576 sense outside the lexical context for which it was set up. Such an
7577 expression is disabled when execution enters a context where one of its
7578 variables is not defined. For example, if you give the command
7579 @code{display last_char} while inside a function with an argument
7580 @code{last_char}, @value{GDBN} displays this argument while your program
7581 continues to stop inside that function. When it stops elsewhere---where
7582 there is no variable @code{last_char}---the display is disabled
7583 automatically. The next time your program stops where @code{last_char}
7584 is meaningful, you can enable the display expression once again.
7586 @node Print Settings
7587 @section Print Settings
7589 @cindex format options
7590 @cindex print settings
7591 @value{GDBN} provides the following ways to control how arrays, structures,
7592 and symbols are printed.
7595 These settings are useful for debugging programs in any language:
7599 @item set print address
7600 @itemx set print address on
7601 @cindex print/don't print memory addresses
7602 @value{GDBN} prints memory addresses showing the location of stack
7603 traces, structure values, pointer values, breakpoints, and so forth,
7604 even when it also displays the contents of those addresses. The default
7605 is @code{on}. For example, this is what a stack frame display looks like with
7606 @code{set print address on}:
7611 #0 set_quotes (lq=0x34c78 "<<", rq=0x34c88 ">>")
7613 530 if (lquote != def_lquote)
7617 @item set print address off
7618 Do not print addresses when displaying their contents. For example,
7619 this is the same stack frame displayed with @code{set print address off}:
7623 (@value{GDBP}) set print addr off
7625 #0 set_quotes (lq="<<", rq=">>") at input.c:530
7626 530 if (lquote != def_lquote)
7630 You can use @samp{set print address off} to eliminate all machine
7631 dependent displays from the @value{GDBN} interface. For example, with
7632 @code{print address off}, you should get the same text for backtraces on
7633 all machines---whether or not they involve pointer arguments.
7636 @item show print address
7637 Show whether or not addresses are to be printed.
7640 When @value{GDBN} prints a symbolic address, it normally prints the
7641 closest earlier symbol plus an offset. If that symbol does not uniquely
7642 identify the address (for example, it is a name whose scope is a single
7643 source file), you may need to clarify. One way to do this is with
7644 @code{info line}, for example @samp{info line *0x4537}. Alternately,
7645 you can set @value{GDBN} to print the source file and line number when
7646 it prints a symbolic address:
7649 @item set print symbol-filename on
7650 @cindex source file and line of a symbol
7651 @cindex symbol, source file and line
7652 Tell @value{GDBN} to print the source file name and line number of a
7653 symbol in the symbolic form of an address.
7655 @item set print symbol-filename off
7656 Do not print source file name and line number of a symbol. This is the
7659 @item show print symbol-filename
7660 Show whether or not @value{GDBN} will print the source file name and
7661 line number of a symbol in the symbolic form of an address.
7664 Another situation where it is helpful to show symbol filenames and line
7665 numbers is when disassembling code; @value{GDBN} shows you the line
7666 number and source file that corresponds to each instruction.
7668 Also, you may wish to see the symbolic form only if the address being
7669 printed is reasonably close to the closest earlier symbol:
7672 @item set print max-symbolic-offset @var{max-offset}
7673 @cindex maximum value for offset of closest symbol
7674 Tell @value{GDBN} to only display the symbolic form of an address if the
7675 offset between the closest earlier symbol and the address is less than
7676 @var{max-offset}. The default is 0, which tells @value{GDBN}
7677 to always print the symbolic form of an address if any symbol precedes it.
7679 @item show print max-symbolic-offset
7680 Ask how large the maximum offset is that @value{GDBN} prints in a
7684 @cindex wild pointer, interpreting
7685 @cindex pointer, finding referent
7686 If you have a pointer and you are not sure where it points, try
7687 @samp{set print symbol-filename on}. Then you can determine the name
7688 and source file location of the variable where it points, using
7689 @samp{p/a @var{pointer}}. This interprets the address in symbolic form.
7690 For example, here @value{GDBN} shows that a variable @code{ptt} points
7691 at another variable @code{t}, defined in @file{hi2.c}:
7694 (@value{GDBP}) set print symbol-filename on
7695 (@value{GDBP}) p/a ptt
7696 $4 = 0xe008 <t in hi2.c>
7700 @emph{Warning:} For pointers that point to a local variable, @samp{p/a}
7701 does not show the symbol name and filename of the referent, even with
7702 the appropriate @code{set print} options turned on.
7705 Other settings control how different kinds of objects are printed:
7708 @item set print array
7709 @itemx set print array on
7710 @cindex pretty print arrays
7711 Pretty print arrays. This format is more convenient to read,
7712 but uses more space. The default is off.
7714 @item set print array off
7715 Return to compressed format for arrays.
7717 @item show print array
7718 Show whether compressed or pretty format is selected for displaying
7721 @cindex print array indexes
7722 @item set print array-indexes
7723 @itemx set print array-indexes on
7724 Print the index of each element when displaying arrays. May be more
7725 convenient to locate a given element in the array or quickly find the
7726 index of a given element in that printed array. The default is off.
7728 @item set print array-indexes off
7729 Stop printing element indexes when displaying arrays.
7731 @item show print array-indexes
7732 Show whether the index of each element is printed when displaying
7735 @item set print elements @var{number-of-elements}
7736 @cindex number of array elements to print
7737 @cindex limit on number of printed array elements
7738 Set a limit on how many elements of an array @value{GDBN} will print.
7739 If @value{GDBN} is printing a large array, it stops printing after it has
7740 printed the number of elements set by the @code{set print elements} command.
7741 This limit also applies to the display of strings.
7742 When @value{GDBN} starts, this limit is set to 200.
7743 Setting @var{number-of-elements} to zero means that the printing is unlimited.
7745 @item show print elements
7746 Display the number of elements of a large array that @value{GDBN} will print.
7747 If the number is 0, then the printing is unlimited.
7749 @item set print frame-arguments @var{value}
7750 @kindex set print frame-arguments
7751 @cindex printing frame argument values
7752 @cindex print all frame argument values
7753 @cindex print frame argument values for scalars only
7754 @cindex do not print frame argument values
7755 This command allows to control how the values of arguments are printed
7756 when the debugger prints a frame (@pxref{Frames}). The possible
7761 The values of all arguments are printed.
7764 Print the value of an argument only if it is a scalar. The value of more
7765 complex arguments such as arrays, structures, unions, etc, is replaced
7766 by @code{@dots{}}. This is the default. Here is an example where
7767 only scalar arguments are shown:
7770 #1 0x08048361 in call_me (i=3, s=@dots{}, ss=0xbf8d508c, u=@dots{}, e=green)
7775 None of the argument values are printed. Instead, the value of each argument
7776 is replaced by @code{@dots{}}. In this case, the example above now becomes:
7779 #1 0x08048361 in call_me (i=@dots{}, s=@dots{}, ss=@dots{}, u=@dots{}, e=@dots{})
7784 By default, only scalar arguments are printed. This command can be used
7785 to configure the debugger to print the value of all arguments, regardless
7786 of their type. However, it is often advantageous to not print the value
7787 of more complex parameters. For instance, it reduces the amount of
7788 information printed in each frame, making the backtrace more readable.
7789 Also, it improves performance when displaying Ada frames, because
7790 the computation of large arguments can sometimes be CPU-intensive,
7791 especially in large applications. Setting @code{print frame-arguments}
7792 to @code{scalars} (the default) or @code{none} avoids this computation,
7793 thus speeding up the display of each Ada frame.
7795 @item show print frame-arguments
7796 Show how the value of arguments should be displayed when printing a frame.
7798 @item set print repeats
7799 @cindex repeated array elements
7800 Set the threshold for suppressing display of repeated array
7801 elements. When the number of consecutive identical elements of an
7802 array exceeds the threshold, @value{GDBN} prints the string
7803 @code{"<repeats @var{n} times>"}, where @var{n} is the number of
7804 identical repetitions, instead of displaying the identical elements
7805 themselves. Setting the threshold to zero will cause all elements to
7806 be individually printed. The default threshold is 10.
7808 @item show print repeats
7809 Display the current threshold for printing repeated identical
7812 @item set print null-stop
7813 @cindex @sc{null} elements in arrays
7814 Cause @value{GDBN} to stop printing the characters of an array when the first
7815 @sc{null} is encountered. This is useful when large arrays actually
7816 contain only short strings.
7819 @item show print null-stop
7820 Show whether @value{GDBN} stops printing an array on the first
7821 @sc{null} character.
7823 @item set print pretty on
7824 @cindex print structures in indented form
7825 @cindex indentation in structure display
7826 Cause @value{GDBN} to print structures in an indented format with one member
7827 per line, like this:
7842 @item set print pretty off
7843 Cause @value{GDBN} to print structures in a compact format, like this:
7847 $1 = @{next = 0x0, flags = @{sweet = 1, sour = 1@}, \
7848 meat = 0x54 "Pork"@}
7853 This is the default format.
7855 @item show print pretty
7856 Show which format @value{GDBN} is using to print structures.
7858 @item set print sevenbit-strings on
7859 @cindex eight-bit characters in strings
7860 @cindex octal escapes in strings
7861 Print using only seven-bit characters; if this option is set,
7862 @value{GDBN} displays any eight-bit characters (in strings or
7863 character values) using the notation @code{\}@var{nnn}. This setting is
7864 best if you are working in English (@sc{ascii}) and you use the
7865 high-order bit of characters as a marker or ``meta'' bit.
7867 @item set print sevenbit-strings off
7868 Print full eight-bit characters. This allows the use of more
7869 international character sets, and is the default.
7871 @item show print sevenbit-strings
7872 Show whether or not @value{GDBN} is printing only seven-bit characters.
7874 @item set print union on
7875 @cindex unions in structures, printing
7876 Tell @value{GDBN} to print unions which are contained in structures
7877 and other unions. This is the default setting.
7879 @item set print union off
7880 Tell @value{GDBN} not to print unions which are contained in
7881 structures and other unions. @value{GDBN} will print @code{"@{...@}"}
7884 @item show print union
7885 Ask @value{GDBN} whether or not it will print unions which are contained in
7886 structures and other unions.
7888 For example, given the declarations
7891 typedef enum @{Tree, Bug@} Species;
7892 typedef enum @{Big_tree, Acorn, Seedling@} Tree_forms;
7893 typedef enum @{Caterpillar, Cocoon, Butterfly@}
7904 struct thing foo = @{Tree, @{Acorn@}@};
7908 with @code{set print union on} in effect @samp{p foo} would print
7911 $1 = @{it = Tree, form = @{tree = Acorn, bug = Cocoon@}@}
7915 and with @code{set print union off} in effect it would print
7918 $1 = @{it = Tree, form = @{...@}@}
7922 @code{set print union} affects programs written in C-like languages
7928 These settings are of interest when debugging C@t{++} programs:
7931 @cindex demangling C@t{++} names
7932 @item set print demangle
7933 @itemx set print demangle on
7934 Print C@t{++} names in their source form rather than in the encoded
7935 (``mangled'') form passed to the assembler and linker for type-safe
7936 linkage. The default is on.
7938 @item show print demangle
7939 Show whether C@t{++} names are printed in mangled or demangled form.
7941 @item set print asm-demangle
7942 @itemx set print asm-demangle on
7943 Print C@t{++} names in their source form rather than their mangled form, even
7944 in assembler code printouts such as instruction disassemblies.
7947 @item show print asm-demangle
7948 Show whether C@t{++} names in assembly listings are printed in mangled
7951 @cindex C@t{++} symbol decoding style
7952 @cindex symbol decoding style, C@t{++}
7953 @kindex set demangle-style
7954 @item set demangle-style @var{style}
7955 Choose among several encoding schemes used by different compilers to
7956 represent C@t{++} names. The choices for @var{style} are currently:
7960 Allow @value{GDBN} to choose a decoding style by inspecting your program.
7963 Decode based on the @sc{gnu} C@t{++} compiler (@code{g++}) encoding algorithm.
7964 This is the default.
7967 Decode based on the HP ANSI C@t{++} (@code{aCC}) encoding algorithm.
7970 Decode based on the Lucid C@t{++} compiler (@code{lcc}) encoding algorithm.
7973 Decode using the algorithm in the @cite{C@t{++} Annotated Reference Manual}.
7974 @strong{Warning:} this setting alone is not sufficient to allow
7975 debugging @code{cfront}-generated executables. @value{GDBN} would
7976 require further enhancement to permit that.
7979 If you omit @var{style}, you will see a list of possible formats.
7981 @item show demangle-style
7982 Display the encoding style currently in use for decoding C@t{++} symbols.
7984 @item set print object
7985 @itemx set print object on
7986 @cindex derived type of an object, printing
7987 @cindex display derived types
7988 When displaying a pointer to an object, identify the @emph{actual}
7989 (derived) type of the object rather than the @emph{declared} type, using
7990 the virtual function table.
7992 @item set print object off
7993 Display only the declared type of objects, without reference to the
7994 virtual function table. This is the default setting.
7996 @item show print object
7997 Show whether actual, or declared, object types are displayed.
7999 @item set print static-members
8000 @itemx set print static-members on
8001 @cindex static members of C@t{++} objects
8002 Print static members when displaying a C@t{++} object. The default is on.
8004 @item set print static-members off
8005 Do not print static members when displaying a C@t{++} object.
8007 @item show print static-members
8008 Show whether C@t{++} static members are printed or not.
8010 @item set print pascal_static-members
8011 @itemx set print pascal_static-members on
8012 @cindex static members of Pascal objects
8013 @cindex Pascal objects, static members display
8014 Print static members when displaying a Pascal object. The default is on.
8016 @item set print pascal_static-members off
8017 Do not print static members when displaying a Pascal object.
8019 @item show print pascal_static-members
8020 Show whether Pascal static members are printed or not.
8022 @c These don't work with HP ANSI C++ yet.
8023 @item set print vtbl
8024 @itemx set print vtbl on
8025 @cindex pretty print C@t{++} virtual function tables
8026 @cindex virtual functions (C@t{++}) display
8027 @cindex VTBL display
8028 Pretty print C@t{++} virtual function tables. The default is off.
8029 (The @code{vtbl} commands do not work on programs compiled with the HP
8030 ANSI C@t{++} compiler (@code{aCC}).)
8032 @item set print vtbl off
8033 Do not pretty print C@t{++} virtual function tables.
8035 @item show print vtbl
8036 Show whether C@t{++} virtual function tables are pretty printed, or not.
8039 @node Pretty Printing
8040 @section Pretty Printing
8042 @value{GDBN} provides a mechanism to allow pretty-printing of values using
8043 Python code. It greatly simplifies the display of complex objects. This
8044 mechanism works for both MI and the CLI.
8046 For example, here is how a C@t{++} @code{std::string} looks without a
8050 (@value{GDBP}) print s
8052 static npos = 4294967295,
8054 <std::allocator<char>> = @{
8055 <__gnu_cxx::new_allocator<char>> = @{
8056 <No data fields>@}, <No data fields>
8058 members of std::basic_string<char, std::char_traits<char>,
8059 std::allocator<char> >::_Alloc_hider:
8060 _M_p = 0x804a014 "abcd"
8065 With a pretty-printer for @code{std::string} only the contents are printed:
8068 (@value{GDBP}) print s
8072 For implementing pretty printers for new types you should read the Python API
8073 details (@pxref{Pretty Printing API}).
8076 @section Value History
8078 @cindex value history
8079 @cindex history of values printed by @value{GDBN}
8080 Values printed by the @code{print} command are saved in the @value{GDBN}
8081 @dfn{value history}. This allows you to refer to them in other expressions.
8082 Values are kept until the symbol table is re-read or discarded
8083 (for example with the @code{file} or @code{symbol-file} commands).
8084 When the symbol table changes, the value history is discarded,
8085 since the values may contain pointers back to the types defined in the
8090 @cindex history number
8091 The values printed are given @dfn{history numbers} by which you can
8092 refer to them. These are successive integers starting with one.
8093 @code{print} shows you the history number assigned to a value by
8094 printing @samp{$@var{num} = } before the value; here @var{num} is the
8097 To refer to any previous value, use @samp{$} followed by the value's
8098 history number. The way @code{print} labels its output is designed to
8099 remind you of this. Just @code{$} refers to the most recent value in
8100 the history, and @code{$$} refers to the value before that.
8101 @code{$$@var{n}} refers to the @var{n}th value from the end; @code{$$2}
8102 is the value just prior to @code{$$}, @code{$$1} is equivalent to
8103 @code{$$}, and @code{$$0} is equivalent to @code{$}.
8105 For example, suppose you have just printed a pointer to a structure and
8106 want to see the contents of the structure. It suffices to type
8112 If you have a chain of structures where the component @code{next} points
8113 to the next one, you can print the contents of the next one with this:
8120 You can print successive links in the chain by repeating this
8121 command---which you can do by just typing @key{RET}.
8123 Note that the history records values, not expressions. If the value of
8124 @code{x} is 4 and you type these commands:
8132 then the value recorded in the value history by the @code{print} command
8133 remains 4 even though the value of @code{x} has changed.
8138 Print the last ten values in the value history, with their item numbers.
8139 This is like @samp{p@ $$9} repeated ten times, except that @code{show
8140 values} does not change the history.
8142 @item show values @var{n}
8143 Print ten history values centered on history item number @var{n}.
8146 Print ten history values just after the values last printed. If no more
8147 values are available, @code{show values +} produces no display.
8150 Pressing @key{RET} to repeat @code{show values @var{n}} has exactly the
8151 same effect as @samp{show values +}.
8153 @node Convenience Vars
8154 @section Convenience Variables
8156 @cindex convenience variables
8157 @cindex user-defined variables
8158 @value{GDBN} provides @dfn{convenience variables} that you can use within
8159 @value{GDBN} to hold on to a value and refer to it later. These variables
8160 exist entirely within @value{GDBN}; they are not part of your program, and
8161 setting a convenience variable has no direct effect on further execution
8162 of your program. That is why you can use them freely.
8164 Convenience variables are prefixed with @samp{$}. Any name preceded by
8165 @samp{$} can be used for a convenience variable, unless it is one of
8166 the predefined machine-specific register names (@pxref{Registers, ,Registers}).
8167 (Value history references, in contrast, are @emph{numbers} preceded
8168 by @samp{$}. @xref{Value History, ,Value History}.)
8170 You can save a value in a convenience variable with an assignment
8171 expression, just as you would set a variable in your program.
8175 set $foo = *object_ptr
8179 would save in @code{$foo} the value contained in the object pointed to by
8182 Using a convenience variable for the first time creates it, but its
8183 value is @code{void} until you assign a new value. You can alter the
8184 value with another assignment at any time.
8186 Convenience variables have no fixed types. You can assign a convenience
8187 variable any type of value, including structures and arrays, even if
8188 that variable already has a value of a different type. The convenience
8189 variable, when used as an expression, has the type of its current value.
8192 @kindex show convenience
8193 @cindex show all user variables
8194 @item show convenience
8195 Print a list of convenience variables used so far, and their values.
8196 Abbreviated @code{show conv}.
8198 @kindex init-if-undefined
8199 @cindex convenience variables, initializing
8200 @item init-if-undefined $@var{variable} = @var{expression}
8201 Set a convenience variable if it has not already been set. This is useful
8202 for user-defined commands that keep some state. It is similar, in concept,
8203 to using local static variables with initializers in C (except that
8204 convenience variables are global). It can also be used to allow users to
8205 override default values used in a command script.
8207 If the variable is already defined then the expression is not evaluated so
8208 any side-effects do not occur.
8211 One of the ways to use a convenience variable is as a counter to be
8212 incremented or a pointer to be advanced. For example, to print
8213 a field from successive elements of an array of structures:
8217 print bar[$i++]->contents
8221 Repeat that command by typing @key{RET}.
8223 Some convenience variables are created automatically by @value{GDBN} and given
8224 values likely to be useful.
8227 @vindex $_@r{, convenience variable}
8229 The variable @code{$_} is automatically set by the @code{x} command to
8230 the last address examined (@pxref{Memory, ,Examining Memory}). Other
8231 commands which provide a default address for @code{x} to examine also
8232 set @code{$_} to that address; these commands include @code{info line}
8233 and @code{info breakpoint}. The type of @code{$_} is @code{void *}
8234 except when set by the @code{x} command, in which case it is a pointer
8235 to the type of @code{$__}.
8237 @vindex $__@r{, convenience variable}
8239 The variable @code{$__} is automatically set by the @code{x} command
8240 to the value found in the last address examined. Its type is chosen
8241 to match the format in which the data was printed.
8244 @vindex $_exitcode@r{, convenience variable}
8245 The variable @code{$_exitcode} is automatically set to the exit code when
8246 the program being debugged terminates.
8249 @vindex $_siginfo@r{, convenience variable}
8250 The variable @code{$_siginfo} contains extra signal information
8251 (@pxref{extra signal information}). Note that @code{$_siginfo}
8252 could be empty, if the application has not yet received any signals.
8253 For example, it will be empty before you execute the @code{run} command.
8256 @vindex $_tlb@r{, convenience variable}
8257 The variable @code{$_tlb} is automatically set when debugging
8258 applications running on MS-Windows in native mode or connected to
8259 gdbserver that supports the @code{qGetTIBAddr} request.
8260 @xref{General Query Packets}.
8261 This variable contains the address of the thread information block.
8265 On HP-UX systems, if you refer to a function or variable name that
8266 begins with a dollar sign, @value{GDBN} searches for a user or system
8267 name first, before it searches for a convenience variable.
8269 @cindex convenience functions
8270 @value{GDBN} also supplies some @dfn{convenience functions}. These
8271 have a syntax similar to convenience variables. A convenience
8272 function can be used in an expression just like an ordinary function;
8273 however, a convenience function is implemented internally to
8278 @kindex help function
8279 @cindex show all convenience functions
8280 Print a list of all convenience functions.
8287 You can refer to machine register contents, in expressions, as variables
8288 with names starting with @samp{$}. The names of registers are different
8289 for each machine; use @code{info registers} to see the names used on
8293 @kindex info registers
8294 @item info registers
8295 Print the names and values of all registers except floating-point
8296 and vector registers (in the selected stack frame).
8298 @kindex info all-registers
8299 @cindex floating point registers
8300 @item info all-registers
8301 Print the names and values of all registers, including floating-point
8302 and vector registers (in the selected stack frame).
8304 @item info registers @var{regname} @dots{}
8305 Print the @dfn{relativized} value of each specified register @var{regname}.
8306 As discussed in detail below, register values are normally relative to
8307 the selected stack frame. @var{regname} may be any register name valid on
8308 the machine you are using, with or without the initial @samp{$}.
8311 @cindex stack pointer register
8312 @cindex program counter register
8313 @cindex process status register
8314 @cindex frame pointer register
8315 @cindex standard registers
8316 @value{GDBN} has four ``standard'' register names that are available (in
8317 expressions) on most machines---whenever they do not conflict with an
8318 architecture's canonical mnemonics for registers. The register names
8319 @code{$pc} and @code{$sp} are used for the program counter register and
8320 the stack pointer. @code{$fp} is used for a register that contains a
8321 pointer to the current stack frame, and @code{$ps} is used for a
8322 register that contains the processor status. For example,
8323 you could print the program counter in hex with
8330 or print the instruction to be executed next with
8337 or add four to the stack pointer@footnote{This is a way of removing
8338 one word from the stack, on machines where stacks grow downward in
8339 memory (most machines, nowadays). This assumes that the innermost
8340 stack frame is selected; setting @code{$sp} is not allowed when other
8341 stack frames are selected. To pop entire frames off the stack,
8342 regardless of machine architecture, use @code{return};
8343 see @ref{Returning, ,Returning from a Function}.} with
8349 Whenever possible, these four standard register names are available on
8350 your machine even though the machine has different canonical mnemonics,
8351 so long as there is no conflict. The @code{info registers} command
8352 shows the canonical names. For example, on the SPARC, @code{info
8353 registers} displays the processor status register as @code{$psr} but you
8354 can also refer to it as @code{$ps}; and on x86-based machines @code{$ps}
8355 is an alias for the @sc{eflags} register.
8357 @value{GDBN} always considers the contents of an ordinary register as an
8358 integer when the register is examined in this way. Some machines have
8359 special registers which can hold nothing but floating point; these
8360 registers are considered to have floating point values. There is no way
8361 to refer to the contents of an ordinary register as floating point value
8362 (although you can @emph{print} it as a floating point value with
8363 @samp{print/f $@var{regname}}).
8365 Some registers have distinct ``raw'' and ``virtual'' data formats. This
8366 means that the data format in which the register contents are saved by
8367 the operating system is not the same one that your program normally
8368 sees. For example, the registers of the 68881 floating point
8369 coprocessor are always saved in ``extended'' (raw) format, but all C
8370 programs expect to work with ``double'' (virtual) format. In such
8371 cases, @value{GDBN} normally works with the virtual format only (the format
8372 that makes sense for your program), but the @code{info registers} command
8373 prints the data in both formats.
8375 @cindex SSE registers (x86)
8376 @cindex MMX registers (x86)
8377 Some machines have special registers whose contents can be interpreted
8378 in several different ways. For example, modern x86-based machines
8379 have SSE and MMX registers that can hold several values packed
8380 together in several different formats. @value{GDBN} refers to such
8381 registers in @code{struct} notation:
8384 (@value{GDBP}) print $xmm1
8386 v4_float = @{0, 3.43859137e-038, 1.54142831e-044, 1.821688e-044@},
8387 v2_double = @{9.92129282474342e-303, 2.7585945287983262e-313@},
8388 v16_int8 = "\000\000\000\000\3706;\001\v\000\000\000\r\000\000",
8389 v8_int16 = @{0, 0, 14072, 315, 11, 0, 13, 0@},
8390 v4_int32 = @{0, 20657912, 11, 13@},
8391 v2_int64 = @{88725056443645952, 55834574859@},
8392 uint128 = 0x0000000d0000000b013b36f800000000
8397 To set values of such registers, you need to tell @value{GDBN} which
8398 view of the register you wish to change, as if you were assigning
8399 value to a @code{struct} member:
8402 (@value{GDBP}) set $xmm1.uint128 = 0x000000000000000000000000FFFFFFFF
8405 Normally, register values are relative to the selected stack frame
8406 (@pxref{Selection, ,Selecting a Frame}). This means that you get the
8407 value that the register would contain if all stack frames farther in
8408 were exited and their saved registers restored. In order to see the
8409 true contents of hardware registers, you must select the innermost
8410 frame (with @samp{frame 0}).
8412 However, @value{GDBN} must deduce where registers are saved, from the machine
8413 code generated by your compiler. If some registers are not saved, or if
8414 @value{GDBN} is unable to locate the saved registers, the selected stack
8415 frame makes no difference.
8417 @node Floating Point Hardware
8418 @section Floating Point Hardware
8419 @cindex floating point
8421 Depending on the configuration, @value{GDBN} may be able to give
8422 you more information about the status of the floating point hardware.
8427 Display hardware-dependent information about the floating
8428 point unit. The exact contents and layout vary depending on the
8429 floating point chip. Currently, @samp{info float} is supported on
8430 the ARM and x86 machines.
8434 @section Vector Unit
8437 Depending on the configuration, @value{GDBN} may be able to give you
8438 more information about the status of the vector unit.
8443 Display information about the vector unit. The exact contents and
8444 layout vary depending on the hardware.
8447 @node OS Information
8448 @section Operating System Auxiliary Information
8449 @cindex OS information
8451 @value{GDBN} provides interfaces to useful OS facilities that can help
8452 you debug your program.
8454 @cindex @code{ptrace} system call
8455 @cindex @code{struct user} contents
8456 When @value{GDBN} runs on a @dfn{Posix system} (such as GNU or Unix
8457 machines), it interfaces with the inferior via the @code{ptrace}
8458 system call. The operating system creates a special sata structure,
8459 called @code{struct user}, for this interface. You can use the
8460 command @code{info udot} to display the contents of this data
8466 Display the contents of the @code{struct user} maintained by the OS
8467 kernel for the program being debugged. @value{GDBN} displays the
8468 contents of @code{struct user} as a list of hex numbers, similar to
8469 the @code{examine} command.
8472 @cindex auxiliary vector
8473 @cindex vector, auxiliary
8474 Some operating systems supply an @dfn{auxiliary vector} to programs at
8475 startup. This is akin to the arguments and environment that you
8476 specify for a program, but contains a system-dependent variety of
8477 binary values that tell system libraries important details about the
8478 hardware, operating system, and process. Each value's purpose is
8479 identified by an integer tag; the meanings are well-known but system-specific.
8480 Depending on the configuration and operating system facilities,
8481 @value{GDBN} may be able to show you this information. For remote
8482 targets, this functionality may further depend on the remote stub's
8483 support of the @samp{qXfer:auxv:read} packet, see
8484 @ref{qXfer auxiliary vector read}.
8489 Display the auxiliary vector of the inferior, which can be either a
8490 live process or a core dump file. @value{GDBN} prints each tag value
8491 numerically, and also shows names and text descriptions for recognized
8492 tags. Some values in the vector are numbers, some bit masks, and some
8493 pointers to strings or other data. @value{GDBN} displays each value in the
8494 most appropriate form for a recognized tag, and in hexadecimal for
8495 an unrecognized tag.
8498 On some targets, @value{GDBN} can access operating-system-specific information
8499 and display it to user, without interpretation. For remote targets,
8500 this functionality depends on the remote stub's support of the
8501 @samp{qXfer:osdata:read} packet, see @ref{qXfer osdata read}.
8504 @kindex info os processes
8505 @item info os processes
8506 Display the list of processes on the target. For each process,
8507 @value{GDBN} prints the process identifier, the name of the user, and
8508 the command corresponding to the process.
8511 @node Memory Region Attributes
8512 @section Memory Region Attributes
8513 @cindex memory region attributes
8515 @dfn{Memory region attributes} allow you to describe special handling
8516 required by regions of your target's memory. @value{GDBN} uses
8517 attributes to determine whether to allow certain types of memory
8518 accesses; whether to use specific width accesses; and whether to cache
8519 target memory. By default the description of memory regions is
8520 fetched from the target (if the current target supports this), but the
8521 user can override the fetched regions.
8523 Defined memory regions can be individually enabled and disabled. When a
8524 memory region is disabled, @value{GDBN} uses the default attributes when
8525 accessing memory in that region. Similarly, if no memory regions have
8526 been defined, @value{GDBN} uses the default attributes when accessing
8529 When a memory region is defined, it is given a number to identify it;
8530 to enable, disable, or remove a memory region, you specify that number.
8534 @item mem @var{lower} @var{upper} @var{attributes}@dots{}
8535 Define a memory region bounded by @var{lower} and @var{upper} with
8536 attributes @var{attributes}@dots{}, and add it to the list of regions
8537 monitored by @value{GDBN}. Note that @var{upper} == 0 is a special
8538 case: it is treated as the target's maximum memory address.
8539 (0xffff on 16 bit targets, 0xffffffff on 32 bit targets, etc.)
8542 Discard any user changes to the memory regions and use target-supplied
8543 regions, if available, or no regions if the target does not support.
8546 @item delete mem @var{nums}@dots{}
8547 Remove memory regions @var{nums}@dots{} from the list of regions
8548 monitored by @value{GDBN}.
8551 @item disable mem @var{nums}@dots{}
8552 Disable monitoring of memory regions @var{nums}@dots{}.
8553 A disabled memory region is not forgotten.
8554 It may be enabled again later.
8557 @item enable mem @var{nums}@dots{}
8558 Enable monitoring of memory regions @var{nums}@dots{}.
8562 Print a table of all defined memory regions, with the following columns
8566 @item Memory Region Number
8567 @item Enabled or Disabled.
8568 Enabled memory regions are marked with @samp{y}.
8569 Disabled memory regions are marked with @samp{n}.
8572 The address defining the inclusive lower bound of the memory region.
8575 The address defining the exclusive upper bound of the memory region.
8578 The list of attributes set for this memory region.
8583 @subsection Attributes
8585 @subsubsection Memory Access Mode
8586 The access mode attributes set whether @value{GDBN} may make read or
8587 write accesses to a memory region.
8589 While these attributes prevent @value{GDBN} from performing invalid
8590 memory accesses, they do nothing to prevent the target system, I/O DMA,
8591 etc.@: from accessing memory.
8595 Memory is read only.
8597 Memory is write only.
8599 Memory is read/write. This is the default.
8602 @subsubsection Memory Access Size
8603 The access size attribute tells @value{GDBN} to use specific sized
8604 accesses in the memory region. Often memory mapped device registers
8605 require specific sized accesses. If no access size attribute is
8606 specified, @value{GDBN} may use accesses of any size.
8610 Use 8 bit memory accesses.
8612 Use 16 bit memory accesses.
8614 Use 32 bit memory accesses.
8616 Use 64 bit memory accesses.
8619 @c @subsubsection Hardware/Software Breakpoints
8620 @c The hardware/software breakpoint attributes set whether @value{GDBN}
8621 @c will use hardware or software breakpoints for the internal breakpoints
8622 @c used by the step, next, finish, until, etc. commands.
8626 @c Always use hardware breakpoints
8627 @c @item swbreak (default)
8630 @subsubsection Data Cache
8631 The data cache attributes set whether @value{GDBN} will cache target
8632 memory. While this generally improves performance by reducing debug
8633 protocol overhead, it can lead to incorrect results because @value{GDBN}
8634 does not know about volatile variables or memory mapped device
8639 Enable @value{GDBN} to cache target memory.
8641 Disable @value{GDBN} from caching target memory. This is the default.
8644 @subsection Memory Access Checking
8645 @value{GDBN} can be instructed to refuse accesses to memory that is
8646 not explicitly described. This can be useful if accessing such
8647 regions has undesired effects for a specific target, or to provide
8648 better error checking. The following commands control this behaviour.
8651 @kindex set mem inaccessible-by-default
8652 @item set mem inaccessible-by-default [on|off]
8653 If @code{on} is specified, make @value{GDBN} treat memory not
8654 explicitly described by the memory ranges as non-existent and refuse accesses
8655 to such memory. The checks are only performed if there's at least one
8656 memory range defined. If @code{off} is specified, make @value{GDBN}
8657 treat the memory not explicitly described by the memory ranges as RAM.
8658 The default value is @code{on}.
8659 @kindex show mem inaccessible-by-default
8660 @item show mem inaccessible-by-default
8661 Show the current handling of accesses to unknown memory.
8665 @c @subsubsection Memory Write Verification
8666 @c The memory write verification attributes set whether @value{GDBN}
8667 @c will re-reads data after each write to verify the write was successful.
8671 @c @item noverify (default)
8674 @node Dump/Restore Files
8675 @section Copy Between Memory and a File
8676 @cindex dump/restore files
8677 @cindex append data to a file
8678 @cindex dump data to a file
8679 @cindex restore data from a file
8681 You can use the commands @code{dump}, @code{append}, and
8682 @code{restore} to copy data between target memory and a file. The
8683 @code{dump} and @code{append} commands write data to a file, and the
8684 @code{restore} command reads data from a file back into the inferior's
8685 memory. Files may be in binary, Motorola S-record, Intel hex, or
8686 Tektronix Hex format; however, @value{GDBN} can only append to binary
8692 @item dump @r{[}@var{format}@r{]} memory @var{filename} @var{start_addr} @var{end_addr}
8693 @itemx dump @r{[}@var{format}@r{]} value @var{filename} @var{expr}
8694 Dump the contents of memory from @var{start_addr} to @var{end_addr},
8695 or the value of @var{expr}, to @var{filename} in the given format.
8697 The @var{format} parameter may be any one of:
8704 Motorola S-record format.
8706 Tektronix Hex format.
8709 @value{GDBN} uses the same definitions of these formats as the
8710 @sc{gnu} binary utilities, like @samp{objdump} and @samp{objcopy}. If
8711 @var{format} is omitted, @value{GDBN} dumps the data in raw binary
8715 @item append @r{[}binary@r{]} memory @var{filename} @var{start_addr} @var{end_addr}
8716 @itemx append @r{[}binary@r{]} value @var{filename} @var{expr}
8717 Append the contents of memory from @var{start_addr} to @var{end_addr},
8718 or the value of @var{expr}, to the file @var{filename}, in raw binary form.
8719 (@value{GDBN} can only append data to files in raw binary form.)
8722 @item restore @var{filename} @r{[}binary@r{]} @var{bias} @var{start} @var{end}
8723 Restore the contents of file @var{filename} into memory. The
8724 @code{restore} command can automatically recognize any known @sc{bfd}
8725 file format, except for raw binary. To restore a raw binary file you
8726 must specify the optional keyword @code{binary} after the filename.
8728 If @var{bias} is non-zero, its value will be added to the addresses
8729 contained in the file. Binary files always start at address zero, so
8730 they will be restored at address @var{bias}. Other bfd files have
8731 a built-in location; they will be restored at offset @var{bias}
8734 If @var{start} and/or @var{end} are non-zero, then only data between
8735 file offset @var{start} and file offset @var{end} will be restored.
8736 These offsets are relative to the addresses in the file, before
8737 the @var{bias} argument is applied.
8741 @node Core File Generation
8742 @section How to Produce a Core File from Your Program
8743 @cindex dump core from inferior
8745 A @dfn{core file} or @dfn{core dump} is a file that records the memory
8746 image of a running process and its process status (register values
8747 etc.). Its primary use is post-mortem debugging of a program that
8748 crashed while it ran outside a debugger. A program that crashes
8749 automatically produces a core file, unless this feature is disabled by
8750 the user. @xref{Files}, for information on invoking @value{GDBN} in
8751 the post-mortem debugging mode.
8753 Occasionally, you may wish to produce a core file of the program you
8754 are debugging in order to preserve a snapshot of its state.
8755 @value{GDBN} has a special command for that.
8759 @kindex generate-core-file
8760 @item generate-core-file [@var{file}]
8761 @itemx gcore [@var{file}]
8762 Produce a core dump of the inferior process. The optional argument
8763 @var{file} specifies the file name where to put the core dump. If not
8764 specified, the file name defaults to @file{core.@var{pid}}, where
8765 @var{pid} is the inferior process ID.
8767 Note that this command is implemented only for some systems (as of
8768 this writing, @sc{gnu}/Linux, FreeBSD, Solaris, Unixware, and S390).
8771 @node Character Sets
8772 @section Character Sets
8773 @cindex character sets
8775 @cindex translating between character sets
8776 @cindex host character set
8777 @cindex target character set
8779 If the program you are debugging uses a different character set to
8780 represent characters and strings than the one @value{GDBN} uses itself,
8781 @value{GDBN} can automatically translate between the character sets for
8782 you. The character set @value{GDBN} uses we call the @dfn{host
8783 character set}; the one the inferior program uses we call the
8784 @dfn{target character set}.
8786 For example, if you are running @value{GDBN} on a @sc{gnu}/Linux system, which
8787 uses the ISO Latin 1 character set, but you are using @value{GDBN}'s
8788 remote protocol (@pxref{Remote Debugging}) to debug a program
8789 running on an IBM mainframe, which uses the @sc{ebcdic} character set,
8790 then the host character set is Latin-1, and the target character set is
8791 @sc{ebcdic}. If you give @value{GDBN} the command @code{set
8792 target-charset EBCDIC-US}, then @value{GDBN} translates between
8793 @sc{ebcdic} and Latin 1 as you print character or string values, or use
8794 character and string literals in expressions.
8796 @value{GDBN} has no way to automatically recognize which character set
8797 the inferior program uses; you must tell it, using the @code{set
8798 target-charset} command, described below.
8800 Here are the commands for controlling @value{GDBN}'s character set
8804 @item set target-charset @var{charset}
8805 @kindex set target-charset
8806 Set the current target character set to @var{charset}. To display the
8807 list of supported target character sets, type
8808 @kbd{@w{set target-charset @key{TAB}@key{TAB}}}.
8810 @item set host-charset @var{charset}
8811 @kindex set host-charset
8812 Set the current host character set to @var{charset}.
8814 By default, @value{GDBN} uses a host character set appropriate to the
8815 system it is running on; you can override that default using the
8816 @code{set host-charset} command. On some systems, @value{GDBN} cannot
8817 automatically determine the appropriate host character set. In this
8818 case, @value{GDBN} uses @samp{UTF-8}.
8820 @value{GDBN} can only use certain character sets as its host character
8821 set. If you type @kbd{@w{set target-charset @key{TAB}@key{TAB}}},
8822 @value{GDBN} will list the host character sets it supports.
8824 @item set charset @var{charset}
8826 Set the current host and target character sets to @var{charset}. As
8827 above, if you type @kbd{@w{set charset @key{TAB}@key{TAB}}},
8828 @value{GDBN} will list the names of the character sets that can be used
8829 for both host and target.
8832 @kindex show charset
8833 Show the names of the current host and target character sets.
8835 @item show host-charset
8836 @kindex show host-charset
8837 Show the name of the current host character set.
8839 @item show target-charset
8840 @kindex show target-charset
8841 Show the name of the current target character set.
8843 @item set target-wide-charset @var{charset}
8844 @kindex set target-wide-charset
8845 Set the current target's wide character set to @var{charset}. This is
8846 the character set used by the target's @code{wchar_t} type. To
8847 display the list of supported wide character sets, type
8848 @kbd{@w{set target-wide-charset @key{TAB}@key{TAB}}}.
8850 @item show target-wide-charset
8851 @kindex show target-wide-charset
8852 Show the name of the current target's wide character set.
8855 Here is an example of @value{GDBN}'s character set support in action.
8856 Assume that the following source code has been placed in the file
8857 @file{charset-test.c}:
8863 = @{72, 101, 108, 108, 111, 44, 32, 119,
8864 111, 114, 108, 100, 33, 10, 0@};
8865 char ibm1047_hello[]
8866 = @{200, 133, 147, 147, 150, 107, 64, 166,
8867 150, 153, 147, 132, 90, 37, 0@};
8871 printf ("Hello, world!\n");
8875 In this program, @code{ascii_hello} and @code{ibm1047_hello} are arrays
8876 containing the string @samp{Hello, world!} followed by a newline,
8877 encoded in the @sc{ascii} and @sc{ibm1047} character sets.
8879 We compile the program, and invoke the debugger on it:
8882 $ gcc -g charset-test.c -o charset-test
8883 $ gdb -nw charset-test
8884 GNU gdb 2001-12-19-cvs
8885 Copyright 2001 Free Software Foundation, Inc.
8890 We can use the @code{show charset} command to see what character sets
8891 @value{GDBN} is currently using to interpret and display characters and
8895 (@value{GDBP}) show charset
8896 The current host and target character set is `ISO-8859-1'.
8900 For the sake of printing this manual, let's use @sc{ascii} as our
8901 initial character set:
8903 (@value{GDBP}) set charset ASCII
8904 (@value{GDBP}) show charset
8905 The current host and target character set is `ASCII'.
8909 Let's assume that @sc{ascii} is indeed the correct character set for our
8910 host system --- in other words, let's assume that if @value{GDBN} prints
8911 characters using the @sc{ascii} character set, our terminal will display
8912 them properly. Since our current target character set is also
8913 @sc{ascii}, the contents of @code{ascii_hello} print legibly:
8916 (@value{GDBP}) print ascii_hello
8917 $1 = 0x401698 "Hello, world!\n"
8918 (@value{GDBP}) print ascii_hello[0]
8923 @value{GDBN} uses the target character set for character and string
8924 literals you use in expressions:
8927 (@value{GDBP}) print '+'
8932 The @sc{ascii} character set uses the number 43 to encode the @samp{+}
8935 @value{GDBN} relies on the user to tell it which character set the
8936 target program uses. If we print @code{ibm1047_hello} while our target
8937 character set is still @sc{ascii}, we get jibberish:
8940 (@value{GDBP}) print ibm1047_hello
8941 $4 = 0x4016a8 "\310\205\223\223\226k@@\246\226\231\223\204Z%"
8942 (@value{GDBP}) print ibm1047_hello[0]
8947 If we invoke the @code{set target-charset} followed by @key{TAB}@key{TAB},
8948 @value{GDBN} tells us the character sets it supports:
8951 (@value{GDBP}) set target-charset
8952 ASCII EBCDIC-US IBM1047 ISO-8859-1
8953 (@value{GDBP}) set target-charset
8956 We can select @sc{ibm1047} as our target character set, and examine the
8957 program's strings again. Now the @sc{ascii} string is wrong, but
8958 @value{GDBN} translates the contents of @code{ibm1047_hello} from the
8959 target character set, @sc{ibm1047}, to the host character set,
8960 @sc{ascii}, and they display correctly:
8963 (@value{GDBP}) set target-charset IBM1047
8964 (@value{GDBP}) show charset
8965 The current host character set is `ASCII'.
8966 The current target character set is `IBM1047'.
8967 (@value{GDBP}) print ascii_hello
8968 $6 = 0x401698 "\110\145%%?\054\040\167?\162%\144\041\012"
8969 (@value{GDBP}) print ascii_hello[0]
8971 (@value{GDBP}) print ibm1047_hello
8972 $8 = 0x4016a8 "Hello, world!\n"
8973 (@value{GDBP}) print ibm1047_hello[0]
8978 As above, @value{GDBN} uses the target character set for character and
8979 string literals you use in expressions:
8982 (@value{GDBP}) print '+'
8987 The @sc{ibm1047} character set uses the number 78 to encode the @samp{+}
8990 @node Caching Remote Data
8991 @section Caching Data of Remote Targets
8992 @cindex caching data of remote targets
8994 @value{GDBN} caches data exchanged between the debugger and a
8995 remote target (@pxref{Remote Debugging}). Such caching generally improves
8996 performance, because it reduces the overhead of the remote protocol by
8997 bundling memory reads and writes into large chunks. Unfortunately, simply
8998 caching everything would lead to incorrect results, since @value{GDBN}
8999 does not necessarily know anything about volatile values, memory-mapped I/O
9000 addresses, etc. Furthermore, in non-stop mode (@pxref{Non-Stop Mode})
9001 memory can be changed @emph{while} a gdb command is executing.
9002 Therefore, by default, @value{GDBN} only caches data
9003 known to be on the stack@footnote{In non-stop mode, it is moderately
9004 rare for a running thread to modify the stack of a stopped thread
9005 in a way that would interfere with a backtrace, and caching of
9006 stack reads provides a significant speed up of remote backtraces.}.
9007 Other regions of memory can be explicitly marked as
9008 cacheable; see @pxref{Memory Region Attributes}.
9011 @kindex set remotecache
9012 @item set remotecache on
9013 @itemx set remotecache off
9014 This option no longer does anything; it exists for compatibility
9017 @kindex show remotecache
9018 @item show remotecache
9019 Show the current state of the obsolete remotecache flag.
9021 @kindex set stack-cache
9022 @item set stack-cache on
9023 @itemx set stack-cache off
9024 Enable or disable caching of stack accesses. When @code{ON}, use
9025 caching. By default, this option is @code{ON}.
9027 @kindex show stack-cache
9028 @item show stack-cache
9029 Show the current state of data caching for memory accesses.
9032 @item info dcache @r{[}line@r{]}
9033 Print the information about the data cache performance. The
9034 information displayed includes the dcache width and depth, and for
9035 each cache line, its number, address, and how many times it was
9036 referenced. This command is useful for debugging the data cache
9039 If a line number is specified, the contents of that line will be
9043 @node Searching Memory
9044 @section Search Memory
9045 @cindex searching memory
9047 Memory can be searched for a particular sequence of bytes with the
9048 @code{find} command.
9052 @item find @r{[}/@var{sn}@r{]} @var{start_addr}, +@var{len}, @var{val1} @r{[}, @var{val2}, @dots{}@r{]}
9053 @itemx find @r{[}/@var{sn}@r{]} @var{start_addr}, @var{end_addr}, @var{val1} @r{[}, @var{val2}, @dots{}@r{]}
9054 Search memory for the sequence of bytes specified by @var{val1}, @var{val2},
9055 etc. The search begins at address @var{start_addr} and continues for either
9056 @var{len} bytes or through to @var{end_addr} inclusive.
9059 @var{s} and @var{n} are optional parameters.
9060 They may be specified in either order, apart or together.
9063 @item @var{s}, search query size
9064 The size of each search query value.
9070 halfwords (two bytes)
9074 giant words (eight bytes)
9077 All values are interpreted in the current language.
9078 This means, for example, that if the current source language is C/C@t{++}
9079 then searching for the string ``hello'' includes the trailing '\0'.
9081 If the value size is not specified, it is taken from the
9082 value's type in the current language.
9083 This is useful when one wants to specify the search
9084 pattern as a mixture of types.
9085 Note that this means, for example, that in the case of C-like languages
9086 a search for an untyped 0x42 will search for @samp{(int) 0x42}
9087 which is typically four bytes.
9089 @item @var{n}, maximum number of finds
9090 The maximum number of matches to print. The default is to print all finds.
9093 You can use strings as search values. Quote them with double-quotes
9095 The string value is copied into the search pattern byte by byte,
9096 regardless of the endianness of the target and the size specification.
9098 The address of each match found is printed as well as a count of the
9099 number of matches found.
9101 The address of the last value found is stored in convenience variable
9103 A count of the number of matches is stored in @samp{$numfound}.
9105 For example, if stopped at the @code{printf} in this function:
9111 static char hello[] = "hello-hello";
9112 static struct @{ char c; short s; int i; @}
9113 __attribute__ ((packed)) mixed
9114 = @{ 'c', 0x1234, 0x87654321 @};
9115 printf ("%s\n", hello);
9120 you get during debugging:
9123 (gdb) find &hello[0], +sizeof(hello), "hello"
9124 0x804956d <hello.1620+6>
9126 (gdb) find &hello[0], +sizeof(hello), 'h', 'e', 'l', 'l', 'o'
9127 0x8049567 <hello.1620>
9128 0x804956d <hello.1620+6>
9130 (gdb) find /b1 &hello[0], +sizeof(hello), 'h', 0x65, 'l'
9131 0x8049567 <hello.1620>
9133 (gdb) find &mixed, +sizeof(mixed), (char) 'c', (short) 0x1234, (int) 0x87654321
9134 0x8049560 <mixed.1625>
9136 (gdb) print $numfound
9139 $2 = (void *) 0x8049560
9142 @node Optimized Code
9143 @chapter Debugging Optimized Code
9144 @cindex optimized code, debugging
9145 @cindex debugging optimized code
9147 Almost all compilers support optimization. With optimization
9148 disabled, the compiler generates assembly code that corresponds
9149 directly to your source code, in a simplistic way. As the compiler
9150 applies more powerful optimizations, the generated assembly code
9151 diverges from your original source code. With help from debugging
9152 information generated by the compiler, @value{GDBN} can map from
9153 the running program back to constructs from your original source.
9155 @value{GDBN} is more accurate with optimization disabled. If you
9156 can recompile without optimization, it is easier to follow the
9157 progress of your program during debugging. But, there are many cases
9158 where you may need to debug an optimized version.
9160 When you debug a program compiled with @samp{-g -O}, remember that the
9161 optimizer has rearranged your code; the debugger shows you what is
9162 really there. Do not be too surprised when the execution path does not
9163 exactly match your source file! An extreme example: if you define a
9164 variable, but never use it, @value{GDBN} never sees that
9165 variable---because the compiler optimizes it out of existence.
9167 Some things do not work as well with @samp{-g -O} as with just
9168 @samp{-g}, particularly on machines with instruction scheduling. If in
9169 doubt, recompile with @samp{-g} alone, and if this fixes the problem,
9170 please report it to us as a bug (including a test case!).
9171 @xref{Variables}, for more information about debugging optimized code.
9174 * Inline Functions:: How @value{GDBN} presents inlining
9177 @node Inline Functions
9178 @section Inline Functions
9179 @cindex inline functions, debugging
9181 @dfn{Inlining} is an optimization that inserts a copy of the function
9182 body directly at each call site, instead of jumping to a shared
9183 routine. @value{GDBN} displays inlined functions just like
9184 non-inlined functions. They appear in backtraces. You can view their
9185 arguments and local variables, step into them with @code{step}, skip
9186 them with @code{next}, and escape from them with @code{finish}.
9187 You can check whether a function was inlined by using the
9188 @code{info frame} command.
9190 For @value{GDBN} to support inlined functions, the compiler must
9191 record information about inlining in the debug information ---
9192 @value{NGCC} using the @sc{dwarf 2} format does this, and several
9193 other compilers do also. @value{GDBN} only supports inlined functions
9194 when using @sc{dwarf 2}. Versions of @value{NGCC} before 4.1
9195 do not emit two required attributes (@samp{DW_AT_call_file} and
9196 @samp{DW_AT_call_line}); @value{GDBN} does not display inlined
9197 function calls with earlier versions of @value{NGCC}. It instead
9198 displays the arguments and local variables of inlined functions as
9199 local variables in the caller.
9201 The body of an inlined function is directly included at its call site;
9202 unlike a non-inlined function, there are no instructions devoted to
9203 the call. @value{GDBN} still pretends that the call site and the
9204 start of the inlined function are different instructions. Stepping to
9205 the call site shows the call site, and then stepping again shows
9206 the first line of the inlined function, even though no additional
9207 instructions are executed.
9209 This makes source-level debugging much clearer; you can see both the
9210 context of the call and then the effect of the call. Only stepping by
9211 a single instruction using @code{stepi} or @code{nexti} does not do
9212 this; single instruction steps always show the inlined body.
9214 There are some ways that @value{GDBN} does not pretend that inlined
9215 function calls are the same as normal calls:
9219 You cannot set breakpoints on inlined functions. @value{GDBN}
9220 either reports that there is no symbol with that name, or else sets the
9221 breakpoint only on non-inlined copies of the function. This limitation
9222 will be removed in a future version of @value{GDBN}; until then,
9223 set a breakpoint by line number on the first line of the inlined
9227 Setting breakpoints at the call site of an inlined function may not
9228 work, because the call site does not contain any code. @value{GDBN}
9229 may incorrectly move the breakpoint to the next line of the enclosing
9230 function, after the call. This limitation will be removed in a future
9231 version of @value{GDBN}; until then, set a breakpoint on an earlier line
9232 or inside the inlined function instead.
9235 @value{GDBN} cannot locate the return value of inlined calls after
9236 using the @code{finish} command. This is a limitation of compiler-generated
9237 debugging information; after @code{finish}, you can step to the next line
9238 and print a variable where your program stored the return value.
9244 @chapter C Preprocessor Macros
9246 Some languages, such as C and C@t{++}, provide a way to define and invoke
9247 ``preprocessor macros'' which expand into strings of tokens.
9248 @value{GDBN} can evaluate expressions containing macro invocations, show
9249 the result of macro expansion, and show a macro's definition, including
9250 where it was defined.
9252 You may need to compile your program specially to provide @value{GDBN}
9253 with information about preprocessor macros. Most compilers do not
9254 include macros in their debugging information, even when you compile
9255 with the @option{-g} flag. @xref{Compilation}.
9257 A program may define a macro at one point, remove that definition later,
9258 and then provide a different definition after that. Thus, at different
9259 points in the program, a macro may have different definitions, or have
9260 no definition at all. If there is a current stack frame, @value{GDBN}
9261 uses the macros in scope at that frame's source code line. Otherwise,
9262 @value{GDBN} uses the macros in scope at the current listing location;
9265 Whenever @value{GDBN} evaluates an expression, it always expands any
9266 macro invocations present in the expression. @value{GDBN} also provides
9267 the following commands for working with macros explicitly.
9271 @kindex macro expand
9272 @cindex macro expansion, showing the results of preprocessor
9273 @cindex preprocessor macro expansion, showing the results of
9274 @cindex expanding preprocessor macros
9275 @item macro expand @var{expression}
9276 @itemx macro exp @var{expression}
9277 Show the results of expanding all preprocessor macro invocations in
9278 @var{expression}. Since @value{GDBN} simply expands macros, but does
9279 not parse the result, @var{expression} need not be a valid expression;
9280 it can be any string of tokens.
9283 @item macro expand-once @var{expression}
9284 @itemx macro exp1 @var{expression}
9285 @cindex expand macro once
9286 @i{(This command is not yet implemented.)} Show the results of
9287 expanding those preprocessor macro invocations that appear explicitly in
9288 @var{expression}. Macro invocations appearing in that expansion are
9289 left unchanged. This command allows you to see the effect of a
9290 particular macro more clearly, without being confused by further
9291 expansions. Since @value{GDBN} simply expands macros, but does not
9292 parse the result, @var{expression} need not be a valid expression; it
9293 can be any string of tokens.
9296 @cindex macro definition, showing
9297 @cindex definition, showing a macro's
9298 @item info macro @var{macro}
9299 Show the definition of the macro named @var{macro}, and describe the
9300 source location or compiler command-line where that definition was established.
9302 @kindex macro define
9303 @cindex user-defined macros
9304 @cindex defining macros interactively
9305 @cindex macros, user-defined
9306 @item macro define @var{macro} @var{replacement-list}
9307 @itemx macro define @var{macro}(@var{arglist}) @var{replacement-list}
9308 Introduce a definition for a preprocessor macro named @var{macro},
9309 invocations of which are replaced by the tokens given in
9310 @var{replacement-list}. The first form of this command defines an
9311 ``object-like'' macro, which takes no arguments; the second form
9312 defines a ``function-like'' macro, which takes the arguments given in
9315 A definition introduced by this command is in scope in every
9316 expression evaluated in @value{GDBN}, until it is removed with the
9317 @code{macro undef} command, described below. The definition overrides
9318 all definitions for @var{macro} present in the program being debugged,
9319 as well as any previous user-supplied definition.
9322 @item macro undef @var{macro}
9323 Remove any user-supplied definition for the macro named @var{macro}.
9324 This command only affects definitions provided with the @code{macro
9325 define} command, described above; it cannot remove definitions present
9326 in the program being debugged.
9330 List all the macros defined using the @code{macro define} command.
9333 @cindex macros, example of debugging with
9334 Here is a transcript showing the above commands in action. First, we
9335 show our source files:
9343 #define ADD(x) (M + x)
9348 printf ("Hello, world!\n");
9350 printf ("We're so creative.\n");
9352 printf ("Goodbye, world!\n");
9359 Now, we compile the program using the @sc{gnu} C compiler, @value{NGCC}.
9360 We pass the @option{-gdwarf-2} and @option{-g3} flags to ensure the
9361 compiler includes information about preprocessor macros in the debugging
9365 $ gcc -gdwarf-2 -g3 sample.c -o sample
9369 Now, we start @value{GDBN} on our sample program:
9373 GNU gdb 2002-05-06-cvs
9374 Copyright 2002 Free Software Foundation, Inc.
9375 GDB is free software, @dots{}
9379 We can expand macros and examine their definitions, even when the
9380 program is not running. @value{GDBN} uses the current listing position
9381 to decide which macro definitions are in scope:
9384 (@value{GDBP}) list main
9387 5 #define ADD(x) (M + x)
9392 10 printf ("Hello, world!\n");
9394 12 printf ("We're so creative.\n");
9395 (@value{GDBP}) info macro ADD
9396 Defined at /home/jimb/gdb/macros/play/sample.c:5
9397 #define ADD(x) (M + x)
9398 (@value{GDBP}) info macro Q
9399 Defined at /home/jimb/gdb/macros/play/sample.h:1
9400 included at /home/jimb/gdb/macros/play/sample.c:2
9402 (@value{GDBP}) macro expand ADD(1)
9403 expands to: (42 + 1)
9404 (@value{GDBP}) macro expand-once ADD(1)
9405 expands to: once (M + 1)
9409 In the example above, note that @code{macro expand-once} expands only
9410 the macro invocation explicit in the original text --- the invocation of
9411 @code{ADD} --- but does not expand the invocation of the macro @code{M},
9412 which was introduced by @code{ADD}.
9414 Once the program is running, @value{GDBN} uses the macro definitions in
9415 force at the source line of the current stack frame:
9418 (@value{GDBP}) break main
9419 Breakpoint 1 at 0x8048370: file sample.c, line 10.
9421 Starting program: /home/jimb/gdb/macros/play/sample
9423 Breakpoint 1, main () at sample.c:10
9424 10 printf ("Hello, world!\n");
9428 At line 10, the definition of the macro @code{N} at line 9 is in force:
9431 (@value{GDBP}) info macro N
9432 Defined at /home/jimb/gdb/macros/play/sample.c:9
9434 (@value{GDBP}) macro expand N Q M
9436 (@value{GDBP}) print N Q M
9441 As we step over directives that remove @code{N}'s definition, and then
9442 give it a new definition, @value{GDBN} finds the definition (or lack
9443 thereof) in force at each point:
9448 12 printf ("We're so creative.\n");
9449 (@value{GDBP}) info macro N
9450 The symbol `N' has no definition as a C/C++ preprocessor macro
9451 at /home/jimb/gdb/macros/play/sample.c:12
9454 14 printf ("Goodbye, world!\n");
9455 (@value{GDBP}) info macro N
9456 Defined at /home/jimb/gdb/macros/play/sample.c:13
9458 (@value{GDBP}) macro expand N Q M
9459 expands to: 1729 < 42
9460 (@value{GDBP}) print N Q M
9465 In addition to source files, macros can be defined on the compilation command
9466 line using the @option{-D@var{name}=@var{value}} syntax. For macros defined in
9467 such a way, @value{GDBN} displays the location of their definition as line zero
9468 of the source file submitted to the compiler.
9471 (@value{GDBP}) info macro __STDC__
9472 Defined at /home/jimb/gdb/macros/play/sample.c:0
9479 @chapter Tracepoints
9480 @c This chapter is based on the documentation written by Michael
9481 @c Snyder, David Taylor, Jim Blandy, and Elena Zannoni.
9484 In some applications, it is not feasible for the debugger to interrupt
9485 the program's execution long enough for the developer to learn
9486 anything helpful about its behavior. If the program's correctness
9487 depends on its real-time behavior, delays introduced by a debugger
9488 might cause the program to change its behavior drastically, or perhaps
9489 fail, even when the code itself is correct. It is useful to be able
9490 to observe the program's behavior without interrupting it.
9492 Using @value{GDBN}'s @code{trace} and @code{collect} commands, you can
9493 specify locations in the program, called @dfn{tracepoints}, and
9494 arbitrary expressions to evaluate when those tracepoints are reached.
9495 Later, using the @code{tfind} command, you can examine the values
9496 those expressions had when the program hit the tracepoints. The
9497 expressions may also denote objects in memory---structures or arrays,
9498 for example---whose values @value{GDBN} should record; while visiting
9499 a particular tracepoint, you may inspect those objects as if they were
9500 in memory at that moment. However, because @value{GDBN} records these
9501 values without interacting with you, it can do so quickly and
9502 unobtrusively, hopefully not disturbing the program's behavior.
9504 The tracepoint facility is currently available only for remote
9505 targets. @xref{Targets}. In addition, your remote target must know
9506 how to collect trace data. This functionality is implemented in the
9507 remote stub; however, none of the stubs distributed with @value{GDBN}
9508 support tracepoints as of this writing. The format of the remote
9509 packets used to implement tracepoints are described in @ref{Tracepoint
9512 It is also possible to get trace data from a file, in a manner reminiscent
9513 of corefiles; you specify the filename, and use @code{tfind} to search
9514 through the file. @xref{Trace Files}, for more details.
9516 This chapter describes the tracepoint commands and features.
9520 * Analyze Collected Data::
9521 * Tracepoint Variables::
9525 @node Set Tracepoints
9526 @section Commands to Set Tracepoints
9528 Before running such a @dfn{trace experiment}, an arbitrary number of
9529 tracepoints can be set. A tracepoint is actually a special type of
9530 breakpoint (@pxref{Set Breaks}), so you can manipulate it using
9531 standard breakpoint commands. For instance, as with breakpoints,
9532 tracepoint numbers are successive integers starting from one, and many
9533 of the commands associated with tracepoints take the tracepoint number
9534 as their argument, to identify which tracepoint to work on.
9536 For each tracepoint, you can specify, in advance, some arbitrary set
9537 of data that you want the target to collect in the trace buffer when
9538 it hits that tracepoint. The collected data can include registers,
9539 local variables, or global data. Later, you can use @value{GDBN}
9540 commands to examine the values these data had at the time the
9543 Tracepoints do not support every breakpoint feature. Ignore counts on
9544 tracepoints have no effect, and tracepoints cannot run @value{GDBN}
9545 commands when they are hit. Tracepoints may not be thread-specific
9548 @cindex fast tracepoints
9549 Some targets may support @dfn{fast tracepoints}, which are inserted in
9550 a different way (such as with a jump instead of a trap), that is
9551 faster but possibly restricted in where they may be installed.
9553 @code{gdbserver} supports tracepoints on some target systems.
9554 @xref{Server,,Tracepoints support in @code{gdbserver}}.
9556 This section describes commands to set tracepoints and associated
9557 conditions and actions.
9560 * Create and Delete Tracepoints::
9561 * Enable and Disable Tracepoints::
9562 * Tracepoint Passcounts::
9563 * Tracepoint Conditions::
9564 * Trace State Variables::
9565 * Tracepoint Actions::
9566 * Listing Tracepoints::
9567 * Starting and Stopping Trace Experiments::
9568 * Tracepoint Restrictions::
9571 @node Create and Delete Tracepoints
9572 @subsection Create and Delete Tracepoints
9575 @cindex set tracepoint
9577 @item trace @var{location}
9578 The @code{trace} command is very similar to the @code{break} command.
9579 Its argument @var{location} can be a source line, a function name, or
9580 an address in the target program. @xref{Specify Location}. The
9581 @code{trace} command defines a tracepoint, which is a point in the
9582 target program where the debugger will briefly stop, collect some
9583 data, and then allow the program to continue. Setting a tracepoint or
9584 changing its actions doesn't take effect until the next @code{tstart}
9585 command, and once a trace experiment is running, further changes will
9586 not have any effect until the next trace experiment starts.
9588 Here are some examples of using the @code{trace} command:
9591 (@value{GDBP}) @b{trace foo.c:121} // a source file and line number
9593 (@value{GDBP}) @b{trace +2} // 2 lines forward
9595 (@value{GDBP}) @b{trace my_function} // first source line of function
9597 (@value{GDBP}) @b{trace *my_function} // EXACT start address of function
9599 (@value{GDBP}) @b{trace *0x2117c4} // an address
9603 You can abbreviate @code{trace} as @code{tr}.
9605 @item trace @var{location} if @var{cond}
9606 Set a tracepoint with condition @var{cond}; evaluate the expression
9607 @var{cond} each time the tracepoint is reached, and collect data only
9608 if the value is nonzero---that is, if @var{cond} evaluates as true.
9609 @xref{Tracepoint Conditions, ,Tracepoint Conditions}, for more
9610 information on tracepoint conditions.
9612 @item ftrace @var{location} [ if @var{cond} ]
9613 @cindex set fast tracepoint
9615 The @code{ftrace} command sets a fast tracepoint. For targets that
9616 support them, fast tracepoints will use a more efficient but possibly
9617 less general technique to trigger data collection, such as a jump
9618 instruction instead of a trap, or some sort of hardware support. It
9619 may not be possible to create a fast tracepoint at the desired
9620 location, in which case the command will exit with an explanatory
9623 @value{GDBN} handles arguments to @code{ftrace} exactly as for
9627 @cindex last tracepoint number
9628 @cindex recent tracepoint number
9629 @cindex tracepoint number
9630 The convenience variable @code{$tpnum} records the tracepoint number
9631 of the most recently set tracepoint.
9633 @kindex delete tracepoint
9634 @cindex tracepoint deletion
9635 @item delete tracepoint @r{[}@var{num}@r{]}
9636 Permanently delete one or more tracepoints. With no argument, the
9637 default is to delete all tracepoints. Note that the regular
9638 @code{delete} command can remove tracepoints also.
9643 (@value{GDBP}) @b{delete trace 1 2 3} // remove three tracepoints
9645 (@value{GDBP}) @b{delete trace} // remove all tracepoints
9649 You can abbreviate this command as @code{del tr}.
9652 @node Enable and Disable Tracepoints
9653 @subsection Enable and Disable Tracepoints
9655 These commands are deprecated; they are equivalent to plain @code{disable} and @code{enable}.
9658 @kindex disable tracepoint
9659 @item disable tracepoint @r{[}@var{num}@r{]}
9660 Disable tracepoint @var{num}, or all tracepoints if no argument
9661 @var{num} is given. A disabled tracepoint will have no effect during
9662 the next trace experiment, but it is not forgotten. You can re-enable
9663 a disabled tracepoint using the @code{enable tracepoint} command.
9665 @kindex enable tracepoint
9666 @item enable tracepoint @r{[}@var{num}@r{]}
9667 Enable tracepoint @var{num}, or all tracepoints. The enabled
9668 tracepoints will become effective the next time a trace experiment is
9672 @node Tracepoint Passcounts
9673 @subsection Tracepoint Passcounts
9677 @cindex tracepoint pass count
9678 @item passcount @r{[}@var{n} @r{[}@var{num}@r{]]}
9679 Set the @dfn{passcount} of a tracepoint. The passcount is a way to
9680 automatically stop a trace experiment. If a tracepoint's passcount is
9681 @var{n}, then the trace experiment will be automatically stopped on
9682 the @var{n}'th time that tracepoint is hit. If the tracepoint number
9683 @var{num} is not specified, the @code{passcount} command sets the
9684 passcount of the most recently defined tracepoint. If no passcount is
9685 given, the trace experiment will run until stopped explicitly by the
9691 (@value{GDBP}) @b{passcount 5 2} // Stop on the 5th execution of
9692 @exdent @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @code{// tracepoint 2}
9694 (@value{GDBP}) @b{passcount 12} // Stop on the 12th execution of the
9695 @exdent @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @code{// most recently defined tracepoint.}
9696 (@value{GDBP}) @b{trace foo}
9697 (@value{GDBP}) @b{pass 3}
9698 (@value{GDBP}) @b{trace bar}
9699 (@value{GDBP}) @b{pass 2}
9700 (@value{GDBP}) @b{trace baz}
9701 (@value{GDBP}) @b{pass 1} // Stop tracing when foo has been
9702 @exdent @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @code{// executed 3 times OR when bar has}
9703 @exdent @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @code{// been executed 2 times}
9704 @exdent @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @code{// OR when baz has been executed 1 time.}
9708 @node Tracepoint Conditions
9709 @subsection Tracepoint Conditions
9710 @cindex conditional tracepoints
9711 @cindex tracepoint conditions
9713 The simplest sort of tracepoint collects data every time your program
9714 reaches a specified place. You can also specify a @dfn{condition} for
9715 a tracepoint. A condition is just a Boolean expression in your
9716 programming language (@pxref{Expressions, ,Expressions}). A
9717 tracepoint with a condition evaluates the expression each time your
9718 program reaches it, and data collection happens only if the condition
9721 Tracepoint conditions can be specified when a tracepoint is set, by
9722 using @samp{if} in the arguments to the @code{trace} command.
9723 @xref{Create and Delete Tracepoints, ,Setting Tracepoints}. They can
9724 also be set or changed at any time with the @code{condition} command,
9725 just as with breakpoints.
9727 Unlike breakpoint conditions, @value{GDBN} does not actually evaluate
9728 the conditional expression itself. Instead, @value{GDBN} encodes the
9729 expression into an agent expression (@pxref{Agent Expressions}
9730 suitable for execution on the target, independently of @value{GDBN}.
9731 Global variables become raw memory locations, locals become stack
9732 accesses, and so forth.
9734 For instance, suppose you have a function that is usually called
9735 frequently, but should not be called after an error has occurred. You
9736 could use the following tracepoint command to collect data about calls
9737 of that function that happen while the error code is propagating
9738 through the program; an unconditional tracepoint could end up
9739 collecting thousands of useless trace frames that you would have to
9743 (@value{GDBP}) @kbd{trace normal_operation if errcode > 0}
9746 @node Trace State Variables
9747 @subsection Trace State Variables
9748 @cindex trace state variables
9750 A @dfn{trace state variable} is a special type of variable that is
9751 created and managed by target-side code. The syntax is the same as
9752 that for GDB's convenience variables (a string prefixed with ``$''),
9753 but they are stored on the target. They must be created explicitly,
9754 using a @code{tvariable} command. They are always 64-bit signed
9757 Trace state variables are remembered by @value{GDBN}, and downloaded
9758 to the target along with tracepoint information when the trace
9759 experiment starts. There are no intrinsic limits on the number of
9760 trace state variables, beyond memory limitations of the target.
9762 @cindex convenience variables, and trace state variables
9763 Although trace state variables are managed by the target, you can use
9764 them in print commands and expressions as if they were convenience
9765 variables; @value{GDBN} will get the current value from the target
9766 while the trace experiment is running. Trace state variables share
9767 the same namespace as other ``$'' variables, which means that you
9768 cannot have trace state variables with names like @code{$23} or
9769 @code{$pc}, nor can you have a trace state variable and a convenience
9770 variable with the same name.
9774 @item tvariable $@var{name} [ = @var{expression} ]
9776 The @code{tvariable} command creates a new trace state variable named
9777 @code{$@var{name}}, and optionally gives it an initial value of
9778 @var{expression}. @var{expression} is evaluated when this command is
9779 entered; the result will be converted to an integer if possible,
9780 otherwise @value{GDBN} will report an error. A subsequent
9781 @code{tvariable} command specifying the same name does not create a
9782 variable, but instead assigns the supplied initial value to the
9783 existing variable of that name, overwriting any previous initial
9784 value. The default initial value is 0.
9786 @item info tvariables
9787 @kindex info tvariables
9788 List all the trace state variables along with their initial values.
9789 Their current values may also be displayed, if the trace experiment is
9792 @item delete tvariable @r{[} $@var{name} @dots{} @r{]}
9793 @kindex delete tvariable
9794 Delete the given trace state variables, or all of them if no arguments
9799 @node Tracepoint Actions
9800 @subsection Tracepoint Action Lists
9804 @cindex tracepoint actions
9805 @item actions @r{[}@var{num}@r{]}
9806 This command will prompt for a list of actions to be taken when the
9807 tracepoint is hit. If the tracepoint number @var{num} is not
9808 specified, this command sets the actions for the one that was most
9809 recently defined (so that you can define a tracepoint and then say
9810 @code{actions} without bothering about its number). You specify the
9811 actions themselves on the following lines, one action at a time, and
9812 terminate the actions list with a line containing just @code{end}. So
9813 far, the only defined actions are @code{collect}, @code{teval}, and
9814 @code{while-stepping}.
9816 @code{actions} is actually equivalent to @code{commands} (@pxref{Break
9817 Commands, ,Breakpoint Command Lists}), except that only the defined
9818 actions are allowed; any other @value{GDBN} command is rejected.
9820 @cindex remove actions from a tracepoint
9821 To remove all actions from a tracepoint, type @samp{actions @var{num}}
9822 and follow it immediately with @samp{end}.
9825 (@value{GDBP}) @b{collect @var{data}} // collect some data
9827 (@value{GDBP}) @b{while-stepping 5} // single-step 5 times, collect data
9829 (@value{GDBP}) @b{end} // signals the end of actions.
9832 In the following example, the action list begins with @code{collect}
9833 commands indicating the things to be collected when the tracepoint is
9834 hit. Then, in order to single-step and collect additional data
9835 following the tracepoint, a @code{while-stepping} command is used,
9836 followed by the list of things to be collected after each step in a
9837 sequence of single steps. The @code{while-stepping} command is
9838 terminated by its own separate @code{end} command. Lastly, the action
9839 list is terminated by an @code{end} command.
9842 (@value{GDBP}) @b{trace foo}
9843 (@value{GDBP}) @b{actions}
9844 Enter actions for tracepoint 1, one per line:
9848 > collect $pc, arr[i]
9853 @kindex collect @r{(tracepoints)}
9854 @item collect @var{expr1}, @var{expr2}, @dots{}
9855 Collect values of the given expressions when the tracepoint is hit.
9856 This command accepts a comma-separated list of any valid expressions.
9857 In addition to global, static, or local variables, the following
9858 special arguments are supported:
9862 collect all registers
9865 collect all function arguments
9868 collect all local variables.
9871 You can give several consecutive @code{collect} commands, each one
9872 with a single argument, or one @code{collect} command with several
9873 arguments separated by commas; the effect is the same.
9875 The command @code{info scope} (@pxref{Symbols, info scope}) is
9876 particularly useful for figuring out what data to collect.
9878 @kindex teval @r{(tracepoints)}
9879 @item teval @var{expr1}, @var{expr2}, @dots{}
9880 Evaluate the given expressions when the tracepoint is hit. This
9881 command accepts a comma-separated list of expressions. The results
9882 are discarded, so this is mainly useful for assigning values to trace
9883 state variables (@pxref{Trace State Variables}) without adding those
9884 values to the trace buffer, as would be the case if the @code{collect}
9887 @kindex while-stepping @r{(tracepoints)}
9888 @item while-stepping @var{n}
9889 Perform @var{n} single-step instruction traces after the tracepoint,
9890 collecting new data after each step. The @code{while-stepping}
9891 command is followed by the list of what to collect while stepping
9892 (followed by its own @code{end} command):
9896 > collect $regs, myglobal
9902 Note that @code{$pc} is not automatically collected by
9903 @code{while-stepping}; you need to explicitly collect that register if
9904 you need it. You may abbreviate @code{while-stepping} as @code{ws} or
9907 @item set default-collect @var{expr1}, @var{expr2}, @dots{}
9908 @kindex set default-collect
9909 @cindex default collection action
9910 This variable is a list of expressions to collect at each tracepoint
9911 hit. It is effectively an additional @code{collect} action prepended
9912 to every tracepoint action list. The expressions are parsed
9913 individually for each tracepoint, so for instance a variable named
9914 @code{xyz} may be interpreted as a global for one tracepoint, and a
9915 local for another, as appropriate to the tracepoint's location.
9917 @item show default-collect
9918 @kindex show default-collect
9919 Show the list of expressions that are collected by default at each
9924 @node Listing Tracepoints
9925 @subsection Listing Tracepoints
9928 @kindex info tracepoints
9930 @cindex information about tracepoints
9931 @item info tracepoints @r{[}@var{num}@r{]}
9932 Display information about the tracepoint @var{num}. If you don't
9933 specify a tracepoint number, displays information about all the
9934 tracepoints defined so far. The format is similar to that used for
9935 @code{info breakpoints}; in fact, @code{info tracepoints} is the same
9936 command, simply restricting itself to tracepoints.
9938 A tracepoint's listing may include additional information specific to
9943 its passcount as given by the @code{passcount @var{n}} command
9947 (@value{GDBP}) @b{info trace}
9948 Num Type Disp Enb Address What
9949 1 tracepoint keep y 0x0804ab57 in foo() at main.cxx:7
9951 collect globfoo, $regs
9960 This command can be abbreviated @code{info tp}.
9963 @node Starting and Stopping Trace Experiments
9964 @subsection Starting and Stopping Trace Experiments
9968 @cindex start a new trace experiment
9969 @cindex collected data discarded
9971 This command takes no arguments. It starts the trace experiment, and
9972 begins collecting data. This has the side effect of discarding all
9973 the data collected in the trace buffer during the previous trace
9977 @cindex stop a running trace experiment
9979 This command takes no arguments. It ends the trace experiment, and
9980 stops collecting data.
9982 @strong{Note}: a trace experiment and data collection may stop
9983 automatically if any tracepoint's passcount is reached
9984 (@pxref{Tracepoint Passcounts}), or if the trace buffer becomes full.
9987 @cindex status of trace data collection
9988 @cindex trace experiment, status of
9990 This command displays the status of the current trace data
9994 Here is an example of the commands we described so far:
9997 (@value{GDBP}) @b{trace gdb_c_test}
9998 (@value{GDBP}) @b{actions}
9999 Enter actions for tracepoint #1, one per line.
10000 > collect $regs,$locals,$args
10001 > while-stepping 11
10005 (@value{GDBP}) @b{tstart}
10006 [time passes @dots{}]
10007 (@value{GDBP}) @b{tstop}
10010 @cindex disconnected tracing
10011 You can choose to continue running the trace experiment even if
10012 @value{GDBN} disconnects from the target, voluntarily or
10013 involuntarily. For commands such as @code{detach}, the debugger will
10014 ask what you want to do with the trace. But for unexpected
10015 terminations (@value{GDBN} crash, network outage), it would be
10016 unfortunate to lose hard-won trace data, so the variable
10017 @code{disconnected-tracing} lets you decide whether the trace should
10018 continue running without @value{GDBN}.
10021 @item set disconnected-tracing on
10022 @itemx set disconnected-tracing off
10023 @kindex set disconnected-tracing
10024 Choose whether a tracing run should continue to run if @value{GDBN}
10025 has disconnected from the target. Note that @code{detach} or
10026 @code{quit} will ask you directly what to do about a running trace no
10027 matter what this variable's setting, so the variable is mainly useful
10028 for handling unexpected situations, such as loss of the network.
10030 @item show disconnected-tracing
10031 @kindex show disconnected-tracing
10032 Show the current choice for disconnected tracing.
10036 When you reconnect to the target, the trace experiment may or may not
10037 still be running; it might have filled the trace buffer in the
10038 meantime, or stopped for one of the other reasons. If it is running,
10039 it will continue after reconnection.
10041 Upon reconnection, the target will upload information about the
10042 tracepoints in effect. @value{GDBN} will then compare that
10043 information to the set of tracepoints currently defined, and attempt
10044 to match them up, allowing for the possibility that the numbers may
10045 have changed due to creation and deletion in the meantime. If one of
10046 the target's tracepoints does not match any in @value{GDBN}, the
10047 debugger will create a new tracepoint, so that you have a number with
10048 which to specify that tracepoint. This matching-up process is
10049 necessarily heuristic, and it may result in useless tracepoints being
10050 created; you may simply delete them if they are of no use.
10052 @cindex circular trace buffer
10053 If your target agent supports a @dfn{circular trace buffer}, then you
10054 can run a trace experiment indefinitely without filling the trace
10055 buffer; when space runs out, the agent deletes already-collected trace
10056 frames, oldest first, until there is enough room to continue
10057 collecting. This is especially useful if your tracepoints are being
10058 hit too often, and your trace gets terminated prematurely because the
10059 buffer is full. To ask for a circular trace buffer, simply set
10060 @samp{circular_trace_buffer} to on. You can set this at any time,
10061 including during tracing; if the agent can do it, it will change
10062 buffer handling on the fly, otherwise it will not take effect until
10066 @item set circular-trace-buffer on
10067 @itemx set circular-trace-buffer off
10068 @kindex set circular-trace-buffer
10069 Choose whether a tracing run should use a linear or circular buffer
10070 for trace data. A linear buffer will not lose any trace data, but may
10071 fill up prematurely, while a circular buffer will discard old trace
10072 data, but it will have always room for the latest tracepoint hits.
10074 @item show circular-trace-buffer
10075 @kindex show circular-trace-buffer
10076 Show the current choice for the trace buffer. Note that this may not
10077 match the agent's current buffer handling, nor is it guaranteed to
10078 match the setting that might have been in effect during a past run,
10079 for instance if you are looking at frames from a trace file.
10083 @node Tracepoint Restrictions
10084 @subsection Tracepoint Restrictions
10086 @cindex tracepoint restrictions
10087 There are a number of restrictions on the use of tracepoints. As
10088 described above, tracepoint data gathering occurs on the target
10089 without interaction from @value{GDBN}. Thus the full capabilities of
10090 the debugger are not available during data gathering, and then at data
10091 examination time, you will be limited by only having what was
10092 collected. The following items describe some common problems, but it
10093 is not exhaustive, and you may run into additional difficulties not
10099 Tracepoint expressions are intended to gather objects (lvalues). Thus
10100 the full flexibility of GDB's expression evaluator is not available.
10101 You cannot call functions, cast objects to aggregate types, access
10102 convenience variables or modify values (except by assignment to trace
10103 state variables). Some language features may implicitly call
10104 functions (for instance Objective-C fields with accessors), and therefore
10105 cannot be collected either.
10108 Collection of local variables, either individually or in bulk with
10109 @code{$locals} or @code{$args}, during @code{while-stepping} may
10110 behave erratically. The stepping action may enter a new scope (for
10111 instance by stepping into a function), or the location of the variable
10112 may change (for instance it is loaded into a register). The
10113 tracepoint data recorded uses the location information for the
10114 variables that is correct for the tracepoint location. When the
10115 tracepoint is created, it is not possible, in general, to determine
10116 where the steps of a @code{while-stepping} sequence will advance the
10117 program---particularly if a conditional branch is stepped.
10120 Collection of an incompletely-initialized or partially-destroyed object
10121 may result in something that @value{GDBN} cannot display, or displays
10122 in a misleading way.
10125 When @value{GDBN} displays a pointer to character it automatically
10126 dereferences the pointer to also display characters of the string
10127 being pointed to. However, collecting the pointer during tracing does
10128 not automatically collect the string. You need to explicitly
10129 dereference the pointer and provide size information if you want to
10130 collect not only the pointer, but the memory pointed to. For example,
10131 @code{*ptr@@50} can be used to collect the 50 element array pointed to
10135 It is not possible to collect a complete stack backtrace at a
10136 tracepoint. Instead, you may collect the registers and a few hundred
10137 bytes from the stack pointer with something like @code{*$esp@@300}
10138 (adjust to use the name of the actual stack pointer register on your
10139 target architecture, and the amount of stack you wish to capture).
10140 Then the @code{backtrace} command will show a partial backtrace when
10141 using a trace frame. The number of stack frames that can be examined
10142 depends on the sizes of the frames in the collected stack. Note that
10143 if you ask for a block so large that it goes past the bottom of the
10144 stack, the target agent may report an error trying to read from an
10148 If you do not collect registers at a tracepoint, @value{GDBN} can
10149 infer that the value of @code{$pc} must be the same as the address of
10150 the tracepoint and use that when you are looking at a trace frame
10151 for that tracepoint. However, this cannot work if the tracepoint has
10152 multiple locations (for instance if it was set in a function that was
10153 inlined), or if it has a @code{while-stepping} loop. In those cases
10154 @value{GDBN} will warn you that it can't infer @code{$pc}, and default
10159 @node Analyze Collected Data
10160 @section Using the Collected Data
10162 After the tracepoint experiment ends, you use @value{GDBN} commands
10163 for examining the trace data. The basic idea is that each tracepoint
10164 collects a trace @dfn{snapshot} every time it is hit and another
10165 snapshot every time it single-steps. All these snapshots are
10166 consecutively numbered from zero and go into a buffer, and you can
10167 examine them later. The way you examine them is to @dfn{focus} on a
10168 specific trace snapshot. When the remote stub is focused on a trace
10169 snapshot, it will respond to all @value{GDBN} requests for memory and
10170 registers by reading from the buffer which belongs to that snapshot,
10171 rather than from @emph{real} memory or registers of the program being
10172 debugged. This means that @strong{all} @value{GDBN} commands
10173 (@code{print}, @code{info registers}, @code{backtrace}, etc.) will
10174 behave as if we were currently debugging the program state as it was
10175 when the tracepoint occurred. Any requests for data that are not in
10176 the buffer will fail.
10179 * tfind:: How to select a trace snapshot
10180 * tdump:: How to display all data for a snapshot
10181 * save tracepoints:: How to save tracepoints for a future run
10185 @subsection @code{tfind @var{n}}
10188 @cindex select trace snapshot
10189 @cindex find trace snapshot
10190 The basic command for selecting a trace snapshot from the buffer is
10191 @code{tfind @var{n}}, which finds trace snapshot number @var{n},
10192 counting from zero. If no argument @var{n} is given, the next
10193 snapshot is selected.
10195 Here are the various forms of using the @code{tfind} command.
10199 Find the first snapshot in the buffer. This is a synonym for
10200 @code{tfind 0} (since 0 is the number of the first snapshot).
10203 Stop debugging trace snapshots, resume @emph{live} debugging.
10206 Same as @samp{tfind none}.
10209 No argument means find the next trace snapshot.
10212 Find the previous trace snapshot before the current one. This permits
10213 retracing earlier steps.
10215 @item tfind tracepoint @var{num}
10216 Find the next snapshot associated with tracepoint @var{num}. Search
10217 proceeds forward from the last examined trace snapshot. If no
10218 argument @var{num} is given, it means find the next snapshot collected
10219 for the same tracepoint as the current snapshot.
10221 @item tfind pc @var{addr}
10222 Find the next snapshot associated with the value @var{addr} of the
10223 program counter. Search proceeds forward from the last examined trace
10224 snapshot. If no argument @var{addr} is given, it means find the next
10225 snapshot with the same value of PC as the current snapshot.
10227 @item tfind outside @var{addr1}, @var{addr2}
10228 Find the next snapshot whose PC is outside the given range of
10229 addresses (exclusive).
10231 @item tfind range @var{addr1}, @var{addr2}
10232 Find the next snapshot whose PC is between @var{addr1} and
10233 @var{addr2} (inclusive).
10235 @item tfind line @r{[}@var{file}:@r{]}@var{n}
10236 Find the next snapshot associated with the source line @var{n}. If
10237 the optional argument @var{file} is given, refer to line @var{n} in
10238 that source file. Search proceeds forward from the last examined
10239 trace snapshot. If no argument @var{n} is given, it means find the
10240 next line other than the one currently being examined; thus saying
10241 @code{tfind line} repeatedly can appear to have the same effect as
10242 stepping from line to line in a @emph{live} debugging session.
10245 The default arguments for the @code{tfind} commands are specifically
10246 designed to make it easy to scan through the trace buffer. For
10247 instance, @code{tfind} with no argument selects the next trace
10248 snapshot, and @code{tfind -} with no argument selects the previous
10249 trace snapshot. So, by giving one @code{tfind} command, and then
10250 simply hitting @key{RET} repeatedly you can examine all the trace
10251 snapshots in order. Or, by saying @code{tfind -} and then hitting
10252 @key{RET} repeatedly you can examine the snapshots in reverse order.
10253 The @code{tfind line} command with no argument selects the snapshot
10254 for the next source line executed. The @code{tfind pc} command with
10255 no argument selects the next snapshot with the same program counter
10256 (PC) as the current frame. The @code{tfind tracepoint} command with
10257 no argument selects the next trace snapshot collected by the same
10258 tracepoint as the current one.
10260 In addition to letting you scan through the trace buffer manually,
10261 these commands make it easy to construct @value{GDBN} scripts that
10262 scan through the trace buffer and print out whatever collected data
10263 you are interested in. Thus, if we want to examine the PC, FP, and SP
10264 registers from each trace frame in the buffer, we can say this:
10267 (@value{GDBP}) @b{tfind start}
10268 (@value{GDBP}) @b{while ($trace_frame != -1)}
10269 > printf "Frame %d, PC = %08X, SP = %08X, FP = %08X\n", \
10270 $trace_frame, $pc, $sp, $fp
10274 Frame 0, PC = 0020DC64, SP = 0030BF3C, FP = 0030BF44
10275 Frame 1, PC = 0020DC6C, SP = 0030BF38, FP = 0030BF44
10276 Frame 2, PC = 0020DC70, SP = 0030BF34, FP = 0030BF44
10277 Frame 3, PC = 0020DC74, SP = 0030BF30, FP = 0030BF44
10278 Frame 4, PC = 0020DC78, SP = 0030BF2C, FP = 0030BF44
10279 Frame 5, PC = 0020DC7C, SP = 0030BF28, FP = 0030BF44
10280 Frame 6, PC = 0020DC80, SP = 0030BF24, FP = 0030BF44
10281 Frame 7, PC = 0020DC84, SP = 0030BF20, FP = 0030BF44
10282 Frame 8, PC = 0020DC88, SP = 0030BF1C, FP = 0030BF44
10283 Frame 9, PC = 0020DC8E, SP = 0030BF18, FP = 0030BF44
10284 Frame 10, PC = 00203F6C, SP = 0030BE3C, FP = 0030BF14
10287 Or, if we want to examine the variable @code{X} at each source line in
10291 (@value{GDBP}) @b{tfind start}
10292 (@value{GDBP}) @b{while ($trace_frame != -1)}
10293 > printf "Frame %d, X == %d\n", $trace_frame, X
10303 @subsection @code{tdump}
10305 @cindex dump all data collected at tracepoint
10306 @cindex tracepoint data, display
10308 This command takes no arguments. It prints all the data collected at
10309 the current trace snapshot.
10312 (@value{GDBP}) @b{trace 444}
10313 (@value{GDBP}) @b{actions}
10314 Enter actions for tracepoint #2, one per line:
10315 > collect $regs, $locals, $args, gdb_long_test
10318 (@value{GDBP}) @b{tstart}
10320 (@value{GDBP}) @b{tfind line 444}
10321 #0 gdb_test (p1=0x11, p2=0x22, p3=0x33, p4=0x44, p5=0x55, p6=0x66)
10323 444 printp( "%s: arguments = 0x%X 0x%X 0x%X 0x%X 0x%X 0x%X\n", )
10325 (@value{GDBP}) @b{tdump}
10326 Data collected at tracepoint 2, trace frame 1:
10327 d0 0xc4aa0085 -995491707
10331 d4 0x71aea3d 119204413
10334 d7 0x380035 3670069
10335 a0 0x19e24a 1696330
10336 a1 0x3000668 50333288
10338 a3 0x322000 3284992
10339 a4 0x3000698 50333336
10340 a5 0x1ad3cc 1758156
10341 fp 0x30bf3c 0x30bf3c
10342 sp 0x30bf34 0x30bf34
10344 pc 0x20b2c8 0x20b2c8
10348 p = 0x20e5b4 "gdb-test"
10355 gdb_long_test = 17 '\021'
10360 @code{tdump} works by scanning the tracepoint's current collection
10361 actions and printing the value of each expression listed. So
10362 @code{tdump} can fail, if after a run, you change the tracepoint's
10363 actions to mention variables that were not collected during the run.
10365 Also, for tracepoints with @code{while-stepping} loops, @code{tdump}
10366 uses the collected value of @code{$pc} to distinguish between trace
10367 frames that were collected at the tracepoint hit, and frames that were
10368 collected while stepping. This allows it to correctly choose whether
10369 to display the basic list of collections, or the collections from the
10370 body of the while-stepping loop. However, if @code{$pc} was not collected,
10371 then @code{tdump} will always attempt to dump using the basic collection
10372 list, and may fail if a while-stepping frame does not include all the
10373 same data that is collected at the tracepoint hit.
10374 @c This is getting pretty arcane, example would be good.
10376 @node save tracepoints
10377 @subsection @code{save tracepoints @var{filename}}
10378 @kindex save tracepoints
10379 @kindex save-tracepoints
10380 @cindex save tracepoints for future sessions
10382 This command saves all current tracepoint definitions together with
10383 their actions and passcounts, into a file @file{@var{filename}}
10384 suitable for use in a later debugging session. To read the saved
10385 tracepoint definitions, use the @code{source} command (@pxref{Command
10386 Files}). The @w{@code{save-tracepoints}} command is a deprecated
10387 alias for @w{@code{save tracepoints}}
10389 @node Tracepoint Variables
10390 @section Convenience Variables for Tracepoints
10391 @cindex tracepoint variables
10392 @cindex convenience variables for tracepoints
10395 @vindex $trace_frame
10396 @item (int) $trace_frame
10397 The current trace snapshot (a.k.a.@: @dfn{frame}) number, or -1 if no
10398 snapshot is selected.
10400 @vindex $tracepoint
10401 @item (int) $tracepoint
10402 The tracepoint for the current trace snapshot.
10404 @vindex $trace_line
10405 @item (int) $trace_line
10406 The line number for the current trace snapshot.
10408 @vindex $trace_file
10409 @item (char []) $trace_file
10410 The source file for the current trace snapshot.
10412 @vindex $trace_func
10413 @item (char []) $trace_func
10414 The name of the function containing @code{$tracepoint}.
10417 Note: @code{$trace_file} is not suitable for use in @code{printf},
10418 use @code{output} instead.
10420 Here's a simple example of using these convenience variables for
10421 stepping through all the trace snapshots and printing some of their
10422 data. Note that these are not the same as trace state variables,
10423 which are managed by the target.
10426 (@value{GDBP}) @b{tfind start}
10428 (@value{GDBP}) @b{while $trace_frame != -1}
10429 > output $trace_file
10430 > printf ", line %d (tracepoint #%d)\n", $trace_line, $tracepoint
10436 @section Using Trace Files
10437 @cindex trace files
10439 In some situations, the target running a trace experiment may no
10440 longer be available; perhaps it crashed, or the hardware was needed
10441 for a different activity. To handle these cases, you can arrange to
10442 dump the trace data into a file, and later use that file as a source
10443 of trace data, via the @code{target tfile} command.
10448 @item tsave [ -r ] @var{filename}
10449 Save the trace data to @var{filename}. By default, this command
10450 assumes that @var{filename} refers to the host filesystem, so if
10451 necessary @value{GDBN} will copy raw trace data up from the target and
10452 then save it. If the target supports it, you can also supply the
10453 optional argument @code{-r} (``remote'') to direct the target to save
10454 the data directly into @var{filename} in its own filesystem, which may be
10455 more efficient if the trace buffer is very large. (Note, however, that
10456 @code{target tfile} can only read from files accessible to the host.)
10458 @kindex target tfile
10460 @item target tfile @var{filename}
10461 Use the file named @var{filename} as a source of trace data. Commands
10462 that examine data work as they do with a live target, but it is not
10463 possible to run any new trace experiments. @code{tstatus} will report
10464 the state of the trace run at the moment the data was saved, as well
10465 as the current trace frame you are examining. @var{filename} must be
10466 on a filesystem accessible to the host.
10471 @chapter Debugging Programs That Use Overlays
10474 If your program is too large to fit completely in your target system's
10475 memory, you can sometimes use @dfn{overlays} to work around this
10476 problem. @value{GDBN} provides some support for debugging programs that
10480 * How Overlays Work:: A general explanation of overlays.
10481 * Overlay Commands:: Managing overlays in @value{GDBN}.
10482 * Automatic Overlay Debugging:: @value{GDBN} can find out which overlays are
10483 mapped by asking the inferior.
10484 * Overlay Sample Program:: A sample program using overlays.
10487 @node How Overlays Work
10488 @section How Overlays Work
10489 @cindex mapped overlays
10490 @cindex unmapped overlays
10491 @cindex load address, overlay's
10492 @cindex mapped address
10493 @cindex overlay area
10495 Suppose you have a computer whose instruction address space is only 64
10496 kilobytes long, but which has much more memory which can be accessed by
10497 other means: special instructions, segment registers, or memory
10498 management hardware, for example. Suppose further that you want to
10499 adapt a program which is larger than 64 kilobytes to run on this system.
10501 One solution is to identify modules of your program which are relatively
10502 independent, and need not call each other directly; call these modules
10503 @dfn{overlays}. Separate the overlays from the main program, and place
10504 their machine code in the larger memory. Place your main program in
10505 instruction memory, but leave at least enough space there to hold the
10506 largest overlay as well.
10508 Now, to call a function located in an overlay, you must first copy that
10509 overlay's machine code from the large memory into the space set aside
10510 for it in the instruction memory, and then jump to its entry point
10513 @c NB: In the below the mapped area's size is greater or equal to the
10514 @c size of all overlays. This is intentional to remind the developer
10515 @c that overlays don't necessarily need to be the same size.
10519 Data Instruction Larger
10520 Address Space Address Space Address Space
10521 +-----------+ +-----------+ +-----------+
10523 +-----------+ +-----------+ +-----------+<-- overlay 1
10524 | program | | main | .----| overlay 1 | load address
10525 | variables | | program | | +-----------+
10526 | and heap | | | | | |
10527 +-----------+ | | | +-----------+<-- overlay 2
10528 | | +-----------+ | | | load address
10529 +-----------+ | | | .-| overlay 2 |
10531 mapped --->+-----------+ | | +-----------+
10532 address | | | | | |
10533 | overlay | <-' | | |
10534 | area | <---' +-----------+<-- overlay 3
10535 | | <---. | | load address
10536 +-----------+ `--| overlay 3 |
10543 @anchor{A code overlay}A code overlay
10547 The diagram (@pxref{A code overlay}) shows a system with separate data
10548 and instruction address spaces. To map an overlay, the program copies
10549 its code from the larger address space to the instruction address space.
10550 Since the overlays shown here all use the same mapped address, only one
10551 may be mapped at a time. For a system with a single address space for
10552 data and instructions, the diagram would be similar, except that the
10553 program variables and heap would share an address space with the main
10554 program and the overlay area.
10556 An overlay loaded into instruction memory and ready for use is called a
10557 @dfn{mapped} overlay; its @dfn{mapped address} is its address in the
10558 instruction memory. An overlay not present (or only partially present)
10559 in instruction memory is called @dfn{unmapped}; its @dfn{load address}
10560 is its address in the larger memory. The mapped address is also called
10561 the @dfn{virtual memory address}, or @dfn{VMA}; the load address is also
10562 called the @dfn{load memory address}, or @dfn{LMA}.
10564 Unfortunately, overlays are not a completely transparent way to adapt a
10565 program to limited instruction memory. They introduce a new set of
10566 global constraints you must keep in mind as you design your program:
10571 Before calling or returning to a function in an overlay, your program
10572 must make sure that overlay is actually mapped. Otherwise, the call or
10573 return will transfer control to the right address, but in the wrong
10574 overlay, and your program will probably crash.
10577 If the process of mapping an overlay is expensive on your system, you
10578 will need to choose your overlays carefully to minimize their effect on
10579 your program's performance.
10582 The executable file you load onto your system must contain each
10583 overlay's instructions, appearing at the overlay's load address, not its
10584 mapped address. However, each overlay's instructions must be relocated
10585 and its symbols defined as if the overlay were at its mapped address.
10586 You can use GNU linker scripts to specify different load and relocation
10587 addresses for pieces of your program; see @ref{Overlay Description,,,
10588 ld.info, Using ld: the GNU linker}.
10591 The procedure for loading executable files onto your system must be able
10592 to load their contents into the larger address space as well as the
10593 instruction and data spaces.
10597 The overlay system described above is rather simple, and could be
10598 improved in many ways:
10603 If your system has suitable bank switch registers or memory management
10604 hardware, you could use those facilities to make an overlay's load area
10605 contents simply appear at their mapped address in instruction space.
10606 This would probably be faster than copying the overlay to its mapped
10607 area in the usual way.
10610 If your overlays are small enough, you could set aside more than one
10611 overlay area, and have more than one overlay mapped at a time.
10614 You can use overlays to manage data, as well as instructions. In
10615 general, data overlays are even less transparent to your design than
10616 code overlays: whereas code overlays only require care when you call or
10617 return to functions, data overlays require care every time you access
10618 the data. Also, if you change the contents of a data overlay, you
10619 must copy its contents back out to its load address before you can copy a
10620 different data overlay into the same mapped area.
10625 @node Overlay Commands
10626 @section Overlay Commands
10628 To use @value{GDBN}'s overlay support, each overlay in your program must
10629 correspond to a separate section of the executable file. The section's
10630 virtual memory address and load memory address must be the overlay's
10631 mapped and load addresses. Identifying overlays with sections allows
10632 @value{GDBN} to determine the appropriate address of a function or
10633 variable, depending on whether the overlay is mapped or not.
10635 @value{GDBN}'s overlay commands all start with the word @code{overlay};
10636 you can abbreviate this as @code{ov} or @code{ovly}. The commands are:
10641 Disable @value{GDBN}'s overlay support. When overlay support is
10642 disabled, @value{GDBN} assumes that all functions and variables are
10643 always present at their mapped addresses. By default, @value{GDBN}'s
10644 overlay support is disabled.
10646 @item overlay manual
10647 @cindex manual overlay debugging
10648 Enable @dfn{manual} overlay debugging. In this mode, @value{GDBN}
10649 relies on you to tell it which overlays are mapped, and which are not,
10650 using the @code{overlay map-overlay} and @code{overlay unmap-overlay}
10651 commands described below.
10653 @item overlay map-overlay @var{overlay}
10654 @itemx overlay map @var{overlay}
10655 @cindex map an overlay
10656 Tell @value{GDBN} that @var{overlay} is now mapped; @var{overlay} must
10657 be the name of the object file section containing the overlay. When an
10658 overlay is mapped, @value{GDBN} assumes it can find the overlay's
10659 functions and variables at their mapped addresses. @value{GDBN} assumes
10660 that any other overlays whose mapped ranges overlap that of
10661 @var{overlay} are now unmapped.
10663 @item overlay unmap-overlay @var{overlay}
10664 @itemx overlay unmap @var{overlay}
10665 @cindex unmap an overlay
10666 Tell @value{GDBN} that @var{overlay} is no longer mapped; @var{overlay}
10667 must be the name of the object file section containing the overlay.
10668 When an overlay is unmapped, @value{GDBN} assumes it can find the
10669 overlay's functions and variables at their load addresses.
10672 Enable @dfn{automatic} overlay debugging. In this mode, @value{GDBN}
10673 consults a data structure the overlay manager maintains in the inferior
10674 to see which overlays are mapped. For details, see @ref{Automatic
10675 Overlay Debugging}.
10677 @item overlay load-target
10678 @itemx overlay load
10679 @cindex reloading the overlay table
10680 Re-read the overlay table from the inferior. Normally, @value{GDBN}
10681 re-reads the table @value{GDBN} automatically each time the inferior
10682 stops, so this command should only be necessary if you have changed the
10683 overlay mapping yourself using @value{GDBN}. This command is only
10684 useful when using automatic overlay debugging.
10686 @item overlay list-overlays
10687 @itemx overlay list
10688 @cindex listing mapped overlays
10689 Display a list of the overlays currently mapped, along with their mapped
10690 addresses, load addresses, and sizes.
10694 Normally, when @value{GDBN} prints a code address, it includes the name
10695 of the function the address falls in:
10698 (@value{GDBP}) print main
10699 $3 = @{int ()@} 0x11a0 <main>
10702 When overlay debugging is enabled, @value{GDBN} recognizes code in
10703 unmapped overlays, and prints the names of unmapped functions with
10704 asterisks around them. For example, if @code{foo} is a function in an
10705 unmapped overlay, @value{GDBN} prints it this way:
10708 (@value{GDBP}) overlay list
10709 No sections are mapped.
10710 (@value{GDBP}) print foo
10711 $5 = @{int (int)@} 0x100000 <*foo*>
10714 When @code{foo}'s overlay is mapped, @value{GDBN} prints the function's
10718 (@value{GDBP}) overlay list
10719 Section .ov.foo.text, loaded at 0x100000 - 0x100034,
10720 mapped at 0x1016 - 0x104a
10721 (@value{GDBP}) print foo
10722 $6 = @{int (int)@} 0x1016 <foo>
10725 When overlay debugging is enabled, @value{GDBN} can find the correct
10726 address for functions and variables in an overlay, whether or not the
10727 overlay is mapped. This allows most @value{GDBN} commands, like
10728 @code{break} and @code{disassemble}, to work normally, even on unmapped
10729 code. However, @value{GDBN}'s breakpoint support has some limitations:
10733 @cindex breakpoints in overlays
10734 @cindex overlays, setting breakpoints in
10735 You can set breakpoints in functions in unmapped overlays, as long as
10736 @value{GDBN} can write to the overlay at its load address.
10738 @value{GDBN} can not set hardware or simulator-based breakpoints in
10739 unmapped overlays. However, if you set a breakpoint at the end of your
10740 overlay manager (and tell @value{GDBN} which overlays are now mapped, if
10741 you are using manual overlay management), @value{GDBN} will re-set its
10742 breakpoints properly.
10746 @node Automatic Overlay Debugging
10747 @section Automatic Overlay Debugging
10748 @cindex automatic overlay debugging
10750 @value{GDBN} can automatically track which overlays are mapped and which
10751 are not, given some simple co-operation from the overlay manager in the
10752 inferior. If you enable automatic overlay debugging with the
10753 @code{overlay auto} command (@pxref{Overlay Commands}), @value{GDBN}
10754 looks in the inferior's memory for certain variables describing the
10755 current state of the overlays.
10757 Here are the variables your overlay manager must define to support
10758 @value{GDBN}'s automatic overlay debugging:
10762 @item @code{_ovly_table}:
10763 This variable must be an array of the following structures:
10768 /* The overlay's mapped address. */
10771 /* The size of the overlay, in bytes. */
10772 unsigned long size;
10774 /* The overlay's load address. */
10777 /* Non-zero if the overlay is currently mapped;
10779 unsigned long mapped;
10783 @item @code{_novlys}:
10784 This variable must be a four-byte signed integer, holding the total
10785 number of elements in @code{_ovly_table}.
10789 To decide whether a particular overlay is mapped or not, @value{GDBN}
10790 looks for an entry in @w{@code{_ovly_table}} whose @code{vma} and
10791 @code{lma} members equal the VMA and LMA of the overlay's section in the
10792 executable file. When @value{GDBN} finds a matching entry, it consults
10793 the entry's @code{mapped} member to determine whether the overlay is
10796 In addition, your overlay manager may define a function called
10797 @code{_ovly_debug_event}. If this function is defined, @value{GDBN}
10798 will silently set a breakpoint there. If the overlay manager then
10799 calls this function whenever it has changed the overlay table, this
10800 will enable @value{GDBN} to accurately keep track of which overlays
10801 are in program memory, and update any breakpoints that may be set
10802 in overlays. This will allow breakpoints to work even if the
10803 overlays are kept in ROM or other non-writable memory while they
10804 are not being executed.
10806 @node Overlay Sample Program
10807 @section Overlay Sample Program
10808 @cindex overlay example program
10810 When linking a program which uses overlays, you must place the overlays
10811 at their load addresses, while relocating them to run at their mapped
10812 addresses. To do this, you must write a linker script (@pxref{Overlay
10813 Description,,, ld.info, Using ld: the GNU linker}). Unfortunately,
10814 since linker scripts are specific to a particular host system, target
10815 architecture, and target memory layout, this manual cannot provide
10816 portable sample code demonstrating @value{GDBN}'s overlay support.
10818 However, the @value{GDBN} source distribution does contain an overlaid
10819 program, with linker scripts for a few systems, as part of its test
10820 suite. The program consists of the following files from
10821 @file{gdb/testsuite/gdb.base}:
10825 The main program file.
10827 A simple overlay manager, used by @file{overlays.c}.
10832 Overlay modules, loaded and used by @file{overlays.c}.
10835 Linker scripts for linking the test program on the @code{d10v-elf}
10836 and @code{m32r-elf} targets.
10839 You can build the test program using the @code{d10v-elf} GCC
10840 cross-compiler like this:
10843 $ d10v-elf-gcc -g -c overlays.c
10844 $ d10v-elf-gcc -g -c ovlymgr.c
10845 $ d10v-elf-gcc -g -c foo.c
10846 $ d10v-elf-gcc -g -c bar.c
10847 $ d10v-elf-gcc -g -c baz.c
10848 $ d10v-elf-gcc -g -c grbx.c
10849 $ d10v-elf-gcc -g overlays.o ovlymgr.o foo.o bar.o \
10850 baz.o grbx.o -Wl,-Td10v.ld -o overlays
10853 The build process is identical for any other architecture, except that
10854 you must substitute the appropriate compiler and linker script for the
10855 target system for @code{d10v-elf-gcc} and @code{d10v.ld}.
10859 @chapter Using @value{GDBN} with Different Languages
10862 Although programming languages generally have common aspects, they are
10863 rarely expressed in the same manner. For instance, in ANSI C,
10864 dereferencing a pointer @code{p} is accomplished by @code{*p}, but in
10865 Modula-2, it is accomplished by @code{p^}. Values can also be
10866 represented (and displayed) differently. Hex numbers in C appear as
10867 @samp{0x1ae}, while in Modula-2 they appear as @samp{1AEH}.
10869 @cindex working language
10870 Language-specific information is built into @value{GDBN} for some languages,
10871 allowing you to express operations like the above in your program's
10872 native language, and allowing @value{GDBN} to output values in a manner
10873 consistent with the syntax of your program's native language. The
10874 language you use to build expressions is called the @dfn{working
10878 * Setting:: Switching between source languages
10879 * Show:: Displaying the language
10880 * Checks:: Type and range checks
10881 * Supported Languages:: Supported languages
10882 * Unsupported Languages:: Unsupported languages
10886 @section Switching Between Source Languages
10888 There are two ways to control the working language---either have @value{GDBN}
10889 set it automatically, or select it manually yourself. You can use the
10890 @code{set language} command for either purpose. On startup, @value{GDBN}
10891 defaults to setting the language automatically. The working language is
10892 used to determine how expressions you type are interpreted, how values
10895 In addition to the working language, every source file that
10896 @value{GDBN} knows about has its own working language. For some object
10897 file formats, the compiler might indicate which language a particular
10898 source file is in. However, most of the time @value{GDBN} infers the
10899 language from the name of the file. The language of a source file
10900 controls whether C@t{++} names are demangled---this way @code{backtrace} can
10901 show each frame appropriately for its own language. There is no way to
10902 set the language of a source file from within @value{GDBN}, but you can
10903 set the language associated with a filename extension. @xref{Show, ,
10904 Displaying the Language}.
10906 This is most commonly a problem when you use a program, such
10907 as @code{cfront} or @code{f2c}, that generates C but is written in
10908 another language. In that case, make the
10909 program use @code{#line} directives in its C output; that way
10910 @value{GDBN} will know the correct language of the source code of the original
10911 program, and will display that source code, not the generated C code.
10914 * Filenames:: Filename extensions and languages.
10915 * Manually:: Setting the working language manually
10916 * Automatically:: Having @value{GDBN} infer the source language
10920 @subsection List of Filename Extensions and Languages
10922 If a source file name ends in one of the following extensions, then
10923 @value{GDBN} infers that its language is the one indicated.
10941 C@t{++} source file
10947 Objective-C source file
10951 Fortran source file
10954 Modula-2 source file
10958 Assembler source file. This actually behaves almost like C, but
10959 @value{GDBN} does not skip over function prologues when stepping.
10962 In addition, you may set the language associated with a filename
10963 extension. @xref{Show, , Displaying the Language}.
10966 @subsection Setting the Working Language
10968 If you allow @value{GDBN} to set the language automatically,
10969 expressions are interpreted the same way in your debugging session and
10972 @kindex set language
10973 If you wish, you may set the language manually. To do this, issue the
10974 command @samp{set language @var{lang}}, where @var{lang} is the name of
10975 a language, such as
10976 @code{c} or @code{modula-2}.
10977 For a list of the supported languages, type @samp{set language}.
10979 Setting the language manually prevents @value{GDBN} from updating the working
10980 language automatically. This can lead to confusion if you try
10981 to debug a program when the working language is not the same as the
10982 source language, when an expression is acceptable to both
10983 languages---but means different things. For instance, if the current
10984 source file were written in C, and @value{GDBN} was parsing Modula-2, a
10992 might not have the effect you intended. In C, this means to add
10993 @code{b} and @code{c} and place the result in @code{a}. The result
10994 printed would be the value of @code{a}. In Modula-2, this means to compare
10995 @code{a} to the result of @code{b+c}, yielding a @code{BOOLEAN} value.
10997 @node Automatically
10998 @subsection Having @value{GDBN} Infer the Source Language
11000 To have @value{GDBN} set the working language automatically, use
11001 @samp{set language local} or @samp{set language auto}. @value{GDBN}
11002 then infers the working language. That is, when your program stops in a
11003 frame (usually by encountering a breakpoint), @value{GDBN} sets the
11004 working language to the language recorded for the function in that
11005 frame. If the language for a frame is unknown (that is, if the function
11006 or block corresponding to the frame was defined in a source file that
11007 does not have a recognized extension), the current working language is
11008 not changed, and @value{GDBN} issues a warning.
11010 This may not seem necessary for most programs, which are written
11011 entirely in one source language. However, program modules and libraries
11012 written in one source language can be used by a main program written in
11013 a different source language. Using @samp{set language auto} in this
11014 case frees you from having to set the working language manually.
11017 @section Displaying the Language
11019 The following commands help you find out which language is the
11020 working language, and also what language source files were written in.
11023 @item show language
11024 @kindex show language
11025 Display the current working language. This is the
11026 language you can use with commands such as @code{print} to
11027 build and compute expressions that may involve variables in your program.
11030 @kindex info frame@r{, show the source language}
11031 Display the source language for this frame. This language becomes the
11032 working language if you use an identifier from this frame.
11033 @xref{Frame Info, ,Information about a Frame}, to identify the other
11034 information listed here.
11037 @kindex info source@r{, show the source language}
11038 Display the source language of this source file.
11039 @xref{Symbols, ,Examining the Symbol Table}, to identify the other
11040 information listed here.
11043 In unusual circumstances, you may have source files with extensions
11044 not in the standard list. You can then set the extension associated
11045 with a language explicitly:
11048 @item set extension-language @var{ext} @var{language}
11049 @kindex set extension-language
11050 Tell @value{GDBN} that source files with extension @var{ext} are to be
11051 assumed as written in the source language @var{language}.
11053 @item info extensions
11054 @kindex info extensions
11055 List all the filename extensions and the associated languages.
11059 @section Type and Range Checking
11062 @emph{Warning:} In this release, the @value{GDBN} commands for type and range
11063 checking are included, but they do not yet have any effect. This
11064 section documents the intended facilities.
11066 @c FIXME remove warning when type/range code added
11068 Some languages are designed to guard you against making seemingly common
11069 errors through a series of compile- and run-time checks. These include
11070 checking the type of arguments to functions and operators, and making
11071 sure mathematical overflows are caught at run time. Checks such as
11072 these help to ensure a program's correctness once it has been compiled
11073 by eliminating type mismatches, and providing active checks for range
11074 errors when your program is running.
11076 @value{GDBN} can check for conditions like the above if you wish.
11077 Although @value{GDBN} does not check the statements in your program,
11078 it can check expressions entered directly into @value{GDBN} for
11079 evaluation via the @code{print} command, for example. As with the
11080 working language, @value{GDBN} can also decide whether or not to check
11081 automatically based on your program's source language.
11082 @xref{Supported Languages, ,Supported Languages}, for the default
11083 settings of supported languages.
11086 * Type Checking:: An overview of type checking
11087 * Range Checking:: An overview of range checking
11090 @cindex type checking
11091 @cindex checks, type
11092 @node Type Checking
11093 @subsection An Overview of Type Checking
11095 Some languages, such as Modula-2, are strongly typed, meaning that the
11096 arguments to operators and functions have to be of the correct type,
11097 otherwise an error occurs. These checks prevent type mismatch
11098 errors from ever causing any run-time problems. For example,
11106 The second example fails because the @code{CARDINAL} 1 is not
11107 type-compatible with the @code{REAL} 2.3.
11109 For the expressions you use in @value{GDBN} commands, you can tell the
11110 @value{GDBN} type checker to skip checking;
11111 to treat any mismatches as errors and abandon the expression;
11112 or to only issue warnings when type mismatches occur,
11113 but evaluate the expression anyway. When you choose the last of
11114 these, @value{GDBN} evaluates expressions like the second example above, but
11115 also issues a warning.
11117 Even if you turn type checking off, there may be other reasons
11118 related to type that prevent @value{GDBN} from evaluating an expression.
11119 For instance, @value{GDBN} does not know how to add an @code{int} and
11120 a @code{struct foo}. These particular type errors have nothing to do
11121 with the language in use, and usually arise from expressions, such as
11122 the one described above, which make little sense to evaluate anyway.
11124 Each language defines to what degree it is strict about type. For
11125 instance, both Modula-2 and C require the arguments to arithmetical
11126 operators to be numbers. In C, enumerated types and pointers can be
11127 represented as numbers, so that they are valid arguments to mathematical
11128 operators. @xref{Supported Languages, ,Supported Languages}, for further
11129 details on specific languages.
11131 @value{GDBN} provides some additional commands for controlling the type checker:
11133 @kindex set check type
11134 @kindex show check type
11136 @item set check type auto
11137 Set type checking on or off based on the current working language.
11138 @xref{Supported Languages, ,Supported Languages}, for the default settings for
11141 @item set check type on
11142 @itemx set check type off
11143 Set type checking on or off, overriding the default setting for the
11144 current working language. Issue a warning if the setting does not
11145 match the language default. If any type mismatches occur in
11146 evaluating an expression while type checking is on, @value{GDBN} prints a
11147 message and aborts evaluation of the expression.
11149 @item set check type warn
11150 Cause the type checker to issue warnings, but to always attempt to
11151 evaluate the expression. Evaluating the expression may still
11152 be impossible for other reasons. For example, @value{GDBN} cannot add
11153 numbers and structures.
11156 Show the current setting of the type checker, and whether or not @value{GDBN}
11157 is setting it automatically.
11160 @cindex range checking
11161 @cindex checks, range
11162 @node Range Checking
11163 @subsection An Overview of Range Checking
11165 In some languages (such as Modula-2), it is an error to exceed the
11166 bounds of a type; this is enforced with run-time checks. Such range
11167 checking is meant to ensure program correctness by making sure
11168 computations do not overflow, or indices on an array element access do
11169 not exceed the bounds of the array.
11171 For expressions you use in @value{GDBN} commands, you can tell
11172 @value{GDBN} to treat range errors in one of three ways: ignore them,
11173 always treat them as errors and abandon the expression, or issue
11174 warnings but evaluate the expression anyway.
11176 A range error can result from numerical overflow, from exceeding an
11177 array index bound, or when you type a constant that is not a member
11178 of any type. Some languages, however, do not treat overflows as an
11179 error. In many implementations of C, mathematical overflow causes the
11180 result to ``wrap around'' to lower values---for example, if @var{m} is
11181 the largest integer value, and @var{s} is the smallest, then
11184 @var{m} + 1 @result{} @var{s}
11187 This, too, is specific to individual languages, and in some cases
11188 specific to individual compilers or machines. @xref{Supported Languages, ,
11189 Supported Languages}, for further details on specific languages.
11191 @value{GDBN} provides some additional commands for controlling the range checker:
11193 @kindex set check range
11194 @kindex show check range
11196 @item set check range auto
11197 Set range checking on or off based on the current working language.
11198 @xref{Supported Languages, ,Supported Languages}, for the default settings for
11201 @item set check range on
11202 @itemx set check range off
11203 Set range checking on or off, overriding the default setting for the
11204 current working language. A warning is issued if the setting does not
11205 match the language default. If a range error occurs and range checking is on,
11206 then a message is printed and evaluation of the expression is aborted.
11208 @item set check range warn
11209 Output messages when the @value{GDBN} range checker detects a range error,
11210 but attempt to evaluate the expression anyway. Evaluating the
11211 expression may still be impossible for other reasons, such as accessing
11212 memory that the process does not own (a typical example from many Unix
11216 Show the current setting of the range checker, and whether or not it is
11217 being set automatically by @value{GDBN}.
11220 @node Supported Languages
11221 @section Supported Languages
11223 @value{GDBN} supports C, C@t{++}, D, Objective-C, Fortran, Java, Pascal,
11224 assembly, Modula-2, and Ada.
11225 @c This is false ...
11226 Some @value{GDBN} features may be used in expressions regardless of the
11227 language you use: the @value{GDBN} @code{@@} and @code{::} operators,
11228 and the @samp{@{type@}addr} construct (@pxref{Expressions,
11229 ,Expressions}) can be used with the constructs of any supported
11232 The following sections detail to what degree each source language is
11233 supported by @value{GDBN}. These sections are not meant to be language
11234 tutorials or references, but serve only as a reference guide to what the
11235 @value{GDBN} expression parser accepts, and what input and output
11236 formats should look like for different languages. There are many good
11237 books written on each of these languages; please look to these for a
11238 language reference or tutorial.
11241 * C:: C and C@t{++}
11243 * Objective-C:: Objective-C
11244 * Fortran:: Fortran
11246 * Modula-2:: Modula-2
11251 @subsection C and C@t{++}
11253 @cindex C and C@t{++}
11254 @cindex expressions in C or C@t{++}
11256 Since C and C@t{++} are so closely related, many features of @value{GDBN} apply
11257 to both languages. Whenever this is the case, we discuss those languages
11261 @cindex @code{g++}, @sc{gnu} C@t{++} compiler
11262 @cindex @sc{gnu} C@t{++}
11263 The C@t{++} debugging facilities are jointly implemented by the C@t{++}
11264 compiler and @value{GDBN}. Therefore, to debug your C@t{++} code
11265 effectively, you must compile your C@t{++} programs with a supported
11266 C@t{++} compiler, such as @sc{gnu} @code{g++}, or the HP ANSI C@t{++}
11267 compiler (@code{aCC}).
11269 For best results when using @sc{gnu} C@t{++}, use the DWARF 2 debugging
11270 format; if it doesn't work on your system, try the stabs+ debugging
11271 format. You can select those formats explicitly with the @code{g++}
11272 command-line options @option{-gdwarf-2} and @option{-gstabs+}.
11273 @xref{Debugging Options,,Options for Debugging Your Program or GCC,
11274 gcc.info, Using the @sc{gnu} Compiler Collection (GCC)}.
11277 * C Operators:: C and C@t{++} operators
11278 * C Constants:: C and C@t{++} constants
11279 * C Plus Plus Expressions:: C@t{++} expressions
11280 * C Defaults:: Default settings for C and C@t{++}
11281 * C Checks:: C and C@t{++} type and range checks
11282 * Debugging C:: @value{GDBN} and C
11283 * Debugging C Plus Plus:: @value{GDBN} features for C@t{++}
11284 * Decimal Floating Point:: Numbers in Decimal Floating Point format
11288 @subsubsection C and C@t{++} Operators
11290 @cindex C and C@t{++} operators
11292 Operators must be defined on values of specific types. For instance,
11293 @code{+} is defined on numbers, but not on structures. Operators are
11294 often defined on groups of types.
11296 For the purposes of C and C@t{++}, the following definitions hold:
11301 @emph{Integral types} include @code{int} with any of its storage-class
11302 specifiers; @code{char}; @code{enum}; and, for C@t{++}, @code{bool}.
11305 @emph{Floating-point types} include @code{float}, @code{double}, and
11306 @code{long double} (if supported by the target platform).
11309 @emph{Pointer types} include all types defined as @code{(@var{type} *)}.
11312 @emph{Scalar types} include all of the above.
11317 The following operators are supported. They are listed here
11318 in order of increasing precedence:
11322 The comma or sequencing operator. Expressions in a comma-separated list
11323 are evaluated from left to right, with the result of the entire
11324 expression being the last expression evaluated.
11327 Assignment. The value of an assignment expression is the value
11328 assigned. Defined on scalar types.
11331 Used in an expression of the form @w{@code{@var{a} @var{op}= @var{b}}},
11332 and translated to @w{@code{@var{a} = @var{a op b}}}.
11333 @w{@code{@var{op}=}} and @code{=} have the same precedence.
11334 @var{op} is any one of the operators @code{|}, @code{^}, @code{&},
11335 @code{<<}, @code{>>}, @code{+}, @code{-}, @code{*}, @code{/}, @code{%}.
11338 The ternary operator. @code{@var{a} ? @var{b} : @var{c}} can be thought
11339 of as: if @var{a} then @var{b} else @var{c}. @var{a} should be of an
11343 Logical @sc{or}. Defined on integral types.
11346 Logical @sc{and}. Defined on integral types.
11349 Bitwise @sc{or}. Defined on integral types.
11352 Bitwise exclusive-@sc{or}. Defined on integral types.
11355 Bitwise @sc{and}. Defined on integral types.
11358 Equality and inequality. Defined on scalar types. The value of these
11359 expressions is 0 for false and non-zero for true.
11361 @item <@r{, }>@r{, }<=@r{, }>=
11362 Less than, greater than, less than or equal, greater than or equal.
11363 Defined on scalar types. The value of these expressions is 0 for false
11364 and non-zero for true.
11367 left shift, and right shift. Defined on integral types.
11370 The @value{GDBN} ``artificial array'' operator (@pxref{Expressions, ,Expressions}).
11373 Addition and subtraction. Defined on integral types, floating-point types and
11376 @item *@r{, }/@r{, }%
11377 Multiplication, division, and modulus. Multiplication and division are
11378 defined on integral and floating-point types. Modulus is defined on
11382 Increment and decrement. When appearing before a variable, the
11383 operation is performed before the variable is used in an expression;
11384 when appearing after it, the variable's value is used before the
11385 operation takes place.
11388 Pointer dereferencing. Defined on pointer types. Same precedence as
11392 Address operator. Defined on variables. Same precedence as @code{++}.
11394 For debugging C@t{++}, @value{GDBN} implements a use of @samp{&} beyond what is
11395 allowed in the C@t{++} language itself: you can use @samp{&(&@var{ref})}
11396 to examine the address
11397 where a C@t{++} reference variable (declared with @samp{&@var{ref}}) is
11401 Negative. Defined on integral and floating-point types. Same
11402 precedence as @code{++}.
11405 Logical negation. Defined on integral types. Same precedence as
11409 Bitwise complement operator. Defined on integral types. Same precedence as
11414 Structure member, and pointer-to-structure member. For convenience,
11415 @value{GDBN} regards the two as equivalent, choosing whether to dereference a
11416 pointer based on the stored type information.
11417 Defined on @code{struct} and @code{union} data.
11420 Dereferences of pointers to members.
11423 Array indexing. @code{@var{a}[@var{i}]} is defined as
11424 @code{*(@var{a}+@var{i})}. Same precedence as @code{->}.
11427 Function parameter list. Same precedence as @code{->}.
11430 C@t{++} scope resolution operator. Defined on @code{struct}, @code{union},
11431 and @code{class} types.
11434 Doubled colons also represent the @value{GDBN} scope operator
11435 (@pxref{Expressions, ,Expressions}). Same precedence as @code{::},
11439 If an operator is redefined in the user code, @value{GDBN} usually
11440 attempts to invoke the redefined version instead of using the operator's
11441 predefined meaning.
11444 @subsubsection C and C@t{++} Constants
11446 @cindex C and C@t{++} constants
11448 @value{GDBN} allows you to express the constants of C and C@t{++} in the
11453 Integer constants are a sequence of digits. Octal constants are
11454 specified by a leading @samp{0} (i.e.@: zero), and hexadecimal constants
11455 by a leading @samp{0x} or @samp{0X}. Constants may also end with a letter
11456 @samp{l}, specifying that the constant should be treated as a
11460 Floating point constants are a sequence of digits, followed by a decimal
11461 point, followed by a sequence of digits, and optionally followed by an
11462 exponent. An exponent is of the form:
11463 @samp{@w{e@r{[[}+@r{]|}-@r{]}@var{nnn}}}, where @var{nnn} is another
11464 sequence of digits. The @samp{+} is optional for positive exponents.
11465 A floating-point constant may also end with a letter @samp{f} or
11466 @samp{F}, specifying that the constant should be treated as being of
11467 the @code{float} (as opposed to the default @code{double}) type; or with
11468 a letter @samp{l} or @samp{L}, which specifies a @code{long double}
11472 Enumerated constants consist of enumerated identifiers, or their
11473 integral equivalents.
11476 Character constants are a single character surrounded by single quotes
11477 (@code{'}), or a number---the ordinal value of the corresponding character
11478 (usually its @sc{ascii} value). Within quotes, the single character may
11479 be represented by a letter or by @dfn{escape sequences}, which are of
11480 the form @samp{\@var{nnn}}, where @var{nnn} is the octal representation
11481 of the character's ordinal value; or of the form @samp{\@var{x}}, where
11482 @samp{@var{x}} is a predefined special character---for example,
11483 @samp{\n} for newline.
11486 String constants are a sequence of character constants surrounded by
11487 double quotes (@code{"}). Any valid character constant (as described
11488 above) may appear. Double quotes within the string must be preceded by
11489 a backslash, so for instance @samp{"a\"b'c"} is a string of five
11493 Pointer constants are an integral value. You can also write pointers
11494 to constants using the C operator @samp{&}.
11497 Array constants are comma-separated lists surrounded by braces @samp{@{}
11498 and @samp{@}}; for example, @samp{@{1,2,3@}} is a three-element array of
11499 integers, @samp{@{@{1,2@}, @{3,4@}, @{5,6@}@}} is a three-by-two array,
11500 and @samp{@{&"hi", &"there", &"fred"@}} is a three-element array of pointers.
11503 @node C Plus Plus Expressions
11504 @subsubsection C@t{++} Expressions
11506 @cindex expressions in C@t{++}
11507 @value{GDBN} expression handling can interpret most C@t{++} expressions.
11509 @cindex debugging C@t{++} programs
11510 @cindex C@t{++} compilers
11511 @cindex debug formats and C@t{++}
11512 @cindex @value{NGCC} and C@t{++}
11514 @emph{Warning:} @value{GDBN} can only debug C@t{++} code if you use the
11515 proper compiler and the proper debug format. Currently, @value{GDBN}
11516 works best when debugging C@t{++} code that is compiled with
11517 @value{NGCC} 2.95.3 or with @value{NGCC} 3.1 or newer, using the options
11518 @option{-gdwarf-2} or @option{-gstabs+}. DWARF 2 is preferred over
11519 stabs+. Most configurations of @value{NGCC} emit either DWARF 2 or
11520 stabs+ as their default debug format, so you usually don't need to
11521 specify a debug format explicitly. Other compilers and/or debug formats
11522 are likely to work badly or not at all when using @value{GDBN} to debug
11528 @cindex member functions
11530 Member function calls are allowed; you can use expressions like
11533 count = aml->GetOriginal(x, y)
11536 @vindex this@r{, inside C@t{++} member functions}
11537 @cindex namespace in C@t{++}
11539 While a member function is active (in the selected stack frame), your
11540 expressions have the same namespace available as the member function;
11541 that is, @value{GDBN} allows implicit references to the class instance
11542 pointer @code{this} following the same rules as C@t{++}.
11544 @cindex call overloaded functions
11545 @cindex overloaded functions, calling
11546 @cindex type conversions in C@t{++}
11548 You can call overloaded functions; @value{GDBN} resolves the function
11549 call to the right definition, with some restrictions. @value{GDBN} does not
11550 perform overload resolution involving user-defined type conversions,
11551 calls to constructors, or instantiations of templates that do not exist
11552 in the program. It also cannot handle ellipsis argument lists or
11555 It does perform integral conversions and promotions, floating-point
11556 promotions, arithmetic conversions, pointer conversions, conversions of
11557 class objects to base classes, and standard conversions such as those of
11558 functions or arrays to pointers; it requires an exact match on the
11559 number of function arguments.
11561 Overload resolution is always performed, unless you have specified
11562 @code{set overload-resolution off}. @xref{Debugging C Plus Plus,
11563 ,@value{GDBN} Features for C@t{++}}.
11565 You must specify @code{set overload-resolution off} in order to use an
11566 explicit function signature to call an overloaded function, as in
11568 p 'foo(char,int)'('x', 13)
11571 The @value{GDBN} command-completion facility can simplify this;
11572 see @ref{Completion, ,Command Completion}.
11574 @cindex reference declarations
11576 @value{GDBN} understands variables declared as C@t{++} references; you can use
11577 them in expressions just as you do in C@t{++} source---they are automatically
11580 In the parameter list shown when @value{GDBN} displays a frame, the values of
11581 reference variables are not displayed (unlike other variables); this
11582 avoids clutter, since references are often used for large structures.
11583 The @emph{address} of a reference variable is always shown, unless
11584 you have specified @samp{set print address off}.
11587 @value{GDBN} supports the C@t{++} name resolution operator @code{::}---your
11588 expressions can use it just as expressions in your program do. Since
11589 one scope may be defined in another, you can use @code{::} repeatedly if
11590 necessary, for example in an expression like
11591 @samp{@var{scope1}::@var{scope2}::@var{name}}. @value{GDBN} also allows
11592 resolving name scope by reference to source files, in both C and C@t{++}
11593 debugging (@pxref{Variables, ,Program Variables}).
11596 In addition, when used with HP's C@t{++} compiler, @value{GDBN} supports
11597 calling virtual functions correctly, printing out virtual bases of
11598 objects, calling functions in a base subobject, casting objects, and
11599 invoking user-defined operators.
11602 @subsubsection C and C@t{++} Defaults
11604 @cindex C and C@t{++} defaults
11606 If you allow @value{GDBN} to set type and range checking automatically, they
11607 both default to @code{off} whenever the working language changes to
11608 C or C@t{++}. This happens regardless of whether you or @value{GDBN}
11609 selects the working language.
11611 If you allow @value{GDBN} to set the language automatically, it
11612 recognizes source files whose names end with @file{.c}, @file{.C}, or
11613 @file{.cc}, etc, and when @value{GDBN} enters code compiled from one of
11614 these files, it sets the working language to C or C@t{++}.
11615 @xref{Automatically, ,Having @value{GDBN} Infer the Source Language},
11616 for further details.
11618 @c Type checking is (a) primarily motivated by Modula-2, and (b)
11619 @c unimplemented. If (b) changes, it might make sense to let this node
11620 @c appear even if Mod-2 does not, but meanwhile ignore it. roland 16jul93.
11623 @subsubsection C and C@t{++} Type and Range Checks
11625 @cindex C and C@t{++} checks
11627 By default, when @value{GDBN} parses C or C@t{++} expressions, type checking
11628 is not used. However, if you turn type checking on, @value{GDBN}
11629 considers two variables type equivalent if:
11633 The two variables are structured and have the same structure, union, or
11637 The two variables have the same type name, or types that have been
11638 declared equivalent through @code{typedef}.
11641 @c leaving this out because neither J Gilmore nor R Pesch understand it.
11644 The two @code{struct}, @code{union}, or @code{enum} variables are
11645 declared in the same declaration. (Note: this may not be true for all C
11650 Range checking, if turned on, is done on mathematical operations. Array
11651 indices are not checked, since they are often used to index a pointer
11652 that is not itself an array.
11655 @subsubsection @value{GDBN} and C
11657 The @code{set print union} and @code{show print union} commands apply to
11658 the @code{union} type. When set to @samp{on}, any @code{union} that is
11659 inside a @code{struct} or @code{class} is also printed. Otherwise, it
11660 appears as @samp{@{...@}}.
11662 The @code{@@} operator aids in the debugging of dynamic arrays, formed
11663 with pointers and a memory allocation function. @xref{Expressions,
11666 @node Debugging C Plus Plus
11667 @subsubsection @value{GDBN} Features for C@t{++}
11669 @cindex commands for C@t{++}
11671 Some @value{GDBN} commands are particularly useful with C@t{++}, and some are
11672 designed specifically for use with C@t{++}. Here is a summary:
11675 @cindex break in overloaded functions
11676 @item @r{breakpoint menus}
11677 When you want a breakpoint in a function whose name is overloaded,
11678 @value{GDBN} has the capability to display a menu of possible breakpoint
11679 locations to help you specify which function definition you want.
11680 @xref{Ambiguous Expressions,,Ambiguous Expressions}.
11682 @cindex overloading in C@t{++}
11683 @item rbreak @var{regex}
11684 Setting breakpoints using regular expressions is helpful for setting
11685 breakpoints on overloaded functions that are not members of any special
11687 @xref{Set Breaks, ,Setting Breakpoints}.
11689 @cindex C@t{++} exception handling
11692 Debug C@t{++} exception handling using these commands. @xref{Set
11693 Catchpoints, , Setting Catchpoints}.
11695 @cindex inheritance
11696 @item ptype @var{typename}
11697 Print inheritance relationships as well as other information for type
11699 @xref{Symbols, ,Examining the Symbol Table}.
11701 @cindex C@t{++} symbol display
11702 @item set print demangle
11703 @itemx show print demangle
11704 @itemx set print asm-demangle
11705 @itemx show print asm-demangle
11706 Control whether C@t{++} symbols display in their source form, both when
11707 displaying code as C@t{++} source and when displaying disassemblies.
11708 @xref{Print Settings, ,Print Settings}.
11710 @item set print object
11711 @itemx show print object
11712 Choose whether to print derived (actual) or declared types of objects.
11713 @xref{Print Settings, ,Print Settings}.
11715 @item set print vtbl
11716 @itemx show print vtbl
11717 Control the format for printing virtual function tables.
11718 @xref{Print Settings, ,Print Settings}.
11719 (The @code{vtbl} commands do not work on programs compiled with the HP
11720 ANSI C@t{++} compiler (@code{aCC}).)
11722 @kindex set overload-resolution
11723 @cindex overloaded functions, overload resolution
11724 @item set overload-resolution on
11725 Enable overload resolution for C@t{++} expression evaluation. The default
11726 is on. For overloaded functions, @value{GDBN} evaluates the arguments
11727 and searches for a function whose signature matches the argument types,
11728 using the standard C@t{++} conversion rules (see @ref{C Plus Plus
11729 Expressions, ,C@t{++} Expressions}, for details).
11730 If it cannot find a match, it emits a message.
11732 @item set overload-resolution off
11733 Disable overload resolution for C@t{++} expression evaluation. For
11734 overloaded functions that are not class member functions, @value{GDBN}
11735 chooses the first function of the specified name that it finds in the
11736 symbol table, whether or not its arguments are of the correct type. For
11737 overloaded functions that are class member functions, @value{GDBN}
11738 searches for a function whose signature @emph{exactly} matches the
11741 @kindex show overload-resolution
11742 @item show overload-resolution
11743 Show the current setting of overload resolution.
11745 @item @r{Overloaded symbol names}
11746 You can specify a particular definition of an overloaded symbol, using
11747 the same notation that is used to declare such symbols in C@t{++}: type
11748 @code{@var{symbol}(@var{types})} rather than just @var{symbol}. You can
11749 also use the @value{GDBN} command-line word completion facilities to list the
11750 available choices, or to finish the type list for you.
11751 @xref{Completion,, Command Completion}, for details on how to do this.
11754 @node Decimal Floating Point
11755 @subsubsection Decimal Floating Point format
11756 @cindex decimal floating point format
11758 @value{GDBN} can examine, set and perform computations with numbers in
11759 decimal floating point format, which in the C language correspond to the
11760 @code{_Decimal32}, @code{_Decimal64} and @code{_Decimal128} types as
11761 specified by the extension to support decimal floating-point arithmetic.
11763 There are two encodings in use, depending on the architecture: BID (Binary
11764 Integer Decimal) for x86 and x86-64, and DPD (Densely Packed Decimal) for
11765 PowerPC. @value{GDBN} will use the appropriate encoding for the configured
11768 Because of a limitation in @file{libdecnumber}, the library used by @value{GDBN}
11769 to manipulate decimal floating point numbers, it is not possible to convert
11770 (using a cast, for example) integers wider than 32-bit to decimal float.
11772 In addition, in order to imitate @value{GDBN}'s behaviour with binary floating
11773 point computations, error checking in decimal float operations ignores
11774 underflow, overflow and divide by zero exceptions.
11776 In the PowerPC architecture, @value{GDBN} provides a set of pseudo-registers
11777 to inspect @code{_Decimal128} values stored in floating point registers.
11778 See @ref{PowerPC,,PowerPC} for more details.
11784 @value{GDBN} can be used to debug programs written in D and compiled with
11785 GDC, LDC or DMD compilers. Currently @value{GDBN} supports only one D
11786 specific feature --- dynamic arrays.
11789 @subsection Objective-C
11791 @cindex Objective-C
11792 This section provides information about some commands and command
11793 options that are useful for debugging Objective-C code. See also
11794 @ref{Symbols, info classes}, and @ref{Symbols, info selectors}, for a
11795 few more commands specific to Objective-C support.
11798 * Method Names in Commands::
11799 * The Print Command with Objective-C::
11802 @node Method Names in Commands
11803 @subsubsection Method Names in Commands
11805 The following commands have been extended to accept Objective-C method
11806 names as line specifications:
11808 @kindex clear@r{, and Objective-C}
11809 @kindex break@r{, and Objective-C}
11810 @kindex info line@r{, and Objective-C}
11811 @kindex jump@r{, and Objective-C}
11812 @kindex list@r{, and Objective-C}
11816 @item @code{info line}
11821 A fully qualified Objective-C method name is specified as
11824 -[@var{Class} @var{methodName}]
11827 where the minus sign is used to indicate an instance method and a
11828 plus sign (not shown) is used to indicate a class method. The class
11829 name @var{Class} and method name @var{methodName} are enclosed in
11830 brackets, similar to the way messages are specified in Objective-C
11831 source code. For example, to set a breakpoint at the @code{create}
11832 instance method of class @code{Fruit} in the program currently being
11836 break -[Fruit create]
11839 To list ten program lines around the @code{initialize} class method,
11843 list +[NSText initialize]
11846 In the current version of @value{GDBN}, the plus or minus sign is
11847 required. In future versions of @value{GDBN}, the plus or minus
11848 sign will be optional, but you can use it to narrow the search. It
11849 is also possible to specify just a method name:
11855 You must specify the complete method name, including any colons. If
11856 your program's source files contain more than one @code{create} method,
11857 you'll be presented with a numbered list of classes that implement that
11858 method. Indicate your choice by number, or type @samp{0} to exit if
11861 As another example, to clear a breakpoint established at the
11862 @code{makeKeyAndOrderFront:} method of the @code{NSWindow} class, enter:
11865 clear -[NSWindow makeKeyAndOrderFront:]
11868 @node The Print Command with Objective-C
11869 @subsubsection The Print Command With Objective-C
11870 @cindex Objective-C, print objects
11871 @kindex print-object
11872 @kindex po @r{(@code{print-object})}
11874 The print command has also been extended to accept methods. For example:
11877 print -[@var{object} hash]
11880 @cindex print an Objective-C object description
11881 @cindex @code{_NSPrintForDebugger}, and printing Objective-C objects
11883 will tell @value{GDBN} to send the @code{hash} message to @var{object}
11884 and print the result. Also, an additional command has been added,
11885 @code{print-object} or @code{po} for short, which is meant to print
11886 the description of an object. However, this command may only work
11887 with certain Objective-C libraries that have a particular hook
11888 function, @code{_NSPrintForDebugger}, defined.
11891 @subsection Fortran
11892 @cindex Fortran-specific support in @value{GDBN}
11894 @value{GDBN} can be used to debug programs written in Fortran, but it
11895 currently supports only the features of Fortran 77 language.
11897 @cindex trailing underscore, in Fortran symbols
11898 Some Fortran compilers (@sc{gnu} Fortran 77 and Fortran 95 compilers
11899 among them) append an underscore to the names of variables and
11900 functions. When you debug programs compiled by those compilers, you
11901 will need to refer to variables and functions with a trailing
11905 * Fortran Operators:: Fortran operators and expressions
11906 * Fortran Defaults:: Default settings for Fortran
11907 * Special Fortran Commands:: Special @value{GDBN} commands for Fortran
11910 @node Fortran Operators
11911 @subsubsection Fortran Operators and Expressions
11913 @cindex Fortran operators and expressions
11915 Operators must be defined on values of specific types. For instance,
11916 @code{+} is defined on numbers, but not on characters or other non-
11917 arithmetic types. Operators are often defined on groups of types.
11921 The exponentiation operator. It raises the first operand to the power
11925 The range operator. Normally used in the form of array(low:high) to
11926 represent a section of array.
11929 The access component operator. Normally used to access elements in derived
11930 types. Also suitable for unions. As unions aren't part of regular Fortran,
11931 this can only happen when accessing a register that uses a gdbarch-defined
11935 @node Fortran Defaults
11936 @subsubsection Fortran Defaults
11938 @cindex Fortran Defaults
11940 Fortran symbols are usually case-insensitive, so @value{GDBN} by
11941 default uses case-insensitive matches for Fortran symbols. You can
11942 change that with the @samp{set case-insensitive} command, see
11943 @ref{Symbols}, for the details.
11945 @node Special Fortran Commands
11946 @subsubsection Special Fortran Commands
11948 @cindex Special Fortran commands
11950 @value{GDBN} has some commands to support Fortran-specific features,
11951 such as displaying common blocks.
11954 @cindex @code{COMMON} blocks, Fortran
11955 @kindex info common
11956 @item info common @r{[}@var{common-name}@r{]}
11957 This command prints the values contained in the Fortran @code{COMMON}
11958 block whose name is @var{common-name}. With no argument, the names of
11959 all @code{COMMON} blocks visible at the current program location are
11966 @cindex Pascal support in @value{GDBN}, limitations
11967 Debugging Pascal programs which use sets, subranges, file variables, or
11968 nested functions does not currently work. @value{GDBN} does not support
11969 entering expressions, printing values, or similar features using Pascal
11972 The Pascal-specific command @code{set print pascal_static-members}
11973 controls whether static members of Pascal objects are displayed.
11974 @xref{Print Settings, pascal_static-members}.
11977 @subsection Modula-2
11979 @cindex Modula-2, @value{GDBN} support
11981 The extensions made to @value{GDBN} to support Modula-2 only support
11982 output from the @sc{gnu} Modula-2 compiler (which is currently being
11983 developed). Other Modula-2 compilers are not currently supported, and
11984 attempting to debug executables produced by them is most likely
11985 to give an error as @value{GDBN} reads in the executable's symbol
11988 @cindex expressions in Modula-2
11990 * M2 Operators:: Built-in operators
11991 * Built-In Func/Proc:: Built-in functions and procedures
11992 * M2 Constants:: Modula-2 constants
11993 * M2 Types:: Modula-2 types
11994 * M2 Defaults:: Default settings for Modula-2
11995 * Deviations:: Deviations from standard Modula-2
11996 * M2 Checks:: Modula-2 type and range checks
11997 * M2 Scope:: The scope operators @code{::} and @code{.}
11998 * GDB/M2:: @value{GDBN} and Modula-2
12002 @subsubsection Operators
12003 @cindex Modula-2 operators
12005 Operators must be defined on values of specific types. For instance,
12006 @code{+} is defined on numbers, but not on structures. Operators are
12007 often defined on groups of types. For the purposes of Modula-2, the
12008 following definitions hold:
12013 @emph{Integral types} consist of @code{INTEGER}, @code{CARDINAL}, and
12017 @emph{Character types} consist of @code{CHAR} and its subranges.
12020 @emph{Floating-point types} consist of @code{REAL}.
12023 @emph{Pointer types} consist of anything declared as @code{POINTER TO
12027 @emph{Scalar types} consist of all of the above.
12030 @emph{Set types} consist of @code{SET} and @code{BITSET} types.
12033 @emph{Boolean types} consist of @code{BOOLEAN}.
12037 The following operators are supported, and appear in order of
12038 increasing precedence:
12042 Function argument or array index separator.
12045 Assignment. The value of @var{var} @code{:=} @var{value} is
12049 Less than, greater than on integral, floating-point, or enumerated
12053 Less than or equal to, greater than or equal to
12054 on integral, floating-point and enumerated types, or set inclusion on
12055 set types. Same precedence as @code{<}.
12057 @item =@r{, }<>@r{, }#
12058 Equality and two ways of expressing inequality, valid on scalar types.
12059 Same precedence as @code{<}. In @value{GDBN} scripts, only @code{<>} is
12060 available for inequality, since @code{#} conflicts with the script
12064 Set membership. Defined on set types and the types of their members.
12065 Same precedence as @code{<}.
12068 Boolean disjunction. Defined on boolean types.
12071 Boolean conjunction. Defined on boolean types.
12074 The @value{GDBN} ``artificial array'' operator (@pxref{Expressions, ,Expressions}).
12077 Addition and subtraction on integral and floating-point types, or union
12078 and difference on set types.
12081 Multiplication on integral and floating-point types, or set intersection
12085 Division on floating-point types, or symmetric set difference on set
12086 types. Same precedence as @code{*}.
12089 Integer division and remainder. Defined on integral types. Same
12090 precedence as @code{*}.
12093 Negative. Defined on @code{INTEGER} and @code{REAL} data.
12096 Pointer dereferencing. Defined on pointer types.
12099 Boolean negation. Defined on boolean types. Same precedence as
12103 @code{RECORD} field selector. Defined on @code{RECORD} data. Same
12104 precedence as @code{^}.
12107 Array indexing. Defined on @code{ARRAY} data. Same precedence as @code{^}.
12110 Procedure argument list. Defined on @code{PROCEDURE} objects. Same precedence
12114 @value{GDBN} and Modula-2 scope operators.
12118 @emph{Warning:} Set expressions and their operations are not yet supported, so @value{GDBN}
12119 treats the use of the operator @code{IN}, or the use of operators
12120 @code{+}, @code{-}, @code{*}, @code{/}, @code{=}, , @code{<>}, @code{#},
12121 @code{<=}, and @code{>=} on sets as an error.
12125 @node Built-In Func/Proc
12126 @subsubsection Built-in Functions and Procedures
12127 @cindex Modula-2 built-ins
12129 Modula-2 also makes available several built-in procedures and functions.
12130 In describing these, the following metavariables are used:
12135 represents an @code{ARRAY} variable.
12138 represents a @code{CHAR} constant or variable.
12141 represents a variable or constant of integral type.
12144 represents an identifier that belongs to a set. Generally used in the
12145 same function with the metavariable @var{s}. The type of @var{s} should
12146 be @code{SET OF @var{mtype}} (where @var{mtype} is the type of @var{m}).
12149 represents a variable or constant of integral or floating-point type.
12152 represents a variable or constant of floating-point type.
12158 represents a variable.
12161 represents a variable or constant of one of many types. See the
12162 explanation of the function for details.
12165 All Modula-2 built-in procedures also return a result, described below.
12169 Returns the absolute value of @var{n}.
12172 If @var{c} is a lower case letter, it returns its upper case
12173 equivalent, otherwise it returns its argument.
12176 Returns the character whose ordinal value is @var{i}.
12179 Decrements the value in the variable @var{v} by one. Returns the new value.
12181 @item DEC(@var{v},@var{i})
12182 Decrements the value in the variable @var{v} by @var{i}. Returns the
12185 @item EXCL(@var{m},@var{s})
12186 Removes the element @var{m} from the set @var{s}. Returns the new
12189 @item FLOAT(@var{i})
12190 Returns the floating point equivalent of the integer @var{i}.
12192 @item HIGH(@var{a})
12193 Returns the index of the last member of @var{a}.
12196 Increments the value in the variable @var{v} by one. Returns the new value.
12198 @item INC(@var{v},@var{i})
12199 Increments the value in the variable @var{v} by @var{i}. Returns the
12202 @item INCL(@var{m},@var{s})
12203 Adds the element @var{m} to the set @var{s} if it is not already
12204 there. Returns the new set.
12207 Returns the maximum value of the type @var{t}.
12210 Returns the minimum value of the type @var{t}.
12213 Returns boolean TRUE if @var{i} is an odd number.
12216 Returns the ordinal value of its argument. For example, the ordinal
12217 value of a character is its @sc{ascii} value (on machines supporting the
12218 @sc{ascii} character set). @var{x} must be of an ordered type, which include
12219 integral, character and enumerated types.
12221 @item SIZE(@var{x})
12222 Returns the size of its argument. @var{x} can be a variable or a type.
12224 @item TRUNC(@var{r})
12225 Returns the integral part of @var{r}.
12227 @item TSIZE(@var{x})
12228 Returns the size of its argument. @var{x} can be a variable or a type.
12230 @item VAL(@var{t},@var{i})
12231 Returns the member of the type @var{t} whose ordinal value is @var{i}.
12235 @emph{Warning:} Sets and their operations are not yet supported, so
12236 @value{GDBN} treats the use of procedures @code{INCL} and @code{EXCL} as
12240 @cindex Modula-2 constants
12242 @subsubsection Constants
12244 @value{GDBN} allows you to express the constants of Modula-2 in the following
12250 Integer constants are simply a sequence of digits. When used in an
12251 expression, a constant is interpreted to be type-compatible with the
12252 rest of the expression. Hexadecimal integers are specified by a
12253 trailing @samp{H}, and octal integers by a trailing @samp{B}.
12256 Floating point constants appear as a sequence of digits, followed by a
12257 decimal point and another sequence of digits. An optional exponent can
12258 then be specified, in the form @samp{E@r{[}+@r{|}-@r{]}@var{nnn}}, where
12259 @samp{@r{[}+@r{|}-@r{]}@var{nnn}} is the desired exponent. All of the
12260 digits of the floating point constant must be valid decimal (base 10)
12264 Character constants consist of a single character enclosed by a pair of
12265 like quotes, either single (@code{'}) or double (@code{"}). They may
12266 also be expressed by their ordinal value (their @sc{ascii} value, usually)
12267 followed by a @samp{C}.
12270 String constants consist of a sequence of characters enclosed by a
12271 pair of like quotes, either single (@code{'}) or double (@code{"}).
12272 Escape sequences in the style of C are also allowed. @xref{C
12273 Constants, ,C and C@t{++} Constants}, for a brief explanation of escape
12277 Enumerated constants consist of an enumerated identifier.
12280 Boolean constants consist of the identifiers @code{TRUE} and
12284 Pointer constants consist of integral values only.
12287 Set constants are not yet supported.
12291 @subsubsection Modula-2 Types
12292 @cindex Modula-2 types
12294 Currently @value{GDBN} can print the following data types in Modula-2
12295 syntax: array types, record types, set types, pointer types, procedure
12296 types, enumerated types, subrange types and base types. You can also
12297 print the contents of variables declared using these type.
12298 This section gives a number of simple source code examples together with
12299 sample @value{GDBN} sessions.
12301 The first example contains the following section of code:
12310 and you can request @value{GDBN} to interrogate the type and value of
12311 @code{r} and @code{s}.
12314 (@value{GDBP}) print s
12316 (@value{GDBP}) ptype s
12318 (@value{GDBP}) print r
12320 (@value{GDBP}) ptype r
12325 Likewise if your source code declares @code{s} as:
12329 s: SET ['A'..'Z'] ;
12333 then you may query the type of @code{s} by:
12336 (@value{GDBP}) ptype s
12337 type = SET ['A'..'Z']
12341 Note that at present you cannot interactively manipulate set
12342 expressions using the debugger.
12344 The following example shows how you might declare an array in Modula-2
12345 and how you can interact with @value{GDBN} to print its type and contents:
12349 s: ARRAY [-10..10] OF CHAR ;
12353 (@value{GDBP}) ptype s
12354 ARRAY [-10..10] OF CHAR
12357 Note that the array handling is not yet complete and although the type
12358 is printed correctly, expression handling still assumes that all
12359 arrays have a lower bound of zero and not @code{-10} as in the example
12362 Here are some more type related Modula-2 examples:
12366 colour = (blue, red, yellow, green) ;
12367 t = [blue..yellow] ;
12375 The @value{GDBN} interaction shows how you can query the data type
12376 and value of a variable.
12379 (@value{GDBP}) print s
12381 (@value{GDBP}) ptype t
12382 type = [blue..yellow]
12386 In this example a Modula-2 array is declared and its contents
12387 displayed. Observe that the contents are written in the same way as
12388 their @code{C} counterparts.
12392 s: ARRAY [1..5] OF CARDINAL ;
12398 (@value{GDBP}) print s
12399 $1 = @{1, 0, 0, 0, 0@}
12400 (@value{GDBP}) ptype s
12401 type = ARRAY [1..5] OF CARDINAL
12404 The Modula-2 language interface to @value{GDBN} also understands
12405 pointer types as shown in this example:
12409 s: POINTER TO ARRAY [1..5] OF CARDINAL ;
12416 and you can request that @value{GDBN} describes the type of @code{s}.
12419 (@value{GDBP}) ptype s
12420 type = POINTER TO ARRAY [1..5] OF CARDINAL
12423 @value{GDBN} handles compound types as we can see in this example.
12424 Here we combine array types, record types, pointer types and subrange
12435 myarray = ARRAY myrange OF CARDINAL ;
12436 myrange = [-2..2] ;
12438 s: POINTER TO ARRAY myrange OF foo ;
12442 and you can ask @value{GDBN} to describe the type of @code{s} as shown
12446 (@value{GDBP}) ptype s
12447 type = POINTER TO ARRAY [-2..2] OF foo = RECORD
12450 f3 : ARRAY [-2..2] OF CARDINAL;
12455 @subsubsection Modula-2 Defaults
12456 @cindex Modula-2 defaults
12458 If type and range checking are set automatically by @value{GDBN}, they
12459 both default to @code{on} whenever the working language changes to
12460 Modula-2. This happens regardless of whether you or @value{GDBN}
12461 selected the working language.
12463 If you allow @value{GDBN} to set the language automatically, then entering
12464 code compiled from a file whose name ends with @file{.mod} sets the
12465 working language to Modula-2. @xref{Automatically, ,Having @value{GDBN}
12466 Infer the Source Language}, for further details.
12469 @subsubsection Deviations from Standard Modula-2
12470 @cindex Modula-2, deviations from
12472 A few changes have been made to make Modula-2 programs easier to debug.
12473 This is done primarily via loosening its type strictness:
12477 Unlike in standard Modula-2, pointer constants can be formed by
12478 integers. This allows you to modify pointer variables during
12479 debugging. (In standard Modula-2, the actual address contained in a
12480 pointer variable is hidden from you; it can only be modified
12481 through direct assignment to another pointer variable or expression that
12482 returned a pointer.)
12485 C escape sequences can be used in strings and characters to represent
12486 non-printable characters. @value{GDBN} prints out strings with these
12487 escape sequences embedded. Single non-printable characters are
12488 printed using the @samp{CHR(@var{nnn})} format.
12491 The assignment operator (@code{:=}) returns the value of its right-hand
12495 All built-in procedures both modify @emph{and} return their argument.
12499 @subsubsection Modula-2 Type and Range Checks
12500 @cindex Modula-2 checks
12503 @emph{Warning:} in this release, @value{GDBN} does not yet perform type or
12506 @c FIXME remove warning when type/range checks added
12508 @value{GDBN} considers two Modula-2 variables type equivalent if:
12512 They are of types that have been declared equivalent via a @code{TYPE
12513 @var{t1} = @var{t2}} statement
12516 They have been declared on the same line. (Note: This is true of the
12517 @sc{gnu} Modula-2 compiler, but it may not be true of other compilers.)
12520 As long as type checking is enabled, any attempt to combine variables
12521 whose types are not equivalent is an error.
12523 Range checking is done on all mathematical operations, assignment, array
12524 index bounds, and all built-in functions and procedures.
12527 @subsubsection The Scope Operators @code{::} and @code{.}
12529 @cindex @code{.}, Modula-2 scope operator
12530 @cindex colon, doubled as scope operator
12532 @vindex colon-colon@r{, in Modula-2}
12533 @c Info cannot handle :: but TeX can.
12536 @vindex ::@r{, in Modula-2}
12539 There are a few subtle differences between the Modula-2 scope operator
12540 (@code{.}) and the @value{GDBN} scope operator (@code{::}). The two have
12545 @var{module} . @var{id}
12546 @var{scope} :: @var{id}
12550 where @var{scope} is the name of a module or a procedure,
12551 @var{module} the name of a module, and @var{id} is any declared
12552 identifier within your program, except another module.
12554 Using the @code{::} operator makes @value{GDBN} search the scope
12555 specified by @var{scope} for the identifier @var{id}. If it is not
12556 found in the specified scope, then @value{GDBN} searches all scopes
12557 enclosing the one specified by @var{scope}.
12559 Using the @code{.} operator makes @value{GDBN} search the current scope for
12560 the identifier specified by @var{id} that was imported from the
12561 definition module specified by @var{module}. With this operator, it is
12562 an error if the identifier @var{id} was not imported from definition
12563 module @var{module}, or if @var{id} is not an identifier in
12567 @subsubsection @value{GDBN} and Modula-2
12569 Some @value{GDBN} commands have little use when debugging Modula-2 programs.
12570 Five subcommands of @code{set print} and @code{show print} apply
12571 specifically to C and C@t{++}: @samp{vtbl}, @samp{demangle},
12572 @samp{asm-demangle}, @samp{object}, and @samp{union}. The first four
12573 apply to C@t{++}, and the last to the C @code{union} type, which has no direct
12574 analogue in Modula-2.
12576 The @code{@@} operator (@pxref{Expressions, ,Expressions}), while available
12577 with any language, is not useful with Modula-2. Its
12578 intent is to aid the debugging of @dfn{dynamic arrays}, which cannot be
12579 created in Modula-2 as they can in C or C@t{++}. However, because an
12580 address can be specified by an integral constant, the construct
12581 @samp{@{@var{type}@}@var{adrexp}} is still useful.
12583 @cindex @code{#} in Modula-2
12584 In @value{GDBN} scripts, the Modula-2 inequality operator @code{#} is
12585 interpreted as the beginning of a comment. Use @code{<>} instead.
12591 The extensions made to @value{GDBN} for Ada only support
12592 output from the @sc{gnu} Ada (GNAT) compiler.
12593 Other Ada compilers are not currently supported, and
12594 attempting to debug executables produced by them is most likely
12598 @cindex expressions in Ada
12600 * Ada Mode Intro:: General remarks on the Ada syntax
12601 and semantics supported by Ada mode
12603 * Omissions from Ada:: Restrictions on the Ada expression syntax.
12604 * Additions to Ada:: Extensions of the Ada expression syntax.
12605 * Stopping Before Main Program:: Debugging the program during elaboration.
12606 * Ada Tasks:: Listing and setting breakpoints in tasks.
12607 * Ada Tasks and Core Files:: Tasking Support when Debugging Core Files
12608 * Ada Glitches:: Known peculiarities of Ada mode.
12611 @node Ada Mode Intro
12612 @subsubsection Introduction
12613 @cindex Ada mode, general
12615 The Ada mode of @value{GDBN} supports a fairly large subset of Ada expression
12616 syntax, with some extensions.
12617 The philosophy behind the design of this subset is
12621 That @value{GDBN} should provide basic literals and access to operations for
12622 arithmetic, dereferencing, field selection, indexing, and subprogram calls,
12623 leaving more sophisticated computations to subprograms written into the
12624 program (which therefore may be called from @value{GDBN}).
12627 That type safety and strict adherence to Ada language restrictions
12628 are not particularly important to the @value{GDBN} user.
12631 That brevity is important to the @value{GDBN} user.
12634 Thus, for brevity, the debugger acts as if all names declared in
12635 user-written packages are directly visible, even if they are not visible
12636 according to Ada rules, thus making it unnecessary to fully qualify most
12637 names with their packages, regardless of context. Where this causes
12638 ambiguity, @value{GDBN} asks the user's intent.
12640 The debugger will start in Ada mode if it detects an Ada main program.
12641 As for other languages, it will enter Ada mode when stopped in a program that
12642 was translated from an Ada source file.
12644 While in Ada mode, you may use `@t{--}' for comments. This is useful
12645 mostly for documenting command files. The standard @value{GDBN} comment
12646 (@samp{#}) still works at the beginning of a line in Ada mode, but not in the
12647 middle (to allow based literals).
12649 The debugger supports limited overloading. Given a subprogram call in which
12650 the function symbol has multiple definitions, it will use the number of
12651 actual parameters and some information about their types to attempt to narrow
12652 the set of definitions. It also makes very limited use of context, preferring
12653 procedures to functions in the context of the @code{call} command, and
12654 functions to procedures elsewhere.
12656 @node Omissions from Ada
12657 @subsubsection Omissions from Ada
12658 @cindex Ada, omissions from
12660 Here are the notable omissions from the subset:
12664 Only a subset of the attributes are supported:
12668 @t{'First}, @t{'Last}, and @t{'Length}
12669 on array objects (not on types and subtypes).
12672 @t{'Min} and @t{'Max}.
12675 @t{'Pos} and @t{'Val}.
12681 @t{'Range} on array objects (not subtypes), but only as the right
12682 operand of the membership (@code{in}) operator.
12685 @t{'Access}, @t{'Unchecked_Access}, and
12686 @t{'Unrestricted_Access} (a GNAT extension).
12694 @code{Characters.Latin_1} are not available and
12695 concatenation is not implemented. Thus, escape characters in strings are
12696 not currently available.
12699 Equality tests (@samp{=} and @samp{/=}) on arrays test for bitwise
12700 equality of representations. They will generally work correctly
12701 for strings and arrays whose elements have integer or enumeration types.
12702 They may not work correctly for arrays whose element
12703 types have user-defined equality, for arrays of real values
12704 (in particular, IEEE-conformant floating point, because of negative
12705 zeroes and NaNs), and for arrays whose elements contain unused bits with
12706 indeterminate values.
12709 The other component-by-component array operations (@code{and}, @code{or},
12710 @code{xor}, @code{not}, and relational tests other than equality)
12711 are not implemented.
12714 @cindex array aggregates (Ada)
12715 @cindex record aggregates (Ada)
12716 @cindex aggregates (Ada)
12717 There is limited support for array and record aggregates. They are
12718 permitted only on the right sides of assignments, as in these examples:
12721 (@value{GDBP}) set An_Array := (1, 2, 3, 4, 5, 6)
12722 (@value{GDBP}) set An_Array := (1, others => 0)
12723 (@value{GDBP}) set An_Array := (0|4 => 1, 1..3 => 2, 5 => 6)
12724 (@value{GDBP}) set A_2D_Array := ((1, 2, 3), (4, 5, 6), (7, 8, 9))
12725 (@value{GDBP}) set A_Record := (1, "Peter", True);
12726 (@value{GDBP}) set A_Record := (Name => "Peter", Id => 1, Alive => True)
12730 discriminant's value by assigning an aggregate has an
12731 undefined effect if that discriminant is used within the record.
12732 However, you can first modify discriminants by directly assigning to
12733 them (which normally would not be allowed in Ada), and then performing an
12734 aggregate assignment. For example, given a variable @code{A_Rec}
12735 declared to have a type such as:
12738 type Rec (Len : Small_Integer := 0) is record
12740 Vals : IntArray (1 .. Len);
12744 you can assign a value with a different size of @code{Vals} with two
12748 (@value{GDBP}) set A_Rec.Len := 4
12749 (@value{GDBP}) set A_Rec := (Id => 42, Vals => (1, 2, 3, 4))
12752 As this example also illustrates, @value{GDBN} is very loose about the usual
12753 rules concerning aggregates. You may leave out some of the
12754 components of an array or record aggregate (such as the @code{Len}
12755 component in the assignment to @code{A_Rec} above); they will retain their
12756 original values upon assignment. You may freely use dynamic values as
12757 indices in component associations. You may even use overlapping or
12758 redundant component associations, although which component values are
12759 assigned in such cases is not defined.
12762 Calls to dispatching subprograms are not implemented.
12765 The overloading algorithm is much more limited (i.e., less selective)
12766 than that of real Ada. It makes only limited use of the context in
12767 which a subexpression appears to resolve its meaning, and it is much
12768 looser in its rules for allowing type matches. As a result, some
12769 function calls will be ambiguous, and the user will be asked to choose
12770 the proper resolution.
12773 The @code{new} operator is not implemented.
12776 Entry calls are not implemented.
12779 Aside from printing, arithmetic operations on the native VAX floating-point
12780 formats are not supported.
12783 It is not possible to slice a packed array.
12786 The names @code{True} and @code{False}, when not part of a qualified name,
12787 are interpreted as if implicitly prefixed by @code{Standard}, regardless of
12789 Should your program
12790 redefine these names in a package or procedure (at best a dubious practice),
12791 you will have to use fully qualified names to access their new definitions.
12794 @node Additions to Ada
12795 @subsubsection Additions to Ada
12796 @cindex Ada, deviations from
12798 As it does for other languages, @value{GDBN} makes certain generic
12799 extensions to Ada (@pxref{Expressions}):
12803 If the expression @var{E} is a variable residing in memory (typically
12804 a local variable or array element) and @var{N} is a positive integer,
12805 then @code{@var{E}@@@var{N}} displays the values of @var{E} and the
12806 @var{N}-1 adjacent variables following it in memory as an array. In
12807 Ada, this operator is generally not necessary, since its prime use is
12808 in displaying parts of an array, and slicing will usually do this in
12809 Ada. However, there are occasional uses when debugging programs in
12810 which certain debugging information has been optimized away.
12813 @code{@var{B}::@var{var}} means ``the variable named @var{var} that
12814 appears in function or file @var{B}.'' When @var{B} is a file name,
12815 you must typically surround it in single quotes.
12818 The expression @code{@{@var{type}@} @var{addr}} means ``the variable of type
12819 @var{type} that appears at address @var{addr}.''
12822 A name starting with @samp{$} is a convenience variable
12823 (@pxref{Convenience Vars}) or a machine register (@pxref{Registers}).
12826 In addition, @value{GDBN} provides a few other shortcuts and outright
12827 additions specific to Ada:
12831 The assignment statement is allowed as an expression, returning
12832 its right-hand operand as its value. Thus, you may enter
12835 (@value{GDBP}) set x := y + 3
12836 (@value{GDBP}) print A(tmp := y + 1)
12840 The semicolon is allowed as an ``operator,'' returning as its value
12841 the value of its right-hand operand.
12842 This allows, for example,
12843 complex conditional breaks:
12846 (@value{GDBP}) break f
12847 (@value{GDBP}) condition 1 (report(i); k += 1; A(k) > 100)
12851 Rather than use catenation and symbolic character names to introduce special
12852 characters into strings, one may instead use a special bracket notation,
12853 which is also used to print strings. A sequence of characters of the form
12854 @samp{["@var{XX}"]} within a string or character literal denotes the
12855 (single) character whose numeric encoding is @var{XX} in hexadecimal. The
12856 sequence of characters @samp{["""]} also denotes a single quotation mark
12857 in strings. For example,
12859 "One line.["0a"]Next line.["0a"]"
12862 contains an ASCII newline character (@code{Ada.Characters.Latin_1.LF})
12866 The subtype used as a prefix for the attributes @t{'Pos}, @t{'Min}, and
12867 @t{'Max} is optional (and is ignored in any case). For example, it is valid
12871 (@value{GDBP}) print 'max(x, y)
12875 When printing arrays, @value{GDBN} uses positional notation when the
12876 array has a lower bound of 1, and uses a modified named notation otherwise.
12877 For example, a one-dimensional array of three integers with a lower bound
12878 of 3 might print as
12885 That is, in contrast to valid Ada, only the first component has a @code{=>}
12889 You may abbreviate attributes in expressions with any unique,
12890 multi-character subsequence of
12891 their names (an exact match gets preference).
12892 For example, you may use @t{a'len}, @t{a'gth}, or @t{a'lh}
12893 in place of @t{a'length}.
12896 @cindex quoting Ada internal identifiers
12897 Since Ada is case-insensitive, the debugger normally maps identifiers you type
12898 to lower case. The GNAT compiler uses upper-case characters for
12899 some of its internal identifiers, which are normally of no interest to users.
12900 For the rare occasions when you actually have to look at them,
12901 enclose them in angle brackets to avoid the lower-case mapping.
12904 (@value{GDBP}) print <JMPBUF_SAVE>[0]
12908 Printing an object of class-wide type or dereferencing an
12909 access-to-class-wide value will display all the components of the object's
12910 specific type (as indicated by its run-time tag). Likewise, component
12911 selection on such a value will operate on the specific type of the
12916 @node Stopping Before Main Program
12917 @subsubsection Stopping at the Very Beginning
12919 @cindex breakpointing Ada elaboration code
12920 It is sometimes necessary to debug the program during elaboration, and
12921 before reaching the main procedure.
12922 As defined in the Ada Reference
12923 Manual, the elaboration code is invoked from a procedure called
12924 @code{adainit}. To run your program up to the beginning of
12925 elaboration, simply use the following two commands:
12926 @code{tbreak adainit} and @code{run}.
12929 @subsubsection Extensions for Ada Tasks
12930 @cindex Ada, tasking
12932 Support for Ada tasks is analogous to that for threads (@pxref{Threads}).
12933 @value{GDBN} provides the following task-related commands:
12938 This command shows a list of current Ada tasks, as in the following example:
12945 (@value{GDBP}) info tasks
12946 ID TID P-ID Pri State Name
12947 1 8088000 0 15 Child Activation Wait main_task
12948 2 80a4000 1 15 Accept Statement b
12949 3 809a800 1 15 Child Activation Wait a
12950 * 4 80ae800 3 15 Runnable c
12955 In this listing, the asterisk before the last task indicates it to be the
12956 task currently being inspected.
12960 Represents @value{GDBN}'s internal task number.
12966 The parent's task ID (@value{GDBN}'s internal task number).
12969 The base priority of the task.
12972 Current state of the task.
12976 The task has been created but has not been activated. It cannot be
12980 The task is not blocked for any reason known to Ada. (It may be waiting
12981 for a mutex, though.) It is conceptually "executing" in normal mode.
12984 The task is terminated, in the sense of ARM 9.3 (5). Any dependents
12985 that were waiting on terminate alternatives have been awakened and have
12986 terminated themselves.
12988 @item Child Activation Wait
12989 The task is waiting for created tasks to complete activation.
12991 @item Accept Statement
12992 The task is waiting on an accept or selective wait statement.
12994 @item Waiting on entry call
12995 The task is waiting on an entry call.
12997 @item Async Select Wait
12998 The task is waiting to start the abortable part of an asynchronous
13002 The task is waiting on a select statement with only a delay
13005 @item Child Termination Wait
13006 The task is sleeping having completed a master within itself, and is
13007 waiting for the tasks dependent on that master to become terminated or
13008 waiting on a terminate Phase.
13010 @item Wait Child in Term Alt
13011 The task is sleeping waiting for tasks on terminate alternatives to
13012 finish terminating.
13014 @item Accepting RV with @var{taskno}
13015 The task is accepting a rendez-vous with the task @var{taskno}.
13019 Name of the task in the program.
13023 @kindex info task @var{taskno}
13024 @item info task @var{taskno}
13025 This command shows detailled informations on the specified task, as in
13026 the following example:
13031 (@value{GDBP}) info tasks
13032 ID TID P-ID Pri State Name
13033 1 8077880 0 15 Child Activation Wait main_task
13034 * 2 807c468 1 15 Runnable task_1
13035 (@value{GDBP}) info task 2
13036 Ada Task: 0x807c468
13039 Parent: 1 (main_task)
13045 @kindex task@r{ (Ada)}
13046 @cindex current Ada task ID
13047 This command prints the ID of the current task.
13053 (@value{GDBP}) info tasks
13054 ID TID P-ID Pri State Name
13055 1 8077870 0 15 Child Activation Wait main_task
13056 * 2 807c458 1 15 Runnable t
13057 (@value{GDBP}) task
13058 [Current task is 2]
13061 @item task @var{taskno}
13062 @cindex Ada task switching
13063 This command is like the @code{thread @var{threadno}}
13064 command (@pxref{Threads}). It switches the context of debugging
13065 from the current task to the given task.
13071 (@value{GDBP}) info tasks
13072 ID TID P-ID Pri State Name
13073 1 8077870 0 15 Child Activation Wait main_task
13074 * 2 807c458 1 15 Runnable t
13075 (@value{GDBP}) task 1
13076 [Switching to task 1]
13077 #0 0x8067726 in pthread_cond_wait ()
13079 #0 0x8067726 in pthread_cond_wait ()
13080 #1 0x8056714 in system.os_interface.pthread_cond_wait ()
13081 #2 0x805cb63 in system.task_primitives.operations.sleep ()
13082 #3 0x806153e in system.tasking.stages.activate_tasks ()
13083 #4 0x804aacc in un () at un.adb:5
13086 @item break @var{linespec} task @var{taskno}
13087 @itemx break @var{linespec} task @var{taskno} if @dots{}
13088 @cindex breakpoints and tasks, in Ada
13089 @cindex task breakpoints, in Ada
13090 @kindex break @dots{} task @var{taskno}@r{ (Ada)}
13091 These commands are like the @code{break @dots{} thread @dots{}}
13092 command (@pxref{Thread Stops}).
13093 @var{linespec} specifies source lines, as described
13094 in @ref{Specify Location}.
13096 Use the qualifier @samp{task @var{taskno}} with a breakpoint command
13097 to specify that you only want @value{GDBN} to stop the program when a
13098 particular Ada task reaches this breakpoint. @var{taskno} is one of the
13099 numeric task identifiers assigned by @value{GDBN}, shown in the first
13100 column of the @samp{info tasks} display.
13102 If you do not specify @samp{task @var{taskno}} when you set a
13103 breakpoint, the breakpoint applies to @emph{all} tasks of your
13106 You can use the @code{task} qualifier on conditional breakpoints as
13107 well; in this case, place @samp{task @var{taskno}} before the
13108 breakpoint condition (before the @code{if}).
13116 (@value{GDBP}) info tasks
13117 ID TID P-ID Pri State Name
13118 1 140022020 0 15 Child Activation Wait main_task
13119 2 140045060 1 15 Accept/Select Wait t2
13120 3 140044840 1 15 Runnable t1
13121 * 4 140056040 1 15 Runnable t3
13122 (@value{GDBP}) b 15 task 2
13123 Breakpoint 5 at 0x120044cb0: file test_task_debug.adb, line 15.
13124 (@value{GDBP}) cont
13129 Breakpoint 5, test_task_debug () at test_task_debug.adb:15
13131 (@value{GDBP}) info tasks
13132 ID TID P-ID Pri State Name
13133 1 140022020 0 15 Child Activation Wait main_task
13134 * 2 140045060 1 15 Runnable t2
13135 3 140044840 1 15 Runnable t1
13136 4 140056040 1 15 Delay Sleep t3
13140 @node Ada Tasks and Core Files
13141 @subsubsection Tasking Support when Debugging Core Files
13142 @cindex Ada tasking and core file debugging
13144 When inspecting a core file, as opposed to debugging a live program,
13145 tasking support may be limited or even unavailable, depending on
13146 the platform being used.
13147 For instance, on x86-linux, the list of tasks is available, but task
13148 switching is not supported. On Tru64, however, task switching will work
13151 On certain platforms, including Tru64, the debugger needs to perform some
13152 memory writes in order to provide Ada tasking support. When inspecting
13153 a core file, this means that the core file must be opened with read-write
13154 privileges, using the command @samp{"set write on"} (@pxref{Patching}).
13155 Under these circumstances, you should make a backup copy of the core
13156 file before inspecting it with @value{GDBN}.
13159 @subsubsection Known Peculiarities of Ada Mode
13160 @cindex Ada, problems
13162 Besides the omissions listed previously (@pxref{Omissions from Ada}),
13163 we know of several problems with and limitations of Ada mode in
13165 some of which will be fixed with planned future releases of the debugger
13166 and the GNU Ada compiler.
13170 Currently, the debugger
13171 has insufficient information to determine whether certain pointers represent
13172 pointers to objects or the objects themselves.
13173 Thus, the user may have to tack an extra @code{.all} after an expression
13174 to get it printed properly.
13177 Static constants that the compiler chooses not to materialize as objects in
13178 storage are invisible to the debugger.
13181 Named parameter associations in function argument lists are ignored (the
13182 argument lists are treated as positional).
13185 Many useful library packages are currently invisible to the debugger.
13188 Fixed-point arithmetic, conversions, input, and output is carried out using
13189 floating-point arithmetic, and may give results that only approximate those on
13193 The GNAT compiler never generates the prefix @code{Standard} for any of
13194 the standard symbols defined by the Ada language. @value{GDBN} knows about
13195 this: it will strip the prefix from names when you use it, and will never
13196 look for a name you have so qualified among local symbols, nor match against
13197 symbols in other packages or subprograms. If you have
13198 defined entities anywhere in your program other than parameters and
13199 local variables whose simple names match names in @code{Standard},
13200 GNAT's lack of qualification here can cause confusion. When this happens,
13201 you can usually resolve the confusion
13202 by qualifying the problematic names with package
13203 @code{Standard} explicitly.
13206 Older versions of the compiler sometimes generate erroneous debugging
13207 information, resulting in the debugger incorrectly printing the value
13208 of affected entities. In some cases, the debugger is able to work
13209 around an issue automatically. In other cases, the debugger is able
13210 to work around the issue, but the work-around has to be specifically
13213 @kindex set ada trust-PAD-over-XVS
13214 @kindex show ada trust-PAD-over-XVS
13217 @item set ada trust-PAD-over-XVS on
13218 Configure GDB to strictly follow the GNAT encoding when computing the
13219 value of Ada entities, particularly when @code{PAD} and @code{PAD___XVS}
13220 types are involved (see @code{ada/exp_dbug.ads} in the GCC sources for
13221 a complete description of the encoding used by the GNAT compiler).
13222 This is the default.
13224 @item set ada trust-PAD-over-XVS off
13225 This is related to the encoding using by the GNAT compiler. If @value{GDBN}
13226 sometimes prints the wrong value for certain entities, changing @code{ada
13227 trust-PAD-over-XVS} to @code{off} activates a work-around which may fix
13228 the issue. It is always safe to set @code{ada trust-PAD-over-XVS} to
13229 @code{off}, but this incurs a slight performance penalty, so it is
13230 recommended to leave this setting to @code{on} unless necessary.
13234 @node Unsupported Languages
13235 @section Unsupported Languages
13237 @cindex unsupported languages
13238 @cindex minimal language
13239 In addition to the other fully-supported programming languages,
13240 @value{GDBN} also provides a pseudo-language, called @code{minimal}.
13241 It does not represent a real programming language, but provides a set
13242 of capabilities close to what the C or assembly languages provide.
13243 This should allow most simple operations to be performed while debugging
13244 an application that uses a language currently not supported by @value{GDBN}.
13246 If the language is set to @code{auto}, @value{GDBN} will automatically
13247 select this language if the current frame corresponds to an unsupported
13251 @chapter Examining the Symbol Table
13253 The commands described in this chapter allow you to inquire about the
13254 symbols (names of variables, functions and types) defined in your
13255 program. This information is inherent in the text of your program and
13256 does not change as your program executes. @value{GDBN} finds it in your
13257 program's symbol table, in the file indicated when you started @value{GDBN}
13258 (@pxref{File Options, ,Choosing Files}), or by one of the
13259 file-management commands (@pxref{Files, ,Commands to Specify Files}).
13261 @cindex symbol names
13262 @cindex names of symbols
13263 @cindex quoting names
13264 Occasionally, you may need to refer to symbols that contain unusual
13265 characters, which @value{GDBN} ordinarily treats as word delimiters. The
13266 most frequent case is in referring to static variables in other
13267 source files (@pxref{Variables,,Program Variables}). File names
13268 are recorded in object files as debugging symbols, but @value{GDBN} would
13269 ordinarily parse a typical file name, like @file{foo.c}, as the three words
13270 @samp{foo} @samp{.} @samp{c}. To allow @value{GDBN} to recognize
13271 @samp{foo.c} as a single symbol, enclose it in single quotes; for example,
13278 looks up the value of @code{x} in the scope of the file @file{foo.c}.
13281 @cindex case-insensitive symbol names
13282 @cindex case sensitivity in symbol names
13283 @kindex set case-sensitive
13284 @item set case-sensitive on
13285 @itemx set case-sensitive off
13286 @itemx set case-sensitive auto
13287 Normally, when @value{GDBN} looks up symbols, it matches their names
13288 with case sensitivity determined by the current source language.
13289 Occasionally, you may wish to control that. The command @code{set
13290 case-sensitive} lets you do that by specifying @code{on} for
13291 case-sensitive matches or @code{off} for case-insensitive ones. If
13292 you specify @code{auto}, case sensitivity is reset to the default
13293 suitable for the source language. The default is case-sensitive
13294 matches for all languages except for Fortran, for which the default is
13295 case-insensitive matches.
13297 @kindex show case-sensitive
13298 @item show case-sensitive
13299 This command shows the current setting of case sensitivity for symbols
13302 @kindex info address
13303 @cindex address of a symbol
13304 @item info address @var{symbol}
13305 Describe where the data for @var{symbol} is stored. For a register
13306 variable, this says which register it is kept in. For a non-register
13307 local variable, this prints the stack-frame offset at which the variable
13310 Note the contrast with @samp{print &@var{symbol}}, which does not work
13311 at all for a register variable, and for a stack local variable prints
13312 the exact address of the current instantiation of the variable.
13314 @kindex info symbol
13315 @cindex symbol from address
13316 @cindex closest symbol and offset for an address
13317 @item info symbol @var{addr}
13318 Print the name of a symbol which is stored at the address @var{addr}.
13319 If no symbol is stored exactly at @var{addr}, @value{GDBN} prints the
13320 nearest symbol and an offset from it:
13323 (@value{GDBP}) info symbol 0x54320
13324 _initialize_vx + 396 in section .text
13328 This is the opposite of the @code{info address} command. You can use
13329 it to find out the name of a variable or a function given its address.
13331 For dynamically linked executables, the name of executable or shared
13332 library containing the symbol is also printed:
13335 (@value{GDBP}) info symbol 0x400225
13336 _start + 5 in section .text of /tmp/a.out
13337 (@value{GDBP}) info symbol 0x2aaaac2811cf
13338 __read_nocancel + 6 in section .text of /usr/lib64/libc.so.6
13342 @item whatis [@var{arg}]
13343 Print the data type of @var{arg}, which can be either an expression or
13344 a data type. With no argument, print the data type of @code{$}, the
13345 last value in the value history. If @var{arg} is an expression, it is
13346 not actually evaluated, and any side-effecting operations (such as
13347 assignments or function calls) inside it do not take place. If
13348 @var{arg} is a type name, it may be the name of a type or typedef, or
13349 for C code it may have the form @samp{class @var{class-name}},
13350 @samp{struct @var{struct-tag}}, @samp{union @var{union-tag}} or
13351 @samp{enum @var{enum-tag}}.
13352 @xref{Expressions, ,Expressions}.
13355 @item ptype [@var{arg}]
13356 @code{ptype} accepts the same arguments as @code{whatis}, but prints a
13357 detailed description of the type, instead of just the name of the type.
13358 @xref{Expressions, ,Expressions}.
13360 For example, for this variable declaration:
13363 struct complex @{double real; double imag;@} v;
13367 the two commands give this output:
13371 (@value{GDBP}) whatis v
13372 type = struct complex
13373 (@value{GDBP}) ptype v
13374 type = struct complex @{
13382 As with @code{whatis}, using @code{ptype} without an argument refers to
13383 the type of @code{$}, the last value in the value history.
13385 @cindex incomplete type
13386 Sometimes, programs use opaque data types or incomplete specifications
13387 of complex data structure. If the debug information included in the
13388 program does not allow @value{GDBN} to display a full declaration of
13389 the data type, it will say @samp{<incomplete type>}. For example,
13390 given these declarations:
13394 struct foo *fooptr;
13398 but no definition for @code{struct foo} itself, @value{GDBN} will say:
13401 (@value{GDBP}) ptype foo
13402 $1 = <incomplete type>
13406 ``Incomplete type'' is C terminology for data types that are not
13407 completely specified.
13410 @item info types @var{regexp}
13412 Print a brief description of all types whose names match the regular
13413 expression @var{regexp} (or all types in your program, if you supply
13414 no argument). Each complete typename is matched as though it were a
13415 complete line; thus, @samp{i type value} gives information on all
13416 types in your program whose names include the string @code{value}, but
13417 @samp{i type ^value$} gives information only on types whose complete
13418 name is @code{value}.
13420 This command differs from @code{ptype} in two ways: first, like
13421 @code{whatis}, it does not print a detailed description; second, it
13422 lists all source files where a type is defined.
13425 @cindex local variables
13426 @item info scope @var{location}
13427 List all the variables local to a particular scope. This command
13428 accepts a @var{location} argument---a function name, a source line, or
13429 an address preceded by a @samp{*}, and prints all the variables local
13430 to the scope defined by that location. (@xref{Specify Location}, for
13431 details about supported forms of @var{location}.) For example:
13434 (@value{GDBP}) @b{info scope command_line_handler}
13435 Scope for command_line_handler:
13436 Symbol rl is an argument at stack/frame offset 8, length 4.
13437 Symbol linebuffer is in static storage at address 0x150a18, length 4.
13438 Symbol linelength is in static storage at address 0x150a1c, length 4.
13439 Symbol p is a local variable in register $esi, length 4.
13440 Symbol p1 is a local variable in register $ebx, length 4.
13441 Symbol nline is a local variable in register $edx, length 4.
13442 Symbol repeat is a local variable at frame offset -8, length 4.
13446 This command is especially useful for determining what data to collect
13447 during a @dfn{trace experiment}, see @ref{Tracepoint Actions,
13450 @kindex info source
13452 Show information about the current source file---that is, the source file for
13453 the function containing the current point of execution:
13456 the name of the source file, and the directory containing it,
13458 the directory it was compiled in,
13460 its length, in lines,
13462 which programming language it is written in,
13464 whether the executable includes debugging information for that file, and
13465 if so, what format the information is in (e.g., STABS, Dwarf 2, etc.), and
13467 whether the debugging information includes information about
13468 preprocessor macros.
13472 @kindex info sources
13474 Print the names of all source files in your program for which there is
13475 debugging information, organized into two lists: files whose symbols
13476 have already been read, and files whose symbols will be read when needed.
13478 @kindex info functions
13479 @item info functions
13480 Print the names and data types of all defined functions.
13482 @item info functions @var{regexp}
13483 Print the names and data types of all defined functions
13484 whose names contain a match for regular expression @var{regexp}.
13485 Thus, @samp{info fun step} finds all functions whose names
13486 include @code{step}; @samp{info fun ^step} finds those whose names
13487 start with @code{step}. If a function name contains characters
13488 that conflict with the regular expression language (e.g.@:
13489 @samp{operator*()}), they may be quoted with a backslash.
13491 @kindex info variables
13492 @item info variables
13493 Print the names and data types of all variables that are defined
13494 outside of functions (i.e.@: excluding local variables).
13496 @item info variables @var{regexp}
13497 Print the names and data types of all variables (except for local
13498 variables) whose names contain a match for regular expression
13501 @kindex info classes
13502 @cindex Objective-C, classes and selectors
13504 @itemx info classes @var{regexp}
13505 Display all Objective-C classes in your program, or
13506 (with the @var{regexp} argument) all those matching a particular regular
13509 @kindex info selectors
13510 @item info selectors
13511 @itemx info selectors @var{regexp}
13512 Display all Objective-C selectors in your program, or
13513 (with the @var{regexp} argument) all those matching a particular regular
13517 This was never implemented.
13518 @kindex info methods
13520 @itemx info methods @var{regexp}
13521 The @code{info methods} command permits the user to examine all defined
13522 methods within C@t{++} program, or (with the @var{regexp} argument) a
13523 specific set of methods found in the various C@t{++} classes. Many
13524 C@t{++} classes provide a large number of methods. Thus, the output
13525 from the @code{ptype} command can be overwhelming and hard to use. The
13526 @code{info-methods} command filters the methods, printing only those
13527 which match the regular-expression @var{regexp}.
13530 @cindex reloading symbols
13531 Some systems allow individual object files that make up your program to
13532 be replaced without stopping and restarting your program. For example,
13533 in VxWorks you can simply recompile a defective object file and keep on
13534 running. If you are running on one of these systems, you can allow
13535 @value{GDBN} to reload the symbols for automatically relinked modules:
13538 @kindex set symbol-reloading
13539 @item set symbol-reloading on
13540 Replace symbol definitions for the corresponding source file when an
13541 object file with a particular name is seen again.
13543 @item set symbol-reloading off
13544 Do not replace symbol definitions when encountering object files of the
13545 same name more than once. This is the default state; if you are not
13546 running on a system that permits automatic relinking of modules, you
13547 should leave @code{symbol-reloading} off, since otherwise @value{GDBN}
13548 may discard symbols when linking large programs, that may contain
13549 several modules (from different directories or libraries) with the same
13552 @kindex show symbol-reloading
13553 @item show symbol-reloading
13554 Show the current @code{on} or @code{off} setting.
13557 @cindex opaque data types
13558 @kindex set opaque-type-resolution
13559 @item set opaque-type-resolution on
13560 Tell @value{GDBN} to resolve opaque types. An opaque type is a type
13561 declared as a pointer to a @code{struct}, @code{class}, or
13562 @code{union}---for example, @code{struct MyType *}---that is used in one
13563 source file although the full declaration of @code{struct MyType} is in
13564 another source file. The default is on.
13566 A change in the setting of this subcommand will not take effect until
13567 the next time symbols for a file are loaded.
13569 @item set opaque-type-resolution off
13570 Tell @value{GDBN} not to resolve opaque types. In this case, the type
13571 is printed as follows:
13573 @{<no data fields>@}
13576 @kindex show opaque-type-resolution
13577 @item show opaque-type-resolution
13578 Show whether opaque types are resolved or not.
13580 @kindex maint print symbols
13581 @cindex symbol dump
13582 @kindex maint print psymbols
13583 @cindex partial symbol dump
13584 @item maint print symbols @var{filename}
13585 @itemx maint print psymbols @var{filename}
13586 @itemx maint print msymbols @var{filename}
13587 Write a dump of debugging symbol data into the file @var{filename}.
13588 These commands are used to debug the @value{GDBN} symbol-reading code. Only
13589 symbols with debugging data are included. If you use @samp{maint print
13590 symbols}, @value{GDBN} includes all the symbols for which it has already
13591 collected full details: that is, @var{filename} reflects symbols for
13592 only those files whose symbols @value{GDBN} has read. You can use the
13593 command @code{info sources} to find out which files these are. If you
13594 use @samp{maint print psymbols} instead, the dump shows information about
13595 symbols that @value{GDBN} only knows partially---that is, symbols defined in
13596 files that @value{GDBN} has skimmed, but not yet read completely. Finally,
13597 @samp{maint print msymbols} dumps just the minimal symbol information
13598 required for each object file from which @value{GDBN} has read some symbols.
13599 @xref{Files, ,Commands to Specify Files}, for a discussion of how
13600 @value{GDBN} reads symbols (in the description of @code{symbol-file}).
13602 @kindex maint info symtabs
13603 @kindex maint info psymtabs
13604 @cindex listing @value{GDBN}'s internal symbol tables
13605 @cindex symbol tables, listing @value{GDBN}'s internal
13606 @cindex full symbol tables, listing @value{GDBN}'s internal
13607 @cindex partial symbol tables, listing @value{GDBN}'s internal
13608 @item maint info symtabs @r{[} @var{regexp} @r{]}
13609 @itemx maint info psymtabs @r{[} @var{regexp} @r{]}
13611 List the @code{struct symtab} or @code{struct partial_symtab}
13612 structures whose names match @var{regexp}. If @var{regexp} is not
13613 given, list them all. The output includes expressions which you can
13614 copy into a @value{GDBN} debugging this one to examine a particular
13615 structure in more detail. For example:
13618 (@value{GDBP}) maint info psymtabs dwarf2read
13619 @{ objfile /home/gnu/build/gdb/gdb
13620 ((struct objfile *) 0x82e69d0)
13621 @{ psymtab /home/gnu/src/gdb/dwarf2read.c
13622 ((struct partial_symtab *) 0x8474b10)
13625 text addresses 0x814d3c8 -- 0x8158074
13626 globals (* (struct partial_symbol **) 0x8507a08 @@ 9)
13627 statics (* (struct partial_symbol **) 0x40e95b78 @@ 2882)
13628 dependencies (none)
13631 (@value{GDBP}) maint info symtabs
13635 We see that there is one partial symbol table whose filename contains
13636 the string @samp{dwarf2read}, belonging to the @samp{gdb} executable;
13637 and we see that @value{GDBN} has not read in any symtabs yet at all.
13638 If we set a breakpoint on a function, that will cause @value{GDBN} to
13639 read the symtab for the compilation unit containing that function:
13642 (@value{GDBP}) break dwarf2_psymtab_to_symtab
13643 Breakpoint 1 at 0x814e5da: file /home/gnu/src/gdb/dwarf2read.c,
13645 (@value{GDBP}) maint info symtabs
13646 @{ objfile /home/gnu/build/gdb/gdb
13647 ((struct objfile *) 0x82e69d0)
13648 @{ symtab /home/gnu/src/gdb/dwarf2read.c
13649 ((struct symtab *) 0x86c1f38)
13652 blockvector ((struct blockvector *) 0x86c1bd0) (primary)
13653 linetable ((struct linetable *) 0x8370fa0)
13654 debugformat DWARF 2
13663 @chapter Altering Execution
13665 Once you think you have found an error in your program, you might want to
13666 find out for certain whether correcting the apparent error would lead to
13667 correct results in the rest of the run. You can find the answer by
13668 experiment, using the @value{GDBN} features for altering execution of the
13671 For example, you can store new values into variables or memory
13672 locations, give your program a signal, restart it at a different
13673 address, or even return prematurely from a function.
13676 * Assignment:: Assignment to variables
13677 * Jumping:: Continuing at a different address
13678 * Signaling:: Giving your program a signal
13679 * Returning:: Returning from a function
13680 * Calling:: Calling your program's functions
13681 * Patching:: Patching your program
13685 @section Assignment to Variables
13688 @cindex setting variables
13689 To alter the value of a variable, evaluate an assignment expression.
13690 @xref{Expressions, ,Expressions}. For example,
13697 stores the value 4 into the variable @code{x}, and then prints the
13698 value of the assignment expression (which is 4).
13699 @xref{Languages, ,Using @value{GDBN} with Different Languages}, for more
13700 information on operators in supported languages.
13702 @kindex set variable
13703 @cindex variables, setting
13704 If you are not interested in seeing the value of the assignment, use the
13705 @code{set} command instead of the @code{print} command. @code{set} is
13706 really the same as @code{print} except that the expression's value is
13707 not printed and is not put in the value history (@pxref{Value History,
13708 ,Value History}). The expression is evaluated only for its effects.
13710 If the beginning of the argument string of the @code{set} command
13711 appears identical to a @code{set} subcommand, use the @code{set
13712 variable} command instead of just @code{set}. This command is identical
13713 to @code{set} except for its lack of subcommands. For example, if your
13714 program has a variable @code{width}, you get an error if you try to set
13715 a new value with just @samp{set width=13}, because @value{GDBN} has the
13716 command @code{set width}:
13719 (@value{GDBP}) whatis width
13721 (@value{GDBP}) p width
13723 (@value{GDBP}) set width=47
13724 Invalid syntax in expression.
13728 The invalid expression, of course, is @samp{=47}. In
13729 order to actually set the program's variable @code{width}, use
13732 (@value{GDBP}) set var width=47
13735 Because the @code{set} command has many subcommands that can conflict
13736 with the names of program variables, it is a good idea to use the
13737 @code{set variable} command instead of just @code{set}. For example, if
13738 your program has a variable @code{g}, you run into problems if you try
13739 to set a new value with just @samp{set g=4}, because @value{GDBN} has
13740 the command @code{set gnutarget}, abbreviated @code{set g}:
13744 (@value{GDBP}) whatis g
13748 (@value{GDBP}) set g=4
13752 The program being debugged has been started already.
13753 Start it from the beginning? (y or n) y
13754 Starting program: /home/smith/cc_progs/a.out
13755 "/home/smith/cc_progs/a.out": can't open to read symbols:
13756 Invalid bfd target.
13757 (@value{GDBP}) show g
13758 The current BFD target is "=4".
13763 The program variable @code{g} did not change, and you silently set the
13764 @code{gnutarget} to an invalid value. In order to set the variable
13768 (@value{GDBP}) set var g=4
13771 @value{GDBN} allows more implicit conversions in assignments than C; you can
13772 freely store an integer value into a pointer variable or vice versa,
13773 and you can convert any structure to any other structure that is the
13774 same length or shorter.
13775 @comment FIXME: how do structs align/pad in these conversions?
13776 @comment /doc@cygnus.com 18dec1990
13778 To store values into arbitrary places in memory, use the @samp{@{@dots{}@}}
13779 construct to generate a value of specified type at a specified address
13780 (@pxref{Expressions, ,Expressions}). For example, @code{@{int@}0x83040} refers
13781 to memory location @code{0x83040} as an integer (which implies a certain size
13782 and representation in memory), and
13785 set @{int@}0x83040 = 4
13789 stores the value 4 into that memory location.
13792 @section Continuing at a Different Address
13794 Ordinarily, when you continue your program, you do so at the place where
13795 it stopped, with the @code{continue} command. You can instead continue at
13796 an address of your own choosing, with the following commands:
13800 @item jump @var{linespec}
13801 @itemx jump @var{location}
13802 Resume execution at line @var{linespec} or at address given by
13803 @var{location}. Execution stops again immediately if there is a
13804 breakpoint there. @xref{Specify Location}, for a description of the
13805 different forms of @var{linespec} and @var{location}. It is common
13806 practice to use the @code{tbreak} command in conjunction with
13807 @code{jump}. @xref{Set Breaks, ,Setting Breakpoints}.
13809 The @code{jump} command does not change the current stack frame, or
13810 the stack pointer, or the contents of any memory location or any
13811 register other than the program counter. If line @var{linespec} is in
13812 a different function from the one currently executing, the results may
13813 be bizarre if the two functions expect different patterns of arguments or
13814 of local variables. For this reason, the @code{jump} command requests
13815 confirmation if the specified line is not in the function currently
13816 executing. However, even bizarre results are predictable if you are
13817 well acquainted with the machine-language code of your program.
13820 @c Doesn't work on HP-UX; have to set $pcoqh and $pcoqt.
13821 On many systems, you can get much the same effect as the @code{jump}
13822 command by storing a new value into the register @code{$pc}. The
13823 difference is that this does not start your program running; it only
13824 changes the address of where it @emph{will} run when you continue. For
13832 makes the next @code{continue} command or stepping command execute at
13833 address @code{0x485}, rather than at the address where your program stopped.
13834 @xref{Continuing and Stepping, ,Continuing and Stepping}.
13836 The most common occasion to use the @code{jump} command is to back
13837 up---perhaps with more breakpoints set---over a portion of a program
13838 that has already executed, in order to examine its execution in more
13843 @section Giving your Program a Signal
13844 @cindex deliver a signal to a program
13848 @item signal @var{signal}
13849 Resume execution where your program stopped, but immediately give it the
13850 signal @var{signal}. @var{signal} can be the name or the number of a
13851 signal. For example, on many systems @code{signal 2} and @code{signal
13852 SIGINT} are both ways of sending an interrupt signal.
13854 Alternatively, if @var{signal} is zero, continue execution without
13855 giving a signal. This is useful when your program stopped on account of
13856 a signal and would ordinary see the signal when resumed with the
13857 @code{continue} command; @samp{signal 0} causes it to resume without a
13860 @code{signal} does not repeat when you press @key{RET} a second time
13861 after executing the command.
13865 Invoking the @code{signal} command is not the same as invoking the
13866 @code{kill} utility from the shell. Sending a signal with @code{kill}
13867 causes @value{GDBN} to decide what to do with the signal depending on
13868 the signal handling tables (@pxref{Signals}). The @code{signal} command
13869 passes the signal directly to your program.
13873 @section Returning from a Function
13876 @cindex returning from a function
13879 @itemx return @var{expression}
13880 You can cancel execution of a function call with the @code{return}
13881 command. If you give an
13882 @var{expression} argument, its value is used as the function's return
13886 When you use @code{return}, @value{GDBN} discards the selected stack frame
13887 (and all frames within it). You can think of this as making the
13888 discarded frame return prematurely. If you wish to specify a value to
13889 be returned, give that value as the argument to @code{return}.
13891 This pops the selected stack frame (@pxref{Selection, ,Selecting a
13892 Frame}), and any other frames inside of it, leaving its caller as the
13893 innermost remaining frame. That frame becomes selected. The
13894 specified value is stored in the registers used for returning values
13897 The @code{return} command does not resume execution; it leaves the
13898 program stopped in the state that would exist if the function had just
13899 returned. In contrast, the @code{finish} command (@pxref{Continuing
13900 and Stepping, ,Continuing and Stepping}) resumes execution until the
13901 selected stack frame returns naturally.
13903 @value{GDBN} needs to know how the @var{expression} argument should be set for
13904 the inferior. The concrete registers assignment depends on the OS ABI and the
13905 type being returned by the selected stack frame. For example it is common for
13906 OS ABI to return floating point values in FPU registers while integer values in
13907 CPU registers. Still some ABIs return even floating point values in CPU
13908 registers. Larger integer widths (such as @code{long long int}) also have
13909 specific placement rules. @value{GDBN} already knows the OS ABI from its
13910 current target so it needs to find out also the type being returned to make the
13911 assignment into the right register(s).
13913 Normally, the selected stack frame has debug info. @value{GDBN} will always
13914 use the debug info instead of the implicit type of @var{expression} when the
13915 debug info is available. For example, if you type @kbd{return -1}, and the
13916 function in the current stack frame is declared to return a @code{long long
13917 int}, @value{GDBN} transparently converts the implicit @code{int} value of -1
13918 into a @code{long long int}:
13921 Breakpoint 1, func () at gdb.base/return-nodebug.c:29
13923 (@value{GDBP}) return -1
13924 Make func return now? (y or n) y
13925 #0 0x004004f6 in main () at gdb.base/return-nodebug.c:43
13926 43 printf ("result=%lld\n", func ());
13930 However, if the selected stack frame does not have a debug info, e.g., if the
13931 function was compiled without debug info, @value{GDBN} has to find out the type
13932 to return from user. Specifying a different type by mistake may set the value
13933 in different inferior registers than the caller code expects. For example,
13934 typing @kbd{return -1} with its implicit type @code{int} would set only a part
13935 of a @code{long long int} result for a debug info less function (on 32-bit
13936 architectures). Therefore the user is required to specify the return type by
13937 an appropriate cast explicitly:
13940 Breakpoint 2, 0x0040050b in func ()
13941 (@value{GDBP}) return -1
13942 Return value type not available for selected stack frame.
13943 Please use an explicit cast of the value to return.
13944 (@value{GDBP}) return (long long int) -1
13945 Make selected stack frame return now? (y or n) y
13946 #0 0x00400526 in main ()
13951 @section Calling Program Functions
13954 @cindex calling functions
13955 @cindex inferior functions, calling
13956 @item print @var{expr}
13957 Evaluate the expression @var{expr} and display the resulting value.
13958 @var{expr} may include calls to functions in the program being
13962 @item call @var{expr}
13963 Evaluate the expression @var{expr} without displaying @code{void}
13966 You can use this variant of the @code{print} command if you want to
13967 execute a function from your program that does not return anything
13968 (a.k.a.@: @dfn{a void function}), but without cluttering the output
13969 with @code{void} returned values that @value{GDBN} will otherwise
13970 print. If the result is not void, it is printed and saved in the
13974 It is possible for the function you call via the @code{print} or
13975 @code{call} command to generate a signal (e.g., if there's a bug in
13976 the function, or if you passed it incorrect arguments). What happens
13977 in that case is controlled by the @code{set unwindonsignal} command.
13979 Similarly, with a C@t{++} program it is possible for the function you
13980 call via the @code{print} or @code{call} command to generate an
13981 exception that is not handled due to the constraints of the dummy
13982 frame. In this case, any exception that is raised in the frame, but has
13983 an out-of-frame exception handler will not be found. GDB builds a
13984 dummy-frame for the inferior function call, and the unwinder cannot
13985 seek for exception handlers outside of this dummy-frame. What happens
13986 in that case is controlled by the
13987 @code{set unwind-on-terminating-exception} command.
13990 @item set unwindonsignal
13991 @kindex set unwindonsignal
13992 @cindex unwind stack in called functions
13993 @cindex call dummy stack unwinding
13994 Set unwinding of the stack if a signal is received while in a function
13995 that @value{GDBN} called in the program being debugged. If set to on,
13996 @value{GDBN} unwinds the stack it created for the call and restores
13997 the context to what it was before the call. If set to off (the
13998 default), @value{GDBN} stops in the frame where the signal was
14001 @item show unwindonsignal
14002 @kindex show unwindonsignal
14003 Show the current setting of stack unwinding in the functions called by
14006 @item set unwind-on-terminating-exception
14007 @kindex set unwind-on-terminating-exception
14008 @cindex unwind stack in called functions with unhandled exceptions
14009 @cindex call dummy stack unwinding on unhandled exception.
14010 Set unwinding of the stack if a C@t{++} exception is raised, but left
14011 unhandled while in a function that @value{GDBN} called in the program being
14012 debugged. If set to on (the default), @value{GDBN} unwinds the stack
14013 it created for the call and restores the context to what it was before
14014 the call. If set to off, @value{GDBN} the exception is delivered to
14015 the default C@t{++} exception handler and the inferior terminated.
14017 @item show unwind-on-terminating-exception
14018 @kindex show unwind-on-terminating-exception
14019 Show the current setting of stack unwinding in the functions called by
14024 @cindex weak alias functions
14025 Sometimes, a function you wish to call is actually a @dfn{weak alias}
14026 for another function. In such case, @value{GDBN} might not pick up
14027 the type information, including the types of the function arguments,
14028 which causes @value{GDBN} to call the inferior function incorrectly.
14029 As a result, the called function will function erroneously and may
14030 even crash. A solution to that is to use the name of the aliased
14034 @section Patching Programs
14036 @cindex patching binaries
14037 @cindex writing into executables
14038 @cindex writing into corefiles
14040 By default, @value{GDBN} opens the file containing your program's
14041 executable code (or the corefile) read-only. This prevents accidental
14042 alterations to machine code; but it also prevents you from intentionally
14043 patching your program's binary.
14045 If you'd like to be able to patch the binary, you can specify that
14046 explicitly with the @code{set write} command. For example, you might
14047 want to turn on internal debugging flags, or even to make emergency
14053 @itemx set write off
14054 If you specify @samp{set write on}, @value{GDBN} opens executable and
14055 core files for both reading and writing; if you specify @kbd{set write
14056 off} (the default), @value{GDBN} opens them read-only.
14058 If you have already loaded a file, you must load it again (using the
14059 @code{exec-file} or @code{core-file} command) after changing @code{set
14060 write}, for your new setting to take effect.
14064 Display whether executable files and core files are opened for writing
14065 as well as reading.
14069 @chapter @value{GDBN} Files
14071 @value{GDBN} needs to know the file name of the program to be debugged,
14072 both in order to read its symbol table and in order to start your
14073 program. To debug a core dump of a previous run, you must also tell
14074 @value{GDBN} the name of the core dump file.
14077 * Files:: Commands to specify files
14078 * Separate Debug Files:: Debugging information in separate files
14079 * Symbol Errors:: Errors reading symbol files
14080 * Data Files:: GDB data files
14084 @section Commands to Specify Files
14086 @cindex symbol table
14087 @cindex core dump file
14089 You may want to specify executable and core dump file names. The usual
14090 way to do this is at start-up time, using the arguments to
14091 @value{GDBN}'s start-up commands (@pxref{Invocation, , Getting In and
14092 Out of @value{GDBN}}).
14094 Occasionally it is necessary to change to a different file during a
14095 @value{GDBN} session. Or you may run @value{GDBN} and forget to
14096 specify a file you want to use. Or you are debugging a remote target
14097 via @code{gdbserver} (@pxref{Server, file, Using the @code{gdbserver}
14098 Program}). In these situations the @value{GDBN} commands to specify
14099 new files are useful.
14102 @cindex executable file
14104 @item file @var{filename}
14105 Use @var{filename} as the program to be debugged. It is read for its
14106 symbols and for the contents of pure memory. It is also the program
14107 executed when you use the @code{run} command. If you do not specify a
14108 directory and the file is not found in the @value{GDBN} working directory,
14109 @value{GDBN} uses the environment variable @code{PATH} as a list of
14110 directories to search, just as the shell does when looking for a program
14111 to run. You can change the value of this variable, for both @value{GDBN}
14112 and your program, using the @code{path} command.
14114 @cindex unlinked object files
14115 @cindex patching object files
14116 You can load unlinked object @file{.o} files into @value{GDBN} using
14117 the @code{file} command. You will not be able to ``run'' an object
14118 file, but you can disassemble functions and inspect variables. Also,
14119 if the underlying BFD functionality supports it, you could use
14120 @kbd{gdb -write} to patch object files using this technique. Note
14121 that @value{GDBN} can neither interpret nor modify relocations in this
14122 case, so branches and some initialized variables will appear to go to
14123 the wrong place. But this feature is still handy from time to time.
14126 @code{file} with no argument makes @value{GDBN} discard any information it
14127 has on both executable file and the symbol table.
14130 @item exec-file @r{[} @var{filename} @r{]}
14131 Specify that the program to be run (but not the symbol table) is found
14132 in @var{filename}. @value{GDBN} searches the environment variable @code{PATH}
14133 if necessary to locate your program. Omitting @var{filename} means to
14134 discard information on the executable file.
14136 @kindex symbol-file
14137 @item symbol-file @r{[} @var{filename} @r{]}
14138 Read symbol table information from file @var{filename}. @code{PATH} is
14139 searched when necessary. Use the @code{file} command to get both symbol
14140 table and program to run from the same file.
14142 @code{symbol-file} with no argument clears out @value{GDBN} information on your
14143 program's symbol table.
14145 The @code{symbol-file} command causes @value{GDBN} to forget the contents of
14146 some breakpoints and auto-display expressions. This is because they may
14147 contain pointers to the internal data recording symbols and data types,
14148 which are part of the old symbol table data being discarded inside
14151 @code{symbol-file} does not repeat if you press @key{RET} again after
14154 When @value{GDBN} is configured for a particular environment, it
14155 understands debugging information in whatever format is the standard
14156 generated for that environment; you may use either a @sc{gnu} compiler, or
14157 other compilers that adhere to the local conventions.
14158 Best results are usually obtained from @sc{gnu} compilers; for example,
14159 using @code{@value{NGCC}} you can generate debugging information for
14162 For most kinds of object files, with the exception of old SVR3 systems
14163 using COFF, the @code{symbol-file} command does not normally read the
14164 symbol table in full right away. Instead, it scans the symbol table
14165 quickly to find which source files and which symbols are present. The
14166 details are read later, one source file at a time, as they are needed.
14168 The purpose of this two-stage reading strategy is to make @value{GDBN}
14169 start up faster. For the most part, it is invisible except for
14170 occasional pauses while the symbol table details for a particular source
14171 file are being read. (The @code{set verbose} command can turn these
14172 pauses into messages if desired. @xref{Messages/Warnings, ,Optional
14173 Warnings and Messages}.)
14175 We have not implemented the two-stage strategy for COFF yet. When the
14176 symbol table is stored in COFF format, @code{symbol-file} reads the
14177 symbol table data in full right away. Note that ``stabs-in-COFF''
14178 still does the two-stage strategy, since the debug info is actually
14182 @cindex reading symbols immediately
14183 @cindex symbols, reading immediately
14184 @item symbol-file @r{[} -readnow @r{]} @var{filename}
14185 @itemx file @r{[} -readnow @r{]} @var{filename}
14186 You can override the @value{GDBN} two-stage strategy for reading symbol
14187 tables by using the @samp{-readnow} option with any of the commands that
14188 load symbol table information, if you want to be sure @value{GDBN} has the
14189 entire symbol table available.
14191 @c FIXME: for now no mention of directories, since this seems to be in
14192 @c flux. 13mar1992 status is that in theory GDB would look either in
14193 @c current dir or in same dir as myprog; but issues like competing
14194 @c GDB's, or clutter in system dirs, mean that in practice right now
14195 @c only current dir is used. FFish says maybe a special GDB hierarchy
14196 @c (eg rooted in val of env var GDBSYMS) could exist for mappable symbol
14200 @item core-file @r{[}@var{filename}@r{]}
14202 Specify the whereabouts of a core dump file to be used as the ``contents
14203 of memory''. Traditionally, core files contain only some parts of the
14204 address space of the process that generated them; @value{GDBN} can access the
14205 executable file itself for other parts.
14207 @code{core-file} with no argument specifies that no core file is
14210 Note that the core file is ignored when your program is actually running
14211 under @value{GDBN}. So, if you have been running your program and you
14212 wish to debug a core file instead, you must kill the subprocess in which
14213 the program is running. To do this, use the @code{kill} command
14214 (@pxref{Kill Process, ,Killing the Child Process}).
14216 @kindex add-symbol-file
14217 @cindex dynamic linking
14218 @item add-symbol-file @var{filename} @var{address}
14219 @itemx add-symbol-file @var{filename} @var{address} @r{[} -readnow @r{]}
14220 @itemx add-symbol-file @var{filename} @r{-s}@var{section} @var{address} @dots{}
14221 The @code{add-symbol-file} command reads additional symbol table
14222 information from the file @var{filename}. You would use this command
14223 when @var{filename} has been dynamically loaded (by some other means)
14224 into the program that is running. @var{address} should be the memory
14225 address at which the file has been loaded; @value{GDBN} cannot figure
14226 this out for itself. You can additionally specify an arbitrary number
14227 of @samp{@r{-s}@var{section} @var{address}} pairs, to give an explicit
14228 section name and base address for that section. You can specify any
14229 @var{address} as an expression.
14231 The symbol table of the file @var{filename} is added to the symbol table
14232 originally read with the @code{symbol-file} command. You can use the
14233 @code{add-symbol-file} command any number of times; the new symbol data
14234 thus read keeps adding to the old. To discard all old symbol data
14235 instead, use the @code{symbol-file} command without any arguments.
14237 @cindex relocatable object files, reading symbols from
14238 @cindex object files, relocatable, reading symbols from
14239 @cindex reading symbols from relocatable object files
14240 @cindex symbols, reading from relocatable object files
14241 @cindex @file{.o} files, reading symbols from
14242 Although @var{filename} is typically a shared library file, an
14243 executable file, or some other object file which has been fully
14244 relocated for loading into a process, you can also load symbolic
14245 information from relocatable @file{.o} files, as long as:
14249 the file's symbolic information refers only to linker symbols defined in
14250 that file, not to symbols defined by other object files,
14252 every section the file's symbolic information refers to has actually
14253 been loaded into the inferior, as it appears in the file, and
14255 you can determine the address at which every section was loaded, and
14256 provide these to the @code{add-symbol-file} command.
14260 Some embedded operating systems, like Sun Chorus and VxWorks, can load
14261 relocatable files into an already running program; such systems
14262 typically make the requirements above easy to meet. However, it's
14263 important to recognize that many native systems use complex link
14264 procedures (@code{.linkonce} section factoring and C@t{++} constructor table
14265 assembly, for example) that make the requirements difficult to meet. In
14266 general, one cannot assume that using @code{add-symbol-file} to read a
14267 relocatable object file's symbolic information will have the same effect
14268 as linking the relocatable object file into the program in the normal
14271 @code{add-symbol-file} does not repeat if you press @key{RET} after using it.
14273 @kindex add-symbol-file-from-memory
14274 @cindex @code{syscall DSO}
14275 @cindex load symbols from memory
14276 @item add-symbol-file-from-memory @var{address}
14277 Load symbols from the given @var{address} in a dynamically loaded
14278 object file whose image is mapped directly into the inferior's memory.
14279 For example, the Linux kernel maps a @code{syscall DSO} into each
14280 process's address space; this DSO provides kernel-specific code for
14281 some system calls. The argument can be any expression whose
14282 evaluation yields the address of the file's shared object file header.
14283 For this command to work, you must have used @code{symbol-file} or
14284 @code{exec-file} commands in advance.
14286 @kindex add-shared-symbol-files
14288 @item add-shared-symbol-files @var{library-file}
14289 @itemx assf @var{library-file}
14290 The @code{add-shared-symbol-files} command can currently be used only
14291 in the Cygwin build of @value{GDBN} on MS-Windows OS, where it is an
14292 alias for the @code{dll-symbols} command (@pxref{Cygwin Native}).
14293 @value{GDBN} automatically looks for shared libraries, however if
14294 @value{GDBN} does not find yours, you can invoke
14295 @code{add-shared-symbol-files}. It takes one argument: the shared
14296 library's file name. @code{assf} is a shorthand alias for
14297 @code{add-shared-symbol-files}.
14300 @item section @var{section} @var{addr}
14301 The @code{section} command changes the base address of the named
14302 @var{section} of the exec file to @var{addr}. This can be used if the
14303 exec file does not contain section addresses, (such as in the
14304 @code{a.out} format), or when the addresses specified in the file
14305 itself are wrong. Each section must be changed separately. The
14306 @code{info files} command, described below, lists all the sections and
14310 @kindex info target
14313 @code{info files} and @code{info target} are synonymous; both print the
14314 current target (@pxref{Targets, ,Specifying a Debugging Target}),
14315 including the names of the executable and core dump files currently in
14316 use by @value{GDBN}, and the files from which symbols were loaded. The
14317 command @code{help target} lists all possible targets rather than
14320 @kindex maint info sections
14321 @item maint info sections
14322 Another command that can give you extra information about program sections
14323 is @code{maint info sections}. In addition to the section information
14324 displayed by @code{info files}, this command displays the flags and file
14325 offset of each section in the executable and core dump files. In addition,
14326 @code{maint info sections} provides the following command options (which
14327 may be arbitrarily combined):
14331 Display sections for all loaded object files, including shared libraries.
14332 @item @var{sections}
14333 Display info only for named @var{sections}.
14334 @item @var{section-flags}
14335 Display info only for sections for which @var{section-flags} are true.
14336 The section flags that @value{GDBN} currently knows about are:
14339 Section will have space allocated in the process when loaded.
14340 Set for all sections except those containing debug information.
14342 Section will be loaded from the file into the child process memory.
14343 Set for pre-initialized code and data, clear for @code{.bss} sections.
14345 Section needs to be relocated before loading.
14347 Section cannot be modified by the child process.
14349 Section contains executable code only.
14351 Section contains data only (no executable code).
14353 Section will reside in ROM.
14355 Section contains data for constructor/destructor lists.
14357 Section is not empty.
14359 An instruction to the linker to not output the section.
14360 @item COFF_SHARED_LIBRARY
14361 A notification to the linker that the section contains
14362 COFF shared library information.
14364 Section contains common symbols.
14367 @kindex set trust-readonly-sections
14368 @cindex read-only sections
14369 @item set trust-readonly-sections on
14370 Tell @value{GDBN} that readonly sections in your object file
14371 really are read-only (i.e.@: that their contents will not change).
14372 In that case, @value{GDBN} can fetch values from these sections
14373 out of the object file, rather than from the target program.
14374 For some targets (notably embedded ones), this can be a significant
14375 enhancement to debugging performance.
14377 The default is off.
14379 @item set trust-readonly-sections off
14380 Tell @value{GDBN} not to trust readonly sections. This means that
14381 the contents of the section might change while the program is running,
14382 and must therefore be fetched from the target when needed.
14384 @item show trust-readonly-sections
14385 Show the current setting of trusting readonly sections.
14388 All file-specifying commands allow both absolute and relative file names
14389 as arguments. @value{GDBN} always converts the file name to an absolute file
14390 name and remembers it that way.
14392 @cindex shared libraries
14393 @anchor{Shared Libraries}
14394 @value{GDBN} supports @sc{gnu}/Linux, MS-Windows, HP-UX, SunOS, SVr4, Irix,
14395 and IBM RS/6000 AIX shared libraries.
14397 On MS-Windows @value{GDBN} must be linked with the Expat library to support
14398 shared libraries. @xref{Expat}.
14400 @value{GDBN} automatically loads symbol definitions from shared libraries
14401 when you use the @code{run} command, or when you examine a core file.
14402 (Before you issue the @code{run} command, @value{GDBN} does not understand
14403 references to a function in a shared library, however---unless you are
14404 debugging a core file).
14406 On HP-UX, if the program loads a library explicitly, @value{GDBN}
14407 automatically loads the symbols at the time of the @code{shl_load} call.
14409 @c FIXME: some @value{GDBN} release may permit some refs to undef
14410 @c FIXME...symbols---eg in a break cmd---assuming they are from a shared
14411 @c FIXME...lib; check this from time to time when updating manual
14413 There are times, however, when you may wish to not automatically load
14414 symbol definitions from shared libraries, such as when they are
14415 particularly large or there are many of them.
14417 To control the automatic loading of shared library symbols, use the
14421 @kindex set auto-solib-add
14422 @item set auto-solib-add @var{mode}
14423 If @var{mode} is @code{on}, symbols from all shared object libraries
14424 will be loaded automatically when the inferior begins execution, you
14425 attach to an independently started inferior, or when the dynamic linker
14426 informs @value{GDBN} that a new library has been loaded. If @var{mode}
14427 is @code{off}, symbols must be loaded manually, using the
14428 @code{sharedlibrary} command. The default value is @code{on}.
14430 @cindex memory used for symbol tables
14431 If your program uses lots of shared libraries with debug info that
14432 takes large amounts of memory, you can decrease the @value{GDBN}
14433 memory footprint by preventing it from automatically loading the
14434 symbols from shared libraries. To that end, type @kbd{set
14435 auto-solib-add off} before running the inferior, then load each
14436 library whose debug symbols you do need with @kbd{sharedlibrary
14437 @var{regexp}}, where @var{regexp} is a regular expression that matches
14438 the libraries whose symbols you want to be loaded.
14440 @kindex show auto-solib-add
14441 @item show auto-solib-add
14442 Display the current autoloading mode.
14445 @cindex load shared library
14446 To explicitly load shared library symbols, use the @code{sharedlibrary}
14450 @kindex info sharedlibrary
14452 @item info share @var{regex}
14453 @itemx info sharedlibrary @var{regex}
14454 Print the names of the shared libraries which are currently loaded
14455 that match @var{regex}. If @var{regex} is omitted then print
14456 all shared libraries that are loaded.
14458 @kindex sharedlibrary
14460 @item sharedlibrary @var{regex}
14461 @itemx share @var{regex}
14462 Load shared object library symbols for files matching a
14463 Unix regular expression.
14464 As with files loaded automatically, it only loads shared libraries
14465 required by your program for a core file or after typing @code{run}. If
14466 @var{regex} is omitted all shared libraries required by your program are
14469 @item nosharedlibrary
14470 @kindex nosharedlibrary
14471 @cindex unload symbols from shared libraries
14472 Unload all shared object library symbols. This discards all symbols
14473 that have been loaded from all shared libraries. Symbols from shared
14474 libraries that were loaded by explicit user requests are not
14478 Sometimes you may wish that @value{GDBN} stops and gives you control
14479 when any of shared library events happen. Use the @code{set
14480 stop-on-solib-events} command for this:
14483 @item set stop-on-solib-events
14484 @kindex set stop-on-solib-events
14485 This command controls whether @value{GDBN} should give you control
14486 when the dynamic linker notifies it about some shared library event.
14487 The most common event of interest is loading or unloading of a new
14490 @item show stop-on-solib-events
14491 @kindex show stop-on-solib-events
14492 Show whether @value{GDBN} stops and gives you control when shared
14493 library events happen.
14496 Shared libraries are also supported in many cross or remote debugging
14497 configurations. @value{GDBN} needs to have access to the target's libraries;
14498 this can be accomplished either by providing copies of the libraries
14499 on the host system, or by asking @value{GDBN} to automatically retrieve the
14500 libraries from the target. If copies of the target libraries are
14501 provided, they need to be the same as the target libraries, although the
14502 copies on the target can be stripped as long as the copies on the host are
14505 @cindex where to look for shared libraries
14506 For remote debugging, you need to tell @value{GDBN} where the target
14507 libraries are, so that it can load the correct copies---otherwise, it
14508 may try to load the host's libraries. @value{GDBN} has two variables
14509 to specify the search directories for target libraries.
14512 @cindex prefix for shared library file names
14513 @cindex system root, alternate
14514 @kindex set solib-absolute-prefix
14515 @kindex set sysroot
14516 @item set sysroot @var{path}
14517 Use @var{path} as the system root for the program being debugged. Any
14518 absolute shared library paths will be prefixed with @var{path}; many
14519 runtime loaders store the absolute paths to the shared library in the
14520 target program's memory. If you use @code{set sysroot} to find shared
14521 libraries, they need to be laid out in the same way that they are on
14522 the target, with e.g.@: a @file{/lib} and @file{/usr/lib} hierarchy
14525 If @var{path} starts with the sequence @file{remote:}, @value{GDBN} will
14526 retrieve the target libraries from the remote system. This is only
14527 supported when using a remote target that supports the @code{remote get}
14528 command (@pxref{File Transfer,,Sending files to a remote system}).
14529 The part of @var{path} following the initial @file{remote:}
14530 (if present) is used as system root prefix on the remote file system.
14531 @footnote{If you want to specify a local system root using a directory
14532 that happens to be named @file{remote:}, you need to use some equivalent
14533 variant of the name like @file{./remote:}.}
14535 For targets with an MS-DOS based filesystem, such as MS-Windows and
14536 SymbianOS, @value{GDBN} tries prefixing a few variants of the target
14537 absolute file name with @var{path}. But first, on Unix hosts,
14538 @value{GDBN} converts all backslash directory separators into forward
14539 slashes, because the backslash is not a directory separator on Unix:
14542 c:\foo\bar.dll @result{} c:/foo/bar.dll
14545 Then, @value{GDBN} attempts prefixing the target file name with
14546 @var{path}, and looks for the resulting file name in the host file
14550 c:/foo/bar.dll @result{} /path/to/sysroot/c:/foo/bar.dll
14553 If that does not find the shared library, @value{GDBN} tries removing
14554 the @samp{:} character from the drive spec, both for convenience, and,
14555 for the case of the host file system not supporting file names with
14559 c:/foo/bar.dll @result{} /path/to/sysroot/c/foo/bar.dll
14562 This makes it possible to have a system root that mirrors a target
14563 with more than one drive. E.g., you may want to setup your local
14564 copies of the target system shared libraries like so (note @samp{c} vs
14568 @file{/path/to/sysroot/c/sys/bin/foo.dll}
14569 @file{/path/to/sysroot/c/sys/bin/bar.dll}
14570 @file{/path/to/sysroot/z/sys/bin/bar.dll}
14574 and point the system root at @file{/path/to/sysroot}, so that
14575 @value{GDBN} can find the correct copies of both
14576 @file{c:\sys\bin\foo.dll}, and @file{z:\sys\bin\bar.dll}.
14578 If that still does not find the shared library, @value{GDBN} tries
14579 removing the whole drive spec from the target file name:
14582 c:/foo/bar.dll @result{} /path/to/sysroot/foo/bar.dll
14585 This last lookup makes it possible to not care about the drive name,
14586 if you don't want or need to.
14588 The @code{set solib-absolute-prefix} command is an alias for @code{set
14591 @cindex default system root
14592 @cindex @samp{--with-sysroot}
14593 You can set the default system root by using the configure-time
14594 @samp{--with-sysroot} option. If the system root is inside
14595 @value{GDBN}'s configured binary prefix (set with @samp{--prefix} or
14596 @samp{--exec-prefix}), then the default system root will be updated
14597 automatically if the installed @value{GDBN} is moved to a new
14600 @kindex show sysroot
14602 Display the current shared library prefix.
14604 @kindex set solib-search-path
14605 @item set solib-search-path @var{path}
14606 If this variable is set, @var{path} is a colon-separated list of
14607 directories to search for shared libraries. @samp{solib-search-path}
14608 is used after @samp{sysroot} fails to locate the library, or if the
14609 path to the library is relative instead of absolute. If you want to
14610 use @samp{solib-search-path} instead of @samp{sysroot}, be sure to set
14611 @samp{sysroot} to a nonexistent directory to prevent @value{GDBN} from
14612 finding your host's libraries. @samp{sysroot} is preferred; setting
14613 it to a nonexistent directory may interfere with automatic loading
14614 of shared library symbols.
14616 @kindex show solib-search-path
14617 @item show solib-search-path
14618 Display the current shared library search path.
14620 @cindex DOS file-name semantics of file names.
14621 @kindex set target-file-system-kind (unix|dos-based|auto)
14622 @kindex show target-file-system-kind
14623 @item set target-file-system-kind @var{kind}
14624 Set assumed file system kind for target reported file names.
14626 Shared library file names as reported by the target system may not
14627 make sense as is on the system @value{GDBN} is running on. For
14628 example, when remote debugging a target that has MS-DOS based file
14629 system semantics, from a Unix host, the target may be reporting to
14630 @value{GDBN} a list of loaded shared libraries with file names such as
14631 @file{c:\Windows\kernel32.dll}. On Unix hosts, there's no concept of
14632 drive letters, so the @samp{c:\} prefix is not normally understood as
14633 indicating an absolute file name, and neither is the backslash
14634 normally considered a directory separator character. In that case,
14635 the native file system would interpret this whole absolute file name
14636 as a relative file name with no directory components. This would make
14637 it impossible to point @value{GDBN} at a copy of the remote target's
14638 shared libraries on the host using @code{set sysroot}, and impractical
14639 with @code{set solib-search-path}. Setting
14640 @code{target-file-system-kind} to @code{dos-based} tells @value{GDBN}
14641 to interpret such file names similarly to how the target would, and to
14642 map them to file names valid on @value{GDBN}'s native file system
14643 semantics. The value of @var{kind} can be @code{"auto"}, in addition
14644 to one of the supported file system kinds. In that case, @value{GDBN}
14645 tries to determine the appropriate file system variant based on the
14646 current target's operating system (@pxref{ABI, ,Configuring the
14647 Current ABI}). The supported file system settings are:
14651 Instruct @value{GDBN} to assume the target file system is of Unix
14652 kind. Only file names starting the forward slash (@samp{/}) character
14653 are considered absolute, and the directory separator character is also
14657 Instruct @value{GDBN} to assume the target file system is DOS based.
14658 File names starting with either a forward slash, or a drive letter
14659 followed by a colon (e.g., @samp{c:}), are considered absolute, and
14660 both the slash (@samp{/}) and the backslash (@samp{\\}) characters are
14661 considered directory separators.
14664 Instruct @value{GDBN} to use the file system kind associated with the
14665 target operating system (@pxref{ABI, ,Configuring the Current ABI}).
14666 This is the default.
14671 @node Separate Debug Files
14672 @section Debugging Information in Separate Files
14673 @cindex separate debugging information files
14674 @cindex debugging information in separate files
14675 @cindex @file{.debug} subdirectories
14676 @cindex debugging information directory, global
14677 @cindex global debugging information directory
14678 @cindex build ID, and separate debugging files
14679 @cindex @file{.build-id} directory
14681 @value{GDBN} allows you to put a program's debugging information in a
14682 file separate from the executable itself, in a way that allows
14683 @value{GDBN} to find and load the debugging information automatically.
14684 Since debugging information can be very large---sometimes larger
14685 than the executable code itself---some systems distribute debugging
14686 information for their executables in separate files, which users can
14687 install only when they need to debug a problem.
14689 @value{GDBN} supports two ways of specifying the separate debug info
14694 The executable contains a @dfn{debug link} that specifies the name of
14695 the separate debug info file. The separate debug file's name is
14696 usually @file{@var{executable}.debug}, where @var{executable} is the
14697 name of the corresponding executable file without leading directories
14698 (e.g., @file{ls.debug} for @file{/usr/bin/ls}). In addition, the
14699 debug link specifies a 32-bit @dfn{Cyclic Redundancy Check} (CRC)
14700 checksum for the debug file, which @value{GDBN} uses to validate that
14701 the executable and the debug file came from the same build.
14704 The executable contains a @dfn{build ID}, a unique bit string that is
14705 also present in the corresponding debug info file. (This is supported
14706 only on some operating systems, notably those which use the ELF format
14707 for binary files and the @sc{gnu} Binutils.) For more details about
14708 this feature, see the description of the @option{--build-id}
14709 command-line option in @ref{Options, , Command Line Options, ld.info,
14710 The GNU Linker}. The debug info file's name is not specified
14711 explicitly by the build ID, but can be computed from the build ID, see
14715 Depending on the way the debug info file is specified, @value{GDBN}
14716 uses two different methods of looking for the debug file:
14720 For the ``debug link'' method, @value{GDBN} looks up the named file in
14721 the directory of the executable file, then in a subdirectory of that
14722 directory named @file{.debug}, and finally under the global debug
14723 directory, in a subdirectory whose name is identical to the leading
14724 directories of the executable's absolute file name.
14727 For the ``build ID'' method, @value{GDBN} looks in the
14728 @file{.build-id} subdirectory of the global debug directory for a file
14729 named @file{@var{nn}/@var{nnnnnnnn}.debug}, where @var{nn} are the
14730 first 2 hex characters of the build ID bit string, and @var{nnnnnnnn}
14731 are the rest of the bit string. (Real build ID strings are 32 or more
14732 hex characters, not 10.)
14735 So, for example, suppose you ask @value{GDBN} to debug
14736 @file{/usr/bin/ls}, which has a debug link that specifies the
14737 file @file{ls.debug}, and a build ID whose value in hex is
14738 @code{abcdef1234}. If the global debug directory is
14739 @file{/usr/lib/debug}, then @value{GDBN} will look for the following
14740 debug information files, in the indicated order:
14744 @file{/usr/lib/debug/.build-id/ab/cdef1234.debug}
14746 @file{/usr/bin/ls.debug}
14748 @file{/usr/bin/.debug/ls.debug}
14750 @file{/usr/lib/debug/usr/bin/ls.debug}.
14753 You can set the global debugging info directory's name, and view the
14754 name @value{GDBN} is currently using.
14758 @kindex set debug-file-directory
14759 @item set debug-file-directory @var{directories}
14760 Set the directories which @value{GDBN} searches for separate debugging
14761 information files to @var{directory}. Multiple directory components can be set
14762 concatenating them by a directory separator.
14764 @kindex show debug-file-directory
14765 @item show debug-file-directory
14766 Show the directories @value{GDBN} searches for separate debugging
14771 @cindex @code{.gnu_debuglink} sections
14772 @cindex debug link sections
14773 A debug link is a special section of the executable file named
14774 @code{.gnu_debuglink}. The section must contain:
14778 A filename, with any leading directory components removed, followed by
14781 zero to three bytes of padding, as needed to reach the next four-byte
14782 boundary within the section, and
14784 a four-byte CRC checksum, stored in the same endianness used for the
14785 executable file itself. The checksum is computed on the debugging
14786 information file's full contents by the function given below, passing
14787 zero as the @var{crc} argument.
14790 Any executable file format can carry a debug link, as long as it can
14791 contain a section named @code{.gnu_debuglink} with the contents
14794 @cindex @code{.note.gnu.build-id} sections
14795 @cindex build ID sections
14796 The build ID is a special section in the executable file (and in other
14797 ELF binary files that @value{GDBN} may consider). This section is
14798 often named @code{.note.gnu.build-id}, but that name is not mandatory.
14799 It contains unique identification for the built files---the ID remains
14800 the same across multiple builds of the same build tree. The default
14801 algorithm SHA1 produces 160 bits (40 hexadecimal characters) of the
14802 content for the build ID string. The same section with an identical
14803 value is present in the original built binary with symbols, in its
14804 stripped variant, and in the separate debugging information file.
14806 The debugging information file itself should be an ordinary
14807 executable, containing a full set of linker symbols, sections, and
14808 debugging information. The sections of the debugging information file
14809 should have the same names, addresses, and sizes as the original file,
14810 but they need not contain any data---much like a @code{.bss} section
14811 in an ordinary executable.
14813 The @sc{gnu} binary utilities (Binutils) package includes the
14814 @samp{objcopy} utility that can produce
14815 the separated executable / debugging information file pairs using the
14816 following commands:
14819 @kbd{objcopy --only-keep-debug foo foo.debug}
14824 These commands remove the debugging
14825 information from the executable file @file{foo} and place it in the file
14826 @file{foo.debug}. You can use the first, second or both methods to link the
14831 The debug link method needs the following additional command to also leave
14832 behind a debug link in @file{foo}:
14835 @kbd{objcopy --add-gnu-debuglink=foo.debug foo}
14838 Ulrich Drepper's @file{elfutils} package, starting with version 0.53, contains
14839 a version of the @code{strip} command such that the command @kbd{strip foo -f
14840 foo.debug} has the same functionality as the two @code{objcopy} commands and
14841 the @code{ln -s} command above, together.
14844 Build ID gets embedded into the main executable using @code{ld --build-id} or
14845 the @value{NGCC} counterpart @code{gcc -Wl,--build-id}. Build ID support plus
14846 compatibility fixes for debug files separation are present in @sc{gnu} binary
14847 utilities (Binutils) package since version 2.18.
14852 @cindex CRC algorithm definition
14853 The CRC used in @code{.gnu_debuglink} is the CRC-32 defined in
14854 IEEE 802.3 using the polynomial:
14856 @c TexInfo requires naked braces for multi-digit exponents for Tex
14857 @c output, but this causes HTML output to barf. HTML has to be set using
14858 @c raw commands. So we end up having to specify this equation in 2
14863 <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>
14864 + <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
14870 @math{x^{32} + x^{26} + x^{23} + x^{22} + x^{16} + x^{12} + x^{11}}
14871 @math{+ x^{10} + x^8 + x^7 + x^5 + x^4 + x^2 + x + 1}
14875 The function is computed byte at a time, taking the least
14876 significant bit of each byte first. The initial pattern
14877 @code{0xffffffff} is used, to ensure leading zeros affect the CRC and
14878 the final result is inverted to ensure trailing zeros also affect the
14881 @emph{Note:} This is the same CRC polynomial as used in handling the
14882 @dfn{Remote Serial Protocol} @code{qCRC} packet (@pxref{Remote Protocol,
14883 , @value{GDBN} Remote Serial Protocol}). However in the
14884 case of the Remote Serial Protocol, the CRC is computed @emph{most}
14885 significant bit first, and the result is not inverted, so trailing
14886 zeros have no effect on the CRC value.
14888 To complete the description, we show below the code of the function
14889 which produces the CRC used in @code{.gnu_debuglink}. Inverting the
14890 initially supplied @code{crc} argument means that an initial call to
14891 this function passing in zero will start computing the CRC using
14894 @kindex gnu_debuglink_crc32
14897 gnu_debuglink_crc32 (unsigned long crc,
14898 unsigned char *buf, size_t len)
14900 static const unsigned long crc32_table[256] =
14902 0x00000000, 0x77073096, 0xee0e612c, 0x990951ba, 0x076dc419,
14903 0x706af48f, 0xe963a535, 0x9e6495a3, 0x0edb8832, 0x79dcb8a4,
14904 0xe0d5e91e, 0x97d2d988, 0x09b64c2b, 0x7eb17cbd, 0xe7b82d07,
14905 0x90bf1d91, 0x1db71064, 0x6ab020f2, 0xf3b97148, 0x84be41de,
14906 0x1adad47d, 0x6ddde4eb, 0xf4d4b551, 0x83d385c7, 0x136c9856,
14907 0x646ba8c0, 0xfd62f97a, 0x8a65c9ec, 0x14015c4f, 0x63066cd9,
14908 0xfa0f3d63, 0x8d080df5, 0x3b6e20c8, 0x4c69105e, 0xd56041e4,
14909 0xa2677172, 0x3c03e4d1, 0x4b04d447, 0xd20d85fd, 0xa50ab56b,
14910 0x35b5a8fa, 0x42b2986c, 0xdbbbc9d6, 0xacbcf940, 0x32d86ce3,
14911 0x45df5c75, 0xdcd60dcf, 0xabd13d59, 0x26d930ac, 0x51de003a,
14912 0xc8d75180, 0xbfd06116, 0x21b4f4b5, 0x56b3c423, 0xcfba9599,
14913 0xb8bda50f, 0x2802b89e, 0x5f058808, 0xc60cd9b2, 0xb10be924,
14914 0x2f6f7c87, 0x58684c11, 0xc1611dab, 0xb6662d3d, 0x76dc4190,
14915 0x01db7106, 0x98d220bc, 0xefd5102a, 0x71b18589, 0x06b6b51f,
14916 0x9fbfe4a5, 0xe8b8d433, 0x7807c9a2, 0x0f00f934, 0x9609a88e,
14917 0xe10e9818, 0x7f6a0dbb, 0x086d3d2d, 0x91646c97, 0xe6635c01,
14918 0x6b6b51f4, 0x1c6c6162, 0x856530d8, 0xf262004e, 0x6c0695ed,
14919 0x1b01a57b, 0x8208f4c1, 0xf50fc457, 0x65b0d9c6, 0x12b7e950,
14920 0x8bbeb8ea, 0xfcb9887c, 0x62dd1ddf, 0x15da2d49, 0x8cd37cf3,
14921 0xfbd44c65, 0x4db26158, 0x3ab551ce, 0xa3bc0074, 0xd4bb30e2,
14922 0x4adfa541, 0x3dd895d7, 0xa4d1c46d, 0xd3d6f4fb, 0x4369e96a,
14923 0x346ed9fc, 0xad678846, 0xda60b8d0, 0x44042d73, 0x33031de5,
14924 0xaa0a4c5f, 0xdd0d7cc9, 0x5005713c, 0x270241aa, 0xbe0b1010,
14925 0xc90c2086, 0x5768b525, 0x206f85b3, 0xb966d409, 0xce61e49f,
14926 0x5edef90e, 0x29d9c998, 0xb0d09822, 0xc7d7a8b4, 0x59b33d17,
14927 0x2eb40d81, 0xb7bd5c3b, 0xc0ba6cad, 0xedb88320, 0x9abfb3b6,
14928 0x03b6e20c, 0x74b1d29a, 0xead54739, 0x9dd277af, 0x04db2615,
14929 0x73dc1683, 0xe3630b12, 0x94643b84, 0x0d6d6a3e, 0x7a6a5aa8,
14930 0xe40ecf0b, 0x9309ff9d, 0x0a00ae27, 0x7d079eb1, 0xf00f9344,
14931 0x8708a3d2, 0x1e01f268, 0x6906c2fe, 0xf762575d, 0x806567cb,
14932 0x196c3671, 0x6e6b06e7, 0xfed41b76, 0x89d32be0, 0x10da7a5a,
14933 0x67dd4acc, 0xf9b9df6f, 0x8ebeeff9, 0x17b7be43, 0x60b08ed5,
14934 0xd6d6a3e8, 0xa1d1937e, 0x38d8c2c4, 0x4fdff252, 0xd1bb67f1,
14935 0xa6bc5767, 0x3fb506dd, 0x48b2364b, 0xd80d2bda, 0xaf0a1b4c,
14936 0x36034af6, 0x41047a60, 0xdf60efc3, 0xa867df55, 0x316e8eef,
14937 0x4669be79, 0xcb61b38c, 0xbc66831a, 0x256fd2a0, 0x5268e236,
14938 0xcc0c7795, 0xbb0b4703, 0x220216b9, 0x5505262f, 0xc5ba3bbe,
14939 0xb2bd0b28, 0x2bb45a92, 0x5cb36a04, 0xc2d7ffa7, 0xb5d0cf31,
14940 0x2cd99e8b, 0x5bdeae1d, 0x9b64c2b0, 0xec63f226, 0x756aa39c,
14941 0x026d930a, 0x9c0906a9, 0xeb0e363f, 0x72076785, 0x05005713,
14942 0x95bf4a82, 0xe2b87a14, 0x7bb12bae, 0x0cb61b38, 0x92d28e9b,
14943 0xe5d5be0d, 0x7cdcefb7, 0x0bdbdf21, 0x86d3d2d4, 0xf1d4e242,
14944 0x68ddb3f8, 0x1fda836e, 0x81be16cd, 0xf6b9265b, 0x6fb077e1,
14945 0x18b74777, 0x88085ae6, 0xff0f6a70, 0x66063bca, 0x11010b5c,
14946 0x8f659eff, 0xf862ae69, 0x616bffd3, 0x166ccf45, 0xa00ae278,
14947 0xd70dd2ee, 0x4e048354, 0x3903b3c2, 0xa7672661, 0xd06016f7,
14948 0x4969474d, 0x3e6e77db, 0xaed16a4a, 0xd9d65adc, 0x40df0b66,
14949 0x37d83bf0, 0xa9bcae53, 0xdebb9ec5, 0x47b2cf7f, 0x30b5ffe9,
14950 0xbdbdf21c, 0xcabac28a, 0x53b39330, 0x24b4a3a6, 0xbad03605,
14951 0xcdd70693, 0x54de5729, 0x23d967bf, 0xb3667a2e, 0xc4614ab8,
14952 0x5d681b02, 0x2a6f2b94, 0xb40bbe37, 0xc30c8ea1, 0x5a05df1b,
14955 unsigned char *end;
14957 crc = ~crc & 0xffffffff;
14958 for (end = buf + len; buf < end; ++buf)
14959 crc = crc32_table[(crc ^ *buf) & 0xff] ^ (crc >> 8);
14960 return ~crc & 0xffffffff;
14965 This computation does not apply to the ``build ID'' method.
14968 @node Symbol Errors
14969 @section Errors Reading Symbol Files
14971 While reading a symbol file, @value{GDBN} occasionally encounters problems,
14972 such as symbol types it does not recognize, or known bugs in compiler
14973 output. By default, @value{GDBN} does not notify you of such problems, since
14974 they are relatively common and primarily of interest to people
14975 debugging compilers. If you are interested in seeing information
14976 about ill-constructed symbol tables, you can either ask @value{GDBN} to print
14977 only one message about each such type of problem, no matter how many
14978 times the problem occurs; or you can ask @value{GDBN} to print more messages,
14979 to see how many times the problems occur, with the @code{set
14980 complaints} command (@pxref{Messages/Warnings, ,Optional Warnings and
14983 The messages currently printed, and their meanings, include:
14986 @item inner block not inside outer block in @var{symbol}
14988 The symbol information shows where symbol scopes begin and end
14989 (such as at the start of a function or a block of statements). This
14990 error indicates that an inner scope block is not fully contained
14991 in its outer scope blocks.
14993 @value{GDBN} circumvents the problem by treating the inner block as if it had
14994 the same scope as the outer block. In the error message, @var{symbol}
14995 may be shown as ``@code{(don't know)}'' if the outer block is not a
14998 @item block at @var{address} out of order
15000 The symbol information for symbol scope blocks should occur in
15001 order of increasing addresses. This error indicates that it does not
15004 @value{GDBN} does not circumvent this problem, and has trouble
15005 locating symbols in the source file whose symbols it is reading. (You
15006 can often determine what source file is affected by specifying
15007 @code{set verbose on}. @xref{Messages/Warnings, ,Optional Warnings and
15010 @item bad block start address patched
15012 The symbol information for a symbol scope block has a start address
15013 smaller than the address of the preceding source line. This is known
15014 to occur in the SunOS 4.1.1 (and earlier) C compiler.
15016 @value{GDBN} circumvents the problem by treating the symbol scope block as
15017 starting on the previous source line.
15019 @item bad string table offset in symbol @var{n}
15022 Symbol number @var{n} contains a pointer into the string table which is
15023 larger than the size of the string table.
15025 @value{GDBN} circumvents the problem by considering the symbol to have the
15026 name @code{foo}, which may cause other problems if many symbols end up
15029 @item unknown symbol type @code{0x@var{nn}}
15031 The symbol information contains new data types that @value{GDBN} does
15032 not yet know how to read. @code{0x@var{nn}} is the symbol type of the
15033 uncomprehended information, in hexadecimal.
15035 @value{GDBN} circumvents the error by ignoring this symbol information.
15036 This usually allows you to debug your program, though certain symbols
15037 are not accessible. If you encounter such a problem and feel like
15038 debugging it, you can debug @code{@value{GDBP}} with itself, breakpoint
15039 on @code{complain}, then go up to the function @code{read_dbx_symtab}
15040 and examine @code{*bufp} to see the symbol.
15042 @item stub type has NULL name
15044 @value{GDBN} could not find the full definition for a struct or class.
15046 @item const/volatile indicator missing (ok if using g++ v1.x), got@dots{}
15047 The symbol information for a C@t{++} member function is missing some
15048 information that recent versions of the compiler should have output for
15051 @item info mismatch between compiler and debugger
15053 @value{GDBN} could not parse a type specification output by the compiler.
15058 @section GDB Data Files
15060 @cindex prefix for data files
15061 @value{GDBN} will sometimes read an auxiliary data file. These files
15062 are kept in a directory known as the @dfn{data directory}.
15064 You can set the data directory's name, and view the name @value{GDBN}
15065 is currently using.
15068 @kindex set data-directory
15069 @item set data-directory @var{directory}
15070 Set the directory which @value{GDBN} searches for auxiliary data files
15071 to @var{directory}.
15073 @kindex show data-directory
15074 @item show data-directory
15075 Show the directory @value{GDBN} searches for auxiliary data files.
15078 @cindex default data directory
15079 @cindex @samp{--with-gdb-datadir}
15080 You can set the default data directory by using the configure-time
15081 @samp{--with-gdb-datadir} option. If the data directory is inside
15082 @value{GDBN}'s configured binary prefix (set with @samp{--prefix} or
15083 @samp{--exec-prefix}), then the default data directory will be updated
15084 automatically if the installed @value{GDBN} is moved to a new
15088 @chapter Specifying a Debugging Target
15090 @cindex debugging target
15091 A @dfn{target} is the execution environment occupied by your program.
15093 Often, @value{GDBN} runs in the same host environment as your program;
15094 in that case, the debugging target is specified as a side effect when
15095 you use the @code{file} or @code{core} commands. When you need more
15096 flexibility---for example, running @value{GDBN} on a physically separate
15097 host, or controlling a standalone system over a serial port or a
15098 realtime system over a TCP/IP connection---you can use the @code{target}
15099 command to specify one of the target types configured for @value{GDBN}
15100 (@pxref{Target Commands, ,Commands for Managing Targets}).
15102 @cindex target architecture
15103 It is possible to build @value{GDBN} for several different @dfn{target
15104 architectures}. When @value{GDBN} is built like that, you can choose
15105 one of the available architectures with the @kbd{set architecture}
15109 @kindex set architecture
15110 @kindex show architecture
15111 @item set architecture @var{arch}
15112 This command sets the current target architecture to @var{arch}. The
15113 value of @var{arch} can be @code{"auto"}, in addition to one of the
15114 supported architectures.
15116 @item show architecture
15117 Show the current target architecture.
15119 @item set processor
15121 @kindex set processor
15122 @kindex show processor
15123 These are alias commands for, respectively, @code{set architecture}
15124 and @code{show architecture}.
15128 * Active Targets:: Active targets
15129 * Target Commands:: Commands for managing targets
15130 * Byte Order:: Choosing target byte order
15133 @node Active Targets
15134 @section Active Targets
15136 @cindex stacking targets
15137 @cindex active targets
15138 @cindex multiple targets
15140 There are three classes of targets: processes, core files, and
15141 executable files. @value{GDBN} can work concurrently on up to three
15142 active targets, one in each class. This allows you to (for example)
15143 start a process and inspect its activity without abandoning your work on
15146 For example, if you execute @samp{gdb a.out}, then the executable file
15147 @code{a.out} is the only active target. If you designate a core file as
15148 well---presumably from a prior run that crashed and coredumped---then
15149 @value{GDBN} has two active targets and uses them in tandem, looking
15150 first in the corefile target, then in the executable file, to satisfy
15151 requests for memory addresses. (Typically, these two classes of target
15152 are complementary, since core files contain only a program's
15153 read-write memory---variables and so on---plus machine status, while
15154 executable files contain only the program text and initialized data.)
15156 When you type @code{run}, your executable file becomes an active process
15157 target as well. When a process target is active, all @value{GDBN}
15158 commands requesting memory addresses refer to that target; addresses in
15159 an active core file or executable file target are obscured while the
15160 process target is active.
15162 Use the @code{core-file} and @code{exec-file} commands to select a new
15163 core file or executable target (@pxref{Files, ,Commands to Specify
15164 Files}). To specify as a target a process that is already running, use
15165 the @code{attach} command (@pxref{Attach, ,Debugging an Already-running
15168 @node Target Commands
15169 @section Commands for Managing Targets
15172 @item target @var{type} @var{parameters}
15173 Connects the @value{GDBN} host environment to a target machine or
15174 process. A target is typically a protocol for talking to debugging
15175 facilities. You use the argument @var{type} to specify the type or
15176 protocol of the target machine.
15178 Further @var{parameters} are interpreted by the target protocol, but
15179 typically include things like device names or host names to connect
15180 with, process numbers, and baud rates.
15182 The @code{target} command does not repeat if you press @key{RET} again
15183 after executing the command.
15185 @kindex help target
15187 Displays the names of all targets available. To display targets
15188 currently selected, use either @code{info target} or @code{info files}
15189 (@pxref{Files, ,Commands to Specify Files}).
15191 @item help target @var{name}
15192 Describe a particular target, including any parameters necessary to
15195 @kindex set gnutarget
15196 @item set gnutarget @var{args}
15197 @value{GDBN} uses its own library BFD to read your files. @value{GDBN}
15198 knows whether it is reading an @dfn{executable},
15199 a @dfn{core}, or a @dfn{.o} file; however, you can specify the file format
15200 with the @code{set gnutarget} command. Unlike most @code{target} commands,
15201 with @code{gnutarget} the @code{target} refers to a program, not a machine.
15204 @emph{Warning:} To specify a file format with @code{set gnutarget},
15205 you must know the actual BFD name.
15209 @xref{Files, , Commands to Specify Files}.
15211 @kindex show gnutarget
15212 @item show gnutarget
15213 Use the @code{show gnutarget} command to display what file format
15214 @code{gnutarget} is set to read. If you have not set @code{gnutarget},
15215 @value{GDBN} will determine the file format for each file automatically,
15216 and @code{show gnutarget} displays @samp{The current BDF target is "auto"}.
15219 @cindex common targets
15220 Here are some common targets (available, or not, depending on the GDB
15225 @item target exec @var{program}
15226 @cindex executable file target
15227 An executable file. @samp{target exec @var{program}} is the same as
15228 @samp{exec-file @var{program}}.
15230 @item target core @var{filename}
15231 @cindex core dump file target
15232 A core dump file. @samp{target core @var{filename}} is the same as
15233 @samp{core-file @var{filename}}.
15235 @item target remote @var{medium}
15236 @cindex remote target
15237 A remote system connected to @value{GDBN} via a serial line or network
15238 connection. This command tells @value{GDBN} to use its own remote
15239 protocol over @var{medium} for debugging. @xref{Remote Debugging}.
15241 For example, if you have a board connected to @file{/dev/ttya} on the
15242 machine running @value{GDBN}, you could say:
15245 target remote /dev/ttya
15248 @code{target remote} supports the @code{load} command. This is only
15249 useful if you have some other way of getting the stub to the target
15250 system, and you can put it somewhere in memory where it won't get
15251 clobbered by the download.
15253 @item target sim @r{[}@var{simargs}@r{]} @dots{}
15254 @cindex built-in simulator target
15255 Builtin CPU simulator. @value{GDBN} includes simulators for most architectures.
15263 works; however, you cannot assume that a specific memory map, device
15264 drivers, or even basic I/O is available, although some simulators do
15265 provide these. For info about any processor-specific simulator details,
15266 see the appropriate section in @ref{Embedded Processors, ,Embedded
15271 Some configurations may include these targets as well:
15275 @item target nrom @var{dev}
15276 @cindex NetROM ROM emulator target
15277 NetROM ROM emulator. This target only supports downloading.
15281 Different targets are available on different configurations of @value{GDBN};
15282 your configuration may have more or fewer targets.
15284 Many remote targets require you to download the executable's code once
15285 you've successfully established a connection. You may wish to control
15286 various aspects of this process.
15291 @kindex set hash@r{, for remote monitors}
15292 @cindex hash mark while downloading
15293 This command controls whether a hash mark @samp{#} is displayed while
15294 downloading a file to the remote monitor. If on, a hash mark is
15295 displayed after each S-record is successfully downloaded to the
15299 @kindex show hash@r{, for remote monitors}
15300 Show the current status of displaying the hash mark.
15302 @item set debug monitor
15303 @kindex set debug monitor
15304 @cindex display remote monitor communications
15305 Enable or disable display of communications messages between
15306 @value{GDBN} and the remote monitor.
15308 @item show debug monitor
15309 @kindex show debug monitor
15310 Show the current status of displaying communications between
15311 @value{GDBN} and the remote monitor.
15316 @kindex load @var{filename}
15317 @item load @var{filename}
15319 Depending on what remote debugging facilities are configured into
15320 @value{GDBN}, the @code{load} command may be available. Where it exists, it
15321 is meant to make @var{filename} (an executable) available for debugging
15322 on the remote system---by downloading, or dynamic linking, for example.
15323 @code{load} also records the @var{filename} symbol table in @value{GDBN}, like
15324 the @code{add-symbol-file} command.
15326 If your @value{GDBN} does not have a @code{load} command, attempting to
15327 execute it gets the error message ``@code{You can't do that when your
15328 target is @dots{}}''
15330 The file is loaded at whatever address is specified in the executable.
15331 For some object file formats, you can specify the load address when you
15332 link the program; for other formats, like a.out, the object file format
15333 specifies a fixed address.
15334 @c FIXME! This would be a good place for an xref to the GNU linker doc.
15336 Depending on the remote side capabilities, @value{GDBN} may be able to
15337 load programs into flash memory.
15339 @code{load} does not repeat if you press @key{RET} again after using it.
15343 @section Choosing Target Byte Order
15345 @cindex choosing target byte order
15346 @cindex target byte order
15348 Some types of processors, such as the MIPS, PowerPC, and Renesas SH,
15349 offer the ability to run either big-endian or little-endian byte
15350 orders. Usually the executable or symbol will include a bit to
15351 designate the endian-ness, and you will not need to worry about
15352 which to use. However, you may still find it useful to adjust
15353 @value{GDBN}'s idea of processor endian-ness manually.
15357 @item set endian big
15358 Instruct @value{GDBN} to assume the target is big-endian.
15360 @item set endian little
15361 Instruct @value{GDBN} to assume the target is little-endian.
15363 @item set endian auto
15364 Instruct @value{GDBN} to use the byte order associated with the
15368 Display @value{GDBN}'s current idea of the target byte order.
15372 Note that these commands merely adjust interpretation of symbolic
15373 data on the host, and that they have absolutely no effect on the
15377 @node Remote Debugging
15378 @chapter Debugging Remote Programs
15379 @cindex remote debugging
15381 If you are trying to debug a program running on a machine that cannot run
15382 @value{GDBN} in the usual way, it is often useful to use remote debugging.
15383 For example, you might use remote debugging on an operating system kernel,
15384 or on a small system which does not have a general purpose operating system
15385 powerful enough to run a full-featured debugger.
15387 Some configurations of @value{GDBN} have special serial or TCP/IP interfaces
15388 to make this work with particular debugging targets. In addition,
15389 @value{GDBN} comes with a generic serial protocol (specific to @value{GDBN},
15390 but not specific to any particular target system) which you can use if you
15391 write the remote stubs---the code that runs on the remote system to
15392 communicate with @value{GDBN}.
15394 Other remote targets may be available in your
15395 configuration of @value{GDBN}; use @code{help target} to list them.
15398 * Connecting:: Connecting to a remote target
15399 * File Transfer:: Sending files to a remote system
15400 * Server:: Using the gdbserver program
15401 * Remote Configuration:: Remote configuration
15402 * Remote Stub:: Implementing a remote stub
15406 @section Connecting to a Remote Target
15408 On the @value{GDBN} host machine, you will need an unstripped copy of
15409 your program, since @value{GDBN} needs symbol and debugging information.
15410 Start up @value{GDBN} as usual, using the name of the local copy of your
15411 program as the first argument.
15413 @cindex @code{target remote}
15414 @value{GDBN} can communicate with the target over a serial line, or
15415 over an @acronym{IP} network using @acronym{TCP} or @acronym{UDP}. In
15416 each case, @value{GDBN} uses the same protocol for debugging your
15417 program; only the medium carrying the debugging packets varies. The
15418 @code{target remote} command establishes a connection to the target.
15419 Its arguments indicate which medium to use:
15423 @item target remote @var{serial-device}
15424 @cindex serial line, @code{target remote}
15425 Use @var{serial-device} to communicate with the target. For example,
15426 to use a serial line connected to the device named @file{/dev/ttyb}:
15429 target remote /dev/ttyb
15432 If you're using a serial line, you may want to give @value{GDBN} the
15433 @w{@samp{--baud}} option, or use the @code{set remotebaud} command
15434 (@pxref{Remote Configuration, set remotebaud}) before the
15435 @code{target} command.
15437 @item target remote @code{@var{host}:@var{port}}
15438 @itemx target remote @code{tcp:@var{host}:@var{port}}
15439 @cindex @acronym{TCP} port, @code{target remote}
15440 Debug using a @acronym{TCP} connection to @var{port} on @var{host}.
15441 The @var{host} may be either a host name or a numeric @acronym{IP}
15442 address; @var{port} must be a decimal number. The @var{host} could be
15443 the target machine itself, if it is directly connected to the net, or
15444 it might be a terminal server which in turn has a serial line to the
15447 For example, to connect to port 2828 on a terminal server named
15451 target remote manyfarms:2828
15454 If your remote target is actually running on the same machine as your
15455 debugger session (e.g.@: a simulator for your target running on the
15456 same host), you can omit the hostname. For example, to connect to
15457 port 1234 on your local machine:
15460 target remote :1234
15464 Note that the colon is still required here.
15466 @item target remote @code{udp:@var{host}:@var{port}}
15467 @cindex @acronym{UDP} port, @code{target remote}
15468 Debug using @acronym{UDP} packets to @var{port} on @var{host}. For example, to
15469 connect to @acronym{UDP} port 2828 on a terminal server named @code{manyfarms}:
15472 target remote udp:manyfarms:2828
15475 When using a @acronym{UDP} connection for remote debugging, you should
15476 keep in mind that the `U' stands for ``Unreliable''. @acronym{UDP}
15477 can silently drop packets on busy or unreliable networks, which will
15478 cause havoc with your debugging session.
15480 @item target remote | @var{command}
15481 @cindex pipe, @code{target remote} to
15482 Run @var{command} in the background and communicate with it using a
15483 pipe. The @var{command} is a shell command, to be parsed and expanded
15484 by the system's command shell, @code{/bin/sh}; it should expect remote
15485 protocol packets on its standard input, and send replies on its
15486 standard output. You could use this to run a stand-alone simulator
15487 that speaks the remote debugging protocol, to make net connections
15488 using programs like @code{ssh}, or for other similar tricks.
15490 If @var{command} closes its standard output (perhaps by exiting),
15491 @value{GDBN} will try to send it a @code{SIGTERM} signal. (If the
15492 program has already exited, this will have no effect.)
15496 Once the connection has been established, you can use all the usual
15497 commands to examine and change data. The remote program is already
15498 running; you can use @kbd{step} and @kbd{continue}, and you do not
15499 need to use @kbd{run}.
15501 @cindex interrupting remote programs
15502 @cindex remote programs, interrupting
15503 Whenever @value{GDBN} is waiting for the remote program, if you type the
15504 interrupt character (often @kbd{Ctrl-c}), @value{GDBN} attempts to stop the
15505 program. This may or may not succeed, depending in part on the hardware
15506 and the serial drivers the remote system uses. If you type the
15507 interrupt character once again, @value{GDBN} displays this prompt:
15510 Interrupted while waiting for the program.
15511 Give up (and stop debugging it)? (y or n)
15514 If you type @kbd{y}, @value{GDBN} abandons the remote debugging session.
15515 (If you decide you want to try again later, you can use @samp{target
15516 remote} again to connect once more.) If you type @kbd{n}, @value{GDBN}
15517 goes back to waiting.
15520 @kindex detach (remote)
15522 When you have finished debugging the remote program, you can use the
15523 @code{detach} command to release it from @value{GDBN} control.
15524 Detaching from the target normally resumes its execution, but the results
15525 will depend on your particular remote stub. After the @code{detach}
15526 command, @value{GDBN} is free to connect to another target.
15530 The @code{disconnect} command behaves like @code{detach}, except that
15531 the target is generally not resumed. It will wait for @value{GDBN}
15532 (this instance or another one) to connect and continue debugging. After
15533 the @code{disconnect} command, @value{GDBN} is again free to connect to
15536 @cindex send command to remote monitor
15537 @cindex extend @value{GDBN} for remote targets
15538 @cindex add new commands for external monitor
15540 @item monitor @var{cmd}
15541 This command allows you to send arbitrary commands directly to the
15542 remote monitor. Since @value{GDBN} doesn't care about the commands it
15543 sends like this, this command is the way to extend @value{GDBN}---you
15544 can add new commands that only the external monitor will understand
15548 @node File Transfer
15549 @section Sending files to a remote system
15550 @cindex remote target, file transfer
15551 @cindex file transfer
15552 @cindex sending files to remote systems
15554 Some remote targets offer the ability to transfer files over the same
15555 connection used to communicate with @value{GDBN}. This is convenient
15556 for targets accessible through other means, e.g.@: @sc{gnu}/Linux systems
15557 running @code{gdbserver} over a network interface. For other targets,
15558 e.g.@: embedded devices with only a single serial port, this may be
15559 the only way to upload or download files.
15561 Not all remote targets support these commands.
15565 @item remote put @var{hostfile} @var{targetfile}
15566 Copy file @var{hostfile} from the host system (the machine running
15567 @value{GDBN}) to @var{targetfile} on the target system.
15570 @item remote get @var{targetfile} @var{hostfile}
15571 Copy file @var{targetfile} from the target system to @var{hostfile}
15572 on the host system.
15574 @kindex remote delete
15575 @item remote delete @var{targetfile}
15576 Delete @var{targetfile} from the target system.
15581 @section Using the @code{gdbserver} Program
15584 @cindex remote connection without stubs
15585 @code{gdbserver} is a control program for Unix-like systems, which
15586 allows you to connect your program with a remote @value{GDBN} via
15587 @code{target remote}---but without linking in the usual debugging stub.
15589 @code{gdbserver} is not a complete replacement for the debugging stubs,
15590 because it requires essentially the same operating-system facilities
15591 that @value{GDBN} itself does. In fact, a system that can run
15592 @code{gdbserver} to connect to a remote @value{GDBN} could also run
15593 @value{GDBN} locally! @code{gdbserver} is sometimes useful nevertheless,
15594 because it is a much smaller program than @value{GDBN} itself. It is
15595 also easier to port than all of @value{GDBN}, so you may be able to get
15596 started more quickly on a new system by using @code{gdbserver}.
15597 Finally, if you develop code for real-time systems, you may find that
15598 the tradeoffs involved in real-time operation make it more convenient to
15599 do as much development work as possible on another system, for example
15600 by cross-compiling. You can use @code{gdbserver} to make a similar
15601 choice for debugging.
15603 @value{GDBN} and @code{gdbserver} communicate via either a serial line
15604 or a TCP connection, using the standard @value{GDBN} remote serial
15608 @emph{Warning:} @code{gdbserver} does not have any built-in security.
15609 Do not run @code{gdbserver} connected to any public network; a
15610 @value{GDBN} connection to @code{gdbserver} provides access to the
15611 target system with the same privileges as the user running
15615 @subsection Running @code{gdbserver}
15616 @cindex arguments, to @code{gdbserver}
15618 Run @code{gdbserver} on the target system. You need a copy of the
15619 program you want to debug, including any libraries it requires.
15620 @code{gdbserver} does not need your program's symbol table, so you can
15621 strip the program if necessary to save space. @value{GDBN} on the host
15622 system does all the symbol handling.
15624 To use the server, you must tell it how to communicate with @value{GDBN};
15625 the name of your program; and the arguments for your program. The usual
15629 target> gdbserver @var{comm} @var{program} [ @var{args} @dots{} ]
15632 @var{comm} is either a device name (to use a serial line) or a TCP
15633 hostname and portnumber. For example, to debug Emacs with the argument
15634 @samp{foo.txt} and communicate with @value{GDBN} over the serial port
15638 target> gdbserver /dev/com1 emacs foo.txt
15641 @code{gdbserver} waits passively for the host @value{GDBN} to communicate
15644 To use a TCP connection instead of a serial line:
15647 target> gdbserver host:2345 emacs foo.txt
15650 The only difference from the previous example is the first argument,
15651 specifying that you are communicating with the host @value{GDBN} via
15652 TCP. The @samp{host:2345} argument means that @code{gdbserver} is to
15653 expect a TCP connection from machine @samp{host} to local TCP port 2345.
15654 (Currently, the @samp{host} part is ignored.) You can choose any number
15655 you want for the port number as long as it does not conflict with any
15656 TCP ports already in use on the target system (for example, @code{23} is
15657 reserved for @code{telnet}).@footnote{If you choose a port number that
15658 conflicts with another service, @code{gdbserver} prints an error message
15659 and exits.} You must use the same port number with the host @value{GDBN}
15660 @code{target remote} command.
15662 @subsubsection Attaching to a Running Program
15664 On some targets, @code{gdbserver} can also attach to running programs.
15665 This is accomplished via the @code{--attach} argument. The syntax is:
15668 target> gdbserver --attach @var{comm} @var{pid}
15671 @var{pid} is the process ID of a currently running process. It isn't necessary
15672 to point @code{gdbserver} at a binary for the running process.
15675 @cindex attach to a program by name
15676 You can debug processes by name instead of process ID if your target has the
15677 @code{pidof} utility:
15680 target> gdbserver --attach @var{comm} `pidof @var{program}`
15683 In case more than one copy of @var{program} is running, or @var{program}
15684 has multiple threads, most versions of @code{pidof} support the
15685 @code{-s} option to only return the first process ID.
15687 @subsubsection Multi-Process Mode for @code{gdbserver}
15688 @cindex gdbserver, multiple processes
15689 @cindex multiple processes with gdbserver
15691 When you connect to @code{gdbserver} using @code{target remote},
15692 @code{gdbserver} debugs the specified program only once. When the
15693 program exits, or you detach from it, @value{GDBN} closes the connection
15694 and @code{gdbserver} exits.
15696 If you connect using @kbd{target extended-remote}, @code{gdbserver}
15697 enters multi-process mode. When the debugged program exits, or you
15698 detach from it, @value{GDBN} stays connected to @code{gdbserver} even
15699 though no program is running. The @code{run} and @code{attach}
15700 commands instruct @code{gdbserver} to run or attach to a new program.
15701 The @code{run} command uses @code{set remote exec-file} (@pxref{set
15702 remote exec-file}) to select the program to run. Command line
15703 arguments are supported, except for wildcard expansion and I/O
15704 redirection (@pxref{Arguments}).
15706 To start @code{gdbserver} without supplying an initial command to run
15707 or process ID to attach, use the @option{--multi} command line option.
15708 Then you can connect using @kbd{target extended-remote} and start
15709 the program you want to debug.
15711 @code{gdbserver} does not automatically exit in multi-process mode.
15712 You can terminate it by using @code{monitor exit}
15713 (@pxref{Monitor Commands for gdbserver}).
15715 @subsubsection Other Command-Line Arguments for @code{gdbserver}
15717 The @option{--debug} option tells @code{gdbserver} to display extra
15718 status information about the debugging process. The
15719 @option{--remote-debug} option tells @code{gdbserver} to display
15720 remote protocol debug output. These options are intended for
15721 @code{gdbserver} development and for bug reports to the developers.
15723 The @option{--wrapper} option specifies a wrapper to launch programs
15724 for debugging. The option should be followed by the name of the
15725 wrapper, then any command-line arguments to pass to the wrapper, then
15726 @kbd{--} indicating the end of the wrapper arguments.
15728 @code{gdbserver} runs the specified wrapper program with a combined
15729 command line including the wrapper arguments, then the name of the
15730 program to debug, then any arguments to the program. The wrapper
15731 runs until it executes your program, and then @value{GDBN} gains control.
15733 You can use any program that eventually calls @code{execve} with
15734 its arguments as a wrapper. Several standard Unix utilities do
15735 this, e.g.@: @code{env} and @code{nohup}. Any Unix shell script ending
15736 with @code{exec "$@@"} will also work.
15738 For example, you can use @code{env} to pass an environment variable to
15739 the debugged program, without setting the variable in @code{gdbserver}'s
15743 $ gdbserver --wrapper env LD_PRELOAD=libtest.so -- :2222 ./testprog
15746 @subsection Connecting to @code{gdbserver}
15748 Run @value{GDBN} on the host system.
15750 First make sure you have the necessary symbol files. Load symbols for
15751 your application using the @code{file} command before you connect. Use
15752 @code{set sysroot} to locate target libraries (unless your @value{GDBN}
15753 was compiled with the correct sysroot using @code{--with-sysroot}).
15755 The symbol file and target libraries must exactly match the executable
15756 and libraries on the target, with one exception: the files on the host
15757 system should not be stripped, even if the files on the target system
15758 are. Mismatched or missing files will lead to confusing results
15759 during debugging. On @sc{gnu}/Linux targets, mismatched or missing
15760 files may also prevent @code{gdbserver} from debugging multi-threaded
15763 Connect to your target (@pxref{Connecting,,Connecting to a Remote Target}).
15764 For TCP connections, you must start up @code{gdbserver} prior to using
15765 the @code{target remote} command. Otherwise you may get an error whose
15766 text depends on the host system, but which usually looks something like
15767 @samp{Connection refused}. Don't use the @code{load}
15768 command in @value{GDBN} when using @code{gdbserver}, since the program is
15769 already on the target.
15771 @subsection Monitor Commands for @code{gdbserver}
15772 @cindex monitor commands, for @code{gdbserver}
15773 @anchor{Monitor Commands for gdbserver}
15775 During a @value{GDBN} session using @code{gdbserver}, you can use the
15776 @code{monitor} command to send special requests to @code{gdbserver}.
15777 Here are the available commands.
15781 List the available monitor commands.
15783 @item monitor set debug 0
15784 @itemx monitor set debug 1
15785 Disable or enable general debugging messages.
15787 @item monitor set remote-debug 0
15788 @itemx monitor set remote-debug 1
15789 Disable or enable specific debugging messages associated with the remote
15790 protocol (@pxref{Remote Protocol}).
15792 @item monitor set libthread-db-search-path [PATH]
15793 @cindex gdbserver, search path for @code{libthread_db}
15794 When this command is issued, @var{path} is a colon-separated list of
15795 directories to search for @code{libthread_db} (@pxref{Threads,,set
15796 libthread-db-search-path}). If you omit @var{path},
15797 @samp{libthread-db-search-path} will be reset to an empty list.
15800 Tell gdbserver to exit immediately. This command should be followed by
15801 @code{disconnect} to close the debugging session. @code{gdbserver} will
15802 detach from any attached processes and kill any processes it created.
15803 Use @code{monitor exit} to terminate @code{gdbserver} at the end
15804 of a multi-process mode debug session.
15808 @subsection Tracepoints support in @code{gdbserver}
15809 @cindex tracepoints support in @code{gdbserver}
15811 On some targets, @code{gdbserver} supports tracepoints and fast
15814 For fast tracepoints to work, a special library called the
15815 @dfn{in-process agent} (IPA), must be loaded in the inferior process.
15816 This library is built and distributed as an integral part of
15819 There are several ways to load the in-process agent in your program:
15822 @item Specifying it as dependency at link time
15824 You can link your program dynamically with the in-process agent
15825 library. On most systems, this is accomplished by adding
15826 @code{-linproctrace} to the link command.
15828 @item Using the system's preloading mechanisms
15830 You can force loading the in-process agent at startup time by using
15831 your system's support for preloading shared libraries. Many Unixes
15832 support the concept of preloading user defined libraries. In most
15833 cases, you do that by specifying @code{LD_PRELOAD=libinproctrace.so}
15834 in the environment. See also the description of @code{gdbserver}'s
15835 @option{--wrapper} command line option.
15837 @item Using @value{GDBN} to force loading the agent at run time
15839 On some systems, you can force the inferior to load a shared library,
15840 by calling a dynamic loader function in the inferior that takes care
15841 of dynamically looking up and loading a shared library. On most Unix
15842 systems, the function is @code{dlopen}. You'll use the @code{call}
15843 command for that. For example:
15846 (@value{GDBP}) call dlopen ("libinproctrace.so", ...)
15849 Note that on most Unix systems, for the @code{dlopen} function to be
15850 available, the program needs to be linked with @code{-ldl}.
15853 On systems that have a userspace dynamic loader, like most Unix
15854 systems, when you connect to @code{gdbserver} using @code{target
15855 remote}, you'll find that the program is stopped at the dynamic
15856 loader's entry point, and no shared library has been loaded in the
15857 program's address space yet, including the in-process agent. In that
15858 case, before being able to use any of the fast tracepoints features,
15859 you need to let the loader run and load the shared libraries. The
15860 most simple way to do that is to run the program to the main
15861 procedure. E.g., if debugging a C or C@t{++} program, start
15862 @code{gdbserver} like so:
15865 $ gdbserver :9999 myprogram
15868 Start GDB and connect to @code{gdbserver} like so, and run to main:
15872 (@value{GDBP}) target remote myhost:9999
15873 0x00007f215893ba60 in ?? () from /lib64/ld-linux-x86-64.so.2
15874 (@value{GDBP}) b main
15875 (@value{GDBP}) continue
15878 The in-process tracing agent library should now be loaded into the
15879 process; you can confirm it with the @code{info sharedlibrary}
15880 command, which will list @file{libinproctrace.so} as loaded in the
15881 process. You are now ready to install fast tracepoints and start
15884 @node Remote Configuration
15885 @section Remote Configuration
15888 @kindex show remote
15889 This section documents the configuration options available when
15890 debugging remote programs. For the options related to the File I/O
15891 extensions of the remote protocol, see @ref{system,
15892 system-call-allowed}.
15895 @item set remoteaddresssize @var{bits}
15896 @cindex address size for remote targets
15897 @cindex bits in remote address
15898 Set the maximum size of address in a memory packet to the specified
15899 number of bits. @value{GDBN} will mask off the address bits above
15900 that number, when it passes addresses to the remote target. The
15901 default value is the number of bits in the target's address.
15903 @item show remoteaddresssize
15904 Show the current value of remote address size in bits.
15906 @item set remotebaud @var{n}
15907 @cindex baud rate for remote targets
15908 Set the baud rate for the remote serial I/O to @var{n} baud. The
15909 value is used to set the speed of the serial port used for debugging
15912 @item show remotebaud
15913 Show the current speed of the remote connection.
15915 @item set remotebreak
15916 @cindex interrupt remote programs
15917 @cindex BREAK signal instead of Ctrl-C
15918 @anchor{set remotebreak}
15919 If set to on, @value{GDBN} sends a @code{BREAK} signal to the remote
15920 when you type @kbd{Ctrl-c} to interrupt the program running
15921 on the remote. If set to off, @value{GDBN} sends the @samp{Ctrl-C}
15922 character instead. The default is off, since most remote systems
15923 expect to see @samp{Ctrl-C} as the interrupt signal.
15925 @item show remotebreak
15926 Show whether @value{GDBN} sends @code{BREAK} or @samp{Ctrl-C} to
15927 interrupt the remote program.
15929 @item set remoteflow on
15930 @itemx set remoteflow off
15931 @kindex set remoteflow
15932 Enable or disable hardware flow control (@code{RTS}/@code{CTS})
15933 on the serial port used to communicate to the remote target.
15935 @item show remoteflow
15936 @kindex show remoteflow
15937 Show the current setting of hardware flow control.
15939 @item set remotelogbase @var{base}
15940 Set the base (a.k.a.@: radix) of logging serial protocol
15941 communications to @var{base}. Supported values of @var{base} are:
15942 @code{ascii}, @code{octal}, and @code{hex}. The default is
15945 @item show remotelogbase
15946 Show the current setting of the radix for logging remote serial
15949 @item set remotelogfile @var{file}
15950 @cindex record serial communications on file
15951 Record remote serial communications on the named @var{file}. The
15952 default is not to record at all.
15954 @item show remotelogfile.
15955 Show the current setting of the file name on which to record the
15956 serial communications.
15958 @item set remotetimeout @var{num}
15959 @cindex timeout for serial communications
15960 @cindex remote timeout
15961 Set the timeout limit to wait for the remote target to respond to
15962 @var{num} seconds. The default is 2 seconds.
15964 @item show remotetimeout
15965 Show the current number of seconds to wait for the remote target
15968 @cindex limit hardware breakpoints and watchpoints
15969 @cindex remote target, limit break- and watchpoints
15970 @anchor{set remote hardware-watchpoint-limit}
15971 @anchor{set remote hardware-breakpoint-limit}
15972 @item set remote hardware-watchpoint-limit @var{limit}
15973 @itemx set remote hardware-breakpoint-limit @var{limit}
15974 Restrict @value{GDBN} to using @var{limit} remote hardware breakpoint or
15975 watchpoints. A limit of -1, the default, is treated as unlimited.
15977 @item set remote exec-file @var{filename}
15978 @itemx show remote exec-file
15979 @anchor{set remote exec-file}
15980 @cindex executable file, for remote target
15981 Select the file used for @code{run} with @code{target
15982 extended-remote}. This should be set to a filename valid on the
15983 target system. If it is not set, the target will use a default
15984 filename (e.g.@: the last program run).
15986 @item set remote interrupt-sequence
15987 @cindex interrupt remote programs
15988 @cindex select Ctrl-C, BREAK or BREAK-g
15989 Allow the user to select one of @samp{Ctrl-C}, a @code{BREAK} or
15990 @samp{BREAK-g} as the
15991 sequence to the remote target in order to interrupt the execution.
15992 @samp{Ctrl-C} is a default. Some system prefers @code{BREAK} which
15993 is high level of serial line for some certain time.
15994 Linux kernel prefers @samp{BREAK-g}, a.k.a Magic SysRq g.
15995 It is @code{BREAK} signal followed by character @code{g}.
15997 @item show interrupt-sequence
15998 Show which of @samp{Ctrl-C}, @code{BREAK} or @code{BREAK-g}
15999 is sent by @value{GDBN} to interrupt the remote program.
16000 @code{BREAK-g} is BREAK signal followed by @code{g} and
16001 also known as Magic SysRq g.
16003 @item set remote interrupt-on-connect
16004 @cindex send interrupt-sequence on start
16005 Specify whether interrupt-sequence is sent to remote target when
16006 @value{GDBN} connects to it. This is mostly needed when you debug
16007 Linux kernel. Linux kernel expects @code{BREAK} followed by @code{g}
16008 which is known as Magic SysRq g in order to connect @value{GDBN}.
16010 @item show interrupt-on-connect
16011 Show whether interrupt-sequence is sent
16012 to remote target when @value{GDBN} connects to it.
16016 @item set tcp auto-retry on
16017 @cindex auto-retry, for remote TCP target
16018 Enable auto-retry for remote TCP connections. This is useful if the remote
16019 debugging agent is launched in parallel with @value{GDBN}; there is a race
16020 condition because the agent may not become ready to accept the connection
16021 before @value{GDBN} attempts to connect. When auto-retry is
16022 enabled, if the initial attempt to connect fails, @value{GDBN} reattempts
16023 to establish the connection using the timeout specified by
16024 @code{set tcp connect-timeout}.
16026 @item set tcp auto-retry off
16027 Do not auto-retry failed TCP connections.
16029 @item show tcp auto-retry
16030 Show the current auto-retry setting.
16032 @item set tcp connect-timeout @var{seconds}
16033 @cindex connection timeout, for remote TCP target
16034 @cindex timeout, for remote target connection
16035 Set the timeout for establishing a TCP connection to the remote target to
16036 @var{seconds}. The timeout affects both polling to retry failed connections
16037 (enabled by @code{set tcp auto-retry on}) and waiting for connections
16038 that are merely slow to complete, and represents an approximate cumulative
16041 @item show tcp connect-timeout
16042 Show the current connection timeout setting.
16045 @cindex remote packets, enabling and disabling
16046 The @value{GDBN} remote protocol autodetects the packets supported by
16047 your debugging stub. If you need to override the autodetection, you
16048 can use these commands to enable or disable individual packets. Each
16049 packet can be set to @samp{on} (the remote target supports this
16050 packet), @samp{off} (the remote target does not support this packet),
16051 or @samp{auto} (detect remote target support for this packet). They
16052 all default to @samp{auto}. For more information about each packet,
16053 see @ref{Remote Protocol}.
16055 During normal use, you should not have to use any of these commands.
16056 If you do, that may be a bug in your remote debugging stub, or a bug
16057 in @value{GDBN}. You may want to report the problem to the
16058 @value{GDBN} developers.
16060 For each packet @var{name}, the command to enable or disable the
16061 packet is @code{set remote @var{name}-packet}. The available settings
16064 @multitable @columnfractions 0.28 0.32 0.25
16067 @tab Related Features
16069 @item @code{fetch-register}
16071 @tab @code{info registers}
16073 @item @code{set-register}
16077 @item @code{binary-download}
16079 @tab @code{load}, @code{set}
16081 @item @code{read-aux-vector}
16082 @tab @code{qXfer:auxv:read}
16083 @tab @code{info auxv}
16085 @item @code{symbol-lookup}
16086 @tab @code{qSymbol}
16087 @tab Detecting multiple threads
16089 @item @code{attach}
16090 @tab @code{vAttach}
16093 @item @code{verbose-resume}
16095 @tab Stepping or resuming multiple threads
16101 @item @code{software-breakpoint}
16105 @item @code{hardware-breakpoint}
16109 @item @code{write-watchpoint}
16113 @item @code{read-watchpoint}
16117 @item @code{access-watchpoint}
16121 @item @code{target-features}
16122 @tab @code{qXfer:features:read}
16123 @tab @code{set architecture}
16125 @item @code{library-info}
16126 @tab @code{qXfer:libraries:read}
16127 @tab @code{info sharedlibrary}
16129 @item @code{memory-map}
16130 @tab @code{qXfer:memory-map:read}
16131 @tab @code{info mem}
16133 @item @code{read-spu-object}
16134 @tab @code{qXfer:spu:read}
16135 @tab @code{info spu}
16137 @item @code{write-spu-object}
16138 @tab @code{qXfer:spu:write}
16139 @tab @code{info spu}
16141 @item @code{read-siginfo-object}
16142 @tab @code{qXfer:siginfo:read}
16143 @tab @code{print $_siginfo}
16145 @item @code{write-siginfo-object}
16146 @tab @code{qXfer:siginfo:write}
16147 @tab @code{set $_siginfo}
16149 @item @code{threads}
16150 @tab @code{qXfer:threads:read}
16151 @tab @code{info threads}
16153 @item @code{get-thread-local-@*storage-address}
16154 @tab @code{qGetTLSAddr}
16155 @tab Displaying @code{__thread} variables
16157 @item @code{get-thread-information-block-address}
16158 @tab @code{qGetTIBAddr}
16159 @tab Display MS-Windows Thread Information Block.
16161 @item @code{search-memory}
16162 @tab @code{qSearch:memory}
16165 @item @code{supported-packets}
16166 @tab @code{qSupported}
16167 @tab Remote communications parameters
16169 @item @code{pass-signals}
16170 @tab @code{QPassSignals}
16171 @tab @code{handle @var{signal}}
16173 @item @code{hostio-close-packet}
16174 @tab @code{vFile:close}
16175 @tab @code{remote get}, @code{remote put}
16177 @item @code{hostio-open-packet}
16178 @tab @code{vFile:open}
16179 @tab @code{remote get}, @code{remote put}
16181 @item @code{hostio-pread-packet}
16182 @tab @code{vFile:pread}
16183 @tab @code{remote get}, @code{remote put}
16185 @item @code{hostio-pwrite-packet}
16186 @tab @code{vFile:pwrite}
16187 @tab @code{remote get}, @code{remote put}
16189 @item @code{hostio-unlink-packet}
16190 @tab @code{vFile:unlink}
16191 @tab @code{remote delete}
16193 @item @code{noack-packet}
16194 @tab @code{QStartNoAckMode}
16195 @tab Packet acknowledgment
16197 @item @code{osdata}
16198 @tab @code{qXfer:osdata:read}
16199 @tab @code{info os}
16201 @item @code{query-attached}
16202 @tab @code{qAttached}
16203 @tab Querying remote process attach state.
16207 @section Implementing a Remote Stub
16209 @cindex debugging stub, example
16210 @cindex remote stub, example
16211 @cindex stub example, remote debugging
16212 The stub files provided with @value{GDBN} implement the target side of the
16213 communication protocol, and the @value{GDBN} side is implemented in the
16214 @value{GDBN} source file @file{remote.c}. Normally, you can simply allow
16215 these subroutines to communicate, and ignore the details. (If you're
16216 implementing your own stub file, you can still ignore the details: start
16217 with one of the existing stub files. @file{sparc-stub.c} is the best
16218 organized, and therefore the easiest to read.)
16220 @cindex remote serial debugging, overview
16221 To debug a program running on another machine (the debugging
16222 @dfn{target} machine), you must first arrange for all the usual
16223 prerequisites for the program to run by itself. For example, for a C
16228 A startup routine to set up the C runtime environment; these usually
16229 have a name like @file{crt0}. The startup routine may be supplied by
16230 your hardware supplier, or you may have to write your own.
16233 A C subroutine library to support your program's
16234 subroutine calls, notably managing input and output.
16237 A way of getting your program to the other machine---for example, a
16238 download program. These are often supplied by the hardware
16239 manufacturer, but you may have to write your own from hardware
16243 The next step is to arrange for your program to use a serial port to
16244 communicate with the machine where @value{GDBN} is running (the @dfn{host}
16245 machine). In general terms, the scheme looks like this:
16249 @value{GDBN} already understands how to use this protocol; when everything
16250 else is set up, you can simply use the @samp{target remote} command
16251 (@pxref{Targets,,Specifying a Debugging Target}).
16253 @item On the target,
16254 you must link with your program a few special-purpose subroutines that
16255 implement the @value{GDBN} remote serial protocol. The file containing these
16256 subroutines is called a @dfn{debugging stub}.
16258 On certain remote targets, you can use an auxiliary program
16259 @code{gdbserver} instead of linking a stub into your program.
16260 @xref{Server,,Using the @code{gdbserver} Program}, for details.
16263 The debugging stub is specific to the architecture of the remote
16264 machine; for example, use @file{sparc-stub.c} to debug programs on
16267 @cindex remote serial stub list
16268 These working remote stubs are distributed with @value{GDBN}:
16273 @cindex @file{i386-stub.c}
16276 For Intel 386 and compatible architectures.
16279 @cindex @file{m68k-stub.c}
16280 @cindex Motorola 680x0
16282 For Motorola 680x0 architectures.
16285 @cindex @file{sh-stub.c}
16288 For Renesas SH architectures.
16291 @cindex @file{sparc-stub.c}
16293 For @sc{sparc} architectures.
16295 @item sparcl-stub.c
16296 @cindex @file{sparcl-stub.c}
16299 For Fujitsu @sc{sparclite} architectures.
16303 The @file{README} file in the @value{GDBN} distribution may list other
16304 recently added stubs.
16307 * Stub Contents:: What the stub can do for you
16308 * Bootstrapping:: What you must do for the stub
16309 * Debug Session:: Putting it all together
16312 @node Stub Contents
16313 @subsection What the Stub Can Do for You
16315 @cindex remote serial stub
16316 The debugging stub for your architecture supplies these three
16320 @item set_debug_traps
16321 @findex set_debug_traps
16322 @cindex remote serial stub, initialization
16323 This routine arranges for @code{handle_exception} to run when your
16324 program stops. You must call this subroutine explicitly near the
16325 beginning of your program.
16327 @item handle_exception
16328 @findex handle_exception
16329 @cindex remote serial stub, main routine
16330 This is the central workhorse, but your program never calls it
16331 explicitly---the setup code arranges for @code{handle_exception} to
16332 run when a trap is triggered.
16334 @code{handle_exception} takes control when your program stops during
16335 execution (for example, on a breakpoint), and mediates communications
16336 with @value{GDBN} on the host machine. This is where the communications
16337 protocol is implemented; @code{handle_exception} acts as the @value{GDBN}
16338 representative on the target machine. It begins by sending summary
16339 information on the state of your program, then continues to execute,
16340 retrieving and transmitting any information @value{GDBN} needs, until you
16341 execute a @value{GDBN} command that makes your program resume; at that point,
16342 @code{handle_exception} returns control to your own code on the target
16346 @cindex @code{breakpoint} subroutine, remote
16347 Use this auxiliary subroutine to make your program contain a
16348 breakpoint. Depending on the particular situation, this may be the only
16349 way for @value{GDBN} to get control. For instance, if your target
16350 machine has some sort of interrupt button, you won't need to call this;
16351 pressing the interrupt button transfers control to
16352 @code{handle_exception}---in effect, to @value{GDBN}. On some machines,
16353 simply receiving characters on the serial port may also trigger a trap;
16354 again, in that situation, you don't need to call @code{breakpoint} from
16355 your own program---simply running @samp{target remote} from the host
16356 @value{GDBN} session gets control.
16358 Call @code{breakpoint} if none of these is true, or if you simply want
16359 to make certain your program stops at a predetermined point for the
16360 start of your debugging session.
16363 @node Bootstrapping
16364 @subsection What You Must Do for the Stub
16366 @cindex remote stub, support routines
16367 The debugging stubs that come with @value{GDBN} are set up for a particular
16368 chip architecture, but they have no information about the rest of your
16369 debugging target machine.
16371 First of all you need to tell the stub how to communicate with the
16375 @item int getDebugChar()
16376 @findex getDebugChar
16377 Write this subroutine to read a single character from the serial port.
16378 It may be identical to @code{getchar} for your target system; a
16379 different name is used to allow you to distinguish the two if you wish.
16381 @item void putDebugChar(int)
16382 @findex putDebugChar
16383 Write this subroutine to write a single character to the serial port.
16384 It may be identical to @code{putchar} for your target system; a
16385 different name is used to allow you to distinguish the two if you wish.
16388 @cindex control C, and remote debugging
16389 @cindex interrupting remote targets
16390 If you want @value{GDBN} to be able to stop your program while it is
16391 running, you need to use an interrupt-driven serial driver, and arrange
16392 for it to stop when it receives a @code{^C} (@samp{\003}, the control-C
16393 character). That is the character which @value{GDBN} uses to tell the
16394 remote system to stop.
16396 Getting the debugging target to return the proper status to @value{GDBN}
16397 probably requires changes to the standard stub; one quick and dirty way
16398 is to just execute a breakpoint instruction (the ``dirty'' part is that
16399 @value{GDBN} reports a @code{SIGTRAP} instead of a @code{SIGINT}).
16401 Other routines you need to supply are:
16404 @item void exceptionHandler (int @var{exception_number}, void *@var{exception_address})
16405 @findex exceptionHandler
16406 Write this function to install @var{exception_address} in the exception
16407 handling tables. You need to do this because the stub does not have any
16408 way of knowing what the exception handling tables on your target system
16409 are like (for example, the processor's table might be in @sc{rom},
16410 containing entries which point to a table in @sc{ram}).
16411 @var{exception_number} is the exception number which should be changed;
16412 its meaning is architecture-dependent (for example, different numbers
16413 might represent divide by zero, misaligned access, etc). When this
16414 exception occurs, control should be transferred directly to
16415 @var{exception_address}, and the processor state (stack, registers,
16416 and so on) should be just as it is when a processor exception occurs. So if
16417 you want to use a jump instruction to reach @var{exception_address}, it
16418 should be a simple jump, not a jump to subroutine.
16420 For the 386, @var{exception_address} should be installed as an interrupt
16421 gate so that interrupts are masked while the handler runs. The gate
16422 should be at privilege level 0 (the most privileged level). The
16423 @sc{sparc} and 68k stubs are able to mask interrupts themselves without
16424 help from @code{exceptionHandler}.
16426 @item void flush_i_cache()
16427 @findex flush_i_cache
16428 On @sc{sparc} and @sc{sparclite} only, write this subroutine to flush the
16429 instruction cache, if any, on your target machine. If there is no
16430 instruction cache, this subroutine may be a no-op.
16432 On target machines that have instruction caches, @value{GDBN} requires this
16433 function to make certain that the state of your program is stable.
16437 You must also make sure this library routine is available:
16440 @item void *memset(void *, int, int)
16442 This is the standard library function @code{memset} that sets an area of
16443 memory to a known value. If you have one of the free versions of
16444 @code{libc.a}, @code{memset} can be found there; otherwise, you must
16445 either obtain it from your hardware manufacturer, or write your own.
16448 If you do not use the GNU C compiler, you may need other standard
16449 library subroutines as well; this varies from one stub to another,
16450 but in general the stubs are likely to use any of the common library
16451 subroutines which @code{@value{NGCC}} generates as inline code.
16454 @node Debug Session
16455 @subsection Putting it All Together
16457 @cindex remote serial debugging summary
16458 In summary, when your program is ready to debug, you must follow these
16463 Make sure you have defined the supporting low-level routines
16464 (@pxref{Bootstrapping,,What You Must Do for the Stub}):
16466 @code{getDebugChar}, @code{putDebugChar},
16467 @code{flush_i_cache}, @code{memset}, @code{exceptionHandler}.
16471 Insert these lines near the top of your program:
16479 For the 680x0 stub only, you need to provide a variable called
16480 @code{exceptionHook}. Normally you just use:
16483 void (*exceptionHook)() = 0;
16487 but if before calling @code{set_debug_traps}, you set it to point to a
16488 function in your program, that function is called when
16489 @code{@value{GDBN}} continues after stopping on a trap (for example, bus
16490 error). The function indicated by @code{exceptionHook} is called with
16491 one parameter: an @code{int} which is the exception number.
16494 Compile and link together: your program, the @value{GDBN} debugging stub for
16495 your target architecture, and the supporting subroutines.
16498 Make sure you have a serial connection between your target machine and
16499 the @value{GDBN} host, and identify the serial port on the host.
16502 @c The "remote" target now provides a `load' command, so we should
16503 @c document that. FIXME.
16504 Download your program to your target machine (or get it there by
16505 whatever means the manufacturer provides), and start it.
16508 Start @value{GDBN} on the host, and connect to the target
16509 (@pxref{Connecting,,Connecting to a Remote Target}).
16513 @node Configurations
16514 @chapter Configuration-Specific Information
16516 While nearly all @value{GDBN} commands are available for all native and
16517 cross versions of the debugger, there are some exceptions. This chapter
16518 describes things that are only available in certain configurations.
16520 There are three major categories of configurations: native
16521 configurations, where the host and target are the same, embedded
16522 operating system configurations, which are usually the same for several
16523 different processor architectures, and bare embedded processors, which
16524 are quite different from each other.
16529 * Embedded Processors::
16536 This section describes details specific to particular native
16541 * BSD libkvm Interface:: Debugging BSD kernel memory images
16542 * SVR4 Process Information:: SVR4 process information
16543 * DJGPP Native:: Features specific to the DJGPP port
16544 * Cygwin Native:: Features specific to the Cygwin port
16545 * Hurd Native:: Features specific to @sc{gnu} Hurd
16546 * Neutrino:: Features specific to QNX Neutrino
16547 * Darwin:: Features specific to Darwin
16553 On HP-UX systems, if you refer to a function or variable name that
16554 begins with a dollar sign, @value{GDBN} searches for a user or system
16555 name first, before it searches for a convenience variable.
16558 @node BSD libkvm Interface
16559 @subsection BSD libkvm Interface
16562 @cindex kernel memory image
16563 @cindex kernel crash dump
16565 BSD-derived systems (FreeBSD/NetBSD/OpenBSD) have a kernel memory
16566 interface that provides a uniform interface for accessing kernel virtual
16567 memory images, including live systems and crash dumps. @value{GDBN}
16568 uses this interface to allow you to debug live kernels and kernel crash
16569 dumps on many native BSD configurations. This is implemented as a
16570 special @code{kvm} debugging target. For debugging a live system, load
16571 the currently running kernel into @value{GDBN} and connect to the
16575 (@value{GDBP}) @b{target kvm}
16578 For debugging crash dumps, provide the file name of the crash dump as an
16582 (@value{GDBP}) @b{target kvm /var/crash/bsd.0}
16585 Once connected to the @code{kvm} target, the following commands are
16591 Set current context from the @dfn{Process Control Block} (PCB) address.
16594 Set current context from proc address. This command isn't available on
16595 modern FreeBSD systems.
16598 @node SVR4 Process Information
16599 @subsection SVR4 Process Information
16601 @cindex examine process image
16602 @cindex process info via @file{/proc}
16604 Many versions of SVR4 and compatible systems provide a facility called
16605 @samp{/proc} that can be used to examine the image of a running
16606 process using file-system subroutines. If @value{GDBN} is configured
16607 for an operating system with this facility, the command @code{info
16608 proc} is available to report information about the process running
16609 your program, or about any process running on your system. @code{info
16610 proc} works only on SVR4 systems that include the @code{procfs} code.
16611 This includes, as of this writing, @sc{gnu}/Linux, OSF/1 (Digital
16612 Unix), Solaris, Irix, and Unixware, but not HP-UX, for example.
16618 @itemx info proc @var{process-id}
16619 Summarize available information about any running process. If a
16620 process ID is specified by @var{process-id}, display information about
16621 that process; otherwise display information about the program being
16622 debugged. The summary includes the debugged process ID, the command
16623 line used to invoke it, its current working directory, and its
16624 executable file's absolute file name.
16626 On some systems, @var{process-id} can be of the form
16627 @samp{[@var{pid}]/@var{tid}} which specifies a certain thread ID
16628 within a process. If the optional @var{pid} part is missing, it means
16629 a thread from the process being debugged (the leading @samp{/} still
16630 needs to be present, or else @value{GDBN} will interpret the number as
16631 a process ID rather than a thread ID).
16633 @item info proc mappings
16634 @cindex memory address space mappings
16635 Report the memory address space ranges accessible in the program, with
16636 information on whether the process has read, write, or execute access
16637 rights to each range. On @sc{gnu}/Linux systems, each memory range
16638 includes the object file which is mapped to that range, instead of the
16639 memory access rights to that range.
16641 @item info proc stat
16642 @itemx info proc status
16643 @cindex process detailed status information
16644 These subcommands are specific to @sc{gnu}/Linux systems. They show
16645 the process-related information, including the user ID and group ID;
16646 how many threads are there in the process; its virtual memory usage;
16647 the signals that are pending, blocked, and ignored; its TTY; its
16648 consumption of system and user time; its stack size; its @samp{nice}
16649 value; etc. For more information, see the @samp{proc} man page
16650 (type @kbd{man 5 proc} from your shell prompt).
16652 @item info proc all
16653 Show all the information about the process described under all of the
16654 above @code{info proc} subcommands.
16657 @comment These sub-options of 'info proc' were not included when
16658 @comment procfs.c was re-written. Keep their descriptions around
16659 @comment against the day when someone finds the time to put them back in.
16660 @kindex info proc times
16661 @item info proc times
16662 Starting time, user CPU time, and system CPU time for your program and
16665 @kindex info proc id
16667 Report on the process IDs related to your program: its own process ID,
16668 the ID of its parent, the process group ID, and the session ID.
16671 @item set procfs-trace
16672 @kindex set procfs-trace
16673 @cindex @code{procfs} API calls
16674 This command enables and disables tracing of @code{procfs} API calls.
16676 @item show procfs-trace
16677 @kindex show procfs-trace
16678 Show the current state of @code{procfs} API call tracing.
16680 @item set procfs-file @var{file}
16681 @kindex set procfs-file
16682 Tell @value{GDBN} to write @code{procfs} API trace to the named
16683 @var{file}. @value{GDBN} appends the trace info to the previous
16684 contents of the file. The default is to display the trace on the
16687 @item show procfs-file
16688 @kindex show procfs-file
16689 Show the file to which @code{procfs} API trace is written.
16691 @item proc-trace-entry
16692 @itemx proc-trace-exit
16693 @itemx proc-untrace-entry
16694 @itemx proc-untrace-exit
16695 @kindex proc-trace-entry
16696 @kindex proc-trace-exit
16697 @kindex proc-untrace-entry
16698 @kindex proc-untrace-exit
16699 These commands enable and disable tracing of entries into and exits
16700 from the @code{syscall} interface.
16703 @kindex info pidlist
16704 @cindex process list, QNX Neutrino
16705 For QNX Neutrino only, this command displays the list of all the
16706 processes and all the threads within each process.
16709 @kindex info meminfo
16710 @cindex mapinfo list, QNX Neutrino
16711 For QNX Neutrino only, this command displays the list of all mapinfos.
16715 @subsection Features for Debugging @sc{djgpp} Programs
16716 @cindex @sc{djgpp} debugging
16717 @cindex native @sc{djgpp} debugging
16718 @cindex MS-DOS-specific commands
16721 @sc{djgpp} is a port of the @sc{gnu} development tools to MS-DOS and
16722 MS-Windows. @sc{djgpp} programs are 32-bit protected-mode programs
16723 that use the @dfn{DPMI} (DOS Protected-Mode Interface) API to run on
16724 top of real-mode DOS systems and their emulations.
16726 @value{GDBN} supports native debugging of @sc{djgpp} programs, and
16727 defines a few commands specific to the @sc{djgpp} port. This
16728 subsection describes those commands.
16733 This is a prefix of @sc{djgpp}-specific commands which print
16734 information about the target system and important OS structures.
16737 @cindex MS-DOS system info
16738 @cindex free memory information (MS-DOS)
16739 @item info dos sysinfo
16740 This command displays assorted information about the underlying
16741 platform: the CPU type and features, the OS version and flavor, the
16742 DPMI version, and the available conventional and DPMI memory.
16747 @cindex segment descriptor tables
16748 @cindex descriptor tables display
16750 @itemx info dos ldt
16751 @itemx info dos idt
16752 These 3 commands display entries from, respectively, Global, Local,
16753 and Interrupt Descriptor Tables (GDT, LDT, and IDT). The descriptor
16754 tables are data structures which store a descriptor for each segment
16755 that is currently in use. The segment's selector is an index into a
16756 descriptor table; the table entry for that index holds the
16757 descriptor's base address and limit, and its attributes and access
16760 A typical @sc{djgpp} program uses 3 segments: a code segment, a data
16761 segment (used for both data and the stack), and a DOS segment (which
16762 allows access to DOS/BIOS data structures and absolute addresses in
16763 conventional memory). However, the DPMI host will usually define
16764 additional segments in order to support the DPMI environment.
16766 @cindex garbled pointers
16767 These commands allow to display entries from the descriptor tables.
16768 Without an argument, all entries from the specified table are
16769 displayed. An argument, which should be an integer expression, means
16770 display a single entry whose index is given by the argument. For
16771 example, here's a convenient way to display information about the
16772 debugged program's data segment:
16775 @exdent @code{(@value{GDBP}) info dos ldt $ds}
16776 @exdent @code{0x13f: base=0x11970000 limit=0x0009ffff 32-Bit Data (Read/Write, Exp-up)}
16780 This comes in handy when you want to see whether a pointer is outside
16781 the data segment's limit (i.e.@: @dfn{garbled}).
16783 @cindex page tables display (MS-DOS)
16785 @itemx info dos pte
16786 These two commands display entries from, respectively, the Page
16787 Directory and the Page Tables. Page Directories and Page Tables are
16788 data structures which control how virtual memory addresses are mapped
16789 into physical addresses. A Page Table includes an entry for every
16790 page of memory that is mapped into the program's address space; there
16791 may be several Page Tables, each one holding up to 4096 entries. A
16792 Page Directory has up to 4096 entries, one each for every Page Table
16793 that is currently in use.
16795 Without an argument, @kbd{info dos pde} displays the entire Page
16796 Directory, and @kbd{info dos pte} displays all the entries in all of
16797 the Page Tables. An argument, an integer expression, given to the
16798 @kbd{info dos pde} command means display only that entry from the Page
16799 Directory table. An argument given to the @kbd{info dos pte} command
16800 means display entries from a single Page Table, the one pointed to by
16801 the specified entry in the Page Directory.
16803 @cindex direct memory access (DMA) on MS-DOS
16804 These commands are useful when your program uses @dfn{DMA} (Direct
16805 Memory Access), which needs physical addresses to program the DMA
16808 These commands are supported only with some DPMI servers.
16810 @cindex physical address from linear address
16811 @item info dos address-pte @var{addr}
16812 This command displays the Page Table entry for a specified linear
16813 address. The argument @var{addr} is a linear address which should
16814 already have the appropriate segment's base address added to it,
16815 because this command accepts addresses which may belong to @emph{any}
16816 segment. For example, here's how to display the Page Table entry for
16817 the page where a variable @code{i} is stored:
16820 @exdent @code{(@value{GDBP}) info dos address-pte __djgpp_base_address + (char *)&i}
16821 @exdent @code{Page Table entry for address 0x11a00d30:}
16822 @exdent @code{Base=0x02698000 Dirty Acc. Not-Cached Write-Back Usr Read-Write +0xd30}
16826 This says that @code{i} is stored at offset @code{0xd30} from the page
16827 whose physical base address is @code{0x02698000}, and shows all the
16828 attributes of that page.
16830 Note that you must cast the addresses of variables to a @code{char *},
16831 since otherwise the value of @code{__djgpp_base_address}, the base
16832 address of all variables and functions in a @sc{djgpp} program, will
16833 be added using the rules of C pointer arithmetics: if @code{i} is
16834 declared an @code{int}, @value{GDBN} will add 4 times the value of
16835 @code{__djgpp_base_address} to the address of @code{i}.
16837 Here's another example, it displays the Page Table entry for the
16841 @exdent @code{(@value{GDBP}) info dos address-pte *((unsigned *)&_go32_info_block + 3)}
16842 @exdent @code{Page Table entry for address 0x29110:}
16843 @exdent @code{Base=0x00029000 Dirty Acc. Not-Cached Write-Back Usr Read-Write +0x110}
16847 (The @code{+ 3} offset is because the transfer buffer's address is the
16848 3rd member of the @code{_go32_info_block} structure.) The output
16849 clearly shows that this DPMI server maps the addresses in conventional
16850 memory 1:1, i.e.@: the physical (@code{0x00029000} + @code{0x110}) and
16851 linear (@code{0x29110}) addresses are identical.
16853 This command is supported only with some DPMI servers.
16856 @cindex DOS serial data link, remote debugging
16857 In addition to native debugging, the DJGPP port supports remote
16858 debugging via a serial data link. The following commands are specific
16859 to remote serial debugging in the DJGPP port of @value{GDBN}.
16862 @kindex set com1base
16863 @kindex set com1irq
16864 @kindex set com2base
16865 @kindex set com2irq
16866 @kindex set com3base
16867 @kindex set com3irq
16868 @kindex set com4base
16869 @kindex set com4irq
16870 @item set com1base @var{addr}
16871 This command sets the base I/O port address of the @file{COM1} serial
16874 @item set com1irq @var{irq}
16875 This command sets the @dfn{Interrupt Request} (@code{IRQ}) line to use
16876 for the @file{COM1} serial port.
16878 There are similar commands @samp{set com2base}, @samp{set com3irq},
16879 etc.@: for setting the port address and the @code{IRQ} lines for the
16882 @kindex show com1base
16883 @kindex show com1irq
16884 @kindex show com2base
16885 @kindex show com2irq
16886 @kindex show com3base
16887 @kindex show com3irq
16888 @kindex show com4base
16889 @kindex show com4irq
16890 The related commands @samp{show com1base}, @samp{show com1irq} etc.@:
16891 display the current settings of the base address and the @code{IRQ}
16892 lines used by the COM ports.
16895 @kindex info serial
16896 @cindex DOS serial port status
16897 This command prints the status of the 4 DOS serial ports. For each
16898 port, it prints whether it's active or not, its I/O base address and
16899 IRQ number, whether it uses a 16550-style FIFO, its baudrate, and the
16900 counts of various errors encountered so far.
16904 @node Cygwin Native
16905 @subsection Features for Debugging MS Windows PE Executables
16906 @cindex MS Windows debugging
16907 @cindex native Cygwin debugging
16908 @cindex Cygwin-specific commands
16910 @value{GDBN} supports native debugging of MS Windows programs, including
16911 DLLs with and without symbolic debugging information.
16913 @cindex Ctrl-BREAK, MS-Windows
16914 @cindex interrupt debuggee on MS-Windows
16915 MS-Windows programs that call @code{SetConsoleMode} to switch off the
16916 special meaning of the @samp{Ctrl-C} keystroke cannot be interrupted
16917 by typing @kbd{C-c}. For this reason, @value{GDBN} on MS-Windows
16918 supports @kbd{C-@key{BREAK}} as an alternative interrupt key
16919 sequence, which can be used to interrupt the debuggee even if it
16922 There are various additional Cygwin-specific commands, described in
16923 this section. Working with DLLs that have no debugging symbols is
16924 described in @ref{Non-debug DLL Symbols}.
16929 This is a prefix of MS Windows-specific commands which print
16930 information about the target system and important OS structures.
16932 @item info w32 selector
16933 This command displays information returned by
16934 the Win32 API @code{GetThreadSelectorEntry} function.
16935 It takes an optional argument that is evaluated to
16936 a long value to give the information about this given selector.
16937 Without argument, this command displays information
16938 about the six segment registers.
16940 @item info w32 thread-information-block
16941 This command displays thread specific information stored in the
16942 Thread Information Block (readable on the X86 CPU family using @code{$fs}
16943 selector for 32-bit programs and @code{$gs} for 64-bit programs).
16947 This is a Cygwin-specific alias of @code{info shared}.
16949 @kindex dll-symbols
16951 This command loads symbols from a dll similarly to
16952 add-sym command but without the need to specify a base address.
16954 @kindex set cygwin-exceptions
16955 @cindex debugging the Cygwin DLL
16956 @cindex Cygwin DLL, debugging
16957 @item set cygwin-exceptions @var{mode}
16958 If @var{mode} is @code{on}, @value{GDBN} will break on exceptions that
16959 happen inside the Cygwin DLL. If @var{mode} is @code{off},
16960 @value{GDBN} will delay recognition of exceptions, and may ignore some
16961 exceptions which seem to be caused by internal Cygwin DLL
16962 ``bookkeeping''. This option is meant primarily for debugging the
16963 Cygwin DLL itself; the default value is @code{off} to avoid annoying
16964 @value{GDBN} users with false @code{SIGSEGV} signals.
16966 @kindex show cygwin-exceptions
16967 @item show cygwin-exceptions
16968 Displays whether @value{GDBN} will break on exceptions that happen
16969 inside the Cygwin DLL itself.
16971 @kindex set new-console
16972 @item set new-console @var{mode}
16973 If @var{mode} is @code{on} the debuggee will
16974 be started in a new console on next start.
16975 If @var{mode} is @code{off}, the debuggee will
16976 be started in the same console as the debugger.
16978 @kindex show new-console
16979 @item show new-console
16980 Displays whether a new console is used
16981 when the debuggee is started.
16983 @kindex set new-group
16984 @item set new-group @var{mode}
16985 This boolean value controls whether the debuggee should
16986 start a new group or stay in the same group as the debugger.
16987 This affects the way the Windows OS handles
16990 @kindex show new-group
16991 @item show new-group
16992 Displays current value of new-group boolean.
16994 @kindex set debugevents
16995 @item set debugevents
16996 This boolean value adds debug output concerning kernel events related
16997 to the debuggee seen by the debugger. This includes events that
16998 signal thread and process creation and exit, DLL loading and
16999 unloading, console interrupts, and debugging messages produced by the
17000 Windows @code{OutputDebugString} API call.
17002 @kindex set debugexec
17003 @item set debugexec
17004 This boolean value adds debug output concerning execute events
17005 (such as resume thread) seen by the debugger.
17007 @kindex set debugexceptions
17008 @item set debugexceptions
17009 This boolean value adds debug output concerning exceptions in the
17010 debuggee seen by the debugger.
17012 @kindex set debugmemory
17013 @item set debugmemory
17014 This boolean value adds debug output concerning debuggee memory reads
17015 and writes by the debugger.
17019 This boolean values specifies whether the debuggee is called
17020 via a shell or directly (default value is on).
17024 Displays if the debuggee will be started with a shell.
17029 * Non-debug DLL Symbols:: Support for DLLs without debugging symbols
17032 @node Non-debug DLL Symbols
17033 @subsubsection Support for DLLs without Debugging Symbols
17034 @cindex DLLs with no debugging symbols
17035 @cindex Minimal symbols and DLLs
17037 Very often on windows, some of the DLLs that your program relies on do
17038 not include symbolic debugging information (for example,
17039 @file{kernel32.dll}). When @value{GDBN} doesn't recognize any debugging
17040 symbols in a DLL, it relies on the minimal amount of symbolic
17041 information contained in the DLL's export table. This section
17042 describes working with such symbols, known internally to @value{GDBN} as
17043 ``minimal symbols''.
17045 Note that before the debugged program has started execution, no DLLs
17046 will have been loaded. The easiest way around this problem is simply to
17047 start the program --- either by setting a breakpoint or letting the
17048 program run once to completion. It is also possible to force
17049 @value{GDBN} to load a particular DLL before starting the executable ---
17050 see the shared library information in @ref{Files}, or the
17051 @code{dll-symbols} command in @ref{Cygwin Native}. Currently,
17052 explicitly loading symbols from a DLL with no debugging information will
17053 cause the symbol names to be duplicated in @value{GDBN}'s lookup table,
17054 which may adversely affect symbol lookup performance.
17056 @subsubsection DLL Name Prefixes
17058 In keeping with the naming conventions used by the Microsoft debugging
17059 tools, DLL export symbols are made available with a prefix based on the
17060 DLL name, for instance @code{KERNEL32!CreateFileA}. The plain name is
17061 also entered into the symbol table, so @code{CreateFileA} is often
17062 sufficient. In some cases there will be name clashes within a program
17063 (particularly if the executable itself includes full debugging symbols)
17064 necessitating the use of the fully qualified name when referring to the
17065 contents of the DLL. Use single-quotes around the name to avoid the
17066 exclamation mark (``!'') being interpreted as a language operator.
17068 Note that the internal name of the DLL may be all upper-case, even
17069 though the file name of the DLL is lower-case, or vice-versa. Since
17070 symbols within @value{GDBN} are @emph{case-sensitive} this may cause
17071 some confusion. If in doubt, try the @code{info functions} and
17072 @code{info variables} commands or even @code{maint print msymbols}
17073 (@pxref{Symbols}). Here's an example:
17076 (@value{GDBP}) info function CreateFileA
17077 All functions matching regular expression "CreateFileA":
17079 Non-debugging symbols:
17080 0x77e885f4 CreateFileA
17081 0x77e885f4 KERNEL32!CreateFileA
17085 (@value{GDBP}) info function !
17086 All functions matching regular expression "!":
17088 Non-debugging symbols:
17089 0x6100114c cygwin1!__assert
17090 0x61004034 cygwin1!_dll_crt0@@0
17091 0x61004240 cygwin1!dll_crt0(per_process *)
17095 @subsubsection Working with Minimal Symbols
17097 Symbols extracted from a DLL's export table do not contain very much
17098 type information. All that @value{GDBN} can do is guess whether a symbol
17099 refers to a function or variable depending on the linker section that
17100 contains the symbol. Also note that the actual contents of the memory
17101 contained in a DLL are not available unless the program is running. This
17102 means that you cannot examine the contents of a variable or disassemble
17103 a function within a DLL without a running program.
17105 Variables are generally treated as pointers and dereferenced
17106 automatically. For this reason, it is often necessary to prefix a
17107 variable name with the address-of operator (``&'') and provide explicit
17108 type information in the command. Here's an example of the type of
17112 (@value{GDBP}) print 'cygwin1!__argv'
17117 (@value{GDBP}) x 'cygwin1!__argv'
17118 0x10021610: "\230y\""
17121 And two possible solutions:
17124 (@value{GDBP}) print ((char **)'cygwin1!__argv')[0]
17125 $2 = 0x22fd98 "/cygdrive/c/mydirectory/myprogram"
17129 (@value{GDBP}) x/2x &'cygwin1!__argv'
17130 0x610c0aa8 <cygwin1!__argv>: 0x10021608 0x00000000
17131 (@value{GDBP}) x/x 0x10021608
17132 0x10021608: 0x0022fd98
17133 (@value{GDBP}) x/s 0x0022fd98
17134 0x22fd98: "/cygdrive/c/mydirectory/myprogram"
17137 Setting a break point within a DLL is possible even before the program
17138 starts execution. However, under these circumstances, @value{GDBN} can't
17139 examine the initial instructions of the function in order to skip the
17140 function's frame set-up code. You can work around this by using ``*&''
17141 to set the breakpoint at a raw memory address:
17144 (@value{GDBP}) break *&'python22!PyOS_Readline'
17145 Breakpoint 1 at 0x1e04eff0
17148 The author of these extensions is not entirely convinced that setting a
17149 break point within a shared DLL like @file{kernel32.dll} is completely
17153 @subsection Commands Specific to @sc{gnu} Hurd Systems
17154 @cindex @sc{gnu} Hurd debugging
17156 This subsection describes @value{GDBN} commands specific to the
17157 @sc{gnu} Hurd native debugging.
17162 @kindex set signals@r{, Hurd command}
17163 @kindex set sigs@r{, Hurd command}
17164 This command toggles the state of inferior signal interception by
17165 @value{GDBN}. Mach exceptions, such as breakpoint traps, are not
17166 affected by this command. @code{sigs} is a shorthand alias for
17171 @kindex show signals@r{, Hurd command}
17172 @kindex show sigs@r{, Hurd command}
17173 Show the current state of intercepting inferior's signals.
17175 @item set signal-thread
17176 @itemx set sigthread
17177 @kindex set signal-thread
17178 @kindex set sigthread
17179 This command tells @value{GDBN} which thread is the @code{libc} signal
17180 thread. That thread is run when a signal is delivered to a running
17181 process. @code{set sigthread} is the shorthand alias of @code{set
17184 @item show signal-thread
17185 @itemx show sigthread
17186 @kindex show signal-thread
17187 @kindex show sigthread
17188 These two commands show which thread will run when the inferior is
17189 delivered a signal.
17192 @kindex set stopped@r{, Hurd command}
17193 This commands tells @value{GDBN} that the inferior process is stopped,
17194 as with the @code{SIGSTOP} signal. The stopped process can be
17195 continued by delivering a signal to it.
17198 @kindex show stopped@r{, Hurd command}
17199 This command shows whether @value{GDBN} thinks the debuggee is
17202 @item set exceptions
17203 @kindex set exceptions@r{, Hurd command}
17204 Use this command to turn off trapping of exceptions in the inferior.
17205 When exception trapping is off, neither breakpoints nor
17206 single-stepping will work. To restore the default, set exception
17209 @item show exceptions
17210 @kindex show exceptions@r{, Hurd command}
17211 Show the current state of trapping exceptions in the inferior.
17213 @item set task pause
17214 @kindex set task@r{, Hurd commands}
17215 @cindex task attributes (@sc{gnu} Hurd)
17216 @cindex pause current task (@sc{gnu} Hurd)
17217 This command toggles task suspension when @value{GDBN} has control.
17218 Setting it to on takes effect immediately, and the task is suspended
17219 whenever @value{GDBN} gets control. Setting it to off will take
17220 effect the next time the inferior is continued. If this option is set
17221 to off, you can use @code{set thread default pause on} or @code{set
17222 thread pause on} (see below) to pause individual threads.
17224 @item show task pause
17225 @kindex show task@r{, Hurd commands}
17226 Show the current state of task suspension.
17228 @item set task detach-suspend-count
17229 @cindex task suspend count
17230 @cindex detach from task, @sc{gnu} Hurd
17231 This command sets the suspend count the task will be left with when
17232 @value{GDBN} detaches from it.
17234 @item show task detach-suspend-count
17235 Show the suspend count the task will be left with when detaching.
17237 @item set task exception-port
17238 @itemx set task excp
17239 @cindex task exception port, @sc{gnu} Hurd
17240 This command sets the task exception port to which @value{GDBN} will
17241 forward exceptions. The argument should be the value of the @dfn{send
17242 rights} of the task. @code{set task excp} is a shorthand alias.
17244 @item set noninvasive
17245 @cindex noninvasive task options
17246 This command switches @value{GDBN} to a mode that is the least
17247 invasive as far as interfering with the inferior is concerned. This
17248 is the same as using @code{set task pause}, @code{set exceptions}, and
17249 @code{set signals} to values opposite to the defaults.
17251 @item info send-rights
17252 @itemx info receive-rights
17253 @itemx info port-rights
17254 @itemx info port-sets
17255 @itemx info dead-names
17258 @cindex send rights, @sc{gnu} Hurd
17259 @cindex receive rights, @sc{gnu} Hurd
17260 @cindex port rights, @sc{gnu} Hurd
17261 @cindex port sets, @sc{gnu} Hurd
17262 @cindex dead names, @sc{gnu} Hurd
17263 These commands display information about, respectively, send rights,
17264 receive rights, port rights, port sets, and dead names of a task.
17265 There are also shorthand aliases: @code{info ports} for @code{info
17266 port-rights} and @code{info psets} for @code{info port-sets}.
17268 @item set thread pause
17269 @kindex set thread@r{, Hurd command}
17270 @cindex thread properties, @sc{gnu} Hurd
17271 @cindex pause current thread (@sc{gnu} Hurd)
17272 This command toggles current thread suspension when @value{GDBN} has
17273 control. Setting it to on takes effect immediately, and the current
17274 thread is suspended whenever @value{GDBN} gets control. Setting it to
17275 off will take effect the next time the inferior is continued.
17276 Normally, this command has no effect, since when @value{GDBN} has
17277 control, the whole task is suspended. However, if you used @code{set
17278 task pause off} (see above), this command comes in handy to suspend
17279 only the current thread.
17281 @item show thread pause
17282 @kindex show thread@r{, Hurd command}
17283 This command shows the state of current thread suspension.
17285 @item set thread run
17286 This command sets whether the current thread is allowed to run.
17288 @item show thread run
17289 Show whether the current thread is allowed to run.
17291 @item set thread detach-suspend-count
17292 @cindex thread suspend count, @sc{gnu} Hurd
17293 @cindex detach from thread, @sc{gnu} Hurd
17294 This command sets the suspend count @value{GDBN} will leave on a
17295 thread when detaching. This number is relative to the suspend count
17296 found by @value{GDBN} when it notices the thread; use @code{set thread
17297 takeover-suspend-count} to force it to an absolute value.
17299 @item show thread detach-suspend-count
17300 Show the suspend count @value{GDBN} will leave on the thread when
17303 @item set thread exception-port
17304 @itemx set thread excp
17305 Set the thread exception port to which to forward exceptions. This
17306 overrides the port set by @code{set task exception-port} (see above).
17307 @code{set thread excp} is the shorthand alias.
17309 @item set thread takeover-suspend-count
17310 Normally, @value{GDBN}'s thread suspend counts are relative to the
17311 value @value{GDBN} finds when it notices each thread. This command
17312 changes the suspend counts to be absolute instead.
17314 @item set thread default
17315 @itemx show thread default
17316 @cindex thread default settings, @sc{gnu} Hurd
17317 Each of the above @code{set thread} commands has a @code{set thread
17318 default} counterpart (e.g., @code{set thread default pause}, @code{set
17319 thread default exception-port}, etc.). The @code{thread default}
17320 variety of commands sets the default thread properties for all
17321 threads; you can then change the properties of individual threads with
17322 the non-default commands.
17327 @subsection QNX Neutrino
17328 @cindex QNX Neutrino
17330 @value{GDBN} provides the following commands specific to the QNX
17334 @item set debug nto-debug
17335 @kindex set debug nto-debug
17336 When set to on, enables debugging messages specific to the QNX
17339 @item show debug nto-debug
17340 @kindex show debug nto-debug
17341 Show the current state of QNX Neutrino messages.
17348 @value{GDBN} provides the following commands specific to the Darwin target:
17351 @item set debug darwin @var{num}
17352 @kindex set debug darwin
17353 When set to a non zero value, enables debugging messages specific to
17354 the Darwin support. Higher values produce more verbose output.
17356 @item show debug darwin
17357 @kindex show debug darwin
17358 Show the current state of Darwin messages.
17360 @item set debug mach-o @var{num}
17361 @kindex set debug mach-o
17362 When set to a non zero value, enables debugging messages while
17363 @value{GDBN} is reading Darwin object files. (@dfn{Mach-O} is the
17364 file format used on Darwin for object and executable files.) Higher
17365 values produce more verbose output. This is a command to diagnose
17366 problems internal to @value{GDBN} and should not be needed in normal
17369 @item show debug mach-o
17370 @kindex show debug mach-o
17371 Show the current state of Mach-O file messages.
17373 @item set mach-exceptions on
17374 @itemx set mach-exceptions off
17375 @kindex set mach-exceptions
17376 On Darwin, faults are first reported as a Mach exception and are then
17377 mapped to a Posix signal. Use this command to turn on trapping of
17378 Mach exceptions in the inferior. This might be sometimes useful to
17379 better understand the cause of a fault. The default is off.
17381 @item show mach-exceptions
17382 @kindex show mach-exceptions
17383 Show the current state of exceptions trapping.
17388 @section Embedded Operating Systems
17390 This section describes configurations involving the debugging of
17391 embedded operating systems that are available for several different
17395 * VxWorks:: Using @value{GDBN} with VxWorks
17398 @value{GDBN} includes the ability to debug programs running on
17399 various real-time operating systems.
17402 @subsection Using @value{GDBN} with VxWorks
17408 @kindex target vxworks
17409 @item target vxworks @var{machinename}
17410 A VxWorks system, attached via TCP/IP. The argument @var{machinename}
17411 is the target system's machine name or IP address.
17415 On VxWorks, @code{load} links @var{filename} dynamically on the
17416 current target system as well as adding its symbols in @value{GDBN}.
17418 @value{GDBN} enables developers to spawn and debug tasks running on networked
17419 VxWorks targets from a Unix host. Already-running tasks spawned from
17420 the VxWorks shell can also be debugged. @value{GDBN} uses code that runs on
17421 both the Unix host and on the VxWorks target. The program
17422 @code{@value{GDBP}} is installed and executed on the Unix host. (It may be
17423 installed with the name @code{vxgdb}, to distinguish it from a
17424 @value{GDBN} for debugging programs on the host itself.)
17427 @item VxWorks-timeout @var{args}
17428 @kindex vxworks-timeout
17429 All VxWorks-based targets now support the option @code{vxworks-timeout}.
17430 This option is set by the user, and @var{args} represents the number of
17431 seconds @value{GDBN} waits for responses to rpc's. You might use this if
17432 your VxWorks target is a slow software simulator or is on the far side
17433 of a thin network line.
17436 The following information on connecting to VxWorks was current when
17437 this manual was produced; newer releases of VxWorks may use revised
17440 @findex INCLUDE_RDB
17441 To use @value{GDBN} with VxWorks, you must rebuild your VxWorks kernel
17442 to include the remote debugging interface routines in the VxWorks
17443 library @file{rdb.a}. To do this, define @code{INCLUDE_RDB} in the
17444 VxWorks configuration file @file{configAll.h} and rebuild your VxWorks
17445 kernel. The resulting kernel contains @file{rdb.a}, and spawns the
17446 source debugging task @code{tRdbTask} when VxWorks is booted. For more
17447 information on configuring and remaking VxWorks, see the manufacturer's
17449 @c VxWorks, see the @cite{VxWorks Programmer's Guide}.
17451 Once you have included @file{rdb.a} in your VxWorks system image and set
17452 your Unix execution search path to find @value{GDBN}, you are ready to
17453 run @value{GDBN}. From your Unix host, run @code{@value{GDBP}} (or
17454 @code{vxgdb}, depending on your installation).
17456 @value{GDBN} comes up showing the prompt:
17463 * VxWorks Connection:: Connecting to VxWorks
17464 * VxWorks Download:: VxWorks download
17465 * VxWorks Attach:: Running tasks
17468 @node VxWorks Connection
17469 @subsubsection Connecting to VxWorks
17471 The @value{GDBN} command @code{target} lets you connect to a VxWorks target on the
17472 network. To connect to a target whose host name is ``@code{tt}'', type:
17475 (vxgdb) target vxworks tt
17479 @value{GDBN} displays messages like these:
17482 Attaching remote machine across net...
17487 @value{GDBN} then attempts to read the symbol tables of any object modules
17488 loaded into the VxWorks target since it was last booted. @value{GDBN} locates
17489 these files by searching the directories listed in the command search
17490 path (@pxref{Environment, ,Your Program's Environment}); if it fails
17491 to find an object file, it displays a message such as:
17494 prog.o: No such file or directory.
17497 When this happens, add the appropriate directory to the search path with
17498 the @value{GDBN} command @code{path}, and execute the @code{target}
17501 @node VxWorks Download
17502 @subsubsection VxWorks Download
17504 @cindex download to VxWorks
17505 If you have connected to the VxWorks target and you want to debug an
17506 object that has not yet been loaded, you can use the @value{GDBN}
17507 @code{load} command to download a file from Unix to VxWorks
17508 incrementally. The object file given as an argument to the @code{load}
17509 command is actually opened twice: first by the VxWorks target in order
17510 to download the code, then by @value{GDBN} in order to read the symbol
17511 table. This can lead to problems if the current working directories on
17512 the two systems differ. If both systems have NFS mounted the same
17513 filesystems, you can avoid these problems by using absolute paths.
17514 Otherwise, it is simplest to set the working directory on both systems
17515 to the directory in which the object file resides, and then to reference
17516 the file by its name, without any path. For instance, a program
17517 @file{prog.o} may reside in @file{@var{vxpath}/vw/demo/rdb} in VxWorks
17518 and in @file{@var{hostpath}/vw/demo/rdb} on the host. To load this
17519 program, type this on VxWorks:
17522 -> cd "@var{vxpath}/vw/demo/rdb"
17526 Then, in @value{GDBN}, type:
17529 (vxgdb) cd @var{hostpath}/vw/demo/rdb
17530 (vxgdb) load prog.o
17533 @value{GDBN} displays a response similar to this:
17536 Reading symbol data from wherever/vw/demo/rdb/prog.o... done.
17539 You can also use the @code{load} command to reload an object module
17540 after editing and recompiling the corresponding source file. Note that
17541 this makes @value{GDBN} delete all currently-defined breakpoints,
17542 auto-displays, and convenience variables, and to clear the value
17543 history. (This is necessary in order to preserve the integrity of
17544 debugger's data structures that reference the target system's symbol
17547 @node VxWorks Attach
17548 @subsubsection Running Tasks
17550 @cindex running VxWorks tasks
17551 You can also attach to an existing task using the @code{attach} command as
17555 (vxgdb) attach @var{task}
17559 where @var{task} is the VxWorks hexadecimal task ID. The task can be running
17560 or suspended when you attach to it. Running tasks are suspended at
17561 the time of attachment.
17563 @node Embedded Processors
17564 @section Embedded Processors
17566 This section goes into details specific to particular embedded
17569 @cindex send command to simulator
17570 Whenever a specific embedded processor has a simulator, @value{GDBN}
17571 allows to send an arbitrary command to the simulator.
17574 @item sim @var{command}
17575 @kindex sim@r{, a command}
17576 Send an arbitrary @var{command} string to the simulator. Consult the
17577 documentation for the specific simulator in use for information about
17578 acceptable commands.
17584 * M32R/D:: Renesas M32R/D
17585 * M68K:: Motorola M68K
17586 * MicroBlaze:: Xilinx MicroBlaze
17587 * MIPS Embedded:: MIPS Embedded
17588 * OpenRISC 1000:: OpenRisc 1000
17589 * PA:: HP PA Embedded
17590 * PowerPC Embedded:: PowerPC Embedded
17591 * Sparclet:: Tsqware Sparclet
17592 * Sparclite:: Fujitsu Sparclite
17593 * Z8000:: Zilog Z8000
17596 * Super-H:: Renesas Super-H
17605 @item target rdi @var{dev}
17606 ARM Angel monitor, via RDI library interface to ADP protocol. You may
17607 use this target to communicate with both boards running the Angel
17608 monitor, or with the EmbeddedICE JTAG debug device.
17611 @item target rdp @var{dev}
17616 @value{GDBN} provides the following ARM-specific commands:
17619 @item set arm disassembler
17621 This commands selects from a list of disassembly styles. The
17622 @code{"std"} style is the standard style.
17624 @item show arm disassembler
17626 Show the current disassembly style.
17628 @item set arm apcs32
17629 @cindex ARM 32-bit mode
17630 This command toggles ARM operation mode between 32-bit and 26-bit.
17632 @item show arm apcs32
17633 Display the current usage of the ARM 32-bit mode.
17635 @item set arm fpu @var{fputype}
17636 This command sets the ARM floating-point unit (FPU) type. The
17637 argument @var{fputype} can be one of these:
17641 Determine the FPU type by querying the OS ABI.
17643 Software FPU, with mixed-endian doubles on little-endian ARM
17646 GCC-compiled FPA co-processor.
17648 Software FPU with pure-endian doubles.
17654 Show the current type of the FPU.
17657 This command forces @value{GDBN} to use the specified ABI.
17660 Show the currently used ABI.
17662 @item set arm fallback-mode (arm|thumb|auto)
17663 @value{GDBN} uses the symbol table, when available, to determine
17664 whether instructions are ARM or Thumb. This command controls
17665 @value{GDBN}'s default behavior when the symbol table is not
17666 available. The default is @samp{auto}, which causes @value{GDBN} to
17667 use the current execution mode (from the @code{T} bit in the @code{CPSR}
17670 @item show arm fallback-mode
17671 Show the current fallback instruction mode.
17673 @item set arm force-mode (arm|thumb|auto)
17674 This command overrides use of the symbol table to determine whether
17675 instructions are ARM or Thumb. The default is @samp{auto}, which
17676 causes @value{GDBN} to use the symbol table and then the setting
17677 of @samp{set arm fallback-mode}.
17679 @item show arm force-mode
17680 Show the current forced instruction mode.
17682 @item set debug arm
17683 Toggle whether to display ARM-specific debugging messages from the ARM
17684 target support subsystem.
17686 @item show debug arm
17687 Show whether ARM-specific debugging messages are enabled.
17690 The following commands are available when an ARM target is debugged
17691 using the RDI interface:
17694 @item rdilogfile @r{[}@var{file}@r{]}
17696 @cindex ADP (Angel Debugger Protocol) logging
17697 Set the filename for the ADP (Angel Debugger Protocol) packet log.
17698 With an argument, sets the log file to the specified @var{file}. With
17699 no argument, show the current log file name. The default log file is
17702 @item rdilogenable @r{[}@var{arg}@r{]}
17703 @kindex rdilogenable
17704 Control logging of ADP packets. With an argument of 1 or @code{"yes"}
17705 enables logging, with an argument 0 or @code{"no"} disables it. With
17706 no arguments displays the current setting. When logging is enabled,
17707 ADP packets exchanged between @value{GDBN} and the RDI target device
17708 are logged to a file.
17710 @item set rdiromatzero
17711 @kindex set rdiromatzero
17712 @cindex ROM at zero address, RDI
17713 Tell @value{GDBN} whether the target has ROM at address 0. If on,
17714 vector catching is disabled, so that zero address can be used. If off
17715 (the default), vector catching is enabled. For this command to take
17716 effect, it needs to be invoked prior to the @code{target rdi} command.
17718 @item show rdiromatzero
17719 @kindex show rdiromatzero
17720 Show the current setting of ROM at zero address.
17722 @item set rdiheartbeat
17723 @kindex set rdiheartbeat
17724 @cindex RDI heartbeat
17725 Enable or disable RDI heartbeat packets. It is not recommended to
17726 turn on this option, since it confuses ARM and EPI JTAG interface, as
17727 well as the Angel monitor.
17729 @item show rdiheartbeat
17730 @kindex show rdiheartbeat
17731 Show the setting of RDI heartbeat packets.
17735 @item target sim @r{[}@var{simargs}@r{]} @dots{}
17736 The @value{GDBN} ARM simulator accepts the following optional arguments.
17739 @item --swi-support=@var{type}
17740 Tell the simulator which SWI interfaces to support.
17741 @var{type} may be a comma separated list of the following values.
17742 The default value is @code{all}.
17755 @subsection Renesas M32R/D and M32R/SDI
17758 @kindex target m32r
17759 @item target m32r @var{dev}
17760 Renesas M32R/D ROM monitor.
17762 @kindex target m32rsdi
17763 @item target m32rsdi @var{dev}
17764 Renesas M32R SDI server, connected via parallel port to the board.
17767 The following @value{GDBN} commands are specific to the M32R monitor:
17770 @item set download-path @var{path}
17771 @kindex set download-path
17772 @cindex find downloadable @sc{srec} files (M32R)
17773 Set the default path for finding downloadable @sc{srec} files.
17775 @item show download-path
17776 @kindex show download-path
17777 Show the default path for downloadable @sc{srec} files.
17779 @item set board-address @var{addr}
17780 @kindex set board-address
17781 @cindex M32-EVA target board address
17782 Set the IP address for the M32R-EVA target board.
17784 @item show board-address
17785 @kindex show board-address
17786 Show the current IP address of the target board.
17788 @item set server-address @var{addr}
17789 @kindex set server-address
17790 @cindex download server address (M32R)
17791 Set the IP address for the download server, which is the @value{GDBN}'s
17794 @item show server-address
17795 @kindex show server-address
17796 Display the IP address of the download server.
17798 @item upload @r{[}@var{file}@r{]}
17799 @kindex upload@r{, M32R}
17800 Upload the specified @sc{srec} @var{file} via the monitor's Ethernet
17801 upload capability. If no @var{file} argument is given, the current
17802 executable file is uploaded.
17804 @item tload @r{[}@var{file}@r{]}
17805 @kindex tload@r{, M32R}
17806 Test the @code{upload} command.
17809 The following commands are available for M32R/SDI:
17814 @cindex reset SDI connection, M32R
17815 This command resets the SDI connection.
17819 This command shows the SDI connection status.
17822 @kindex debug_chaos
17823 @cindex M32R/Chaos debugging
17824 Instructs the remote that M32R/Chaos debugging is to be used.
17826 @item use_debug_dma
17827 @kindex use_debug_dma
17828 Instructs the remote to use the DEBUG_DMA method of accessing memory.
17831 @kindex use_mon_code
17832 Instructs the remote to use the MON_CODE method of accessing memory.
17835 @kindex use_ib_break
17836 Instructs the remote to set breakpoints by IB break.
17838 @item use_dbt_break
17839 @kindex use_dbt_break
17840 Instructs the remote to set breakpoints by DBT.
17846 The Motorola m68k configuration includes ColdFire support, and a
17847 target command for the following ROM monitor.
17851 @kindex target dbug
17852 @item target dbug @var{dev}
17853 dBUG ROM monitor for Motorola ColdFire.
17858 @subsection MicroBlaze
17859 @cindex Xilinx MicroBlaze
17860 @cindex XMD, Xilinx Microprocessor Debugger
17862 The MicroBlaze is a soft-core processor supported on various Xilinx
17863 FPGAs, such as Spartan or Virtex series. Boards with these processors
17864 usually have JTAG ports which connect to a host system running the Xilinx
17865 Embedded Development Kit (EDK) or Software Development Kit (SDK).
17866 This host system is used to download the configuration bitstream to
17867 the target FPGA. The Xilinx Microprocessor Debugger (XMD) program
17868 communicates with the target board using the JTAG interface and
17869 presents a @code{gdbserver} interface to the board. By default
17870 @code{xmd} uses port @code{1234}. (While it is possible to change
17871 this default port, it requires the use of undocumented @code{xmd}
17872 commands. Contact Xilinx support if you need to do this.)
17874 Use these GDB commands to connect to the MicroBlaze target processor.
17877 @item target remote :1234
17878 Use this command to connect to the target if you are running @value{GDBN}
17879 on the same system as @code{xmd}.
17881 @item target remote @var{xmd-host}:1234
17882 Use this command to connect to the target if it is connected to @code{xmd}
17883 running on a different system named @var{xmd-host}.
17886 Use this command to download a program to the MicroBlaze target.
17888 @item set debug microblaze @var{n}
17889 Enable MicroBlaze-specific debugging messages if non-zero.
17891 @item show debug microblaze @var{n}
17892 Show MicroBlaze-specific debugging level.
17895 @node MIPS Embedded
17896 @subsection MIPS Embedded
17898 @cindex MIPS boards
17899 @value{GDBN} can use the MIPS remote debugging protocol to talk to a
17900 MIPS board attached to a serial line. This is available when
17901 you configure @value{GDBN} with @samp{--target=mips-idt-ecoff}.
17904 Use these @value{GDBN} commands to specify the connection to your target board:
17907 @item target mips @var{port}
17908 @kindex target mips @var{port}
17909 To run a program on the board, start up @code{@value{GDBP}} with the
17910 name of your program as the argument. To connect to the board, use the
17911 command @samp{target mips @var{port}}, where @var{port} is the name of
17912 the serial port connected to the board. If the program has not already
17913 been downloaded to the board, you may use the @code{load} command to
17914 download it. You can then use all the usual @value{GDBN} commands.
17916 For example, this sequence connects to the target board through a serial
17917 port, and loads and runs a program called @var{prog} through the
17921 host$ @value{GDBP} @var{prog}
17922 @value{GDBN} is free software and @dots{}
17923 (@value{GDBP}) target mips /dev/ttyb
17924 (@value{GDBP}) load @var{prog}
17928 @item target mips @var{hostname}:@var{portnumber}
17929 On some @value{GDBN} host configurations, you can specify a TCP
17930 connection (for instance, to a serial line managed by a terminal
17931 concentrator) instead of a serial port, using the syntax
17932 @samp{@var{hostname}:@var{portnumber}}.
17934 @item target pmon @var{port}
17935 @kindex target pmon @var{port}
17938 @item target ddb @var{port}
17939 @kindex target ddb @var{port}
17940 NEC's DDB variant of PMON for Vr4300.
17942 @item target lsi @var{port}
17943 @kindex target lsi @var{port}
17944 LSI variant of PMON.
17946 @kindex target r3900
17947 @item target r3900 @var{dev}
17948 Densan DVE-R3900 ROM monitor for Toshiba R3900 Mips.
17950 @kindex target array
17951 @item target array @var{dev}
17952 Array Tech LSI33K RAID controller board.
17958 @value{GDBN} also supports these special commands for MIPS targets:
17961 @item set mipsfpu double
17962 @itemx set mipsfpu single
17963 @itemx set mipsfpu none
17964 @itemx set mipsfpu auto
17965 @itemx show mipsfpu
17966 @kindex set mipsfpu
17967 @kindex show mipsfpu
17968 @cindex MIPS remote floating point
17969 @cindex floating point, MIPS remote
17970 If your target board does not support the MIPS floating point
17971 coprocessor, you should use the command @samp{set mipsfpu none} (if you
17972 need this, you may wish to put the command in your @value{GDBN} init
17973 file). This tells @value{GDBN} how to find the return value of
17974 functions which return floating point values. It also allows
17975 @value{GDBN} to avoid saving the floating point registers when calling
17976 functions on the board. If you are using a floating point coprocessor
17977 with only single precision floating point support, as on the @sc{r4650}
17978 processor, use the command @samp{set mipsfpu single}. The default
17979 double precision floating point coprocessor may be selected using
17980 @samp{set mipsfpu double}.
17982 In previous versions the only choices were double precision or no
17983 floating point, so @samp{set mipsfpu on} will select double precision
17984 and @samp{set mipsfpu off} will select no floating point.
17986 As usual, you can inquire about the @code{mipsfpu} variable with
17987 @samp{show mipsfpu}.
17989 @item set timeout @var{seconds}
17990 @itemx set retransmit-timeout @var{seconds}
17991 @itemx show timeout
17992 @itemx show retransmit-timeout
17993 @cindex @code{timeout}, MIPS protocol
17994 @cindex @code{retransmit-timeout}, MIPS protocol
17995 @kindex set timeout
17996 @kindex show timeout
17997 @kindex set retransmit-timeout
17998 @kindex show retransmit-timeout
17999 You can control the timeout used while waiting for a packet, in the MIPS
18000 remote protocol, with the @code{set timeout @var{seconds}} command. The
18001 default is 5 seconds. Similarly, you can control the timeout used while
18002 waiting for an acknowledgment of a packet with the @code{set
18003 retransmit-timeout @var{seconds}} command. The default is 3 seconds.
18004 You can inspect both values with @code{show timeout} and @code{show
18005 retransmit-timeout}. (These commands are @emph{only} available when
18006 @value{GDBN} is configured for @samp{--target=mips-idt-ecoff}.)
18008 The timeout set by @code{set timeout} does not apply when @value{GDBN}
18009 is waiting for your program to stop. In that case, @value{GDBN} waits
18010 forever because it has no way of knowing how long the program is going
18011 to run before stopping.
18013 @item set syn-garbage-limit @var{num}
18014 @kindex set syn-garbage-limit@r{, MIPS remote}
18015 @cindex synchronize with remote MIPS target
18016 Limit the maximum number of characters @value{GDBN} should ignore when
18017 it tries to synchronize with the remote target. The default is 10
18018 characters. Setting the limit to -1 means there's no limit.
18020 @item show syn-garbage-limit
18021 @kindex show syn-garbage-limit@r{, MIPS remote}
18022 Show the current limit on the number of characters to ignore when
18023 trying to synchronize with the remote system.
18025 @item set monitor-prompt @var{prompt}
18026 @kindex set monitor-prompt@r{, MIPS remote}
18027 @cindex remote monitor prompt
18028 Tell @value{GDBN} to expect the specified @var{prompt} string from the
18029 remote monitor. The default depends on the target:
18039 @item show monitor-prompt
18040 @kindex show monitor-prompt@r{, MIPS remote}
18041 Show the current strings @value{GDBN} expects as the prompt from the
18044 @item set monitor-warnings
18045 @kindex set monitor-warnings@r{, MIPS remote}
18046 Enable or disable monitor warnings about hardware breakpoints. This
18047 has effect only for the @code{lsi} target. When on, @value{GDBN} will
18048 display warning messages whose codes are returned by the @code{lsi}
18049 PMON monitor for breakpoint commands.
18051 @item show monitor-warnings
18052 @kindex show monitor-warnings@r{, MIPS remote}
18053 Show the current setting of printing monitor warnings.
18055 @item pmon @var{command}
18056 @kindex pmon@r{, MIPS remote}
18057 @cindex send PMON command
18058 This command allows sending an arbitrary @var{command} string to the
18059 monitor. The monitor must be in debug mode for this to work.
18062 @node OpenRISC 1000
18063 @subsection OpenRISC 1000
18064 @cindex OpenRISC 1000
18066 @cindex or1k boards
18067 See OR1k Architecture document (@uref{www.opencores.org}) for more information
18068 about platform and commands.
18072 @kindex target jtag
18073 @item target jtag jtag://@var{host}:@var{port}
18075 Connects to remote JTAG server.
18076 JTAG remote server can be either an or1ksim or JTAG server,
18077 connected via parallel port to the board.
18079 Example: @code{target jtag jtag://localhost:9999}
18082 @item or1ksim @var{command}
18083 If connected to @code{or1ksim} OpenRISC 1000 Architectural
18084 Simulator, proprietary commands can be executed.
18086 @kindex info or1k spr
18087 @item info or1k spr
18088 Displays spr groups.
18090 @item info or1k spr @var{group}
18091 @itemx info or1k spr @var{groupno}
18092 Displays register names in selected group.
18094 @item info or1k spr @var{group} @var{register}
18095 @itemx info or1k spr @var{register}
18096 @itemx info or1k spr @var{groupno} @var{registerno}
18097 @itemx info or1k spr @var{registerno}
18098 Shows information about specified spr register.
18101 @item spr @var{group} @var{register} @var{value}
18102 @itemx spr @var{register @var{value}}
18103 @itemx spr @var{groupno} @var{registerno @var{value}}
18104 @itemx spr @var{registerno @var{value}}
18105 Writes @var{value} to specified spr register.
18108 Some implementations of OpenRISC 1000 Architecture also have hardware trace.
18109 It is very similar to @value{GDBN} trace, except it does not interfere with normal
18110 program execution and is thus much faster. Hardware breakpoints/watchpoint
18111 triggers can be set using:
18114 Load effective address/data
18116 Store effective address/data
18118 Access effective address ($SEA or $LEA) or data ($SDATA/$LDATA)
18123 When triggered, it can capture low level data, like: @code{PC}, @code{LSEA},
18124 @code{LDATA}, @code{SDATA}, @code{READSPR}, @code{WRITESPR}, @code{INSTR}.
18126 @code{htrace} commands:
18127 @cindex OpenRISC 1000 htrace
18130 @item hwatch @var{conditional}
18131 Set hardware watchpoint on combination of Load/Store Effective Address(es)
18132 or Data. For example:
18134 @code{hwatch ($LEA == my_var) && ($LDATA < 50) || ($SEA == my_var) && ($SDATA >= 50)}
18136 @code{hwatch ($LEA == my_var) && ($LDATA < 50) || ($SEA == my_var) && ($SDATA >= 50)}
18140 Display information about current HW trace configuration.
18142 @item htrace trigger @var{conditional}
18143 Set starting criteria for HW trace.
18145 @item htrace qualifier @var{conditional}
18146 Set acquisition qualifier for HW trace.
18148 @item htrace stop @var{conditional}
18149 Set HW trace stopping criteria.
18151 @item htrace record [@var{data}]*
18152 Selects the data to be recorded, when qualifier is met and HW trace was
18155 @item htrace enable
18156 @itemx htrace disable
18157 Enables/disables the HW trace.
18159 @item htrace rewind [@var{filename}]
18160 Clears currently recorded trace data.
18162 If filename is specified, new trace file is made and any newly collected data
18163 will be written there.
18165 @item htrace print [@var{start} [@var{len}]]
18166 Prints trace buffer, using current record configuration.
18168 @item htrace mode continuous
18169 Set continuous trace mode.
18171 @item htrace mode suspend
18172 Set suspend trace mode.
18176 @node PowerPC Embedded
18177 @subsection PowerPC Embedded
18179 @value{GDBN} provides the following PowerPC-specific commands:
18182 @kindex set powerpc
18183 @item set powerpc soft-float
18184 @itemx show powerpc soft-float
18185 Force @value{GDBN} to use (or not use) a software floating point calling
18186 convention. By default, @value{GDBN} selects the calling convention based
18187 on the selected architecture and the provided executable file.
18189 @item set powerpc vector-abi
18190 @itemx show powerpc vector-abi
18191 Force @value{GDBN} to use the specified calling convention for vector
18192 arguments and return values. The valid options are @samp{auto};
18193 @samp{generic}, to avoid vector registers even if they are present;
18194 @samp{altivec}, to use AltiVec registers; and @samp{spe} to use SPE
18195 registers. By default, @value{GDBN} selects the calling convention
18196 based on the selected architecture and the provided executable file.
18198 @kindex target dink32
18199 @item target dink32 @var{dev}
18200 DINK32 ROM monitor.
18202 @kindex target ppcbug
18203 @item target ppcbug @var{dev}
18204 @kindex target ppcbug1
18205 @item target ppcbug1 @var{dev}
18206 PPCBUG ROM monitor for PowerPC.
18209 @item target sds @var{dev}
18210 SDS monitor, running on a PowerPC board (such as Motorola's ADS).
18213 @cindex SDS protocol
18214 The following commands specific to the SDS protocol are supported
18218 @item set sdstimeout @var{nsec}
18219 @kindex set sdstimeout
18220 Set the timeout for SDS protocol reads to be @var{nsec} seconds. The
18221 default is 2 seconds.
18223 @item show sdstimeout
18224 @kindex show sdstimeout
18225 Show the current value of the SDS timeout.
18227 @item sds @var{command}
18228 @kindex sds@r{, a command}
18229 Send the specified @var{command} string to the SDS monitor.
18234 @subsection HP PA Embedded
18238 @kindex target op50n
18239 @item target op50n @var{dev}
18240 OP50N monitor, running on an OKI HPPA board.
18242 @kindex target w89k
18243 @item target w89k @var{dev}
18244 W89K monitor, running on a Winbond HPPA board.
18249 @subsection Tsqware Sparclet
18253 @value{GDBN} enables developers to debug tasks running on
18254 Sparclet targets from a Unix host.
18255 @value{GDBN} uses code that runs on
18256 both the Unix host and on the Sparclet target. The program
18257 @code{@value{GDBP}} is installed and executed on the Unix host.
18260 @item remotetimeout @var{args}
18261 @kindex remotetimeout
18262 @value{GDBN} supports the option @code{remotetimeout}.
18263 This option is set by the user, and @var{args} represents the number of
18264 seconds @value{GDBN} waits for responses.
18267 @cindex compiling, on Sparclet
18268 When compiling for debugging, include the options @samp{-g} to get debug
18269 information and @samp{-Ttext} to relocate the program to where you wish to
18270 load it on the target. You may also want to add the options @samp{-n} or
18271 @samp{-N} in order to reduce the size of the sections. Example:
18274 sparclet-aout-gcc prog.c -Ttext 0x12010000 -g -o prog -N
18277 You can use @code{objdump} to verify that the addresses are what you intended:
18280 sparclet-aout-objdump --headers --syms prog
18283 @cindex running, on Sparclet
18285 your Unix execution search path to find @value{GDBN}, you are ready to
18286 run @value{GDBN}. From your Unix host, run @code{@value{GDBP}}
18287 (or @code{sparclet-aout-gdb}, depending on your installation).
18289 @value{GDBN} comes up showing the prompt:
18296 * Sparclet File:: Setting the file to debug
18297 * Sparclet Connection:: Connecting to Sparclet
18298 * Sparclet Download:: Sparclet download
18299 * Sparclet Execution:: Running and debugging
18302 @node Sparclet File
18303 @subsubsection Setting File to Debug
18305 The @value{GDBN} command @code{file} lets you choose with program to debug.
18308 (gdbslet) file prog
18312 @value{GDBN} then attempts to read the symbol table of @file{prog}.
18313 @value{GDBN} locates
18314 the file by searching the directories listed in the command search
18316 If the file was compiled with debug information (option @samp{-g}), source
18317 files will be searched as well.
18318 @value{GDBN} locates
18319 the source files by searching the directories listed in the directory search
18320 path (@pxref{Environment, ,Your Program's Environment}).
18322 to find a file, it displays a message such as:
18325 prog: No such file or directory.
18328 When this happens, add the appropriate directories to the search paths with
18329 the @value{GDBN} commands @code{path} and @code{dir}, and execute the
18330 @code{target} command again.
18332 @node Sparclet Connection
18333 @subsubsection Connecting to Sparclet
18335 The @value{GDBN} command @code{target} lets you connect to a Sparclet target.
18336 To connect to a target on serial port ``@code{ttya}'', type:
18339 (gdbslet) target sparclet /dev/ttya
18340 Remote target sparclet connected to /dev/ttya
18341 main () at ../prog.c:3
18345 @value{GDBN} displays messages like these:
18351 @node Sparclet Download
18352 @subsubsection Sparclet Download
18354 @cindex download to Sparclet
18355 Once connected to the Sparclet target,
18356 you can use the @value{GDBN}
18357 @code{load} command to download the file from the host to the target.
18358 The file name and load offset should be given as arguments to the @code{load}
18360 Since the file format is aout, the program must be loaded to the starting
18361 address. You can use @code{objdump} to find out what this value is. The load
18362 offset is an offset which is added to the VMA (virtual memory address)
18363 of each of the file's sections.
18364 For instance, if the program
18365 @file{prog} was linked to text address 0x1201000, with data at 0x12010160
18366 and bss at 0x12010170, in @value{GDBN}, type:
18369 (gdbslet) load prog 0x12010000
18370 Loading section .text, size 0xdb0 vma 0x12010000
18373 If the code is loaded at a different address then what the program was linked
18374 to, you may need to use the @code{section} and @code{add-symbol-file} commands
18375 to tell @value{GDBN} where to map the symbol table.
18377 @node Sparclet Execution
18378 @subsubsection Running and Debugging
18380 @cindex running and debugging Sparclet programs
18381 You can now begin debugging the task using @value{GDBN}'s execution control
18382 commands, @code{b}, @code{step}, @code{run}, etc. See the @value{GDBN}
18383 manual for the list of commands.
18387 Breakpoint 1 at 0x12010000: file prog.c, line 3.
18389 Starting program: prog
18390 Breakpoint 1, main (argc=1, argv=0xeffff21c) at prog.c:3
18391 3 char *symarg = 0;
18393 4 char *execarg = "hello!";
18398 @subsection Fujitsu Sparclite
18402 @kindex target sparclite
18403 @item target sparclite @var{dev}
18404 Fujitsu sparclite boards, used only for the purpose of loading.
18405 You must use an additional command to debug the program.
18406 For example: target remote @var{dev} using @value{GDBN} standard
18412 @subsection Zilog Z8000
18415 @cindex simulator, Z8000
18416 @cindex Zilog Z8000 simulator
18418 When configured for debugging Zilog Z8000 targets, @value{GDBN} includes
18421 For the Z8000 family, @samp{target sim} simulates either the Z8002 (the
18422 unsegmented variant of the Z8000 architecture) or the Z8001 (the
18423 segmented variant). The simulator recognizes which architecture is
18424 appropriate by inspecting the object code.
18427 @item target sim @var{args}
18429 @kindex target sim@r{, with Z8000}
18430 Debug programs on a simulated CPU. If the simulator supports setup
18431 options, specify them via @var{args}.
18435 After specifying this target, you can debug programs for the simulated
18436 CPU in the same style as programs for your host computer; use the
18437 @code{file} command to load a new program image, the @code{run} command
18438 to run your program, and so on.
18440 As well as making available all the usual machine registers
18441 (@pxref{Registers, ,Registers}), the Z8000 simulator provides three
18442 additional items of information as specially named registers:
18447 Counts clock-ticks in the simulator.
18450 Counts instructions run in the simulator.
18453 Execution time in 60ths of a second.
18457 You can refer to these values in @value{GDBN} expressions with the usual
18458 conventions; for example, @w{@samp{b fputc if $cycles>5000}} sets a
18459 conditional breakpoint that suspends only after at least 5000
18460 simulated clock ticks.
18463 @subsection Atmel AVR
18466 When configured for debugging the Atmel AVR, @value{GDBN} supports the
18467 following AVR-specific commands:
18470 @item info io_registers
18471 @kindex info io_registers@r{, AVR}
18472 @cindex I/O registers (Atmel AVR)
18473 This command displays information about the AVR I/O registers. For
18474 each register, @value{GDBN} prints its number and value.
18481 When configured for debugging CRIS, @value{GDBN} provides the
18482 following CRIS-specific commands:
18485 @item set cris-version @var{ver}
18486 @cindex CRIS version
18487 Set the current CRIS version to @var{ver}, either @samp{10} or @samp{32}.
18488 The CRIS version affects register names and sizes. This command is useful in
18489 case autodetection of the CRIS version fails.
18491 @item show cris-version
18492 Show the current CRIS version.
18494 @item set cris-dwarf2-cfi
18495 @cindex DWARF-2 CFI and CRIS
18496 Set the usage of DWARF-2 CFI for CRIS debugging. The default is @samp{on}.
18497 Change to @samp{off} when using @code{gcc-cris} whose version is below
18500 @item show cris-dwarf2-cfi
18501 Show the current state of using DWARF-2 CFI.
18503 @item set cris-mode @var{mode}
18505 Set the current CRIS mode to @var{mode}. It should only be changed when
18506 debugging in guru mode, in which case it should be set to
18507 @samp{guru} (the default is @samp{normal}).
18509 @item show cris-mode
18510 Show the current CRIS mode.
18514 @subsection Renesas Super-H
18517 For the Renesas Super-H processor, @value{GDBN} provides these
18522 @kindex regs@r{, Super-H}
18523 Show the values of all Super-H registers.
18525 @item set sh calling-convention @var{convention}
18526 @kindex set sh calling-convention
18527 Set the calling-convention used when calling functions from @value{GDBN}.
18528 Allowed values are @samp{gcc}, which is the default setting, and @samp{renesas}.
18529 With the @samp{gcc} setting, functions are called using the @value{NGCC} calling
18530 convention. If the DWARF-2 information of the called function specifies
18531 that the function follows the Renesas calling convention, the function
18532 is called using the Renesas calling convention. If the calling convention
18533 is set to @samp{renesas}, the Renesas calling convention is always used,
18534 regardless of the DWARF-2 information. This can be used to override the
18535 default of @samp{gcc} if debug information is missing, or the compiler
18536 does not emit the DWARF-2 calling convention entry for a function.
18538 @item show sh calling-convention
18539 @kindex show sh calling-convention
18540 Show the current calling convention setting.
18545 @node Architectures
18546 @section Architectures
18548 This section describes characteristics of architectures that affect
18549 all uses of @value{GDBN} with the architecture, both native and cross.
18556 * HPPA:: HP PA architecture
18557 * SPU:: Cell Broadband Engine SPU architecture
18562 @subsection x86 Architecture-specific Issues
18565 @item set struct-convention @var{mode}
18566 @kindex set struct-convention
18567 @cindex struct return convention
18568 @cindex struct/union returned in registers
18569 Set the convention used by the inferior to return @code{struct}s and
18570 @code{union}s from functions to @var{mode}. Possible values of
18571 @var{mode} are @code{"pcc"}, @code{"reg"}, and @code{"default"} (the
18572 default). @code{"default"} or @code{"pcc"} means that @code{struct}s
18573 are returned on the stack, while @code{"reg"} means that a
18574 @code{struct} or a @code{union} whose size is 1, 2, 4, or 8 bytes will
18575 be returned in a register.
18577 @item show struct-convention
18578 @kindex show struct-convention
18579 Show the current setting of the convention to return @code{struct}s
18588 @kindex set rstack_high_address
18589 @cindex AMD 29K register stack
18590 @cindex register stack, AMD29K
18591 @item set rstack_high_address @var{address}
18592 On AMD 29000 family processors, registers are saved in a separate
18593 @dfn{register stack}. There is no way for @value{GDBN} to determine the
18594 extent of this stack. Normally, @value{GDBN} just assumes that the
18595 stack is ``large enough''. This may result in @value{GDBN} referencing
18596 memory locations that do not exist. If necessary, you can get around
18597 this problem by specifying the ending address of the register stack with
18598 the @code{set rstack_high_address} command. The argument should be an
18599 address, which you probably want to precede with @samp{0x} to specify in
18602 @kindex show rstack_high_address
18603 @item show rstack_high_address
18604 Display the current limit of the register stack, on AMD 29000 family
18612 See the following section.
18617 @cindex stack on Alpha
18618 @cindex stack on MIPS
18619 @cindex Alpha stack
18621 Alpha- and MIPS-based computers use an unusual stack frame, which
18622 sometimes requires @value{GDBN} to search backward in the object code to
18623 find the beginning of a function.
18625 @cindex response time, MIPS debugging
18626 To improve response time (especially for embedded applications, where
18627 @value{GDBN} may be restricted to a slow serial line for this search)
18628 you may want to limit the size of this search, using one of these
18632 @cindex @code{heuristic-fence-post} (Alpha, MIPS)
18633 @item set heuristic-fence-post @var{limit}
18634 Restrict @value{GDBN} to examining at most @var{limit} bytes in its
18635 search for the beginning of a function. A value of @var{0} (the
18636 default) means there is no limit. However, except for @var{0}, the
18637 larger the limit the more bytes @code{heuristic-fence-post} must search
18638 and therefore the longer it takes to run. You should only need to use
18639 this command when debugging a stripped executable.
18641 @item show heuristic-fence-post
18642 Display the current limit.
18646 These commands are available @emph{only} when @value{GDBN} is configured
18647 for debugging programs on Alpha or MIPS processors.
18649 Several MIPS-specific commands are available when debugging MIPS
18653 @item set mips abi @var{arg}
18654 @kindex set mips abi
18655 @cindex set ABI for MIPS
18656 Tell @value{GDBN} which MIPS ABI is used by the inferior. Possible
18657 values of @var{arg} are:
18661 The default ABI associated with the current binary (this is the
18672 @item show mips abi
18673 @kindex show mips abi
18674 Show the MIPS ABI used by @value{GDBN} to debug the inferior.
18677 @itemx show mipsfpu
18678 @xref{MIPS Embedded, set mipsfpu}.
18680 @item set mips mask-address @var{arg}
18681 @kindex set mips mask-address
18682 @cindex MIPS addresses, masking
18683 This command determines whether the most-significant 32 bits of 64-bit
18684 MIPS addresses are masked off. The argument @var{arg} can be
18685 @samp{on}, @samp{off}, or @samp{auto}. The latter is the default
18686 setting, which lets @value{GDBN} determine the correct value.
18688 @item show mips mask-address
18689 @kindex show mips mask-address
18690 Show whether the upper 32 bits of MIPS addresses are masked off or
18693 @item set remote-mips64-transfers-32bit-regs
18694 @kindex set remote-mips64-transfers-32bit-regs
18695 This command controls compatibility with 64-bit MIPS targets that
18696 transfer data in 32-bit quantities. If you have an old MIPS 64 target
18697 that transfers 32 bits for some registers, like @sc{sr} and @sc{fsr},
18698 and 64 bits for other registers, set this option to @samp{on}.
18700 @item show remote-mips64-transfers-32bit-regs
18701 @kindex show remote-mips64-transfers-32bit-regs
18702 Show the current setting of compatibility with older MIPS 64 targets.
18704 @item set debug mips
18705 @kindex set debug mips
18706 This command turns on and off debugging messages for the MIPS-specific
18707 target code in @value{GDBN}.
18709 @item show debug mips
18710 @kindex show debug mips
18711 Show the current setting of MIPS debugging messages.
18717 @cindex HPPA support
18719 When @value{GDBN} is debugging the HP PA architecture, it provides the
18720 following special commands:
18723 @item set debug hppa
18724 @kindex set debug hppa
18725 This command determines whether HPPA architecture-specific debugging
18726 messages are to be displayed.
18728 @item show debug hppa
18729 Show whether HPPA debugging messages are displayed.
18731 @item maint print unwind @var{address}
18732 @kindex maint print unwind@r{, HPPA}
18733 This command displays the contents of the unwind table entry at the
18734 given @var{address}.
18740 @subsection Cell Broadband Engine SPU architecture
18741 @cindex Cell Broadband Engine
18744 When @value{GDBN} is debugging the Cell Broadband Engine SPU architecture,
18745 it provides the following special commands:
18748 @item info spu event
18750 Display SPU event facility status. Shows current event mask
18751 and pending event status.
18753 @item info spu signal
18754 Display SPU signal notification facility status. Shows pending
18755 signal-control word and signal notification mode of both signal
18756 notification channels.
18758 @item info spu mailbox
18759 Display SPU mailbox facility status. Shows all pending entries,
18760 in order of processing, in each of the SPU Write Outbound,
18761 SPU Write Outbound Interrupt, and SPU Read Inbound mailboxes.
18764 Display MFC DMA status. Shows all pending commands in the MFC
18765 DMA queue. For each entry, opcode, tag, class IDs, effective
18766 and local store addresses and transfer size are shown.
18768 @item info spu proxydma
18769 Display MFC Proxy-DMA status. Shows all pending commands in the MFC
18770 Proxy-DMA queue. For each entry, opcode, tag, class IDs, effective
18771 and local store addresses and transfer size are shown.
18775 When @value{GDBN} is debugging a combined PowerPC/SPU application
18776 on the Cell Broadband Engine, it provides in addition the following
18780 @item set spu stop-on-load @var{arg}
18782 Set whether to stop for new SPE threads. When set to @code{on}, @value{GDBN}
18783 will give control to the user when a new SPE thread enters its @code{main}
18784 function. The default is @code{off}.
18786 @item show spu stop-on-load
18788 Show whether to stop for new SPE threads.
18790 @item set spu auto-flush-cache @var{arg}
18791 Set whether to automatically flush the software-managed cache. When set to
18792 @code{on}, @value{GDBN} will automatically cause the SPE software-managed
18793 cache to be flushed whenever SPE execution stops. This provides a consistent
18794 view of PowerPC memory that is accessed via the cache. If an application
18795 does not use the software-managed cache, this option has no effect.
18797 @item show spu auto-flush-cache
18798 Show whether to automatically flush the software-managed cache.
18803 @subsection PowerPC
18804 @cindex PowerPC architecture
18806 When @value{GDBN} is debugging the PowerPC architecture, it provides a set of
18807 pseudo-registers to enable inspection of 128-bit wide Decimal Floating Point
18808 numbers stored in the floating point registers. These values must be stored
18809 in two consecutive registers, always starting at an even register like
18810 @code{f0} or @code{f2}.
18812 The pseudo-registers go from @code{$dl0} through @code{$dl15}, and are formed
18813 by joining the even/odd register pairs @code{f0} and @code{f1} for @code{$dl0},
18814 @code{f2} and @code{f3} for @code{$dl1} and so on.
18816 For POWER7 processors, @value{GDBN} provides a set of pseudo-registers, the 64-bit
18817 wide Extended Floating Point Registers (@samp{f32} through @samp{f63}).
18820 @node Controlling GDB
18821 @chapter Controlling @value{GDBN}
18823 You can alter the way @value{GDBN} interacts with you by using the
18824 @code{set} command. For commands controlling how @value{GDBN} displays
18825 data, see @ref{Print Settings, ,Print Settings}. Other settings are
18830 * Editing:: Command editing
18831 * Command History:: Command history
18832 * Screen Size:: Screen size
18833 * Numbers:: Numbers
18834 * ABI:: Configuring the current ABI
18835 * Messages/Warnings:: Optional warnings and messages
18836 * Debugging Output:: Optional messages about internal happenings
18837 * Other Misc Settings:: Other Miscellaneous Settings
18845 @value{GDBN} indicates its readiness to read a command by printing a string
18846 called the @dfn{prompt}. This string is normally @samp{(@value{GDBP})}. You
18847 can change the prompt string with the @code{set prompt} command. For
18848 instance, when debugging @value{GDBN} with @value{GDBN}, it is useful to change
18849 the prompt in one of the @value{GDBN} sessions so that you can always tell
18850 which one you are talking to.
18852 @emph{Note:} @code{set prompt} does not add a space for you after the
18853 prompt you set. This allows you to set a prompt which ends in a space
18854 or a prompt that does not.
18858 @item set prompt @var{newprompt}
18859 Directs @value{GDBN} to use @var{newprompt} as its prompt string henceforth.
18861 @kindex show prompt
18863 Prints a line of the form: @samp{Gdb's prompt is: @var{your-prompt}}
18867 @section Command Editing
18869 @cindex command line editing
18871 @value{GDBN} reads its input commands via the @dfn{Readline} interface. This
18872 @sc{gnu} library provides consistent behavior for programs which provide a
18873 command line interface to the user. Advantages are @sc{gnu} Emacs-style
18874 or @dfn{vi}-style inline editing of commands, @code{csh}-like history
18875 substitution, and a storage and recall of command history across
18876 debugging sessions.
18878 You may control the behavior of command line editing in @value{GDBN} with the
18879 command @code{set}.
18882 @kindex set editing
18885 @itemx set editing on
18886 Enable command line editing (enabled by default).
18888 @item set editing off
18889 Disable command line editing.
18891 @kindex show editing
18893 Show whether command line editing is enabled.
18896 @xref{Command Line Editing}, for more details about the Readline
18897 interface. Users unfamiliar with @sc{gnu} Emacs or @code{vi} are
18898 encouraged to read that chapter.
18900 @node Command History
18901 @section Command History
18902 @cindex command history
18904 @value{GDBN} can keep track of the commands you type during your
18905 debugging sessions, so that you can be certain of precisely what
18906 happened. Use these commands to manage the @value{GDBN} command
18909 @value{GDBN} uses the @sc{gnu} History library, a part of the Readline
18910 package, to provide the history facility. @xref{Using History
18911 Interactively}, for the detailed description of the History library.
18913 To issue a command to @value{GDBN} without affecting certain aspects of
18914 the state which is seen by users, prefix it with @samp{server }
18915 (@pxref{Server Prefix}). This
18916 means that this command will not affect the command history, nor will it
18917 affect @value{GDBN}'s notion of which command to repeat if @key{RET} is
18918 pressed on a line by itself.
18920 @cindex @code{server}, command prefix
18921 The server prefix does not affect the recording of values into the value
18922 history; to print a value without recording it into the value history,
18923 use the @code{output} command instead of the @code{print} command.
18925 Here is the description of @value{GDBN} commands related to command
18929 @cindex history substitution
18930 @cindex history file
18931 @kindex set history filename
18932 @cindex @env{GDBHISTFILE}, environment variable
18933 @item set history filename @var{fname}
18934 Set the name of the @value{GDBN} command history file to @var{fname}.
18935 This is the file where @value{GDBN} reads an initial command history
18936 list, and where it writes the command history from this session when it
18937 exits. You can access this list through history expansion or through
18938 the history command editing characters listed below. This file defaults
18939 to the value of the environment variable @code{GDBHISTFILE}, or to
18940 @file{./.gdb_history} (@file{./_gdb_history} on MS-DOS) if this variable
18943 @cindex save command history
18944 @kindex set history save
18945 @item set history save
18946 @itemx set history save on
18947 Record command history in a file, whose name may be specified with the
18948 @code{set history filename} command. By default, this option is disabled.
18950 @item set history save off
18951 Stop recording command history in a file.
18953 @cindex history size
18954 @kindex set history size
18955 @cindex @env{HISTSIZE}, environment variable
18956 @item set history size @var{size}
18957 Set the number of commands which @value{GDBN} keeps in its history list.
18958 This defaults to the value of the environment variable
18959 @code{HISTSIZE}, or to 256 if this variable is not set.
18962 History expansion assigns special meaning to the character @kbd{!}.
18963 @xref{Event Designators}, for more details.
18965 @cindex history expansion, turn on/off
18966 Since @kbd{!} is also the logical not operator in C, history expansion
18967 is off by default. If you decide to enable history expansion with the
18968 @code{set history expansion on} command, you may sometimes need to
18969 follow @kbd{!} (when it is used as logical not, in an expression) with
18970 a space or a tab to prevent it from being expanded. The readline
18971 history facilities do not attempt substitution on the strings
18972 @kbd{!=} and @kbd{!(}, even when history expansion is enabled.
18974 The commands to control history expansion are:
18977 @item set history expansion on
18978 @itemx set history expansion
18979 @kindex set history expansion
18980 Enable history expansion. History expansion is off by default.
18982 @item set history expansion off
18983 Disable history expansion.
18986 @kindex show history
18988 @itemx show history filename
18989 @itemx show history save
18990 @itemx show history size
18991 @itemx show history expansion
18992 These commands display the state of the @value{GDBN} history parameters.
18993 @code{show history} by itself displays all four states.
18998 @kindex show commands
18999 @cindex show last commands
19000 @cindex display command history
19001 @item show commands
19002 Display the last ten commands in the command history.
19004 @item show commands @var{n}
19005 Print ten commands centered on command number @var{n}.
19007 @item show commands +
19008 Print ten commands just after the commands last printed.
19012 @section Screen Size
19013 @cindex size of screen
19014 @cindex pauses in output
19016 Certain commands to @value{GDBN} may produce large amounts of
19017 information output to the screen. To help you read all of it,
19018 @value{GDBN} pauses and asks you for input at the end of each page of
19019 output. Type @key{RET} when you want to continue the output, or @kbd{q}
19020 to discard the remaining output. Also, the screen width setting
19021 determines when to wrap lines of output. Depending on what is being
19022 printed, @value{GDBN} tries to break the line at a readable place,
19023 rather than simply letting it overflow onto the following line.
19025 Normally @value{GDBN} knows the size of the screen from the terminal
19026 driver software. For example, on Unix @value{GDBN} uses the termcap data base
19027 together with the value of the @code{TERM} environment variable and the
19028 @code{stty rows} and @code{stty cols} settings. If this is not correct,
19029 you can override it with the @code{set height} and @code{set
19036 @kindex show height
19037 @item set height @var{lpp}
19039 @itemx set width @var{cpl}
19041 These @code{set} commands specify a screen height of @var{lpp} lines and
19042 a screen width of @var{cpl} characters. The associated @code{show}
19043 commands display the current settings.
19045 If you specify a height of zero lines, @value{GDBN} does not pause during
19046 output no matter how long the output is. This is useful if output is to a
19047 file or to an editor buffer.
19049 Likewise, you can specify @samp{set width 0} to prevent @value{GDBN}
19050 from wrapping its output.
19052 @item set pagination on
19053 @itemx set pagination off
19054 @kindex set pagination
19055 Turn the output pagination on or off; the default is on. Turning
19056 pagination off is the alternative to @code{set height 0}. Note that
19057 running @value{GDBN} with the @option{--batch} option (@pxref{Mode
19058 Options, -batch}) also automatically disables pagination.
19060 @item show pagination
19061 @kindex show pagination
19062 Show the current pagination mode.
19067 @cindex number representation
19068 @cindex entering numbers
19070 You can always enter numbers in octal, decimal, or hexadecimal in
19071 @value{GDBN} by the usual conventions: octal numbers begin with
19072 @samp{0}, decimal numbers end with @samp{.}, and hexadecimal numbers
19073 begin with @samp{0x}. Numbers that neither begin with @samp{0} or
19074 @samp{0x}, nor end with a @samp{.} are, by default, entered in base
19075 10; likewise, the default display for numbers---when no particular
19076 format is specified---is base 10. You can change the default base for
19077 both input and output with the commands described below.
19080 @kindex set input-radix
19081 @item set input-radix @var{base}
19082 Set the default base for numeric input. Supported choices
19083 for @var{base} are decimal 8, 10, or 16. @var{base} must itself be
19084 specified either unambiguously or using the current input radix; for
19088 set input-radix 012
19089 set input-radix 10.
19090 set input-radix 0xa
19094 sets the input base to decimal. On the other hand, @samp{set input-radix 10}
19095 leaves the input radix unchanged, no matter what it was, since
19096 @samp{10}, being without any leading or trailing signs of its base, is
19097 interpreted in the current radix. Thus, if the current radix is 16,
19098 @samp{10} is interpreted in hex, i.e.@: as 16 decimal, which doesn't
19101 @kindex set output-radix
19102 @item set output-radix @var{base}
19103 Set the default base for numeric display. Supported choices
19104 for @var{base} are decimal 8, 10, or 16. @var{base} must itself be
19105 specified either unambiguously or using the current input radix.
19107 @kindex show input-radix
19108 @item show input-radix
19109 Display the current default base for numeric input.
19111 @kindex show output-radix
19112 @item show output-radix
19113 Display the current default base for numeric display.
19115 @item set radix @r{[}@var{base}@r{]}
19119 These commands set and show the default base for both input and output
19120 of numbers. @code{set radix} sets the radix of input and output to
19121 the same base; without an argument, it resets the radix back to its
19122 default value of 10.
19127 @section Configuring the Current ABI
19129 @value{GDBN} can determine the @dfn{ABI} (Application Binary Interface) of your
19130 application automatically. However, sometimes you need to override its
19131 conclusions. Use these commands to manage @value{GDBN}'s view of the
19138 One @value{GDBN} configuration can debug binaries for multiple operating
19139 system targets, either via remote debugging or native emulation.
19140 @value{GDBN} will autodetect the @dfn{OS ABI} (Operating System ABI) in use,
19141 but you can override its conclusion using the @code{set osabi} command.
19142 One example where this is useful is in debugging of binaries which use
19143 an alternate C library (e.g.@: @sc{uClibc} for @sc{gnu}/Linux) which does
19144 not have the same identifying marks that the standard C library for your
19149 Show the OS ABI currently in use.
19152 With no argument, show the list of registered available OS ABI's.
19154 @item set osabi @var{abi}
19155 Set the current OS ABI to @var{abi}.
19158 @cindex float promotion
19160 Generally, the way that an argument of type @code{float} is passed to a
19161 function depends on whether the function is prototyped. For a prototyped
19162 (i.e.@: ANSI/ISO style) function, @code{float} arguments are passed unchanged,
19163 according to the architecture's convention for @code{float}. For unprototyped
19164 (i.e.@: K&R style) functions, @code{float} arguments are first promoted to type
19165 @code{double} and then passed.
19167 Unfortunately, some forms of debug information do not reliably indicate whether
19168 a function is prototyped. If @value{GDBN} calls a function that is not marked
19169 as prototyped, it consults @kbd{set coerce-float-to-double}.
19172 @kindex set coerce-float-to-double
19173 @item set coerce-float-to-double
19174 @itemx set coerce-float-to-double on
19175 Arguments of type @code{float} will be promoted to @code{double} when passed
19176 to an unprototyped function. This is the default setting.
19178 @item set coerce-float-to-double off
19179 Arguments of type @code{float} will be passed directly to unprototyped
19182 @kindex show coerce-float-to-double
19183 @item show coerce-float-to-double
19184 Show the current setting of promoting @code{float} to @code{double}.
19188 @kindex show cp-abi
19189 @value{GDBN} needs to know the ABI used for your program's C@t{++}
19190 objects. The correct C@t{++} ABI depends on which C@t{++} compiler was
19191 used to build your application. @value{GDBN} only fully supports
19192 programs with a single C@t{++} ABI; if your program contains code using
19193 multiple C@t{++} ABI's or if @value{GDBN} can not identify your
19194 program's ABI correctly, you can tell @value{GDBN} which ABI to use.
19195 Currently supported ABI's include ``gnu-v2'', for @code{g++} versions
19196 before 3.0, ``gnu-v3'', for @code{g++} versions 3.0 and later, and
19197 ``hpaCC'' for the HP ANSI C@t{++} compiler. Other C@t{++} compilers may
19198 use the ``gnu-v2'' or ``gnu-v3'' ABI's as well. The default setting is
19203 Show the C@t{++} ABI currently in use.
19206 With no argument, show the list of supported C@t{++} ABI's.
19208 @item set cp-abi @var{abi}
19209 @itemx set cp-abi auto
19210 Set the current C@t{++} ABI to @var{abi}, or return to automatic detection.
19213 @node Messages/Warnings
19214 @section Optional Warnings and Messages
19216 @cindex verbose operation
19217 @cindex optional warnings
19218 By default, @value{GDBN} is silent about its inner workings. If you are
19219 running on a slow machine, you may want to use the @code{set verbose}
19220 command. This makes @value{GDBN} tell you when it does a lengthy
19221 internal operation, so you will not think it has crashed.
19223 Currently, the messages controlled by @code{set verbose} are those
19224 which announce that the symbol table for a source file is being read;
19225 see @code{symbol-file} in @ref{Files, ,Commands to Specify Files}.
19228 @kindex set verbose
19229 @item set verbose on
19230 Enables @value{GDBN} output of certain informational messages.
19232 @item set verbose off
19233 Disables @value{GDBN} output of certain informational messages.
19235 @kindex show verbose
19237 Displays whether @code{set verbose} is on or off.
19240 By default, if @value{GDBN} encounters bugs in the symbol table of an
19241 object file, it is silent; but if you are debugging a compiler, you may
19242 find this information useful (@pxref{Symbol Errors, ,Errors Reading
19247 @kindex set complaints
19248 @item set complaints @var{limit}
19249 Permits @value{GDBN} to output @var{limit} complaints about each type of
19250 unusual symbols before becoming silent about the problem. Set
19251 @var{limit} to zero to suppress all complaints; set it to a large number
19252 to prevent complaints from being suppressed.
19254 @kindex show complaints
19255 @item show complaints
19256 Displays how many symbol complaints @value{GDBN} is permitted to produce.
19260 @anchor{confirmation requests}
19261 By default, @value{GDBN} is cautious, and asks what sometimes seems to be a
19262 lot of stupid questions to confirm certain commands. For example, if
19263 you try to run a program which is already running:
19267 The program being debugged has been started already.
19268 Start it from the beginning? (y or n)
19271 If you are willing to unflinchingly face the consequences of your own
19272 commands, you can disable this ``feature'':
19276 @kindex set confirm
19278 @cindex confirmation
19279 @cindex stupid questions
19280 @item set confirm off
19281 Disables confirmation requests. Note that running @value{GDBN} with
19282 the @option{--batch} option (@pxref{Mode Options, -batch}) also
19283 automatically disables confirmation requests.
19285 @item set confirm on
19286 Enables confirmation requests (the default).
19288 @kindex show confirm
19290 Displays state of confirmation requests.
19294 @cindex command tracing
19295 If you need to debug user-defined commands or sourced files you may find it
19296 useful to enable @dfn{command tracing}. In this mode each command will be
19297 printed as it is executed, prefixed with one or more @samp{+} symbols, the
19298 quantity denoting the call depth of each command.
19301 @kindex set trace-commands
19302 @cindex command scripts, debugging
19303 @item set trace-commands on
19304 Enable command tracing.
19305 @item set trace-commands off
19306 Disable command tracing.
19307 @item show trace-commands
19308 Display the current state of command tracing.
19311 @node Debugging Output
19312 @section Optional Messages about Internal Happenings
19313 @cindex optional debugging messages
19315 @value{GDBN} has commands that enable optional debugging messages from
19316 various @value{GDBN} subsystems; normally these commands are of
19317 interest to @value{GDBN} maintainers, or when reporting a bug. This
19318 section documents those commands.
19321 @kindex set exec-done-display
19322 @item set exec-done-display
19323 Turns on or off the notification of asynchronous commands'
19324 completion. When on, @value{GDBN} will print a message when an
19325 asynchronous command finishes its execution. The default is off.
19326 @kindex show exec-done-display
19327 @item show exec-done-display
19328 Displays the current setting of asynchronous command completion
19331 @cindex gdbarch debugging info
19332 @cindex architecture debugging info
19333 @item set debug arch
19334 Turns on or off display of gdbarch debugging info. The default is off
19336 @item show debug arch
19337 Displays the current state of displaying gdbarch debugging info.
19338 @item set debug aix-thread
19339 @cindex AIX threads
19340 Display debugging messages about inner workings of the AIX thread
19342 @item show debug aix-thread
19343 Show the current state of AIX thread debugging info display.
19344 @item set debug dwarf2-die
19345 @cindex DWARF2 DIEs
19346 Dump DWARF2 DIEs after they are read in.
19347 The value is the number of nesting levels to print.
19348 A value of zero turns off the display.
19349 @item show debug dwarf2-die
19350 Show the current state of DWARF2 DIE debugging.
19351 @item set debug displaced
19352 @cindex displaced stepping debugging info
19353 Turns on or off display of @value{GDBN} debugging info for the
19354 displaced stepping support. The default is off.
19355 @item show debug displaced
19356 Displays the current state of displaying @value{GDBN} debugging info
19357 related to displaced stepping.
19358 @item set debug event
19359 @cindex event debugging info
19360 Turns on or off display of @value{GDBN} event debugging info. The
19362 @item show debug event
19363 Displays the current state of displaying @value{GDBN} event debugging
19365 @item set debug expression
19366 @cindex expression debugging info
19367 Turns on or off display of debugging info about @value{GDBN}
19368 expression parsing. The default is off.
19369 @item show debug expression
19370 Displays the current state of displaying debugging info about
19371 @value{GDBN} expression parsing.
19372 @item set debug frame
19373 @cindex frame debugging info
19374 Turns on or off display of @value{GDBN} frame debugging info. The
19376 @item show debug frame
19377 Displays the current state of displaying @value{GDBN} frame debugging
19379 @item set debug gnu-nat
19380 @cindex @sc{gnu}/Hurd debug messages
19381 Turns on or off debugging messages from the @sc{gnu}/Hurd debug support.
19382 @item show debug gnu-nat
19383 Show the current state of @sc{gnu}/Hurd debugging messages.
19384 @item set debug infrun
19385 @cindex inferior debugging info
19386 Turns on or off display of @value{GDBN} debugging info for running the inferior.
19387 The default is off. @file{infrun.c} contains GDB's runtime state machine used
19388 for implementing operations such as single-stepping the inferior.
19389 @item show debug infrun
19390 Displays the current state of @value{GDBN} inferior debugging.
19391 @item set debug lin-lwp
19392 @cindex @sc{gnu}/Linux LWP debug messages
19393 @cindex Linux lightweight processes
19394 Turns on or off debugging messages from the Linux LWP debug support.
19395 @item show debug lin-lwp
19396 Show the current state of Linux LWP debugging messages.
19397 @item set debug lin-lwp-async
19398 @cindex @sc{gnu}/Linux LWP async debug messages
19399 @cindex Linux lightweight processes
19400 Turns on or off debugging messages from the Linux LWP async debug support.
19401 @item show debug lin-lwp-async
19402 Show the current state of Linux LWP async debugging messages.
19403 @item set debug observer
19404 @cindex observer debugging info
19405 Turns on or off display of @value{GDBN} observer debugging. This
19406 includes info such as the notification of observable events.
19407 @item show debug observer
19408 Displays the current state of observer debugging.
19409 @item set debug overload
19410 @cindex C@t{++} overload debugging info
19411 Turns on or off display of @value{GDBN} C@t{++} overload debugging
19412 info. This includes info such as ranking of functions, etc. The default
19414 @item show debug overload
19415 Displays the current state of displaying @value{GDBN} C@t{++} overload
19417 @cindex expression parser, debugging info
19418 @cindex debug expression parser
19419 @item set debug parser
19420 Turns on or off the display of expression parser debugging output.
19421 Internally, this sets the @code{yydebug} variable in the expression
19422 parser. @xref{Tracing, , Tracing Your Parser, bison, Bison}, for
19423 details. The default is off.
19424 @item show debug parser
19425 Show the current state of expression parser debugging.
19426 @cindex packets, reporting on stdout
19427 @cindex serial connections, debugging
19428 @cindex debug remote protocol
19429 @cindex remote protocol debugging
19430 @cindex display remote packets
19431 @item set debug remote
19432 Turns on or off display of reports on all packets sent back and forth across
19433 the serial line to the remote machine. The info is printed on the
19434 @value{GDBN} standard output stream. The default is off.
19435 @item show debug remote
19436 Displays the state of display of remote packets.
19437 @item set debug serial
19438 Turns on or off display of @value{GDBN} serial debugging info. The
19440 @item show debug serial
19441 Displays the current state of displaying @value{GDBN} serial debugging
19443 @item set debug solib-frv
19444 @cindex FR-V shared-library debugging
19445 Turns on or off debugging messages for FR-V shared-library code.
19446 @item show debug solib-frv
19447 Display the current state of FR-V shared-library code debugging
19449 @item set debug target
19450 @cindex target debugging info
19451 Turns on or off display of @value{GDBN} target debugging info. This info
19452 includes what is going on at the target level of GDB, as it happens. The
19453 default is 0. Set it to 1 to track events, and to 2 to also track the
19454 value of large memory transfers. Changes to this flag do not take effect
19455 until the next time you connect to a target or use the @code{run} command.
19456 @item show debug target
19457 Displays the current state of displaying @value{GDBN} target debugging
19459 @item set debug timestamp
19460 @cindex timestampping debugging info
19461 Turns on or off display of timestamps with @value{GDBN} debugging info.
19462 When enabled, seconds and microseconds are displayed before each debugging
19464 @item show debug timestamp
19465 Displays the current state of displaying timestamps with @value{GDBN}
19467 @item set debugvarobj
19468 @cindex variable object debugging info
19469 Turns on or off display of @value{GDBN} variable object debugging
19470 info. The default is off.
19471 @item show debugvarobj
19472 Displays the current state of displaying @value{GDBN} variable object
19474 @item set debug xml
19475 @cindex XML parser debugging
19476 Turns on or off debugging messages for built-in XML parsers.
19477 @item show debug xml
19478 Displays the current state of XML debugging messages.
19481 @node Other Misc Settings
19482 @section Other Miscellaneous Settings
19483 @cindex miscellaneous settings
19486 @kindex set interactive-mode
19487 @item set interactive-mode
19488 If @code{on}, forces @value{GDBN} to operate interactively.
19489 If @code{off}, forces @value{GDBN} to operate non-interactively,
19490 If @code{auto} (the default), @value{GDBN} guesses which mode to use,
19491 based on whether the debugger was started in a terminal or not.
19493 In the vast majority of cases, the debugger should be able to guess
19494 correctly which mode should be used. But this setting can be useful
19495 in certain specific cases, such as running a MinGW @value{GDBN}
19496 inside a cygwin window.
19498 @kindex show interactive-mode
19499 @item show interactive-mode
19500 Displays whether the debugger is operating in interactive mode or not.
19503 @node Extending GDB
19504 @chapter Extending @value{GDBN}
19505 @cindex extending GDB
19507 @value{GDBN} provides two mechanisms for extension. The first is based
19508 on composition of @value{GDBN} commands, and the second is based on the
19509 Python scripting language.
19511 To facilitate the use of these extensions, @value{GDBN} is capable
19512 of evaluating the contents of a file. When doing so, @value{GDBN}
19513 can recognize which scripting language is being used by looking at
19514 the filename extension. Files with an unrecognized filename extension
19515 are always treated as a @value{GDBN} Command Files.
19516 @xref{Command Files,, Command files}.
19518 You can control how @value{GDBN} evaluates these files with the following
19522 @kindex set script-extension
19523 @kindex show script-extension
19524 @item set script-extension off
19525 All scripts are always evaluated as @value{GDBN} Command Files.
19527 @item set script-extension soft
19528 The debugger determines the scripting language based on filename
19529 extension. If this scripting language is supported, @value{GDBN}
19530 evaluates the script using that language. Otherwise, it evaluates
19531 the file as a @value{GDBN} Command File.
19533 @item set script-extension strict
19534 The debugger determines the scripting language based on filename
19535 extension, and evaluates the script using that language. If the
19536 language is not supported, then the evaluation fails.
19538 @item show script-extension
19539 Display the current value of the @code{script-extension} option.
19544 * Sequences:: Canned Sequences of Commands
19545 * Python:: Scripting @value{GDBN} using Python
19549 @section Canned Sequences of Commands
19551 Aside from breakpoint commands (@pxref{Break Commands, ,Breakpoint
19552 Command Lists}), @value{GDBN} provides two ways to store sequences of
19553 commands for execution as a unit: user-defined commands and command
19557 * Define:: How to define your own commands
19558 * Hooks:: Hooks for user-defined commands
19559 * Command Files:: How to write scripts of commands to be stored in a file
19560 * Output:: Commands for controlled output
19564 @subsection User-defined Commands
19566 @cindex user-defined command
19567 @cindex arguments, to user-defined commands
19568 A @dfn{user-defined command} is a sequence of @value{GDBN} commands to
19569 which you assign a new name as a command. This is done with the
19570 @code{define} command. User commands may accept up to 10 arguments
19571 separated by whitespace. Arguments are accessed within the user command
19572 via @code{$arg0@dots{}$arg9}. A trivial example:
19576 print $arg0 + $arg1 + $arg2
19581 To execute the command use:
19588 This defines the command @code{adder}, which prints the sum of
19589 its three arguments. Note the arguments are text substitutions, so they may
19590 reference variables, use complex expressions, or even perform inferior
19593 @cindex argument count in user-defined commands
19594 @cindex how many arguments (user-defined commands)
19595 In addition, @code{$argc} may be used to find out how many arguments have
19596 been passed. This expands to a number in the range 0@dots{}10.
19601 print $arg0 + $arg1
19604 print $arg0 + $arg1 + $arg2
19612 @item define @var{commandname}
19613 Define a command named @var{commandname}. If there is already a command
19614 by that name, you are asked to confirm that you want to redefine it.
19615 @var{commandname} may be a bare command name consisting of letters,
19616 numbers, dashes, and underscores. It may also start with any predefined
19617 prefix command. For example, @samp{define target my-target} creates
19618 a user-defined @samp{target my-target} command.
19620 The definition of the command is made up of other @value{GDBN} command lines,
19621 which are given following the @code{define} command. The end of these
19622 commands is marked by a line containing @code{end}.
19625 @kindex end@r{ (user-defined commands)}
19626 @item document @var{commandname}
19627 Document the user-defined command @var{commandname}, so that it can be
19628 accessed by @code{help}. The command @var{commandname} must already be
19629 defined. This command reads lines of documentation just as @code{define}
19630 reads the lines of the command definition, ending with @code{end}.
19631 After the @code{document} command is finished, @code{help} on command
19632 @var{commandname} displays the documentation you have written.
19634 You may use the @code{document} command again to change the
19635 documentation of a command. Redefining the command with @code{define}
19636 does not change the documentation.
19638 @kindex dont-repeat
19639 @cindex don't repeat command
19641 Used inside a user-defined command, this tells @value{GDBN} that this
19642 command should not be repeated when the user hits @key{RET}
19643 (@pxref{Command Syntax, repeat last command}).
19645 @kindex help user-defined
19646 @item help user-defined
19647 List all user-defined commands, with the first line of the documentation
19652 @itemx show user @var{commandname}
19653 Display the @value{GDBN} commands used to define @var{commandname} (but
19654 not its documentation). If no @var{commandname} is given, display the
19655 definitions for all user-defined commands.
19657 @cindex infinite recursion in user-defined commands
19658 @kindex show max-user-call-depth
19659 @kindex set max-user-call-depth
19660 @item show max-user-call-depth
19661 @itemx set max-user-call-depth
19662 The value of @code{max-user-call-depth} controls how many recursion
19663 levels are allowed in user-defined commands before @value{GDBN} suspects an
19664 infinite recursion and aborts the command.
19667 In addition to the above commands, user-defined commands frequently
19668 use control flow commands, described in @ref{Command Files}.
19670 When user-defined commands are executed, the
19671 commands of the definition are not printed. An error in any command
19672 stops execution of the user-defined command.
19674 If used interactively, commands that would ask for confirmation proceed
19675 without asking when used inside a user-defined command. Many @value{GDBN}
19676 commands that normally print messages to say what they are doing omit the
19677 messages when used in a user-defined command.
19680 @subsection User-defined Command Hooks
19681 @cindex command hooks
19682 @cindex hooks, for commands
19683 @cindex hooks, pre-command
19686 You may define @dfn{hooks}, which are a special kind of user-defined
19687 command. Whenever you run the command @samp{foo}, if the user-defined
19688 command @samp{hook-foo} exists, it is executed (with no arguments)
19689 before that command.
19691 @cindex hooks, post-command
19693 A hook may also be defined which is run after the command you executed.
19694 Whenever you run the command @samp{foo}, if the user-defined command
19695 @samp{hookpost-foo} exists, it is executed (with no arguments) after
19696 that command. Post-execution hooks may exist simultaneously with
19697 pre-execution hooks, for the same command.
19699 It is valid for a hook to call the command which it hooks. If this
19700 occurs, the hook is not re-executed, thereby avoiding infinite recursion.
19702 @c It would be nice if hookpost could be passed a parameter indicating
19703 @c if the command it hooks executed properly or not. FIXME!
19705 @kindex stop@r{, a pseudo-command}
19706 In addition, a pseudo-command, @samp{stop} exists. Defining
19707 (@samp{hook-stop}) makes the associated commands execute every time
19708 execution stops in your program: before breakpoint commands are run,
19709 displays are printed, or the stack frame is printed.
19711 For example, to ignore @code{SIGALRM} signals while
19712 single-stepping, but treat them normally during normal execution,
19717 handle SIGALRM nopass
19721 handle SIGALRM pass
19724 define hook-continue
19725 handle SIGALRM pass
19729 As a further example, to hook at the beginning and end of the @code{echo}
19730 command, and to add extra text to the beginning and end of the message,
19738 define hookpost-echo
19742 (@value{GDBP}) echo Hello World
19743 <<<---Hello World--->>>
19748 You can define a hook for any single-word command in @value{GDBN}, but
19749 not for command aliases; you should define a hook for the basic command
19750 name, e.g.@: @code{backtrace} rather than @code{bt}.
19751 @c FIXME! So how does Joe User discover whether a command is an alias
19753 You can hook a multi-word command by adding @code{hook-} or
19754 @code{hookpost-} to the last word of the command, e.g.@:
19755 @samp{define target hook-remote} to add a hook to @samp{target remote}.
19757 If an error occurs during the execution of your hook, execution of
19758 @value{GDBN} commands stops and @value{GDBN} issues a prompt
19759 (before the command that you actually typed had a chance to run).
19761 If you try to define a hook which does not match any known command, you
19762 get a warning from the @code{define} command.
19764 @node Command Files
19765 @subsection Command Files
19767 @cindex command files
19768 @cindex scripting commands
19769 A command file for @value{GDBN} is a text file made of lines that are
19770 @value{GDBN} commands. Comments (lines starting with @kbd{#}) may
19771 also be included. An empty line in a command file does nothing; it
19772 does not mean to repeat the last command, as it would from the
19775 You can request the execution of a command file with the @code{source}
19776 command. Note that the @code{source} command is also used to evaluate
19777 scripts that are not Command Files. The exact behavior can be configured
19778 using the @code{script-extension} setting.
19779 @xref{Extending GDB,, Extending GDB}.
19783 @cindex execute commands from a file
19784 @item source [-s] [-v] @var{filename}
19785 Execute the command file @var{filename}.
19788 The lines in a command file are generally executed sequentially,
19789 unless the order of execution is changed by one of the
19790 @emph{flow-control commands} described below. The commands are not
19791 printed as they are executed. An error in any command terminates
19792 execution of the command file and control is returned to the console.
19794 @value{GDBN} first searches for @var{filename} in the current directory.
19795 If the file is not found there, and @var{filename} does not specify a
19796 directory, then @value{GDBN} also looks for the file on the source search path
19797 (specified with the @samp{directory} command);
19798 except that @file{$cdir} is not searched because the compilation directory
19799 is not relevant to scripts.
19801 If @code{-s} is specified, then @value{GDBN} searches for @var{filename}
19802 on the search path even if @var{filename} specifies a directory.
19803 The search is done by appending @var{filename} to each element of the
19804 search path. So, for example, if @var{filename} is @file{mylib/myscript}
19805 and the search path contains @file{/home/user} then @value{GDBN} will
19806 look for the script @file{/home/user/mylib/myscript}.
19807 The search is also done if @var{filename} is an absolute path.
19808 For example, if @var{filename} is @file{/tmp/myscript} and
19809 the search path contains @file{/home/user} then @value{GDBN} will
19810 look for the script @file{/home/user/tmp/myscript}.
19811 For DOS-like systems, if @var{filename} contains a drive specification,
19812 it is stripped before concatenation. For example, if @var{filename} is
19813 @file{d:myscript} and the search path contains @file{c:/tmp} then @value{GDBN}
19814 will look for the script @file{c:/tmp/myscript}.
19816 If @code{-v}, for verbose mode, is given then @value{GDBN} displays
19817 each command as it is executed. The option must be given before
19818 @var{filename}, and is interpreted as part of the filename anywhere else.
19820 Commands that would ask for confirmation if used interactively proceed
19821 without asking when used in a command file. Many @value{GDBN} commands that
19822 normally print messages to say what they are doing omit the messages
19823 when called from command files.
19825 @value{GDBN} also accepts command input from standard input. In this
19826 mode, normal output goes to standard output and error output goes to
19827 standard error. Errors in a command file supplied on standard input do
19828 not terminate execution of the command file---execution continues with
19832 gdb < cmds > log 2>&1
19835 (The syntax above will vary depending on the shell used.) This example
19836 will execute commands from the file @file{cmds}. All output and errors
19837 would be directed to @file{log}.
19839 Since commands stored on command files tend to be more general than
19840 commands typed interactively, they frequently need to deal with
19841 complicated situations, such as different or unexpected values of
19842 variables and symbols, changes in how the program being debugged is
19843 built, etc. @value{GDBN} provides a set of flow-control commands to
19844 deal with these complexities. Using these commands, you can write
19845 complex scripts that loop over data structures, execute commands
19846 conditionally, etc.
19853 This command allows to include in your script conditionally executed
19854 commands. The @code{if} command takes a single argument, which is an
19855 expression to evaluate. It is followed by a series of commands that
19856 are executed only if the expression is true (its value is nonzero).
19857 There can then optionally be an @code{else} line, followed by a series
19858 of commands that are only executed if the expression was false. The
19859 end of the list is marked by a line containing @code{end}.
19863 This command allows to write loops. Its syntax is similar to
19864 @code{if}: the command takes a single argument, which is an expression
19865 to evaluate, and must be followed by the commands to execute, one per
19866 line, terminated by an @code{end}. These commands are called the
19867 @dfn{body} of the loop. The commands in the body of @code{while} are
19868 executed repeatedly as long as the expression evaluates to true.
19872 This command exits the @code{while} loop in whose body it is included.
19873 Execution of the script continues after that @code{while}s @code{end}
19876 @kindex loop_continue
19877 @item loop_continue
19878 This command skips the execution of the rest of the body of commands
19879 in the @code{while} loop in whose body it is included. Execution
19880 branches to the beginning of the @code{while} loop, where it evaluates
19881 the controlling expression.
19883 @kindex end@r{ (if/else/while commands)}
19885 Terminate the block of commands that are the body of @code{if},
19886 @code{else}, or @code{while} flow-control commands.
19891 @subsection Commands for Controlled Output
19893 During the execution of a command file or a user-defined command, normal
19894 @value{GDBN} output is suppressed; the only output that appears is what is
19895 explicitly printed by the commands in the definition. This section
19896 describes three commands useful for generating exactly the output you
19901 @item echo @var{text}
19902 @c I do not consider backslash-space a standard C escape sequence
19903 @c because it is not in ANSI.
19904 Print @var{text}. Nonprinting characters can be included in
19905 @var{text} using C escape sequences, such as @samp{\n} to print a
19906 newline. @strong{No newline is printed unless you specify one.}
19907 In addition to the standard C escape sequences, a backslash followed
19908 by a space stands for a space. This is useful for displaying a
19909 string with spaces at the beginning or the end, since leading and
19910 trailing spaces are otherwise trimmed from all arguments.
19911 To print @samp{@w{ }and foo =@w{ }}, use the command
19912 @samp{echo \@w{ }and foo = \@w{ }}.
19914 A backslash at the end of @var{text} can be used, as in C, to continue
19915 the command onto subsequent lines. For example,
19918 echo This is some text\n\
19919 which is continued\n\
19920 onto several lines.\n
19923 produces the same output as
19926 echo This is some text\n
19927 echo which is continued\n
19928 echo onto several lines.\n
19932 @item output @var{expression}
19933 Print the value of @var{expression} and nothing but that value: no
19934 newlines, no @samp{$@var{nn} = }. The value is not entered in the
19935 value history either. @xref{Expressions, ,Expressions}, for more information
19938 @item output/@var{fmt} @var{expression}
19939 Print the value of @var{expression} in format @var{fmt}. You can use
19940 the same formats as for @code{print}. @xref{Output Formats,,Output
19941 Formats}, for more information.
19944 @item printf @var{template}, @var{expressions}@dots{}
19945 Print the values of one or more @var{expressions} under the control of
19946 the string @var{template}. To print several values, make
19947 @var{expressions} be a comma-separated list of individual expressions,
19948 which may be either numbers or pointers. Their values are printed as
19949 specified by @var{template}, exactly as a C program would do by
19950 executing the code below:
19953 printf (@var{template}, @var{expressions}@dots{});
19956 As in @code{C} @code{printf}, ordinary characters in @var{template}
19957 are printed verbatim, while @dfn{conversion specification} introduced
19958 by the @samp{%} character cause subsequent @var{expressions} to be
19959 evaluated, their values converted and formatted according to type and
19960 style information encoded in the conversion specifications, and then
19963 For example, you can print two values in hex like this:
19966 printf "foo, bar-foo = 0x%x, 0x%x\n", foo, bar-foo
19969 @code{printf} supports all the standard @code{C} conversion
19970 specifications, including the flags and modifiers between the @samp{%}
19971 character and the conversion letter, with the following exceptions:
19975 The argument-ordering modifiers, such as @samp{2$}, are not supported.
19978 The modifier @samp{*} is not supported for specifying precision or
19982 The @samp{'} flag (for separation of digits into groups according to
19983 @code{LC_NUMERIC'}) is not supported.
19986 The type modifiers @samp{hh}, @samp{j}, @samp{t}, and @samp{z} are not
19990 The conversion letter @samp{n} (as in @samp{%n}) is not supported.
19993 The conversion letters @samp{a} and @samp{A} are not supported.
19997 Note that the @samp{ll} type modifier is supported only if the
19998 underlying @code{C} implementation used to build @value{GDBN} supports
19999 the @code{long long int} type, and the @samp{L} type modifier is
20000 supported only if @code{long double} type is available.
20002 As in @code{C}, @code{printf} supports simple backslash-escape
20003 sequences, such as @code{\n}, @samp{\t}, @samp{\\}, @samp{\"},
20004 @samp{\a}, and @samp{\f}, that consist of backslash followed by a
20005 single character. Octal and hexadecimal escape sequences are not
20008 Additionally, @code{printf} supports conversion specifications for DFP
20009 (@dfn{Decimal Floating Point}) types using the following length modifiers
20010 together with a floating point specifier.
20015 @samp{H} for printing @code{Decimal32} types.
20018 @samp{D} for printing @code{Decimal64} types.
20021 @samp{DD} for printing @code{Decimal128} types.
20024 If the underlying @code{C} implementation used to build @value{GDBN} has
20025 support for the three length modifiers for DFP types, other modifiers
20026 such as width and precision will also be available for @value{GDBN} to use.
20028 In case there is no such @code{C} support, no additional modifiers will be
20029 available and the value will be printed in the standard way.
20031 Here's an example of printing DFP types using the above conversion letters:
20033 printf "D32: %Hf - D64: %Df - D128: %DDf\n",1.2345df,1.2E10dd,1.2E1dl
20039 @section Scripting @value{GDBN} using Python
20040 @cindex python scripting
20041 @cindex scripting with python
20043 You can script @value{GDBN} using the @uref{http://www.python.org/,
20044 Python programming language}. This feature is available only if
20045 @value{GDBN} was configured using @option{--with-python}.
20048 * Python Commands:: Accessing Python from @value{GDBN}.
20049 * Python API:: Accessing @value{GDBN} from Python.
20050 * Auto-loading:: Automatically loading Python code.
20053 @node Python Commands
20054 @subsection Python Commands
20055 @cindex python commands
20056 @cindex commands to access python
20058 @value{GDBN} provides one command for accessing the Python interpreter,
20059 and one related setting:
20063 @item python @r{[}@var{code}@r{]}
20064 The @code{python} command can be used to evaluate Python code.
20066 If given an argument, the @code{python} command will evaluate the
20067 argument as a Python command. For example:
20070 (@value{GDBP}) python print 23
20074 If you do not provide an argument to @code{python}, it will act as a
20075 multi-line command, like @code{define}. In this case, the Python
20076 script is made up of subsequent command lines, given after the
20077 @code{python} command. This command list is terminated using a line
20078 containing @code{end}. For example:
20081 (@value{GDBP}) python
20083 End with a line saying just "end".
20089 @kindex maint set python print-stack
20090 @item maint set python print-stack
20091 By default, @value{GDBN} will print a stack trace when an error occurs
20092 in a Python script. This can be controlled using @code{maint set
20093 python print-stack}: if @code{on}, the default, then Python stack
20094 printing is enabled; if @code{off}, then Python stack printing is
20098 It is also possible to execute a Python script from the @value{GDBN}
20102 @item source @file{script-name}
20103 The script name must end with @samp{.py} and @value{GDBN} must be configured
20104 to recognize the script language based on filename extension using
20105 the @code{script-extension} setting. @xref{Extending GDB, ,Extending GDB}.
20107 @item python execfile ("script-name")
20108 This method is based on the @code{execfile} Python built-in function,
20109 and thus is always available.
20113 @subsection Python API
20115 @cindex programming in python
20117 @cindex python stdout
20118 @cindex python pagination
20119 At startup, @value{GDBN} overrides Python's @code{sys.stdout} and
20120 @code{sys.stderr} to print using @value{GDBN}'s output-paging streams.
20121 A Python program which outputs to one of these streams may have its
20122 output interrupted by the user (@pxref{Screen Size}). In this
20123 situation, a Python @code{KeyboardInterrupt} exception is thrown.
20126 * Basic Python:: Basic Python Functions.
20127 * Exception Handling::
20128 * Values From Inferior::
20129 * Types In Python:: Python representation of types.
20130 * Pretty Printing API:: Pretty-printing values.
20131 * Selecting Pretty-Printers:: How GDB chooses a pretty-printer.
20132 * Disabling Pretty-Printers:: Disabling broken printers.
20133 * Commands In Python:: Implementing new commands in Python.
20134 * Parameters In Python:: Adding new @value{GDBN} parameters.
20135 * Functions In Python:: Writing new convenience functions.
20136 * Progspaces In Python:: Program spaces.
20137 * Objfiles In Python:: Object files.
20138 * Frames In Python:: Accessing inferior stack frames from Python.
20139 * Blocks In Python:: Accessing frame blocks from Python.
20140 * Symbols In Python:: Python representation of symbols.
20141 * Symbol Tables In Python:: Python representation of symbol tables.
20142 * Lazy Strings In Python:: Python representation of lazy strings.
20143 * Breakpoints In Python:: Manipulating breakpoints using Python.
20147 @subsubsection Basic Python
20149 @cindex python functions
20150 @cindex python module
20152 @value{GDBN} introduces a new Python module, named @code{gdb}. All
20153 methods and classes added by @value{GDBN} are placed in this module.
20154 @value{GDBN} automatically @code{import}s the @code{gdb} module for
20155 use in all scripts evaluated by the @code{python} command.
20157 @findex gdb.execute
20158 @defun execute command [from_tty]
20159 Evaluate @var{command}, a string, as a @value{GDBN} CLI command.
20160 If a GDB exception happens while @var{command} runs, it is
20161 translated as described in @ref{Exception Handling,,Exception Handling}.
20162 If no exceptions occur, this function returns @code{None}.
20164 @var{from_tty} specifies whether @value{GDBN} ought to consider this
20165 command as having originated from the user invoking it interactively.
20166 It must be a boolean value. If omitted, it defaults to @code{False}.
20169 @findex gdb.breakpoints
20171 Return a sequence holding all of @value{GDBN}'s breakpoints.
20172 @xref{Breakpoints In Python}, for more information.
20175 @findex gdb.parameter
20176 @defun parameter parameter
20177 Return the value of a @value{GDBN} parameter. @var{parameter} is a
20178 string naming the parameter to look up; @var{parameter} may contain
20179 spaces if the parameter has a multi-part name. For example,
20180 @samp{print object} is a valid parameter name.
20182 If the named parameter does not exist, this function throws a
20183 @code{RuntimeError}. Otherwise, the parameter's value is converted to
20184 a Python value of the appropriate type, and returned.
20187 @findex gdb.history
20188 @defun history number
20189 Return a value from @value{GDBN}'s value history (@pxref{Value
20190 History}). @var{number} indicates which history element to return.
20191 If @var{number} is negative, then @value{GDBN} will take its absolute value
20192 and count backward from the last element (i.e., the most recent element) to
20193 find the value to return. If @var{number} is zero, then @value{GDBN} will
20194 return the most recent element. If the element specified by @var{number}
20195 doesn't exist in the value history, a @code{RuntimeError} exception will be
20198 If no exception is raised, the return value is always an instance of
20199 @code{gdb.Value} (@pxref{Values From Inferior}).
20202 @findex gdb.parse_and_eval
20203 @defun parse_and_eval expression
20204 Parse @var{expression} as an expression in the current language,
20205 evaluate it, and return the result as a @code{gdb.Value}.
20206 @var{expression} must be a string.
20208 This function can be useful when implementing a new command
20209 (@pxref{Commands In Python}), as it provides a way to parse the
20210 command's argument as an expression. It is also useful simply to
20211 compute values, for example, it is the only way to get the value of a
20212 convenience variable (@pxref{Convenience Vars}) as a @code{gdb.Value}.
20216 @defun write string
20217 Print a string to @value{GDBN}'s paginated standard output stream.
20218 Writing to @code{sys.stdout} or @code{sys.stderr} will automatically
20219 call this function.
20224 Flush @value{GDBN}'s paginated standard output stream. Flushing
20225 @code{sys.stdout} or @code{sys.stderr} will automatically call this
20229 @findex gdb.target_charset
20230 @defun target_charset
20231 Return the name of the current target character set (@pxref{Character
20232 Sets}). This differs from @code{gdb.parameter('target-charset')} in
20233 that @samp{auto} is never returned.
20236 @findex gdb.target_wide_charset
20237 @defun target_wide_charset
20238 Return the name of the current target wide character set
20239 (@pxref{Character Sets}). This differs from
20240 @code{gdb.parameter('target-wide-charset')} in that @samp{auto} is
20244 @node Exception Handling
20245 @subsubsection Exception Handling
20246 @cindex python exceptions
20247 @cindex exceptions, python
20249 When executing the @code{python} command, Python exceptions
20250 uncaught within the Python code are translated to calls to
20251 @value{GDBN} error-reporting mechanism. If the command that called
20252 @code{python} does not handle the error, @value{GDBN} will
20253 terminate it and print an error message containing the Python
20254 exception name, the associated value, and the Python call stack
20255 backtrace at the point where the exception was raised. Example:
20258 (@value{GDBP}) python print foo
20259 Traceback (most recent call last):
20260 File "<string>", line 1, in <module>
20261 NameError: name 'foo' is not defined
20264 @value{GDBN} errors that happen in @value{GDBN} commands invoked by Python
20265 code are converted to Python @code{RuntimeError} exceptions. User
20266 interrupt (via @kbd{C-c} or by typing @kbd{q} at a pagination
20267 prompt) is translated to a Python @code{KeyboardInterrupt}
20268 exception. If you catch these exceptions in your Python code, your
20269 exception handler will see @code{RuntimeError} or
20270 @code{KeyboardInterrupt} as the exception type, the @value{GDBN} error
20271 message as its value, and the Python call stack backtrace at the
20272 Python statement closest to where the @value{GDBN} error occured as the
20275 @findex gdb.GdbError
20276 When implementing @value{GDBN} commands in Python via @code{gdb.Command},
20277 it is useful to be able to throw an exception that doesn't cause a
20278 traceback to be printed. For example, the user may have invoked the
20279 command incorrectly. Use the @code{gdb.GdbError} exception
20280 to handle this case. Example:
20284 >class HelloWorld (gdb.Command):
20285 > """Greet the whole world."""
20286 > def __init__ (self):
20287 > super (HelloWorld, self).__init__ ("hello-world", gdb.COMMAND_OBSCURE)
20288 > def invoke (self, args, from_tty):
20289 > argv = gdb.string_to_argv (args)
20290 > if len (argv) != 0:
20291 > raise gdb.GdbError ("hello-world takes no arguments")
20292 > print "Hello, World!"
20295 (gdb) hello-world 42
20296 hello-world takes no arguments
20299 @node Values From Inferior
20300 @subsubsection Values From Inferior
20301 @cindex values from inferior, with Python
20302 @cindex python, working with values from inferior
20304 @cindex @code{gdb.Value}
20305 @value{GDBN} provides values it obtains from the inferior program in
20306 an object of type @code{gdb.Value}. @value{GDBN} uses this object
20307 for its internal bookkeeping of the inferior's values, and for
20308 fetching values when necessary.
20310 Inferior values that are simple scalars can be used directly in
20311 Python expressions that are valid for the value's data type. Here's
20312 an example for an integer or floating-point value @code{some_val}:
20319 As result of this, @code{bar} will also be a @code{gdb.Value} object
20320 whose values are of the same type as those of @code{some_val}.
20322 Inferior values that are structures or instances of some class can
20323 be accessed using the Python @dfn{dictionary syntax}. For example, if
20324 @code{some_val} is a @code{gdb.Value} instance holding a structure, you
20325 can access its @code{foo} element with:
20328 bar = some_val['foo']
20331 Again, @code{bar} will also be a @code{gdb.Value} object.
20333 The following attributes are provided:
20336 @defivar Value address
20337 If this object is addressable, this read-only attribute holds a
20338 @code{gdb.Value} object representing the address. Otherwise,
20339 this attribute holds @code{None}.
20342 @cindex optimized out value in Python
20343 @defivar Value is_optimized_out
20344 This read-only boolean attribute is true if the compiler optimized out
20345 this value, thus it is not available for fetching from the inferior.
20348 @defivar Value type
20349 The type of this @code{gdb.Value}. The value of this attribute is a
20350 @code{gdb.Type} object.
20354 The following methods are provided:
20357 @defmethod Value cast type
20358 Return a new instance of @code{gdb.Value} that is the result of
20359 casting this instance to the type described by @var{type}, which must
20360 be a @code{gdb.Type} object. If the cast cannot be performed for some
20361 reason, this method throws an exception.
20364 @defmethod Value dereference
20365 For pointer data types, this method returns a new @code{gdb.Value} object
20366 whose contents is the object pointed to by the pointer. For example, if
20367 @code{foo} is a C pointer to an @code{int}, declared in your C program as
20374 then you can use the corresponding @code{gdb.Value} to access what
20375 @code{foo} points to like this:
20378 bar = foo.dereference ()
20381 The result @code{bar} will be a @code{gdb.Value} object holding the
20382 value pointed to by @code{foo}.
20385 @defmethod Value string @r{[}encoding@r{]} @r{[}errors@r{]} @r{[}length@r{]}
20386 If this @code{gdb.Value} represents a string, then this method
20387 converts the contents to a Python string. Otherwise, this method will
20388 throw an exception.
20390 Strings are recognized in a language-specific way; whether a given
20391 @code{gdb.Value} represents a string is determined by the current
20394 For C-like languages, a value is a string if it is a pointer to or an
20395 array of characters or ints. The string is assumed to be terminated
20396 by a zero of the appropriate width. However if the optional length
20397 argument is given, the string will be converted to that given length,
20398 ignoring any embedded zeros that the string may contain.
20400 If the optional @var{encoding} argument is given, it must be a string
20401 naming the encoding of the string in the @code{gdb.Value}, such as
20402 @code{"ascii"}, @code{"iso-8859-6"} or @code{"utf-8"}. It accepts
20403 the same encodings as the corresponding argument to Python's
20404 @code{string.decode} method, and the Python codec machinery will be used
20405 to convert the string. If @var{encoding} is not given, or if
20406 @var{encoding} is the empty string, then either the @code{target-charset}
20407 (@pxref{Character Sets}) will be used, or a language-specific encoding
20408 will be used, if the current language is able to supply one.
20410 The optional @var{errors} argument is the same as the corresponding
20411 argument to Python's @code{string.decode} method.
20413 If the optional @var{length} argument is given, the string will be
20414 fetched and converted to the given length.
20417 @defmethod Value lazy_string @r{[}encoding@r{]} @r{[}length@r{]}
20418 If this @code{gdb.Value} represents a string, then this method
20419 converts the contents to a @code{gdb.LazyString} (@pxref{Lazy Strings
20420 In Python}). Otherwise, this method will throw an exception.
20422 If the optional @var{encoding} argument is given, it must be a string
20423 naming the encoding of the @code{gdb.LazyString}. Some examples are:
20424 @samp{ascii}, @samp{iso-8859-6} or @samp{utf-8}. If the
20425 @var{encoding} argument is an encoding that @value{GDBN} does
20426 recognize, @value{GDBN} will raise an error.
20428 When a lazy string is printed, the @value{GDBN} encoding machinery is
20429 used to convert the string during printing. If the optional
20430 @var{encoding} argument is not provided, or is an empty string,
20431 @value{GDBN} will automatically select the encoding most suitable for
20432 the string type. For further information on encoding in @value{GDBN}
20433 please see @ref{Character Sets}.
20435 If the optional @var{length} argument is given, the string will be
20436 fetched and encoded to the length of characters specified. If
20437 the @var{length} argument is not provided, the string will be fetched
20438 and encoded until a null of appropriate width is found.
20442 @node Types In Python
20443 @subsubsection Types In Python
20444 @cindex types in Python
20445 @cindex Python, working with types
20448 @value{GDBN} represents types from the inferior using the class
20451 The following type-related functions are available in the @code{gdb}
20454 @findex gdb.lookup_type
20455 @defun lookup_type name [block]
20456 This function looks up a type by name. @var{name} is the name of the
20457 type to look up. It must be a string.
20459 If @var{block} is given, then @var{name} is looked up in that scope.
20460 Otherwise, it is searched for globally.
20462 Ordinarily, this function will return an instance of @code{gdb.Type}.
20463 If the named type cannot be found, it will throw an exception.
20466 An instance of @code{Type} has the following attributes:
20470 The type code for this type. The type code will be one of the
20471 @code{TYPE_CODE_} constants defined below.
20474 @defivar Type sizeof
20475 The size of this type, in target @code{char} units. Usually, a
20476 target's @code{char} type will be an 8-bit byte. However, on some
20477 unusual platforms, this type may have a different size.
20481 The tag name for this type. The tag name is the name after
20482 @code{struct}, @code{union}, or @code{enum} in C and C@t{++}; not all
20483 languages have this concept. If this type has no tag name, then
20484 @code{None} is returned.
20488 The following methods are provided:
20491 @defmethod Type fields
20492 For structure and union types, this method returns the fields. Range
20493 types have two fields, the minimum and maximum values. Enum types
20494 have one field per enum constant. Function and method types have one
20495 field per parameter. The base types of C@t{++} classes are also
20496 represented as fields. If the type has no fields, or does not fit
20497 into one of these categories, an empty sequence will be returned.
20499 Each field is an object, with some pre-defined attributes:
20502 This attribute is not available for @code{static} fields (as in
20503 C@t{++} or Java). For non-@code{static} fields, the value is the bit
20504 position of the field.
20507 The name of the field, or @code{None} for anonymous fields.
20510 This is @code{True} if the field is artificial, usually meaning that
20511 it was provided by the compiler and not the user. This attribute is
20512 always provided, and is @code{False} if the field is not artificial.
20514 @item is_base_class
20515 This is @code{True} if the field represents a base class of a C@t{++}
20516 structure. This attribute is always provided, and is @code{False}
20517 if the field is not a base class of the type that is the argument of
20518 @code{fields}, or if that type was not a C@t{++} class.
20521 If the field is packed, or is a bitfield, then this will have a
20522 non-zero value, which is the size of the field in bits. Otherwise,
20523 this will be zero; in this case the field's size is given by its type.
20526 The type of the field. This is usually an instance of @code{Type},
20527 but it can be @code{None} in some situations.
20531 @defmethod Type const
20532 Return a new @code{gdb.Type} object which represents a
20533 @code{const}-qualified variant of this type.
20536 @defmethod Type volatile
20537 Return a new @code{gdb.Type} object which represents a
20538 @code{volatile}-qualified variant of this type.
20541 @defmethod Type unqualified
20542 Return a new @code{gdb.Type} object which represents an unqualified
20543 variant of this type. That is, the result is neither @code{const} nor
20547 @defmethod Type range
20548 Return a Python @code{Tuple} object that contains two elements: the
20549 low bound of the argument type and the high bound of that type. If
20550 the type does not have a range, @value{GDBN} will raise a
20551 @code{RuntimeError} exception.
20554 @defmethod Type reference
20555 Return a new @code{gdb.Type} object which represents a reference to this
20559 @defmethod Type pointer
20560 Return a new @code{gdb.Type} object which represents a pointer to this
20564 @defmethod Type strip_typedefs
20565 Return a new @code{gdb.Type} that represents the real type,
20566 after removing all layers of typedefs.
20569 @defmethod Type target
20570 Return a new @code{gdb.Type} object which represents the target type
20573 For a pointer type, the target type is the type of the pointed-to
20574 object. For an array type (meaning C-like arrays), the target type is
20575 the type of the elements of the array. For a function or method type,
20576 the target type is the type of the return value. For a complex type,
20577 the target type is the type of the elements. For a typedef, the
20578 target type is the aliased type.
20580 If the type does not have a target, this method will throw an
20584 @defmethod Type template_argument n [block]
20585 If this @code{gdb.Type} is an instantiation of a template, this will
20586 return a new @code{gdb.Type} which represents the type of the
20587 @var{n}th template argument.
20589 If this @code{gdb.Type} is not a template type, this will throw an
20590 exception. Ordinarily, only C@t{++} code will have template types.
20592 If @var{block} is given, then @var{name} is looked up in that scope.
20593 Otherwise, it is searched for globally.
20598 Each type has a code, which indicates what category this type falls
20599 into. The available type categories are represented by constants
20600 defined in the @code{gdb} module:
20603 @findex TYPE_CODE_PTR
20604 @findex gdb.TYPE_CODE_PTR
20605 @item TYPE_CODE_PTR
20606 The type is a pointer.
20608 @findex TYPE_CODE_ARRAY
20609 @findex gdb.TYPE_CODE_ARRAY
20610 @item TYPE_CODE_ARRAY
20611 The type is an array.
20613 @findex TYPE_CODE_STRUCT
20614 @findex gdb.TYPE_CODE_STRUCT
20615 @item TYPE_CODE_STRUCT
20616 The type is a structure.
20618 @findex TYPE_CODE_UNION
20619 @findex gdb.TYPE_CODE_UNION
20620 @item TYPE_CODE_UNION
20621 The type is a union.
20623 @findex TYPE_CODE_ENUM
20624 @findex gdb.TYPE_CODE_ENUM
20625 @item TYPE_CODE_ENUM
20626 The type is an enum.
20628 @findex TYPE_CODE_FLAGS
20629 @findex gdb.TYPE_CODE_FLAGS
20630 @item TYPE_CODE_FLAGS
20631 A bit flags type, used for things such as status registers.
20633 @findex TYPE_CODE_FUNC
20634 @findex gdb.TYPE_CODE_FUNC
20635 @item TYPE_CODE_FUNC
20636 The type is a function.
20638 @findex TYPE_CODE_INT
20639 @findex gdb.TYPE_CODE_INT
20640 @item TYPE_CODE_INT
20641 The type is an integer type.
20643 @findex TYPE_CODE_FLT
20644 @findex gdb.TYPE_CODE_FLT
20645 @item TYPE_CODE_FLT
20646 A floating point type.
20648 @findex TYPE_CODE_VOID
20649 @findex gdb.TYPE_CODE_VOID
20650 @item TYPE_CODE_VOID
20651 The special type @code{void}.
20653 @findex TYPE_CODE_SET
20654 @findex gdb.TYPE_CODE_SET
20655 @item TYPE_CODE_SET
20658 @findex TYPE_CODE_RANGE
20659 @findex gdb.TYPE_CODE_RANGE
20660 @item TYPE_CODE_RANGE
20661 A range type, that is, an integer type with bounds.
20663 @findex TYPE_CODE_STRING
20664 @findex gdb.TYPE_CODE_STRING
20665 @item TYPE_CODE_STRING
20666 A string type. Note that this is only used for certain languages with
20667 language-defined string types; C strings are not represented this way.
20669 @findex TYPE_CODE_BITSTRING
20670 @findex gdb.TYPE_CODE_BITSTRING
20671 @item TYPE_CODE_BITSTRING
20674 @findex TYPE_CODE_ERROR
20675 @findex gdb.TYPE_CODE_ERROR
20676 @item TYPE_CODE_ERROR
20677 An unknown or erroneous type.
20679 @findex TYPE_CODE_METHOD
20680 @findex gdb.TYPE_CODE_METHOD
20681 @item TYPE_CODE_METHOD
20682 A method type, as found in C@t{++} or Java.
20684 @findex TYPE_CODE_METHODPTR
20685 @findex gdb.TYPE_CODE_METHODPTR
20686 @item TYPE_CODE_METHODPTR
20687 A pointer-to-member-function.
20689 @findex TYPE_CODE_MEMBERPTR
20690 @findex gdb.TYPE_CODE_MEMBERPTR
20691 @item TYPE_CODE_MEMBERPTR
20692 A pointer-to-member.
20694 @findex TYPE_CODE_REF
20695 @findex gdb.TYPE_CODE_REF
20696 @item TYPE_CODE_REF
20699 @findex TYPE_CODE_CHAR
20700 @findex gdb.TYPE_CODE_CHAR
20701 @item TYPE_CODE_CHAR
20704 @findex TYPE_CODE_BOOL
20705 @findex gdb.TYPE_CODE_BOOL
20706 @item TYPE_CODE_BOOL
20709 @findex TYPE_CODE_COMPLEX
20710 @findex gdb.TYPE_CODE_COMPLEX
20711 @item TYPE_CODE_COMPLEX
20712 A complex float type.
20714 @findex TYPE_CODE_TYPEDEF
20715 @findex gdb.TYPE_CODE_TYPEDEF
20716 @item TYPE_CODE_TYPEDEF
20717 A typedef to some other type.
20719 @findex TYPE_CODE_NAMESPACE
20720 @findex gdb.TYPE_CODE_NAMESPACE
20721 @item TYPE_CODE_NAMESPACE
20722 A C@t{++} namespace.
20724 @findex TYPE_CODE_DECFLOAT
20725 @findex gdb.TYPE_CODE_DECFLOAT
20726 @item TYPE_CODE_DECFLOAT
20727 A decimal floating point type.
20729 @findex TYPE_CODE_INTERNAL_FUNCTION
20730 @findex gdb.TYPE_CODE_INTERNAL_FUNCTION
20731 @item TYPE_CODE_INTERNAL_FUNCTION
20732 A function internal to @value{GDBN}. This is the type used to represent
20733 convenience functions.
20736 @node Pretty Printing API
20737 @subsubsection Pretty Printing API
20739 An example output is provided (@pxref{Pretty Printing}).
20741 A pretty-printer is just an object that holds a value and implements a
20742 specific interface, defined here.
20744 @defop Operation {pretty printer} children (self)
20745 @value{GDBN} will call this method on a pretty-printer to compute the
20746 children of the pretty-printer's value.
20748 This method must return an object conforming to the Python iterator
20749 protocol. Each item returned by the iterator must be a tuple holding
20750 two elements. The first element is the ``name'' of the child; the
20751 second element is the child's value. The value can be any Python
20752 object which is convertible to a @value{GDBN} value.
20754 This method is optional. If it does not exist, @value{GDBN} will act
20755 as though the value has no children.
20758 @defop Operation {pretty printer} display_hint (self)
20759 The CLI may call this method and use its result to change the
20760 formatting of a value. The result will also be supplied to an MI
20761 consumer as a @samp{displayhint} attribute of the variable being
20764 This method is optional. If it does exist, this method must return a
20767 Some display hints are predefined by @value{GDBN}:
20771 Indicate that the object being printed is ``array-like''. The CLI
20772 uses this to respect parameters such as @code{set print elements} and
20773 @code{set print array}.
20776 Indicate that the object being printed is ``map-like'', and that the
20777 children of this value can be assumed to alternate between keys and
20781 Indicate that the object being printed is ``string-like''. If the
20782 printer's @code{to_string} method returns a Python string of some
20783 kind, then @value{GDBN} will call its internal language-specific
20784 string-printing function to format the string. For the CLI this means
20785 adding quotation marks, possibly escaping some characters, respecting
20786 @code{set print elements}, and the like.
20790 @defop Operation {pretty printer} to_string (self)
20791 @value{GDBN} will call this method to display the string
20792 representation of the value passed to the object's constructor.
20794 When printing from the CLI, if the @code{to_string} method exists,
20795 then @value{GDBN} will prepend its result to the values returned by
20796 @code{children}. Exactly how this formatting is done is dependent on
20797 the display hint, and may change as more hints are added. Also,
20798 depending on the print settings (@pxref{Print Settings}), the CLI may
20799 print just the result of @code{to_string} in a stack trace, omitting
20800 the result of @code{children}.
20802 If this method returns a string, it is printed verbatim.
20804 Otherwise, if this method returns an instance of @code{gdb.Value},
20805 then @value{GDBN} prints this value. This may result in a call to
20806 another pretty-printer.
20808 If instead the method returns a Python value which is convertible to a
20809 @code{gdb.Value}, then @value{GDBN} performs the conversion and prints
20810 the resulting value. Again, this may result in a call to another
20811 pretty-printer. Python scalars (integers, floats, and booleans) and
20812 strings are convertible to @code{gdb.Value}; other types are not.
20814 Finally, if this method returns @code{None} then no further operations
20815 are peformed in this method and nothing is printed.
20817 If the result is not one of these types, an exception is raised.
20820 @node Selecting Pretty-Printers
20821 @subsubsection Selecting Pretty-Printers
20823 The Python list @code{gdb.pretty_printers} contains an array of
20824 functions or callable objects that have been registered via addition
20825 as a pretty-printer.
20826 Each @code{gdb.Progspace} contains a @code{pretty_printers} attribute.
20827 Each @code{gdb.Objfile} also contains a @code{pretty_printers}
20830 A function on one of these lists is passed a single @code{gdb.Value}
20831 argument and should return a pretty-printer object conforming to the
20832 interface definition above (@pxref{Pretty Printing API}). If a function
20833 cannot create a pretty-printer for the value, it should return
20836 @value{GDBN} first checks the @code{pretty_printers} attribute of each
20837 @code{gdb.Objfile} in the current program space and iteratively calls
20838 each enabled function (@pxref{Disabling Pretty-Printers})
20839 in the list for that @code{gdb.Objfile} until it receives
20840 a pretty-printer object.
20841 If no pretty-printer is found in the objfile lists, @value{GDBN} then
20842 searches the pretty-printer list of the current program space,
20843 calling each enabled function until an object is returned.
20844 After these lists have been exhausted, it tries the global
20845 @code{gdb.pretty_printers} list, again calling each enabled function until an
20846 object is returned.
20848 The order in which the objfiles are searched is not specified. For a
20849 given list, functions are always invoked from the head of the list,
20850 and iterated over sequentially until the end of the list, or a printer
20851 object is returned.
20853 Here is an example showing how a @code{std::string} printer might be
20857 class StdStringPrinter:
20858 "Print a std::string"
20860 def __init__ (self, val):
20863 def to_string (self):
20864 return self.val['_M_dataplus']['_M_p']
20866 def display_hint (self):
20870 And here is an example showing how a lookup function for the printer
20871 example above might be written.
20874 def str_lookup_function (val):
20876 lookup_tag = val.type.tag
20877 regex = re.compile ("^std::basic_string<char,.*>$")
20878 if lookup_tag == None:
20880 if regex.match (lookup_tag):
20881 return StdStringPrinter (val)
20886 The example lookup function extracts the value's type, and attempts to
20887 match it to a type that it can pretty-print. If it is a type the
20888 printer can pretty-print, it will return a printer object. If not, it
20889 returns @code{None}.
20891 We recommend that you put your core pretty-printers into a Python
20892 package. If your pretty-printers are for use with a library, we
20893 further recommend embedding a version number into the package name.
20894 This practice will enable @value{GDBN} to load multiple versions of
20895 your pretty-printers at the same time, because they will have
20898 You should write auto-loaded code (@pxref{Auto-loading}) such that it
20899 can be evaluated multiple times without changing its meaning. An
20900 ideal auto-load file will consist solely of @code{import}s of your
20901 printer modules, followed by a call to a register pretty-printers with
20902 the current objfile.
20904 Taken as a whole, this approach will scale nicely to multiple
20905 inferiors, each potentially using a different library version.
20906 Embedding a version number in the Python package name will ensure that
20907 @value{GDBN} is able to load both sets of printers simultaneously.
20908 Then, because the search for pretty-printers is done by objfile, and
20909 because your auto-loaded code took care to register your library's
20910 printers with a specific objfile, @value{GDBN} will find the correct
20911 printers for the specific version of the library used by each
20914 To continue the @code{std::string} example (@pxref{Pretty Printing API}),
20915 this code might appear in @code{gdb.libstdcxx.v6}:
20918 def register_printers (objfile):
20919 objfile.pretty_printers.add (str_lookup_function)
20923 And then the corresponding contents of the auto-load file would be:
20926 import gdb.libstdcxx.v6
20927 gdb.libstdcxx.v6.register_printers (gdb.current_objfile ())
20930 @node Disabling Pretty-Printers
20931 @subsubsection Disabling Pretty-Printers
20932 @cindex disabling pretty-printers
20934 For various reasons a pretty-printer may not work.
20935 For example, the underlying data structure may have changed and
20936 the pretty-printer is out of date.
20938 The consequences of a broken pretty-printer are severe enough that
20939 @value{GDBN} provides support for enabling and disabling individual
20940 printers. For example, if @code{print frame-arguments} is on,
20941 a backtrace can become highly illegible if any argument is printed
20942 with a broken printer.
20944 Pretty-printers are enabled and disabled by attaching an @code{enabled}
20945 attribute to the registered function or callable object. If this attribute
20946 is present and its value is @code{False}, the printer is disabled, otherwise
20947 the printer is enabled.
20949 @node Commands In Python
20950 @subsubsection Commands In Python
20952 @cindex commands in python
20953 @cindex python commands
20954 You can implement new @value{GDBN} CLI commands in Python. A CLI
20955 command is implemented using an instance of the @code{gdb.Command}
20956 class, most commonly using a subclass.
20958 @defmethod Command __init__ name @var{command_class} @r{[}@var{completer_class}@r{]} @r{[}@var{prefix}@r{]}
20959 The object initializer for @code{Command} registers the new command
20960 with @value{GDBN}. This initializer is normally invoked from the
20961 subclass' own @code{__init__} method.
20963 @var{name} is the name of the command. If @var{name} consists of
20964 multiple words, then the initial words are looked for as prefix
20965 commands. In this case, if one of the prefix commands does not exist,
20966 an exception is raised.
20968 There is no support for multi-line commands.
20970 @var{command_class} should be one of the @samp{COMMAND_} constants
20971 defined below. This argument tells @value{GDBN} how to categorize the
20972 new command in the help system.
20974 @var{completer_class} is an optional argument. If given, it should be
20975 one of the @samp{COMPLETE_} constants defined below. This argument
20976 tells @value{GDBN} how to perform completion for this command. If not
20977 given, @value{GDBN} will attempt to complete using the object's
20978 @code{complete} method (see below); if no such method is found, an
20979 error will occur when completion is attempted.
20981 @var{prefix} is an optional argument. If @code{True}, then the new
20982 command is a prefix command; sub-commands of this command may be
20985 The help text for the new command is taken from the Python
20986 documentation string for the command's class, if there is one. If no
20987 documentation string is provided, the default value ``This command is
20988 not documented.'' is used.
20991 @cindex don't repeat Python command
20992 @defmethod Command dont_repeat
20993 By default, a @value{GDBN} command is repeated when the user enters a
20994 blank line at the command prompt. A command can suppress this
20995 behavior by invoking the @code{dont_repeat} method. This is similar
20996 to the user command @code{dont-repeat}, see @ref{Define, dont-repeat}.
20999 @defmethod Command invoke argument from_tty
21000 This method is called by @value{GDBN} when this command is invoked.
21002 @var{argument} is a string. It is the argument to the command, after
21003 leading and trailing whitespace has been stripped.
21005 @var{from_tty} is a boolean argument. When true, this means that the
21006 command was entered by the user at the terminal; when false it means
21007 that the command came from elsewhere.
21009 If this method throws an exception, it is turned into a @value{GDBN}
21010 @code{error} call. Otherwise, the return value is ignored.
21012 @findex gdb.string_to_argv
21013 To break @var{argument} up into an argv-like string use
21014 @code{gdb.string_to_argv}. This function behaves identically to
21015 @value{GDBN}'s internal argument lexer @code{buildargv}.
21016 It is recommended to use this for consistency.
21017 Arguments are separated by spaces and may be quoted.
21021 print gdb.string_to_argv ("1 2\ \\\"3 '4 \"5' \"6 '7\"")
21022 ['1', '2 "3', '4 "5', "6 '7"]
21027 @cindex completion of Python commands
21028 @defmethod Command complete text word
21029 This method is called by @value{GDBN} when the user attempts
21030 completion on this command. All forms of completion are handled by
21031 this method, that is, the @key{TAB} and @key{M-?} key bindings
21032 (@pxref{Completion}), and the @code{complete} command (@pxref{Help,
21035 The arguments @var{text} and @var{word} are both strings. @var{text}
21036 holds the complete command line up to the cursor's location.
21037 @var{word} holds the last word of the command line; this is computed
21038 using a word-breaking heuristic.
21040 The @code{complete} method can return several values:
21043 If the return value is a sequence, the contents of the sequence are
21044 used as the completions. It is up to @code{complete} to ensure that the
21045 contents actually do complete the word. A zero-length sequence is
21046 allowed, it means that there were no completions available. Only
21047 string elements of the sequence are used; other elements in the
21048 sequence are ignored.
21051 If the return value is one of the @samp{COMPLETE_} constants defined
21052 below, then the corresponding @value{GDBN}-internal completion
21053 function is invoked, and its result is used.
21056 All other results are treated as though there were no available
21061 When a new command is registered, it must be declared as a member of
21062 some general class of commands. This is used to classify top-level
21063 commands in the on-line help system; note that prefix commands are not
21064 listed under their own category but rather that of their top-level
21065 command. The available classifications are represented by constants
21066 defined in the @code{gdb} module:
21069 @findex COMMAND_NONE
21070 @findex gdb.COMMAND_NONE
21072 The command does not belong to any particular class. A command in
21073 this category will not be displayed in any of the help categories.
21075 @findex COMMAND_RUNNING
21076 @findex gdb.COMMAND_RUNNING
21077 @item COMMAND_RUNNING
21078 The command is related to running the inferior. For example,
21079 @code{start}, @code{step}, and @code{continue} are in this category.
21080 Type @kbd{help running} at the @value{GDBN} prompt to see a list of
21081 commands in this category.
21083 @findex COMMAND_DATA
21084 @findex gdb.COMMAND_DATA
21086 The command is related to data or variables. For example,
21087 @code{call}, @code{find}, and @code{print} are in this category. Type
21088 @kbd{help data} at the @value{GDBN} prompt to see a list of commands
21091 @findex COMMAND_STACK
21092 @findex gdb.COMMAND_STACK
21093 @item COMMAND_STACK
21094 The command has to do with manipulation of the stack. For example,
21095 @code{backtrace}, @code{frame}, and @code{return} are in this
21096 category. Type @kbd{help stack} at the @value{GDBN} prompt to see a
21097 list of commands in this category.
21099 @findex COMMAND_FILES
21100 @findex gdb.COMMAND_FILES
21101 @item COMMAND_FILES
21102 This class is used for file-related commands. For example,
21103 @code{file}, @code{list} and @code{section} are in this category.
21104 Type @kbd{help files} at the @value{GDBN} prompt to see a list of
21105 commands in this category.
21107 @findex COMMAND_SUPPORT
21108 @findex gdb.COMMAND_SUPPORT
21109 @item COMMAND_SUPPORT
21110 This should be used for ``support facilities'', generally meaning
21111 things that are useful to the user when interacting with @value{GDBN},
21112 but not related to the state of the inferior. For example,
21113 @code{help}, @code{make}, and @code{shell} are in this category. Type
21114 @kbd{help support} at the @value{GDBN} prompt to see a list of
21115 commands in this category.
21117 @findex COMMAND_STATUS
21118 @findex gdb.COMMAND_STATUS
21119 @item COMMAND_STATUS
21120 The command is an @samp{info}-related command, that is, related to the
21121 state of @value{GDBN} itself. For example, @code{info}, @code{macro},
21122 and @code{show} are in this category. Type @kbd{help status} at the
21123 @value{GDBN} prompt to see a list of commands in this category.
21125 @findex COMMAND_BREAKPOINTS
21126 @findex gdb.COMMAND_BREAKPOINTS
21127 @item COMMAND_BREAKPOINTS
21128 The command has to do with breakpoints. For example, @code{break},
21129 @code{clear}, and @code{delete} are in this category. Type @kbd{help
21130 breakpoints} at the @value{GDBN} prompt to see a list of commands in
21133 @findex COMMAND_TRACEPOINTS
21134 @findex gdb.COMMAND_TRACEPOINTS
21135 @item COMMAND_TRACEPOINTS
21136 The command has to do with tracepoints. For example, @code{trace},
21137 @code{actions}, and @code{tfind} are in this category. Type
21138 @kbd{help tracepoints} at the @value{GDBN} prompt to see a list of
21139 commands in this category.
21141 @findex COMMAND_OBSCURE
21142 @findex gdb.COMMAND_OBSCURE
21143 @item COMMAND_OBSCURE
21144 The command is only used in unusual circumstances, or is not of
21145 general interest to users. For example, @code{checkpoint},
21146 @code{fork}, and @code{stop} are in this category. Type @kbd{help
21147 obscure} at the @value{GDBN} prompt to see a list of commands in this
21150 @findex COMMAND_MAINTENANCE
21151 @findex gdb.COMMAND_MAINTENANCE
21152 @item COMMAND_MAINTENANCE
21153 The command is only useful to @value{GDBN} maintainers. The
21154 @code{maintenance} and @code{flushregs} commands are in this category.
21155 Type @kbd{help internals} at the @value{GDBN} prompt to see a list of
21156 commands in this category.
21159 A new command can use a predefined completion function, either by
21160 specifying it via an argument at initialization, or by returning it
21161 from the @code{complete} method. These predefined completion
21162 constants are all defined in the @code{gdb} module:
21165 @findex COMPLETE_NONE
21166 @findex gdb.COMPLETE_NONE
21167 @item COMPLETE_NONE
21168 This constant means that no completion should be done.
21170 @findex COMPLETE_FILENAME
21171 @findex gdb.COMPLETE_FILENAME
21172 @item COMPLETE_FILENAME
21173 This constant means that filename completion should be performed.
21175 @findex COMPLETE_LOCATION
21176 @findex gdb.COMPLETE_LOCATION
21177 @item COMPLETE_LOCATION
21178 This constant means that location completion should be done.
21179 @xref{Specify Location}.
21181 @findex COMPLETE_COMMAND
21182 @findex gdb.COMPLETE_COMMAND
21183 @item COMPLETE_COMMAND
21184 This constant means that completion should examine @value{GDBN}
21187 @findex COMPLETE_SYMBOL
21188 @findex gdb.COMPLETE_SYMBOL
21189 @item COMPLETE_SYMBOL
21190 This constant means that completion should be done using symbol names
21194 The following code snippet shows how a trivial CLI command can be
21195 implemented in Python:
21198 class HelloWorld (gdb.Command):
21199 """Greet the whole world."""
21201 def __init__ (self):
21202 super (HelloWorld, self).__init__ ("hello-world", gdb.COMMAND_OBSCURE)
21204 def invoke (self, arg, from_tty):
21205 print "Hello, World!"
21210 The last line instantiates the class, and is necessary to trigger the
21211 registration of the command with @value{GDBN}. Depending on how the
21212 Python code is read into @value{GDBN}, you may need to import the
21213 @code{gdb} module explicitly.
21215 @node Parameters In Python
21216 @subsubsection Parameters In Python
21218 @cindex parameters in python
21219 @cindex python parameters
21220 @tindex gdb.Parameter
21222 You can implement new @value{GDBN} parameters using Python. A new
21223 parameter is implemented as an instance of the @code{gdb.Parameter}
21226 Parameters are exposed to the user via the @code{set} and
21227 @code{show} commands. @xref{Help}.
21229 There are many parameters that already exist and can be set in
21230 @value{GDBN}. Two examples are: @code{set follow fork} and
21231 @code{set charset}. Setting these parameters influences certain
21232 behavior in @value{GDBN}. Similarly, you can define parameters that
21233 can be used to influence behavior in custom Python scripts and commands.
21235 @defmethod Parameter __init__ name @var{command-class} @var{parameter-class} @r{[}@var{enum-sequence}@r{]}
21236 The object initializer for @code{Parameter} registers the new
21237 parameter with @value{GDBN}. This initializer is normally invoked
21238 from the subclass' own @code{__init__} method.
21240 @var{name} is the name of the new parameter. If @var{name} consists
21241 of multiple words, then the initial words are looked for as prefix
21242 parameters. An example of this can be illustrated with the
21243 @code{set print} set of parameters. If @var{name} is
21244 @code{print foo}, then @code{print} will be searched as the prefix
21245 parameter. In this case the parameter can subsequently be accessed in
21246 @value{GDBN} as @code{set print foo}.
21248 If @var{name} consists of multiple words, and no prefix parameter group
21249 can be found, an exception is raised.
21251 @var{command-class} should be one of the @samp{COMMAND_} constants
21252 (@pxref{Commands In Python}). This argument tells @value{GDBN} how to
21253 categorize the new parameter in the help system.
21255 @var{parameter-class} should be one of the @samp{PARAM_} constants
21256 defined below. This argument tells @value{GDBN} the type of the new
21257 parameter; this information is used for input validation and
21260 If @var{parameter-class} is @code{PARAM_ENUM}, then
21261 @var{enum-sequence} must be a sequence of strings. These strings
21262 represent the possible values for the parameter.
21264 If @var{parameter-class} is not @code{PARAM_ENUM}, then the presence
21265 of a fourth argument will cause an exception to be thrown.
21267 The help text for the new parameter is taken from the Python
21268 documentation string for the parameter's class, if there is one. If
21269 there is no documentation string, a default value is used.
21272 @defivar Parameter set_doc
21273 If this attribute exists, and is a string, then its value is used as
21274 the help text for this parameter's @code{set} command. The value is
21275 examined when @code{Parameter.__init__} is invoked; subsequent changes
21279 @defivar Parameter show_doc
21280 If this attribute exists, and is a string, then its value is used as
21281 the help text for this parameter's @code{show} command. The value is
21282 examined when @code{Parameter.__init__} is invoked; subsequent changes
21286 @defivar Parameter value
21287 The @code{value} attribute holds the underlying value of the
21288 parameter. It can be read and assigned to just as any other
21289 attribute. @value{GDBN} does validation when assignments are made.
21293 When a new parameter is defined, its type must be specified. The
21294 available types are represented by constants defined in the @code{gdb}
21298 @findex PARAM_BOOLEAN
21299 @findex gdb.PARAM_BOOLEAN
21300 @item PARAM_BOOLEAN
21301 The value is a plain boolean. The Python boolean values, @code{True}
21302 and @code{False} are the only valid values.
21304 @findex PARAM_AUTO_BOOLEAN
21305 @findex gdb.PARAM_AUTO_BOOLEAN
21306 @item PARAM_AUTO_BOOLEAN
21307 The value has three possible states: true, false, and @samp{auto}. In
21308 Python, true and false are represented using boolean constants, and
21309 @samp{auto} is represented using @code{None}.
21311 @findex PARAM_UINTEGER
21312 @findex gdb.PARAM_UINTEGER
21313 @item PARAM_UINTEGER
21314 The value is an unsigned integer. The value of 0 should be
21315 interpreted to mean ``unlimited''.
21317 @findex PARAM_INTEGER
21318 @findex gdb.PARAM_INTEGER
21319 @item PARAM_INTEGER
21320 The value is a signed integer. The value of 0 should be interpreted
21321 to mean ``unlimited''.
21323 @findex PARAM_STRING
21324 @findex gdb.PARAM_STRING
21326 The value is a string. When the user modifies the string, any escape
21327 sequences, such as @samp{\t}, @samp{\f}, and octal escapes, are
21328 translated into corresponding characters and encoded into the current
21331 @findex PARAM_STRING_NOESCAPE
21332 @findex gdb.PARAM_STRING_NOESCAPE
21333 @item PARAM_STRING_NOESCAPE
21334 The value is a string. When the user modifies the string, escapes are
21335 passed through untranslated.
21337 @findex PARAM_OPTIONAL_FILENAME
21338 @findex gdb.PARAM_OPTIONAL_FILENAME
21339 @item PARAM_OPTIONAL_FILENAME
21340 The value is a either a filename (a string), or @code{None}.
21342 @findex PARAM_FILENAME
21343 @findex gdb.PARAM_FILENAME
21344 @item PARAM_FILENAME
21345 The value is a filename. This is just like
21346 @code{PARAM_STRING_NOESCAPE}, but uses file names for completion.
21348 @findex PARAM_ZINTEGER
21349 @findex gdb.PARAM_ZINTEGER
21350 @item PARAM_ZINTEGER
21351 The value is an integer. This is like @code{PARAM_INTEGER}, except 0
21352 is interpreted as itself.
21355 @findex gdb.PARAM_ENUM
21357 The value is a string, which must be one of a collection string
21358 constants provided when the parameter is created.
21361 @node Functions In Python
21362 @subsubsection Writing new convenience functions
21364 @cindex writing convenience functions
21365 @cindex convenience functions in python
21366 @cindex python convenience functions
21367 @tindex gdb.Function
21369 You can implement new convenience functions (@pxref{Convenience Vars})
21370 in Python. A convenience function is an instance of a subclass of the
21371 class @code{gdb.Function}.
21373 @defmethod Function __init__ name
21374 The initializer for @code{Function} registers the new function with
21375 @value{GDBN}. The argument @var{name} is the name of the function,
21376 a string. The function will be visible to the user as a convenience
21377 variable of type @code{internal function}, whose name is the same as
21378 the given @var{name}.
21380 The documentation for the new function is taken from the documentation
21381 string for the new class.
21384 @defmethod Function invoke @var{*args}
21385 When a convenience function is evaluated, its arguments are converted
21386 to instances of @code{gdb.Value}, and then the function's
21387 @code{invoke} method is called. Note that @value{GDBN} does not
21388 predetermine the arity of convenience functions. Instead, all
21389 available arguments are passed to @code{invoke}, following the
21390 standard Python calling convention. In particular, a convenience
21391 function can have default values for parameters without ill effect.
21393 The return value of this method is used as its value in the enclosing
21394 expression. If an ordinary Python value is returned, it is converted
21395 to a @code{gdb.Value} following the usual rules.
21398 The following code snippet shows how a trivial convenience function can
21399 be implemented in Python:
21402 class Greet (gdb.Function):
21403 """Return string to greet someone.
21404 Takes a name as argument."""
21406 def __init__ (self):
21407 super (Greet, self).__init__ ("greet")
21409 def invoke (self, name):
21410 return "Hello, %s!" % name.string ()
21415 The last line instantiates the class, and is necessary to trigger the
21416 registration of the function with @value{GDBN}. Depending on how the
21417 Python code is read into @value{GDBN}, you may need to import the
21418 @code{gdb} module explicitly.
21420 @node Progspaces In Python
21421 @subsubsection Program Spaces In Python
21423 @cindex progspaces in python
21424 @tindex gdb.Progspace
21426 A program space, or @dfn{progspace}, represents a symbolic view
21427 of an address space.
21428 It consists of all of the objfiles of the program.
21429 @xref{Objfiles In Python}.
21430 @xref{Inferiors and Programs, program spaces}, for more details
21431 about program spaces.
21433 The following progspace-related functions are available in the
21436 @findex gdb.current_progspace
21437 @defun current_progspace
21438 This function returns the program space of the currently selected inferior.
21439 @xref{Inferiors and Programs}.
21442 @findex gdb.progspaces
21444 Return a sequence of all the progspaces currently known to @value{GDBN}.
21447 Each progspace is represented by an instance of the @code{gdb.Progspace}
21450 @defivar Progspace filename
21451 The file name of the progspace as a string.
21454 @defivar Progspace pretty_printers
21455 The @code{pretty_printers} attribute is a list of functions. It is
21456 used to look up pretty-printers. A @code{Value} is passed to each
21457 function in order; if the function returns @code{None}, then the
21458 search continues. Otherwise, the return value should be an object
21459 which is used to format the value. @xref{Pretty Printing API}, for more
21463 @node Objfiles In Python
21464 @subsubsection Objfiles In Python
21466 @cindex objfiles in python
21467 @tindex gdb.Objfile
21469 @value{GDBN} loads symbols for an inferior from various
21470 symbol-containing files (@pxref{Files}). These include the primary
21471 executable file, any shared libraries used by the inferior, and any
21472 separate debug info files (@pxref{Separate Debug Files}).
21473 @value{GDBN} calls these symbol-containing files @dfn{objfiles}.
21475 The following objfile-related functions are available in the
21478 @findex gdb.current_objfile
21479 @defun current_objfile
21480 When auto-loading a Python script (@pxref{Auto-loading}), @value{GDBN}
21481 sets the ``current objfile'' to the corresponding objfile. This
21482 function returns the current objfile. If there is no current objfile,
21483 this function returns @code{None}.
21486 @findex gdb.objfiles
21488 Return a sequence of all the objfiles current known to @value{GDBN}.
21489 @xref{Objfiles In Python}.
21492 Each objfile is represented by an instance of the @code{gdb.Objfile}
21495 @defivar Objfile filename
21496 The file name of the objfile as a string.
21499 @defivar Objfile pretty_printers
21500 The @code{pretty_printers} attribute is a list of functions. It is
21501 used to look up pretty-printers. A @code{Value} is passed to each
21502 function in order; if the function returns @code{None}, then the
21503 search continues. Otherwise, the return value should be an object
21504 which is used to format the value. @xref{Pretty Printing API}, for more
21508 @node Frames In Python
21509 @subsubsection Accessing inferior stack frames from Python.
21511 @cindex frames in python
21512 When the debugged program stops, @value{GDBN} is able to analyze its call
21513 stack (@pxref{Frames,,Stack frames}). The @code{gdb.Frame} class
21514 represents a frame in the stack. A @code{gdb.Frame} object is only valid
21515 while its corresponding frame exists in the inferior's stack. If you try
21516 to use an invalid frame object, @value{GDBN} will throw a @code{RuntimeError}
21519 Two @code{gdb.Frame} objects can be compared for equality with the @code{==}
21523 (@value{GDBP}) python print gdb.newest_frame() == gdb.selected_frame ()
21527 The following frame-related functions are available in the @code{gdb} module:
21529 @findex gdb.selected_frame
21530 @defun selected_frame
21531 Return the selected frame object. (@pxref{Selection,,Selecting a Frame}).
21534 @defun frame_stop_reason_string reason
21535 Return a string explaining the reason why @value{GDBN} stopped unwinding
21536 frames, as expressed by the given @var{reason} code (an integer, see the
21537 @code{unwind_stop_reason} method further down in this section).
21540 A @code{gdb.Frame} object has the following methods:
21543 @defmethod Frame is_valid
21544 Returns true if the @code{gdb.Frame} object is valid, false if not.
21545 A frame object can become invalid if the frame it refers to doesn't
21546 exist anymore in the inferior. All @code{gdb.Frame} methods will throw
21547 an exception if it is invalid at the time the method is called.
21550 @defmethod Frame name
21551 Returns the function name of the frame, or @code{None} if it can't be
21555 @defmethod Frame type
21556 Returns the type of the frame. The value can be one of
21557 @code{gdb.NORMAL_FRAME}, @code{gdb.DUMMY_FRAME}, @code{gdb.SIGTRAMP_FRAME}
21558 or @code{gdb.SENTINEL_FRAME}.
21561 @defmethod Frame unwind_stop_reason
21562 Return an integer representing the reason why it's not possible to find
21563 more frames toward the outermost frame. Use
21564 @code{gdb.frame_stop_reason_string} to convert the value returned by this
21565 function to a string.
21568 @defmethod Frame pc
21569 Returns the frame's resume address.
21572 @defmethod Frame block
21573 Return the frame's code block. @xref{Blocks In Python}.
21576 @defmethod Frame function
21577 Return the symbol for the function corresponding to this frame.
21578 @xref{Symbols In Python}.
21581 @defmethod Frame older
21582 Return the frame that called this frame.
21585 @defmethod Frame newer
21586 Return the frame called by this frame.
21589 @defmethod Frame find_sal
21590 Return the frame's symtab and line object.
21591 @xref{Symbol Tables In Python}.
21594 @defmethod Frame read_var variable @r{[}block@r{]}
21595 Return the value of @var{variable} in this frame. If the optional
21596 argument @var{block} is provided, search for the variable from that
21597 block; otherwise start at the frame's current block (which is
21598 determined by the frame's current program counter). @var{variable}
21599 must be a string or a @code{gdb.Symbol} object. @var{block} must be a
21600 @code{gdb.Block} object.
21603 @defmethod Frame select
21604 Set this frame to be the selected frame. @xref{Stack, ,Examining the
21609 @node Blocks In Python
21610 @subsubsection Accessing frame blocks from Python.
21612 @cindex blocks in python
21615 Within each frame, @value{GDBN} maintains information on each block
21616 stored in that frame. These blocks are organized hierarchically, and
21617 are represented individually in Python as a @code{gdb.Block}.
21618 Please see @ref{Frames In Python}, for a more in-depth discussion on
21619 frames. Furthermore, see @ref{Stack, ,Examining the Stack}, for more
21620 detailed technical information on @value{GDBN}'s book-keeping of the
21623 The following block-related functions are available in the @code{gdb}
21626 @findex gdb.block_for_pc
21627 @defun block_for_pc pc
21628 Return the @code{gdb.Block} containing the given @var{pc} value. If the
21629 block cannot be found for the @var{pc} value specified, the function
21630 will return @code{None}.
21633 A @code{gdb.Block} object has the following attributes:
21636 @defivar Block start
21637 The start address of the block. This attribute is not writable.
21641 The end address of the block. This attribute is not writable.
21644 @defivar Block function
21645 The name of the block represented as a @code{gdb.Symbol}. If the
21646 block is not named, then this attribute holds @code{None}. This
21647 attribute is not writable.
21650 @defivar Block superblock
21651 The block containing this block. If this parent block does not exist,
21652 this attribute holds @code{None}. This attribute is not writable.
21656 @node Symbols In Python
21657 @subsubsection Python representation of Symbols.
21659 @cindex symbols in python
21662 @value{GDBN} represents every variable, function and type as an
21663 entry in a symbol table. @xref{Symbols, ,Examining the Symbol Table}.
21664 Similarly, Python represents these symbols in @value{GDBN} with the
21665 @code{gdb.Symbol} object.
21667 The following symbol-related functions are available in the @code{gdb}
21670 @findex gdb.lookup_symbol
21671 @defun lookup_symbol name [block] [domain]
21672 This function searches for a symbol by name. The search scope can be
21673 restricted to the parameters defined in the optional domain and block
21676 @var{name} is the name of the symbol. It must be a string. The
21677 optional @var{block} argument restricts the search to symbols visible
21678 in that @var{block}. The @var{block} argument must be a
21679 @code{gdb.Block} object. The optional @var{domain} argument restricts
21680 the search to the domain type. The @var{domain} argument must be a
21681 domain constant defined in the @code{gdb} module and described later
21685 A @code{gdb.Symbol} object has the following attributes:
21688 @defivar Symbol symtab
21689 The symbol table in which the symbol appears. This attribute is
21690 represented as a @code{gdb.Symtab} object. @xref{Symbol Tables In
21691 Python}. This attribute is not writable.
21694 @defivar Symbol name
21695 The name of the symbol as a string. This attribute is not writable.
21698 @defivar Symbol linkage_name
21699 The name of the symbol, as used by the linker (i.e., may be mangled).
21700 This attribute is not writable.
21703 @defivar Symbol print_name
21704 The name of the symbol in a form suitable for output. This is either
21705 @code{name} or @code{linkage_name}, depending on whether the user
21706 asked @value{GDBN} to display demangled or mangled names.
21709 @defivar Symbol addr_class
21710 The address class of the symbol. This classifies how to find the value
21711 of a symbol. Each address class is a constant defined in the
21712 @code{gdb} module and described later in this chapter.
21715 @defivar Symbol is_argument
21716 @code{True} if the symbol is an argument of a function.
21719 @defivar Symbol is_constant
21720 @code{True} if the symbol is a constant.
21723 @defivar Symbol is_function
21724 @code{True} if the symbol is a function or a method.
21727 @defivar Symbol is_variable
21728 @code{True} if the symbol is a variable.
21732 The available domain categories in @code{gdb.Symbol} are represented
21733 as constants in the @code{gdb} module:
21736 @findex SYMBOL_UNDEF_DOMAIN
21737 @findex gdb.SYMBOL_UNDEF_DOMAIN
21738 @item SYMBOL_UNDEF_DOMAIN
21739 This is used when a domain has not been discovered or none of the
21740 following domains apply. This usually indicates an error either
21741 in the symbol information or in @value{GDBN}'s handling of symbols.
21742 @findex SYMBOL_VAR_DOMAIN
21743 @findex gdb.SYMBOL_VAR_DOMAIN
21744 @item SYMBOL_VAR_DOMAIN
21745 This domain contains variables, function names, typedef names and enum
21747 @findex SYMBOL_STRUCT_DOMAIN
21748 @findex gdb.SYMBOL_STRUCT_DOMAIN
21749 @item SYMBOL_STRUCT_DOMAIN
21750 This domain holds struct, union and enum type names.
21751 @findex SYMBOL_LABEL_DOMAIN
21752 @findex gdb.SYMBOL_LABEL_DOMAIN
21753 @item SYMBOL_LABEL_DOMAIN
21754 This domain contains names of labels (for gotos).
21755 @findex SYMBOL_VARIABLES_DOMAIN
21756 @findex gdb.SYMBOL_VARIABLES_DOMAIN
21757 @item SYMBOL_VARIABLES_DOMAIN
21758 This domain holds a subset of the @code{SYMBOLS_VAR_DOMAIN}; it
21759 contains everything minus functions and types.
21760 @findex SYMBOL_FUNCTIONS_DOMAIN
21761 @findex gdb.SYMBOL_FUNCTIONS_DOMAIN
21762 @item SYMBOL_FUNCTION_DOMAIN
21763 This domain contains all functions.
21764 @findex SYMBOL_TYPES_DOMAIN
21765 @findex gdb.SYMBOL_TYPES_DOMAIN
21766 @item SYMBOL_TYPES_DOMAIN
21767 This domain contains all types.
21770 The available address class categories in @code{gdb.Symbol} are represented
21771 as constants in the @code{gdb} module:
21774 @findex SYMBOL_LOC_UNDEF
21775 @findex gdb.SYMBOL_LOC_UNDEF
21776 @item SYMBOL_LOC_UNDEF
21777 If this is returned by address class, it indicates an error either in
21778 the symbol information or in @value{GDBN}'s handling of symbols.
21779 @findex SYMBOL_LOC_CONST
21780 @findex gdb.SYMBOL_LOC_CONST
21781 @item SYMBOL_LOC_CONST
21782 Value is constant int.
21783 @findex SYMBOL_LOC_STATIC
21784 @findex gdb.SYMBOL_LOC_STATIC
21785 @item SYMBOL_LOC_STATIC
21786 Value is at a fixed address.
21787 @findex SYMBOL_LOC_REGISTER
21788 @findex gdb.SYMBOL_LOC_REGISTER
21789 @item SYMBOL_LOC_REGISTER
21790 Value is in a register.
21791 @findex SYMBOL_LOC_ARG
21792 @findex gdb.SYMBOL_LOC_ARG
21793 @item SYMBOL_LOC_ARG
21794 Value is an argument. This value is at the offset stored within the
21795 symbol inside the frame's argument list.
21796 @findex SYMBOL_LOC_REF_ARG
21797 @findex gdb.SYMBOL_LOC_REF_ARG
21798 @item SYMBOL_LOC_REF_ARG
21799 Value address is stored in the frame's argument list. Just like
21800 @code{LOC_ARG} except that the value's address is stored at the
21801 offset, not the value itself.
21802 @findex SYMBOL_LOC_REGPARM_ADDR
21803 @findex gdb.SYMBOL_LOC_REGPARM_ADDR
21804 @item SYMBOL_LOC_REGPARM_ADDR
21805 Value is a specified register. Just like @code{LOC_REGISTER} except
21806 the register holds the address of the argument instead of the argument
21808 @findex SYMBOL_LOC_LOCAL
21809 @findex gdb.SYMBOL_LOC_LOCAL
21810 @item SYMBOL_LOC_LOCAL
21811 Value is a local variable.
21812 @findex SYMBOL_LOC_TYPEDEF
21813 @findex gdb.SYMBOL_LOC_TYPEDEF
21814 @item SYMBOL_LOC_TYPEDEF
21815 Value not used. Symbols in the domain @code{SYMBOL_STRUCT_DOMAIN} all
21817 @findex SYMBOL_LOC_BLOCK
21818 @findex gdb.SYMBOL_LOC_BLOCK
21819 @item SYMBOL_LOC_BLOCK
21821 @findex SYMBOL_LOC_CONST_BYTES
21822 @findex gdb.SYMBOL_LOC_CONST_BYTES
21823 @item SYMBOL_LOC_CONST_BYTES
21824 Value is a byte-sequence.
21825 @findex SYMBOL_LOC_UNRESOLVED
21826 @findex gdb.SYMBOL_LOC_UNRESOLVED
21827 @item SYMBOL_LOC_UNRESOLVED
21828 Value is at a fixed address, but the address of the variable has to be
21829 determined from the minimal symbol table whenever the variable is
21831 @findex SYMBOL_LOC_OPTIMIZED_OUT
21832 @findex gdb.SYMBOL_LOC_OPTIMIZED_OUT
21833 @item SYMBOL_LOC_OPTIMIZED_OUT
21834 The value does not actually exist in the program.
21835 @findex SYMBOL_LOC_COMPUTED
21836 @findex gdb.SYMBOL_LOC_COMPUTED
21837 @item SYMBOL_LOC_COMPUTED
21838 The value's address is a computed location.
21841 @node Symbol Tables In Python
21842 @subsubsection Symbol table representation in Python.
21844 @cindex symbol tables in python
21846 @tindex gdb.Symtab_and_line
21848 Access to symbol table data maintained by @value{GDBN} on the inferior
21849 is exposed to Python via two objects: @code{gdb.Symtab_and_line} and
21850 @code{gdb.Symtab}. Symbol table and line data for a frame is returned
21851 from the @code{find_sal} method in @code{gdb.Frame} object.
21852 @xref{Frames In Python}.
21854 For more information on @value{GDBN}'s symbol table management, see
21855 @ref{Symbols, ,Examining the Symbol Table}, for more information.
21857 A @code{gdb.Symtab_and_line} object has the following attributes:
21860 @defivar Symtab_and_line symtab
21861 The symbol table object (@code{gdb.Symtab}) for this frame.
21862 This attribute is not writable.
21865 @defivar Symtab_and_line pc
21866 Indicates the current program counter address. This attribute is not
21870 @defivar Symtab_and_line line
21871 Indicates the current line number for this object. This
21872 attribute is not writable.
21876 A @code{gdb.Symtab} object has the following attributes:
21879 @defivar Symtab filename
21880 The symbol table's source filename. This attribute is not writable.
21883 @defivar Symtab objfile
21884 The symbol table's backing object file. @xref{Objfiles In Python}.
21885 This attribute is not writable.
21889 The following methods are provided:
21892 @defmethod Symtab fullname
21893 Return the symbol table's source absolute file name.
21897 @node Breakpoints In Python
21898 @subsubsection Manipulating breakpoints using Python
21900 @cindex breakpoints in python
21901 @tindex gdb.Breakpoint
21903 Python code can manipulate breakpoints via the @code{gdb.Breakpoint}
21906 @defmethod Breakpoint __init__ spec @r{[}type@r{]} @r{[}wp_class@r{]}
21907 Create a new breakpoint. @var{spec} is a string naming the
21908 location of the breakpoint, or an expression that defines a
21909 watchpoint. The contents can be any location recognized by the
21910 @code{break} command, or in the case of a watchpoint, by the @code{watch}
21911 command. The optional @var{type} denotes the breakpoint to create
21912 from the types defined later in this chapter. This argument can be
21913 either: @code{BP_BREAKPOINT} or @code{BP_WATCHPOINT}. @var{type}
21914 defaults to @code{BP_BREAKPOINT}. The optional @var{wp_class}
21915 argument defines the class of watchpoint to create, if @var{type} is
21916 defined as @code{BP_WATCHPOINT}. If a watchpoint class is not
21917 provided, it is assumed to be a @var{WP_WRITE} class.
21920 The available watchpoint types represented by constants are defined in the
21925 @findex gdb.WP_READ
21927 Read only watchpoint.
21930 @findex gdb.WP_WRITE
21932 Write only watchpoint.
21935 @findex gdb.WP_ACCESS
21937 Read/Write watchpoint.
21940 @defmethod Breakpoint is_valid
21941 Return @code{True} if this @code{Breakpoint} object is valid,
21942 @code{False} otherwise. A @code{Breakpoint} object can become invalid
21943 if the user deletes the breakpoint. In this case, the object still
21944 exists, but the underlying breakpoint does not. In the cases of
21945 watchpoint scope, the watchpoint remains valid even if execution of the
21946 inferior leaves the scope of that watchpoint.
21949 @defivar Breakpoint enabled
21950 This attribute is @code{True} if the breakpoint is enabled, and
21951 @code{False} otherwise. This attribute is writable.
21954 @defivar Breakpoint silent
21955 This attribute is @code{True} if the breakpoint is silent, and
21956 @code{False} otherwise. This attribute is writable.
21958 Note that a breakpoint can also be silent if it has commands and the
21959 first command is @code{silent}. This is not reported by the
21960 @code{silent} attribute.
21963 @defivar Breakpoint thread
21964 If the breakpoint is thread-specific, this attribute holds the thread
21965 id. If the breakpoint is not thread-specific, this attribute is
21966 @code{None}. This attribute is writable.
21969 @defivar Breakpoint task
21970 If the breakpoint is Ada task-specific, this attribute holds the Ada task
21971 id. If the breakpoint is not task-specific (or the underlying
21972 language is not Ada), this attribute is @code{None}. This attribute
21976 @defivar Breakpoint ignore_count
21977 This attribute holds the ignore count for the breakpoint, an integer.
21978 This attribute is writable.
21981 @defivar Breakpoint number
21982 This attribute holds the breakpoint's number --- the identifier used by
21983 the user to manipulate the breakpoint. This attribute is not writable.
21986 @defivar Breakpoint type
21987 This attribute holds the breakpoint's type --- the identifier used to
21988 determine the actual breakpoint type or use-case. This attribute is not
21992 The available types are represented by constants defined in the @code{gdb}
21996 @findex BP_BREAKPOINT
21997 @findex gdb.BP_BREAKPOINT
21998 @item BP_BREAKPOINT
21999 Normal code breakpoint.
22001 @findex BP_WATCHPOINT
22002 @findex gdb.BP_WATCHPOINT
22003 @item BP_WATCHPOINT
22004 Watchpoint breakpoint.
22006 @findex BP_HARDWARE_WATCHPOINT
22007 @findex gdb.BP_HARDWARE_WATCHPOINT
22008 @item BP_HARDWARE_WATCHPOINT
22009 Hardware assisted watchpoint.
22011 @findex BP_READ_WATCHPOINT
22012 @findex gdb.BP_READ_WATCHPOINT
22013 @item BP_READ_WATCHPOINT
22014 Hardware assisted read watchpoint.
22016 @findex BP_ACCESS_WATCHPOINT
22017 @findex gdb.BP_ACCESS_WATCHPOINT
22018 @item BP_ACCESS_WATCHPOINT
22019 Hardware assisted access watchpoint.
22022 @defivar Breakpoint hit_count
22023 This attribute holds the hit count for the breakpoint, an integer.
22024 This attribute is writable, but currently it can only be set to zero.
22027 @defivar Breakpoint location
22028 This attribute holds the location of the breakpoint, as specified by
22029 the user. It is a string. If the breakpoint does not have a location
22030 (that is, it is a watchpoint) the attribute's value is @code{None}. This
22031 attribute is not writable.
22034 @defivar Breakpoint expression
22035 This attribute holds a breakpoint expression, as specified by
22036 the user. It is a string. If the breakpoint does not have an
22037 expression (the breakpoint is not a watchpoint) the attribute's value
22038 is @code{None}. This attribute is not writable.
22041 @defivar Breakpoint condition
22042 This attribute holds the condition of the breakpoint, as specified by
22043 the user. It is a string. If there is no condition, this attribute's
22044 value is @code{None}. This attribute is writable.
22047 @defivar Breakpoint commands
22048 This attribute holds the commands attached to the breakpoint. If
22049 there are commands, this attribute's value is a string holding all the
22050 commands, separated by newlines. If there are no commands, this
22051 attribute is @code{None}. This attribute is not writable.
22054 @node Lazy Strings In Python
22055 @subsubsection Python representation of lazy strings.
22057 @cindex lazy strings in python
22058 @tindex gdb.LazyString
22060 A @dfn{lazy string} is a string whose contents is not retrieved or
22061 encoded until it is needed.
22063 A @code{gdb.LazyString} is represented in @value{GDBN} as an
22064 @code{address} that points to a region of memory, an @code{encoding}
22065 that will be used to encode that region of memory, and a @code{length}
22066 to delimit the region of memory that represents the string. The
22067 difference between a @code{gdb.LazyString} and a string wrapped within
22068 a @code{gdb.Value} is that a @code{gdb.LazyString} will be treated
22069 differently by @value{GDBN} when printing. A @code{gdb.LazyString} is
22070 retrieved and encoded during printing, while a @code{gdb.Value}
22071 wrapping a string is immediately retrieved and encoded on creation.
22073 A @code{gdb.LazyString} object has the following functions:
22075 @defmethod LazyString value
22076 Convert the @code{gdb.LazyString} to a @code{gdb.Value}. This value
22077 will point to the string in memory, but will lose all the delayed
22078 retrieval, encoding and handling that @value{GDBN} applies to a
22079 @code{gdb.LazyString}.
22082 @defivar LazyString address
22083 This attribute holds the address of the string. This attribute is not
22087 @defivar LazyString length
22088 This attribute holds the length of the string in characters. If the
22089 length is -1, then the string will be fetched and encoded up to the
22090 first null of appropriate width. This attribute is not writable.
22093 @defivar LazyString encoding
22094 This attribute holds the encoding that will be applied to the string
22095 when the string is printed by @value{GDBN}. If the encoding is not
22096 set, or contains an empty string, then @value{GDBN} will select the
22097 most appropriate encoding when the string is printed. This attribute
22101 @defivar LazyString type
22102 This attribute holds the type that is represented by the lazy string's
22103 type. For a lazy string this will always be a pointer type. To
22104 resolve this to the lazy string's character type, use the type's
22105 @code{target} method. @xref{Types In Python}. This attribute is not
22110 @subsection Auto-loading
22111 @cindex auto-loading, Python
22113 When a new object file is read (for example, due to the @code{file}
22114 command, or because the inferior has loaded a shared library),
22115 @value{GDBN} will look for Python support scripts in several ways:
22116 @file{@var{objfile}-gdb.py} and @code{.debug_gdb_scripts} section.
22119 * objfile-gdb.py file:: The @file{@var{objfile}-gdb.py} file
22120 * .debug_gdb_scripts section:: The @code{.debug_gdb_scripts} section
22121 * Which flavor to choose?::
22124 The auto-loading feature is useful for supplying application-specific
22125 debugging commands and scripts.
22127 Auto-loading can be enabled or disabled.
22130 @kindex maint set python auto-load
22131 @item maint set python auto-load [yes|no]
22132 Enable or disable the Python auto-loading feature.
22134 @kindex maint show python auto-load
22135 @item maint show python auto-load
22136 Show whether Python auto-loading is enabled or disabled.
22139 When reading an auto-loaded file, @value{GDBN} sets the
22140 @dfn{current objfile}. This is available via the @code{gdb.current_objfile}
22141 function (@pxref{Objfiles In Python}). This can be useful for
22142 registering objfile-specific pretty-printers.
22144 @node objfile-gdb.py file
22145 @subsubsection The @file{@var{objfile}-gdb.py} file
22146 @cindex @file{@var{objfile}-gdb.py}
22148 When a new object file is read, @value{GDBN} looks for
22149 a file named @file{@var{objfile}-gdb.py},
22150 where @var{objfile} is the object file's real name, formed by ensuring
22151 that the file name is absolute, following all symlinks, and resolving
22152 @code{.} and @code{..} components. If this file exists and is
22153 readable, @value{GDBN} will evaluate it as a Python script.
22155 If this file does not exist, and if the parameter
22156 @code{debug-file-directory} is set (@pxref{Separate Debug Files}),
22157 then @value{GDBN} will look for @var{real-name} in all of the
22158 directories mentioned in the value of @code{debug-file-directory}.
22160 Finally, if this file does not exist, then @value{GDBN} will look for
22161 a file named @file{@var{data-directory}/python/auto-load/@var{real-name}}, where
22162 @var{data-directory} is @value{GDBN}'s data directory (available via
22163 @code{show data-directory}, @pxref{Data Files}), and @var{real-name}
22164 is the object file's real name, as described above.
22166 @value{GDBN} does not track which files it has already auto-loaded this way.
22167 @value{GDBN} will load the associated script every time the corresponding
22168 @var{objfile} is opened.
22169 So your @file{-gdb.py} file should be careful to avoid errors if it
22170 is evaluated more than once.
22172 @node .debug_gdb_scripts section
22173 @subsubsection The @code{.debug_gdb_scripts} section
22174 @cindex @code{.debug_gdb_scripts} section
22176 For systems using file formats like ELF and COFF,
22177 when @value{GDBN} loads a new object file
22178 it will look for a special section named @samp{.debug_gdb_scripts}.
22179 If this section exists, its contents is a list of names of scripts to load.
22181 @value{GDBN} will look for each specified script file first in the
22182 current directory and then along the source search path
22183 (@pxref{Source Path, ,Specifying Source Directories}),
22184 except that @file{$cdir} is not searched, since the compilation
22185 directory is not relevant to scripts.
22187 Entries can be placed in section @code{.debug_gdb_scripts} with,
22188 for example, this GCC macro:
22191 /* Note: The "MS" section flags are to remote duplicates. */
22192 #define DEFINE_GDB_SCRIPT(script_name) \
22194 .pushsection \".debug_gdb_scripts\", \"MS\",@@progbits,1\n\
22196 .asciz \"" script_name "\"\n\
22202 Then one can reference the macro in a header or source file like this:
22205 DEFINE_GDB_SCRIPT ("my-app-scripts.py")
22208 The script name may include directories if desired.
22210 If the macro is put in a header, any application or library
22211 using this header will get a reference to the specified script.
22213 @node Which flavor to choose?
22214 @subsubsection Which flavor to choose?
22216 Given the multiple ways of auto-loading Python scripts, it might not always
22217 be clear which one to choose. This section provides some guidance.
22219 Benefits of the @file{-gdb.py} way:
22223 Can be used with file formats that don't support multiple sections.
22226 Ease of finding scripts for public libraries.
22228 Scripts specified in the @code{.debug_gdb_scripts} section are searched for
22229 in the source search path.
22230 For publicly installed libraries, e.g., @file{libstdc++}, there typically
22231 isn't a source directory in which to find the script.
22234 Doesn't require source code additions.
22237 Benefits of the @code{.debug_gdb_scripts} way:
22241 Works with static linking.
22243 Scripts for libraries done the @file{-gdb.py} way require an objfile to
22244 trigger their loading. When an application is statically linked the only
22245 objfile available is the executable, and it is cumbersome to attach all the
22246 scripts from all the input libraries to the executable's @file{-gdb.py} script.
22249 Works with classes that are entirely inlined.
22251 Some classes can be entirely inlined, and thus there may not be an associated
22252 shared library to attach a @file{-gdb.py} script to.
22255 Scripts needn't be copied out of the source tree.
22257 In some circumstances, apps can be built out of large collections of internal
22258 libraries, and the build infrastructure necessary to install the
22259 @file{-gdb.py} scripts in a place where @value{GDBN} can find them is
22260 cumbersome. It may be easier to specify the scripts in the
22261 @code{.debug_gdb_scripts} section as relative paths, and add a path to the
22262 top of the source tree to the source search path.
22266 @chapter Command Interpreters
22267 @cindex command interpreters
22269 @value{GDBN} supports multiple command interpreters, and some command
22270 infrastructure to allow users or user interface writers to switch
22271 between interpreters or run commands in other interpreters.
22273 @value{GDBN} currently supports two command interpreters, the console
22274 interpreter (sometimes called the command-line interpreter or @sc{cli})
22275 and the machine interface interpreter (or @sc{gdb/mi}). This manual
22276 describes both of these interfaces in great detail.
22278 By default, @value{GDBN} will start with the console interpreter.
22279 However, the user may choose to start @value{GDBN} with another
22280 interpreter by specifying the @option{-i} or @option{--interpreter}
22281 startup options. Defined interpreters include:
22285 @cindex console interpreter
22286 The traditional console or command-line interpreter. This is the most often
22287 used interpreter with @value{GDBN}. With no interpreter specified at runtime,
22288 @value{GDBN} will use this interpreter.
22291 @cindex mi interpreter
22292 The newest @sc{gdb/mi} interface (currently @code{mi2}). Used primarily
22293 by programs wishing to use @value{GDBN} as a backend for a debugger GUI
22294 or an IDE. For more information, see @ref{GDB/MI, ,The @sc{gdb/mi}
22298 @cindex mi2 interpreter
22299 The current @sc{gdb/mi} interface.
22302 @cindex mi1 interpreter
22303 The @sc{gdb/mi} interface included in @value{GDBN} 5.1, 5.2, and 5.3.
22307 @cindex invoke another interpreter
22308 The interpreter being used by @value{GDBN} may not be dynamically
22309 switched at runtime. Although possible, this could lead to a very
22310 precarious situation. Consider an IDE using @sc{gdb/mi}. If a user
22311 enters the command "interpreter-set console" in a console view,
22312 @value{GDBN} would switch to using the console interpreter, rendering
22313 the IDE inoperable!
22315 @kindex interpreter-exec
22316 Although you may only choose a single interpreter at startup, you may execute
22317 commands in any interpreter from the current interpreter using the appropriate
22318 command. If you are running the console interpreter, simply use the
22319 @code{interpreter-exec} command:
22322 interpreter-exec mi "-data-list-register-names"
22325 @sc{gdb/mi} has a similar command, although it is only available in versions of
22326 @value{GDBN} which support @sc{gdb/mi} version 2 (or greater).
22329 @chapter @value{GDBN} Text User Interface
22331 @cindex Text User Interface
22334 * TUI Overview:: TUI overview
22335 * TUI Keys:: TUI key bindings
22336 * TUI Single Key Mode:: TUI single key mode
22337 * TUI Commands:: TUI-specific commands
22338 * TUI Configuration:: TUI configuration variables
22341 The @value{GDBN} Text User Interface (TUI) is a terminal
22342 interface which uses the @code{curses} library to show the source
22343 file, the assembly output, the program registers and @value{GDBN}
22344 commands in separate text windows. The TUI mode is supported only
22345 on platforms where a suitable version of the @code{curses} library
22348 @pindex @value{GDBTUI}
22349 The TUI mode is enabled by default when you invoke @value{GDBN} as
22350 either @samp{@value{GDBTUI}} or @samp{@value{GDBP} -tui}.
22351 You can also switch in and out of TUI mode while @value{GDBN} runs by
22352 using various TUI commands and key bindings, such as @kbd{C-x C-a}.
22353 @xref{TUI Keys, ,TUI Key Bindings}.
22356 @section TUI Overview
22358 In TUI mode, @value{GDBN} can display several text windows:
22362 This window is the @value{GDBN} command window with the @value{GDBN}
22363 prompt and the @value{GDBN} output. The @value{GDBN} input is still
22364 managed using readline.
22367 The source window shows the source file of the program. The current
22368 line and active breakpoints are displayed in this window.
22371 The assembly window shows the disassembly output of the program.
22374 This window shows the processor registers. Registers are highlighted
22375 when their values change.
22378 The source and assembly windows show the current program position
22379 by highlighting the current line and marking it with a @samp{>} marker.
22380 Breakpoints are indicated with two markers. The first marker
22381 indicates the breakpoint type:
22385 Breakpoint which was hit at least once.
22388 Breakpoint which was never hit.
22391 Hardware breakpoint which was hit at least once.
22394 Hardware breakpoint which was never hit.
22397 The second marker indicates whether the breakpoint is enabled or not:
22401 Breakpoint is enabled.
22404 Breakpoint is disabled.
22407 The source, assembly and register windows are updated when the current
22408 thread changes, when the frame changes, or when the program counter
22411 These windows are not all visible at the same time. The command
22412 window is always visible. The others can be arranged in several
22423 source and assembly,
22426 source and registers, or
22429 assembly and registers.
22432 A status line above the command window shows the following information:
22436 Indicates the current @value{GDBN} target.
22437 (@pxref{Targets, ,Specifying a Debugging Target}).
22440 Gives the current process or thread number.
22441 When no process is being debugged, this field is set to @code{No process}.
22444 Gives the current function name for the selected frame.
22445 The name is demangled if demangling is turned on (@pxref{Print Settings}).
22446 When there is no symbol corresponding to the current program counter,
22447 the string @code{??} is displayed.
22450 Indicates the current line number for the selected frame.
22451 When the current line number is not known, the string @code{??} is displayed.
22454 Indicates the current program counter address.
22458 @section TUI Key Bindings
22459 @cindex TUI key bindings
22461 The TUI installs several key bindings in the readline keymaps
22462 (@pxref{Command Line Editing}). The following key bindings
22463 are installed for both TUI mode and the @value{GDBN} standard mode.
22472 Enter or leave the TUI mode. When leaving the TUI mode,
22473 the curses window management stops and @value{GDBN} operates using
22474 its standard mode, writing on the terminal directly. When reentering
22475 the TUI mode, control is given back to the curses windows.
22476 The screen is then refreshed.
22480 Use a TUI layout with only one window. The layout will
22481 either be @samp{source} or @samp{assembly}. When the TUI mode
22482 is not active, it will switch to the TUI mode.
22484 Think of this key binding as the Emacs @kbd{C-x 1} binding.
22488 Use a TUI layout with at least two windows. When the current
22489 layout already has two windows, the next layout with two windows is used.
22490 When a new layout is chosen, one window will always be common to the
22491 previous layout and the new one.
22493 Think of it as the Emacs @kbd{C-x 2} binding.
22497 Change the active window. The TUI associates several key bindings
22498 (like scrolling and arrow keys) with the active window. This command
22499 gives the focus to the next TUI window.
22501 Think of it as the Emacs @kbd{C-x o} binding.
22505 Switch in and out of the TUI SingleKey mode that binds single
22506 keys to @value{GDBN} commands (@pxref{TUI Single Key Mode}).
22509 The following key bindings only work in the TUI mode:
22514 Scroll the active window one page up.
22518 Scroll the active window one page down.
22522 Scroll the active window one line up.
22526 Scroll the active window one line down.
22530 Scroll the active window one column left.
22534 Scroll the active window one column right.
22538 Refresh the screen.
22541 Because the arrow keys scroll the active window in the TUI mode, they
22542 are not available for their normal use by readline unless the command
22543 window has the focus. When another window is active, you must use
22544 other readline key bindings such as @kbd{C-p}, @kbd{C-n}, @kbd{C-b}
22545 and @kbd{C-f} to control the command window.
22547 @node TUI Single Key Mode
22548 @section TUI Single Key Mode
22549 @cindex TUI single key mode
22551 The TUI also provides a @dfn{SingleKey} mode, which binds several
22552 frequently used @value{GDBN} commands to single keys. Type @kbd{C-x s} to
22553 switch into this mode, where the following key bindings are used:
22556 @kindex c @r{(SingleKey TUI key)}
22560 @kindex d @r{(SingleKey TUI key)}
22564 @kindex f @r{(SingleKey TUI key)}
22568 @kindex n @r{(SingleKey TUI key)}
22572 @kindex q @r{(SingleKey TUI key)}
22574 exit the SingleKey mode.
22576 @kindex r @r{(SingleKey TUI key)}
22580 @kindex s @r{(SingleKey TUI key)}
22584 @kindex u @r{(SingleKey TUI key)}
22588 @kindex v @r{(SingleKey TUI key)}
22592 @kindex w @r{(SingleKey TUI key)}
22597 Other keys temporarily switch to the @value{GDBN} command prompt.
22598 The key that was pressed is inserted in the editing buffer so that
22599 it is possible to type most @value{GDBN} commands without interaction
22600 with the TUI SingleKey mode. Once the command is entered the TUI
22601 SingleKey mode is restored. The only way to permanently leave
22602 this mode is by typing @kbd{q} or @kbd{C-x s}.
22606 @section TUI-specific Commands
22607 @cindex TUI commands
22609 The TUI has specific commands to control the text windows.
22610 These commands are always available, even when @value{GDBN} is not in
22611 the TUI mode. When @value{GDBN} is in the standard mode, most
22612 of these commands will automatically switch to the TUI mode.
22614 Note that if @value{GDBN}'s @code{stdout} is not connected to a
22615 terminal, or @value{GDBN} has been started with the machine interface
22616 interpreter (@pxref{GDB/MI, ,The @sc{gdb/mi} Interface}), most of
22617 these commands will fail with an error, because it would not be
22618 possible or desirable to enable curses window management.
22623 List and give the size of all displayed windows.
22627 Display the next layout.
22630 Display the previous layout.
22633 Display the source window only.
22636 Display the assembly window only.
22639 Display the source and assembly window.
22642 Display the register window together with the source or assembly window.
22646 Make the next window active for scrolling.
22649 Make the previous window active for scrolling.
22652 Make the source window active for scrolling.
22655 Make the assembly window active for scrolling.
22658 Make the register window active for scrolling.
22661 Make the command window active for scrolling.
22665 Refresh the screen. This is similar to typing @kbd{C-L}.
22667 @item tui reg float
22669 Show the floating point registers in the register window.
22671 @item tui reg general
22672 Show the general registers in the register window.
22675 Show the next register group. The list of register groups as well as
22676 their order is target specific. The predefined register groups are the
22677 following: @code{general}, @code{float}, @code{system}, @code{vector},
22678 @code{all}, @code{save}, @code{restore}.
22680 @item tui reg system
22681 Show the system registers in the register window.
22685 Update the source window and the current execution point.
22687 @item winheight @var{name} +@var{count}
22688 @itemx winheight @var{name} -@var{count}
22690 Change the height of the window @var{name} by @var{count}
22691 lines. Positive counts increase the height, while negative counts
22694 @item tabset @var{nchars}
22696 Set the width of tab stops to be @var{nchars} characters.
22699 @node TUI Configuration
22700 @section TUI Configuration Variables
22701 @cindex TUI configuration variables
22703 Several configuration variables control the appearance of TUI windows.
22706 @item set tui border-kind @var{kind}
22707 @kindex set tui border-kind
22708 Select the border appearance for the source, assembly and register windows.
22709 The possible values are the following:
22712 Use a space character to draw the border.
22715 Use @sc{ascii} characters @samp{+}, @samp{-} and @samp{|} to draw the border.
22718 Use the Alternate Character Set to draw the border. The border is
22719 drawn using character line graphics if the terminal supports them.
22722 @item set tui border-mode @var{mode}
22723 @kindex set tui border-mode
22724 @itemx set tui active-border-mode @var{mode}
22725 @kindex set tui active-border-mode
22726 Select the display attributes for the borders of the inactive windows
22727 or the active window. The @var{mode} can be one of the following:
22730 Use normal attributes to display the border.
22736 Use reverse video mode.
22739 Use half bright mode.
22741 @item half-standout
22742 Use half bright and standout mode.
22745 Use extra bright or bold mode.
22747 @item bold-standout
22748 Use extra bright or bold and standout mode.
22753 @chapter Using @value{GDBN} under @sc{gnu} Emacs
22756 @cindex @sc{gnu} Emacs
22757 A special interface allows you to use @sc{gnu} Emacs to view (and
22758 edit) the source files for the program you are debugging with
22761 To use this interface, use the command @kbd{M-x gdb} in Emacs. Give the
22762 executable file you want to debug as an argument. This command starts
22763 @value{GDBN} as a subprocess of Emacs, with input and output through a newly
22764 created Emacs buffer.
22765 @c (Do not use the @code{-tui} option to run @value{GDBN} from Emacs.)
22767 Running @value{GDBN} under Emacs can be just like running @value{GDBN} normally except for two
22772 All ``terminal'' input and output goes through an Emacs buffer, called
22775 This applies both to @value{GDBN} commands and their output, and to the input
22776 and output done by the program you are debugging.
22778 This is useful because it means that you can copy the text of previous
22779 commands and input them again; you can even use parts of the output
22782 All the facilities of Emacs' Shell mode are available for interacting
22783 with your program. In particular, you can send signals the usual
22784 way---for example, @kbd{C-c C-c} for an interrupt, @kbd{C-c C-z} for a
22788 @value{GDBN} displays source code through Emacs.
22790 Each time @value{GDBN} displays a stack frame, Emacs automatically finds the
22791 source file for that frame and puts an arrow (@samp{=>}) at the
22792 left margin of the current line. Emacs uses a separate buffer for
22793 source display, and splits the screen to show both your @value{GDBN} session
22796 Explicit @value{GDBN} @code{list} or search commands still produce output as
22797 usual, but you probably have no reason to use them from Emacs.
22800 We call this @dfn{text command mode}. Emacs 22.1, and later, also uses
22801 a graphical mode, enabled by default, which provides further buffers
22802 that can control the execution and describe the state of your program.
22803 @xref{GDB Graphical Interface,,, Emacs, The @sc{gnu} Emacs Manual}.
22805 If you specify an absolute file name when prompted for the @kbd{M-x
22806 gdb} argument, then Emacs sets your current working directory to where
22807 your program resides. If you only specify the file name, then Emacs
22808 sets your current working directory to to the directory associated
22809 with the previous buffer. In this case, @value{GDBN} may find your
22810 program by searching your environment's @code{PATH} variable, but on
22811 some operating systems it might not find the source. So, although the
22812 @value{GDBN} input and output session proceeds normally, the auxiliary
22813 buffer does not display the current source and line of execution.
22815 The initial working directory of @value{GDBN} is printed on the top
22816 line of the GUD buffer and this serves as a default for the commands
22817 that specify files for @value{GDBN} to operate on. @xref{Files,
22818 ,Commands to Specify Files}.
22820 By default, @kbd{M-x gdb} calls the program called @file{gdb}. If you
22821 need to call @value{GDBN} by a different name (for example, if you
22822 keep several configurations around, with different names) you can
22823 customize the Emacs variable @code{gud-gdb-command-name} to run the
22826 In the GUD buffer, you can use these special Emacs commands in
22827 addition to the standard Shell mode commands:
22831 Describe the features of Emacs' GUD Mode.
22834 Execute to another source line, like the @value{GDBN} @code{step} command; also
22835 update the display window to show the current file and location.
22838 Execute to next source line in this function, skipping all function
22839 calls, like the @value{GDBN} @code{next} command. Then update the display window
22840 to show the current file and location.
22843 Execute one instruction, like the @value{GDBN} @code{stepi} command; update
22844 display window accordingly.
22847 Execute until exit from the selected stack frame, like the @value{GDBN}
22848 @code{finish} command.
22851 Continue execution of your program, like the @value{GDBN} @code{continue}
22855 Go up the number of frames indicated by the numeric argument
22856 (@pxref{Arguments, , Numeric Arguments, Emacs, The @sc{gnu} Emacs Manual}),
22857 like the @value{GDBN} @code{up} command.
22860 Go down the number of frames indicated by the numeric argument, like the
22861 @value{GDBN} @code{down} command.
22864 In any source file, the Emacs command @kbd{C-x @key{SPC}} (@code{gud-break})
22865 tells @value{GDBN} to set a breakpoint on the source line point is on.
22867 In text command mode, if you type @kbd{M-x speedbar}, Emacs displays a
22868 separate frame which shows a backtrace when the GUD buffer is current.
22869 Move point to any frame in the stack and type @key{RET} to make it
22870 become the current frame and display the associated source in the
22871 source buffer. Alternatively, click @kbd{Mouse-2} to make the
22872 selected frame become the current one. In graphical mode, the
22873 speedbar displays watch expressions.
22875 If you accidentally delete the source-display buffer, an easy way to get
22876 it back is to type the command @code{f} in the @value{GDBN} buffer, to
22877 request a frame display; when you run under Emacs, this recreates
22878 the source buffer if necessary to show you the context of the current
22881 The source files displayed in Emacs are in ordinary Emacs buffers
22882 which are visiting the source files in the usual way. You can edit
22883 the files with these buffers if you wish; but keep in mind that @value{GDBN}
22884 communicates with Emacs in terms of line numbers. If you add or
22885 delete lines from the text, the line numbers that @value{GDBN} knows cease
22886 to correspond properly with the code.
22888 A more detailed description of Emacs' interaction with @value{GDBN} is
22889 given in the Emacs manual (@pxref{Debuggers,,, Emacs, The @sc{gnu}
22892 @c The following dropped because Epoch is nonstandard. Reactivate
22893 @c if/when v19 does something similar. ---doc@cygnus.com 19dec1990
22895 @kindex Emacs Epoch environment
22899 Version 18 of @sc{gnu} Emacs has a built-in window system
22900 called the @code{epoch}
22901 environment. Users of this environment can use a new command,
22902 @code{inspect} which performs identically to @code{print} except that
22903 each value is printed in its own window.
22908 @chapter The @sc{gdb/mi} Interface
22910 @unnumberedsec Function and Purpose
22912 @cindex @sc{gdb/mi}, its purpose
22913 @sc{gdb/mi} is a line based machine oriented text interface to
22914 @value{GDBN} and is activated by specifying using the
22915 @option{--interpreter} command line option (@pxref{Mode Options}). It
22916 is specifically intended to support the development of systems which
22917 use the debugger as just one small component of a larger system.
22919 This chapter is a specification of the @sc{gdb/mi} interface. It is written
22920 in the form of a reference manual.
22922 Note that @sc{gdb/mi} is still under construction, so some of the
22923 features described below are incomplete and subject to change
22924 (@pxref{GDB/MI Development and Front Ends, , @sc{gdb/mi} Development and Front Ends}).
22926 @unnumberedsec Notation and Terminology
22928 @cindex notational conventions, for @sc{gdb/mi}
22929 This chapter uses the following notation:
22933 @code{|} separates two alternatives.
22936 @code{[ @var{something} ]} indicates that @var{something} is optional:
22937 it may or may not be given.
22940 @code{( @var{group} )*} means that @var{group} inside the parentheses
22941 may repeat zero or more times.
22944 @code{( @var{group} )+} means that @var{group} inside the parentheses
22945 may repeat one or more times.
22948 @code{"@var{string}"} means a literal @var{string}.
22952 @heading Dependencies
22956 * GDB/MI General Design::
22957 * GDB/MI Command Syntax::
22958 * GDB/MI Compatibility with CLI::
22959 * GDB/MI Development and Front Ends::
22960 * GDB/MI Output Records::
22961 * GDB/MI Simple Examples::
22962 * GDB/MI Command Description Format::
22963 * GDB/MI Breakpoint Commands::
22964 * GDB/MI Program Context::
22965 * GDB/MI Thread Commands::
22966 * GDB/MI Program Execution::
22967 * GDB/MI Stack Manipulation::
22968 * GDB/MI Variable Objects::
22969 * GDB/MI Data Manipulation::
22970 * GDB/MI Tracepoint Commands::
22971 * GDB/MI Symbol Query::
22972 * GDB/MI File Commands::
22974 * GDB/MI Kod Commands::
22975 * GDB/MI Memory Overlay Commands::
22976 * GDB/MI Signal Handling Commands::
22978 * GDB/MI Target Manipulation::
22979 * GDB/MI File Transfer Commands::
22980 * GDB/MI Miscellaneous Commands::
22983 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
22984 @node GDB/MI General Design
22985 @section @sc{gdb/mi} General Design
22986 @cindex GDB/MI General Design
22988 Interaction of a @sc{GDB/MI} frontend with @value{GDBN} involves three
22989 parts---commands sent to @value{GDBN}, responses to those commands
22990 and notifications. Each command results in exactly one response,
22991 indicating either successful completion of the command, or an error.
22992 For the commands that do not resume the target, the response contains the
22993 requested information. For the commands that resume the target, the
22994 response only indicates whether the target was successfully resumed.
22995 Notifications is the mechanism for reporting changes in the state of the
22996 target, or in @value{GDBN} state, that cannot conveniently be associated with
22997 a command and reported as part of that command response.
22999 The important examples of notifications are:
23003 Exec notifications. These are used to report changes in
23004 target state---when a target is resumed, or stopped. It would not
23005 be feasible to include this information in response of resuming
23006 commands, because one resume commands can result in multiple events in
23007 different threads. Also, quite some time may pass before any event
23008 happens in the target, while a frontend needs to know whether the resuming
23009 command itself was successfully executed.
23012 Console output, and status notifications. Console output
23013 notifications are used to report output of CLI commands, as well as
23014 diagnostics for other commands. Status notifications are used to
23015 report the progress of a long-running operation. Naturally, including
23016 this information in command response would mean no output is produced
23017 until the command is finished, which is undesirable.
23020 General notifications. Commands may have various side effects on
23021 the @value{GDBN} or target state beyond their official purpose. For example,
23022 a command may change the selected thread. Although such changes can
23023 be included in command response, using notification allows for more
23024 orthogonal frontend design.
23028 There's no guarantee that whenever an MI command reports an error,
23029 @value{GDBN} or the target are in any specific state, and especially,
23030 the state is not reverted to the state before the MI command was
23031 processed. Therefore, whenever an MI command results in an error,
23032 we recommend that the frontend refreshes all the information shown in
23033 the user interface.
23037 * Context management::
23038 * Asynchronous and non-stop modes::
23042 @node Context management
23043 @subsection Context management
23045 In most cases when @value{GDBN} accesses the target, this access is
23046 done in context of a specific thread and frame (@pxref{Frames}).
23047 Often, even when accessing global data, the target requires that a thread
23048 be specified. The CLI interface maintains the selected thread and frame,
23049 and supplies them to target on each command. This is convenient,
23050 because a command line user would not want to specify that information
23051 explicitly on each command, and because user interacts with
23052 @value{GDBN} via a single terminal, so no confusion is possible as
23053 to what thread and frame are the current ones.
23055 In the case of MI, the concept of selected thread and frame is less
23056 useful. First, a frontend can easily remember this information
23057 itself. Second, a graphical frontend can have more than one window,
23058 each one used for debugging a different thread, and the frontend might
23059 want to access additional threads for internal purposes. This
23060 increases the risk that by relying on implicitly selected thread, the
23061 frontend may be operating on a wrong one. Therefore, each MI command
23062 should explicitly specify which thread and frame to operate on. To
23063 make it possible, each MI command accepts the @samp{--thread} and
23064 @samp{--frame} options, the value to each is @value{GDBN} identifier
23065 for thread and frame to operate on.
23067 Usually, each top-level window in a frontend allows the user to select
23068 a thread and a frame, and remembers the user selection for further
23069 operations. However, in some cases @value{GDBN} may suggest that the
23070 current thread be changed. For example, when stopping on a breakpoint
23071 it is reasonable to switch to the thread where breakpoint is hit. For
23072 another example, if the user issues the CLI @samp{thread} command via
23073 the frontend, it is desirable to change the frontend's selected thread to the
23074 one specified by user. @value{GDBN} communicates the suggestion to
23075 change current thread using the @samp{=thread-selected} notification.
23076 No such notification is available for the selected frame at the moment.
23078 Note that historically, MI shares the selected thread with CLI, so
23079 frontends used the @code{-thread-select} to execute commands in the
23080 right context. However, getting this to work right is cumbersome. The
23081 simplest way is for frontend to emit @code{-thread-select} command
23082 before every command. This doubles the number of commands that need
23083 to be sent. The alternative approach is to suppress @code{-thread-select}
23084 if the selected thread in @value{GDBN} is supposed to be identical to the
23085 thread the frontend wants to operate on. However, getting this
23086 optimization right can be tricky. In particular, if the frontend
23087 sends several commands to @value{GDBN}, and one of the commands changes the
23088 selected thread, then the behaviour of subsequent commands will
23089 change. So, a frontend should either wait for response from such
23090 problematic commands, or explicitly add @code{-thread-select} for
23091 all subsequent commands. No frontend is known to do this exactly
23092 right, so it is suggested to just always pass the @samp{--thread} and
23093 @samp{--frame} options.
23095 @node Asynchronous and non-stop modes
23096 @subsection Asynchronous command execution and non-stop mode
23098 On some targets, @value{GDBN} is capable of processing MI commands
23099 even while the target is running. This is called @dfn{asynchronous
23100 command execution} (@pxref{Background Execution}). The frontend may
23101 specify a preferrence for asynchronous execution using the
23102 @code{-gdb-set target-async 1} command, which should be emitted before
23103 either running the executable or attaching to the target. After the
23104 frontend has started the executable or attached to the target, it can
23105 find if asynchronous execution is enabled using the
23106 @code{-list-target-features} command.
23108 Even if @value{GDBN} can accept a command while target is running,
23109 many commands that access the target do not work when the target is
23110 running. Therefore, asynchronous command execution is most useful
23111 when combined with non-stop mode (@pxref{Non-Stop Mode}). Then,
23112 it is possible to examine the state of one thread, while other threads
23115 When a given thread is running, MI commands that try to access the
23116 target in the context of that thread may not work, or may work only on
23117 some targets. In particular, commands that try to operate on thread's
23118 stack will not work, on any target. Commands that read memory, or
23119 modify breakpoints, may work or not work, depending on the target. Note
23120 that even commands that operate on global state, such as @code{print},
23121 @code{set}, and breakpoint commands, still access the target in the
23122 context of a specific thread, so frontend should try to find a
23123 stopped thread and perform the operation on that thread (using the
23124 @samp{--thread} option).
23126 Which commands will work in the context of a running thread is
23127 highly target dependent. However, the two commands
23128 @code{-exec-interrupt}, to stop a thread, and @code{-thread-info},
23129 to find the state of a thread, will always work.
23131 @node Thread groups
23132 @subsection Thread groups
23133 @value{GDBN} may be used to debug several processes at the same time.
23134 On some platfroms, @value{GDBN} may support debugging of several
23135 hardware systems, each one having several cores with several different
23136 processes running on each core. This section describes the MI
23137 mechanism to support such debugging scenarios.
23139 The key observation is that regardless of the structure of the
23140 target, MI can have a global list of threads, because most commands that
23141 accept the @samp{--thread} option do not need to know what process that
23142 thread belongs to. Therefore, it is not necessary to introduce
23143 neither additional @samp{--process} option, nor an notion of the
23144 current process in the MI interface. The only strictly new feature
23145 that is required is the ability to find how the threads are grouped
23148 To allow the user to discover such grouping, and to support arbitrary
23149 hierarchy of machines/cores/processes, MI introduces the concept of a
23150 @dfn{thread group}. Thread group is a collection of threads and other
23151 thread groups. A thread group always has a string identifier, a type,
23152 and may have additional attributes specific to the type. A new
23153 command, @code{-list-thread-groups}, returns the list of top-level
23154 thread groups, which correspond to processes that @value{GDBN} is
23155 debugging at the moment. By passing an identifier of a thread group
23156 to the @code{-list-thread-groups} command, it is possible to obtain
23157 the members of specific thread group.
23159 To allow the user to easily discover processes, and other objects, he
23160 wishes to debug, a concept of @dfn{available thread group} is
23161 introduced. Available thread group is an thread group that
23162 @value{GDBN} is not debugging, but that can be attached to, using the
23163 @code{-target-attach} command. The list of available top-level thread
23164 groups can be obtained using @samp{-list-thread-groups --available}.
23165 In general, the content of a thread group may be only retrieved only
23166 after attaching to that thread group.
23168 Thread groups are related to inferiors (@pxref{Inferiors and
23169 Programs}). Each inferior corresponds to a thread group of a special
23170 type @samp{process}, and some additional operations are permitted on
23171 such thread groups.
23173 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
23174 @node GDB/MI Command Syntax
23175 @section @sc{gdb/mi} Command Syntax
23178 * GDB/MI Input Syntax::
23179 * GDB/MI Output Syntax::
23182 @node GDB/MI Input Syntax
23183 @subsection @sc{gdb/mi} Input Syntax
23185 @cindex input syntax for @sc{gdb/mi}
23186 @cindex @sc{gdb/mi}, input syntax
23188 @item @var{command} @expansion{}
23189 @code{@var{cli-command} | @var{mi-command}}
23191 @item @var{cli-command} @expansion{}
23192 @code{[ @var{token} ] @var{cli-command} @var{nl}}, where
23193 @var{cli-command} is any existing @value{GDBN} CLI command.
23195 @item @var{mi-command} @expansion{}
23196 @code{[ @var{token} ] "-" @var{operation} ( " " @var{option} )*
23197 @code{[} " --" @code{]} ( " " @var{parameter} )* @var{nl}}
23199 @item @var{token} @expansion{}
23200 "any sequence of digits"
23202 @item @var{option} @expansion{}
23203 @code{"-" @var{parameter} [ " " @var{parameter} ]}
23205 @item @var{parameter} @expansion{}
23206 @code{@var{non-blank-sequence} | @var{c-string}}
23208 @item @var{operation} @expansion{}
23209 @emph{any of the operations described in this chapter}
23211 @item @var{non-blank-sequence} @expansion{}
23212 @emph{anything, provided it doesn't contain special characters such as
23213 "-", @var{nl}, """ and of course " "}
23215 @item @var{c-string} @expansion{}
23216 @code{""" @var{seven-bit-iso-c-string-content} """}
23218 @item @var{nl} @expansion{}
23227 The CLI commands are still handled by the @sc{mi} interpreter; their
23228 output is described below.
23231 The @code{@var{token}}, when present, is passed back when the command
23235 Some @sc{mi} commands accept optional arguments as part of the parameter
23236 list. Each option is identified by a leading @samp{-} (dash) and may be
23237 followed by an optional argument parameter. Options occur first in the
23238 parameter list and can be delimited from normal parameters using
23239 @samp{--} (this is useful when some parameters begin with a dash).
23246 We want easy access to the existing CLI syntax (for debugging).
23249 We want it to be easy to spot a @sc{mi} operation.
23252 @node GDB/MI Output Syntax
23253 @subsection @sc{gdb/mi} Output Syntax
23255 @cindex output syntax of @sc{gdb/mi}
23256 @cindex @sc{gdb/mi}, output syntax
23257 The output from @sc{gdb/mi} consists of zero or more out-of-band records
23258 followed, optionally, by a single result record. This result record
23259 is for the most recent command. The sequence of output records is
23260 terminated by @samp{(gdb)}.
23262 If an input command was prefixed with a @code{@var{token}} then the
23263 corresponding output for that command will also be prefixed by that same
23267 @item @var{output} @expansion{}
23268 @code{( @var{out-of-band-record} )* [ @var{result-record} ] "(gdb)" @var{nl}}
23270 @item @var{result-record} @expansion{}
23271 @code{ [ @var{token} ] "^" @var{result-class} ( "," @var{result} )* @var{nl}}
23273 @item @var{out-of-band-record} @expansion{}
23274 @code{@var{async-record} | @var{stream-record}}
23276 @item @var{async-record} @expansion{}
23277 @code{@var{exec-async-output} | @var{status-async-output} | @var{notify-async-output}}
23279 @item @var{exec-async-output} @expansion{}
23280 @code{[ @var{token} ] "*" @var{async-output}}
23282 @item @var{status-async-output} @expansion{}
23283 @code{[ @var{token} ] "+" @var{async-output}}
23285 @item @var{notify-async-output} @expansion{}
23286 @code{[ @var{token} ] "=" @var{async-output}}
23288 @item @var{async-output} @expansion{}
23289 @code{@var{async-class} ( "," @var{result} )* @var{nl}}
23291 @item @var{result-class} @expansion{}
23292 @code{"done" | "running" | "connected" | "error" | "exit"}
23294 @item @var{async-class} @expansion{}
23295 @code{"stopped" | @var{others}} (where @var{others} will be added
23296 depending on the needs---this is still in development).
23298 @item @var{result} @expansion{}
23299 @code{ @var{variable} "=" @var{value}}
23301 @item @var{variable} @expansion{}
23302 @code{ @var{string} }
23304 @item @var{value} @expansion{}
23305 @code{ @var{const} | @var{tuple} | @var{list} }
23307 @item @var{const} @expansion{}
23308 @code{@var{c-string}}
23310 @item @var{tuple} @expansion{}
23311 @code{ "@{@}" | "@{" @var{result} ( "," @var{result} )* "@}" }
23313 @item @var{list} @expansion{}
23314 @code{ "[]" | "[" @var{value} ( "," @var{value} )* "]" | "["
23315 @var{result} ( "," @var{result} )* "]" }
23317 @item @var{stream-record} @expansion{}
23318 @code{@var{console-stream-output} | @var{target-stream-output} | @var{log-stream-output}}
23320 @item @var{console-stream-output} @expansion{}
23321 @code{"~" @var{c-string}}
23323 @item @var{target-stream-output} @expansion{}
23324 @code{"@@" @var{c-string}}
23326 @item @var{log-stream-output} @expansion{}
23327 @code{"&" @var{c-string}}
23329 @item @var{nl} @expansion{}
23332 @item @var{token} @expansion{}
23333 @emph{any sequence of digits}.
23341 All output sequences end in a single line containing a period.
23344 The @code{@var{token}} is from the corresponding request. Note that
23345 for all async output, while the token is allowed by the grammar and
23346 may be output by future versions of @value{GDBN} for select async
23347 output messages, it is generally omitted. Frontends should treat
23348 all async output as reporting general changes in the state of the
23349 target and there should be no need to associate async output to any
23353 @cindex status output in @sc{gdb/mi}
23354 @var{status-async-output} contains on-going status information about the
23355 progress of a slow operation. It can be discarded. All status output is
23356 prefixed by @samp{+}.
23359 @cindex async output in @sc{gdb/mi}
23360 @var{exec-async-output} contains asynchronous state change on the target
23361 (stopped, started, disappeared). All async output is prefixed by
23365 @cindex notify output in @sc{gdb/mi}
23366 @var{notify-async-output} contains supplementary information that the
23367 client should handle (e.g., a new breakpoint information). All notify
23368 output is prefixed by @samp{=}.
23371 @cindex console output in @sc{gdb/mi}
23372 @var{console-stream-output} is output that should be displayed as is in the
23373 console. It is the textual response to a CLI command. All the console
23374 output is prefixed by @samp{~}.
23377 @cindex target output in @sc{gdb/mi}
23378 @var{target-stream-output} is the output produced by the target program.
23379 All the target output is prefixed by @samp{@@}.
23382 @cindex log output in @sc{gdb/mi}
23383 @var{log-stream-output} is output text coming from @value{GDBN}'s internals, for
23384 instance messages that should be displayed as part of an error log. All
23385 the log output is prefixed by @samp{&}.
23388 @cindex list output in @sc{gdb/mi}
23389 New @sc{gdb/mi} commands should only output @var{lists} containing
23395 @xref{GDB/MI Stream Records, , @sc{gdb/mi} Stream Records}, for more
23396 details about the various output records.
23398 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
23399 @node GDB/MI Compatibility with CLI
23400 @section @sc{gdb/mi} Compatibility with CLI
23402 @cindex compatibility, @sc{gdb/mi} and CLI
23403 @cindex @sc{gdb/mi}, compatibility with CLI
23405 For the developers convenience CLI commands can be entered directly,
23406 but there may be some unexpected behaviour. For example, commands
23407 that query the user will behave as if the user replied yes, breakpoint
23408 command lists are not executed and some CLI commands, such as
23409 @code{if}, @code{when} and @code{define}, prompt for further input with
23410 @samp{>}, which is not valid MI output.
23412 This feature may be removed at some stage in the future and it is
23413 recommended that front ends use the @code{-interpreter-exec} command
23414 (@pxref{-interpreter-exec}).
23416 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
23417 @node GDB/MI Development and Front Ends
23418 @section @sc{gdb/mi} Development and Front Ends
23419 @cindex @sc{gdb/mi} development
23421 The application which takes the MI output and presents the state of the
23422 program being debugged to the user is called a @dfn{front end}.
23424 Although @sc{gdb/mi} is still incomplete, it is currently being used
23425 by a variety of front ends to @value{GDBN}. This makes it difficult
23426 to introduce new functionality without breaking existing usage. This
23427 section tries to minimize the problems by describing how the protocol
23430 Some changes in MI need not break a carefully designed front end, and
23431 for these the MI version will remain unchanged. The following is a
23432 list of changes that may occur within one level, so front ends should
23433 parse MI output in a way that can handle them:
23437 New MI commands may be added.
23440 New fields may be added to the output of any MI command.
23443 The range of values for fields with specified values, e.g.,
23444 @code{in_scope} (@pxref{-var-update}) may be extended.
23446 @c The format of field's content e.g type prefix, may change so parse it
23447 @c at your own risk. Yes, in general?
23449 @c The order of fields may change? Shouldn't really matter but it might
23450 @c resolve inconsistencies.
23453 If the changes are likely to break front ends, the MI version level
23454 will be increased by one. This will allow the front end to parse the
23455 output according to the MI version. Apart from mi0, new versions of
23456 @value{GDBN} will not support old versions of MI and it will be the
23457 responsibility of the front end to work with the new one.
23459 @c Starting with mi3, add a new command -mi-version that prints the MI
23462 The best way to avoid unexpected changes in MI that might break your front
23463 end is to make your project known to @value{GDBN} developers and
23464 follow development on @email{gdb@@sourceware.org} and
23465 @email{gdb-patches@@sourceware.org}.
23466 @cindex mailing lists
23468 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
23469 @node GDB/MI Output Records
23470 @section @sc{gdb/mi} Output Records
23473 * GDB/MI Result Records::
23474 * GDB/MI Stream Records::
23475 * GDB/MI Async Records::
23476 * GDB/MI Frame Information::
23477 * GDB/MI Thread Information::
23480 @node GDB/MI Result Records
23481 @subsection @sc{gdb/mi} Result Records
23483 @cindex result records in @sc{gdb/mi}
23484 @cindex @sc{gdb/mi}, result records
23485 In addition to a number of out-of-band notifications, the response to a
23486 @sc{gdb/mi} command includes one of the following result indications:
23490 @item "^done" [ "," @var{results} ]
23491 The synchronous operation was successful, @code{@var{results}} are the return
23496 This result record is equivalent to @samp{^done}. Historically, it
23497 was output instead of @samp{^done} if the command has resumed the
23498 target. This behaviour is maintained for backward compatibility, but
23499 all frontends should treat @samp{^done} and @samp{^running}
23500 identically and rely on the @samp{*running} output record to determine
23501 which threads are resumed.
23505 @value{GDBN} has connected to a remote target.
23507 @item "^error" "," @var{c-string}
23509 The operation failed. The @code{@var{c-string}} contains the corresponding
23514 @value{GDBN} has terminated.
23518 @node GDB/MI Stream Records
23519 @subsection @sc{gdb/mi} Stream Records
23521 @cindex @sc{gdb/mi}, stream records
23522 @cindex stream records in @sc{gdb/mi}
23523 @value{GDBN} internally maintains a number of output streams: the console, the
23524 target, and the log. The output intended for each of these streams is
23525 funneled through the @sc{gdb/mi} interface using @dfn{stream records}.
23527 Each stream record begins with a unique @dfn{prefix character} which
23528 identifies its stream (@pxref{GDB/MI Output Syntax, , @sc{gdb/mi} Output
23529 Syntax}). In addition to the prefix, each stream record contains a
23530 @code{@var{string-output}}. This is either raw text (with an implicit new
23531 line) or a quoted C string (which does not contain an implicit newline).
23534 @item "~" @var{string-output}
23535 The console output stream contains text that should be displayed in the
23536 CLI console window. It contains the textual responses to CLI commands.
23538 @item "@@" @var{string-output}
23539 The target output stream contains any textual output from the running
23540 target. This is only present when GDB's event loop is truly
23541 asynchronous, which is currently only the case for remote targets.
23543 @item "&" @var{string-output}
23544 The log stream contains debugging messages being produced by @value{GDBN}'s
23548 @node GDB/MI Async Records
23549 @subsection @sc{gdb/mi} Async Records
23551 @cindex async records in @sc{gdb/mi}
23552 @cindex @sc{gdb/mi}, async records
23553 @dfn{Async} records are used to notify the @sc{gdb/mi} client of
23554 additional changes that have occurred. Those changes can either be a
23555 consequence of @sc{gdb/mi} commands (e.g., a breakpoint modified) or a result of
23556 target activity (e.g., target stopped).
23558 The following is the list of possible async records:
23562 @item *running,thread-id="@var{thread}"
23563 The target is now running. The @var{thread} field tells which
23564 specific thread is now running, and can be @samp{all} if all threads
23565 are running. The frontend should assume that no interaction with a
23566 running thread is possible after this notification is produced.
23567 The frontend should not assume that this notification is output
23568 only once for any command. @value{GDBN} may emit this notification
23569 several times, either for different threads, because it cannot resume
23570 all threads together, or even for a single thread, if the thread must
23571 be stepped though some code before letting it run freely.
23573 @item *stopped,reason="@var{reason}",thread-id="@var{id}",stopped-threads="@var{stopped}",core="@var{core}"
23574 The target has stopped. The @var{reason} field can have one of the
23578 @item breakpoint-hit
23579 A breakpoint was reached.
23580 @item watchpoint-trigger
23581 A watchpoint was triggered.
23582 @item read-watchpoint-trigger
23583 A read watchpoint was triggered.
23584 @item access-watchpoint-trigger
23585 An access watchpoint was triggered.
23586 @item function-finished
23587 An -exec-finish or similar CLI command was accomplished.
23588 @item location-reached
23589 An -exec-until or similar CLI command was accomplished.
23590 @item watchpoint-scope
23591 A watchpoint has gone out of scope.
23592 @item end-stepping-range
23593 An -exec-next, -exec-next-instruction, -exec-step, -exec-step-instruction or
23594 similar CLI command was accomplished.
23595 @item exited-signalled
23596 The inferior exited because of a signal.
23598 The inferior exited.
23599 @item exited-normally
23600 The inferior exited normally.
23601 @item signal-received
23602 A signal was received by the inferior.
23605 The @var{id} field identifies the thread that directly caused the stop
23606 -- for example by hitting a breakpoint. Depending on whether all-stop
23607 mode is in effect (@pxref{All-Stop Mode}), @value{GDBN} may either
23608 stop all threads, or only the thread that directly triggered the stop.
23609 If all threads are stopped, the @var{stopped} field will have the
23610 value of @code{"all"}. Otherwise, the value of the @var{stopped}
23611 field will be a list of thread identifiers. Presently, this list will
23612 always include a single thread, but frontend should be prepared to see
23613 several threads in the list. The @var{core} field reports the
23614 processor core on which the stop event has happened. This field may be absent
23615 if such information is not available.
23617 @item =thread-group-added,id="@var{id}"
23618 @itemx =thread-group-removed,id="@var{id}"
23619 A thread group was either added or removed. The @var{id} field
23620 contains the @value{GDBN} identifier of the thread group. When a thread
23621 group is added, it generally might not be associated with a running
23622 process. When a thread group is removed, its id becomes invalid and
23623 cannot be used in any way.
23625 @item =thread-group-started,id="@var{id}",pid="@var{pid}"
23626 A thread group became associated with a running program,
23627 either because the program was just started or the thread group
23628 was attached to a program. The @var{id} field contains the
23629 @value{GDBN} identifier of the thread group. The @var{pid} field
23630 contains process identifier, specific to the operating system.
23632 @itemx =thread-group-exited,id="@var{id}"
23633 A thread group is no longer associated with a running program,
23634 either because the program has exited, or because it was detached
23635 from. The @var{id} field contains the @value{GDBN} identifier of the
23638 @item =thread-created,id="@var{id}",group-id="@var{gid}"
23639 @itemx =thread-exited,id="@var{id}",group-id="@var{gid}"
23640 A thread either was created, or has exited. The @var{id} field
23641 contains the @value{GDBN} identifier of the thread. The @var{gid}
23642 field identifies the thread group this thread belongs to.
23644 @item =thread-selected,id="@var{id}"
23645 Informs that the selected thread was changed as result of the last
23646 command. This notification is not emitted as result of @code{-thread-select}
23647 command but is emitted whenever an MI command that is not documented
23648 to change the selected thread actually changes it. In particular,
23649 invoking, directly or indirectly (via user-defined command), the CLI
23650 @code{thread} command, will generate this notification.
23652 We suggest that in response to this notification, front ends
23653 highlight the selected thread and cause subsequent commands to apply to
23656 @item =library-loaded,...
23657 Reports that a new library file was loaded by the program. This
23658 notification has 4 fields---@var{id}, @var{target-name},
23659 @var{host-name}, and @var{symbols-loaded}. The @var{id} field is an
23660 opaque identifier of the library. For remote debugging case,
23661 @var{target-name} and @var{host-name} fields give the name of the
23662 library file on the target, and on the host respectively. For native
23663 debugging, both those fields have the same value. The
23664 @var{symbols-loaded} field reports if the debug symbols for this
23665 library are loaded. The @var{thread-group} field, if present,
23666 specifies the id of the thread group in whose context the library was loaded.
23667 If the field is absent, it means the library was loaded in the context
23668 of all present thread groups.
23670 @item =library-unloaded,...
23671 Reports that a library was unloaded by the program. This notification
23672 has 3 fields---@var{id}, @var{target-name} and @var{host-name} with
23673 the same meaning as for the @code{=library-loaded} notification.
23674 The @var{thread-group} field, if present, specifies the id of the
23675 thread group in whose context the library was unloaded. If the field is
23676 absent, it means the library was unloaded in the context of all present
23681 @node GDB/MI Frame Information
23682 @subsection @sc{gdb/mi} Frame Information
23684 Response from many MI commands includes an information about stack
23685 frame. This information is a tuple that may have the following
23690 The level of the stack frame. The innermost frame has the level of
23691 zero. This field is always present.
23694 The name of the function corresponding to the frame. This field may
23695 be absent if @value{GDBN} is unable to determine the function name.
23698 The code address for the frame. This field is always present.
23701 The name of the source files that correspond to the frame's code
23702 address. This field may be absent.
23705 The source line corresponding to the frames' code address. This field
23709 The name of the binary file (either executable or shared library) the
23710 corresponds to the frame's code address. This field may be absent.
23714 @node GDB/MI Thread Information
23715 @subsection @sc{gdb/mi} Thread Information
23717 Whenever @value{GDBN} has to report an information about a thread, it
23718 uses a tuple with the following fields:
23722 The numeric id assigned to the thread by @value{GDBN}. This field is
23726 Target-specific string identifying the thread. This field is always present.
23729 Additional information about the thread provided by the target.
23730 It is supposed to be human-readable and not interpreted by the
23731 frontend. This field is optional.
23734 Either @samp{stopped} or @samp{running}, depending on whether the
23735 thread is presently running. This field is always present.
23738 The value of this field is an integer number of the processor core the
23739 thread was last seen on. This field is optional.
23743 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
23744 @node GDB/MI Simple Examples
23745 @section Simple Examples of @sc{gdb/mi} Interaction
23746 @cindex @sc{gdb/mi}, simple examples
23748 This subsection presents several simple examples of interaction using
23749 the @sc{gdb/mi} interface. In these examples, @samp{->} means that the
23750 following line is passed to @sc{gdb/mi} as input, while @samp{<-} means
23751 the output received from @sc{gdb/mi}.
23753 Note the line breaks shown in the examples are here only for
23754 readability, they don't appear in the real output.
23756 @subheading Setting a Breakpoint
23758 Setting a breakpoint generates synchronous output which contains detailed
23759 information of the breakpoint.
23762 -> -break-insert main
23763 <- ^done,bkpt=@{number="1",type="breakpoint",disp="keep",
23764 enabled="y",addr="0x08048564",func="main",file="myprog.c",
23765 fullname="/home/nickrob/myprog.c",line="68",times="0"@}
23769 @subheading Program Execution
23771 Program execution generates asynchronous records and MI gives the
23772 reason that execution stopped.
23778 <- *stopped,reason="breakpoint-hit",disp="keep",bkptno="1",thread-id="0",
23779 frame=@{addr="0x08048564",func="main",
23780 args=[@{name="argc",value="1"@},@{name="argv",value="0xbfc4d4d4"@}],
23781 file="myprog.c",fullname="/home/nickrob/myprog.c",line="68"@}
23786 <- *stopped,reason="exited-normally"
23790 @subheading Quitting @value{GDBN}
23792 Quitting @value{GDBN} just prints the result class @samp{^exit}.
23800 Please note that @samp{^exit} is printed immediately, but it might
23801 take some time for @value{GDBN} to actually exit. During that time, @value{GDBN}
23802 performs necessary cleanups, including killing programs being debugged
23803 or disconnecting from debug hardware, so the frontend should wait till
23804 @value{GDBN} exits and should only forcibly kill @value{GDBN} if it
23805 fails to exit in reasonable time.
23807 @subheading A Bad Command
23809 Here's what happens if you pass a non-existent command:
23813 <- ^error,msg="Undefined MI command: rubbish"
23818 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
23819 @node GDB/MI Command Description Format
23820 @section @sc{gdb/mi} Command Description Format
23822 The remaining sections describe blocks of commands. Each block of
23823 commands is laid out in a fashion similar to this section.
23825 @subheading Motivation
23827 The motivation for this collection of commands.
23829 @subheading Introduction
23831 A brief introduction to this collection of commands as a whole.
23833 @subheading Commands
23835 For each command in the block, the following is described:
23837 @subsubheading Synopsis
23840 -command @var{args}@dots{}
23843 @subsubheading Result
23845 @subsubheading @value{GDBN} Command
23847 The corresponding @value{GDBN} CLI command(s), if any.
23849 @subsubheading Example
23851 Example(s) formatted for readability. Some of the described commands have
23852 not been implemented yet and these are labeled N.A.@: (not available).
23855 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
23856 @node GDB/MI Breakpoint Commands
23857 @section @sc{gdb/mi} Breakpoint Commands
23859 @cindex breakpoint commands for @sc{gdb/mi}
23860 @cindex @sc{gdb/mi}, breakpoint commands
23861 This section documents @sc{gdb/mi} commands for manipulating
23864 @subheading The @code{-break-after} Command
23865 @findex -break-after
23867 @subsubheading Synopsis
23870 -break-after @var{number} @var{count}
23873 The breakpoint number @var{number} is not in effect until it has been
23874 hit @var{count} times. To see how this is reflected in the output of
23875 the @samp{-break-list} command, see the description of the
23876 @samp{-break-list} command below.
23878 @subsubheading @value{GDBN} Command
23880 The corresponding @value{GDBN} command is @samp{ignore}.
23882 @subsubheading Example
23887 ^done,bkpt=@{number="1",type="breakpoint",disp="keep",
23888 enabled="y",addr="0x000100d0",func="main",file="hello.c",
23889 fullname="/home/foo/hello.c",line="5",times="0"@}
23896 ^done,BreakpointTable=@{nr_rows="1",nr_cols="6",
23897 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
23898 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
23899 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
23900 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
23901 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
23902 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
23903 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
23904 addr="0x000100d0",func="main",file="hello.c",fullname="/home/foo/hello.c",
23905 line="5",times="0",ignore="3"@}]@}
23910 @subheading The @code{-break-catch} Command
23911 @findex -break-catch
23914 @subheading The @code{-break-commands} Command
23915 @findex -break-commands
23917 @subsubheading Synopsis
23920 -break-commands @var{number} [ @var{command1} ... @var{commandN} ]
23923 Specifies the CLI commands that should be executed when breakpoint
23924 @var{number} is hit. The parameters @var{command1} to @var{commandN}
23925 are the commands. If no command is specified, any previously-set
23926 commands are cleared. @xref{Break Commands}. Typical use of this
23927 functionality is tracing a program, that is, printing of values of
23928 some variables whenever breakpoint is hit and then continuing.
23930 @subsubheading @value{GDBN} Command
23932 The corresponding @value{GDBN} command is @samp{commands}.
23934 @subsubheading Example
23939 ^done,bkpt=@{number="1",type="breakpoint",disp="keep",
23940 enabled="y",addr="0x000100d0",func="main",file="hello.c",
23941 fullname="/home/foo/hello.c",line="5",times="0"@}
23943 -break-commands 1 "print v" "continue"
23948 @subheading The @code{-break-condition} Command
23949 @findex -break-condition
23951 @subsubheading Synopsis
23954 -break-condition @var{number} @var{expr}
23957 Breakpoint @var{number} will stop the program only if the condition in
23958 @var{expr} is true. The condition becomes part of the
23959 @samp{-break-list} output (see the description of the @samp{-break-list}
23962 @subsubheading @value{GDBN} Command
23964 The corresponding @value{GDBN} command is @samp{condition}.
23966 @subsubheading Example
23970 -break-condition 1 1
23974 ^done,BreakpointTable=@{nr_rows="1",nr_cols="6",
23975 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
23976 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
23977 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
23978 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
23979 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
23980 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
23981 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
23982 addr="0x000100d0",func="main",file="hello.c",fullname="/home/foo/hello.c",
23983 line="5",cond="1",times="0",ignore="3"@}]@}
23987 @subheading The @code{-break-delete} Command
23988 @findex -break-delete
23990 @subsubheading Synopsis
23993 -break-delete ( @var{breakpoint} )+
23996 Delete the breakpoint(s) whose number(s) are specified in the argument
23997 list. This is obviously reflected in the breakpoint list.
23999 @subsubheading @value{GDBN} Command
24001 The corresponding @value{GDBN} command is @samp{delete}.
24003 @subsubheading Example
24011 ^done,BreakpointTable=@{nr_rows="0",nr_cols="6",
24012 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
24013 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
24014 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
24015 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
24016 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
24017 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
24022 @subheading The @code{-break-disable} Command
24023 @findex -break-disable
24025 @subsubheading Synopsis
24028 -break-disable ( @var{breakpoint} )+
24031 Disable the named @var{breakpoint}(s). The field @samp{enabled} in the
24032 break list is now set to @samp{n} for the named @var{breakpoint}(s).
24034 @subsubheading @value{GDBN} Command
24036 The corresponding @value{GDBN} command is @samp{disable}.
24038 @subsubheading Example
24046 ^done,BreakpointTable=@{nr_rows="1",nr_cols="6",
24047 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
24048 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
24049 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
24050 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
24051 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
24052 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
24053 body=[bkpt=@{number="2",type="breakpoint",disp="keep",enabled="n",
24054 addr="0x000100d0",func="main",file="hello.c",fullname="/home/foo/hello.c",
24055 line="5",times="0"@}]@}
24059 @subheading The @code{-break-enable} Command
24060 @findex -break-enable
24062 @subsubheading Synopsis
24065 -break-enable ( @var{breakpoint} )+
24068 Enable (previously disabled) @var{breakpoint}(s).
24070 @subsubheading @value{GDBN} Command
24072 The corresponding @value{GDBN} command is @samp{enable}.
24074 @subsubheading Example
24082 ^done,BreakpointTable=@{nr_rows="1",nr_cols="6",
24083 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
24084 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
24085 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
24086 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
24087 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
24088 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
24089 body=[bkpt=@{number="2",type="breakpoint",disp="keep",enabled="y",
24090 addr="0x000100d0",func="main",file="hello.c",fullname="/home/foo/hello.c",
24091 line="5",times="0"@}]@}
24095 @subheading The @code{-break-info} Command
24096 @findex -break-info
24098 @subsubheading Synopsis
24101 -break-info @var{breakpoint}
24105 Get information about a single breakpoint.
24107 @subsubheading @value{GDBN} Command
24109 The corresponding @value{GDBN} command is @samp{info break @var{breakpoint}}.
24111 @subsubheading Example
24114 @subheading The @code{-break-insert} Command
24115 @findex -break-insert
24117 @subsubheading Synopsis
24120 -break-insert [ -t ] [ -h ] [ -f ] [ -d ] [ -a ]
24121 [ -c @var{condition} ] [ -i @var{ignore-count} ]
24122 [ -p @var{thread} ] [ @var{location} ]
24126 If specified, @var{location}, can be one of:
24133 @item filename:linenum
24134 @item filename:function
24138 The possible optional parameters of this command are:
24142 Insert a temporary breakpoint.
24144 Insert a hardware breakpoint.
24145 @item -c @var{condition}
24146 Make the breakpoint conditional on @var{condition}.
24147 @item -i @var{ignore-count}
24148 Initialize the @var{ignore-count}.
24150 If @var{location} cannot be parsed (for example if it
24151 refers to unknown files or functions), create a pending
24152 breakpoint. Without this flag, @value{GDBN} will report
24153 an error, and won't create a breakpoint, if @var{location}
24156 Create a disabled breakpoint.
24158 Create a tracepoint. @xref{Tracepoints}. When this parameter
24159 is used together with @samp{-h}, a fast tracepoint is created.
24162 @subsubheading Result
24164 The result is in the form:
24167 ^done,bkpt=@{number="@var{number}",type="@var{type}",disp="del"|"keep",
24168 enabled="y"|"n",addr="@var{hex}",func="@var{funcname}",file="@var{filename}",
24169 fullname="@var{full_filename}",line="@var{lineno}",[thread="@var{threadno},]
24170 times="@var{times}"@}
24174 where @var{number} is the @value{GDBN} number for this breakpoint,
24175 @var{funcname} is the name of the function where the breakpoint was
24176 inserted, @var{filename} is the name of the source file which contains
24177 this function, @var{lineno} is the source line number within that file
24178 and @var{times} the number of times that the breakpoint has been hit
24179 (always 0 for -break-insert but may be greater for -break-info or -break-list
24180 which use the same output).
24182 Note: this format is open to change.
24183 @c An out-of-band breakpoint instead of part of the result?
24185 @subsubheading @value{GDBN} Command
24187 The corresponding @value{GDBN} commands are @samp{break}, @samp{tbreak},
24188 @samp{hbreak}, @samp{thbreak}, and @samp{rbreak}.
24190 @subsubheading Example
24195 ^done,bkpt=@{number="1",addr="0x0001072c",file="recursive2.c",
24196 fullname="/home/foo/recursive2.c,line="4",times="0"@}
24198 -break-insert -t foo
24199 ^done,bkpt=@{number="2",addr="0x00010774",file="recursive2.c",
24200 fullname="/home/foo/recursive2.c,line="11",times="0"@}
24203 ^done,BreakpointTable=@{nr_rows="2",nr_cols="6",
24204 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
24205 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
24206 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
24207 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
24208 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
24209 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
24210 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
24211 addr="0x0001072c", func="main",file="recursive2.c",
24212 fullname="/home/foo/recursive2.c,"line="4",times="0"@},
24213 bkpt=@{number="2",type="breakpoint",disp="del",enabled="y",
24214 addr="0x00010774",func="foo",file="recursive2.c",
24215 fullname="/home/foo/recursive2.c",line="11",times="0"@}]@}
24217 -break-insert -r foo.*
24218 ~int foo(int, int);
24219 ^done,bkpt=@{number="3",addr="0x00010774",file="recursive2.c,
24220 "fullname="/home/foo/recursive2.c",line="11",times="0"@}
24224 @subheading The @code{-break-list} Command
24225 @findex -break-list
24227 @subsubheading Synopsis
24233 Displays the list of inserted breakpoints, showing the following fields:
24237 number of the breakpoint
24239 type of the breakpoint: @samp{breakpoint} or @samp{watchpoint}
24241 should the breakpoint be deleted or disabled when it is hit: @samp{keep}
24244 is the breakpoint enabled or no: @samp{y} or @samp{n}
24246 memory location at which the breakpoint is set
24248 logical location of the breakpoint, expressed by function name, file
24251 number of times the breakpoint has been hit
24254 If there are no breakpoints or watchpoints, the @code{BreakpointTable}
24255 @code{body} field is an empty list.
24257 @subsubheading @value{GDBN} Command
24259 The corresponding @value{GDBN} command is @samp{info break}.
24261 @subsubheading Example
24266 ^done,BreakpointTable=@{nr_rows="2",nr_cols="6",
24267 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
24268 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
24269 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
24270 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
24271 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
24272 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
24273 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
24274 addr="0x000100d0",func="main",file="hello.c",line="5",times="0"@},
24275 bkpt=@{number="2",type="breakpoint",disp="keep",enabled="y",
24276 addr="0x00010114",func="foo",file="hello.c",fullname="/home/foo/hello.c",
24277 line="13",times="0"@}]@}
24281 Here's an example of the result when there are no breakpoints:
24286 ^done,BreakpointTable=@{nr_rows="0",nr_cols="6",
24287 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
24288 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
24289 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
24290 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
24291 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
24292 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
24297 @subheading The @code{-break-passcount} Command
24298 @findex -break-passcount
24300 @subsubheading Synopsis
24303 -break-passcount @var{tracepoint-number} @var{passcount}
24306 Set the passcount for tracepoint @var{tracepoint-number} to
24307 @var{passcount}. If the breakpoint referred to by @var{tracepoint-number}
24308 is not a tracepoint, error is emitted. This corresponds to CLI
24309 command @samp{passcount}.
24311 @subheading The @code{-break-watch} Command
24312 @findex -break-watch
24314 @subsubheading Synopsis
24317 -break-watch [ -a | -r ]
24320 Create a watchpoint. With the @samp{-a} option it will create an
24321 @dfn{access} watchpoint, i.e., a watchpoint that triggers either on a
24322 read from or on a write to the memory location. With the @samp{-r}
24323 option, the watchpoint created is a @dfn{read} watchpoint, i.e., it will
24324 trigger only when the memory location is accessed for reading. Without
24325 either of the options, the watchpoint created is a regular watchpoint,
24326 i.e., it will trigger when the memory location is accessed for writing.
24327 @xref{Set Watchpoints, , Setting Watchpoints}.
24329 Note that @samp{-break-list} will report a single list of watchpoints and
24330 breakpoints inserted.
24332 @subsubheading @value{GDBN} Command
24334 The corresponding @value{GDBN} commands are @samp{watch}, @samp{awatch}, and
24337 @subsubheading Example
24339 Setting a watchpoint on a variable in the @code{main} function:
24344 ^done,wpt=@{number="2",exp="x"@}
24349 *stopped,reason="watchpoint-trigger",wpt=@{number="2",exp="x"@},
24350 value=@{old="-268439212",new="55"@},
24351 frame=@{func="main",args=[],file="recursive2.c",
24352 fullname="/home/foo/bar/recursive2.c",line="5"@}
24356 Setting a watchpoint on a variable local to a function. @value{GDBN} will stop
24357 the program execution twice: first for the variable changing value, then
24358 for the watchpoint going out of scope.
24363 ^done,wpt=@{number="5",exp="C"@}
24368 *stopped,reason="watchpoint-trigger",
24369 wpt=@{number="5",exp="C"@},value=@{old="-276895068",new="3"@},
24370 frame=@{func="callee4",args=[],
24371 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
24372 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="13"@}
24377 *stopped,reason="watchpoint-scope",wpnum="5",
24378 frame=@{func="callee3",args=[@{name="strarg",
24379 value="0x11940 \"A string argument.\""@}],
24380 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
24381 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="18"@}
24385 Listing breakpoints and watchpoints, at different points in the program
24386 execution. Note that once the watchpoint goes out of scope, it is
24392 ^done,wpt=@{number="2",exp="C"@}
24395 ^done,BreakpointTable=@{nr_rows="2",nr_cols="6",
24396 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
24397 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
24398 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
24399 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
24400 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
24401 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
24402 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
24403 addr="0x00010734",func="callee4",
24404 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
24405 fullname="/home/foo/devo/gdb/testsuite/gdb.mi/basics.c"line="8",times="1"@},
24406 bkpt=@{number="2",type="watchpoint",disp="keep",
24407 enabled="y",addr="",what="C",times="0"@}]@}
24412 *stopped,reason="watchpoint-trigger",wpt=@{number="2",exp="C"@},
24413 value=@{old="-276895068",new="3"@},
24414 frame=@{func="callee4",args=[],
24415 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
24416 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="13"@}
24419 ^done,BreakpointTable=@{nr_rows="2",nr_cols="6",
24420 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
24421 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
24422 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
24423 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
24424 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
24425 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
24426 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
24427 addr="0x00010734",func="callee4",
24428 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
24429 fullname="/home/foo/devo/gdb/testsuite/gdb.mi/basics.c",line="8",times="1"@},
24430 bkpt=@{number="2",type="watchpoint",disp="keep",
24431 enabled="y",addr="",what="C",times="-5"@}]@}
24435 ^done,reason="watchpoint-scope",wpnum="2",
24436 frame=@{func="callee3",args=[@{name="strarg",
24437 value="0x11940 \"A string argument.\""@}],
24438 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
24439 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="18"@}
24442 ^done,BreakpointTable=@{nr_rows="1",nr_cols="6",
24443 hdr=[@{width="3",alignment="-1",col_name="number",colhdr="Num"@},
24444 @{width="14",alignment="-1",col_name="type",colhdr="Type"@},
24445 @{width="4",alignment="-1",col_name="disp",colhdr="Disp"@},
24446 @{width="3",alignment="-1",col_name="enabled",colhdr="Enb"@},
24447 @{width="10",alignment="-1",col_name="addr",colhdr="Address"@},
24448 @{width="40",alignment="2",col_name="what",colhdr="What"@}],
24449 body=[bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
24450 addr="0x00010734",func="callee4",
24451 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
24452 fullname="/home/foo/devo/gdb/testsuite/gdb.mi/basics.c",line="8",
24457 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
24458 @node GDB/MI Program Context
24459 @section @sc{gdb/mi} Program Context
24461 @subheading The @code{-exec-arguments} Command
24462 @findex -exec-arguments
24465 @subsubheading Synopsis
24468 -exec-arguments @var{args}
24471 Set the inferior program arguments, to be used in the next
24474 @subsubheading @value{GDBN} Command
24476 The corresponding @value{GDBN} command is @samp{set args}.
24478 @subsubheading Example
24482 -exec-arguments -v word
24489 @subheading The @code{-exec-show-arguments} Command
24490 @findex -exec-show-arguments
24492 @subsubheading Synopsis
24495 -exec-show-arguments
24498 Print the arguments of the program.
24500 @subsubheading @value{GDBN} Command
24502 The corresponding @value{GDBN} command is @samp{show args}.
24504 @subsubheading Example
24509 @subheading The @code{-environment-cd} Command
24510 @findex -environment-cd
24512 @subsubheading Synopsis
24515 -environment-cd @var{pathdir}
24518 Set @value{GDBN}'s working directory.
24520 @subsubheading @value{GDBN} Command
24522 The corresponding @value{GDBN} command is @samp{cd}.
24524 @subsubheading Example
24528 -environment-cd /kwikemart/marge/ezannoni/flathead-dev/devo/gdb
24534 @subheading The @code{-environment-directory} Command
24535 @findex -environment-directory
24537 @subsubheading Synopsis
24540 -environment-directory [ -r ] [ @var{pathdir} ]+
24543 Add directories @var{pathdir} to beginning of search path for source files.
24544 If the @samp{-r} option is used, the search path is reset to the default
24545 search path. If directories @var{pathdir} are supplied in addition to the
24546 @samp{-r} option, the search path is first reset and then addition
24548 Multiple directories may be specified, separated by blanks. Specifying
24549 multiple directories in a single command
24550 results in the directories added to the beginning of the
24551 search path in the same order they were presented in the command.
24552 If blanks are needed as
24553 part of a directory name, double-quotes should be used around
24554 the name. In the command output, the path will show up separated
24555 by the system directory-separator character. The directory-separator
24556 character must not be used
24557 in any directory name.
24558 If no directories are specified, the current search path is displayed.
24560 @subsubheading @value{GDBN} Command
24562 The corresponding @value{GDBN} command is @samp{dir}.
24564 @subsubheading Example
24568 -environment-directory /kwikemart/marge/ezannoni/flathead-dev/devo/gdb
24569 ^done,source-path="/kwikemart/marge/ezannoni/flathead-dev/devo/gdb:$cdir:$cwd"
24571 -environment-directory ""
24572 ^done,source-path="/kwikemart/marge/ezannoni/flathead-dev/devo/gdb:$cdir:$cwd"
24574 -environment-directory -r /home/jjohnstn/src/gdb /usr/src
24575 ^done,source-path="/home/jjohnstn/src/gdb:/usr/src:$cdir:$cwd"
24577 -environment-directory -r
24578 ^done,source-path="$cdir:$cwd"
24583 @subheading The @code{-environment-path} Command
24584 @findex -environment-path
24586 @subsubheading Synopsis
24589 -environment-path [ -r ] [ @var{pathdir} ]+
24592 Add directories @var{pathdir} to beginning of search path for object files.
24593 If the @samp{-r} option is used, the search path is reset to the original
24594 search path that existed at gdb start-up. If directories @var{pathdir} are
24595 supplied in addition to the
24596 @samp{-r} option, the search path is first reset and then addition
24598 Multiple directories may be specified, separated by blanks. Specifying
24599 multiple directories in a single command
24600 results in the directories added to the beginning of the
24601 search path in the same order they were presented in the command.
24602 If blanks are needed as
24603 part of a directory name, double-quotes should be used around
24604 the name. In the command output, the path will show up separated
24605 by the system directory-separator character. The directory-separator
24606 character must not be used
24607 in any directory name.
24608 If no directories are specified, the current path is displayed.
24611 @subsubheading @value{GDBN} Command
24613 The corresponding @value{GDBN} command is @samp{path}.
24615 @subsubheading Example
24620 ^done,path="/usr/bin"
24622 -environment-path /kwikemart/marge/ezannoni/flathead-dev/ppc-eabi/gdb /bin
24623 ^done,path="/kwikemart/marge/ezannoni/flathead-dev/ppc-eabi/gdb:/bin:/usr/bin"
24625 -environment-path -r /usr/local/bin
24626 ^done,path="/usr/local/bin:/usr/bin"
24631 @subheading The @code{-environment-pwd} Command
24632 @findex -environment-pwd
24634 @subsubheading Synopsis
24640 Show the current working directory.
24642 @subsubheading @value{GDBN} Command
24644 The corresponding @value{GDBN} command is @samp{pwd}.
24646 @subsubheading Example
24651 ^done,cwd="/kwikemart/marge/ezannoni/flathead-dev/devo/gdb"
24655 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
24656 @node GDB/MI Thread Commands
24657 @section @sc{gdb/mi} Thread Commands
24660 @subheading The @code{-thread-info} Command
24661 @findex -thread-info
24663 @subsubheading Synopsis
24666 -thread-info [ @var{thread-id} ]
24669 Reports information about either a specific thread, if
24670 the @var{thread-id} parameter is present, or about all
24671 threads. When printing information about all threads,
24672 also reports the current thread.
24674 @subsubheading @value{GDBN} Command
24676 The @samp{info thread} command prints the same information
24679 @subsubheading Example
24684 @{id="2",target-id="Thread 0xb7e14b90 (LWP 21257)",
24685 frame=@{level="0",addr="0xffffe410",func="__kernel_vsyscall",args=[]@},state="running"@},
24686 @{id="1",target-id="Thread 0xb7e156b0 (LWP 21254)",
24687 frame=@{level="0",addr="0x0804891f",func="foo",args=[@{name="i",value="10"@}],
24688 file="/tmp/a.c",fullname="/tmp/a.c",line="158"@},state="running"@}],
24689 current-thread-id="1"
24693 The @samp{state} field may have the following values:
24697 The thread is stopped. Frame information is available for stopped
24701 The thread is running. There's no frame information for running
24706 @subheading The @code{-thread-list-ids} Command
24707 @findex -thread-list-ids
24709 @subsubheading Synopsis
24715 Produces a list of the currently known @value{GDBN} thread ids. At the
24716 end of the list it also prints the total number of such threads.
24718 This command is retained for historical reasons, the
24719 @code{-thread-info} command should be used instead.
24721 @subsubheading @value{GDBN} Command
24723 Part of @samp{info threads} supplies the same information.
24725 @subsubheading Example
24730 ^done,thread-ids=@{thread-id="3",thread-id="2",thread-id="1"@},
24731 current-thread-id="1",number-of-threads="3"
24736 @subheading The @code{-thread-select} Command
24737 @findex -thread-select
24739 @subsubheading Synopsis
24742 -thread-select @var{threadnum}
24745 Make @var{threadnum} the current thread. It prints the number of the new
24746 current thread, and the topmost frame for that thread.
24748 This command is deprecated in favor of explicitly using the
24749 @samp{--thread} option to each command.
24751 @subsubheading @value{GDBN} Command
24753 The corresponding @value{GDBN} command is @samp{thread}.
24755 @subsubheading Example
24762 *stopped,reason="end-stepping-range",thread-id="2",line="187",
24763 file="../../../devo/gdb/testsuite/gdb.threads/linux-dp.c"
24767 thread-ids=@{thread-id="3",thread-id="2",thread-id="1"@},
24768 number-of-threads="3"
24771 ^done,new-thread-id="3",
24772 frame=@{level="0",func="vprintf",
24773 args=[@{name="format",value="0x8048e9c \"%*s%c %d %c\\n\""@},
24774 @{name="arg",value="0x2"@}],file="vprintf.c",line="31"@}
24778 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
24779 @node GDB/MI Program Execution
24780 @section @sc{gdb/mi} Program Execution
24782 These are the asynchronous commands which generate the out-of-band
24783 record @samp{*stopped}. Currently @value{GDBN} only really executes
24784 asynchronously with remote targets and this interaction is mimicked in
24787 @subheading The @code{-exec-continue} Command
24788 @findex -exec-continue
24790 @subsubheading Synopsis
24793 -exec-continue [--reverse] [--all|--thread-group N]
24796 Resumes the execution of the inferior program, which will continue
24797 to execute until it reaches a debugger stop event. If the
24798 @samp{--reverse} option is specified, execution resumes in reverse until
24799 it reaches a stop event. Stop events may include
24802 breakpoints or watchpoints
24804 signals or exceptions
24806 the end of the process (or its beginning under @samp{--reverse})
24808 the end or beginning of a replay log if one is being used.
24810 In all-stop mode (@pxref{All-Stop
24811 Mode}), may resume only one thread, or all threads, depending on the
24812 value of the @samp{scheduler-locking} variable. If @samp{--all} is
24813 specified, all threads (in all inferiors) will be resumed. The @samp{--all} option is
24814 ignored in all-stop mode. If the @samp{--thread-group} options is
24815 specified, then all threads in that thread group are resumed.
24817 @subsubheading @value{GDBN} Command
24819 The corresponding @value{GDBN} corresponding is @samp{continue}.
24821 @subsubheading Example
24828 *stopped,reason="breakpoint-hit",disp="keep",bkptno="2",frame=@{
24829 func="foo",args=[],file="hello.c",fullname="/home/foo/bar/hello.c",
24835 @subheading The @code{-exec-finish} Command
24836 @findex -exec-finish
24838 @subsubheading Synopsis
24841 -exec-finish [--reverse]
24844 Resumes the execution of the inferior program until the current
24845 function is exited. Displays the results returned by the function.
24846 If the @samp{--reverse} option is specified, resumes the reverse
24847 execution of the inferior program until the point where current
24848 function was called.
24850 @subsubheading @value{GDBN} Command
24852 The corresponding @value{GDBN} command is @samp{finish}.
24854 @subsubheading Example
24856 Function returning @code{void}.
24863 *stopped,reason="function-finished",frame=@{func="main",args=[],
24864 file="hello.c",fullname="/home/foo/bar/hello.c",line="7"@}
24868 Function returning other than @code{void}. The name of the internal
24869 @value{GDBN} variable storing the result is printed, together with the
24876 *stopped,reason="function-finished",frame=@{addr="0x000107b0",func="foo",
24877 args=[@{name="a",value="1"],@{name="b",value="9"@}@},
24878 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
24879 gdb-result-var="$1",return-value="0"
24884 @subheading The @code{-exec-interrupt} Command
24885 @findex -exec-interrupt
24887 @subsubheading Synopsis
24890 -exec-interrupt [--all|--thread-group N]
24893 Interrupts the background execution of the target. Note how the token
24894 associated with the stop message is the one for the execution command
24895 that has been interrupted. The token for the interrupt itself only
24896 appears in the @samp{^done} output. If the user is trying to
24897 interrupt a non-running program, an error message will be printed.
24899 Note that when asynchronous execution is enabled, this command is
24900 asynchronous just like other execution commands. That is, first the
24901 @samp{^done} response will be printed, and the target stop will be
24902 reported after that using the @samp{*stopped} notification.
24904 In non-stop mode, only the context thread is interrupted by default.
24905 All threads (in all inferiors) will be interrupted if the
24906 @samp{--all} option is specified. If the @samp{--thread-group}
24907 option is specified, all threads in that group will be interrupted.
24909 @subsubheading @value{GDBN} Command
24911 The corresponding @value{GDBN} command is @samp{interrupt}.
24913 @subsubheading Example
24924 111*stopped,signal-name="SIGINT",signal-meaning="Interrupt",
24925 frame=@{addr="0x00010140",func="foo",args=[],file="try.c",
24926 fullname="/home/foo/bar/try.c",line="13"@}
24931 ^error,msg="mi_cmd_exec_interrupt: Inferior not executing."
24935 @subheading The @code{-exec-jump} Command
24938 @subsubheading Synopsis
24941 -exec-jump @var{location}
24944 Resumes execution of the inferior program at the location specified by
24945 parameter. @xref{Specify Location}, for a description of the
24946 different forms of @var{location}.
24948 @subsubheading @value{GDBN} Command
24950 The corresponding @value{GDBN} command is @samp{jump}.
24952 @subsubheading Example
24955 -exec-jump foo.c:10
24956 *running,thread-id="all"
24961 @subheading The @code{-exec-next} Command
24964 @subsubheading Synopsis
24967 -exec-next [--reverse]
24970 Resumes execution of the inferior program, stopping when the beginning
24971 of the next source line is reached.
24973 If the @samp{--reverse} option is specified, resumes reverse execution
24974 of the inferior program, stopping at the beginning of the previous
24975 source line. If you issue this command on the first line of a
24976 function, it will take you back to the caller of that function, to the
24977 source line where the function was called.
24980 @subsubheading @value{GDBN} Command
24982 The corresponding @value{GDBN} command is @samp{next}.
24984 @subsubheading Example
24990 *stopped,reason="end-stepping-range",line="8",file="hello.c"
24995 @subheading The @code{-exec-next-instruction} Command
24996 @findex -exec-next-instruction
24998 @subsubheading Synopsis
25001 -exec-next-instruction [--reverse]
25004 Executes one machine instruction. If the instruction is a function
25005 call, continues until the function returns. If the program stops at an
25006 instruction in the middle of a source line, the address will be
25009 If the @samp{--reverse} option is specified, resumes reverse execution
25010 of the inferior program, stopping at the previous instruction. If the
25011 previously executed instruction was a return from another function,
25012 it will continue to execute in reverse until the call to that function
25013 (from the current stack frame) is reached.
25015 @subsubheading @value{GDBN} Command
25017 The corresponding @value{GDBN} command is @samp{nexti}.
25019 @subsubheading Example
25023 -exec-next-instruction
25027 *stopped,reason="end-stepping-range",
25028 addr="0x000100d4",line="5",file="hello.c"
25033 @subheading The @code{-exec-return} Command
25034 @findex -exec-return
25036 @subsubheading Synopsis
25042 Makes current function return immediately. Doesn't execute the inferior.
25043 Displays the new current frame.
25045 @subsubheading @value{GDBN} Command
25047 The corresponding @value{GDBN} command is @samp{return}.
25049 @subsubheading Example
25053 200-break-insert callee4
25054 200^done,bkpt=@{number="1",addr="0x00010734",
25055 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",line="8"@}
25060 000*stopped,reason="breakpoint-hit",disp="keep",bkptno="1",
25061 frame=@{func="callee4",args=[],
25062 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
25063 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="8"@}
25069 111^done,frame=@{level="0",func="callee3",
25070 args=[@{name="strarg",
25071 value="0x11940 \"A string argument.\""@}],
25072 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
25073 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="18"@}
25078 @subheading The @code{-exec-run} Command
25081 @subsubheading Synopsis
25084 -exec-run [--all | --thread-group N]
25087 Starts execution of the inferior from the beginning. The inferior
25088 executes until either a breakpoint is encountered or the program
25089 exits. In the latter case the output will include an exit code, if
25090 the program has exited exceptionally.
25092 When no option is specified, the current inferior is started. If the
25093 @samp{--thread-group} option is specified, it should refer to a thread
25094 group of type @samp{process}, and that thread group will be started.
25095 If the @samp{--all} option is specified, then all inferiors will be started.
25097 @subsubheading @value{GDBN} Command
25099 The corresponding @value{GDBN} command is @samp{run}.
25101 @subsubheading Examples
25106 ^done,bkpt=@{number="1",addr="0x0001072c",file="recursive2.c",line="4"@}
25111 *stopped,reason="breakpoint-hit",disp="keep",bkptno="1",
25112 frame=@{func="main",args=[],file="recursive2.c",
25113 fullname="/home/foo/bar/recursive2.c",line="4"@}
25118 Program exited normally:
25126 *stopped,reason="exited-normally"
25131 Program exited exceptionally:
25139 *stopped,reason="exited",exit-code="01"
25143 Another way the program can terminate is if it receives a signal such as
25144 @code{SIGINT}. In this case, @sc{gdb/mi} displays this:
25148 *stopped,reason="exited-signalled",signal-name="SIGINT",
25149 signal-meaning="Interrupt"
25153 @c @subheading -exec-signal
25156 @subheading The @code{-exec-step} Command
25159 @subsubheading Synopsis
25162 -exec-step [--reverse]
25165 Resumes execution of the inferior program, stopping when the beginning
25166 of the next source line is reached, if the next source line is not a
25167 function call. If it is, stop at the first instruction of the called
25168 function. If the @samp{--reverse} option is specified, resumes reverse
25169 execution of the inferior program, stopping at the beginning of the
25170 previously executed source line.
25172 @subsubheading @value{GDBN} Command
25174 The corresponding @value{GDBN} command is @samp{step}.
25176 @subsubheading Example
25178 Stepping into a function:
25184 *stopped,reason="end-stepping-range",
25185 frame=@{func="foo",args=[@{name="a",value="10"@},
25186 @{name="b",value="0"@}],file="recursive2.c",
25187 fullname="/home/foo/bar/recursive2.c",line="11"@}
25197 *stopped,reason="end-stepping-range",line="14",file="recursive2.c"
25202 @subheading The @code{-exec-step-instruction} Command
25203 @findex -exec-step-instruction
25205 @subsubheading Synopsis
25208 -exec-step-instruction [--reverse]
25211 Resumes the inferior which executes one machine instruction. If the
25212 @samp{--reverse} option is specified, resumes reverse execution of the
25213 inferior program, stopping at the previously executed instruction.
25214 The output, once @value{GDBN} has stopped, will vary depending on
25215 whether we have stopped in the middle of a source line or not. In the
25216 former case, the address at which the program stopped will be printed
25219 @subsubheading @value{GDBN} Command
25221 The corresponding @value{GDBN} command is @samp{stepi}.
25223 @subsubheading Example
25227 -exec-step-instruction
25231 *stopped,reason="end-stepping-range",
25232 frame=@{func="foo",args=[],file="try.c",
25233 fullname="/home/foo/bar/try.c",line="10"@}
25235 -exec-step-instruction
25239 *stopped,reason="end-stepping-range",
25240 frame=@{addr="0x000100f4",func="foo",args=[],file="try.c",
25241 fullname="/home/foo/bar/try.c",line="10"@}
25246 @subheading The @code{-exec-until} Command
25247 @findex -exec-until
25249 @subsubheading Synopsis
25252 -exec-until [ @var{location} ]
25255 Executes the inferior until the @var{location} specified in the
25256 argument is reached. If there is no argument, the inferior executes
25257 until a source line greater than the current one is reached. The
25258 reason for stopping in this case will be @samp{location-reached}.
25260 @subsubheading @value{GDBN} Command
25262 The corresponding @value{GDBN} command is @samp{until}.
25264 @subsubheading Example
25268 -exec-until recursive2.c:6
25272 *stopped,reason="location-reached",frame=@{func="main",args=[],
25273 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="6"@}
25278 @subheading -file-clear
25279 Is this going away????
25282 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
25283 @node GDB/MI Stack Manipulation
25284 @section @sc{gdb/mi} Stack Manipulation Commands
25287 @subheading The @code{-stack-info-frame} Command
25288 @findex -stack-info-frame
25290 @subsubheading Synopsis
25296 Get info on the selected frame.
25298 @subsubheading @value{GDBN} Command
25300 The corresponding @value{GDBN} command is @samp{info frame} or @samp{frame}
25301 (without arguments).
25303 @subsubheading Example
25308 ^done,frame=@{level="1",addr="0x0001076c",func="callee3",
25309 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
25310 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="17"@}
25314 @subheading The @code{-stack-info-depth} Command
25315 @findex -stack-info-depth
25317 @subsubheading Synopsis
25320 -stack-info-depth [ @var{max-depth} ]
25323 Return the depth of the stack. If the integer argument @var{max-depth}
25324 is specified, do not count beyond @var{max-depth} frames.
25326 @subsubheading @value{GDBN} Command
25328 There's no equivalent @value{GDBN} command.
25330 @subsubheading Example
25332 For a stack with frame levels 0 through 11:
25339 -stack-info-depth 4
25342 -stack-info-depth 12
25345 -stack-info-depth 11
25348 -stack-info-depth 13
25353 @subheading The @code{-stack-list-arguments} Command
25354 @findex -stack-list-arguments
25356 @subsubheading Synopsis
25359 -stack-list-arguments @var{print-values}
25360 [ @var{low-frame} @var{high-frame} ]
25363 Display a list of the arguments for the frames between @var{low-frame}
25364 and @var{high-frame} (inclusive). If @var{low-frame} and
25365 @var{high-frame} are not provided, list the arguments for the whole
25366 call stack. If the two arguments are equal, show the single frame
25367 at the corresponding level. It is an error if @var{low-frame} is
25368 larger than the actual number of frames. On the other hand,
25369 @var{high-frame} may be larger than the actual number of frames, in
25370 which case only existing frames will be returned.
25372 If @var{print-values} is 0 or @code{--no-values}, print only the names of
25373 the variables; if it is 1 or @code{--all-values}, print also their
25374 values; and if it is 2 or @code{--simple-values}, print the name,
25375 type and value for simple data types, and the name and type for arrays,
25376 structures and unions.
25378 Use of this command to obtain arguments in a single frame is
25379 deprecated in favor of the @samp{-stack-list-variables} command.
25381 @subsubheading @value{GDBN} Command
25383 @value{GDBN} does not have an equivalent command. @code{gdbtk} has a
25384 @samp{gdb_get_args} command which partially overlaps with the
25385 functionality of @samp{-stack-list-arguments}.
25387 @subsubheading Example
25394 frame=@{level="0",addr="0x00010734",func="callee4",
25395 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
25396 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="8"@},
25397 frame=@{level="1",addr="0x0001076c",func="callee3",
25398 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
25399 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="17"@},
25400 frame=@{level="2",addr="0x0001078c",func="callee2",
25401 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
25402 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="22"@},
25403 frame=@{level="3",addr="0x000107b4",func="callee1",
25404 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
25405 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="27"@},
25406 frame=@{level="4",addr="0x000107e0",func="main",
25407 file="../../../devo/gdb/testsuite/gdb.mi/basics.c",
25408 fullname="/home/foo/bar/devo/gdb/testsuite/gdb.mi/basics.c",line="32"@}]
25410 -stack-list-arguments 0
25413 frame=@{level="0",args=[]@},
25414 frame=@{level="1",args=[name="strarg"]@},
25415 frame=@{level="2",args=[name="intarg",name="strarg"]@},
25416 frame=@{level="3",args=[name="intarg",name="strarg",name="fltarg"]@},
25417 frame=@{level="4",args=[]@}]
25419 -stack-list-arguments 1
25422 frame=@{level="0",args=[]@},
25424 args=[@{name="strarg",value="0x11940 \"A string argument.\""@}]@},
25425 frame=@{level="2",args=[
25426 @{name="intarg",value="2"@},
25427 @{name="strarg",value="0x11940 \"A string argument.\""@}]@},
25428 @{frame=@{level="3",args=[
25429 @{name="intarg",value="2"@},
25430 @{name="strarg",value="0x11940 \"A string argument.\""@},
25431 @{name="fltarg",value="3.5"@}]@},
25432 frame=@{level="4",args=[]@}]
25434 -stack-list-arguments 0 2 2
25435 ^done,stack-args=[frame=@{level="2",args=[name="intarg",name="strarg"]@}]
25437 -stack-list-arguments 1 2 2
25438 ^done,stack-args=[frame=@{level="2",
25439 args=[@{name="intarg",value="2"@},
25440 @{name="strarg",value="0x11940 \"A string argument.\""@}]@}]
25444 @c @subheading -stack-list-exception-handlers
25447 @subheading The @code{-stack-list-frames} Command
25448 @findex -stack-list-frames
25450 @subsubheading Synopsis
25453 -stack-list-frames [ @var{low-frame} @var{high-frame} ]
25456 List the frames currently on the stack. For each frame it displays the
25461 The frame number, 0 being the topmost frame, i.e., the innermost function.
25463 The @code{$pc} value for that frame.
25467 File name of the source file where the function lives.
25469 Line number corresponding to the @code{$pc}.
25472 If invoked without arguments, this command prints a backtrace for the
25473 whole stack. If given two integer arguments, it shows the frames whose
25474 levels are between the two arguments (inclusive). If the two arguments
25475 are equal, it shows the single frame at the corresponding level. It is
25476 an error if @var{low-frame} is larger than the actual number of
25477 frames. On the other hand, @var{high-frame} may be larger than the
25478 actual number of frames, in which case only existing frames will be returned.
25480 @subsubheading @value{GDBN} Command
25482 The corresponding @value{GDBN} commands are @samp{backtrace} and @samp{where}.
25484 @subsubheading Example
25486 Full stack backtrace:
25492 [frame=@{level="0",addr="0x0001076c",func="foo",
25493 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="11"@},
25494 frame=@{level="1",addr="0x000107a4",func="foo",
25495 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
25496 frame=@{level="2",addr="0x000107a4",func="foo",
25497 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
25498 frame=@{level="3",addr="0x000107a4",func="foo",
25499 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
25500 frame=@{level="4",addr="0x000107a4",func="foo",
25501 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
25502 frame=@{level="5",addr="0x000107a4",func="foo",
25503 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
25504 frame=@{level="6",addr="0x000107a4",func="foo",
25505 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
25506 frame=@{level="7",addr="0x000107a4",func="foo",
25507 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
25508 frame=@{level="8",addr="0x000107a4",func="foo",
25509 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
25510 frame=@{level="9",addr="0x000107a4",func="foo",
25511 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
25512 frame=@{level="10",addr="0x000107a4",func="foo",
25513 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
25514 frame=@{level="11",addr="0x00010738",func="main",
25515 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="4"@}]
25519 Show frames between @var{low_frame} and @var{high_frame}:
25523 -stack-list-frames 3 5
25525 [frame=@{level="3",addr="0x000107a4",func="foo",
25526 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
25527 frame=@{level="4",addr="0x000107a4",func="foo",
25528 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@},
25529 frame=@{level="5",addr="0x000107a4",func="foo",
25530 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@}]
25534 Show a single frame:
25538 -stack-list-frames 3 3
25540 [frame=@{level="3",addr="0x000107a4",func="foo",
25541 file="recursive2.c",fullname="/home/foo/bar/recursive2.c",line="14"@}]
25546 @subheading The @code{-stack-list-locals} Command
25547 @findex -stack-list-locals
25549 @subsubheading Synopsis
25552 -stack-list-locals @var{print-values}
25555 Display the local variable names for the selected frame. If
25556 @var{print-values} is 0 or @code{--no-values}, print only the names of
25557 the variables; if it is 1 or @code{--all-values}, print also their
25558 values; and if it is 2 or @code{--simple-values}, print the name,
25559 type and value for simple data types, and the name and type for arrays,
25560 structures and unions. In this last case, a frontend can immediately
25561 display the value of simple data types and create variable objects for
25562 other data types when the user wishes to explore their values in
25565 This command is deprecated in favor of the
25566 @samp{-stack-list-variables} command.
25568 @subsubheading @value{GDBN} Command
25570 @samp{info locals} in @value{GDBN}, @samp{gdb_get_locals} in @code{gdbtk}.
25572 @subsubheading Example
25576 -stack-list-locals 0
25577 ^done,locals=[name="A",name="B",name="C"]
25579 -stack-list-locals --all-values
25580 ^done,locals=[@{name="A",value="1"@},@{name="B",value="2"@},
25581 @{name="C",value="@{1, 2, 3@}"@}]
25582 -stack-list-locals --simple-values
25583 ^done,locals=[@{name="A",type="int",value="1"@},
25584 @{name="B",type="int",value="2"@},@{name="C",type="int [3]"@}]
25588 @subheading The @code{-stack-list-variables} Command
25589 @findex -stack-list-variables
25591 @subsubheading Synopsis
25594 -stack-list-variables @var{print-values}
25597 Display the names of local variables and function arguments for the selected frame. If
25598 @var{print-values} is 0 or @code{--no-values}, print only the names of
25599 the variables; if it is 1 or @code{--all-values}, print also their
25600 values; and if it is 2 or @code{--simple-values}, print the name,
25601 type and value for simple data types, and the name and type for arrays,
25602 structures and unions.
25604 @subsubheading Example
25608 -stack-list-variables --thread 1 --frame 0 --all-values
25609 ^done,variables=[@{name="x",value="11"@},@{name="s",value="@{a = 1, b = 2@}"@}]
25614 @subheading The @code{-stack-select-frame} Command
25615 @findex -stack-select-frame
25617 @subsubheading Synopsis
25620 -stack-select-frame @var{framenum}
25623 Change the selected frame. Select a different frame @var{framenum} on
25626 This command in deprecated in favor of passing the @samp{--frame}
25627 option to every command.
25629 @subsubheading @value{GDBN} Command
25631 The corresponding @value{GDBN} commands are @samp{frame}, @samp{up},
25632 @samp{down}, @samp{select-frame}, @samp{up-silent}, and @samp{down-silent}.
25634 @subsubheading Example
25638 -stack-select-frame 2
25643 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
25644 @node GDB/MI Variable Objects
25645 @section @sc{gdb/mi} Variable Objects
25649 @subheading Motivation for Variable Objects in @sc{gdb/mi}
25651 For the implementation of a variable debugger window (locals, watched
25652 expressions, etc.), we are proposing the adaptation of the existing code
25653 used by @code{Insight}.
25655 The two main reasons for that are:
25659 It has been proven in practice (it is already on its second generation).
25662 It will shorten development time (needless to say how important it is
25666 The original interface was designed to be used by Tcl code, so it was
25667 slightly changed so it could be used through @sc{gdb/mi}. This section
25668 describes the @sc{gdb/mi} operations that will be available and gives some
25669 hints about their use.
25671 @emph{Note}: In addition to the set of operations described here, we
25672 expect the @sc{gui} implementation of a variable window to require, at
25673 least, the following operations:
25676 @item @code{-gdb-show} @code{output-radix}
25677 @item @code{-stack-list-arguments}
25678 @item @code{-stack-list-locals}
25679 @item @code{-stack-select-frame}
25684 @subheading Introduction to Variable Objects
25686 @cindex variable objects in @sc{gdb/mi}
25688 Variable objects are "object-oriented" MI interface for examining and
25689 changing values of expressions. Unlike some other MI interfaces that
25690 work with expressions, variable objects are specifically designed for
25691 simple and efficient presentation in the frontend. A variable object
25692 is identified by string name. When a variable object is created, the
25693 frontend specifies the expression for that variable object. The
25694 expression can be a simple variable, or it can be an arbitrary complex
25695 expression, and can even involve CPU registers. After creating a
25696 variable object, the frontend can invoke other variable object
25697 operations---for example to obtain or change the value of a variable
25698 object, or to change display format.
25700 Variable objects have hierarchical tree structure. Any variable object
25701 that corresponds to a composite type, such as structure in C, has
25702 a number of child variable objects, for example corresponding to each
25703 element of a structure. A child variable object can itself have
25704 children, recursively. Recursion ends when we reach
25705 leaf variable objects, which always have built-in types. Child variable
25706 objects are created only by explicit request, so if a frontend
25707 is not interested in the children of a particular variable object, no
25708 child will be created.
25710 For a leaf variable object it is possible to obtain its value as a
25711 string, or set the value from a string. String value can be also
25712 obtained for a non-leaf variable object, but it's generally a string
25713 that only indicates the type of the object, and does not list its
25714 contents. Assignment to a non-leaf variable object is not allowed.
25716 A frontend does not need to read the values of all variable objects each time
25717 the program stops. Instead, MI provides an update command that lists all
25718 variable objects whose values has changed since the last update
25719 operation. This considerably reduces the amount of data that must
25720 be transferred to the frontend. As noted above, children variable
25721 objects are created on demand, and only leaf variable objects have a
25722 real value. As result, gdb will read target memory only for leaf
25723 variables that frontend has created.
25725 The automatic update is not always desirable. For example, a frontend
25726 might want to keep a value of some expression for future reference,
25727 and never update it. For another example, fetching memory is
25728 relatively slow for embedded targets, so a frontend might want
25729 to disable automatic update for the variables that are either not
25730 visible on the screen, or ``closed''. This is possible using so
25731 called ``frozen variable objects''. Such variable objects are never
25732 implicitly updated.
25734 Variable objects can be either @dfn{fixed} or @dfn{floating}. For the
25735 fixed variable object, the expression is parsed when the variable
25736 object is created, including associating identifiers to specific
25737 variables. The meaning of expression never changes. For a floating
25738 variable object the values of variables whose names appear in the
25739 expressions are re-evaluated every time in the context of the current
25740 frame. Consider this example:
25745 struct work_state state;
25752 If a fixed variable object for the @code{state} variable is created in
25753 this function, and we enter the recursive call, the the variable
25754 object will report the value of @code{state} in the top-level
25755 @code{do_work} invocation. On the other hand, a floating variable
25756 object will report the value of @code{state} in the current frame.
25758 If an expression specified when creating a fixed variable object
25759 refers to a local variable, the variable object becomes bound to the
25760 thread and frame in which the variable object is created. When such
25761 variable object is updated, @value{GDBN} makes sure that the
25762 thread/frame combination the variable object is bound to still exists,
25763 and re-evaluates the variable object in context of that thread/frame.
25765 The following is the complete set of @sc{gdb/mi} operations defined to
25766 access this functionality:
25768 @multitable @columnfractions .4 .6
25769 @item @strong{Operation}
25770 @tab @strong{Description}
25772 @item @code{-enable-pretty-printing}
25773 @tab enable Python-based pretty-printing
25774 @item @code{-var-create}
25775 @tab create a variable object
25776 @item @code{-var-delete}
25777 @tab delete the variable object and/or its children
25778 @item @code{-var-set-format}
25779 @tab set the display format of this variable
25780 @item @code{-var-show-format}
25781 @tab show the display format of this variable
25782 @item @code{-var-info-num-children}
25783 @tab tells how many children this object has
25784 @item @code{-var-list-children}
25785 @tab return a list of the object's children
25786 @item @code{-var-info-type}
25787 @tab show the type of this variable object
25788 @item @code{-var-info-expression}
25789 @tab print parent-relative expression that this variable object represents
25790 @item @code{-var-info-path-expression}
25791 @tab print full expression that this variable object represents
25792 @item @code{-var-show-attributes}
25793 @tab is this variable editable? does it exist here?
25794 @item @code{-var-evaluate-expression}
25795 @tab get the value of this variable
25796 @item @code{-var-assign}
25797 @tab set the value of this variable
25798 @item @code{-var-update}
25799 @tab update the variable and its children
25800 @item @code{-var-set-frozen}
25801 @tab set frozeness attribute
25802 @item @code{-var-set-update-range}
25803 @tab set range of children to display on update
25806 In the next subsection we describe each operation in detail and suggest
25807 how it can be used.
25809 @subheading Description And Use of Operations on Variable Objects
25811 @subheading The @code{-enable-pretty-printing} Command
25812 @findex -enable-pretty-printing
25815 -enable-pretty-printing
25818 @value{GDBN} allows Python-based visualizers to affect the output of the
25819 MI variable object commands. However, because there was no way to
25820 implement this in a fully backward-compatible way, a front end must
25821 request that this functionality be enabled.
25823 Once enabled, this feature cannot be disabled.
25825 Note that if Python support has not been compiled into @value{GDBN},
25826 this command will still succeed (and do nothing).
25828 This feature is currently (as of @value{GDBN} 7.0) experimental, and
25829 may work differently in future versions of @value{GDBN}.
25831 @subheading The @code{-var-create} Command
25832 @findex -var-create
25834 @subsubheading Synopsis
25837 -var-create @{@var{name} | "-"@}
25838 @{@var{frame-addr} | "*" | "@@"@} @var{expression}
25841 This operation creates a variable object, which allows the monitoring of
25842 a variable, the result of an expression, a memory cell or a CPU
25845 The @var{name} parameter is the string by which the object can be
25846 referenced. It must be unique. If @samp{-} is specified, the varobj
25847 system will generate a string ``varNNNNNN'' automatically. It will be
25848 unique provided that one does not specify @var{name} of that format.
25849 The command fails if a duplicate name is found.
25851 The frame under which the expression should be evaluated can be
25852 specified by @var{frame-addr}. A @samp{*} indicates that the current
25853 frame should be used. A @samp{@@} indicates that a floating variable
25854 object must be created.
25856 @var{expression} is any expression valid on the current language set (must not
25857 begin with a @samp{*}), or one of the following:
25861 @samp{*@var{addr}}, where @var{addr} is the address of a memory cell
25864 @samp{*@var{addr}-@var{addr}} --- a memory address range (TBD)
25867 @samp{$@var{regname}} --- a CPU register name
25870 @cindex dynamic varobj
25871 A varobj's contents may be provided by a Python-based pretty-printer. In this
25872 case the varobj is known as a @dfn{dynamic varobj}. Dynamic varobjs
25873 have slightly different semantics in some cases. If the
25874 @code{-enable-pretty-printing} command is not sent, then @value{GDBN}
25875 will never create a dynamic varobj. This ensures backward
25876 compatibility for existing clients.
25878 @subsubheading Result
25880 This operation returns attributes of the newly-created varobj. These
25885 The name of the varobj.
25888 The number of children of the varobj. This number is not necessarily
25889 reliable for a dynamic varobj. Instead, you must examine the
25890 @samp{has_more} attribute.
25893 The varobj's scalar value. For a varobj whose type is some sort of
25894 aggregate (e.g., a @code{struct}), or for a dynamic varobj, this value
25895 will not be interesting.
25898 The varobj's type. This is a string representation of the type, as
25899 would be printed by the @value{GDBN} CLI.
25902 If a variable object is bound to a specific thread, then this is the
25903 thread's identifier.
25906 For a dynamic varobj, this indicates whether there appear to be any
25907 children available. For a non-dynamic varobj, this will be 0.
25910 This attribute will be present and have the value @samp{1} if the
25911 varobj is a dynamic varobj. If the varobj is not a dynamic varobj,
25912 then this attribute will not be present.
25915 A dynamic varobj can supply a display hint to the front end. The
25916 value comes directly from the Python pretty-printer object's
25917 @code{display_hint} method. @xref{Pretty Printing API}.
25920 Typical output will look like this:
25923 name="@var{name}",numchild="@var{N}",type="@var{type}",thread-id="@var{M}",
25924 has_more="@var{has_more}"
25928 @subheading The @code{-var-delete} Command
25929 @findex -var-delete
25931 @subsubheading Synopsis
25934 -var-delete [ -c ] @var{name}
25937 Deletes a previously created variable object and all of its children.
25938 With the @samp{-c} option, just deletes the children.
25940 Returns an error if the object @var{name} is not found.
25943 @subheading The @code{-var-set-format} Command
25944 @findex -var-set-format
25946 @subsubheading Synopsis
25949 -var-set-format @var{name} @var{format-spec}
25952 Sets the output format for the value of the object @var{name} to be
25955 @anchor{-var-set-format}
25956 The syntax for the @var{format-spec} is as follows:
25959 @var{format-spec} @expansion{}
25960 @{binary | decimal | hexadecimal | octal | natural@}
25963 The natural format is the default format choosen automatically
25964 based on the variable type (like decimal for an @code{int}, hex
25965 for pointers, etc.).
25967 For a variable with children, the format is set only on the
25968 variable itself, and the children are not affected.
25970 @subheading The @code{-var-show-format} Command
25971 @findex -var-show-format
25973 @subsubheading Synopsis
25976 -var-show-format @var{name}
25979 Returns the format used to display the value of the object @var{name}.
25982 @var{format} @expansion{}
25987 @subheading The @code{-var-info-num-children} Command
25988 @findex -var-info-num-children
25990 @subsubheading Synopsis
25993 -var-info-num-children @var{name}
25996 Returns the number of children of a variable object @var{name}:
26002 Note that this number is not completely reliable for a dynamic varobj.
26003 It will return the current number of children, but more children may
26007 @subheading The @code{-var-list-children} Command
26008 @findex -var-list-children
26010 @subsubheading Synopsis
26013 -var-list-children [@var{print-values}] @var{name} [@var{from} @var{to}]
26015 @anchor{-var-list-children}
26017 Return a list of the children of the specified variable object and
26018 create variable objects for them, if they do not already exist. With
26019 a single argument or if @var{print-values} has a value for of 0 or
26020 @code{--no-values}, print only the names of the variables; if
26021 @var{print-values} is 1 or @code{--all-values}, also print their
26022 values; and if it is 2 or @code{--simple-values} print the name and
26023 value for simple data types and just the name for arrays, structures
26026 @var{from} and @var{to}, if specified, indicate the range of children
26027 to report. If @var{from} or @var{to} is less than zero, the range is
26028 reset and all children will be reported. Otherwise, children starting
26029 at @var{from} (zero-based) and up to and excluding @var{to} will be
26032 If a child range is requested, it will only affect the current call to
26033 @code{-var-list-children}, but not future calls to @code{-var-update}.
26034 For this, you must instead use @code{-var-set-update-range}. The
26035 intent of this approach is to enable a front end to implement any
26036 update approach it likes; for example, scrolling a view may cause the
26037 front end to request more children with @code{-var-list-children}, and
26038 then the front end could call @code{-var-set-update-range} with a
26039 different range to ensure that future updates are restricted to just
26042 For each child the following results are returned:
26047 Name of the variable object created for this child.
26050 The expression to be shown to the user by the front end to designate this child.
26051 For example this may be the name of a structure member.
26053 For a dynamic varobj, this value cannot be used to form an
26054 expression. There is no way to do this at all with a dynamic varobj.
26056 For C/C@t{++} structures there are several pseudo children returned to
26057 designate access qualifiers. For these pseudo children @var{exp} is
26058 @samp{public}, @samp{private}, or @samp{protected}. In this case the
26059 type and value are not present.
26061 A dynamic varobj will not report the access qualifying
26062 pseudo-children, regardless of the language. This information is not
26063 available at all with a dynamic varobj.
26066 Number of children this child has. For a dynamic varobj, this will be
26070 The type of the child.
26073 If values were requested, this is the value.
26076 If this variable object is associated with a thread, this is the thread id.
26077 Otherwise this result is not present.
26080 If the variable object is frozen, this variable will be present with a value of 1.
26083 The result may have its own attributes:
26087 A dynamic varobj can supply a display hint to the front end. The
26088 value comes directly from the Python pretty-printer object's
26089 @code{display_hint} method. @xref{Pretty Printing API}.
26092 This is an integer attribute which is nonzero if there are children
26093 remaining after the end of the selected range.
26096 @subsubheading Example
26100 -var-list-children n
26101 ^done,numchild=@var{n},children=[child=@{name=@var{name},exp=@var{exp},
26102 numchild=@var{n},type=@var{type}@},@r{(repeats N times)}]
26104 -var-list-children --all-values n
26105 ^done,numchild=@var{n},children=[child=@{name=@var{name},exp=@var{exp},
26106 numchild=@var{n},value=@var{value},type=@var{type}@},@r{(repeats N times)}]
26110 @subheading The @code{-var-info-type} Command
26111 @findex -var-info-type
26113 @subsubheading Synopsis
26116 -var-info-type @var{name}
26119 Returns the type of the specified variable @var{name}. The type is
26120 returned as a string in the same format as it is output by the
26124 type=@var{typename}
26128 @subheading The @code{-var-info-expression} Command
26129 @findex -var-info-expression
26131 @subsubheading Synopsis
26134 -var-info-expression @var{name}
26137 Returns a string that is suitable for presenting this
26138 variable object in user interface. The string is generally
26139 not valid expression in the current language, and cannot be evaluated.
26141 For example, if @code{a} is an array, and variable object
26142 @code{A} was created for @code{a}, then we'll get this output:
26145 (gdb) -var-info-expression A.1
26146 ^done,lang="C",exp="1"
26150 Here, the values of @code{lang} can be @code{@{"C" | "C++" | "Java"@}}.
26152 Note that the output of the @code{-var-list-children} command also
26153 includes those expressions, so the @code{-var-info-expression} command
26156 @subheading The @code{-var-info-path-expression} Command
26157 @findex -var-info-path-expression
26159 @subsubheading Synopsis
26162 -var-info-path-expression @var{name}
26165 Returns an expression that can be evaluated in the current
26166 context and will yield the same value that a variable object has.
26167 Compare this with the @code{-var-info-expression} command, which
26168 result can be used only for UI presentation. Typical use of
26169 the @code{-var-info-path-expression} command is creating a
26170 watchpoint from a variable object.
26172 This command is currently not valid for children of a dynamic varobj,
26173 and will give an error when invoked on one.
26175 For example, suppose @code{C} is a C@t{++} class, derived from class
26176 @code{Base}, and that the @code{Base} class has a member called
26177 @code{m_size}. Assume a variable @code{c} is has the type of
26178 @code{C} and a variable object @code{C} was created for variable
26179 @code{c}. Then, we'll get this output:
26181 (gdb) -var-info-path-expression C.Base.public.m_size
26182 ^done,path_expr=((Base)c).m_size)
26185 @subheading The @code{-var-show-attributes} Command
26186 @findex -var-show-attributes
26188 @subsubheading Synopsis
26191 -var-show-attributes @var{name}
26194 List attributes of the specified variable object @var{name}:
26197 status=@var{attr} [ ( ,@var{attr} )* ]
26201 where @var{attr} is @code{@{ @{ editable | noneditable @} | TBD @}}.
26203 @subheading The @code{-var-evaluate-expression} Command
26204 @findex -var-evaluate-expression
26206 @subsubheading Synopsis
26209 -var-evaluate-expression [-f @var{format-spec}] @var{name}
26212 Evaluates the expression that is represented by the specified variable
26213 object and returns its value as a string. The format of the string
26214 can be specified with the @samp{-f} option. The possible values of
26215 this option are the same as for @code{-var-set-format}
26216 (@pxref{-var-set-format}). If the @samp{-f} option is not specified,
26217 the current display format will be used. The current display format
26218 can be changed using the @code{-var-set-format} command.
26224 Note that one must invoke @code{-var-list-children} for a variable
26225 before the value of a child variable can be evaluated.
26227 @subheading The @code{-var-assign} Command
26228 @findex -var-assign
26230 @subsubheading Synopsis
26233 -var-assign @var{name} @var{expression}
26236 Assigns the value of @var{expression} to the variable object specified
26237 by @var{name}. The object must be @samp{editable}. If the variable's
26238 value is altered by the assign, the variable will show up in any
26239 subsequent @code{-var-update} list.
26241 @subsubheading Example
26249 ^done,changelist=[@{name="var1",in_scope="true",type_changed="false"@}]
26253 @subheading The @code{-var-update} Command
26254 @findex -var-update
26256 @subsubheading Synopsis
26259 -var-update [@var{print-values}] @{@var{name} | "*"@}
26262 Reevaluate the expressions corresponding to the variable object
26263 @var{name} and all its direct and indirect children, and return the
26264 list of variable objects whose values have changed; @var{name} must
26265 be a root variable object. Here, ``changed'' means that the result of
26266 @code{-var-evaluate-expression} before and after the
26267 @code{-var-update} is different. If @samp{*} is used as the variable
26268 object names, all existing variable objects are updated, except
26269 for frozen ones (@pxref{-var-set-frozen}). The option
26270 @var{print-values} determines whether both names and values, or just
26271 names are printed. The possible values of this option are the same
26272 as for @code{-var-list-children} (@pxref{-var-list-children}). It is
26273 recommended to use the @samp{--all-values} option, to reduce the
26274 number of MI commands needed on each program stop.
26276 With the @samp{*} parameter, if a variable object is bound to a
26277 currently running thread, it will not be updated, without any
26280 If @code{-var-set-update-range} was previously used on a varobj, then
26281 only the selected range of children will be reported.
26283 @code{-var-update} reports all the changed varobjs in a tuple named
26286 Each item in the change list is itself a tuple holding:
26290 The name of the varobj.
26293 If values were requested for this update, then this field will be
26294 present and will hold the value of the varobj.
26297 @anchor{-var-update}
26298 This field is a string which may take one of three values:
26302 The variable object's current value is valid.
26305 The variable object does not currently hold a valid value but it may
26306 hold one in the future if its associated expression comes back into
26310 The variable object no longer holds a valid value.
26311 This can occur when the executable file being debugged has changed,
26312 either through recompilation or by using the @value{GDBN} @code{file}
26313 command. The front end should normally choose to delete these variable
26317 In the future new values may be added to this list so the front should
26318 be prepared for this possibility. @xref{GDB/MI Development and Front Ends, ,@sc{GDB/MI} Development and Front Ends}.
26321 This is only present if the varobj is still valid. If the type
26322 changed, then this will be the string @samp{true}; otherwise it will
26326 If the varobj's type changed, then this field will be present and will
26329 @item new_num_children
26330 For a dynamic varobj, if the number of children changed, or if the
26331 type changed, this will be the new number of children.
26333 The @samp{numchild} field in other varobj responses is generally not
26334 valid for a dynamic varobj -- it will show the number of children that
26335 @value{GDBN} knows about, but because dynamic varobjs lazily
26336 instantiate their children, this will not reflect the number of
26337 children which may be available.
26339 The @samp{new_num_children} attribute only reports changes to the
26340 number of children known by @value{GDBN}. This is the only way to
26341 detect whether an update has removed children (which necessarily can
26342 only happen at the end of the update range).
26345 The display hint, if any.
26348 This is an integer value, which will be 1 if there are more children
26349 available outside the varobj's update range.
26352 This attribute will be present and have the value @samp{1} if the
26353 varobj is a dynamic varobj. If the varobj is not a dynamic varobj,
26354 then this attribute will not be present.
26357 If new children were added to a dynamic varobj within the selected
26358 update range (as set by @code{-var-set-update-range}), then they will
26359 be listed in this attribute.
26362 @subsubheading Example
26369 -var-update --all-values var1
26370 ^done,changelist=[@{name="var1",value="3",in_scope="true",
26371 type_changed="false"@}]
26375 @subheading The @code{-var-set-frozen} Command
26376 @findex -var-set-frozen
26377 @anchor{-var-set-frozen}
26379 @subsubheading Synopsis
26382 -var-set-frozen @var{name} @var{flag}
26385 Set the frozenness flag on the variable object @var{name}. The
26386 @var{flag} parameter should be either @samp{1} to make the variable
26387 frozen or @samp{0} to make it unfrozen. If a variable object is
26388 frozen, then neither itself, nor any of its children, are
26389 implicitly updated by @code{-var-update} of
26390 a parent variable or by @code{-var-update *}. Only
26391 @code{-var-update} of the variable itself will update its value and
26392 values of its children. After a variable object is unfrozen, it is
26393 implicitly updated by all subsequent @code{-var-update} operations.
26394 Unfreezing a variable does not update it, only subsequent
26395 @code{-var-update} does.
26397 @subsubheading Example
26401 -var-set-frozen V 1
26406 @subheading The @code{-var-set-update-range} command
26407 @findex -var-set-update-range
26408 @anchor{-var-set-update-range}
26410 @subsubheading Synopsis
26413 -var-set-update-range @var{name} @var{from} @var{to}
26416 Set the range of children to be returned by future invocations of
26417 @code{-var-update}.
26419 @var{from} and @var{to} indicate the range of children to report. If
26420 @var{from} or @var{to} is less than zero, the range is reset and all
26421 children will be reported. Otherwise, children starting at @var{from}
26422 (zero-based) and up to and excluding @var{to} will be reported.
26424 @subsubheading Example
26428 -var-set-update-range V 1 2
26432 @subheading The @code{-var-set-visualizer} command
26433 @findex -var-set-visualizer
26434 @anchor{-var-set-visualizer}
26436 @subsubheading Synopsis
26439 -var-set-visualizer @var{name} @var{visualizer}
26442 Set a visualizer for the variable object @var{name}.
26444 @var{visualizer} is the visualizer to use. The special value
26445 @samp{None} means to disable any visualizer in use.
26447 If not @samp{None}, @var{visualizer} must be a Python expression.
26448 This expression must evaluate to a callable object which accepts a
26449 single argument. @value{GDBN} will call this object with the value of
26450 the varobj @var{name} as an argument (this is done so that the same
26451 Python pretty-printing code can be used for both the CLI and MI).
26452 When called, this object must return an object which conforms to the
26453 pretty-printing interface (@pxref{Pretty Printing API}).
26455 The pre-defined function @code{gdb.default_visualizer} may be used to
26456 select a visualizer by following the built-in process
26457 (@pxref{Selecting Pretty-Printers}). This is done automatically when
26458 a varobj is created, and so ordinarily is not needed.
26460 This feature is only available if Python support is enabled. The MI
26461 command @code{-list-features} (@pxref{GDB/MI Miscellaneous Commands})
26462 can be used to check this.
26464 @subsubheading Example
26466 Resetting the visualizer:
26470 -var-set-visualizer V None
26474 Reselecting the default (type-based) visualizer:
26478 -var-set-visualizer V gdb.default_visualizer
26482 Suppose @code{SomeClass} is a visualizer class. A lambda expression
26483 can be used to instantiate this class for a varobj:
26487 -var-set-visualizer V "lambda val: SomeClass()"
26491 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
26492 @node GDB/MI Data Manipulation
26493 @section @sc{gdb/mi} Data Manipulation
26495 @cindex data manipulation, in @sc{gdb/mi}
26496 @cindex @sc{gdb/mi}, data manipulation
26497 This section describes the @sc{gdb/mi} commands that manipulate data:
26498 examine memory and registers, evaluate expressions, etc.
26500 @c REMOVED FROM THE INTERFACE.
26501 @c @subheading -data-assign
26502 @c Change the value of a program variable. Plenty of side effects.
26503 @c @subsubheading GDB Command
26505 @c @subsubheading Example
26508 @subheading The @code{-data-disassemble} Command
26509 @findex -data-disassemble
26511 @subsubheading Synopsis
26515 [ -s @var{start-addr} -e @var{end-addr} ]
26516 | [ -f @var{filename} -l @var{linenum} [ -n @var{lines} ] ]
26524 @item @var{start-addr}
26525 is the beginning address (or @code{$pc})
26526 @item @var{end-addr}
26528 @item @var{filename}
26529 is the name of the file to disassemble
26530 @item @var{linenum}
26531 is the line number to disassemble around
26533 is the number of disassembly lines to be produced. If it is -1,
26534 the whole function will be disassembled, in case no @var{end-addr} is
26535 specified. If @var{end-addr} is specified as a non-zero value, and
26536 @var{lines} is lower than the number of disassembly lines between
26537 @var{start-addr} and @var{end-addr}, only @var{lines} lines are
26538 displayed; if @var{lines} is higher than the number of lines between
26539 @var{start-addr} and @var{end-addr}, only the lines up to @var{end-addr}
26542 is either 0 (meaning only disassembly) or 1 (meaning mixed source and
26546 @subsubheading Result
26548 The output for each instruction is composed of four fields:
26557 Note that whatever included in the instruction field, is not manipulated
26558 directly by @sc{gdb/mi}, i.e., it is not possible to adjust its format.
26560 @subsubheading @value{GDBN} Command
26562 There's no direct mapping from this command to the CLI.
26564 @subsubheading Example
26566 Disassemble from the current value of @code{$pc} to @code{$pc + 20}:
26570 -data-disassemble -s $pc -e "$pc + 20" -- 0
26573 @{address="0x000107c0",func-name="main",offset="4",
26574 inst="mov 2, %o0"@},
26575 @{address="0x000107c4",func-name="main",offset="8",
26576 inst="sethi %hi(0x11800), %o2"@},
26577 @{address="0x000107c8",func-name="main",offset="12",
26578 inst="or %o2, 0x140, %o1\t! 0x11940 <_lib_version+8>"@},
26579 @{address="0x000107cc",func-name="main",offset="16",
26580 inst="sethi %hi(0x11800), %o2"@},
26581 @{address="0x000107d0",func-name="main",offset="20",
26582 inst="or %o2, 0x168, %o4\t! 0x11968 <_lib_version+48>"@}]
26586 Disassemble the whole @code{main} function. Line 32 is part of
26590 -data-disassemble -f basics.c -l 32 -- 0
26592 @{address="0x000107bc",func-name="main",offset="0",
26593 inst="save %sp, -112, %sp"@},
26594 @{address="0x000107c0",func-name="main",offset="4",
26595 inst="mov 2, %o0"@},
26596 @{address="0x000107c4",func-name="main",offset="8",
26597 inst="sethi %hi(0x11800), %o2"@},
26599 @{address="0x0001081c",func-name="main",offset="96",inst="ret "@},
26600 @{address="0x00010820",func-name="main",offset="100",inst="restore "@}]
26604 Disassemble 3 instructions from the start of @code{main}:
26608 -data-disassemble -f basics.c -l 32 -n 3 -- 0
26610 @{address="0x000107bc",func-name="main",offset="0",
26611 inst="save %sp, -112, %sp"@},
26612 @{address="0x000107c0",func-name="main",offset="4",
26613 inst="mov 2, %o0"@},
26614 @{address="0x000107c4",func-name="main",offset="8",
26615 inst="sethi %hi(0x11800), %o2"@}]
26619 Disassemble 3 instructions from the start of @code{main} in mixed mode:
26623 -data-disassemble -f basics.c -l 32 -n 3 -- 1
26625 src_and_asm_line=@{line="31",
26626 file="/kwikemart/marge/ezannoni/flathead-dev/devo/gdb/ \
26627 testsuite/gdb.mi/basics.c",line_asm_insn=[
26628 @{address="0x000107bc",func-name="main",offset="0",
26629 inst="save %sp, -112, %sp"@}]@},
26630 src_and_asm_line=@{line="32",
26631 file="/kwikemart/marge/ezannoni/flathead-dev/devo/gdb/ \
26632 testsuite/gdb.mi/basics.c",line_asm_insn=[
26633 @{address="0x000107c0",func-name="main",offset="4",
26634 inst="mov 2, %o0"@},
26635 @{address="0x000107c4",func-name="main",offset="8",
26636 inst="sethi %hi(0x11800), %o2"@}]@}]
26641 @subheading The @code{-data-evaluate-expression} Command
26642 @findex -data-evaluate-expression
26644 @subsubheading Synopsis
26647 -data-evaluate-expression @var{expr}
26650 Evaluate @var{expr} as an expression. The expression could contain an
26651 inferior function call. The function call will execute synchronously.
26652 If the expression contains spaces, it must be enclosed in double quotes.
26654 @subsubheading @value{GDBN} Command
26656 The corresponding @value{GDBN} commands are @samp{print}, @samp{output}, and
26657 @samp{call}. In @code{gdbtk} only, there's a corresponding
26658 @samp{gdb_eval} command.
26660 @subsubheading Example
26662 In the following example, the numbers that precede the commands are the
26663 @dfn{tokens} described in @ref{GDB/MI Command Syntax, ,@sc{gdb/mi}
26664 Command Syntax}. Notice how @sc{gdb/mi} returns the same tokens in its
26668 211-data-evaluate-expression A
26671 311-data-evaluate-expression &A
26672 311^done,value="0xefffeb7c"
26674 411-data-evaluate-expression A+3
26677 511-data-evaluate-expression "A + 3"
26683 @subheading The @code{-data-list-changed-registers} Command
26684 @findex -data-list-changed-registers
26686 @subsubheading Synopsis
26689 -data-list-changed-registers
26692 Display a list of the registers that have changed.
26694 @subsubheading @value{GDBN} Command
26696 @value{GDBN} doesn't have a direct analog for this command; @code{gdbtk}
26697 has the corresponding command @samp{gdb_changed_register_list}.
26699 @subsubheading Example
26701 On a PPC MBX board:
26709 *stopped,reason="breakpoint-hit",disp="keep",bkptno="1",frame=@{
26710 func="main",args=[],file="try.c",fullname="/home/foo/bar/try.c",
26713 -data-list-changed-registers
26714 ^done,changed-registers=["0","1","2","4","5","6","7","8","9",
26715 "10","11","13","14","15","16","17","18","19","20","21","22","23",
26716 "24","25","26","27","28","30","31","64","65","66","67","69"]
26721 @subheading The @code{-data-list-register-names} Command
26722 @findex -data-list-register-names
26724 @subsubheading Synopsis
26727 -data-list-register-names [ ( @var{regno} )+ ]
26730 Show a list of register names for the current target. If no arguments
26731 are given, it shows a list of the names of all the registers. If
26732 integer numbers are given as arguments, it will print a list of the
26733 names of the registers corresponding to the arguments. To ensure
26734 consistency between a register name and its number, the output list may
26735 include empty register names.
26737 @subsubheading @value{GDBN} Command
26739 @value{GDBN} does not have a command which corresponds to
26740 @samp{-data-list-register-names}. In @code{gdbtk} there is a
26741 corresponding command @samp{gdb_regnames}.
26743 @subsubheading Example
26745 For the PPC MBX board:
26748 -data-list-register-names
26749 ^done,register-names=["r0","r1","r2","r3","r4","r5","r6","r7",
26750 "r8","r9","r10","r11","r12","r13","r14","r15","r16","r17","r18",
26751 "r19","r20","r21","r22","r23","r24","r25","r26","r27","r28","r29",
26752 "r30","r31","f0","f1","f2","f3","f4","f5","f6","f7","f8","f9",
26753 "f10","f11","f12","f13","f14","f15","f16","f17","f18","f19","f20",
26754 "f21","f22","f23","f24","f25","f26","f27","f28","f29","f30","f31",
26755 "", "pc","ps","cr","lr","ctr","xer"]
26757 -data-list-register-names 1 2 3
26758 ^done,register-names=["r1","r2","r3"]
26762 @subheading The @code{-data-list-register-values} Command
26763 @findex -data-list-register-values
26765 @subsubheading Synopsis
26768 -data-list-register-values @var{fmt} [ ( @var{regno} )*]
26771 Display the registers' contents. @var{fmt} is the format according to
26772 which the registers' contents are to be returned, followed by an optional
26773 list of numbers specifying the registers to display. A missing list of
26774 numbers indicates that the contents of all the registers must be returned.
26776 Allowed formats for @var{fmt} are:
26793 @subsubheading @value{GDBN} Command
26795 The corresponding @value{GDBN} commands are @samp{info reg}, @samp{info
26796 all-reg}, and (in @code{gdbtk}) @samp{gdb_fetch_registers}.
26798 @subsubheading Example
26800 For a PPC MBX board (note: line breaks are for readability only, they
26801 don't appear in the actual output):
26805 -data-list-register-values r 64 65
26806 ^done,register-values=[@{number="64",value="0xfe00a300"@},
26807 @{number="65",value="0x00029002"@}]
26809 -data-list-register-values x
26810 ^done,register-values=[@{number="0",value="0xfe0043c8"@},
26811 @{number="1",value="0x3fff88"@},@{number="2",value="0xfffffffe"@},
26812 @{number="3",value="0x0"@},@{number="4",value="0xa"@},
26813 @{number="5",value="0x3fff68"@},@{number="6",value="0x3fff58"@},
26814 @{number="7",value="0xfe011e98"@},@{number="8",value="0x2"@},
26815 @{number="9",value="0xfa202820"@},@{number="10",value="0xfa202808"@},
26816 @{number="11",value="0x1"@},@{number="12",value="0x0"@},
26817 @{number="13",value="0x4544"@},@{number="14",value="0xffdfffff"@},
26818 @{number="15",value="0xffffffff"@},@{number="16",value="0xfffffeff"@},
26819 @{number="17",value="0xefffffed"@},@{number="18",value="0xfffffffe"@},
26820 @{number="19",value="0xffffffff"@},@{number="20",value="0xffffffff"@},
26821 @{number="21",value="0xffffffff"@},@{number="22",value="0xfffffff7"@},
26822 @{number="23",value="0xffffffff"@},@{number="24",value="0xffffffff"@},
26823 @{number="25",value="0xffffffff"@},@{number="26",value="0xfffffffb"@},
26824 @{number="27",value="0xffffffff"@},@{number="28",value="0xf7bfffff"@},
26825 @{number="29",value="0x0"@},@{number="30",value="0xfe010000"@},
26826 @{number="31",value="0x0"@},@{number="32",value="0x0"@},
26827 @{number="33",value="0x0"@},@{number="34",value="0x0"@},
26828 @{number="35",value="0x0"@},@{number="36",value="0x0"@},
26829 @{number="37",value="0x0"@},@{number="38",value="0x0"@},
26830 @{number="39",value="0x0"@},@{number="40",value="0x0"@},
26831 @{number="41",value="0x0"@},@{number="42",value="0x0"@},
26832 @{number="43",value="0x0"@},@{number="44",value="0x0"@},
26833 @{number="45",value="0x0"@},@{number="46",value="0x0"@},
26834 @{number="47",value="0x0"@},@{number="48",value="0x0"@},
26835 @{number="49",value="0x0"@},@{number="50",value="0x0"@},
26836 @{number="51",value="0x0"@},@{number="52",value="0x0"@},
26837 @{number="53",value="0x0"@},@{number="54",value="0x0"@},
26838 @{number="55",value="0x0"@},@{number="56",value="0x0"@},
26839 @{number="57",value="0x0"@},@{number="58",value="0x0"@},
26840 @{number="59",value="0x0"@},@{number="60",value="0x0"@},
26841 @{number="61",value="0x0"@},@{number="62",value="0x0"@},
26842 @{number="63",value="0x0"@},@{number="64",value="0xfe00a300"@},
26843 @{number="65",value="0x29002"@},@{number="66",value="0x202f04b5"@},
26844 @{number="67",value="0xfe0043b0"@},@{number="68",value="0xfe00b3e4"@},
26845 @{number="69",value="0x20002b03"@}]
26850 @subheading The @code{-data-read-memory} Command
26851 @findex -data-read-memory
26853 @subsubheading Synopsis
26856 -data-read-memory [ -o @var{byte-offset} ]
26857 @var{address} @var{word-format} @var{word-size}
26858 @var{nr-rows} @var{nr-cols} [ @var{aschar} ]
26865 @item @var{address}
26866 An expression specifying the address of the first memory word to be
26867 read. Complex expressions containing embedded white space should be
26868 quoted using the C convention.
26870 @item @var{word-format}
26871 The format to be used to print the memory words. The notation is the
26872 same as for @value{GDBN}'s @code{print} command (@pxref{Output Formats,
26875 @item @var{word-size}
26876 The size of each memory word in bytes.
26878 @item @var{nr-rows}
26879 The number of rows in the output table.
26881 @item @var{nr-cols}
26882 The number of columns in the output table.
26885 If present, indicates that each row should include an @sc{ascii} dump. The
26886 value of @var{aschar} is used as a padding character when a byte is not a
26887 member of the printable @sc{ascii} character set (printable @sc{ascii}
26888 characters are those whose code is between 32 and 126, inclusively).
26890 @item @var{byte-offset}
26891 An offset to add to the @var{address} before fetching memory.
26894 This command displays memory contents as a table of @var{nr-rows} by
26895 @var{nr-cols} words, each word being @var{word-size} bytes. In total,
26896 @code{@var{nr-rows} * @var{nr-cols} * @var{word-size}} bytes are read
26897 (returned as @samp{total-bytes}). Should less than the requested number
26898 of bytes be returned by the target, the missing words are identified
26899 using @samp{N/A}. The number of bytes read from the target is returned
26900 in @samp{nr-bytes} and the starting address used to read memory in
26903 The address of the next/previous row or page is available in
26904 @samp{next-row} and @samp{prev-row}, @samp{next-page} and
26907 @subsubheading @value{GDBN} Command
26909 The corresponding @value{GDBN} command is @samp{x}. @code{gdbtk} has
26910 @samp{gdb_get_mem} memory read command.
26912 @subsubheading Example
26914 Read six bytes of memory starting at @code{bytes+6} but then offset by
26915 @code{-6} bytes. Format as three rows of two columns. One byte per
26916 word. Display each word in hex.
26920 9-data-read-memory -o -6 -- bytes+6 x 1 3 2
26921 9^done,addr="0x00001390",nr-bytes="6",total-bytes="6",
26922 next-row="0x00001396",prev-row="0x0000138e",next-page="0x00001396",
26923 prev-page="0x0000138a",memory=[
26924 @{addr="0x00001390",data=["0x00","0x01"]@},
26925 @{addr="0x00001392",data=["0x02","0x03"]@},
26926 @{addr="0x00001394",data=["0x04","0x05"]@}]
26930 Read two bytes of memory starting at address @code{shorts + 64} and
26931 display as a single word formatted in decimal.
26935 5-data-read-memory shorts+64 d 2 1 1
26936 5^done,addr="0x00001510",nr-bytes="2",total-bytes="2",
26937 next-row="0x00001512",prev-row="0x0000150e",
26938 next-page="0x00001512",prev-page="0x0000150e",memory=[
26939 @{addr="0x00001510",data=["128"]@}]
26943 Read thirty two bytes of memory starting at @code{bytes+16} and format
26944 as eight rows of four columns. Include a string encoding with @samp{x}
26945 used as the non-printable character.
26949 4-data-read-memory bytes+16 x 1 8 4 x
26950 4^done,addr="0x000013a0",nr-bytes="32",total-bytes="32",
26951 next-row="0x000013c0",prev-row="0x0000139c",
26952 next-page="0x000013c0",prev-page="0x00001380",memory=[
26953 @{addr="0x000013a0",data=["0x10","0x11","0x12","0x13"],ascii="xxxx"@},
26954 @{addr="0x000013a4",data=["0x14","0x15","0x16","0x17"],ascii="xxxx"@},
26955 @{addr="0x000013a8",data=["0x18","0x19","0x1a","0x1b"],ascii="xxxx"@},
26956 @{addr="0x000013ac",data=["0x1c","0x1d","0x1e","0x1f"],ascii="xxxx"@},
26957 @{addr="0x000013b0",data=["0x20","0x21","0x22","0x23"],ascii=" !\"#"@},
26958 @{addr="0x000013b4",data=["0x24","0x25","0x26","0x27"],ascii="$%&'"@},
26959 @{addr="0x000013b8",data=["0x28","0x29","0x2a","0x2b"],ascii="()*+"@},
26960 @{addr="0x000013bc",data=["0x2c","0x2d","0x2e","0x2f"],ascii=",-./"@}]
26964 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
26965 @node GDB/MI Tracepoint Commands
26966 @section @sc{gdb/mi} Tracepoint Commands
26968 The commands defined in this section implement MI support for
26969 tracepoints. For detailed introduction, see @ref{Tracepoints}.
26971 @subheading The @code{-trace-find} Command
26972 @findex -trace-find
26974 @subsubheading Synopsis
26977 -trace-find @var{mode} [@var{parameters}@dots{}]
26980 Find a trace frame using criteria defined by @var{mode} and
26981 @var{parameters}. The following table lists permissible
26982 modes and their parameters. For details of operation, see @ref{tfind}.
26987 No parameters are required. Stops examining trace frames.
26990 An integer is required as parameter. Selects tracepoint frame with
26993 @item tracepoint-number
26994 An integer is required as parameter. Finds next
26995 trace frame that corresponds to tracepoint with the specified number.
26998 An address is required as parameter. Finds
26999 next trace frame that corresponds to any tracepoint at the specified
27002 @item pc-inside-range
27003 Two addresses are required as parameters. Finds next trace
27004 frame that corresponds to a tracepoint at an address inside the
27005 specified range. Both bounds are considered to be inside the range.
27007 @item pc-outside-range
27008 Two addresses are required as parameters. Finds
27009 next trace frame that corresponds to a tracepoint at an address outside
27010 the specified range. Both bounds are considered to be inside the range.
27013 Line specification is required as parameter. @xref{Specify Location}.
27014 Finds next trace frame that corresponds to a tracepoint at
27015 the specified location.
27019 If @samp{none} was passed as @var{mode}, the response does not
27020 have fields. Otherwise, the response may have the following fields:
27024 This field has either @samp{0} or @samp{1} as the value, depending
27025 on whether a matching tracepoint was found.
27028 The index of the found traceframe. This field is present iff
27029 the @samp{found} field has value of @samp{1}.
27032 The index of the found tracepoint. This field is present iff
27033 the @samp{found} field has value of @samp{1}.
27036 The information about the frame corresponding to the found trace
27037 frame. This field is present only if a trace frame was found.
27038 @xref{GDB/MI Frame Information}, for description of this field.
27042 @subsubheading @value{GDBN} Command
27044 The corresponding @value{GDBN} command is @samp{tfind}.
27046 @subheading -trace-define-variable
27047 @findex -trace-define-variable
27049 @subsubheading Synopsis
27052 -trace-define-variable @var{name} [ @var{value} ]
27055 Create trace variable @var{name} if it does not exist. If
27056 @var{value} is specified, sets the initial value of the specified
27057 trace variable to that value. Note that the @var{name} should start
27058 with the @samp{$} character.
27060 @subsubheading @value{GDBN} Command
27062 The corresponding @value{GDBN} command is @samp{tvariable}.
27064 @subheading -trace-list-variables
27065 @findex -trace-list-variables
27067 @subsubheading Synopsis
27070 -trace-list-variables
27073 Return a table of all defined trace variables. Each element of the
27074 table has the following fields:
27078 The name of the trace variable. This field is always present.
27081 The initial value. This is a 64-bit signed integer. This
27082 field is always present.
27085 The value the trace variable has at the moment. This is a 64-bit
27086 signed integer. This field is absent iff current value is
27087 not defined, for example if the trace was never run, or is
27092 @subsubheading @value{GDBN} Command
27094 The corresponding @value{GDBN} command is @samp{tvariables}.
27096 @subsubheading Example
27100 -trace-list-variables
27101 ^done,trace-variables=@{nr_rows="1",nr_cols="3",
27102 hdr=[@{width="15",alignment="-1",col_name="name",colhdr="Name"@},
27103 @{width="11",alignment="-1",col_name="initial",colhdr="Initial"@},
27104 @{width="11",alignment="-1",col_name="current",colhdr="Current"@}],
27105 body=[variable=@{name="$trace_timestamp",initial="0"@}
27106 variable=@{name="$foo",initial="10",current="15"@}]@}
27110 @subheading -trace-save
27111 @findex -trace-save
27113 @subsubheading Synopsis
27116 -trace-save [-r ] @var{filename}
27119 Saves the collected trace data to @var{filename}. Without the
27120 @samp{-r} option, the data is downloaded from the target and saved
27121 in a local file. With the @samp{-r} option the target is asked
27122 to perform the save.
27124 @subsubheading @value{GDBN} Command
27126 The corresponding @value{GDBN} command is @samp{tsave}.
27129 @subheading -trace-start
27130 @findex -trace-start
27132 @subsubheading Synopsis
27138 Starts a tracing experiments. The result of this command does not
27141 @subsubheading @value{GDBN} Command
27143 The corresponding @value{GDBN} command is @samp{tstart}.
27145 @subheading -trace-status
27146 @findex -trace-status
27148 @subsubheading Synopsis
27154 Obtains the status of a tracing experiment. The result may include
27155 the following fields:
27160 May have a value of either @samp{0}, when no tracing operations are
27161 supported, @samp{1}, when all tracing operations are supported, or
27162 @samp{file} when examining trace file. In the latter case, examining
27163 of trace frame is possible but new tracing experiement cannot be
27164 started. This field is always present.
27167 May have a value of either @samp{0} or @samp{1} depending on whether
27168 tracing experiement is in progress on target. This field is present
27169 if @samp{supported} field is not @samp{0}.
27172 Report the reason why the tracing was stopped last time. This field
27173 may be absent iff tracing was never stopped on target yet. The
27174 value of @samp{request} means the tracing was stopped as result of
27175 the @code{-trace-stop} command. The value of @samp{overflow} means
27176 the tracing buffer is full. The value of @samp{disconnection} means
27177 tracing was automatically stopped when @value{GDBN} has disconnected.
27178 The value of @samp{passcount} means tracing was stopped when a
27179 tracepoint was passed a maximal number of times for that tracepoint.
27180 This field is present if @samp{supported} field is not @samp{0}.
27182 @item stopping-tracepoint
27183 The number of tracepoint whose passcount as exceeded. This field is
27184 present iff the @samp{stop-reason} field has the value of
27188 @itemx frames-created
27189 The @samp{frames} field is a count of the total number of trace frames
27190 in the trace buffer, while @samp{frames-created} is the total created
27191 during the run, including ones that were discarded, such as when a
27192 circular trace buffer filled up. Both fields are optional.
27196 These fields tell the current size of the tracing buffer and the
27197 remaining space. These fields are optional.
27200 The value of the circular trace buffer flag. @code{1} means that the
27201 trace buffer is circular and old trace frames will be discarded if
27202 necessary to make room, @code{0} means that the trace buffer is linear
27206 The value of the disconnected tracing flag. @code{1} means that
27207 tracing will continue after @value{GDBN} disconnects, @code{0} means
27208 that the trace run will stop.
27212 @subsubheading @value{GDBN} Command
27214 The corresponding @value{GDBN} command is @samp{tstatus}.
27216 @subheading -trace-stop
27217 @findex -trace-stop
27219 @subsubheading Synopsis
27225 Stops a tracing experiment. The result of this command has the same
27226 fields as @code{-trace-status}, except that the @samp{supported} and
27227 @samp{running} fields are not output.
27229 @subsubheading @value{GDBN} Command
27231 The corresponding @value{GDBN} command is @samp{tstop}.
27234 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27235 @node GDB/MI Symbol Query
27236 @section @sc{gdb/mi} Symbol Query Commands
27240 @subheading The @code{-symbol-info-address} Command
27241 @findex -symbol-info-address
27243 @subsubheading Synopsis
27246 -symbol-info-address @var{symbol}
27249 Describe where @var{symbol} is stored.
27251 @subsubheading @value{GDBN} Command
27253 The corresponding @value{GDBN} command is @samp{info address}.
27255 @subsubheading Example
27259 @subheading The @code{-symbol-info-file} Command
27260 @findex -symbol-info-file
27262 @subsubheading Synopsis
27268 Show the file for the symbol.
27270 @subsubheading @value{GDBN} Command
27272 There's no equivalent @value{GDBN} command. @code{gdbtk} has
27273 @samp{gdb_find_file}.
27275 @subsubheading Example
27279 @subheading The @code{-symbol-info-function} Command
27280 @findex -symbol-info-function
27282 @subsubheading Synopsis
27285 -symbol-info-function
27288 Show which function the symbol lives in.
27290 @subsubheading @value{GDBN} Command
27292 @samp{gdb_get_function} in @code{gdbtk}.
27294 @subsubheading Example
27298 @subheading The @code{-symbol-info-line} Command
27299 @findex -symbol-info-line
27301 @subsubheading Synopsis
27307 Show the core addresses of the code for a source line.
27309 @subsubheading @value{GDBN} Command
27311 The corresponding @value{GDBN} command is @samp{info line}.
27312 @code{gdbtk} has the @samp{gdb_get_line} and @samp{gdb_get_file} commands.
27314 @subsubheading Example
27318 @subheading The @code{-symbol-info-symbol} Command
27319 @findex -symbol-info-symbol
27321 @subsubheading Synopsis
27324 -symbol-info-symbol @var{addr}
27327 Describe what symbol is at location @var{addr}.
27329 @subsubheading @value{GDBN} Command
27331 The corresponding @value{GDBN} command is @samp{info symbol}.
27333 @subsubheading Example
27337 @subheading The @code{-symbol-list-functions} Command
27338 @findex -symbol-list-functions
27340 @subsubheading Synopsis
27343 -symbol-list-functions
27346 List the functions in the executable.
27348 @subsubheading @value{GDBN} Command
27350 @samp{info functions} in @value{GDBN}, @samp{gdb_listfunc} and
27351 @samp{gdb_search} in @code{gdbtk}.
27353 @subsubheading Example
27358 @subheading The @code{-symbol-list-lines} Command
27359 @findex -symbol-list-lines
27361 @subsubheading Synopsis
27364 -symbol-list-lines @var{filename}
27367 Print the list of lines that contain code and their associated program
27368 addresses for the given source filename. The entries are sorted in
27369 ascending PC order.
27371 @subsubheading @value{GDBN} Command
27373 There is no corresponding @value{GDBN} command.
27375 @subsubheading Example
27378 -symbol-list-lines basics.c
27379 ^done,lines=[@{pc="0x08048554",line="7"@},@{pc="0x0804855a",line="8"@}]
27385 @subheading The @code{-symbol-list-types} Command
27386 @findex -symbol-list-types
27388 @subsubheading Synopsis
27394 List all the type names.
27396 @subsubheading @value{GDBN} Command
27398 The corresponding commands are @samp{info types} in @value{GDBN},
27399 @samp{gdb_search} in @code{gdbtk}.
27401 @subsubheading Example
27405 @subheading The @code{-symbol-list-variables} Command
27406 @findex -symbol-list-variables
27408 @subsubheading Synopsis
27411 -symbol-list-variables
27414 List all the global and static variable names.
27416 @subsubheading @value{GDBN} Command
27418 @samp{info variables} in @value{GDBN}, @samp{gdb_search} in @code{gdbtk}.
27420 @subsubheading Example
27424 @subheading The @code{-symbol-locate} Command
27425 @findex -symbol-locate
27427 @subsubheading Synopsis
27433 @subsubheading @value{GDBN} Command
27435 @samp{gdb_loc} in @code{gdbtk}.
27437 @subsubheading Example
27441 @subheading The @code{-symbol-type} Command
27442 @findex -symbol-type
27444 @subsubheading Synopsis
27447 -symbol-type @var{variable}
27450 Show type of @var{variable}.
27452 @subsubheading @value{GDBN} Command
27454 The corresponding @value{GDBN} command is @samp{ptype}, @code{gdbtk} has
27455 @samp{gdb_obj_variable}.
27457 @subsubheading Example
27462 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27463 @node GDB/MI File Commands
27464 @section @sc{gdb/mi} File Commands
27466 This section describes the GDB/MI commands to specify executable file names
27467 and to read in and obtain symbol table information.
27469 @subheading The @code{-file-exec-and-symbols} Command
27470 @findex -file-exec-and-symbols
27472 @subsubheading Synopsis
27475 -file-exec-and-symbols @var{file}
27478 Specify the executable file to be debugged. This file is the one from
27479 which the symbol table is also read. If no file is specified, the
27480 command clears the executable and symbol information. If breakpoints
27481 are set when using this command with no arguments, @value{GDBN} will produce
27482 error messages. Otherwise, no output is produced, except a completion
27485 @subsubheading @value{GDBN} Command
27487 The corresponding @value{GDBN} command is @samp{file}.
27489 @subsubheading Example
27493 -file-exec-and-symbols /kwikemart/marge/ezannoni/TRUNK/mbx/hello.mbx
27499 @subheading The @code{-file-exec-file} Command
27500 @findex -file-exec-file
27502 @subsubheading Synopsis
27505 -file-exec-file @var{file}
27508 Specify the executable file to be debugged. Unlike
27509 @samp{-file-exec-and-symbols}, the symbol table is @emph{not} read
27510 from this file. If used without argument, @value{GDBN} clears the information
27511 about the executable file. No output is produced, except a completion
27514 @subsubheading @value{GDBN} Command
27516 The corresponding @value{GDBN} command is @samp{exec-file}.
27518 @subsubheading Example
27522 -file-exec-file /kwikemart/marge/ezannoni/TRUNK/mbx/hello.mbx
27529 @subheading The @code{-file-list-exec-sections} Command
27530 @findex -file-list-exec-sections
27532 @subsubheading Synopsis
27535 -file-list-exec-sections
27538 List the sections of the current executable file.
27540 @subsubheading @value{GDBN} Command
27542 The @value{GDBN} command @samp{info file} shows, among the rest, the same
27543 information as this command. @code{gdbtk} has a corresponding command
27544 @samp{gdb_load_info}.
27546 @subsubheading Example
27551 @subheading The @code{-file-list-exec-source-file} Command
27552 @findex -file-list-exec-source-file
27554 @subsubheading Synopsis
27557 -file-list-exec-source-file
27560 List the line number, the current source file, and the absolute path
27561 to the current source file for the current executable. The macro
27562 information field has a value of @samp{1} or @samp{0} depending on
27563 whether or not the file includes preprocessor macro information.
27565 @subsubheading @value{GDBN} Command
27567 The @value{GDBN} equivalent is @samp{info source}
27569 @subsubheading Example
27573 123-file-list-exec-source-file
27574 123^done,line="1",file="foo.c",fullname="/home/bar/foo.c,macro-info="1"
27579 @subheading The @code{-file-list-exec-source-files} Command
27580 @findex -file-list-exec-source-files
27582 @subsubheading Synopsis
27585 -file-list-exec-source-files
27588 List the source files for the current executable.
27590 It will always output the filename, but only when @value{GDBN} can find
27591 the absolute file name of a source file, will it output the fullname.
27593 @subsubheading @value{GDBN} Command
27595 The @value{GDBN} equivalent is @samp{info sources}.
27596 @code{gdbtk} has an analogous command @samp{gdb_listfiles}.
27598 @subsubheading Example
27601 -file-list-exec-source-files
27603 @{file=foo.c,fullname=/home/foo.c@},
27604 @{file=/home/bar.c,fullname=/home/bar.c@},
27605 @{file=gdb_could_not_find_fullpath.c@}]
27610 @subheading The @code{-file-list-shared-libraries} Command
27611 @findex -file-list-shared-libraries
27613 @subsubheading Synopsis
27616 -file-list-shared-libraries
27619 List the shared libraries in the program.
27621 @subsubheading @value{GDBN} Command
27623 The corresponding @value{GDBN} command is @samp{info shared}.
27625 @subsubheading Example
27629 @subheading The @code{-file-list-symbol-files} Command
27630 @findex -file-list-symbol-files
27632 @subsubheading Synopsis
27635 -file-list-symbol-files
27640 @subsubheading @value{GDBN} Command
27642 The corresponding @value{GDBN} command is @samp{info file} (part of it).
27644 @subsubheading Example
27649 @subheading The @code{-file-symbol-file} Command
27650 @findex -file-symbol-file
27652 @subsubheading Synopsis
27655 -file-symbol-file @var{file}
27658 Read symbol table info from the specified @var{file} argument. When
27659 used without arguments, clears @value{GDBN}'s symbol table info. No output is
27660 produced, except for a completion notification.
27662 @subsubheading @value{GDBN} Command
27664 The corresponding @value{GDBN} command is @samp{symbol-file}.
27666 @subsubheading Example
27670 -file-symbol-file /kwikemart/marge/ezannoni/TRUNK/mbx/hello.mbx
27676 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27677 @node GDB/MI Memory Overlay Commands
27678 @section @sc{gdb/mi} Memory Overlay Commands
27680 The memory overlay commands are not implemented.
27682 @c @subheading -overlay-auto
27684 @c @subheading -overlay-list-mapping-state
27686 @c @subheading -overlay-list-overlays
27688 @c @subheading -overlay-map
27690 @c @subheading -overlay-off
27692 @c @subheading -overlay-on
27694 @c @subheading -overlay-unmap
27696 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27697 @node GDB/MI Signal Handling Commands
27698 @section @sc{gdb/mi} Signal Handling Commands
27700 Signal handling commands are not implemented.
27702 @c @subheading -signal-handle
27704 @c @subheading -signal-list-handle-actions
27706 @c @subheading -signal-list-signal-types
27710 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27711 @node GDB/MI Target Manipulation
27712 @section @sc{gdb/mi} Target Manipulation Commands
27715 @subheading The @code{-target-attach} Command
27716 @findex -target-attach
27718 @subsubheading Synopsis
27721 -target-attach @var{pid} | @var{gid} | @var{file}
27724 Attach to a process @var{pid} or a file @var{file} outside of
27725 @value{GDBN}, or a thread group @var{gid}. If attaching to a thread
27726 group, the id previously returned by
27727 @samp{-list-thread-groups --available} must be used.
27729 @subsubheading @value{GDBN} Command
27731 The corresponding @value{GDBN} command is @samp{attach}.
27733 @subsubheading Example
27737 =thread-created,id="1"
27738 *stopped,thread-id="1",frame=@{addr="0xb7f7e410",func="bar",args=[]@}
27744 @subheading The @code{-target-compare-sections} Command
27745 @findex -target-compare-sections
27747 @subsubheading Synopsis
27750 -target-compare-sections [ @var{section} ]
27753 Compare data of section @var{section} on target to the exec file.
27754 Without the argument, all sections are compared.
27756 @subsubheading @value{GDBN} Command
27758 The @value{GDBN} equivalent is @samp{compare-sections}.
27760 @subsubheading Example
27765 @subheading The @code{-target-detach} Command
27766 @findex -target-detach
27768 @subsubheading Synopsis
27771 -target-detach [ @var{pid} | @var{gid} ]
27774 Detach from the remote target which normally resumes its execution.
27775 If either @var{pid} or @var{gid} is specified, detaches from either
27776 the specified process, or specified thread group. There's no output.
27778 @subsubheading @value{GDBN} Command
27780 The corresponding @value{GDBN} command is @samp{detach}.
27782 @subsubheading Example
27792 @subheading The @code{-target-disconnect} Command
27793 @findex -target-disconnect
27795 @subsubheading Synopsis
27801 Disconnect from the remote target. There's no output and the target is
27802 generally not resumed.
27804 @subsubheading @value{GDBN} Command
27806 The corresponding @value{GDBN} command is @samp{disconnect}.
27808 @subsubheading Example
27818 @subheading The @code{-target-download} Command
27819 @findex -target-download
27821 @subsubheading Synopsis
27827 Loads the executable onto the remote target.
27828 It prints out an update message every half second, which includes the fields:
27832 The name of the section.
27834 The size of what has been sent so far for that section.
27836 The size of the section.
27838 The total size of what was sent so far (the current and the previous sections).
27840 The size of the overall executable to download.
27844 Each message is sent as status record (@pxref{GDB/MI Output Syntax, ,
27845 @sc{gdb/mi} Output Syntax}).
27847 In addition, it prints the name and size of the sections, as they are
27848 downloaded. These messages include the following fields:
27852 The name of the section.
27854 The size of the section.
27856 The size of the overall executable to download.
27860 At the end, a summary is printed.
27862 @subsubheading @value{GDBN} Command
27864 The corresponding @value{GDBN} command is @samp{load}.
27866 @subsubheading Example
27868 Note: each status message appears on a single line. Here the messages
27869 have been broken down so that they can fit onto a page.
27874 +download,@{section=".text",section-size="6668",total-size="9880"@}
27875 +download,@{section=".text",section-sent="512",section-size="6668",
27876 total-sent="512",total-size="9880"@}
27877 +download,@{section=".text",section-sent="1024",section-size="6668",
27878 total-sent="1024",total-size="9880"@}
27879 +download,@{section=".text",section-sent="1536",section-size="6668",
27880 total-sent="1536",total-size="9880"@}
27881 +download,@{section=".text",section-sent="2048",section-size="6668",
27882 total-sent="2048",total-size="9880"@}
27883 +download,@{section=".text",section-sent="2560",section-size="6668",
27884 total-sent="2560",total-size="9880"@}
27885 +download,@{section=".text",section-sent="3072",section-size="6668",
27886 total-sent="3072",total-size="9880"@}
27887 +download,@{section=".text",section-sent="3584",section-size="6668",
27888 total-sent="3584",total-size="9880"@}
27889 +download,@{section=".text",section-sent="4096",section-size="6668",
27890 total-sent="4096",total-size="9880"@}
27891 +download,@{section=".text",section-sent="4608",section-size="6668",
27892 total-sent="4608",total-size="9880"@}
27893 +download,@{section=".text",section-sent="5120",section-size="6668",
27894 total-sent="5120",total-size="9880"@}
27895 +download,@{section=".text",section-sent="5632",section-size="6668",
27896 total-sent="5632",total-size="9880"@}
27897 +download,@{section=".text",section-sent="6144",section-size="6668",
27898 total-sent="6144",total-size="9880"@}
27899 +download,@{section=".text",section-sent="6656",section-size="6668",
27900 total-sent="6656",total-size="9880"@}
27901 +download,@{section=".init",section-size="28",total-size="9880"@}
27902 +download,@{section=".fini",section-size="28",total-size="9880"@}
27903 +download,@{section=".data",section-size="3156",total-size="9880"@}
27904 +download,@{section=".data",section-sent="512",section-size="3156",
27905 total-sent="7236",total-size="9880"@}
27906 +download,@{section=".data",section-sent="1024",section-size="3156",
27907 total-sent="7748",total-size="9880"@}
27908 +download,@{section=".data",section-sent="1536",section-size="3156",
27909 total-sent="8260",total-size="9880"@}
27910 +download,@{section=".data",section-sent="2048",section-size="3156",
27911 total-sent="8772",total-size="9880"@}
27912 +download,@{section=".data",section-sent="2560",section-size="3156",
27913 total-sent="9284",total-size="9880"@}
27914 +download,@{section=".data",section-sent="3072",section-size="3156",
27915 total-sent="9796",total-size="9880"@}
27916 ^done,address="0x10004",load-size="9880",transfer-rate="6586",
27923 @subheading The @code{-target-exec-status} Command
27924 @findex -target-exec-status
27926 @subsubheading Synopsis
27929 -target-exec-status
27932 Provide information on the state of the target (whether it is running or
27933 not, for instance).
27935 @subsubheading @value{GDBN} Command
27937 There's no equivalent @value{GDBN} command.
27939 @subsubheading Example
27943 @subheading The @code{-target-list-available-targets} Command
27944 @findex -target-list-available-targets
27946 @subsubheading Synopsis
27949 -target-list-available-targets
27952 List the possible targets to connect to.
27954 @subsubheading @value{GDBN} Command
27956 The corresponding @value{GDBN} command is @samp{help target}.
27958 @subsubheading Example
27962 @subheading The @code{-target-list-current-targets} Command
27963 @findex -target-list-current-targets
27965 @subsubheading Synopsis
27968 -target-list-current-targets
27971 Describe the current target.
27973 @subsubheading @value{GDBN} Command
27975 The corresponding information is printed by @samp{info file} (among
27978 @subsubheading Example
27982 @subheading The @code{-target-list-parameters} Command
27983 @findex -target-list-parameters
27985 @subsubheading Synopsis
27988 -target-list-parameters
27994 @subsubheading @value{GDBN} Command
27998 @subsubheading Example
28002 @subheading The @code{-target-select} Command
28003 @findex -target-select
28005 @subsubheading Synopsis
28008 -target-select @var{type} @var{parameters @dots{}}
28011 Connect @value{GDBN} to the remote target. This command takes two args:
28015 The type of target, for instance @samp{remote}, etc.
28016 @item @var{parameters}
28017 Device names, host names and the like. @xref{Target Commands, ,
28018 Commands for Managing Targets}, for more details.
28021 The output is a connection notification, followed by the address at
28022 which the target program is, in the following form:
28025 ^connected,addr="@var{address}",func="@var{function name}",
28026 args=[@var{arg list}]
28029 @subsubheading @value{GDBN} Command
28031 The corresponding @value{GDBN} command is @samp{target}.
28033 @subsubheading Example
28037 -target-select remote /dev/ttya
28038 ^connected,addr="0xfe00a300",func="??",args=[]
28042 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28043 @node GDB/MI File Transfer Commands
28044 @section @sc{gdb/mi} File Transfer Commands
28047 @subheading The @code{-target-file-put} Command
28048 @findex -target-file-put
28050 @subsubheading Synopsis
28053 -target-file-put @var{hostfile} @var{targetfile}
28056 Copy file @var{hostfile} from the host system (the machine running
28057 @value{GDBN}) to @var{targetfile} on the target system.
28059 @subsubheading @value{GDBN} Command
28061 The corresponding @value{GDBN} command is @samp{remote put}.
28063 @subsubheading Example
28067 -target-file-put localfile remotefile
28073 @subheading The @code{-target-file-get} Command
28074 @findex -target-file-get
28076 @subsubheading Synopsis
28079 -target-file-get @var{targetfile} @var{hostfile}
28082 Copy file @var{targetfile} from the target system to @var{hostfile}
28083 on the host system.
28085 @subsubheading @value{GDBN} Command
28087 The corresponding @value{GDBN} command is @samp{remote get}.
28089 @subsubheading Example
28093 -target-file-get remotefile localfile
28099 @subheading The @code{-target-file-delete} Command
28100 @findex -target-file-delete
28102 @subsubheading Synopsis
28105 -target-file-delete @var{targetfile}
28108 Delete @var{targetfile} from the target system.
28110 @subsubheading @value{GDBN} Command
28112 The corresponding @value{GDBN} command is @samp{remote delete}.
28114 @subsubheading Example
28118 -target-file-delete remotefile
28124 @c %%%%%%%%%%%%%%%%%%%%%%%%%%%% SECTION %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28125 @node GDB/MI Miscellaneous Commands
28126 @section Miscellaneous @sc{gdb/mi} Commands
28128 @c @subheading -gdb-complete
28130 @subheading The @code{-gdb-exit} Command
28133 @subsubheading Synopsis
28139 Exit @value{GDBN} immediately.
28141 @subsubheading @value{GDBN} Command
28143 Approximately corresponds to @samp{quit}.
28145 @subsubheading Example
28155 @subheading The @code{-exec-abort} Command
28156 @findex -exec-abort
28158 @subsubheading Synopsis
28164 Kill the inferior running program.
28166 @subsubheading @value{GDBN} Command
28168 The corresponding @value{GDBN} command is @samp{kill}.
28170 @subsubheading Example
28175 @subheading The @code{-gdb-set} Command
28178 @subsubheading Synopsis
28184 Set an internal @value{GDBN} variable.
28185 @c IS THIS A DOLLAR VARIABLE? OR SOMETHING LIKE ANNOTATE ?????
28187 @subsubheading @value{GDBN} Command
28189 The corresponding @value{GDBN} command is @samp{set}.
28191 @subsubheading Example
28201 @subheading The @code{-gdb-show} Command
28204 @subsubheading Synopsis
28210 Show the current value of a @value{GDBN} variable.
28212 @subsubheading @value{GDBN} Command
28214 The corresponding @value{GDBN} command is @samp{show}.
28216 @subsubheading Example
28225 @c @subheading -gdb-source
28228 @subheading The @code{-gdb-version} Command
28229 @findex -gdb-version
28231 @subsubheading Synopsis
28237 Show version information for @value{GDBN}. Used mostly in testing.
28239 @subsubheading @value{GDBN} Command
28241 The @value{GDBN} equivalent is @samp{show version}. @value{GDBN} by
28242 default shows this information when you start an interactive session.
28244 @subsubheading Example
28246 @c This example modifies the actual output from GDB to avoid overfull
28252 ~Copyright 2000 Free Software Foundation, Inc.
28253 ~GDB is free software, covered by the GNU General Public License, and
28254 ~you are welcome to change it and/or distribute copies of it under
28255 ~ certain conditions.
28256 ~Type "show copying" to see the conditions.
28257 ~There is absolutely no warranty for GDB. Type "show warranty" for
28259 ~This GDB was configured as
28260 "--host=sparc-sun-solaris2.5.1 --target=ppc-eabi".
28265 @subheading The @code{-list-features} Command
28266 @findex -list-features
28268 Returns a list of particular features of the MI protocol that
28269 this version of gdb implements. A feature can be a command,
28270 or a new field in an output of some command, or even an
28271 important bugfix. While a frontend can sometimes detect presence
28272 of a feature at runtime, it is easier to perform detection at debugger
28275 The command returns a list of strings, with each string naming an
28276 available feature. Each returned string is just a name, it does not
28277 have any internal structure. The list of possible feature names
28283 (gdb) -list-features
28284 ^done,result=["feature1","feature2"]
28287 The current list of features is:
28290 @item frozen-varobjs
28291 Indicates presence of the @code{-var-set-frozen} command, as well
28292 as possible presense of the @code{frozen} field in the output
28293 of @code{-varobj-create}.
28294 @item pending-breakpoints
28295 Indicates presence of the @option{-f} option to the @code{-break-insert} command.
28297 Indicates presence of Python scripting support, Python-based
28298 pretty-printing commands, and possible presence of the
28299 @samp{display_hint} field in the output of @code{-var-list-children}
28301 Indicates presence of the @code{-thread-info} command.
28305 @subheading The @code{-list-target-features} Command
28306 @findex -list-target-features
28308 Returns a list of particular features that are supported by the
28309 target. Those features affect the permitted MI commands, but
28310 unlike the features reported by the @code{-list-features} command, the
28311 features depend on which target GDB is using at the moment. Whenever
28312 a target can change, due to commands such as @code{-target-select},
28313 @code{-target-attach} or @code{-exec-run}, the list of target features
28314 may change, and the frontend should obtain it again.
28318 (gdb) -list-features
28319 ^done,result=["async"]
28322 The current list of features is:
28326 Indicates that the target is capable of asynchronous command
28327 execution, which means that @value{GDBN} will accept further commands
28328 while the target is running.
28332 @subheading The @code{-list-thread-groups} Command
28333 @findex -list-thread-groups
28335 @subheading Synopsis
28338 -list-thread-groups [ --available ] [ --recurse 1 ] [ @var{group} ... ]
28341 Lists thread groups (@pxref{Thread groups}). When a single thread
28342 group is passed as the argument, lists the children of that group.
28343 When several thread group are passed, lists information about those
28344 thread groups. Without any parameters, lists information about all
28345 top-level thread groups.
28347 Normally, thread groups that are being debugged are reported.
28348 With the @samp{--available} option, @value{GDBN} reports thread groups
28349 available on the target.
28351 The output of this command may have either a @samp{threads} result or
28352 a @samp{groups} result. The @samp{thread} result has a list of tuples
28353 as value, with each tuple describing a thread (@pxref{GDB/MI Thread
28354 Information}). The @samp{groups} result has a list of tuples as value,
28355 each tuple describing a thread group. If top-level groups are
28356 requested (that is, no parameter is passed), or when several groups
28357 are passed, the output always has a @samp{groups} result. The format
28358 of the @samp{group} result is described below.
28360 To reduce the number of roundtrips it's possible to list thread groups
28361 together with their children, by passing the @samp{--recurse} option
28362 and the recursion depth. Presently, only recursion depth of 1 is
28363 permitted. If this option is present, then every reported thread group
28364 will also include its children, either as @samp{group} or
28365 @samp{threads} field.
28367 In general, any combination of option and parameters is permitted, with
28368 the following caveats:
28372 When a single thread group is passed, the output will typically
28373 be the @samp{threads} result. Because threads may not contain
28374 anything, the @samp{recurse} option will be ignored.
28377 When the @samp{--available} option is passed, limited information may
28378 be available. In particular, the list of threads of a process might
28379 be inaccessible. Further, specifying specific thread groups might
28380 not give any performance advantage over listing all thread groups.
28381 The frontend should assume that @samp{-list-thread-groups --available}
28382 is always an expensive operation and cache the results.
28386 The @samp{groups} result is a list of tuples, where each tuple may
28387 have the following fields:
28391 Identifier of the thread group. This field is always present.
28392 The identifier is an opaque string; frontends should not try to
28393 convert it to an integer, even though it might look like one.
28396 The type of the thread group. At present, only @samp{process} is a
28400 The target-specific process identifier. This field is only present
28401 for thread groups of type @samp{process} and only if the process exists.
28404 The number of children this thread group has. This field may be
28405 absent for an available thread group.
28408 This field has a list of tuples as value, each tuple describing a
28409 thread. It may be present if the @samp{--recurse} option is
28410 specified, and it's actually possible to obtain the threads.
28413 This field is a list of integers, each identifying a core that one
28414 thread of the group is running on. This field may be absent if
28415 such information is not available.
28418 The name of the executable file that corresponds to this thread group.
28419 The field is only present for thread groups of type @samp{process},
28420 and only if there is a corresponding executable file.
28424 @subheading Example
28428 -list-thread-groups
28429 ^done,groups=[@{id="17",type="process",pid="yyy",num_children="2"@}]
28430 -list-thread-groups 17
28431 ^done,threads=[@{id="2",target-id="Thread 0xb7e14b90 (LWP 21257)",
28432 frame=@{level="0",addr="0xffffe410",func="__kernel_vsyscall",args=[]@},state="running"@},
28433 @{id="1",target-id="Thread 0xb7e156b0 (LWP 21254)",
28434 frame=@{level="0",addr="0x0804891f",func="foo",args=[@{name="i",value="10"@}],
28435 file="/tmp/a.c",fullname="/tmp/a.c",line="158"@},state="running"@}]]
28436 -list-thread-groups --available
28437 ^done,groups=[@{id="17",type="process",pid="yyy",num_children="2",cores=[1,2]@}]
28438 -list-thread-groups --available --recurse 1
28439 ^done,groups=[@{id="17", types="process",pid="yyy",num_children="2",cores=[1,2],
28440 threads=[@{id="1",target-id="Thread 0xb7e14b90",cores=[1]@},
28441 @{id="2",target-id="Thread 0xb7e14b90",cores=[2]@}]@},..]
28442 -list-thread-groups --available --recurse 1 17 18
28443 ^done,groups=[@{id="17", types="process",pid="yyy",num_children="2",cores=[1,2],
28444 threads=[@{id="1",target-id="Thread 0xb7e14b90",cores=[1]@},
28445 @{id="2",target-id="Thread 0xb7e14b90",cores=[2]@}]@},...]
28449 @subheading The @code{-add-inferior} Command
28450 @findex -add-inferior
28452 @subheading Synopsis
28458 Creates a new inferior (@pxref{Inferiors and Programs}). The created
28459 inferior is not associated with any executable. Such association may
28460 be established with the @samp{-file-exec-and-symbols} command
28461 (@pxref{GDB/MI File Commands}). The command response has a single
28462 field, @samp{thread-group}, whose value is the identifier of the
28463 thread group corresponding to the new inferior.
28465 @subheading Example
28470 ^done,thread-group="i3"
28473 @subheading The @code{-interpreter-exec} Command
28474 @findex -interpreter-exec
28476 @subheading Synopsis
28479 -interpreter-exec @var{interpreter} @var{command}
28481 @anchor{-interpreter-exec}
28483 Execute the specified @var{command} in the given @var{interpreter}.
28485 @subheading @value{GDBN} Command
28487 The corresponding @value{GDBN} command is @samp{interpreter-exec}.
28489 @subheading Example
28493 -interpreter-exec console "break main"
28494 &"During symbol reading, couldn't parse type; debugger out of date?.\n"
28495 &"During symbol reading, bad structure-type format.\n"
28496 ~"Breakpoint 1 at 0x8074fc6: file ../../src/gdb/main.c, line 743.\n"
28501 @subheading The @code{-inferior-tty-set} Command
28502 @findex -inferior-tty-set
28504 @subheading Synopsis
28507 -inferior-tty-set /dev/pts/1
28510 Set terminal for future runs of the program being debugged.
28512 @subheading @value{GDBN} Command
28514 The corresponding @value{GDBN} command is @samp{set inferior-tty} /dev/pts/1.
28516 @subheading Example
28520 -inferior-tty-set /dev/pts/1
28525 @subheading The @code{-inferior-tty-show} Command
28526 @findex -inferior-tty-show
28528 @subheading Synopsis
28534 Show terminal for future runs of program being debugged.
28536 @subheading @value{GDBN} Command
28538 The corresponding @value{GDBN} command is @samp{show inferior-tty}.
28540 @subheading Example
28544 -inferior-tty-set /dev/pts/1
28548 ^done,inferior_tty_terminal="/dev/pts/1"
28552 @subheading The @code{-enable-timings} Command
28553 @findex -enable-timings
28555 @subheading Synopsis
28558 -enable-timings [yes | no]
28561 Toggle the printing of the wallclock, user and system times for an MI
28562 command as a field in its output. This command is to help frontend
28563 developers optimize the performance of their code. No argument is
28564 equivalent to @samp{yes}.
28566 @subheading @value{GDBN} Command
28570 @subheading Example
28578 ^done,bkpt=@{number="1",type="breakpoint",disp="keep",enabled="y",
28579 addr="0x080484ed",func="main",file="myprog.c",
28580 fullname="/home/nickrob/myprog.c",line="73",times="0"@},
28581 time=@{wallclock="0.05185",user="0.00800",system="0.00000"@}
28589 *stopped,reason="breakpoint-hit",disp="keep",bkptno="1",thread-id="0",
28590 frame=@{addr="0x080484ed",func="main",args=[@{name="argc",value="1"@},
28591 @{name="argv",value="0xbfb60364"@}],file="myprog.c",
28592 fullname="/home/nickrob/myprog.c",line="73"@}
28597 @chapter @value{GDBN} Annotations
28599 This chapter describes annotations in @value{GDBN}. Annotations were
28600 designed to interface @value{GDBN} to graphical user interfaces or other
28601 similar programs which want to interact with @value{GDBN} at a
28602 relatively high level.
28604 The annotation mechanism has largely been superseded by @sc{gdb/mi}
28608 This is Edition @value{EDITION}, @value{DATE}.
28612 * Annotations Overview:: What annotations are; the general syntax.
28613 * Server Prefix:: Issuing a command without affecting user state.
28614 * Prompting:: Annotations marking @value{GDBN}'s need for input.
28615 * Errors:: Annotations for error messages.
28616 * Invalidation:: Some annotations describe things now invalid.
28617 * Annotations for Running::
28618 Whether the program is running, how it stopped, etc.
28619 * Source Annotations:: Annotations describing source code.
28622 @node Annotations Overview
28623 @section What is an Annotation?
28624 @cindex annotations
28626 Annotations start with a newline character, two @samp{control-z}
28627 characters, and the name of the annotation. If there is no additional
28628 information associated with this annotation, the name of the annotation
28629 is followed immediately by a newline. If there is additional
28630 information, the name of the annotation is followed by a space, the
28631 additional information, and a newline. The additional information
28632 cannot contain newline characters.
28634 Any output not beginning with a newline and two @samp{control-z}
28635 characters denotes literal output from @value{GDBN}. Currently there is
28636 no need for @value{GDBN} to output a newline followed by two
28637 @samp{control-z} characters, but if there was such a need, the
28638 annotations could be extended with an @samp{escape} annotation which
28639 means those three characters as output.
28641 The annotation @var{level}, which is specified using the
28642 @option{--annotate} command line option (@pxref{Mode Options}), controls
28643 how much information @value{GDBN} prints together with its prompt,
28644 values of expressions, source lines, and other types of output. Level 0
28645 is for no annotations, level 1 is for use when @value{GDBN} is run as a
28646 subprocess of @sc{gnu} Emacs, level 3 is the maximum annotation suitable
28647 for programs that control @value{GDBN}, and level 2 annotations have
28648 been made obsolete (@pxref{Limitations, , Limitations of the Annotation
28649 Interface, annotate, GDB's Obsolete Annotations}).
28652 @kindex set annotate
28653 @item set annotate @var{level}
28654 The @value{GDBN} command @code{set annotate} sets the level of
28655 annotations to the specified @var{level}.
28657 @item show annotate
28658 @kindex show annotate
28659 Show the current annotation level.
28662 This chapter describes level 3 annotations.
28664 A simple example of starting up @value{GDBN} with annotations is:
28667 $ @kbd{gdb --annotate=3}
28669 Copyright 2003 Free Software Foundation, Inc.
28670 GDB is free software, covered by the GNU General Public License,
28671 and you are welcome to change it and/or distribute copies of it
28672 under certain conditions.
28673 Type "show copying" to see the conditions.
28674 There is absolutely no warranty for GDB. Type "show warranty"
28676 This GDB was configured as "i386-pc-linux-gnu"
28687 Here @samp{quit} is input to @value{GDBN}; the rest is output from
28688 @value{GDBN}. The three lines beginning @samp{^Z^Z} (where @samp{^Z}
28689 denotes a @samp{control-z} character) are annotations; the rest is
28690 output from @value{GDBN}.
28692 @node Server Prefix
28693 @section The Server Prefix
28694 @cindex server prefix
28696 If you prefix a command with @samp{server } then it will not affect
28697 the command history, nor will it affect @value{GDBN}'s notion of which
28698 command to repeat if @key{RET} is pressed on a line by itself. This
28699 means that commands can be run behind a user's back by a front-end in
28700 a transparent manner.
28702 The @code{server } prefix does not affect the recording of values into
28703 the value history; to print a value without recording it into the
28704 value history, use the @code{output} command instead of the
28705 @code{print} command.
28707 Using this prefix also disables confirmation requests
28708 (@pxref{confirmation requests}).
28711 @section Annotation for @value{GDBN} Input
28713 @cindex annotations for prompts
28714 When @value{GDBN} prompts for input, it annotates this fact so it is possible
28715 to know when to send output, when the output from a given command is
28718 Different kinds of input each have a different @dfn{input type}. Each
28719 input type has three annotations: a @code{pre-} annotation, which
28720 denotes the beginning of any prompt which is being output, a plain
28721 annotation, which denotes the end of the prompt, and then a @code{post-}
28722 annotation which denotes the end of any echo which may (or may not) be
28723 associated with the input. For example, the @code{prompt} input type
28724 features the following annotations:
28732 The input types are
28735 @findex pre-prompt annotation
28736 @findex prompt annotation
28737 @findex post-prompt annotation
28739 When @value{GDBN} is prompting for a command (the main @value{GDBN} prompt).
28741 @findex pre-commands annotation
28742 @findex commands annotation
28743 @findex post-commands annotation
28745 When @value{GDBN} prompts for a set of commands, like in the @code{commands}
28746 command. The annotations are repeated for each command which is input.
28748 @findex pre-overload-choice annotation
28749 @findex overload-choice annotation
28750 @findex post-overload-choice annotation
28751 @item overload-choice
28752 When @value{GDBN} wants the user to select between various overloaded functions.
28754 @findex pre-query annotation
28755 @findex query annotation
28756 @findex post-query annotation
28758 When @value{GDBN} wants the user to confirm a potentially dangerous operation.
28760 @findex pre-prompt-for-continue annotation
28761 @findex prompt-for-continue annotation
28762 @findex post-prompt-for-continue annotation
28763 @item prompt-for-continue
28764 When @value{GDBN} is asking the user to press return to continue. Note: Don't
28765 expect this to work well; instead use @code{set height 0} to disable
28766 prompting. This is because the counting of lines is buggy in the
28767 presence of annotations.
28772 @cindex annotations for errors, warnings and interrupts
28774 @findex quit annotation
28779 This annotation occurs right before @value{GDBN} responds to an interrupt.
28781 @findex error annotation
28786 This annotation occurs right before @value{GDBN} responds to an error.
28788 Quit and error annotations indicate that any annotations which @value{GDBN} was
28789 in the middle of may end abruptly. For example, if a
28790 @code{value-history-begin} annotation is followed by a @code{error}, one
28791 cannot expect to receive the matching @code{value-history-end}. One
28792 cannot expect not to receive it either, however; an error annotation
28793 does not necessarily mean that @value{GDBN} is immediately returning all the way
28796 @findex error-begin annotation
28797 A quit or error annotation may be preceded by
28803 Any output between that and the quit or error annotation is the error
28806 Warning messages are not yet annotated.
28807 @c If we want to change that, need to fix warning(), type_error(),
28808 @c range_error(), and possibly other places.
28811 @section Invalidation Notices
28813 @cindex annotations for invalidation messages
28814 The following annotations say that certain pieces of state may have
28818 @findex frames-invalid annotation
28819 @item ^Z^Zframes-invalid
28821 The frames (for example, output from the @code{backtrace} command) may
28824 @findex breakpoints-invalid annotation
28825 @item ^Z^Zbreakpoints-invalid
28827 The breakpoints may have changed. For example, the user just added or
28828 deleted a breakpoint.
28831 @node Annotations for Running
28832 @section Running the Program
28833 @cindex annotations for running programs
28835 @findex starting annotation
28836 @findex stopping annotation
28837 When the program starts executing due to a @value{GDBN} command such as
28838 @code{step} or @code{continue},
28844 is output. When the program stops,
28850 is output. Before the @code{stopped} annotation, a variety of
28851 annotations describe how the program stopped.
28854 @findex exited annotation
28855 @item ^Z^Zexited @var{exit-status}
28856 The program exited, and @var{exit-status} is the exit status (zero for
28857 successful exit, otherwise nonzero).
28859 @findex signalled annotation
28860 @findex signal-name annotation
28861 @findex signal-name-end annotation
28862 @findex signal-string annotation
28863 @findex signal-string-end annotation
28864 @item ^Z^Zsignalled
28865 The program exited with a signal. After the @code{^Z^Zsignalled}, the
28866 annotation continues:
28872 ^Z^Zsignal-name-end
28876 ^Z^Zsignal-string-end
28881 where @var{name} is the name of the signal, such as @code{SIGILL} or
28882 @code{SIGSEGV}, and @var{string} is the explanation of the signal, such
28883 as @code{Illegal Instruction} or @code{Segmentation fault}.
28884 @var{intro-text}, @var{middle-text}, and @var{end-text} are for the
28885 user's benefit and have no particular format.
28887 @findex signal annotation
28889 The syntax of this annotation is just like @code{signalled}, but @value{GDBN} is
28890 just saying that the program received the signal, not that it was
28891 terminated with it.
28893 @findex breakpoint annotation
28894 @item ^Z^Zbreakpoint @var{number}
28895 The program hit breakpoint number @var{number}.
28897 @findex watchpoint annotation
28898 @item ^Z^Zwatchpoint @var{number}
28899 The program hit watchpoint number @var{number}.
28902 @node Source Annotations
28903 @section Displaying Source
28904 @cindex annotations for source display
28906 @findex source annotation
28907 The following annotation is used instead of displaying source code:
28910 ^Z^Zsource @var{filename}:@var{line}:@var{character}:@var{middle}:@var{addr}
28913 where @var{filename} is an absolute file name indicating which source
28914 file, @var{line} is the line number within that file (where 1 is the
28915 first line in the file), @var{character} is the character position
28916 within the file (where 0 is the first character in the file) (for most
28917 debug formats this will necessarily point to the beginning of a line),
28918 @var{middle} is @samp{middle} if @var{addr} is in the middle of the
28919 line, or @samp{beg} if @var{addr} is at the beginning of the line, and
28920 @var{addr} is the address in the target program associated with the
28921 source which is being displayed. @var{addr} is in the form @samp{0x}
28922 followed by one or more lowercase hex digits (note that this does not
28923 depend on the language).
28925 @node JIT Interface
28926 @chapter JIT Compilation Interface
28927 @cindex just-in-time compilation
28928 @cindex JIT compilation interface
28930 This chapter documents @value{GDBN}'s @dfn{just-in-time} (JIT) compilation
28931 interface. A JIT compiler is a program or library that generates native
28932 executable code at runtime and executes it, usually in order to achieve good
28933 performance while maintaining platform independence.
28935 Programs that use JIT compilation are normally difficult to debug because
28936 portions of their code are generated at runtime, instead of being loaded from
28937 object files, which is where @value{GDBN} normally finds the program's symbols
28938 and debug information. In order to debug programs that use JIT compilation,
28939 @value{GDBN} has an interface that allows the program to register in-memory
28940 symbol files with @value{GDBN} at runtime.
28942 If you are using @value{GDBN} to debug a program that uses this interface, then
28943 it should work transparently so long as you have not stripped the binary. If
28944 you are developing a JIT compiler, then the interface is documented in the rest
28945 of this chapter. At this time, the only known client of this interface is the
28948 Broadly speaking, the JIT interface mirrors the dynamic loader interface. The
28949 JIT compiler communicates with @value{GDBN} by writing data into a global
28950 variable and calling a fuction at a well-known symbol. When @value{GDBN}
28951 attaches, it reads a linked list of symbol files from the global variable to
28952 find existing code, and puts a breakpoint in the function so that it can find
28953 out about additional code.
28956 * Declarations:: Relevant C struct declarations
28957 * Registering Code:: Steps to register code
28958 * Unregistering Code:: Steps to unregister code
28962 @section JIT Declarations
28964 These are the relevant struct declarations that a C program should include to
28965 implement the interface:
28975 struct jit_code_entry
28977 struct jit_code_entry *next_entry;
28978 struct jit_code_entry *prev_entry;
28979 const char *symfile_addr;
28980 uint64_t symfile_size;
28983 struct jit_descriptor
28986 /* This type should be jit_actions_t, but we use uint32_t
28987 to be explicit about the bitwidth. */
28988 uint32_t action_flag;
28989 struct jit_code_entry *relevant_entry;
28990 struct jit_code_entry *first_entry;
28993 /* GDB puts a breakpoint in this function. */
28994 void __attribute__((noinline)) __jit_debug_register_code() @{ @};
28996 /* Make sure to specify the version statically, because the
28997 debugger may check the version before we can set it. */
28998 struct jit_descriptor __jit_debug_descriptor = @{ 1, 0, 0, 0 @};
29001 If the JIT is multi-threaded, then it is important that the JIT synchronize any
29002 modifications to this global data properly, which can easily be done by putting
29003 a global mutex around modifications to these structures.
29005 @node Registering Code
29006 @section Registering Code
29008 To register code with @value{GDBN}, the JIT should follow this protocol:
29012 Generate an object file in memory with symbols and other desired debug
29013 information. The file must include the virtual addresses of the sections.
29016 Create a code entry for the file, which gives the start and size of the symbol
29020 Add it to the linked list in the JIT descriptor.
29023 Point the relevant_entry field of the descriptor at the entry.
29026 Set @code{action_flag} to @code{JIT_REGISTER} and call
29027 @code{__jit_debug_register_code}.
29030 When @value{GDBN} is attached and the breakpoint fires, @value{GDBN} uses the
29031 @code{relevant_entry} pointer so it doesn't have to walk the list looking for
29032 new code. However, the linked list must still be maintained in order to allow
29033 @value{GDBN} to attach to a running process and still find the symbol files.
29035 @node Unregistering Code
29036 @section Unregistering Code
29038 If code is freed, then the JIT should use the following protocol:
29042 Remove the code entry corresponding to the code from the linked list.
29045 Point the @code{relevant_entry} field of the descriptor at the code entry.
29048 Set @code{action_flag} to @code{JIT_UNREGISTER} and call
29049 @code{__jit_debug_register_code}.
29052 If the JIT frees or recompiles code without unregistering it, then @value{GDBN}
29053 and the JIT will leak the memory used for the associated symbol files.
29056 @chapter Reporting Bugs in @value{GDBN}
29057 @cindex bugs in @value{GDBN}
29058 @cindex reporting bugs in @value{GDBN}
29060 Your bug reports play an essential role in making @value{GDBN} reliable.
29062 Reporting a bug may help you by bringing a solution to your problem, or it
29063 may not. But in any case the principal function of a bug report is to help
29064 the entire community by making the next version of @value{GDBN} work better. Bug
29065 reports are your contribution to the maintenance of @value{GDBN}.
29067 In order for a bug report to serve its purpose, you must include the
29068 information that enables us to fix the bug.
29071 * Bug Criteria:: Have you found a bug?
29072 * Bug Reporting:: How to report bugs
29076 @section Have You Found a Bug?
29077 @cindex bug criteria
29079 If you are not sure whether you have found a bug, here are some guidelines:
29082 @cindex fatal signal
29083 @cindex debugger crash
29084 @cindex crash of debugger
29086 If the debugger gets a fatal signal, for any input whatever, that is a
29087 @value{GDBN} bug. Reliable debuggers never crash.
29089 @cindex error on valid input
29091 If @value{GDBN} produces an error message for valid input, that is a
29092 bug. (Note that if you're cross debugging, the problem may also be
29093 somewhere in the connection to the target.)
29095 @cindex invalid input
29097 If @value{GDBN} does not produce an error message for invalid input,
29098 that is a bug. However, you should note that your idea of
29099 ``invalid input'' might be our idea of ``an extension'' or ``support
29100 for traditional practice''.
29103 If you are an experienced user of debugging tools, your suggestions
29104 for improvement of @value{GDBN} are welcome in any case.
29107 @node Bug Reporting
29108 @section How to Report Bugs
29109 @cindex bug reports
29110 @cindex @value{GDBN} bugs, reporting
29112 A number of companies and individuals offer support for @sc{gnu} products.
29113 If you obtained @value{GDBN} from a support organization, we recommend you
29114 contact that organization first.
29116 You can find contact information for many support companies and
29117 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
29119 @c should add a web page ref...
29122 @ifset BUGURL_DEFAULT
29123 In any event, we also recommend that you submit bug reports for
29124 @value{GDBN}. The preferred method is to submit them directly using
29125 @uref{http://www.gnu.org/software/gdb/bugs/, @value{GDBN}'s Bugs web
29126 page}. Alternatively, the @email{bug-gdb@@gnu.org, e-mail gateway} can
29129 @strong{Do not send bug reports to @samp{info-gdb}, or to
29130 @samp{help-gdb}, or to any newsgroups.} Most users of @value{GDBN} do
29131 not want to receive bug reports. Those that do have arranged to receive
29134 The mailing list @samp{bug-gdb} has a newsgroup @samp{gnu.gdb.bug} which
29135 serves as a repeater. The mailing list and the newsgroup carry exactly
29136 the same messages. Often people think of posting bug reports to the
29137 newsgroup instead of mailing them. This appears to work, but it has one
29138 problem which can be crucial: a newsgroup posting often lacks a mail
29139 path back to the sender. Thus, if we need to ask for more information,
29140 we may be unable to reach you. For this reason, it is better to send
29141 bug reports to the mailing list.
29143 @ifclear BUGURL_DEFAULT
29144 In any event, we also recommend that you submit bug reports for
29145 @value{GDBN} to @value{BUGURL}.
29149 The fundamental principle of reporting bugs usefully is this:
29150 @strong{report all the facts}. If you are not sure whether to state a
29151 fact or leave it out, state it!
29153 Often people omit facts because they think they know what causes the
29154 problem and assume that some details do not matter. Thus, you might
29155 assume that the name of the variable you use in an example does not matter.
29156 Well, probably it does not, but one cannot be sure. Perhaps the bug is a
29157 stray memory reference which happens to fetch from the location where that
29158 name is stored in memory; perhaps, if the name were different, the contents
29159 of that location would fool the debugger into doing the right thing despite
29160 the bug. Play it safe and give a specific, complete example. That is the
29161 easiest thing for you to do, and the most helpful.
29163 Keep in mind that the purpose of a bug report is to enable us to fix the
29164 bug. It may be that the bug has been reported previously, but neither
29165 you nor we can know that unless your bug report is complete and
29168 Sometimes people give a few sketchy facts and ask, ``Does this ring a
29169 bell?'' Those bug reports are useless, and we urge everyone to
29170 @emph{refuse to respond to them} except to chide the sender to report
29173 To enable us to fix the bug, you should include all these things:
29177 The version of @value{GDBN}. @value{GDBN} announces it if you start
29178 with no arguments; you can also print it at any time using @code{show
29181 Without this, we will not know whether there is any point in looking for
29182 the bug in the current version of @value{GDBN}.
29185 The type of machine you are using, and the operating system name and
29189 What compiler (and its version) was used to compile @value{GDBN}---e.g.@:
29190 ``@value{GCC}--2.8.1''.
29193 What compiler (and its version) was used to compile the program you are
29194 debugging---e.g.@: ``@value{GCC}--2.8.1'', or ``HP92453-01 A.10.32.03 HP
29195 C Compiler''. For @value{NGCC}, you can say @kbd{@value{GCC} --version}
29196 to get this information; for other compilers, see the documentation for
29200 The command arguments you gave the compiler to compile your example and
29201 observe the bug. For example, did you use @samp{-O}? To guarantee
29202 you will not omit something important, list them all. A copy of the
29203 Makefile (or the output from make) is sufficient.
29205 If we were to try to guess the arguments, we would probably guess wrong
29206 and then we might not encounter the bug.
29209 A complete input script, and all necessary source files, that will
29213 A description of what behavior you observe that you believe is
29214 incorrect. For example, ``It gets a fatal signal.''
29216 Of course, if the bug is that @value{GDBN} gets a fatal signal, then we
29217 will certainly notice it. But if the bug is incorrect output, we might
29218 not notice unless it is glaringly wrong. You might as well not give us
29219 a chance to make a mistake.
29221 Even if the problem you experience is a fatal signal, you should still
29222 say so explicitly. Suppose something strange is going on, such as, your
29223 copy of @value{GDBN} is out of synch, or you have encountered a bug in
29224 the C library on your system. (This has happened!) Your copy might
29225 crash and ours would not. If you told us to expect a crash, then when
29226 ours fails to crash, we would know that the bug was not happening for
29227 us. If you had not told us to expect a crash, then we would not be able
29228 to draw any conclusion from our observations.
29231 @cindex recording a session script
29232 To collect all this information, you can use a session recording program
29233 such as @command{script}, which is available on many Unix systems.
29234 Just run your @value{GDBN} session inside @command{script} and then
29235 include the @file{typescript} file with your bug report.
29237 Another way to record a @value{GDBN} session is to run @value{GDBN}
29238 inside Emacs and then save the entire buffer to a file.
29241 If you wish to suggest changes to the @value{GDBN} source, send us context
29242 diffs. If you even discuss something in the @value{GDBN} source, refer to
29243 it by context, not by line number.
29245 The line numbers in our development sources will not match those in your
29246 sources. Your line numbers would convey no useful information to us.
29250 Here are some things that are not necessary:
29254 A description of the envelope of the bug.
29256 Often people who encounter a bug spend a lot of time investigating
29257 which changes to the input file will make the bug go away and which
29258 changes will not affect it.
29260 This is often time consuming and not very useful, because the way we
29261 will find the bug is by running a single example under the debugger
29262 with breakpoints, not by pure deduction from a series of examples.
29263 We recommend that you save your time for something else.
29265 Of course, if you can find a simpler example to report @emph{instead}
29266 of the original one, that is a convenience for us. Errors in the
29267 output will be easier to spot, running under the debugger will take
29268 less time, and so on.
29270 However, simplification is not vital; if you do not want to do this,
29271 report the bug anyway and send us the entire test case you used.
29274 A patch for the bug.
29276 A patch for the bug does help us if it is a good one. But do not omit
29277 the necessary information, such as the test case, on the assumption that
29278 a patch is all we need. We might see problems with your patch and decide
29279 to fix the problem another way, or we might not understand it at all.
29281 Sometimes with a program as complicated as @value{GDBN} it is very hard to
29282 construct an example that will make the program follow a certain path
29283 through the code. If you do not send us the example, we will not be able
29284 to construct one, so we will not be able to verify that the bug is fixed.
29286 And if we cannot understand what bug you are trying to fix, or why your
29287 patch should be an improvement, we will not install it. A test case will
29288 help us to understand.
29291 A guess about what the bug is or what it depends on.
29293 Such guesses are usually wrong. Even we cannot guess right about such
29294 things without first using the debugger to find the facts.
29297 @c The readline documentation is distributed with the readline code
29298 @c and consists of the two following files:
29300 @c inc-hist.texinfo
29301 @c Use -I with makeinfo to point to the appropriate directory,
29302 @c environment var TEXINPUTS with TeX.
29303 @include rluser.texi
29304 @include inc-hist.texinfo
29307 @node Formatting Documentation
29308 @appendix Formatting Documentation
29310 @cindex @value{GDBN} reference card
29311 @cindex reference card
29312 The @value{GDBN} 4 release includes an already-formatted reference card, ready
29313 for printing with PostScript or Ghostscript, in the @file{gdb}
29314 subdirectory of the main source directory@footnote{In
29315 @file{gdb-@value{GDBVN}/gdb/refcard.ps} of the version @value{GDBVN}
29316 release.}. If you can use PostScript or Ghostscript with your printer,
29317 you can print the reference card immediately with @file{refcard.ps}.
29319 The release also includes the source for the reference card. You
29320 can format it, using @TeX{}, by typing:
29326 The @value{GDBN} reference card is designed to print in @dfn{landscape}
29327 mode on US ``letter'' size paper;
29328 that is, on a sheet 11 inches wide by 8.5 inches
29329 high. You will need to specify this form of printing as an option to
29330 your @sc{dvi} output program.
29332 @cindex documentation
29334 All the documentation for @value{GDBN} comes as part of the machine-readable
29335 distribution. The documentation is written in Texinfo format, which is
29336 a documentation system that uses a single source file to produce both
29337 on-line information and a printed manual. You can use one of the Info
29338 formatting commands to create the on-line version of the documentation
29339 and @TeX{} (or @code{texi2roff}) to typeset the printed version.
29341 @value{GDBN} includes an already formatted copy of the on-line Info
29342 version of this manual in the @file{gdb} subdirectory. The main Info
29343 file is @file{gdb-@value{GDBVN}/gdb/gdb.info}, and it refers to
29344 subordinate files matching @samp{gdb.info*} in the same directory. If
29345 necessary, you can print out these files, or read them with any editor;
29346 but they are easier to read using the @code{info} subsystem in @sc{gnu}
29347 Emacs or the standalone @code{info} program, available as part of the
29348 @sc{gnu} Texinfo distribution.
29350 If you want to format these Info files yourself, you need one of the
29351 Info formatting programs, such as @code{texinfo-format-buffer} or
29354 If you have @code{makeinfo} installed, and are in the top level
29355 @value{GDBN} source directory (@file{gdb-@value{GDBVN}}, in the case of
29356 version @value{GDBVN}), you can make the Info file by typing:
29363 If you want to typeset and print copies of this manual, you need @TeX{},
29364 a program to print its @sc{dvi} output files, and @file{texinfo.tex}, the
29365 Texinfo definitions file.
29367 @TeX{} is a typesetting program; it does not print files directly, but
29368 produces output files called @sc{dvi} files. To print a typeset
29369 document, you need a program to print @sc{dvi} files. If your system
29370 has @TeX{} installed, chances are it has such a program. The precise
29371 command to use depends on your system; @kbd{lpr -d} is common; another
29372 (for PostScript devices) is @kbd{dvips}. The @sc{dvi} print command may
29373 require a file name without any extension or a @samp{.dvi} extension.
29375 @TeX{} also requires a macro definitions file called
29376 @file{texinfo.tex}. This file tells @TeX{} how to typeset a document
29377 written in Texinfo format. On its own, @TeX{} cannot either read or
29378 typeset a Texinfo file. @file{texinfo.tex} is distributed with GDB
29379 and is located in the @file{gdb-@var{version-number}/texinfo}
29382 If you have @TeX{} and a @sc{dvi} printer program installed, you can
29383 typeset and print this manual. First switch to the @file{gdb}
29384 subdirectory of the main source directory (for example, to
29385 @file{gdb-@value{GDBVN}/gdb}) and type:
29391 Then give @file{gdb.dvi} to your @sc{dvi} printing program.
29393 @node Installing GDB
29394 @appendix Installing @value{GDBN}
29395 @cindex installation
29398 * Requirements:: Requirements for building @value{GDBN}
29399 * Running Configure:: Invoking the @value{GDBN} @file{configure} script
29400 * Separate Objdir:: Compiling @value{GDBN} in another directory
29401 * Config Names:: Specifying names for hosts and targets
29402 * Configure Options:: Summary of options for configure
29403 * System-wide configuration:: Having a system-wide init file
29407 @section Requirements for Building @value{GDBN}
29408 @cindex building @value{GDBN}, requirements for
29410 Building @value{GDBN} requires various tools and packages to be available.
29411 Other packages will be used only if they are found.
29413 @heading Tools/Packages Necessary for Building @value{GDBN}
29415 @item ISO C90 compiler
29416 @value{GDBN} is written in ISO C90. It should be buildable with any
29417 working C90 compiler, e.g.@: GCC.
29421 @heading Tools/Packages Optional for Building @value{GDBN}
29425 @value{GDBN} can use the Expat XML parsing library. This library may be
29426 included with your operating system distribution; if it is not, you
29427 can get the latest version from @url{http://expat.sourceforge.net}.
29428 The @file{configure} script will search for this library in several
29429 standard locations; if it is installed in an unusual path, you can
29430 use the @option{--with-libexpat-prefix} option to specify its location.
29436 Remote protocol memory maps (@pxref{Memory Map Format})
29438 Target descriptions (@pxref{Target Descriptions})
29440 Remote shared library lists (@pxref{Library List Format})
29442 MS-Windows shared libraries (@pxref{Shared Libraries})
29446 @cindex compressed debug sections
29447 @value{GDBN} will use the @samp{zlib} library, if available, to read
29448 compressed debug sections. Some linkers, such as GNU gold, are capable
29449 of producing binaries with compressed debug sections. If @value{GDBN}
29450 is compiled with @samp{zlib}, it will be able to read the debug
29451 information in such binaries.
29453 The @samp{zlib} library is likely included with your operating system
29454 distribution; if it is not, you can get the latest version from
29455 @url{http://zlib.net}.
29458 @value{GDBN}'s features related to character sets (@pxref{Character
29459 Sets}) require a functioning @code{iconv} implementation. If you are
29460 on a GNU system, then this is provided by the GNU C Library. Some
29461 other systems also provide a working @code{iconv}.
29463 On systems with @code{iconv}, you can install GNU Libiconv. If you
29464 have previously installed Libiconv, you can use the
29465 @option{--with-libiconv-prefix} option to configure.
29467 @value{GDBN}'s top-level @file{configure} and @file{Makefile} will
29468 arrange to build Libiconv if a directory named @file{libiconv} appears
29469 in the top-most source directory. If Libiconv is built this way, and
29470 if the operating system does not provide a suitable @code{iconv}
29471 implementation, then the just-built library will automatically be used
29472 by @value{GDBN}. One easy way to set this up is to download GNU
29473 Libiconv, unpack it, and then rename the directory holding the
29474 Libiconv source code to @samp{libiconv}.
29477 @node Running Configure
29478 @section Invoking the @value{GDBN} @file{configure} Script
29479 @cindex configuring @value{GDBN}
29480 @value{GDBN} comes with a @file{configure} script that automates the process
29481 of preparing @value{GDBN} for installation; you can then use @code{make} to
29482 build the @code{gdb} program.
29484 @c irrelevant in info file; it's as current as the code it lives with.
29485 @footnote{If you have a more recent version of @value{GDBN} than @value{GDBVN},
29486 look at the @file{README} file in the sources; we may have improved the
29487 installation procedures since publishing this manual.}
29490 The @value{GDBN} distribution includes all the source code you need for
29491 @value{GDBN} in a single directory, whose name is usually composed by
29492 appending the version number to @samp{gdb}.
29494 For example, the @value{GDBN} version @value{GDBVN} distribution is in the
29495 @file{gdb-@value{GDBVN}} directory. That directory contains:
29498 @item gdb-@value{GDBVN}/configure @r{(and supporting files)}
29499 script for configuring @value{GDBN} and all its supporting libraries
29501 @item gdb-@value{GDBVN}/gdb
29502 the source specific to @value{GDBN} itself
29504 @item gdb-@value{GDBVN}/bfd
29505 source for the Binary File Descriptor library
29507 @item gdb-@value{GDBVN}/include
29508 @sc{gnu} include files
29510 @item gdb-@value{GDBVN}/libiberty
29511 source for the @samp{-liberty} free software library
29513 @item gdb-@value{GDBVN}/opcodes
29514 source for the library of opcode tables and disassemblers
29516 @item gdb-@value{GDBVN}/readline
29517 source for the @sc{gnu} command-line interface
29519 @item gdb-@value{GDBVN}/glob
29520 source for the @sc{gnu} filename pattern-matching subroutine
29522 @item gdb-@value{GDBVN}/mmalloc
29523 source for the @sc{gnu} memory-mapped malloc package
29526 The simplest way to configure and build @value{GDBN} is to run @file{configure}
29527 from the @file{gdb-@var{version-number}} source directory, which in
29528 this example is the @file{gdb-@value{GDBVN}} directory.
29530 First switch to the @file{gdb-@var{version-number}} source directory
29531 if you are not already in it; then run @file{configure}. Pass the
29532 identifier for the platform on which @value{GDBN} will run as an
29538 cd gdb-@value{GDBVN}
29539 ./configure @var{host}
29544 where @var{host} is an identifier such as @samp{sun4} or
29545 @samp{decstation}, that identifies the platform where @value{GDBN} will run.
29546 (You can often leave off @var{host}; @file{configure} tries to guess the
29547 correct value by examining your system.)
29549 Running @samp{configure @var{host}} and then running @code{make} builds the
29550 @file{bfd}, @file{readline}, @file{mmalloc}, and @file{libiberty}
29551 libraries, then @code{gdb} itself. The configured source files, and the
29552 binaries, are left in the corresponding source directories.
29555 @file{configure} is a Bourne-shell (@code{/bin/sh}) script; if your
29556 system does not recognize this automatically when you run a different
29557 shell, you may need to run @code{sh} on it explicitly:
29560 sh configure @var{host}
29563 If you run @file{configure} from a directory that contains source
29564 directories for multiple libraries or programs, such as the
29565 @file{gdb-@value{GDBVN}} source directory for version @value{GDBVN},
29567 creates configuration files for every directory level underneath (unless
29568 you tell it not to, with the @samp{--norecursion} option).
29570 You should run the @file{configure} script from the top directory in the
29571 source tree, the @file{gdb-@var{version-number}} directory. If you run
29572 @file{configure} from one of the subdirectories, you will configure only
29573 that subdirectory. That is usually not what you want. In particular,
29574 if you run the first @file{configure} from the @file{gdb} subdirectory
29575 of the @file{gdb-@var{version-number}} directory, you will omit the
29576 configuration of @file{bfd}, @file{readline}, and other sibling
29577 directories of the @file{gdb} subdirectory. This leads to build errors
29578 about missing include files such as @file{bfd/bfd.h}.
29580 You can install @code{@value{GDBP}} anywhere; it has no hardwired paths.
29581 However, you should make sure that the shell on your path (named by
29582 the @samp{SHELL} environment variable) is publicly readable. Remember
29583 that @value{GDBN} uses the shell to start your program---some systems refuse to
29584 let @value{GDBN} debug child processes whose programs are not readable.
29586 @node Separate Objdir
29587 @section Compiling @value{GDBN} in Another Directory
29589 If you want to run @value{GDBN} versions for several host or target machines,
29590 you need a different @code{gdb} compiled for each combination of
29591 host and target. @file{configure} is designed to make this easy by
29592 allowing you to generate each configuration in a separate subdirectory,
29593 rather than in the source directory. If your @code{make} program
29594 handles the @samp{VPATH} feature (@sc{gnu} @code{make} does), running
29595 @code{make} in each of these directories builds the @code{gdb}
29596 program specified there.
29598 To build @code{gdb} in a separate directory, run @file{configure}
29599 with the @samp{--srcdir} option to specify where to find the source.
29600 (You also need to specify a path to find @file{configure}
29601 itself from your working directory. If the path to @file{configure}
29602 would be the same as the argument to @samp{--srcdir}, you can leave out
29603 the @samp{--srcdir} option; it is assumed.)
29605 For example, with version @value{GDBVN}, you can build @value{GDBN} in a
29606 separate directory for a Sun 4 like this:
29610 cd gdb-@value{GDBVN}
29613 ../gdb-@value{GDBVN}/configure sun4
29618 When @file{configure} builds a configuration using a remote source
29619 directory, it creates a tree for the binaries with the same structure
29620 (and using the same names) as the tree under the source directory. In
29621 the example, you'd find the Sun 4 library @file{libiberty.a} in the
29622 directory @file{gdb-sun4/libiberty}, and @value{GDBN} itself in
29623 @file{gdb-sun4/gdb}.
29625 Make sure that your path to the @file{configure} script has just one
29626 instance of @file{gdb} in it. If your path to @file{configure} looks
29627 like @file{../gdb-@value{GDBVN}/gdb/configure}, you are configuring only
29628 one subdirectory of @value{GDBN}, not the whole package. This leads to
29629 build errors about missing include files such as @file{bfd/bfd.h}.
29631 One popular reason to build several @value{GDBN} configurations in separate
29632 directories is to configure @value{GDBN} for cross-compiling (where
29633 @value{GDBN} runs on one machine---the @dfn{host}---while debugging
29634 programs that run on another machine---the @dfn{target}).
29635 You specify a cross-debugging target by
29636 giving the @samp{--target=@var{target}} option to @file{configure}.
29638 When you run @code{make} to build a program or library, you must run
29639 it in a configured directory---whatever directory you were in when you
29640 called @file{configure} (or one of its subdirectories).
29642 The @code{Makefile} that @file{configure} generates in each source
29643 directory also runs recursively. If you type @code{make} in a source
29644 directory such as @file{gdb-@value{GDBVN}} (or in a separate configured
29645 directory configured with @samp{--srcdir=@var{dirname}/gdb-@value{GDBVN}}), you
29646 will build all the required libraries, and then build GDB.
29648 When you have multiple hosts or targets configured in separate
29649 directories, you can run @code{make} on them in parallel (for example,
29650 if they are NFS-mounted on each of the hosts); they will not interfere
29654 @section Specifying Names for Hosts and Targets
29656 The specifications used for hosts and targets in the @file{configure}
29657 script are based on a three-part naming scheme, but some short predefined
29658 aliases are also supported. The full naming scheme encodes three pieces
29659 of information in the following pattern:
29662 @var{architecture}-@var{vendor}-@var{os}
29665 For example, you can use the alias @code{sun4} as a @var{host} argument,
29666 or as the value for @var{target} in a @code{--target=@var{target}}
29667 option. The equivalent full name is @samp{sparc-sun-sunos4}.
29669 The @file{configure} script accompanying @value{GDBN} does not provide
29670 any query facility to list all supported host and target names or
29671 aliases. @file{configure} calls the Bourne shell script
29672 @code{config.sub} to map abbreviations to full names; you can read the
29673 script, if you wish, or you can use it to test your guesses on
29674 abbreviations---for example:
29677 % sh config.sub i386-linux
29679 % sh config.sub alpha-linux
29680 alpha-unknown-linux-gnu
29681 % sh config.sub hp9k700
29683 % sh config.sub sun4
29684 sparc-sun-sunos4.1.1
29685 % sh config.sub sun3
29686 m68k-sun-sunos4.1.1
29687 % sh config.sub i986v
29688 Invalid configuration `i986v': machine `i986v' not recognized
29692 @code{config.sub} is also distributed in the @value{GDBN} source
29693 directory (@file{gdb-@value{GDBVN}}, for version @value{GDBVN}).
29695 @node Configure Options
29696 @section @file{configure} Options
29698 Here is a summary of the @file{configure} options and arguments that
29699 are most often useful for building @value{GDBN}. @file{configure} also has
29700 several other options not listed here. @inforef{What Configure
29701 Does,,configure.info}, for a full explanation of @file{configure}.
29704 configure @r{[}--help@r{]}
29705 @r{[}--prefix=@var{dir}@r{]}
29706 @r{[}--exec-prefix=@var{dir}@r{]}
29707 @r{[}--srcdir=@var{dirname}@r{]}
29708 @r{[}--norecursion@r{]} @r{[}--rm@r{]}
29709 @r{[}--target=@var{target}@r{]}
29714 You may introduce options with a single @samp{-} rather than
29715 @samp{--} if you prefer; but you may abbreviate option names if you use
29720 Display a quick summary of how to invoke @file{configure}.
29722 @item --prefix=@var{dir}
29723 Configure the source to install programs and files under directory
29726 @item --exec-prefix=@var{dir}
29727 Configure the source to install programs under directory
29730 @c avoid splitting the warning from the explanation:
29732 @item --srcdir=@var{dirname}
29733 @strong{Warning: using this option requires @sc{gnu} @code{make}, or another
29734 @code{make} that implements the @code{VPATH} feature.}@*
29735 Use this option to make configurations in directories separate from the
29736 @value{GDBN} source directories. Among other things, you can use this to
29737 build (or maintain) several configurations simultaneously, in separate
29738 directories. @file{configure} writes configuration-specific files in
29739 the current directory, but arranges for them to use the source in the
29740 directory @var{dirname}. @file{configure} creates directories under
29741 the working directory in parallel to the source directories below
29744 @item --norecursion
29745 Configure only the directory level where @file{configure} is executed; do not
29746 propagate configuration to subdirectories.
29748 @item --target=@var{target}
29749 Configure @value{GDBN} for cross-debugging programs running on the specified
29750 @var{target}. Without this option, @value{GDBN} is configured to debug
29751 programs that run on the same machine (@var{host}) as @value{GDBN} itself.
29753 There is no convenient way to generate a list of all available targets.
29755 @item @var{host} @dots{}
29756 Configure @value{GDBN} to run on the specified @var{host}.
29758 There is no convenient way to generate a list of all available hosts.
29761 There are many other options available as well, but they are generally
29762 needed for special purposes only.
29764 @node System-wide configuration
29765 @section System-wide configuration and settings
29766 @cindex system-wide init file
29768 @value{GDBN} can be configured to have a system-wide init file;
29769 this file will be read and executed at startup (@pxref{Startup, , What
29770 @value{GDBN} does during startup}).
29772 Here is the corresponding configure option:
29775 @item --with-system-gdbinit=@var{file}
29776 Specify that the default location of the system-wide init file is
29780 If @value{GDBN} has been configured with the option @option{--prefix=$prefix},
29781 it may be subject to relocation. Two possible cases:
29785 If the default location of this init file contains @file{$prefix},
29786 it will be subject to relocation. Suppose that the configure options
29787 are @option{--prefix=$prefix --with-system-gdbinit=$prefix/etc/gdbinit};
29788 if @value{GDBN} is moved from @file{$prefix} to @file{$install}, the system
29789 init file is looked for as @file{$install/etc/gdbinit} instead of
29790 @file{$prefix/etc/gdbinit}.
29793 By contrast, if the default location does not contain the prefix,
29794 it will not be relocated. E.g.@: if @value{GDBN} has been configured with
29795 @option{--prefix=/usr/local --with-system-gdbinit=/usr/share/gdb/gdbinit},
29796 then @value{GDBN} will always look for @file{/usr/share/gdb/gdbinit},
29797 wherever @value{GDBN} is installed.
29800 @node Maintenance Commands
29801 @appendix Maintenance Commands
29802 @cindex maintenance commands
29803 @cindex internal commands
29805 In addition to commands intended for @value{GDBN} users, @value{GDBN}
29806 includes a number of commands intended for @value{GDBN} developers,
29807 that are not documented elsewhere in this manual. These commands are
29808 provided here for reference. (For commands that turn on debugging
29809 messages, see @ref{Debugging Output}.)
29812 @kindex maint agent
29813 @kindex maint agent-eval
29814 @item maint agent @var{expression}
29815 @itemx maint agent-eval @var{expression}
29816 Translate the given @var{expression} into remote agent bytecodes.
29817 This command is useful for debugging the Agent Expression mechanism
29818 (@pxref{Agent Expressions}). The @samp{agent} version produces an
29819 expression useful for data collection, such as by tracepoints, while
29820 @samp{maint agent-eval} produces an expression that evaluates directly
29821 to a result. For instance, a collection expression for @code{globa +
29822 globb} will include bytecodes to record four bytes of memory at each
29823 of the addresses of @code{globa} and @code{globb}, while discarding
29824 the result of the addition, while an evaluation expression will do the
29825 addition and return the sum.
29827 @kindex maint info breakpoints
29828 @item @anchor{maint info breakpoints}maint info breakpoints
29829 Using the same format as @samp{info breakpoints}, display both the
29830 breakpoints you've set explicitly, and those @value{GDBN} is using for
29831 internal purposes. Internal breakpoints are shown with negative
29832 breakpoint numbers. The type column identifies what kind of breakpoint
29837 Normal, explicitly set breakpoint.
29840 Normal, explicitly set watchpoint.
29843 Internal breakpoint, used to handle correctly stepping through
29844 @code{longjmp} calls.
29846 @item longjmp resume
29847 Internal breakpoint at the target of a @code{longjmp}.
29850 Temporary internal breakpoint used by the @value{GDBN} @code{until} command.
29853 Temporary internal breakpoint used by the @value{GDBN} @code{finish} command.
29856 Shared library events.
29860 @kindex set displaced-stepping
29861 @kindex show displaced-stepping
29862 @cindex displaced stepping support
29863 @cindex out-of-line single-stepping
29864 @item set displaced-stepping
29865 @itemx show displaced-stepping
29866 Control whether or not @value{GDBN} will do @dfn{displaced stepping}
29867 if the target supports it. Displaced stepping is a way to single-step
29868 over breakpoints without removing them from the inferior, by executing
29869 an out-of-line copy of the instruction that was originally at the
29870 breakpoint location. It is also known as out-of-line single-stepping.
29873 @item set displaced-stepping on
29874 If the target architecture supports it, @value{GDBN} will use
29875 displaced stepping to step over breakpoints.
29877 @item set displaced-stepping off
29878 @value{GDBN} will not use displaced stepping to step over breakpoints,
29879 even if such is supported by the target architecture.
29881 @cindex non-stop mode, and @samp{set displaced-stepping}
29882 @item set displaced-stepping auto
29883 This is the default mode. @value{GDBN} will use displaced stepping
29884 only if non-stop mode is active (@pxref{Non-Stop Mode}) and the target
29885 architecture supports displaced stepping.
29888 @kindex maint check-symtabs
29889 @item maint check-symtabs
29890 Check the consistency of psymtabs and symtabs.
29892 @kindex maint cplus first_component
29893 @item maint cplus first_component @var{name}
29894 Print the first C@t{++} class/namespace component of @var{name}.
29896 @kindex maint cplus namespace
29897 @item maint cplus namespace
29898 Print the list of possible C@t{++} namespaces.
29900 @kindex maint demangle
29901 @item maint demangle @var{name}
29902 Demangle a C@t{++} or Objective-C mangled @var{name}.
29904 @kindex maint deprecate
29905 @kindex maint undeprecate
29906 @cindex deprecated commands
29907 @item maint deprecate @var{command} @r{[}@var{replacement}@r{]}
29908 @itemx maint undeprecate @var{command}
29909 Deprecate or undeprecate the named @var{command}. Deprecated commands
29910 cause @value{GDBN} to issue a warning when you use them. The optional
29911 argument @var{replacement} says which newer command should be used in
29912 favor of the deprecated one; if it is given, @value{GDBN} will mention
29913 the replacement as part of the warning.
29915 @kindex maint dump-me
29916 @item maint dump-me
29917 @cindex @code{SIGQUIT} signal, dump core of @value{GDBN}
29918 Cause a fatal signal in the debugger and force it to dump its core.
29919 This is supported only on systems which support aborting a program
29920 with the @code{SIGQUIT} signal.
29922 @kindex maint internal-error
29923 @kindex maint internal-warning
29924 @item maint internal-error @r{[}@var{message-text}@r{]}
29925 @itemx maint internal-warning @r{[}@var{message-text}@r{]}
29926 Cause @value{GDBN} to call the internal function @code{internal_error}
29927 or @code{internal_warning} and hence behave as though an internal error
29928 or internal warning has been detected. In addition to reporting the
29929 internal problem, these functions give the user the opportunity to
29930 either quit @value{GDBN} or create a core file of the current
29931 @value{GDBN} session.
29933 These commands take an optional parameter @var{message-text} that is
29934 used as the text of the error or warning message.
29936 Here's an example of using @code{internal-error}:
29939 (@value{GDBP}) @kbd{maint internal-error testing, 1, 2}
29940 @dots{}/maint.c:121: internal-error: testing, 1, 2
29941 A problem internal to GDB has been detected. Further
29942 debugging may prove unreliable.
29943 Quit this debugging session? (y or n) @kbd{n}
29944 Create a core file? (y or n) @kbd{n}
29948 @cindex @value{GDBN} internal error
29949 @cindex internal errors, control of @value{GDBN} behavior
29951 @kindex maint set internal-error
29952 @kindex maint show internal-error
29953 @kindex maint set internal-warning
29954 @kindex maint show internal-warning
29955 @item maint set internal-error @var{action} [ask|yes|no]
29956 @itemx maint show internal-error @var{action}
29957 @itemx maint set internal-warning @var{action} [ask|yes|no]
29958 @itemx maint show internal-warning @var{action}
29959 When @value{GDBN} reports an internal problem (error or warning) it
29960 gives the user the opportunity to both quit @value{GDBN} and create a
29961 core file of the current @value{GDBN} session. These commands let you
29962 override the default behaviour for each particular @var{action},
29963 described in the table below.
29967 You can specify that @value{GDBN} should always (yes) or never (no)
29968 quit. The default is to ask the user what to do.
29971 You can specify that @value{GDBN} should always (yes) or never (no)
29972 create a core file. The default is to ask the user what to do.
29975 @kindex maint packet
29976 @item maint packet @var{text}
29977 If @value{GDBN} is talking to an inferior via the serial protocol,
29978 then this command sends the string @var{text} to the inferior, and
29979 displays the response packet. @value{GDBN} supplies the initial
29980 @samp{$} character, the terminating @samp{#} character, and the
29983 @kindex maint print architecture
29984 @item maint print architecture @r{[}@var{file}@r{]}
29985 Print the entire architecture configuration. The optional argument
29986 @var{file} names the file where the output goes.
29988 @kindex maint print c-tdesc
29989 @item maint print c-tdesc
29990 Print the current target description (@pxref{Target Descriptions}) as
29991 a C source file. The created source file can be used in @value{GDBN}
29992 when an XML parser is not available to parse the description.
29994 @kindex maint print dummy-frames
29995 @item maint print dummy-frames
29996 Prints the contents of @value{GDBN}'s internal dummy-frame stack.
29999 (@value{GDBP}) @kbd{b add}
30001 (@value{GDBP}) @kbd{print add(2,3)}
30002 Breakpoint 2, add (a=2, b=3) at @dots{}
30004 The program being debugged stopped while in a function called from GDB.
30006 (@value{GDBP}) @kbd{maint print dummy-frames}
30007 0x1a57c80: pc=0x01014068 fp=0x0200bddc sp=0x0200bdd6
30008 top=0x0200bdd4 id=@{stack=0x200bddc,code=0x101405c@}
30009 call_lo=0x01014000 call_hi=0x01014001
30013 Takes an optional file parameter.
30015 @kindex maint print registers
30016 @kindex maint print raw-registers
30017 @kindex maint print cooked-registers
30018 @kindex maint print register-groups
30019 @item maint print registers @r{[}@var{file}@r{]}
30020 @itemx maint print raw-registers @r{[}@var{file}@r{]}
30021 @itemx maint print cooked-registers @r{[}@var{file}@r{]}
30022 @itemx maint print register-groups @r{[}@var{file}@r{]}
30023 Print @value{GDBN}'s internal register data structures.
30025 The command @code{maint print raw-registers} includes the contents of
30026 the raw register cache; the command @code{maint print cooked-registers}
30027 includes the (cooked) value of all registers, including registers which
30028 aren't available on the target nor visible to user; and the
30029 command @code{maint print register-groups} includes the groups that each
30030 register is a member of. @xref{Registers,, Registers, gdbint,
30031 @value{GDBN} Internals}.
30033 These commands take an optional parameter, a file name to which to
30034 write the information.
30036 @kindex maint print reggroups
30037 @item maint print reggroups @r{[}@var{file}@r{]}
30038 Print @value{GDBN}'s internal register group data structures. The
30039 optional argument @var{file} tells to what file to write the
30042 The register groups info looks like this:
30045 (@value{GDBP}) @kbd{maint print reggroups}
30058 This command forces @value{GDBN} to flush its internal register cache.
30060 @kindex maint print objfiles
30061 @cindex info for known object files
30062 @item maint print objfiles
30063 Print a dump of all known object files. For each object file, this
30064 command prints its name, address in memory, and all of its psymtabs
30067 @kindex maint print section-scripts
30068 @cindex info for known .debug_gdb_scripts-loaded scripts
30069 @item maint print section-scripts [@var{regexp}]
30070 Print a dump of scripts specified in the @code{.debug_gdb_section} section.
30071 If @var{regexp} is specified, only print scripts loaded by object files
30072 matching @var{regexp}.
30073 For each script, this command prints its name as specified in the objfile,
30074 and the full path if known.
30075 @xref{.debug_gdb_scripts section}.
30077 @kindex maint print statistics
30078 @cindex bcache statistics
30079 @item maint print statistics
30080 This command prints, for each object file in the program, various data
30081 about that object file followed by the byte cache (@dfn{bcache})
30082 statistics for the object file. The objfile data includes the number
30083 of minimal, partial, full, and stabs symbols, the number of types
30084 defined by the objfile, the number of as yet unexpanded psym tables,
30085 the number of line tables and string tables, and the amount of memory
30086 used by the various tables. The bcache statistics include the counts,
30087 sizes, and counts of duplicates of all and unique objects, max,
30088 average, and median entry size, total memory used and its overhead and
30089 savings, and various measures of the hash table size and chain
30092 @kindex maint print target-stack
30093 @cindex target stack description
30094 @item maint print target-stack
30095 A @dfn{target} is an interface between the debugger and a particular
30096 kind of file or process. Targets can be stacked in @dfn{strata},
30097 so that more than one target can potentially respond to a request.
30098 In particular, memory accesses will walk down the stack of targets
30099 until they find a target that is interested in handling that particular
30102 This command prints a short description of each layer that was pushed on
30103 the @dfn{target stack}, starting from the top layer down to the bottom one.
30105 @kindex maint print type
30106 @cindex type chain of a data type
30107 @item maint print type @var{expr}
30108 Print the type chain for a type specified by @var{expr}. The argument
30109 can be either a type name or a symbol. If it is a symbol, the type of
30110 that symbol is described. The type chain produced by this command is
30111 a recursive definition of the data type as stored in @value{GDBN}'s
30112 data structures, including its flags and contained types.
30114 @kindex maint set dwarf2 always-disassemble
30115 @kindex maint show dwarf2 always-disassemble
30116 @item maint set dwarf2 always-disassemble
30117 @item maint show dwarf2 always-disassemble
30118 Control the behavior of @code{info address} when using DWARF debugging
30121 The default is @code{off}, which means that @value{GDBN} should try to
30122 describe a variable's location in an easily readable format. When
30123 @code{on}, @value{GDBN} will instead display the DWARF location
30124 expression in an assembly-like format. Note that some locations are
30125 too complex for @value{GDBN} to describe simply; in this case you will
30126 always see the disassembly form.
30128 Here is an example of the resulting disassembly:
30131 (gdb) info addr argc
30132 Symbol "argc" is a complex DWARF expression:
30136 For more information on these expressions, see
30137 @uref{http://www.dwarfstd.org/, the DWARF standard}.
30139 @kindex maint set dwarf2 max-cache-age
30140 @kindex maint show dwarf2 max-cache-age
30141 @item maint set dwarf2 max-cache-age
30142 @itemx maint show dwarf2 max-cache-age
30143 Control the DWARF 2 compilation unit cache.
30145 @cindex DWARF 2 compilation units cache
30146 In object files with inter-compilation-unit references, such as those
30147 produced by the GCC option @samp{-feliminate-dwarf2-dups}, the DWARF 2
30148 reader needs to frequently refer to previously read compilation units.
30149 This setting controls how long a compilation unit will remain in the
30150 cache if it is not referenced. A higher limit means that cached
30151 compilation units will be stored in memory longer, and more total
30152 memory will be used. Setting it to zero disables caching, which will
30153 slow down @value{GDBN} startup, but reduce memory consumption.
30155 @kindex maint set profile
30156 @kindex maint show profile
30157 @cindex profiling GDB
30158 @item maint set profile
30159 @itemx maint show profile
30160 Control profiling of @value{GDBN}.
30162 Profiling will be disabled until you use the @samp{maint set profile}
30163 command to enable it. When you enable profiling, the system will begin
30164 collecting timing and execution count data; when you disable profiling or
30165 exit @value{GDBN}, the results will be written to a log file. Remember that
30166 if you use profiling, @value{GDBN} will overwrite the profiling log file
30167 (often called @file{gmon.out}). If you have a record of important profiling
30168 data in a @file{gmon.out} file, be sure to move it to a safe location.
30170 Configuring with @samp{--enable-profiling} arranges for @value{GDBN} to be
30171 compiled with the @samp{-pg} compiler option.
30173 @kindex maint set show-debug-regs
30174 @kindex maint show show-debug-regs
30175 @cindex hardware debug registers
30176 @item maint set show-debug-regs
30177 @itemx maint show show-debug-regs
30178 Control whether to show variables that mirror the hardware debug
30179 registers. Use @code{ON} to enable, @code{OFF} to disable. If
30180 enabled, the debug registers values are shown when @value{GDBN} inserts or
30181 removes a hardware breakpoint or watchpoint, and when the inferior
30182 triggers a hardware-assisted breakpoint or watchpoint.
30184 @kindex maint set show-all-tib
30185 @kindex maint show show-all-tib
30186 @item maint set show-all-tib
30187 @itemx maint show show-all-tib
30188 Control whether to show all non zero areas within a 1k block starting
30189 at thread local base, when using the @samp{info w32 thread-information-block}
30192 @kindex maint space
30193 @cindex memory used by commands
30195 Control whether to display memory usage for each command. If set to a
30196 nonzero value, @value{GDBN} will display how much memory each command
30197 took, following the command's own output. This can also be requested
30198 by invoking @value{GDBN} with the @option{--statistics} command-line
30199 switch (@pxref{Mode Options}).
30202 @cindex time of command execution
30204 Control whether to display the execution time for each command. If
30205 set to a nonzero value, @value{GDBN} will display how much time it
30206 took to execute each command, following the command's own output.
30207 The time is not printed for the commands that run the target, since
30208 there's no mechanism currently to compute how much time was spend
30209 by @value{GDBN} and how much time was spend by the program been debugged.
30210 it's not possibly currently
30211 This can also be requested by invoking @value{GDBN} with the
30212 @option{--statistics} command-line switch (@pxref{Mode Options}).
30214 @kindex maint translate-address
30215 @item maint translate-address @r{[}@var{section}@r{]} @var{addr}
30216 Find the symbol stored at the location specified by the address
30217 @var{addr} and an optional section name @var{section}. If found,
30218 @value{GDBN} prints the name of the closest symbol and an offset from
30219 the symbol's location to the specified address. This is similar to
30220 the @code{info address} command (@pxref{Symbols}), except that this
30221 command also allows to find symbols in other sections.
30223 If section was not specified, the section in which the symbol was found
30224 is also printed. For dynamically linked executables, the name of
30225 executable or shared library containing the symbol is printed as well.
30229 The following command is useful for non-interactive invocations of
30230 @value{GDBN}, such as in the test suite.
30233 @item set watchdog @var{nsec}
30234 @kindex set watchdog
30235 @cindex watchdog timer
30236 @cindex timeout for commands
30237 Set the maximum number of seconds @value{GDBN} will wait for the
30238 target operation to finish. If this time expires, @value{GDBN}
30239 reports and error and the command is aborted.
30241 @item show watchdog
30242 Show the current setting of the target wait timeout.
30245 @node Remote Protocol
30246 @appendix @value{GDBN} Remote Serial Protocol
30251 * Stop Reply Packets::
30252 * General Query Packets::
30253 * Architecture-Specific Protocol Details::
30254 * Tracepoint Packets::
30255 * Host I/O Packets::
30257 * Notification Packets::
30258 * Remote Non-Stop::
30259 * Packet Acknowledgment::
30261 * File-I/O Remote Protocol Extension::
30262 * Library List Format::
30263 * Memory Map Format::
30264 * Thread List Format::
30270 There may be occasions when you need to know something about the
30271 protocol---for example, if there is only one serial port to your target
30272 machine, you might want your program to do something special if it
30273 recognizes a packet meant for @value{GDBN}.
30275 In the examples below, @samp{->} and @samp{<-} are used to indicate
30276 transmitted and received data, respectively.
30278 @cindex protocol, @value{GDBN} remote serial
30279 @cindex serial protocol, @value{GDBN} remote
30280 @cindex remote serial protocol
30281 All @value{GDBN} commands and responses (other than acknowledgments
30282 and notifications, see @ref{Notification Packets}) are sent as a
30283 @var{packet}. A @var{packet} is introduced with the character
30284 @samp{$}, the actual @var{packet-data}, and the terminating character
30285 @samp{#} followed by a two-digit @var{checksum}:
30288 @code{$}@var{packet-data}@code{#}@var{checksum}
30292 @cindex checksum, for @value{GDBN} remote
30294 The two-digit @var{checksum} is computed as the modulo 256 sum of all
30295 characters between the leading @samp{$} and the trailing @samp{#} (an
30296 eight bit unsigned checksum).
30298 Implementors should note that prior to @value{GDBN} 5.0 the protocol
30299 specification also included an optional two-digit @var{sequence-id}:
30302 @code{$}@var{sequence-id}@code{:}@var{packet-data}@code{#}@var{checksum}
30305 @cindex sequence-id, for @value{GDBN} remote
30307 That @var{sequence-id} was appended to the acknowledgment. @value{GDBN}
30308 has never output @var{sequence-id}s. Stubs that handle packets added
30309 since @value{GDBN} 5.0 must not accept @var{sequence-id}.
30311 When either the host or the target machine receives a packet, the first
30312 response expected is an acknowledgment: either @samp{+} (to indicate
30313 the package was received correctly) or @samp{-} (to request
30317 -> @code{$}@var{packet-data}@code{#}@var{checksum}
30322 The @samp{+}/@samp{-} acknowledgments can be disabled
30323 once a connection is established.
30324 @xref{Packet Acknowledgment}, for details.
30326 The host (@value{GDBN}) sends @var{command}s, and the target (the
30327 debugging stub incorporated in your program) sends a @var{response}. In
30328 the case of step and continue @var{command}s, the response is only sent
30329 when the operation has completed, and the target has again stopped all
30330 threads in all attached processes. This is the default all-stop mode
30331 behavior, but the remote protocol also supports @value{GDBN}'s non-stop
30332 execution mode; see @ref{Remote Non-Stop}, for details.
30334 @var{packet-data} consists of a sequence of characters with the
30335 exception of @samp{#} and @samp{$} (see @samp{X} packet for additional
30338 @cindex remote protocol, field separator
30339 Fields within the packet should be separated using @samp{,} @samp{;} or
30340 @samp{:}. Except where otherwise noted all numbers are represented in
30341 @sc{hex} with leading zeros suppressed.
30343 Implementors should note that prior to @value{GDBN} 5.0, the character
30344 @samp{:} could not appear as the third character in a packet (as it
30345 would potentially conflict with the @var{sequence-id}).
30347 @cindex remote protocol, binary data
30348 @anchor{Binary Data}
30349 Binary data in most packets is encoded either as two hexadecimal
30350 digits per byte of binary data. This allowed the traditional remote
30351 protocol to work over connections which were only seven-bit clean.
30352 Some packets designed more recently assume an eight-bit clean
30353 connection, and use a more efficient encoding to send and receive
30356 The binary data representation uses @code{7d} (@sc{ascii} @samp{@}})
30357 as an escape character. Any escaped byte is transmitted as the escape
30358 character followed by the original character XORed with @code{0x20}.
30359 For example, the byte @code{0x7d} would be transmitted as the two
30360 bytes @code{0x7d 0x5d}. The bytes @code{0x23} (@sc{ascii} @samp{#}),
30361 @code{0x24} (@sc{ascii} @samp{$}), and @code{0x7d} (@sc{ascii}
30362 @samp{@}}) must always be escaped. Responses sent by the stub
30363 must also escape @code{0x2a} (@sc{ascii} @samp{*}), so that it
30364 is not interpreted as the start of a run-length encoded sequence
30367 Response @var{data} can be run-length encoded to save space.
30368 Run-length encoding replaces runs of identical characters with one
30369 instance of the repeated character, followed by a @samp{*} and a
30370 repeat count. The repeat count is itself sent encoded, to avoid
30371 binary characters in @var{data}: a value of @var{n} is sent as
30372 @code{@var{n}+29}. For a repeat count greater or equal to 3, this
30373 produces a printable @sc{ascii} character, e.g.@: a space (@sc{ascii}
30374 code 32) for a repeat count of 3. (This is because run-length
30375 encoding starts to win for counts 3 or more.) Thus, for example,
30376 @samp{0* } is a run-length encoding of ``0000'': the space character
30377 after @samp{*} means repeat the leading @code{0} @w{@code{32 - 29 =
30380 The printable characters @samp{#} and @samp{$} or with a numeric value
30381 greater than 126 must not be used. Runs of six repeats (@samp{#}) or
30382 seven repeats (@samp{$}) can be expanded using a repeat count of only
30383 five (@samp{"}). For example, @samp{00000000} can be encoded as
30386 The error response returned for some packets includes a two character
30387 error number. That number is not well defined.
30389 @cindex empty response, for unsupported packets
30390 For any @var{command} not supported by the stub, an empty response
30391 (@samp{$#00}) should be returned. That way it is possible to extend the
30392 protocol. A newer @value{GDBN} can tell if a packet is supported based
30395 A stub is required to support the @samp{g}, @samp{G}, @samp{m}, @samp{M},
30396 @samp{c}, and @samp{s} @var{command}s. All other @var{command}s are
30402 The following table provides a complete list of all currently defined
30403 @var{command}s and their corresponding response @var{data}.
30404 @xref{File-I/O Remote Protocol Extension}, for details about the File
30405 I/O extension of the remote protocol.
30407 Each packet's description has a template showing the packet's overall
30408 syntax, followed by an explanation of the packet's meaning. We
30409 include spaces in some of the templates for clarity; these are not
30410 part of the packet's syntax. No @value{GDBN} packet uses spaces to
30411 separate its components. For example, a template like @samp{foo
30412 @var{bar} @var{baz}} describes a packet beginning with the three ASCII
30413 bytes @samp{foo}, followed by a @var{bar}, followed directly by a
30414 @var{baz}. @value{GDBN} does not transmit a space character between the
30415 @samp{foo} and the @var{bar}, or between the @var{bar} and the
30418 @cindex @var{thread-id}, in remote protocol
30419 @anchor{thread-id syntax}
30420 Several packets and replies include a @var{thread-id} field to identify
30421 a thread. Normally these are positive numbers with a target-specific
30422 interpretation, formatted as big-endian hex strings. A @var{thread-id}
30423 can also be a literal @samp{-1} to indicate all threads, or @samp{0} to
30426 In addition, the remote protocol supports a multiprocess feature in
30427 which the @var{thread-id} syntax is extended to optionally include both
30428 process and thread ID fields, as @samp{p@var{pid}.@var{tid}}.
30429 The @var{pid} (process) and @var{tid} (thread) components each have the
30430 format described above: a positive number with target-specific
30431 interpretation formatted as a big-endian hex string, literal @samp{-1}
30432 to indicate all processes or threads (respectively), or @samp{0} to
30433 indicate an arbitrary process or thread. Specifying just a process, as
30434 @samp{p@var{pid}}, is equivalent to @samp{p@var{pid}.-1}. It is an
30435 error to specify all processes but a specific thread, such as
30436 @samp{p-1.@var{tid}}. Note that the @samp{p} prefix is @emph{not} used
30437 for those packets and replies explicitly documented to include a process
30438 ID, rather than a @var{thread-id}.
30440 The multiprocess @var{thread-id} syntax extensions are only used if both
30441 @value{GDBN} and the stub report support for the @samp{multiprocess}
30442 feature using @samp{qSupported}. @xref{multiprocess extensions}, for
30445 Note that all packet forms beginning with an upper- or lower-case
30446 letter, other than those described here, are reserved for future use.
30448 Here are the packet descriptions.
30453 @cindex @samp{!} packet
30454 @anchor{extended mode}
30455 Enable extended mode. In extended mode, the remote server is made
30456 persistent. The @samp{R} packet is used to restart the program being
30462 The remote target both supports and has enabled extended mode.
30466 @cindex @samp{?} packet
30467 Indicate the reason the target halted. The reply is the same as for
30468 step and continue. This packet has a special interpretation when the
30469 target is in non-stop mode; see @ref{Remote Non-Stop}.
30472 @xref{Stop Reply Packets}, for the reply specifications.
30474 @item A @var{arglen},@var{argnum},@var{arg},@dots{}
30475 @cindex @samp{A} packet
30476 Initialized @code{argv[]} array passed into program. @var{arglen}
30477 specifies the number of bytes in the hex encoded byte stream
30478 @var{arg}. See @code{gdbserver} for more details.
30483 The arguments were set.
30489 @cindex @samp{b} packet
30490 (Don't use this packet; its behavior is not well-defined.)
30491 Change the serial line speed to @var{baud}.
30493 JTC: @emph{When does the transport layer state change? When it's
30494 received, or after the ACK is transmitted. In either case, there are
30495 problems if the command or the acknowledgment packet is dropped.}
30497 Stan: @emph{If people really wanted to add something like this, and get
30498 it working for the first time, they ought to modify ser-unix.c to send
30499 some kind of out-of-band message to a specially-setup stub and have the
30500 switch happen "in between" packets, so that from remote protocol's point
30501 of view, nothing actually happened.}
30503 @item B @var{addr},@var{mode}
30504 @cindex @samp{B} packet
30505 Set (@var{mode} is @samp{S}) or clear (@var{mode} is @samp{C}) a
30506 breakpoint at @var{addr}.
30508 Don't use this packet. Use the @samp{Z} and @samp{z} packets instead
30509 (@pxref{insert breakpoint or watchpoint packet}).
30511 @cindex @samp{bc} packet
30514 Backward continue. Execute the target system in reverse. No parameter.
30515 @xref{Reverse Execution}, for more information.
30518 @xref{Stop Reply Packets}, for the reply specifications.
30520 @cindex @samp{bs} packet
30523 Backward single step. Execute one instruction in reverse. No parameter.
30524 @xref{Reverse Execution}, for more information.
30527 @xref{Stop Reply Packets}, for the reply specifications.
30529 @item c @r{[}@var{addr}@r{]}
30530 @cindex @samp{c} packet
30531 Continue. @var{addr} is address to resume. If @var{addr} is omitted,
30532 resume at current address.
30535 @xref{Stop Reply Packets}, for the reply specifications.
30537 @item C @var{sig}@r{[};@var{addr}@r{]}
30538 @cindex @samp{C} packet
30539 Continue with signal @var{sig} (hex signal number). If
30540 @samp{;@var{addr}} is omitted, resume at same address.
30543 @xref{Stop Reply Packets}, for the reply specifications.
30546 @cindex @samp{d} packet
30549 Don't use this packet; instead, define a general set packet
30550 (@pxref{General Query Packets}).
30554 @cindex @samp{D} packet
30555 The first form of the packet is used to detach @value{GDBN} from the
30556 remote system. It is sent to the remote target
30557 before @value{GDBN} disconnects via the @code{detach} command.
30559 The second form, including a process ID, is used when multiprocess
30560 protocol extensions are enabled (@pxref{multiprocess extensions}), to
30561 detach only a specific process. The @var{pid} is specified as a
30562 big-endian hex string.
30572 @item F @var{RC},@var{EE},@var{CF};@var{XX}
30573 @cindex @samp{F} packet
30574 A reply from @value{GDBN} to an @samp{F} packet sent by the target.
30575 This is part of the File-I/O protocol extension. @xref{File-I/O
30576 Remote Protocol Extension}, for the specification.
30579 @anchor{read registers packet}
30580 @cindex @samp{g} packet
30581 Read general registers.
30585 @item @var{XX@dots{}}
30586 Each byte of register data is described by two hex digits. The bytes
30587 with the register are transmitted in target byte order. The size of
30588 each register and their position within the @samp{g} packet are
30589 determined by the @value{GDBN} internal gdbarch functions
30590 @code{DEPRECATED_REGISTER_RAW_SIZE} and @code{gdbarch_register_name}. The
30591 specification of several standard @samp{g} packets is specified below.
30596 @item G @var{XX@dots{}}
30597 @cindex @samp{G} packet
30598 Write general registers. @xref{read registers packet}, for a
30599 description of the @var{XX@dots{}} data.
30609 @item H @var{c} @var{thread-id}
30610 @cindex @samp{H} packet
30611 Set thread for subsequent operations (@samp{m}, @samp{M}, @samp{g},
30612 @samp{G}, et.al.). @var{c} depends on the operation to be performed: it
30613 should be @samp{c} for step and continue operations, @samp{g} for other
30614 operations. The thread designator @var{thread-id} has the format and
30615 interpretation described in @ref{thread-id syntax}.
30626 @c 'H': How restrictive (or permissive) is the thread model. If a
30627 @c thread is selected and stopped, are other threads allowed
30628 @c to continue to execute? As I mentioned above, I think the
30629 @c semantics of each command when a thread is selected must be
30630 @c described. For example:
30632 @c 'g': If the stub supports threads and a specific thread is
30633 @c selected, returns the register block from that thread;
30634 @c otherwise returns current registers.
30636 @c 'G' If the stub supports threads and a specific thread is
30637 @c selected, sets the registers of the register block of
30638 @c that thread; otherwise sets current registers.
30640 @item i @r{[}@var{addr}@r{[},@var{nnn}@r{]]}
30641 @anchor{cycle step packet}
30642 @cindex @samp{i} packet
30643 Step the remote target by a single clock cycle. If @samp{,@var{nnn}} is
30644 present, cycle step @var{nnn} cycles. If @var{addr} is present, cycle
30645 step starting at that address.
30648 @cindex @samp{I} packet
30649 Signal, then cycle step. @xref{step with signal packet}. @xref{cycle
30653 @cindex @samp{k} packet
30656 FIXME: @emph{There is no description of how to operate when a specific
30657 thread context has been selected (i.e.@: does 'k' kill only that
30660 @item m @var{addr},@var{length}
30661 @cindex @samp{m} packet
30662 Read @var{length} bytes of memory starting at address @var{addr}.
30663 Note that @var{addr} may not be aligned to any particular boundary.
30665 The stub need not use any particular size or alignment when gathering
30666 data from memory for the response; even if @var{addr} is word-aligned
30667 and @var{length} is a multiple of the word size, the stub is free to
30668 use byte accesses, or not. For this reason, this packet may not be
30669 suitable for accessing memory-mapped I/O devices.
30670 @cindex alignment of remote memory accesses
30671 @cindex size of remote memory accesses
30672 @cindex memory, alignment and size of remote accesses
30676 @item @var{XX@dots{}}
30677 Memory contents; each byte is transmitted as a two-digit hexadecimal
30678 number. The reply may contain fewer bytes than requested if the
30679 server was able to read only part of the region of memory.
30684 @item M @var{addr},@var{length}:@var{XX@dots{}}
30685 @cindex @samp{M} packet
30686 Write @var{length} bytes of memory starting at address @var{addr}.
30687 @var{XX@dots{}} is the data; each byte is transmitted as a two-digit
30688 hexadecimal number.
30695 for an error (this includes the case where only part of the data was
30700 @cindex @samp{p} packet
30701 Read the value of register @var{n}; @var{n} is in hex.
30702 @xref{read registers packet}, for a description of how the returned
30703 register value is encoded.
30707 @item @var{XX@dots{}}
30708 the register's value
30712 Indicating an unrecognized @var{query}.
30715 @item P @var{n@dots{}}=@var{r@dots{}}
30716 @anchor{write register packet}
30717 @cindex @samp{P} packet
30718 Write register @var{n@dots{}} with value @var{r@dots{}}. The register
30719 number @var{n} is in hexadecimal, and @var{r@dots{}} contains two hex
30720 digits for each byte in the register (target byte order).
30730 @item q @var{name} @var{params}@dots{}
30731 @itemx Q @var{name} @var{params}@dots{}
30732 @cindex @samp{q} packet
30733 @cindex @samp{Q} packet
30734 General query (@samp{q}) and set (@samp{Q}). These packets are
30735 described fully in @ref{General Query Packets}.
30738 @cindex @samp{r} packet
30739 Reset the entire system.
30741 Don't use this packet; use the @samp{R} packet instead.
30744 @cindex @samp{R} packet
30745 Restart the program being debugged. @var{XX}, while needed, is ignored.
30746 This packet is only available in extended mode (@pxref{extended mode}).
30748 The @samp{R} packet has no reply.
30750 @item s @r{[}@var{addr}@r{]}
30751 @cindex @samp{s} packet
30752 Single step. @var{addr} is the address at which to resume. If
30753 @var{addr} is omitted, resume at same address.
30756 @xref{Stop Reply Packets}, for the reply specifications.
30758 @item S @var{sig}@r{[};@var{addr}@r{]}
30759 @anchor{step with signal packet}
30760 @cindex @samp{S} packet
30761 Step with signal. This is analogous to the @samp{C} packet, but
30762 requests a single-step, rather than a normal resumption of execution.
30765 @xref{Stop Reply Packets}, for the reply specifications.
30767 @item t @var{addr}:@var{PP},@var{MM}
30768 @cindex @samp{t} packet
30769 Search backwards starting at address @var{addr} for a match with pattern
30770 @var{PP} and mask @var{MM}. @var{PP} and @var{MM} are 4 bytes.
30771 @var{addr} must be at least 3 digits.
30773 @item T @var{thread-id}
30774 @cindex @samp{T} packet
30775 Find out if the thread @var{thread-id} is alive. @xref{thread-id syntax}.
30780 thread is still alive
30786 Packets starting with @samp{v} are identified by a multi-letter name,
30787 up to the first @samp{;} or @samp{?} (or the end of the packet).
30789 @item vAttach;@var{pid}
30790 @cindex @samp{vAttach} packet
30791 Attach to a new process with the specified process ID @var{pid}.
30792 The process ID is a
30793 hexadecimal integer identifying the process. In all-stop mode, all
30794 threads in the attached process are stopped; in non-stop mode, it may be
30795 attached without being stopped if that is supported by the target.
30797 @c In non-stop mode, on a successful vAttach, the stub should set the
30798 @c current thread to a thread of the newly-attached process. After
30799 @c attaching, GDB queries for the attached process's thread ID with qC.
30800 @c Also note that, from a user perspective, whether or not the
30801 @c target is stopped on attach in non-stop mode depends on whether you
30802 @c use the foreground or background version of the attach command, not
30803 @c on what vAttach does; GDB does the right thing with respect to either
30804 @c stopping or restarting threads.
30806 This packet is only available in extended mode (@pxref{extended mode}).
30812 @item @r{Any stop packet}
30813 for success in all-stop mode (@pxref{Stop Reply Packets})
30815 for success in non-stop mode (@pxref{Remote Non-Stop})
30818 @item vCont@r{[};@var{action}@r{[}:@var{thread-id}@r{]]}@dots{}
30819 @cindex @samp{vCont} packet
30820 Resume the inferior, specifying different actions for each thread.
30821 If an action is specified with no @var{thread-id}, then it is applied to any
30822 threads that don't have a specific action specified; if no default action is
30823 specified then other threads should remain stopped in all-stop mode and
30824 in their current state in non-stop mode.
30825 Specifying multiple
30826 default actions is an error; specifying no actions is also an error.
30827 Thread IDs are specified using the syntax described in @ref{thread-id syntax}.
30829 Currently supported actions are:
30835 Continue with signal @var{sig}. The signal @var{sig} should be two hex digits.
30839 Step with signal @var{sig}. The signal @var{sig} should be two hex digits.
30844 The optional argument @var{addr} normally associated with the
30845 @samp{c}, @samp{C}, @samp{s}, and @samp{S} packets is
30846 not supported in @samp{vCont}.
30848 The @samp{t} action is only relevant in non-stop mode
30849 (@pxref{Remote Non-Stop}) and may be ignored by the stub otherwise.
30850 A stop reply should be generated for any affected thread not already stopped.
30851 When a thread is stopped by means of a @samp{t} action,
30852 the corresponding stop reply should indicate that the thread has stopped with
30853 signal @samp{0}, regardless of whether the target uses some other signal
30854 as an implementation detail.
30857 @xref{Stop Reply Packets}, for the reply specifications.
30860 @cindex @samp{vCont?} packet
30861 Request a list of actions supported by the @samp{vCont} packet.
30865 @item vCont@r{[};@var{action}@dots{}@r{]}
30866 The @samp{vCont} packet is supported. Each @var{action} is a supported
30867 command in the @samp{vCont} packet.
30869 The @samp{vCont} packet is not supported.
30872 @item vFile:@var{operation}:@var{parameter}@dots{}
30873 @cindex @samp{vFile} packet
30874 Perform a file operation on the target system. For details,
30875 see @ref{Host I/O Packets}.
30877 @item vFlashErase:@var{addr},@var{length}
30878 @cindex @samp{vFlashErase} packet
30879 Direct the stub to erase @var{length} bytes of flash starting at
30880 @var{addr}. The region may enclose any number of flash blocks, but
30881 its start and end must fall on block boundaries, as indicated by the
30882 flash block size appearing in the memory map (@pxref{Memory Map
30883 Format}). @value{GDBN} groups flash memory programming operations
30884 together, and sends a @samp{vFlashDone} request after each group; the
30885 stub is allowed to delay erase operation until the @samp{vFlashDone}
30886 packet is received.
30888 The stub must support @samp{vCont} if it reports support for
30889 multiprocess extensions (@pxref{multiprocess extensions}). Note that in
30890 this case @samp{vCont} actions can be specified to apply to all threads
30891 in a process by using the @samp{p@var{pid}.-1} form of the
30902 @item vFlashWrite:@var{addr}:@var{XX@dots{}}
30903 @cindex @samp{vFlashWrite} packet
30904 Direct the stub to write data to flash address @var{addr}. The data
30905 is passed in binary form using the same encoding as for the @samp{X}
30906 packet (@pxref{Binary Data}). The memory ranges specified by
30907 @samp{vFlashWrite} packets preceding a @samp{vFlashDone} packet must
30908 not overlap, and must appear in order of increasing addresses
30909 (although @samp{vFlashErase} packets for higher addresses may already
30910 have been received; the ordering is guaranteed only between
30911 @samp{vFlashWrite} packets). If a packet writes to an address that was
30912 neither erased by a preceding @samp{vFlashErase} packet nor by some other
30913 target-specific method, the results are unpredictable.
30921 for vFlashWrite addressing non-flash memory
30927 @cindex @samp{vFlashDone} packet
30928 Indicate to the stub that flash programming operation is finished.
30929 The stub is permitted to delay or batch the effects of a group of
30930 @samp{vFlashErase} and @samp{vFlashWrite} packets until a
30931 @samp{vFlashDone} packet is received. The contents of the affected
30932 regions of flash memory are unpredictable until the @samp{vFlashDone}
30933 request is completed.
30935 @item vKill;@var{pid}
30936 @cindex @samp{vKill} packet
30937 Kill the process with the specified process ID. @var{pid} is a
30938 hexadecimal integer identifying the process. This packet is used in
30939 preference to @samp{k} when multiprocess protocol extensions are
30940 supported; see @ref{multiprocess extensions}.
30950 @item vRun;@var{filename}@r{[};@var{argument}@r{]}@dots{}
30951 @cindex @samp{vRun} packet
30952 Run the program @var{filename}, passing it each @var{argument} on its
30953 command line. The file and arguments are hex-encoded strings. If
30954 @var{filename} is an empty string, the stub may use a default program
30955 (e.g.@: the last program run). The program is created in the stopped
30958 @c FIXME: What about non-stop mode?
30960 This packet is only available in extended mode (@pxref{extended mode}).
30966 @item @r{Any stop packet}
30967 for success (@pxref{Stop Reply Packets})
30971 @anchor{vStopped packet}
30972 @cindex @samp{vStopped} packet
30974 In non-stop mode (@pxref{Remote Non-Stop}), acknowledge a previous stop
30975 reply and prompt for the stub to report another one.
30979 @item @r{Any stop packet}
30980 if there is another unreported stop event (@pxref{Stop Reply Packets})
30982 if there are no unreported stop events
30985 @item X @var{addr},@var{length}:@var{XX@dots{}}
30987 @cindex @samp{X} packet
30988 Write data to memory, where the data is transmitted in binary.
30989 @var{addr} is address, @var{length} is number of bytes,
30990 @samp{@var{XX}@dots{}} is binary data (@pxref{Binary Data}).
31000 @item z @var{type},@var{addr},@var{kind}
31001 @itemx Z @var{type},@var{addr},@var{kind}
31002 @anchor{insert breakpoint or watchpoint packet}
31003 @cindex @samp{z} packet
31004 @cindex @samp{Z} packets
31005 Insert (@samp{Z}) or remove (@samp{z}) a @var{type} breakpoint or
31006 watchpoint starting at address @var{address} of kind @var{kind}.
31008 Each breakpoint and watchpoint packet @var{type} is documented
31011 @emph{Implementation notes: A remote target shall return an empty string
31012 for an unrecognized breakpoint or watchpoint packet @var{type}. A
31013 remote target shall support either both or neither of a given
31014 @samp{Z@var{type}@dots{}} and @samp{z@var{type}@dots{}} packet pair. To
31015 avoid potential problems with duplicate packets, the operations should
31016 be implemented in an idempotent way.}
31018 @item z0,@var{addr},@var{kind}
31019 @itemx Z0,@var{addr},@var{kind}
31020 @cindex @samp{z0} packet
31021 @cindex @samp{Z0} packet
31022 Insert (@samp{Z0}) or remove (@samp{z0}) a memory breakpoint at address
31023 @var{addr} of type @var{kind}.
31025 A memory breakpoint is implemented by replacing the instruction at
31026 @var{addr} with a software breakpoint or trap instruction. The
31027 @var{kind} is target-specific and typically indicates the size of
31028 the breakpoint in bytes that should be inserted. E.g., the @sc{arm}
31029 and @sc{mips} can insert either a 2 or 4 byte breakpoint. Some
31030 architectures have additional meanings for @var{kind};
31031 see @ref{Architecture-Specific Protocol Details}.
31033 @emph{Implementation note: It is possible for a target to copy or move
31034 code that contains memory breakpoints (e.g., when implementing
31035 overlays). The behavior of this packet, in the presence of such a
31036 target, is not defined.}
31048 @item z1,@var{addr},@var{kind}
31049 @itemx Z1,@var{addr},@var{kind}
31050 @cindex @samp{z1} packet
31051 @cindex @samp{Z1} packet
31052 Insert (@samp{Z1}) or remove (@samp{z1}) a hardware breakpoint at
31053 address @var{addr}.
31055 A hardware breakpoint is implemented using a mechanism that is not
31056 dependant on being able to modify the target's memory. @var{kind}
31057 has the same meaning as in @samp{Z0} packets.
31059 @emph{Implementation note: A hardware breakpoint is not affected by code
31072 @item z2,@var{addr},@var{kind}
31073 @itemx Z2,@var{addr},@var{kind}
31074 @cindex @samp{z2} packet
31075 @cindex @samp{Z2} packet
31076 Insert (@samp{Z2}) or remove (@samp{z2}) a write watchpoint at @var{addr}.
31077 @var{kind} is interpreted as the number of bytes to watch.
31089 @item z3,@var{addr},@var{kind}
31090 @itemx Z3,@var{addr},@var{kind}
31091 @cindex @samp{z3} packet
31092 @cindex @samp{Z3} packet
31093 Insert (@samp{Z3}) or remove (@samp{z3}) a read watchpoint at @var{addr}.
31094 @var{kind} is interpreted as the number of bytes to watch.
31106 @item z4,@var{addr},@var{kind}
31107 @itemx Z4,@var{addr},@var{kind}
31108 @cindex @samp{z4} packet
31109 @cindex @samp{Z4} packet
31110 Insert (@samp{Z4}) or remove (@samp{z4}) an access watchpoint at @var{addr}.
31111 @var{kind} is interpreted as the number of bytes to watch.
31125 @node Stop Reply Packets
31126 @section Stop Reply Packets
31127 @cindex stop reply packets
31129 The @samp{C}, @samp{c}, @samp{S}, @samp{s}, @samp{vCont},
31130 @samp{vAttach}, @samp{vRun}, @samp{vStopped}, and @samp{?} packets can
31131 receive any of the below as a reply. Except for @samp{?}
31132 and @samp{vStopped}, that reply is only returned
31133 when the target halts. In the below the exact meaning of @dfn{signal
31134 number} is defined by the header @file{include/gdb/signals.h} in the
31135 @value{GDBN} source code.
31137 As in the description of request packets, we include spaces in the
31138 reply templates for clarity; these are not part of the reply packet's
31139 syntax. No @value{GDBN} stop reply packet uses spaces to separate its
31145 The program received signal number @var{AA} (a two-digit hexadecimal
31146 number). This is equivalent to a @samp{T} response with no
31147 @var{n}:@var{r} pairs.
31149 @item T @var{AA} @var{n1}:@var{r1};@var{n2}:@var{r2};@dots{}
31150 @cindex @samp{T} packet reply
31151 The program received signal number @var{AA} (a two-digit hexadecimal
31152 number). This is equivalent to an @samp{S} response, except that the
31153 @samp{@var{n}:@var{r}} pairs can carry values of important registers
31154 and other information directly in the stop reply packet, reducing
31155 round-trip latency. Single-step and breakpoint traps are reported
31156 this way. Each @samp{@var{n}:@var{r}} pair is interpreted as follows:
31160 If @var{n} is a hexadecimal number, it is a register number, and the
31161 corresponding @var{r} gives that register's value. @var{r} is a
31162 series of bytes in target byte order, with each byte given by a
31163 two-digit hex number.
31166 If @var{n} is @samp{thread}, then @var{r} is the @var{thread-id} of
31167 the stopped thread, as specified in @ref{thread-id syntax}.
31170 If @var{n} is @samp{core}, then @var{r} is the hexadecimal number of
31171 the core on which the stop event was detected.
31174 If @var{n} is a recognized @dfn{stop reason}, it describes a more
31175 specific event that stopped the target. The currently defined stop
31176 reasons are listed below. @var{aa} should be @samp{05}, the trap
31177 signal. At most one stop reason should be present.
31180 Otherwise, @value{GDBN} should ignore this @samp{@var{n}:@var{r}} pair
31181 and go on to the next; this allows us to extend the protocol in the
31185 The currently defined stop reasons are:
31191 The packet indicates a watchpoint hit, and @var{r} is the data address, in
31194 @cindex shared library events, remote reply
31196 The packet indicates that the loaded libraries have changed.
31197 @value{GDBN} should use @samp{qXfer:libraries:read} to fetch a new
31198 list of loaded libraries. @var{r} is ignored.
31200 @cindex replay log events, remote reply
31202 The packet indicates that the target cannot continue replaying
31203 logged execution events, because it has reached the end (or the
31204 beginning when executing backward) of the log. The value of @var{r}
31205 will be either @samp{begin} or @samp{end}. @xref{Reverse Execution},
31206 for more information.
31210 @itemx W @var{AA} ; process:@var{pid}
31211 The process exited, and @var{AA} is the exit status. This is only
31212 applicable to certain targets.
31214 The second form of the response, including the process ID of the exited
31215 process, can be used only when @value{GDBN} has reported support for
31216 multiprocess protocol extensions; see @ref{multiprocess extensions}.
31217 The @var{pid} is formatted as a big-endian hex string.
31220 @itemx X @var{AA} ; process:@var{pid}
31221 The process terminated with signal @var{AA}.
31223 The second form of the response, including the process ID of the
31224 terminated process, can be used only when @value{GDBN} has reported
31225 support for multiprocess protocol extensions; see @ref{multiprocess
31226 extensions}. The @var{pid} is formatted as a big-endian hex string.
31228 @item O @var{XX}@dots{}
31229 @samp{@var{XX}@dots{}} is hex encoding of @sc{ascii} data, to be
31230 written as the program's console output. This can happen at any time
31231 while the program is running and the debugger should continue to wait
31232 for @samp{W}, @samp{T}, etc. This reply is not permitted in non-stop mode.
31234 @item F @var{call-id},@var{parameter}@dots{}
31235 @var{call-id} is the identifier which says which host system call should
31236 be called. This is just the name of the function. Translation into the
31237 correct system call is only applicable as it's defined in @value{GDBN}.
31238 @xref{File-I/O Remote Protocol Extension}, for a list of implemented
31241 @samp{@var{parameter}@dots{}} is a list of parameters as defined for
31242 this very system call.
31244 The target replies with this packet when it expects @value{GDBN} to
31245 call a host system call on behalf of the target. @value{GDBN} replies
31246 with an appropriate @samp{F} packet and keeps up waiting for the next
31247 reply packet from the target. The latest @samp{C}, @samp{c}, @samp{S}
31248 or @samp{s} action is expected to be continued. @xref{File-I/O Remote
31249 Protocol Extension}, for more details.
31253 @node General Query Packets
31254 @section General Query Packets
31255 @cindex remote query requests
31257 Packets starting with @samp{q} are @dfn{general query packets};
31258 packets starting with @samp{Q} are @dfn{general set packets}. General
31259 query and set packets are a semi-unified form for retrieving and
31260 sending information to and from the stub.
31262 The initial letter of a query or set packet is followed by a name
31263 indicating what sort of thing the packet applies to. For example,
31264 @value{GDBN} may use a @samp{qSymbol} packet to exchange symbol
31265 definitions with the stub. These packet names follow some
31270 The name must not contain commas, colons or semicolons.
31272 Most @value{GDBN} query and set packets have a leading upper case
31275 The names of custom vendor packets should use a company prefix, in
31276 lower case, followed by a period. For example, packets designed at
31277 the Acme Corporation might begin with @samp{qacme.foo} (for querying
31278 foos) or @samp{Qacme.bar} (for setting bars).
31281 The name of a query or set packet should be separated from any
31282 parameters by a @samp{:}; the parameters themselves should be
31283 separated by @samp{,} or @samp{;}. Stubs must be careful to match the
31284 full packet name, and check for a separator or the end of the packet,
31285 in case two packet names share a common prefix. New packets should not begin
31286 with @samp{qC}, @samp{qP}, or @samp{qL}@footnote{The @samp{qP} and @samp{qL}
31287 packets predate these conventions, and have arguments without any terminator
31288 for the packet name; we suspect they are in widespread use in places that
31289 are difficult to upgrade. The @samp{qC} packet has no arguments, but some
31290 existing stubs (e.g.@: RedBoot) are known to not check for the end of the
31293 Like the descriptions of the other packets, each description here
31294 has a template showing the packet's overall syntax, followed by an
31295 explanation of the packet's meaning. We include spaces in some of the
31296 templates for clarity; these are not part of the packet's syntax. No
31297 @value{GDBN} packet uses spaces to separate its components.
31299 Here are the currently defined query and set packets:
31303 @item QAllow:@var{op}:@var{val}@dots{}
31304 @cindex @samp{QAllow} packet
31305 Specify which operations @value{GDBN} expects to request of the
31306 target, as a semicolon-separated list of operation name and value
31307 pairs. Possible values for @var{op} include @samp{WriteReg},
31308 @samp{WriteMem}, @samp{InsertBreak}, @samp{InsertTrace},
31309 @samp{InsertFastTrace}, and @samp{Stop}. @var{val} is either 0,
31310 indicating that @value{GDBN} will not request the operation, or 1,
31311 indicating that it may. (The target can then use this to set up its
31312 own internals optimally, for instance if the debugger never expects to
31313 insert breakpoints, it may not need to install its own trap handler.)
31316 @cindex current thread, remote request
31317 @cindex @samp{qC} packet
31318 Return the current thread ID.
31322 @item QC @var{thread-id}
31323 Where @var{thread-id} is a thread ID as documented in
31324 @ref{thread-id syntax}.
31325 @item @r{(anything else)}
31326 Any other reply implies the old thread ID.
31329 @item qCRC:@var{addr},@var{length}
31330 @cindex CRC of memory block, remote request
31331 @cindex @samp{qCRC} packet
31332 Compute the CRC checksum of a block of memory using CRC-32 defined in
31333 IEEE 802.3. The CRC is computed byte at a time, taking the most
31334 significant bit of each byte first. The initial pattern code
31335 @code{0xffffffff} is used to ensure leading zeros affect the CRC.
31337 @emph{Note:} This is the same CRC used in validating separate debug
31338 files (@pxref{Separate Debug Files, , Debugging Information in Separate
31339 Files}). However the algorithm is slightly different. When validating
31340 separate debug files, the CRC is computed taking the @emph{least}
31341 significant bit of each byte first, and the final result is inverted to
31342 detect trailing zeros.
31347 An error (such as memory fault)
31348 @item C @var{crc32}
31349 The specified memory region's checksum is @var{crc32}.
31353 @itemx qsThreadInfo
31354 @cindex list active threads, remote request
31355 @cindex @samp{qfThreadInfo} packet
31356 @cindex @samp{qsThreadInfo} packet
31357 Obtain a list of all active thread IDs from the target (OS). Since there
31358 may be too many active threads to fit into one reply packet, this query
31359 works iteratively: it may require more than one query/reply sequence to
31360 obtain the entire list of threads. The first query of the sequence will
31361 be the @samp{qfThreadInfo} query; subsequent queries in the
31362 sequence will be the @samp{qsThreadInfo} query.
31364 NOTE: This packet replaces the @samp{qL} query (see below).
31368 @item m @var{thread-id}
31370 @item m @var{thread-id},@var{thread-id}@dots{}
31371 a comma-separated list of thread IDs
31373 (lower case letter @samp{L}) denotes end of list.
31376 In response to each query, the target will reply with a list of one or
31377 more thread IDs, separated by commas.
31378 @value{GDBN} will respond to each reply with a request for more thread
31379 ids (using the @samp{qs} form of the query), until the target responds
31380 with @samp{l} (lower-case el, for @dfn{last}).
31381 Refer to @ref{thread-id syntax}, for the format of the @var{thread-id}
31384 @item qGetTLSAddr:@var{thread-id},@var{offset},@var{lm}
31385 @cindex get thread-local storage address, remote request
31386 @cindex @samp{qGetTLSAddr} packet
31387 Fetch the address associated with thread local storage specified
31388 by @var{thread-id}, @var{offset}, and @var{lm}.
31390 @var{thread-id} is the thread ID associated with the
31391 thread for which to fetch the TLS address. @xref{thread-id syntax}.
31393 @var{offset} is the (big endian, hex encoded) offset associated with the
31394 thread local variable. (This offset is obtained from the debug
31395 information associated with the variable.)
31397 @var{lm} is the (big endian, hex encoded) OS/ABI-specific encoding of the
31398 the load module associated with the thread local storage. For example,
31399 a @sc{gnu}/Linux system will pass the link map address of the shared
31400 object associated with the thread local storage under consideration.
31401 Other operating environments may choose to represent the load module
31402 differently, so the precise meaning of this parameter will vary.
31406 @item @var{XX}@dots{}
31407 Hex encoded (big endian) bytes representing the address of the thread
31408 local storage requested.
31411 An error occurred. @var{nn} are hex digits.
31414 An empty reply indicates that @samp{qGetTLSAddr} is not supported by the stub.
31417 @item qGetTIBAddr:@var{thread-id}
31418 @cindex get thread information block address
31419 @cindex @samp{qGetTIBAddr} packet
31420 Fetch address of the Windows OS specific Thread Information Block.
31422 @var{thread-id} is the thread ID associated with the thread.
31426 @item @var{XX}@dots{}
31427 Hex encoded (big endian) bytes representing the linear address of the
31428 thread information block.
31431 An error occured. This means that either the thread was not found, or the
31432 address could not be retrieved.
31435 An empty reply indicates that @samp{qGetTIBAddr} is not supported by the stub.
31438 @item qL @var{startflag} @var{threadcount} @var{nextthread}
31439 Obtain thread information from RTOS. Where: @var{startflag} (one hex
31440 digit) is one to indicate the first query and zero to indicate a
31441 subsequent query; @var{threadcount} (two hex digits) is the maximum
31442 number of threads the response packet can contain; and @var{nextthread}
31443 (eight hex digits), for subsequent queries (@var{startflag} is zero), is
31444 returned in the response as @var{argthread}.
31446 Don't use this packet; use the @samp{qfThreadInfo} query instead (see above).
31450 @item qM @var{count} @var{done} @var{argthread} @var{thread}@dots{}
31451 Where: @var{count} (two hex digits) is the number of threads being
31452 returned; @var{done} (one hex digit) is zero to indicate more threads
31453 and one indicates no further threads; @var{argthreadid} (eight hex
31454 digits) is @var{nextthread} from the request packet; @var{thread}@dots{}
31455 is a sequence of thread IDs from the target. @var{threadid} (eight hex
31456 digits). See @code{remote.c:parse_threadlist_response()}.
31460 @cindex section offsets, remote request
31461 @cindex @samp{qOffsets} packet
31462 Get section offsets that the target used when relocating the downloaded
31467 @item Text=@var{xxx};Data=@var{yyy}@r{[};Bss=@var{zzz}@r{]}
31468 Relocate the @code{Text} section by @var{xxx} from its original address.
31469 Relocate the @code{Data} section by @var{yyy} from its original address.
31470 If the object file format provides segment information (e.g.@: @sc{elf}
31471 @samp{PT_LOAD} program headers), @value{GDBN} will relocate entire
31472 segments by the supplied offsets.
31474 @emph{Note: while a @code{Bss} offset may be included in the response,
31475 @value{GDBN} ignores this and instead applies the @code{Data} offset
31476 to the @code{Bss} section.}
31478 @item TextSeg=@var{xxx}@r{[};DataSeg=@var{yyy}@r{]}
31479 Relocate the first segment of the object file, which conventionally
31480 contains program code, to a starting address of @var{xxx}. If
31481 @samp{DataSeg} is specified, relocate the second segment, which
31482 conventionally contains modifiable data, to a starting address of
31483 @var{yyy}. @value{GDBN} will report an error if the object file
31484 does not contain segment information, or does not contain at least
31485 as many segments as mentioned in the reply. Extra segments are
31486 kept at fixed offsets relative to the last relocated segment.
31489 @item qP @var{mode} @var{thread-id}
31490 @cindex thread information, remote request
31491 @cindex @samp{qP} packet
31492 Returns information on @var{thread-id}. Where: @var{mode} is a hex
31493 encoded 32 bit mode; @var{thread-id} is a thread ID
31494 (@pxref{thread-id syntax}).
31496 Don't use this packet; use the @samp{qThreadExtraInfo} query instead
31499 Reply: see @code{remote.c:remote_unpack_thread_info_response()}.
31503 @cindex non-stop mode, remote request
31504 @cindex @samp{QNonStop} packet
31506 Enter non-stop (@samp{QNonStop:1}) or all-stop (@samp{QNonStop:0}) mode.
31507 @xref{Remote Non-Stop}, for more information.
31512 The request succeeded.
31515 An error occurred. @var{nn} are hex digits.
31518 An empty reply indicates that @samp{QNonStop} is not supported by
31522 This packet is not probed by default; the remote stub must request it,
31523 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
31524 Use of this packet is controlled by the @code{set non-stop} command;
31525 @pxref{Non-Stop Mode}.
31527 @item QPassSignals: @var{signal} @r{[};@var{signal}@r{]}@dots{}
31528 @cindex pass signals to inferior, remote request
31529 @cindex @samp{QPassSignals} packet
31530 @anchor{QPassSignals}
31531 Each listed @var{signal} should be passed directly to the inferior process.
31532 Signals are numbered identically to continue packets and stop replies
31533 (@pxref{Stop Reply Packets}). Each @var{signal} list item should be
31534 strictly greater than the previous item. These signals do not need to stop
31535 the inferior, or be reported to @value{GDBN}. All other signals should be
31536 reported to @value{GDBN}. Multiple @samp{QPassSignals} packets do not
31537 combine; any earlier @samp{QPassSignals} list is completely replaced by the
31538 new list. This packet improves performance when using @samp{handle
31539 @var{signal} nostop noprint pass}.
31544 The request succeeded.
31547 An error occurred. @var{nn} are hex digits.
31550 An empty reply indicates that @samp{QPassSignals} is not supported by
31554 Use of this packet is controlled by the @code{set remote pass-signals}
31555 command (@pxref{Remote Configuration, set remote pass-signals}).
31556 This packet is not probed by default; the remote stub must request it,
31557 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
31559 @item qRcmd,@var{command}
31560 @cindex execute remote command, remote request
31561 @cindex @samp{qRcmd} packet
31562 @var{command} (hex encoded) is passed to the local interpreter for
31563 execution. Invalid commands should be reported using the output
31564 string. Before the final result packet, the target may also respond
31565 with a number of intermediate @samp{O@var{output}} console output
31566 packets. @emph{Implementors should note that providing access to a
31567 stubs's interpreter may have security implications}.
31572 A command response with no output.
31574 A command response with the hex encoded output string @var{OUTPUT}.
31576 Indicate a badly formed request.
31578 An empty reply indicates that @samp{qRcmd} is not recognized.
31581 (Note that the @code{qRcmd} packet's name is separated from the
31582 command by a @samp{,}, not a @samp{:}, contrary to the naming
31583 conventions above. Please don't use this packet as a model for new
31586 @item qSearch:memory:@var{address};@var{length};@var{search-pattern}
31587 @cindex searching memory, in remote debugging
31588 @cindex @samp{qSearch:memory} packet
31589 @anchor{qSearch memory}
31590 Search @var{length} bytes at @var{address} for @var{search-pattern}.
31591 @var{address} and @var{length} are encoded in hex.
31592 @var{search-pattern} is a sequence of bytes, hex encoded.
31597 The pattern was not found.
31599 The pattern was found at @var{address}.
31601 A badly formed request or an error was encountered while searching memory.
31603 An empty reply indicates that @samp{qSearch:memory} is not recognized.
31606 @item QStartNoAckMode
31607 @cindex @samp{QStartNoAckMode} packet
31608 @anchor{QStartNoAckMode}
31609 Request that the remote stub disable the normal @samp{+}/@samp{-}
31610 protocol acknowledgments (@pxref{Packet Acknowledgment}).
31615 The stub has switched to no-acknowledgment mode.
31616 @value{GDBN} acknowledges this reponse,
31617 but neither the stub nor @value{GDBN} shall send or expect further
31618 @samp{+}/@samp{-} acknowledgments in the current connection.
31620 An empty reply indicates that the stub does not support no-acknowledgment mode.
31623 @item qSupported @r{[}:@var{gdbfeature} @r{[};@var{gdbfeature}@r{]}@dots{} @r{]}
31624 @cindex supported packets, remote query
31625 @cindex features of the remote protocol
31626 @cindex @samp{qSupported} packet
31627 @anchor{qSupported}
31628 Tell the remote stub about features supported by @value{GDBN}, and
31629 query the stub for features it supports. This packet allows
31630 @value{GDBN} and the remote stub to take advantage of each others'
31631 features. @samp{qSupported} also consolidates multiple feature probes
31632 at startup, to improve @value{GDBN} performance---a single larger
31633 packet performs better than multiple smaller probe packets on
31634 high-latency links. Some features may enable behavior which must not
31635 be on by default, e.g.@: because it would confuse older clients or
31636 stubs. Other features may describe packets which could be
31637 automatically probed for, but are not. These features must be
31638 reported before @value{GDBN} will use them. This ``default
31639 unsupported'' behavior is not appropriate for all packets, but it
31640 helps to keep the initial connection time under control with new
31641 versions of @value{GDBN} which support increasing numbers of packets.
31645 @item @var{stubfeature} @r{[};@var{stubfeature}@r{]}@dots{}
31646 The stub supports or does not support each returned @var{stubfeature},
31647 depending on the form of each @var{stubfeature} (see below for the
31650 An empty reply indicates that @samp{qSupported} is not recognized,
31651 or that no features needed to be reported to @value{GDBN}.
31654 The allowed forms for each feature (either a @var{gdbfeature} in the
31655 @samp{qSupported} packet, or a @var{stubfeature} in the response)
31659 @item @var{name}=@var{value}
31660 The remote protocol feature @var{name} is supported, and associated
31661 with the specified @var{value}. The format of @var{value} depends
31662 on the feature, but it must not include a semicolon.
31664 The remote protocol feature @var{name} is supported, and does not
31665 need an associated value.
31667 The remote protocol feature @var{name} is not supported.
31669 The remote protocol feature @var{name} may be supported, and
31670 @value{GDBN} should auto-detect support in some other way when it is
31671 needed. This form will not be used for @var{gdbfeature} notifications,
31672 but may be used for @var{stubfeature} responses.
31675 Whenever the stub receives a @samp{qSupported} request, the
31676 supplied set of @value{GDBN} features should override any previous
31677 request. This allows @value{GDBN} to put the stub in a known
31678 state, even if the stub had previously been communicating with
31679 a different version of @value{GDBN}.
31681 The following values of @var{gdbfeature} (for the packet sent by @value{GDBN})
31686 This feature indicates whether @value{GDBN} supports multiprocess
31687 extensions to the remote protocol. @value{GDBN} does not use such
31688 extensions unless the stub also reports that it supports them by
31689 including @samp{multiprocess+} in its @samp{qSupported} reply.
31690 @xref{multiprocess extensions}, for details.
31693 This feature indicates that @value{GDBN} supports the XML target
31694 description. If the stub sees @samp{xmlRegisters=} with target
31695 specific strings separated by a comma, it will report register
31699 This feature indicates whether @value{GDBN} supports the
31700 @samp{qRelocInsn} packet (@pxref{Tracepoint Packets,,Relocate
31701 instruction reply packet}).
31704 Stubs should ignore any unknown values for
31705 @var{gdbfeature}. Any @value{GDBN} which sends a @samp{qSupported}
31706 packet supports receiving packets of unlimited length (earlier
31707 versions of @value{GDBN} may reject overly long responses). Additional values
31708 for @var{gdbfeature} may be defined in the future to let the stub take
31709 advantage of new features in @value{GDBN}, e.g.@: incompatible
31710 improvements in the remote protocol---the @samp{multiprocess} feature is
31711 an example of such a feature. The stub's reply should be independent
31712 of the @var{gdbfeature} entries sent by @value{GDBN}; first @value{GDBN}
31713 describes all the features it supports, and then the stub replies with
31714 all the features it supports.
31716 Similarly, @value{GDBN} will silently ignore unrecognized stub feature
31717 responses, as long as each response uses one of the standard forms.
31719 Some features are flags. A stub which supports a flag feature
31720 should respond with a @samp{+} form response. Other features
31721 require values, and the stub should respond with an @samp{=}
31724 Each feature has a default value, which @value{GDBN} will use if
31725 @samp{qSupported} is not available or if the feature is not mentioned
31726 in the @samp{qSupported} response. The default values are fixed; a
31727 stub is free to omit any feature responses that match the defaults.
31729 Not all features can be probed, but for those which can, the probing
31730 mechanism is useful: in some cases, a stub's internal
31731 architecture may not allow the protocol layer to know some information
31732 about the underlying target in advance. This is especially common in
31733 stubs which may be configured for multiple targets.
31735 These are the currently defined stub features and their properties:
31737 @multitable @columnfractions 0.35 0.2 0.12 0.2
31738 @c NOTE: The first row should be @headitem, but we do not yet require
31739 @c a new enough version of Texinfo (4.7) to use @headitem.
31741 @tab Value Required
31745 @item @samp{PacketSize}
31750 @item @samp{qXfer:auxv:read}
31755 @item @samp{qXfer:features:read}
31760 @item @samp{qXfer:libraries:read}
31765 @item @samp{qXfer:memory-map:read}
31770 @item @samp{qXfer:spu:read}
31775 @item @samp{qXfer:spu:write}
31780 @item @samp{qXfer:siginfo:read}
31785 @item @samp{qXfer:siginfo:write}
31790 @item @samp{qXfer:threads:read}
31796 @item @samp{QNonStop}
31801 @item @samp{QPassSignals}
31806 @item @samp{QStartNoAckMode}
31811 @item @samp{multiprocess}
31816 @item @samp{ConditionalTracepoints}
31821 @item @samp{ReverseContinue}
31826 @item @samp{ReverseStep}
31831 @item @samp{TracepointSource}
31836 @item @samp{QAllow}
31843 These are the currently defined stub features, in more detail:
31846 @cindex packet size, remote protocol
31847 @item PacketSize=@var{bytes}
31848 The remote stub can accept packets up to at least @var{bytes} in
31849 length. @value{GDBN} will send packets up to this size for bulk
31850 transfers, and will never send larger packets. This is a limit on the
31851 data characters in the packet, including the frame and checksum.
31852 There is no trailing NUL byte in a remote protocol packet; if the stub
31853 stores packets in a NUL-terminated format, it should allow an extra
31854 byte in its buffer for the NUL. If this stub feature is not supported,
31855 @value{GDBN} guesses based on the size of the @samp{g} packet response.
31857 @item qXfer:auxv:read
31858 The remote stub understands the @samp{qXfer:auxv:read} packet
31859 (@pxref{qXfer auxiliary vector read}).
31861 @item qXfer:features:read
31862 The remote stub understands the @samp{qXfer:features:read} packet
31863 (@pxref{qXfer target description read}).
31865 @item qXfer:libraries:read
31866 The remote stub understands the @samp{qXfer:libraries:read} packet
31867 (@pxref{qXfer library list read}).
31869 @item qXfer:memory-map:read
31870 The remote stub understands the @samp{qXfer:memory-map:read} packet
31871 (@pxref{qXfer memory map read}).
31873 @item qXfer:spu:read
31874 The remote stub understands the @samp{qXfer:spu:read} packet
31875 (@pxref{qXfer spu read}).
31877 @item qXfer:spu:write
31878 The remote stub understands the @samp{qXfer:spu:write} packet
31879 (@pxref{qXfer spu write}).
31881 @item qXfer:siginfo:read
31882 The remote stub understands the @samp{qXfer:siginfo:read} packet
31883 (@pxref{qXfer siginfo read}).
31885 @item qXfer:siginfo:write
31886 The remote stub understands the @samp{qXfer:siginfo:write} packet
31887 (@pxref{qXfer siginfo write}).
31889 @item qXfer:threads:read
31890 The remote stub understands the @samp{qXfer:threads:read} packet
31891 (@pxref{qXfer threads read}).
31894 The remote stub understands the @samp{QNonStop} packet
31895 (@pxref{QNonStop}).
31898 The remote stub understands the @samp{QPassSignals} packet
31899 (@pxref{QPassSignals}).
31901 @item QStartNoAckMode
31902 The remote stub understands the @samp{QStartNoAckMode} packet and
31903 prefers to operate in no-acknowledgment mode. @xref{Packet Acknowledgment}.
31906 @anchor{multiprocess extensions}
31907 @cindex multiprocess extensions, in remote protocol
31908 The remote stub understands the multiprocess extensions to the remote
31909 protocol syntax. The multiprocess extensions affect the syntax of
31910 thread IDs in both packets and replies (@pxref{thread-id syntax}), and
31911 add process IDs to the @samp{D} packet and @samp{W} and @samp{X}
31912 replies. Note that reporting this feature indicates support for the
31913 syntactic extensions only, not that the stub necessarily supports
31914 debugging of more than one process at a time. The stub must not use
31915 multiprocess extensions in packet replies unless @value{GDBN} has also
31916 indicated it supports them in its @samp{qSupported} request.
31918 @item qXfer:osdata:read
31919 The remote stub understands the @samp{qXfer:osdata:read} packet
31920 ((@pxref{qXfer osdata read}).
31922 @item ConditionalTracepoints
31923 The remote stub accepts and implements conditional expressions defined
31924 for tracepoints (@pxref{Tracepoint Conditions}).
31926 @item ReverseContinue
31927 The remote stub accepts and implements the reverse continue packet
31931 The remote stub accepts and implements the reverse step packet
31934 @item TracepointSource
31935 The remote stub understands the @samp{QTDPsrc} packet that supplies
31936 the source form of tracepoint definitions.
31939 The remote stub understands the @samp{QAllow} packet.
31944 @cindex symbol lookup, remote request
31945 @cindex @samp{qSymbol} packet
31946 Notify the target that @value{GDBN} is prepared to serve symbol lookup
31947 requests. Accept requests from the target for the values of symbols.
31952 The target does not need to look up any (more) symbols.
31953 @item qSymbol:@var{sym_name}
31954 The target requests the value of symbol @var{sym_name} (hex encoded).
31955 @value{GDBN} may provide the value by using the
31956 @samp{qSymbol:@var{sym_value}:@var{sym_name}} message, described
31960 @item qSymbol:@var{sym_value}:@var{sym_name}
31961 Set the value of @var{sym_name} to @var{sym_value}.
31963 @var{sym_name} (hex encoded) is the name of a symbol whose value the
31964 target has previously requested.
31966 @var{sym_value} (hex) is the value for symbol @var{sym_name}. If
31967 @value{GDBN} cannot supply a value for @var{sym_name}, then this field
31973 The target does not need to look up any (more) symbols.
31974 @item qSymbol:@var{sym_name}
31975 The target requests the value of a new symbol @var{sym_name} (hex
31976 encoded). @value{GDBN} will continue to supply the values of symbols
31977 (if available), until the target ceases to request them.
31982 @item QTDisconnected
31989 @xref{Tracepoint Packets}.
31991 @item qThreadExtraInfo,@var{thread-id}
31992 @cindex thread attributes info, remote request
31993 @cindex @samp{qThreadExtraInfo} packet
31994 Obtain a printable string description of a thread's attributes from
31995 the target OS. @var{thread-id} is a thread ID;
31996 see @ref{thread-id syntax}. This
31997 string may contain anything that the target OS thinks is interesting
31998 for @value{GDBN} to tell the user about the thread. The string is
31999 displayed in @value{GDBN}'s @code{info threads} display. Some
32000 examples of possible thread extra info strings are @samp{Runnable}, or
32001 @samp{Blocked on Mutex}.
32005 @item @var{XX}@dots{}
32006 Where @samp{@var{XX}@dots{}} is a hex encoding of @sc{ascii} data,
32007 comprising the printable string containing the extra information about
32008 the thread's attributes.
32011 (Note that the @code{qThreadExtraInfo} packet's name is separated from
32012 the command by a @samp{,}, not a @samp{:}, contrary to the naming
32013 conventions above. Please don't use this packet as a model for new
32025 @xref{Tracepoint Packets}.
32027 @item qXfer:@var{object}:read:@var{annex}:@var{offset},@var{length}
32028 @cindex read special object, remote request
32029 @cindex @samp{qXfer} packet
32030 @anchor{qXfer read}
32031 Read uninterpreted bytes from the target's special data area
32032 identified by the keyword @var{object}. Request @var{length} bytes
32033 starting at @var{offset} bytes into the data. The content and
32034 encoding of @var{annex} is specific to @var{object}; it can supply
32035 additional details about what data to access.
32037 Here are the specific requests of this form defined so far. All
32038 @samp{qXfer:@var{object}:read:@dots{}} requests use the same reply
32039 formats, listed below.
32042 @item qXfer:auxv:read::@var{offset},@var{length}
32043 @anchor{qXfer auxiliary vector read}
32044 Access the target's @dfn{auxiliary vector}. @xref{OS Information,
32045 auxiliary vector}. Note @var{annex} must be empty.
32047 This packet is not probed by default; the remote stub must request it,
32048 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
32050 @item qXfer:features:read:@var{annex}:@var{offset},@var{length}
32051 @anchor{qXfer target description read}
32052 Access the @dfn{target description}. @xref{Target Descriptions}. The
32053 annex specifies which XML document to access. The main description is
32054 always loaded from the @samp{target.xml} annex.
32056 This packet is not probed by default; the remote stub must request it,
32057 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
32059 @item qXfer:libraries:read:@var{annex}:@var{offset},@var{length}
32060 @anchor{qXfer library list read}
32061 Access the target's list of loaded libraries. @xref{Library List Format}.
32062 The annex part of the generic @samp{qXfer} packet must be empty
32063 (@pxref{qXfer read}).
32065 Targets which maintain a list of libraries in the program's memory do
32066 not need to implement this packet; it is designed for platforms where
32067 the operating system manages the list of loaded libraries.
32069 This packet is not probed by default; the remote stub must request it,
32070 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
32072 @item qXfer:memory-map:read::@var{offset},@var{length}
32073 @anchor{qXfer memory map read}
32074 Access the target's @dfn{memory-map}. @xref{Memory Map Format}. The
32075 annex part of the generic @samp{qXfer} packet must be empty
32076 (@pxref{qXfer read}).
32078 This packet is not probed by default; the remote stub must request it,
32079 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
32081 @item qXfer:siginfo:read::@var{offset},@var{length}
32082 @anchor{qXfer siginfo read}
32083 Read contents of the extra signal information on the target
32084 system. The annex part of the generic @samp{qXfer} packet must be
32085 empty (@pxref{qXfer read}).
32087 This packet is not probed by default; the remote stub must request it,
32088 by supplying an appropriate @samp{qSupported} response
32089 (@pxref{qSupported}).
32091 @item qXfer:spu:read:@var{annex}:@var{offset},@var{length}
32092 @anchor{qXfer spu read}
32093 Read contents of an @code{spufs} file on the target system. The
32094 annex specifies which file to read; it must be of the form
32095 @file{@var{id}/@var{name}}, where @var{id} specifies an SPU context ID
32096 in the target process, and @var{name} identifes the @code{spufs} file
32097 in that context to be accessed.
32099 This packet is not probed by default; the remote stub must request it,
32100 by supplying an appropriate @samp{qSupported} response
32101 (@pxref{qSupported}).
32103 @item qXfer:threads:read::@var{offset},@var{length}
32104 @anchor{qXfer threads read}
32105 Access the list of threads on target. @xref{Thread List Format}. The
32106 annex part of the generic @samp{qXfer} packet must be empty
32107 (@pxref{qXfer read}).
32109 This packet is not probed by default; the remote stub must request it,
32110 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
32112 @item qXfer:osdata:read::@var{offset},@var{length}
32113 @anchor{qXfer osdata read}
32114 Access the target's @dfn{operating system information}.
32115 @xref{Operating System Information}.
32122 Data @var{data} (@pxref{Binary Data}) has been read from the
32123 target. There may be more data at a higher address (although
32124 it is permitted to return @samp{m} even for the last valid
32125 block of data, as long as at least one byte of data was read).
32126 @var{data} may have fewer bytes than the @var{length} in the
32130 Data @var{data} (@pxref{Binary Data}) has been read from the target.
32131 There is no more data to be read. @var{data} may have fewer bytes
32132 than the @var{length} in the request.
32135 The @var{offset} in the request is at the end of the data.
32136 There is no more data to be read.
32139 The request was malformed, or @var{annex} was invalid.
32142 The offset was invalid, or there was an error encountered reading the data.
32143 @var{nn} is a hex-encoded @code{errno} value.
32146 An empty reply indicates the @var{object} string was not recognized by
32147 the stub, or that the object does not support reading.
32150 @item qXfer:@var{object}:write:@var{annex}:@var{offset}:@var{data}@dots{}
32151 @cindex write data into object, remote request
32152 @anchor{qXfer write}
32153 Write uninterpreted bytes into the target's special data area
32154 identified by the keyword @var{object}, starting at @var{offset} bytes
32155 into the data. @var{data}@dots{} is the binary-encoded data
32156 (@pxref{Binary Data}) to be written. The content and encoding of @var{annex}
32157 is specific to @var{object}; it can supply additional details about what data
32160 Here are the specific requests of this form defined so far. All
32161 @samp{qXfer:@var{object}:write:@dots{}} requests use the same reply
32162 formats, listed below.
32165 @item qXfer:siginfo:write::@var{offset}:@var{data}@dots{}
32166 @anchor{qXfer siginfo write}
32167 Write @var{data} to the extra signal information on the target system.
32168 The annex part of the generic @samp{qXfer} packet must be
32169 empty (@pxref{qXfer write}).
32171 This packet is not probed by default; the remote stub must request it,
32172 by supplying an appropriate @samp{qSupported} response
32173 (@pxref{qSupported}).
32175 @item qXfer:spu:write:@var{annex}:@var{offset}:@var{data}@dots{}
32176 @anchor{qXfer spu write}
32177 Write @var{data} to an @code{spufs} file on the target system. The
32178 annex specifies which file to write; it must be of the form
32179 @file{@var{id}/@var{name}}, where @var{id} specifies an SPU context ID
32180 in the target process, and @var{name} identifes the @code{spufs} file
32181 in that context to be accessed.
32183 This packet is not probed by default; the remote stub must request it,
32184 by supplying an appropriate @samp{qSupported} response (@pxref{qSupported}).
32190 @var{nn} (hex encoded) is the number of bytes written.
32191 This may be fewer bytes than supplied in the request.
32194 The request was malformed, or @var{annex} was invalid.
32197 The offset was invalid, or there was an error encountered writing the data.
32198 @var{nn} is a hex-encoded @code{errno} value.
32201 An empty reply indicates the @var{object} string was not
32202 recognized by the stub, or that the object does not support writing.
32205 @item qXfer:@var{object}:@var{operation}:@dots{}
32206 Requests of this form may be added in the future. When a stub does
32207 not recognize the @var{object} keyword, or its support for
32208 @var{object} does not recognize the @var{operation} keyword, the stub
32209 must respond with an empty packet.
32211 @item qAttached:@var{pid}
32212 @cindex query attached, remote request
32213 @cindex @samp{qAttached} packet
32214 Return an indication of whether the remote server attached to an
32215 existing process or created a new process. When the multiprocess
32216 protocol extensions are supported (@pxref{multiprocess extensions}),
32217 @var{pid} is an integer in hexadecimal format identifying the target
32218 process. Otherwise, @value{GDBN} will omit the @var{pid} field and
32219 the query packet will be simplified as @samp{qAttached}.
32221 This query is used, for example, to know whether the remote process
32222 should be detached or killed when a @value{GDBN} session is ended with
32223 the @code{quit} command.
32228 The remote server attached to an existing process.
32230 The remote server created a new process.
32232 A badly formed request or an error was encountered.
32237 @node Architecture-Specific Protocol Details
32238 @section Architecture-Specific Protocol Details
32240 This section describes how the remote protocol is applied to specific
32241 target architectures. Also see @ref{Standard Target Features}, for
32242 details of XML target descriptions for each architecture.
32246 @subsubsection Breakpoint Kinds
32248 These breakpoint kinds are defined for the @samp{Z0} and @samp{Z1} packets.
32253 16-bit Thumb mode breakpoint.
32256 32-bit Thumb mode (Thumb-2) breakpoint.
32259 32-bit ARM mode breakpoint.
32265 @subsubsection Register Packet Format
32267 The following @code{g}/@code{G} packets have previously been defined.
32268 In the below, some thirty-two bit registers are transferred as
32269 sixty-four bits. Those registers should be zero/sign extended (which?)
32270 to fill the space allocated. Register bytes are transferred in target
32271 byte order. The two nibbles within a register byte are transferred
32272 most-significant - least-significant.
32278 All registers are transferred as thirty-two bit quantities in the order:
32279 32 general-purpose; sr; lo; hi; bad; cause; pc; 32 floating-point
32280 registers; fsr; fir; fp.
32284 All registers are transferred as sixty-four bit quantities (including
32285 thirty-two bit registers such as @code{sr}). The ordering is the same
32290 @node Tracepoint Packets
32291 @section Tracepoint Packets
32292 @cindex tracepoint packets
32293 @cindex packets, tracepoint
32295 Here we describe the packets @value{GDBN} uses to implement
32296 tracepoints (@pxref{Tracepoints}).
32300 @item QTDP:@var{n}:@var{addr}:@var{ena}:@var{step}:@var{pass}[:F@var{flen}][:X@var{len},@var{bytes}]@r{[}-@r{]}
32301 Create a new tracepoint, number @var{n}, at @var{addr}. If @var{ena}
32302 is @samp{E}, then the tracepoint is enabled; if it is @samp{D}, then
32303 the tracepoint is disabled. @var{step} is the tracepoint's step
32304 count, and @var{pass} is its pass count. If an @samp{F} is present,
32305 then the tracepoint is to be a fast tracepoint, and the @var{flen} is
32306 the number of bytes that the target should copy elsewhere to make room
32307 for the tracepoint. If an @samp{X} is present, it introduces a
32308 tracepoint condition, which consists of a hexadecimal length, followed
32309 by a comma and hex-encoded bytes, in a manner similar to action
32310 encodings as described below. If the trailing @samp{-} is present,
32311 further @samp{QTDP} packets will follow to specify this tracepoint's
32317 The packet was understood and carried out.
32319 @xref{Tracepoint Packets,,Relocate instruction reply packet}.
32321 The packet was not recognized.
32324 @item QTDP:-@var{n}:@var{addr}:@r{[}S@r{]}@var{action}@dots{}@r{[}-@r{]}
32325 Define actions to be taken when a tracepoint is hit. @var{n} and
32326 @var{addr} must be the same as in the initial @samp{QTDP} packet for
32327 this tracepoint. This packet may only be sent immediately after
32328 another @samp{QTDP} packet that ended with a @samp{-}. If the
32329 trailing @samp{-} is present, further @samp{QTDP} packets will follow,
32330 specifying more actions for this tracepoint.
32332 In the series of action packets for a given tracepoint, at most one
32333 can have an @samp{S} before its first @var{action}. If such a packet
32334 is sent, it and the following packets define ``while-stepping''
32335 actions. Any prior packets define ordinary actions --- that is, those
32336 taken when the tracepoint is first hit. If no action packet has an
32337 @samp{S}, then all the packets in the series specify ordinary
32338 tracepoint actions.
32340 The @samp{@var{action}@dots{}} portion of the packet is a series of
32341 actions, concatenated without separators. Each action has one of the
32347 Collect the registers whose bits are set in @var{mask}. @var{mask} is
32348 a hexadecimal number whose @var{i}'th bit is set if register number
32349 @var{i} should be collected. (The least significant bit is numbered
32350 zero.) Note that @var{mask} may be any number of digits long; it may
32351 not fit in a 32-bit word.
32353 @item M @var{basereg},@var{offset},@var{len}
32354 Collect @var{len} bytes of memory starting at the address in register
32355 number @var{basereg}, plus @var{offset}. If @var{basereg} is
32356 @samp{-1}, then the range has a fixed address: @var{offset} is the
32357 address of the lowest byte to collect. The @var{basereg},
32358 @var{offset}, and @var{len} parameters are all unsigned hexadecimal
32359 values (the @samp{-1} value for @var{basereg} is a special case).
32361 @item X @var{len},@var{expr}
32362 Evaluate @var{expr}, whose length is @var{len}, and collect memory as
32363 it directs. @var{expr} is an agent expression, as described in
32364 @ref{Agent Expressions}. Each byte of the expression is encoded as a
32365 two-digit hex number in the packet; @var{len} is the number of bytes
32366 in the expression (and thus one-half the number of hex digits in the
32371 Any number of actions may be packed together in a single @samp{QTDP}
32372 packet, as long as the packet does not exceed the maximum packet
32373 length (400 bytes, for many stubs). There may be only one @samp{R}
32374 action per tracepoint, and it must precede any @samp{M} or @samp{X}
32375 actions. Any registers referred to by @samp{M} and @samp{X} actions
32376 must be collected by a preceding @samp{R} action. (The
32377 ``while-stepping'' actions are treated as if they were attached to a
32378 separate tracepoint, as far as these restrictions are concerned.)
32383 The packet was understood and carried out.
32385 @xref{Tracepoint Packets,,Relocate instruction reply packet}.
32387 The packet was not recognized.
32390 @item QTDPsrc:@var{n}:@var{addr}:@var{type}:@var{start}:@var{slen}:@var{bytes}
32391 @cindex @samp{QTDPsrc} packet
32392 Specify a source string of tracepoint @var{n} at address @var{addr}.
32393 This is useful to get accurate reproduction of the tracepoints
32394 originally downloaded at the beginning of the trace run. @var{type}
32395 is the name of the tracepoint part, such as @samp{cond} for the
32396 tracepoint's conditional expression (see below for a list of types), while
32397 @var{bytes} is the string, encoded in hexadecimal.
32399 @var{start} is the offset of the @var{bytes} within the overall source
32400 string, while @var{slen} is the total length of the source string.
32401 This is intended for handling source strings that are longer than will
32402 fit in a single packet.
32403 @c Add detailed example when this info is moved into a dedicated
32404 @c tracepoint descriptions section.
32406 The available string types are @samp{at} for the location,
32407 @samp{cond} for the conditional, and @samp{cmd} for an action command.
32408 @value{GDBN} sends a separate packet for each command in the action
32409 list, in the same order in which the commands are stored in the list.
32411 The target does not need to do anything with source strings except
32412 report them back as part of the replies to the @samp{qTfP}/@samp{qTsP}
32415 Although this packet is optional, and @value{GDBN} will only send it
32416 if the target replies with @samp{TracepointSource} @xref{General
32417 Query Packets}, it makes both disconnected tracing and trace files
32418 much easier to use. Otherwise the user must be careful that the
32419 tracepoints in effect while looking at trace frames are identical to
32420 the ones in effect during the trace run; even a small discrepancy
32421 could cause @samp{tdump} not to work, or a particular trace frame not
32424 @item QTDV:@var{n}:@var{value}
32425 @cindex define trace state variable, remote request
32426 @cindex @samp{QTDV} packet
32427 Create a new trace state variable, number @var{n}, with an initial
32428 value of @var{value}, which is a 64-bit signed integer. Both @var{n}
32429 and @var{value} are encoded as hexadecimal values. @value{GDBN} has
32430 the option of not using this packet for initial values of zero; the
32431 target should simply create the trace state variables as they are
32432 mentioned in expressions.
32434 @item QTFrame:@var{n}
32435 Select the @var{n}'th tracepoint frame from the buffer, and use the
32436 register and memory contents recorded there to answer subsequent
32437 request packets from @value{GDBN}.
32439 A successful reply from the stub indicates that the stub has found the
32440 requested frame. The response is a series of parts, concatenated
32441 without separators, describing the frame we selected. Each part has
32442 one of the following forms:
32446 The selected frame is number @var{n} in the trace frame buffer;
32447 @var{f} is a hexadecimal number. If @var{f} is @samp{-1}, then there
32448 was no frame matching the criteria in the request packet.
32451 The selected trace frame records a hit of tracepoint number @var{t};
32452 @var{t} is a hexadecimal number.
32456 @item QTFrame:pc:@var{addr}
32457 Like @samp{QTFrame:@var{n}}, but select the first tracepoint frame after the
32458 currently selected frame whose PC is @var{addr};
32459 @var{addr} is a hexadecimal number.
32461 @item QTFrame:tdp:@var{t}
32462 Like @samp{QTFrame:@var{n}}, but select the first tracepoint frame after the
32463 currently selected frame that is a hit of tracepoint @var{t}; @var{t}
32464 is a hexadecimal number.
32466 @item QTFrame:range:@var{start}:@var{end}
32467 Like @samp{QTFrame:@var{n}}, but select the first tracepoint frame after the
32468 currently selected frame whose PC is between @var{start} (inclusive)
32469 and @var{end} (inclusive); @var{start} and @var{end} are hexadecimal
32472 @item QTFrame:outside:@var{start}:@var{end}
32473 Like @samp{QTFrame:range:@var{start}:@var{end}}, but select the first
32474 frame @emph{outside} the given range of addresses (exclusive).
32477 Begin the tracepoint experiment. Begin collecting data from
32478 tracepoint hits in the trace frame buffer. This packet supports the
32479 @samp{qRelocInsn} reply (@pxref{Tracepoint Packets,,Relocate
32480 instruction reply packet}).
32483 End the tracepoint experiment. Stop collecting trace frames.
32486 Clear the table of tracepoints, and empty the trace frame buffer.
32488 @item QTro:@var{start1},@var{end1}:@var{start2},@var{end2}:@dots{}
32489 Establish the given ranges of memory as ``transparent''. The stub
32490 will answer requests for these ranges from memory's current contents,
32491 if they were not collected as part of the tracepoint hit.
32493 @value{GDBN} uses this to mark read-only regions of memory, like those
32494 containing program code. Since these areas never change, they should
32495 still have the same contents they did when the tracepoint was hit, so
32496 there's no reason for the stub to refuse to provide their contents.
32498 @item QTDisconnected:@var{value}
32499 Set the choice to what to do with the tracing run when @value{GDBN}
32500 disconnects from the target. A @var{value} of 1 directs the target to
32501 continue the tracing run, while 0 tells the target to stop tracing if
32502 @value{GDBN} is no longer in the picture.
32505 Ask the stub if there is a trace experiment running right now.
32507 The reply has the form:
32511 @item T@var{running}@r{[};@var{field}@r{]}@dots{}
32512 @var{running} is a single digit @code{1} if the trace is presently
32513 running, or @code{0} if not. It is followed by semicolon-separated
32514 optional fields that an agent may use to report additional status.
32518 If the trace is not running, the agent may report any of several
32519 explanations as one of the optional fields:
32524 No trace has been run yet.
32527 The trace was stopped by a user-originated stop command.
32530 The trace stopped because the trace buffer filled up.
32532 @item tdisconnected:0
32533 The trace stopped because @value{GDBN} disconnected from the target.
32535 @item tpasscount:@var{tpnum}
32536 The trace stopped because tracepoint @var{tpnum} exceeded its pass count.
32538 @item terror:@var{text}:@var{tpnum}
32539 The trace stopped because tracepoint @var{tpnum} had an error. The
32540 string @var{text} is available to describe the nature of the error
32541 (for instance, a divide by zero in the condition expression).
32542 @var{text} is hex encoded.
32545 The trace stopped for some other reason.
32549 Additional optional fields supply statistical and other information.
32550 Although not required, they are extremely useful for users monitoring
32551 the progress of a trace run. If a trace has stopped, and these
32552 numbers are reported, they must reflect the state of the just-stopped
32557 @item tframes:@var{n}
32558 The number of trace frames in the buffer.
32560 @item tcreated:@var{n}
32561 The total number of trace frames created during the run. This may
32562 be larger than the trace frame count, if the buffer is circular.
32564 @item tsize:@var{n}
32565 The total size of the trace buffer, in bytes.
32567 @item tfree:@var{n}
32568 The number of bytes still unused in the buffer.
32570 @item circular:@var{n}
32571 The value of the circular trace buffer flag. @code{1} means that the
32572 trace buffer is circular and old trace frames will be discarded if
32573 necessary to make room, @code{0} means that the trace buffer is linear
32576 @item disconn:@var{n}
32577 The value of the disconnected tracing flag. @code{1} means that
32578 tracing will continue after @value{GDBN} disconnects, @code{0} means
32579 that the trace run will stop.
32583 @item qTV:@var{var}
32584 @cindex trace state variable value, remote request
32585 @cindex @samp{qTV} packet
32586 Ask the stub for the value of the trace state variable number @var{var}.
32591 The value of the variable is @var{value}. This will be the current
32592 value of the variable if the user is examining a running target, or a
32593 saved value if the variable was collected in the trace frame that the
32594 user is looking at. Note that multiple requests may result in
32595 different reply values, such as when requesting values while the
32596 program is running.
32599 The value of the variable is unknown. This would occur, for example,
32600 if the user is examining a trace frame in which the requested variable
32606 These packets request data about tracepoints that are being used by
32607 the target. @value{GDBN} sends @code{qTfP} to get the first piece
32608 of data, and multiple @code{qTsP} to get additional pieces. Replies
32609 to these packets generally take the form of the @code{QTDP} packets
32610 that define tracepoints. (FIXME add detailed syntax)
32614 These packets request data about trace state variables that are on the
32615 target. @value{GDBN} sends @code{qTfV} to get the first vari of data,
32616 and multiple @code{qTsV} to get additional variables. Replies to
32617 these packets follow the syntax of the @code{QTDV} packets that define
32618 trace state variables.
32620 @item QTSave:@var{filename}
32621 This packet directs the target to save trace data to the file name
32622 @var{filename} in the target's filesystem. @var{filename} is encoded
32623 as a hex string; the interpretation of the file name (relative vs
32624 absolute, wild cards, etc) is up to the target.
32626 @item qTBuffer:@var{offset},@var{len}
32627 Return up to @var{len} bytes of the current contents of trace buffer,
32628 starting at @var{offset}. The trace buffer is treated as if it were
32629 a contiguous collection of traceframes, as per the trace file format.
32630 The reply consists as many hex-encoded bytes as the target can deliver
32631 in a packet; it is not an error to return fewer than were asked for.
32632 A reply consisting of just @code{l} indicates that no bytes are
32635 @item QTBuffer:circular:@var{value}
32636 This packet directs the target to use a circular trace buffer if
32637 @var{value} is 1, or a linear buffer if the value is 0.
32641 @subsection Relocate instruction reply packet
32642 When installing fast tracepoints in memory, the target may need to
32643 relocate the instruction currently at the tracepoint address to a
32644 different address in memory. For most instructions, a simple copy is
32645 enough, but, for example, call instructions that implicitly push the
32646 return address on the stack, and relative branches or other
32647 PC-relative instructions require offset adjustment, so that the effect
32648 of executing the instruction at a different address is the same as if
32649 it had executed in the original location.
32651 In response to several of the tracepoint packets, the target may also
32652 respond with a number of intermediate @samp{qRelocInsn} request
32653 packets before the final result packet, to have @value{GDBN} handle
32654 this relocation operation. If a packet supports this mechanism, its
32655 documentation will explicitly say so. See for example the above
32656 descriptions for the @samp{QTStart} and @samp{QTDP} packets. The
32657 format of the request is:
32660 @item qRelocInsn:@var{from};@var{to}
32662 This requests @value{GDBN} to copy instruction at address @var{from}
32663 to address @var{to}, possibly adjusted so that executing the
32664 instruction at @var{to} has the same effect as executing it at
32665 @var{from}. @value{GDBN} writes the adjusted instruction to target
32666 memory starting at @var{to}.
32671 @item qRelocInsn:@var{adjusted_size}
32672 Informs the stub the relocation is complete. @var{adjusted_size} is
32673 the length in bytes of resulting relocated instruction sequence.
32675 A badly formed request was detected, or an error was encountered while
32676 relocating the instruction.
32679 @node Host I/O Packets
32680 @section Host I/O Packets
32681 @cindex Host I/O, remote protocol
32682 @cindex file transfer, remote protocol
32684 The @dfn{Host I/O} packets allow @value{GDBN} to perform I/O
32685 operations on the far side of a remote link. For example, Host I/O is
32686 used to upload and download files to a remote target with its own
32687 filesystem. Host I/O uses the same constant values and data structure
32688 layout as the target-initiated File-I/O protocol. However, the
32689 Host I/O packets are structured differently. The target-initiated
32690 protocol relies on target memory to store parameters and buffers.
32691 Host I/O requests are initiated by @value{GDBN}, and the
32692 target's memory is not involved. @xref{File-I/O Remote Protocol
32693 Extension}, for more details on the target-initiated protocol.
32695 The Host I/O request packets all encode a single operation along with
32696 its arguments. They have this format:
32700 @item vFile:@var{operation}: @var{parameter}@dots{}
32701 @var{operation} is the name of the particular request; the target
32702 should compare the entire packet name up to the second colon when checking
32703 for a supported operation. The format of @var{parameter} depends on
32704 the operation. Numbers are always passed in hexadecimal. Negative
32705 numbers have an explicit minus sign (i.e.@: two's complement is not
32706 used). Strings (e.g.@: filenames) are encoded as a series of
32707 hexadecimal bytes. The last argument to a system call may be a
32708 buffer of escaped binary data (@pxref{Binary Data}).
32712 The valid responses to Host I/O packets are:
32716 @item F @var{result} [, @var{errno}] [; @var{attachment}]
32717 @var{result} is the integer value returned by this operation, usually
32718 non-negative for success and -1 for errors. If an error has occured,
32719 @var{errno} will be included in the result. @var{errno} will have a
32720 value defined by the File-I/O protocol (@pxref{Errno Values}). For
32721 operations which return data, @var{attachment} supplies the data as a
32722 binary buffer. Binary buffers in response packets are escaped in the
32723 normal way (@pxref{Binary Data}). See the individual packet
32724 documentation for the interpretation of @var{result} and
32728 An empty response indicates that this operation is not recognized.
32732 These are the supported Host I/O operations:
32735 @item vFile:open: @var{pathname}, @var{flags}, @var{mode}
32736 Open a file at @var{pathname} and return a file descriptor for it, or
32737 return -1 if an error occurs. @var{pathname} is a string,
32738 @var{flags} is an integer indicating a mask of open flags
32739 (@pxref{Open Flags}), and @var{mode} is an integer indicating a mask
32740 of mode bits to use if the file is created (@pxref{mode_t Values}).
32741 @xref{open}, for details of the open flags and mode values.
32743 @item vFile:close: @var{fd}
32744 Close the open file corresponding to @var{fd} and return 0, or
32745 -1 if an error occurs.
32747 @item vFile:pread: @var{fd}, @var{count}, @var{offset}
32748 Read data from the open file corresponding to @var{fd}. Up to
32749 @var{count} bytes will be read from the file, starting at @var{offset}
32750 relative to the start of the file. The target may read fewer bytes;
32751 common reasons include packet size limits and an end-of-file
32752 condition. The number of bytes read is returned. Zero should only be
32753 returned for a successful read at the end of the file, or if
32754 @var{count} was zero.
32756 The data read should be returned as a binary attachment on success.
32757 If zero bytes were read, the response should include an empty binary
32758 attachment (i.e.@: a trailing semicolon). The return value is the
32759 number of target bytes read; the binary attachment may be longer if
32760 some characters were escaped.
32762 @item vFile:pwrite: @var{fd}, @var{offset}, @var{data}
32763 Write @var{data} (a binary buffer) to the open file corresponding
32764 to @var{fd}. Start the write at @var{offset} from the start of the
32765 file. Unlike many @code{write} system calls, there is no
32766 separate @var{count} argument; the length of @var{data} in the
32767 packet is used. @samp{vFile:write} returns the number of bytes written,
32768 which may be shorter than the length of @var{data}, or -1 if an
32771 @item vFile:unlink: @var{pathname}
32772 Delete the file at @var{pathname} on the target. Return 0,
32773 or -1 if an error occurs. @var{pathname} is a string.
32778 @section Interrupts
32779 @cindex interrupts (remote protocol)
32781 When a program on the remote target is running, @value{GDBN} may
32782 attempt to interrupt it by sending a @samp{Ctrl-C}, @code{BREAK} or
32783 a @code{BREAK} followed by @code{g},
32784 control of which is specified via @value{GDBN}'s @samp{interrupt-sequence}.
32786 The precise meaning of @code{BREAK} is defined by the transport
32787 mechanism and may, in fact, be undefined. @value{GDBN} does not
32788 currently define a @code{BREAK} mechanism for any of the network
32789 interfaces except for TCP, in which case @value{GDBN} sends the
32790 @code{telnet} BREAK sequence.
32792 @samp{Ctrl-C}, on the other hand, is defined and implemented for all
32793 transport mechanisms. It is represented by sending the single byte
32794 @code{0x03} without any of the usual packet overhead described in
32795 the Overview section (@pxref{Overview}). When a @code{0x03} byte is
32796 transmitted as part of a packet, it is considered to be packet data
32797 and does @emph{not} represent an interrupt. E.g., an @samp{X} packet
32798 (@pxref{X packet}), used for binary downloads, may include an unescaped
32799 @code{0x03} as part of its packet.
32801 @code{BREAK} followed by @code{g} is also known as Magic SysRq g.
32802 When Linux kernel receives this sequence from serial port,
32803 it stops execution and connects to gdb.
32805 Stubs are not required to recognize these interrupt mechanisms and the
32806 precise meaning associated with receipt of the interrupt is
32807 implementation defined. If the target supports debugging of multiple
32808 threads and/or processes, it should attempt to interrupt all
32809 currently-executing threads and processes.
32810 If the stub is successful at interrupting the
32811 running program, it should send one of the stop
32812 reply packets (@pxref{Stop Reply Packets}) to @value{GDBN} as a result
32813 of successfully stopping the program in all-stop mode, and a stop reply
32814 for each stopped thread in non-stop mode.
32815 Interrupts received while the
32816 program is stopped are discarded.
32818 @node Notification Packets
32819 @section Notification Packets
32820 @cindex notification packets
32821 @cindex packets, notification
32823 The @value{GDBN} remote serial protocol includes @dfn{notifications},
32824 packets that require no acknowledgment. Both the GDB and the stub
32825 may send notifications (although the only notifications defined at
32826 present are sent by the stub). Notifications carry information
32827 without incurring the round-trip latency of an acknowledgment, and so
32828 are useful for low-impact communications where occasional packet loss
32831 A notification packet has the form @samp{% @var{data} #
32832 @var{checksum}}, where @var{data} is the content of the notification,
32833 and @var{checksum} is a checksum of @var{data}, computed and formatted
32834 as for ordinary @value{GDBN} packets. A notification's @var{data}
32835 never contains @samp{$}, @samp{%} or @samp{#} characters. Upon
32836 receiving a notification, the recipient sends no @samp{+} or @samp{-}
32837 to acknowledge the notification's receipt or to report its corruption.
32839 Every notification's @var{data} begins with a name, which contains no
32840 colon characters, followed by a colon character.
32842 Recipients should silently ignore corrupted notifications and
32843 notifications they do not understand. Recipients should restart
32844 timeout periods on receipt of a well-formed notification, whether or
32845 not they understand it.
32847 Senders should only send the notifications described here when this
32848 protocol description specifies that they are permitted. In the
32849 future, we may extend the protocol to permit existing notifications in
32850 new contexts; this rule helps older senders avoid confusing newer
32853 (Older versions of @value{GDBN} ignore bytes received until they see
32854 the @samp{$} byte that begins an ordinary packet, so new stubs may
32855 transmit notifications without fear of confusing older clients. There
32856 are no notifications defined for @value{GDBN} to send at the moment, but we
32857 assume that most older stubs would ignore them, as well.)
32859 The following notification packets from the stub to @value{GDBN} are
32863 @item Stop: @var{reply}
32864 Report an asynchronous stop event in non-stop mode.
32865 The @var{reply} has the form of a stop reply, as
32866 described in @ref{Stop Reply Packets}. Refer to @ref{Remote Non-Stop},
32867 for information on how these notifications are acknowledged by
32871 @node Remote Non-Stop
32872 @section Remote Protocol Support for Non-Stop Mode
32874 @value{GDBN}'s remote protocol supports non-stop debugging of
32875 multi-threaded programs, as described in @ref{Non-Stop Mode}. If the stub
32876 supports non-stop mode, it should report that to @value{GDBN} by including
32877 @samp{QNonStop+} in its @samp{qSupported} response (@pxref{qSupported}).
32879 @value{GDBN} typically sends a @samp{QNonStop} packet only when
32880 establishing a new connection with the stub. Entering non-stop mode
32881 does not alter the state of any currently-running threads, but targets
32882 must stop all threads in any already-attached processes when entering
32883 all-stop mode. @value{GDBN} uses the @samp{?} packet as necessary to
32884 probe the target state after a mode change.
32886 In non-stop mode, when an attached process encounters an event that
32887 would otherwise be reported with a stop reply, it uses the
32888 asynchronous notification mechanism (@pxref{Notification Packets}) to
32889 inform @value{GDBN}. In contrast to all-stop mode, where all threads
32890 in all processes are stopped when a stop reply is sent, in non-stop
32891 mode only the thread reporting the stop event is stopped. That is,
32892 when reporting a @samp{S} or @samp{T} response to indicate completion
32893 of a step operation, hitting a breakpoint, or a fault, only the
32894 affected thread is stopped; any other still-running threads continue
32895 to run. When reporting a @samp{W} or @samp{X} response, all running
32896 threads belonging to other attached processes continue to run.
32898 Only one stop reply notification at a time may be pending; if
32899 additional stop events occur before @value{GDBN} has acknowledged the
32900 previous notification, they must be queued by the stub for later
32901 synchronous transmission in response to @samp{vStopped} packets from
32902 @value{GDBN}. Because the notification mechanism is unreliable,
32903 the stub is permitted to resend a stop reply notification
32904 if it believes @value{GDBN} may not have received it. @value{GDBN}
32905 ignores additional stop reply notifications received before it has
32906 finished processing a previous notification and the stub has completed
32907 sending any queued stop events.
32909 Otherwise, @value{GDBN} must be prepared to receive a stop reply
32910 notification at any time. Specifically, they may appear when
32911 @value{GDBN} is not otherwise reading input from the stub, or when
32912 @value{GDBN} is expecting to read a normal synchronous response or a
32913 @samp{+}/@samp{-} acknowledgment to a packet it has sent.
32914 Notification packets are distinct from any other communication from
32915 the stub so there is no ambiguity.
32917 After receiving a stop reply notification, @value{GDBN} shall
32918 acknowledge it by sending a @samp{vStopped} packet (@pxref{vStopped packet})
32919 as a regular, synchronous request to the stub. Such acknowledgment
32920 is not required to happen immediately, as @value{GDBN} is permitted to
32921 send other, unrelated packets to the stub first, which the stub should
32924 Upon receiving a @samp{vStopped} packet, if the stub has other queued
32925 stop events to report to @value{GDBN}, it shall respond by sending a
32926 normal stop reply response. @value{GDBN} shall then send another
32927 @samp{vStopped} packet to solicit further responses; again, it is
32928 permitted to send other, unrelated packets as well which the stub
32929 should process normally.
32931 If the stub receives a @samp{vStopped} packet and there are no
32932 additional stop events to report, the stub shall return an @samp{OK}
32933 response. At this point, if further stop events occur, the stub shall
32934 send a new stop reply notification, @value{GDBN} shall accept the
32935 notification, and the process shall be repeated.
32937 In non-stop mode, the target shall respond to the @samp{?} packet as
32938 follows. First, any incomplete stop reply notification/@samp{vStopped}
32939 sequence in progress is abandoned. The target must begin a new
32940 sequence reporting stop events for all stopped threads, whether or not
32941 it has previously reported those events to @value{GDBN}. The first
32942 stop reply is sent as a synchronous reply to the @samp{?} packet, and
32943 subsequent stop replies are sent as responses to @samp{vStopped} packets
32944 using the mechanism described above. The target must not send
32945 asynchronous stop reply notifications until the sequence is complete.
32946 If all threads are running when the target receives the @samp{?} packet,
32947 or if the target is not attached to any process, it shall respond
32950 @node Packet Acknowledgment
32951 @section Packet Acknowledgment
32953 @cindex acknowledgment, for @value{GDBN} remote
32954 @cindex packet acknowledgment, for @value{GDBN} remote
32955 By default, when either the host or the target machine receives a packet,
32956 the first response expected is an acknowledgment: either @samp{+} (to indicate
32957 the package was received correctly) or @samp{-} (to request retransmission).
32958 This mechanism allows the @value{GDBN} remote protocol to operate over
32959 unreliable transport mechanisms, such as a serial line.
32961 In cases where the transport mechanism is itself reliable (such as a pipe or
32962 TCP connection), the @samp{+}/@samp{-} acknowledgments are redundant.
32963 It may be desirable to disable them in that case to reduce communication
32964 overhead, or for other reasons. This can be accomplished by means of the
32965 @samp{QStartNoAckMode} packet; @pxref{QStartNoAckMode}.
32967 When in no-acknowledgment mode, neither the stub nor @value{GDBN} shall send or
32968 expect @samp{+}/@samp{-} protocol acknowledgments. The packet
32969 and response format still includes the normal checksum, as described in
32970 @ref{Overview}, but the checksum may be ignored by the receiver.
32972 If the stub supports @samp{QStartNoAckMode} and prefers to operate in
32973 no-acknowledgment mode, it should report that to @value{GDBN}
32974 by including @samp{QStartNoAckMode+} in its response to @samp{qSupported};
32975 @pxref{qSupported}.
32976 If @value{GDBN} also supports @samp{QStartNoAckMode} and it has not been
32977 disabled via the @code{set remote noack-packet off} command
32978 (@pxref{Remote Configuration}),
32979 @value{GDBN} may then send a @samp{QStartNoAckMode} packet to the stub.
32980 Only then may the stub actually turn off packet acknowledgments.
32981 @value{GDBN} sends a final @samp{+} acknowledgment of the stub's @samp{OK}
32982 response, which can be safely ignored by the stub.
32984 Note that @code{set remote noack-packet} command only affects negotiation
32985 between @value{GDBN} and the stub when subsequent connections are made;
32986 it does not affect the protocol acknowledgment state for any current
32988 Since @samp{+}/@samp{-} acknowledgments are enabled by default when a
32989 new connection is established,
32990 there is also no protocol request to re-enable the acknowledgments
32991 for the current connection, once disabled.
32996 Example sequence of a target being re-started. Notice how the restart
32997 does not get any direct output:
33002 @emph{target restarts}
33005 <- @code{T001:1234123412341234}
33009 Example sequence of a target being stepped by a single instruction:
33012 -> @code{G1445@dots{}}
33017 <- @code{T001:1234123412341234}
33021 <- @code{1455@dots{}}
33025 @node File-I/O Remote Protocol Extension
33026 @section File-I/O Remote Protocol Extension
33027 @cindex File-I/O remote protocol extension
33030 * File-I/O Overview::
33031 * Protocol Basics::
33032 * The F Request Packet::
33033 * The F Reply Packet::
33034 * The Ctrl-C Message::
33036 * List of Supported Calls::
33037 * Protocol-specific Representation of Datatypes::
33039 * File-I/O Examples::
33042 @node File-I/O Overview
33043 @subsection File-I/O Overview
33044 @cindex file-i/o overview
33046 The @dfn{File I/O remote protocol extension} (short: File-I/O) allows the
33047 target to use the host's file system and console I/O to perform various
33048 system calls. System calls on the target system are translated into a
33049 remote protocol packet to the host system, which then performs the needed
33050 actions and returns a response packet to the target system.
33051 This simulates file system operations even on targets that lack file systems.
33053 The protocol is defined to be independent of both the host and target systems.
33054 It uses its own internal representation of datatypes and values. Both
33055 @value{GDBN} and the target's @value{GDBN} stub are responsible for
33056 translating the system-dependent value representations into the internal
33057 protocol representations when data is transmitted.
33059 The communication is synchronous. A system call is possible only when
33060 @value{GDBN} is waiting for a response from the @samp{C}, @samp{c}, @samp{S}
33061 or @samp{s} packets. While @value{GDBN} handles the request for a system call,
33062 the target is stopped to allow deterministic access to the target's
33063 memory. Therefore File-I/O is not interruptible by target signals. On
33064 the other hand, it is possible to interrupt File-I/O by a user interrupt
33065 (@samp{Ctrl-C}) within @value{GDBN}.
33067 The target's request to perform a host system call does not finish
33068 the latest @samp{C}, @samp{c}, @samp{S} or @samp{s} action. That means,
33069 after finishing the system call, the target returns to continuing the
33070 previous activity (continue, step). No additional continue or step
33071 request from @value{GDBN} is required.
33074 (@value{GDBP}) continue
33075 <- target requests 'system call X'
33076 target is stopped, @value{GDBN} executes system call
33077 -> @value{GDBN} returns result
33078 ... target continues, @value{GDBN} returns to wait for the target
33079 <- target hits breakpoint and sends a Txx packet
33082 The protocol only supports I/O on the console and to regular files on
33083 the host file system. Character or block special devices, pipes,
33084 named pipes, sockets or any other communication method on the host
33085 system are not supported by this protocol.
33087 File I/O is not supported in non-stop mode.
33089 @node Protocol Basics
33090 @subsection Protocol Basics
33091 @cindex protocol basics, file-i/o
33093 The File-I/O protocol uses the @code{F} packet as the request as well
33094 as reply packet. Since a File-I/O system call can only occur when
33095 @value{GDBN} is waiting for a response from the continuing or stepping target,
33096 the File-I/O request is a reply that @value{GDBN} has to expect as a result
33097 of a previous @samp{C}, @samp{c}, @samp{S} or @samp{s} packet.
33098 This @code{F} packet contains all information needed to allow @value{GDBN}
33099 to call the appropriate host system call:
33103 A unique identifier for the requested system call.
33106 All parameters to the system call. Pointers are given as addresses
33107 in the target memory address space. Pointers to strings are given as
33108 pointer/length pair. Numerical values are given as they are.
33109 Numerical control flags are given in a protocol-specific representation.
33113 At this point, @value{GDBN} has to perform the following actions.
33117 If the parameters include pointer values to data needed as input to a
33118 system call, @value{GDBN} requests this data from the target with a
33119 standard @code{m} packet request. This additional communication has to be
33120 expected by the target implementation and is handled as any other @code{m}
33124 @value{GDBN} translates all value from protocol representation to host
33125 representation as needed. Datatypes are coerced into the host types.
33128 @value{GDBN} calls the system call.
33131 It then coerces datatypes back to protocol representation.
33134 If the system call is expected to return data in buffer space specified
33135 by pointer parameters to the call, the data is transmitted to the
33136 target using a @code{M} or @code{X} packet. This packet has to be expected
33137 by the target implementation and is handled as any other @code{M} or @code{X}
33142 Eventually @value{GDBN} replies with another @code{F} packet which contains all
33143 necessary information for the target to continue. This at least contains
33150 @code{errno}, if has been changed by the system call.
33157 After having done the needed type and value coercion, the target continues
33158 the latest continue or step action.
33160 @node The F Request Packet
33161 @subsection The @code{F} Request Packet
33162 @cindex file-i/o request packet
33163 @cindex @code{F} request packet
33165 The @code{F} request packet has the following format:
33168 @item F@var{call-id},@var{parameter@dots{}}
33170 @var{call-id} is the identifier to indicate the host system call to be called.
33171 This is just the name of the function.
33173 @var{parameter@dots{}} are the parameters to the system call.
33174 Parameters are hexadecimal integer values, either the actual values in case
33175 of scalar datatypes, pointers to target buffer space in case of compound
33176 datatypes and unspecified memory areas, or pointer/length pairs in case
33177 of string parameters. These are appended to the @var{call-id} as a
33178 comma-delimited list. All values are transmitted in ASCII
33179 string representation, pointer/length pairs separated by a slash.
33185 @node The F Reply Packet
33186 @subsection The @code{F} Reply Packet
33187 @cindex file-i/o reply packet
33188 @cindex @code{F} reply packet
33190 The @code{F} reply packet has the following format:
33194 @item F@var{retcode},@var{errno},@var{Ctrl-C flag};@var{call-specific attachment}
33196 @var{retcode} is the return code of the system call as hexadecimal value.
33198 @var{errno} is the @code{errno} set by the call, in protocol-specific
33200 This parameter can be omitted if the call was successful.
33202 @var{Ctrl-C flag} is only sent if the user requested a break. In this
33203 case, @var{errno} must be sent as well, even if the call was successful.
33204 The @var{Ctrl-C flag} itself consists of the character @samp{C}:
33211 or, if the call was interrupted before the host call has been performed:
33218 assuming 4 is the protocol-specific representation of @code{EINTR}.
33223 @node The Ctrl-C Message
33224 @subsection The @samp{Ctrl-C} Message
33225 @cindex ctrl-c message, in file-i/o protocol
33227 If the @samp{Ctrl-C} flag is set in the @value{GDBN}
33228 reply packet (@pxref{The F Reply Packet}),
33229 the target should behave as if it had
33230 gotten a break message. The meaning for the target is ``system call
33231 interrupted by @code{SIGINT}''. Consequentially, the target should actually stop
33232 (as with a break message) and return to @value{GDBN} with a @code{T02}
33235 It's important for the target to know in which
33236 state the system call was interrupted. There are two possible cases:
33240 The system call hasn't been performed on the host yet.
33243 The system call on the host has been finished.
33247 These two states can be distinguished by the target by the value of the
33248 returned @code{errno}. If it's the protocol representation of @code{EINTR}, the system
33249 call hasn't been performed. This is equivalent to the @code{EINTR} handling
33250 on POSIX systems. In any other case, the target may presume that the
33251 system call has been finished --- successfully or not --- and should behave
33252 as if the break message arrived right after the system call.
33254 @value{GDBN} must behave reliably. If the system call has not been called
33255 yet, @value{GDBN} may send the @code{F} reply immediately, setting @code{EINTR} as
33256 @code{errno} in the packet. If the system call on the host has been finished
33257 before the user requests a break, the full action must be finished by
33258 @value{GDBN}. This requires sending @code{M} or @code{X} packets as necessary.
33259 The @code{F} packet may only be sent when either nothing has happened
33260 or the full action has been completed.
33263 @subsection Console I/O
33264 @cindex console i/o as part of file-i/o
33266 By default and if not explicitly closed by the target system, the file
33267 descriptors 0, 1 and 2 are connected to the @value{GDBN} console. Output
33268 on the @value{GDBN} console is handled as any other file output operation
33269 (@code{write(1, @dots{})} or @code{write(2, @dots{})}). Console input is handled
33270 by @value{GDBN} so that after the target read request from file descriptor
33271 0 all following typing is buffered until either one of the following
33276 The user types @kbd{Ctrl-c}. The behaviour is as explained above, and the
33278 system call is treated as finished.
33281 The user presses @key{RET}. This is treated as end of input with a trailing
33285 The user types @kbd{Ctrl-d}. This is treated as end of input. No trailing
33286 character (neither newline nor @samp{Ctrl-D}) is appended to the input.
33290 If the user has typed more characters than fit in the buffer given to
33291 the @code{read} call, the trailing characters are buffered in @value{GDBN} until
33292 either another @code{read(0, @dots{})} is requested by the target, or debugging
33293 is stopped at the user's request.
33296 @node List of Supported Calls
33297 @subsection List of Supported Calls
33298 @cindex list of supported file-i/o calls
33315 @unnumberedsubsubsec open
33316 @cindex open, file-i/o system call
33321 int open(const char *pathname, int flags);
33322 int open(const char *pathname, int flags, mode_t mode);
33326 @samp{Fopen,@var{pathptr}/@var{len},@var{flags},@var{mode}}
33329 @var{flags} is the bitwise @code{OR} of the following values:
33333 If the file does not exist it will be created. The host
33334 rules apply as far as file ownership and time stamps
33338 When used with @code{O_CREAT}, if the file already exists it is
33339 an error and open() fails.
33342 If the file already exists and the open mode allows
33343 writing (@code{O_RDWR} or @code{O_WRONLY} is given) it will be
33344 truncated to zero length.
33347 The file is opened in append mode.
33350 The file is opened for reading only.
33353 The file is opened for writing only.
33356 The file is opened for reading and writing.
33360 Other bits are silently ignored.
33364 @var{mode} is the bitwise @code{OR} of the following values:
33368 User has read permission.
33371 User has write permission.
33374 Group has read permission.
33377 Group has write permission.
33380 Others have read permission.
33383 Others have write permission.
33387 Other bits are silently ignored.
33390 @item Return value:
33391 @code{open} returns the new file descriptor or -1 if an error
33398 @var{pathname} already exists and @code{O_CREAT} and @code{O_EXCL} were used.
33401 @var{pathname} refers to a directory.
33404 The requested access is not allowed.
33407 @var{pathname} was too long.
33410 A directory component in @var{pathname} does not exist.
33413 @var{pathname} refers to a device, pipe, named pipe or socket.
33416 @var{pathname} refers to a file on a read-only filesystem and
33417 write access was requested.
33420 @var{pathname} is an invalid pointer value.
33423 No space on device to create the file.
33426 The process already has the maximum number of files open.
33429 The limit on the total number of files open on the system
33433 The call was interrupted by the user.
33439 @unnumberedsubsubsec close
33440 @cindex close, file-i/o system call
33449 @samp{Fclose,@var{fd}}
33451 @item Return value:
33452 @code{close} returns zero on success, or -1 if an error occurred.
33458 @var{fd} isn't a valid open file descriptor.
33461 The call was interrupted by the user.
33467 @unnumberedsubsubsec read
33468 @cindex read, file-i/o system call
33473 int read(int fd, void *buf, unsigned int count);
33477 @samp{Fread,@var{fd},@var{bufptr},@var{count}}
33479 @item Return value:
33480 On success, the number of bytes read is returned.
33481 Zero indicates end of file. If count is zero, read
33482 returns zero as well. On error, -1 is returned.
33488 @var{fd} is not a valid file descriptor or is not open for
33492 @var{bufptr} is an invalid pointer value.
33495 The call was interrupted by the user.
33501 @unnumberedsubsubsec write
33502 @cindex write, file-i/o system call
33507 int write(int fd, const void *buf, unsigned int count);
33511 @samp{Fwrite,@var{fd},@var{bufptr},@var{count}}
33513 @item Return value:
33514 On success, the number of bytes written are returned.
33515 Zero indicates nothing was written. On error, -1
33522 @var{fd} is not a valid file descriptor or is not open for
33526 @var{bufptr} is an invalid pointer value.
33529 An attempt was made to write a file that exceeds the
33530 host-specific maximum file size allowed.
33533 No space on device to write the data.
33536 The call was interrupted by the user.
33542 @unnumberedsubsubsec lseek
33543 @cindex lseek, file-i/o system call
33548 long lseek (int fd, long offset, int flag);
33552 @samp{Flseek,@var{fd},@var{offset},@var{flag}}
33554 @var{flag} is one of:
33558 The offset is set to @var{offset} bytes.
33561 The offset is set to its current location plus @var{offset}
33565 The offset is set to the size of the file plus @var{offset}
33569 @item Return value:
33570 On success, the resulting unsigned offset in bytes from
33571 the beginning of the file is returned. Otherwise, a
33572 value of -1 is returned.
33578 @var{fd} is not a valid open file descriptor.
33581 @var{fd} is associated with the @value{GDBN} console.
33584 @var{flag} is not a proper value.
33587 The call was interrupted by the user.
33593 @unnumberedsubsubsec rename
33594 @cindex rename, file-i/o system call
33599 int rename(const char *oldpath, const char *newpath);
33603 @samp{Frename,@var{oldpathptr}/@var{len},@var{newpathptr}/@var{len}}
33605 @item Return value:
33606 On success, zero is returned. On error, -1 is returned.
33612 @var{newpath} is an existing directory, but @var{oldpath} is not a
33616 @var{newpath} is a non-empty directory.
33619 @var{oldpath} or @var{newpath} is a directory that is in use by some
33623 An attempt was made to make a directory a subdirectory
33627 A component used as a directory in @var{oldpath} or new
33628 path is not a directory. Or @var{oldpath} is a directory
33629 and @var{newpath} exists but is not a directory.
33632 @var{oldpathptr} or @var{newpathptr} are invalid pointer values.
33635 No access to the file or the path of the file.
33639 @var{oldpath} or @var{newpath} was too long.
33642 A directory component in @var{oldpath} or @var{newpath} does not exist.
33645 The file is on a read-only filesystem.
33648 The device containing the file has no room for the new
33652 The call was interrupted by the user.
33658 @unnumberedsubsubsec unlink
33659 @cindex unlink, file-i/o system call
33664 int unlink(const char *pathname);
33668 @samp{Funlink,@var{pathnameptr}/@var{len}}
33670 @item Return value:
33671 On success, zero is returned. On error, -1 is returned.
33677 No access to the file or the path of the file.
33680 The system does not allow unlinking of directories.
33683 The file @var{pathname} cannot be unlinked because it's
33684 being used by another process.
33687 @var{pathnameptr} is an invalid pointer value.
33690 @var{pathname} was too long.
33693 A directory component in @var{pathname} does not exist.
33696 A component of the path is not a directory.
33699 The file is on a read-only filesystem.
33702 The call was interrupted by the user.
33708 @unnumberedsubsubsec stat/fstat
33709 @cindex fstat, file-i/o system call
33710 @cindex stat, file-i/o system call
33715 int stat(const char *pathname, struct stat *buf);
33716 int fstat(int fd, struct stat *buf);
33720 @samp{Fstat,@var{pathnameptr}/@var{len},@var{bufptr}}@*
33721 @samp{Ffstat,@var{fd},@var{bufptr}}
33723 @item Return value:
33724 On success, zero is returned. On error, -1 is returned.
33730 @var{fd} is not a valid open file.
33733 A directory component in @var{pathname} does not exist or the
33734 path is an empty string.
33737 A component of the path is not a directory.
33740 @var{pathnameptr} is an invalid pointer value.
33743 No access to the file or the path of the file.
33746 @var{pathname} was too long.
33749 The call was interrupted by the user.
33755 @unnumberedsubsubsec gettimeofday
33756 @cindex gettimeofday, file-i/o system call
33761 int gettimeofday(struct timeval *tv, void *tz);
33765 @samp{Fgettimeofday,@var{tvptr},@var{tzptr}}
33767 @item Return value:
33768 On success, 0 is returned, -1 otherwise.
33774 @var{tz} is a non-NULL pointer.
33777 @var{tvptr} and/or @var{tzptr} is an invalid pointer value.
33783 @unnumberedsubsubsec isatty
33784 @cindex isatty, file-i/o system call
33789 int isatty(int fd);
33793 @samp{Fisatty,@var{fd}}
33795 @item Return value:
33796 Returns 1 if @var{fd} refers to the @value{GDBN} console, 0 otherwise.
33802 The call was interrupted by the user.
33807 Note that the @code{isatty} call is treated as a special case: it returns
33808 1 to the target if the file descriptor is attached
33809 to the @value{GDBN} console, 0 otherwise. Implementing through system calls
33810 would require implementing @code{ioctl} and would be more complex than
33815 @unnumberedsubsubsec system
33816 @cindex system, file-i/o system call
33821 int system(const char *command);
33825 @samp{Fsystem,@var{commandptr}/@var{len}}
33827 @item Return value:
33828 If @var{len} is zero, the return value indicates whether a shell is
33829 available. A zero return value indicates a shell is not available.
33830 For non-zero @var{len}, the value returned is -1 on error and the
33831 return status of the command otherwise. Only the exit status of the
33832 command is returned, which is extracted from the host's @code{system}
33833 return value by calling @code{WEXITSTATUS(retval)}. In case
33834 @file{/bin/sh} could not be executed, 127 is returned.
33840 The call was interrupted by the user.
33845 @value{GDBN} takes over the full task of calling the necessary host calls
33846 to perform the @code{system} call. The return value of @code{system} on
33847 the host is simplified before it's returned
33848 to the target. Any termination signal information from the child process
33849 is discarded, and the return value consists
33850 entirely of the exit status of the called command.
33852 Due to security concerns, the @code{system} call is by default refused
33853 by @value{GDBN}. The user has to allow this call explicitly with the
33854 @code{set remote system-call-allowed 1} command.
33857 @item set remote system-call-allowed
33858 @kindex set remote system-call-allowed
33859 Control whether to allow the @code{system} calls in the File I/O
33860 protocol for the remote target. The default is zero (disabled).
33862 @item show remote system-call-allowed
33863 @kindex show remote system-call-allowed
33864 Show whether the @code{system} calls are allowed in the File I/O
33868 @node Protocol-specific Representation of Datatypes
33869 @subsection Protocol-specific Representation of Datatypes
33870 @cindex protocol-specific representation of datatypes, in file-i/o protocol
33873 * Integral Datatypes::
33875 * Memory Transfer::
33880 @node Integral Datatypes
33881 @unnumberedsubsubsec Integral Datatypes
33882 @cindex integral datatypes, in file-i/o protocol
33884 The integral datatypes used in the system calls are @code{int},
33885 @code{unsigned int}, @code{long}, @code{unsigned long},
33886 @code{mode_t}, and @code{time_t}.
33888 @code{int}, @code{unsigned int}, @code{mode_t} and @code{time_t} are
33889 implemented as 32 bit values in this protocol.
33891 @code{long} and @code{unsigned long} are implemented as 64 bit types.
33893 @xref{Limits}, for corresponding MIN and MAX values (similar to those
33894 in @file{limits.h}) to allow range checking on host and target.
33896 @code{time_t} datatypes are defined as seconds since the Epoch.
33898 All integral datatypes transferred as part of a memory read or write of a
33899 structured datatype e.g.@: a @code{struct stat} have to be given in big endian
33902 @node Pointer Values
33903 @unnumberedsubsubsec Pointer Values
33904 @cindex pointer values, in file-i/o protocol
33906 Pointers to target data are transmitted as they are. An exception
33907 is made for pointers to buffers for which the length isn't
33908 transmitted as part of the function call, namely strings. Strings
33909 are transmitted as a pointer/length pair, both as hex values, e.g.@:
33916 which is a pointer to data of length 18 bytes at position 0x1aaf.
33917 The length is defined as the full string length in bytes, including
33918 the trailing null byte. For example, the string @code{"hello world"}
33919 at address 0x123456 is transmitted as
33925 @node Memory Transfer
33926 @unnumberedsubsubsec Memory Transfer
33927 @cindex memory transfer, in file-i/o protocol
33929 Structured data which is transferred using a memory read or write (for
33930 example, a @code{struct stat}) is expected to be in a protocol-specific format
33931 with all scalar multibyte datatypes being big endian. Translation to
33932 this representation needs to be done both by the target before the @code{F}
33933 packet is sent, and by @value{GDBN} before
33934 it transfers memory to the target. Transferred pointers to structured
33935 data should point to the already-coerced data at any time.
33939 @unnumberedsubsubsec struct stat
33940 @cindex struct stat, in file-i/o protocol
33942 The buffer of type @code{struct stat} used by the target and @value{GDBN}
33943 is defined as follows:
33947 unsigned int st_dev; /* device */
33948 unsigned int st_ino; /* inode */
33949 mode_t st_mode; /* protection */
33950 unsigned int st_nlink; /* number of hard links */
33951 unsigned int st_uid; /* user ID of owner */
33952 unsigned int st_gid; /* group ID of owner */
33953 unsigned int st_rdev; /* device type (if inode device) */
33954 unsigned long st_size; /* total size, in bytes */
33955 unsigned long st_blksize; /* blocksize for filesystem I/O */
33956 unsigned long st_blocks; /* number of blocks allocated */
33957 time_t st_atime; /* time of last access */
33958 time_t st_mtime; /* time of last modification */
33959 time_t st_ctime; /* time of last change */
33963 The integral datatypes conform to the definitions given in the
33964 appropriate section (see @ref{Integral Datatypes}, for details) so this
33965 structure is of size 64 bytes.
33967 The values of several fields have a restricted meaning and/or
33973 A value of 0 represents a file, 1 the console.
33976 No valid meaning for the target. Transmitted unchanged.
33979 Valid mode bits are described in @ref{Constants}. Any other
33980 bits have currently no meaning for the target.
33985 No valid meaning for the target. Transmitted unchanged.
33990 These values have a host and file system dependent
33991 accuracy. Especially on Windows hosts, the file system may not
33992 support exact timing values.
33995 The target gets a @code{struct stat} of the above representation and is
33996 responsible for coercing it to the target representation before
33999 Note that due to size differences between the host, target, and protocol
34000 representations of @code{struct stat} members, these members could eventually
34001 get truncated on the target.
34003 @node struct timeval
34004 @unnumberedsubsubsec struct timeval
34005 @cindex struct timeval, in file-i/o protocol
34007 The buffer of type @code{struct timeval} used by the File-I/O protocol
34008 is defined as follows:
34012 time_t tv_sec; /* second */
34013 long tv_usec; /* microsecond */
34017 The integral datatypes conform to the definitions given in the
34018 appropriate section (see @ref{Integral Datatypes}, for details) so this
34019 structure is of size 8 bytes.
34022 @subsection Constants
34023 @cindex constants, in file-i/o protocol
34025 The following values are used for the constants inside of the
34026 protocol. @value{GDBN} and target are responsible for translating these
34027 values before and after the call as needed.
34038 @unnumberedsubsubsec Open Flags
34039 @cindex open flags, in file-i/o protocol
34041 All values are given in hexadecimal representation.
34053 @node mode_t Values
34054 @unnumberedsubsubsec mode_t Values
34055 @cindex mode_t values, in file-i/o protocol
34057 All values are given in octal representation.
34074 @unnumberedsubsubsec Errno Values
34075 @cindex errno values, in file-i/o protocol
34077 All values are given in decimal representation.
34102 @code{EUNKNOWN} is used as a fallback error value if a host system returns
34103 any error value not in the list of supported error numbers.
34106 @unnumberedsubsubsec Lseek Flags
34107 @cindex lseek flags, in file-i/o protocol
34116 @unnumberedsubsubsec Limits
34117 @cindex limits, in file-i/o protocol
34119 All values are given in decimal representation.
34122 INT_MIN -2147483648
34124 UINT_MAX 4294967295
34125 LONG_MIN -9223372036854775808
34126 LONG_MAX 9223372036854775807
34127 ULONG_MAX 18446744073709551615
34130 @node File-I/O Examples
34131 @subsection File-I/O Examples
34132 @cindex file-i/o examples
34134 Example sequence of a write call, file descriptor 3, buffer is at target
34135 address 0x1234, 6 bytes should be written:
34138 <- @code{Fwrite,3,1234,6}
34139 @emph{request memory read from target}
34142 @emph{return "6 bytes written"}
34146 Example sequence of a read call, file descriptor 3, buffer is at target
34147 address 0x1234, 6 bytes should be read:
34150 <- @code{Fread,3,1234,6}
34151 @emph{request memory write to target}
34152 -> @code{X1234,6:XXXXXX}
34153 @emph{return "6 bytes read"}
34157 Example sequence of a read call, call fails on the host due to invalid
34158 file descriptor (@code{EBADF}):
34161 <- @code{Fread,3,1234,6}
34165 Example sequence of a read call, user presses @kbd{Ctrl-c} before syscall on
34169 <- @code{Fread,3,1234,6}
34174 Example sequence of a read call, user presses @kbd{Ctrl-c} after syscall on
34178 <- @code{Fread,3,1234,6}
34179 -> @code{X1234,6:XXXXXX}
34183 @node Library List Format
34184 @section Library List Format
34185 @cindex library list format, remote protocol
34187 On some platforms, a dynamic loader (e.g.@: @file{ld.so}) runs in the
34188 same process as your application to manage libraries. In this case,
34189 @value{GDBN} can use the loader's symbol table and normal memory
34190 operations to maintain a list of shared libraries. On other
34191 platforms, the operating system manages loaded libraries.
34192 @value{GDBN} can not retrieve the list of currently loaded libraries
34193 through memory operations, so it uses the @samp{qXfer:libraries:read}
34194 packet (@pxref{qXfer library list read}) instead. The remote stub
34195 queries the target's operating system and reports which libraries
34198 The @samp{qXfer:libraries:read} packet returns an XML document which
34199 lists loaded libraries and their offsets. Each library has an
34200 associated name and one or more segment or section base addresses,
34201 which report where the library was loaded in memory.
34203 For the common case of libraries that are fully linked binaries, the
34204 library should have a list of segments. If the target supports
34205 dynamic linking of a relocatable object file, its library XML element
34206 should instead include a list of allocated sections. The segment or
34207 section bases are start addresses, not relocation offsets; they do not
34208 depend on the library's link-time base addresses.
34210 @value{GDBN} must be linked with the Expat library to support XML
34211 library lists. @xref{Expat}.
34213 A simple memory map, with one loaded library relocated by a single
34214 offset, looks like this:
34218 <library name="/lib/libc.so.6">
34219 <segment address="0x10000000"/>
34224 Another simple memory map, with one loaded library with three
34225 allocated sections (.text, .data, .bss), looks like this:
34229 <library name="sharedlib.o">
34230 <section address="0x10000000"/>
34231 <section address="0x20000000"/>
34232 <section address="0x30000000"/>
34237 The format of a library list is described by this DTD:
34240 <!-- library-list: Root element with versioning -->
34241 <!ELEMENT library-list (library)*>
34242 <!ATTLIST library-list version CDATA #FIXED "1.0">
34243 <!ELEMENT library (segment*, section*)>
34244 <!ATTLIST library name CDATA #REQUIRED>
34245 <!ELEMENT segment EMPTY>
34246 <!ATTLIST segment address CDATA #REQUIRED>
34247 <!ELEMENT section EMPTY>
34248 <!ATTLIST section address CDATA #REQUIRED>
34251 In addition, segments and section descriptors cannot be mixed within a
34252 single library element, and you must supply at least one segment or
34253 section for each library.
34255 @node Memory Map Format
34256 @section Memory Map Format
34257 @cindex memory map format
34259 To be able to write into flash memory, @value{GDBN} needs to obtain a
34260 memory map from the target. This section describes the format of the
34263 The memory map is obtained using the @samp{qXfer:memory-map:read}
34264 (@pxref{qXfer memory map read}) packet and is an XML document that
34265 lists memory regions.
34267 @value{GDBN} must be linked with the Expat library to support XML
34268 memory maps. @xref{Expat}.
34270 The top-level structure of the document is shown below:
34273 <?xml version="1.0"?>
34274 <!DOCTYPE memory-map
34275 PUBLIC "+//IDN gnu.org//DTD GDB Memory Map V1.0//EN"
34276 "http://sourceware.org/gdb/gdb-memory-map.dtd">
34282 Each region can be either:
34287 A region of RAM starting at @var{addr} and extending for @var{length}
34291 <memory type="ram" start="@var{addr}" length="@var{length}"/>
34296 A region of read-only memory:
34299 <memory type="rom" start="@var{addr}" length="@var{length}"/>
34304 A region of flash memory, with erasure blocks @var{blocksize}
34308 <memory type="flash" start="@var{addr}" length="@var{length}">
34309 <property name="blocksize">@var{blocksize}</property>
34315 Regions must not overlap. @value{GDBN} assumes that areas of memory not covered
34316 by the memory map are RAM, and uses the ordinary @samp{M} and @samp{X}
34317 packets to write to addresses in such ranges.
34319 The formal DTD for memory map format is given below:
34322 <!-- ................................................... -->
34323 <!-- Memory Map XML DTD ................................ -->
34324 <!-- File: memory-map.dtd .............................. -->
34325 <!-- .................................... .............. -->
34326 <!-- memory-map.dtd -->
34327 <!-- memory-map: Root element with versioning -->
34328 <!ELEMENT memory-map (memory | property)>
34329 <!ATTLIST memory-map version CDATA #FIXED "1.0.0">
34330 <!ELEMENT memory (property)>
34331 <!-- memory: Specifies a memory region,
34332 and its type, or device. -->
34333 <!ATTLIST memory type CDATA #REQUIRED
34334 start CDATA #REQUIRED
34335 length CDATA #REQUIRED
34336 device CDATA #IMPLIED>
34337 <!-- property: Generic attribute tag -->
34338 <!ELEMENT property (#PCDATA | property)*>
34339 <!ATTLIST property name CDATA #REQUIRED>
34342 @node Thread List Format
34343 @section Thread List Format
34344 @cindex thread list format
34346 To efficiently update the list of threads and their attributes,
34347 @value{GDBN} issues the @samp{qXfer:threads:read} packet
34348 (@pxref{qXfer threads read}) and obtains the XML document with
34349 the following structure:
34352 <?xml version="1.0"?>
34354 <thread id="id" core="0">
34355 ... description ...
34360 Each @samp{thread} element must have the @samp{id} attribute that
34361 identifies the thread (@pxref{thread-id syntax}). The
34362 @samp{core} attribute, if present, specifies which processor core
34363 the thread was last executing on. The content of the of @samp{thread}
34364 element is interpreted as human-readable auxilliary information.
34366 @include agentexpr.texi
34368 @node Trace File Format
34369 @appendix Trace File Format
34370 @cindex trace file format
34372 The trace file comes in three parts: a header, a textual description
34373 section, and a trace frame section with binary data.
34375 The header has the form @code{\x7fTRACE0\n}. The first byte is
34376 @code{0x7f} so as to indicate that the file contains binary data,
34377 while the @code{0} is a version number that may have different values
34380 The description section consists of multiple lines of @sc{ascii} text
34381 separated by newline characters (@code{0xa}). The lines may include a
34382 variety of optional descriptive or context-setting information, such
34383 as tracepoint definitions or register set size. @value{GDBN} will
34384 ignore any line that it does not recognize. An empty line marks the end
34387 @c FIXME add some specific types of data
34389 The trace frame section consists of a number of consecutive frames.
34390 Each frame begins with a two-byte tracepoint number, followed by a
34391 four-byte size giving the amount of data in the frame. The data in
34392 the frame consists of a number of blocks, each introduced by a
34393 character indicating its type (at least register, memory, and trace
34394 state variable). The data in this section is raw binary, not a
34395 hexadecimal or other encoding; its endianness matches the target's
34398 @c FIXME bi-arch may require endianness/arch info in description section
34401 @item R @var{bytes}
34402 Register block. The number and ordering of bytes matches that of a
34403 @code{g} packet in the remote protocol. Note that these are the
34404 actual bytes, in target order and @value{GDBN} register order, not a
34405 hexadecimal encoding.
34407 @item M @var{address} @var{length} @var{bytes}...
34408 Memory block. This is a contiguous block of memory, at the 8-byte
34409 address @var{address}, with a 2-byte length @var{length}, followed by
34410 @var{length} bytes.
34412 @item V @var{number} @var{value}
34413 Trace state variable block. This records the 8-byte signed value
34414 @var{value} of trace state variable numbered @var{number}.
34418 Future enhancements of the trace file format may include additional types
34421 @node Target Descriptions
34422 @appendix Target Descriptions
34423 @cindex target descriptions
34425 @strong{Warning:} target descriptions are still under active development,
34426 and the contents and format may change between @value{GDBN} releases.
34427 The format is expected to stabilize in the future.
34429 One of the challenges of using @value{GDBN} to debug embedded systems
34430 is that there are so many minor variants of each processor
34431 architecture in use. It is common practice for vendors to start with
34432 a standard processor core --- ARM, PowerPC, or MIPS, for example ---
34433 and then make changes to adapt it to a particular market niche. Some
34434 architectures have hundreds of variants, available from dozens of
34435 vendors. This leads to a number of problems:
34439 With so many different customized processors, it is difficult for
34440 the @value{GDBN} maintainers to keep up with the changes.
34442 Since individual variants may have short lifetimes or limited
34443 audiences, it may not be worthwhile to carry information about every
34444 variant in the @value{GDBN} source tree.
34446 When @value{GDBN} does support the architecture of the embedded system
34447 at hand, the task of finding the correct architecture name to give the
34448 @command{set architecture} command can be error-prone.
34451 To address these problems, the @value{GDBN} remote protocol allows a
34452 target system to not only identify itself to @value{GDBN}, but to
34453 actually describe its own features. This lets @value{GDBN} support
34454 processor variants it has never seen before --- to the extent that the
34455 descriptions are accurate, and that @value{GDBN} understands them.
34457 @value{GDBN} must be linked with the Expat library to support XML
34458 target descriptions. @xref{Expat}.
34461 * Retrieving Descriptions:: How descriptions are fetched from a target.
34462 * Target Description Format:: The contents of a target description.
34463 * Predefined Target Types:: Standard types available for target
34465 * Standard Target Features:: Features @value{GDBN} knows about.
34468 @node Retrieving Descriptions
34469 @section Retrieving Descriptions
34471 Target descriptions can be read from the target automatically, or
34472 specified by the user manually. The default behavior is to read the
34473 description from the target. @value{GDBN} retrieves it via the remote
34474 protocol using @samp{qXfer} requests (@pxref{General Query Packets,
34475 qXfer}). The @var{annex} in the @samp{qXfer} packet will be
34476 @samp{target.xml}. The contents of the @samp{target.xml} annex are an
34477 XML document, of the form described in @ref{Target Description
34480 Alternatively, you can specify a file to read for the target description.
34481 If a file is set, the target will not be queried. The commands to
34482 specify a file are:
34485 @cindex set tdesc filename
34486 @item set tdesc filename @var{path}
34487 Read the target description from @var{path}.
34489 @cindex unset tdesc filename
34490 @item unset tdesc filename
34491 Do not read the XML target description from a file. @value{GDBN}
34492 will use the description supplied by the current target.
34494 @cindex show tdesc filename
34495 @item show tdesc filename
34496 Show the filename to read for a target description, if any.
34500 @node Target Description Format
34501 @section Target Description Format
34502 @cindex target descriptions, XML format
34504 A target description annex is an @uref{http://www.w3.org/XML/, XML}
34505 document which complies with the Document Type Definition provided in
34506 the @value{GDBN} sources in @file{gdb/features/gdb-target.dtd}. This
34507 means you can use generally available tools like @command{xmllint} to
34508 check that your feature descriptions are well-formed and valid.
34509 However, to help people unfamiliar with XML write descriptions for
34510 their targets, we also describe the grammar here.
34512 Target descriptions can identify the architecture of the remote target
34513 and (for some architectures) provide information about custom register
34514 sets. They can also identify the OS ABI of the remote target.
34515 @value{GDBN} can use this information to autoconfigure for your
34516 target, or to warn you if you connect to an unsupported target.
34518 Here is a simple target description:
34521 <target version="1.0">
34522 <architecture>i386:x86-64</architecture>
34527 This minimal description only says that the target uses
34528 the x86-64 architecture.
34530 A target description has the following overall form, with [ ] marking
34531 optional elements and @dots{} marking repeatable elements. The elements
34532 are explained further below.
34535 <?xml version="1.0"?>
34536 <!DOCTYPE target SYSTEM "gdb-target.dtd">
34537 <target version="1.0">
34538 @r{[}@var{architecture}@r{]}
34539 @r{[}@var{osabi}@r{]}
34540 @r{[}@var{compatible}@r{]}
34541 @r{[}@var{feature}@dots{}@r{]}
34546 The description is generally insensitive to whitespace and line
34547 breaks, under the usual common-sense rules. The XML version
34548 declaration and document type declaration can generally be omitted
34549 (@value{GDBN} does not require them), but specifying them may be
34550 useful for XML validation tools. The @samp{version} attribute for
34551 @samp{<target>} may also be omitted, but we recommend
34552 including it; if future versions of @value{GDBN} use an incompatible
34553 revision of @file{gdb-target.dtd}, they will detect and report
34554 the version mismatch.
34556 @subsection Inclusion
34557 @cindex target descriptions, inclusion
34560 @cindex <xi:include>
34563 It can sometimes be valuable to split a target description up into
34564 several different annexes, either for organizational purposes, or to
34565 share files between different possible target descriptions. You can
34566 divide a description into multiple files by replacing any element of
34567 the target description with an inclusion directive of the form:
34570 <xi:include href="@var{document}"/>
34574 When @value{GDBN} encounters an element of this form, it will retrieve
34575 the named XML @var{document}, and replace the inclusion directive with
34576 the contents of that document. If the current description was read
34577 using @samp{qXfer}, then so will be the included document;
34578 @var{document} will be interpreted as the name of an annex. If the
34579 current description was read from a file, @value{GDBN} will look for
34580 @var{document} as a file in the same directory where it found the
34581 original description.
34583 @subsection Architecture
34584 @cindex <architecture>
34586 An @samp{<architecture>} element has this form:
34589 <architecture>@var{arch}</architecture>
34592 @var{arch} is one of the architectures from the set accepted by
34593 @code{set architecture} (@pxref{Targets, ,Specifying a Debugging Target}).
34596 @cindex @code{<osabi>}
34598 This optional field was introduced in @value{GDBN} version 7.0.
34599 Previous versions of @value{GDBN} ignore it.
34601 An @samp{<osabi>} element has this form:
34604 <osabi>@var{abi-name}</osabi>
34607 @var{abi-name} is an OS ABI name from the same selection accepted by
34608 @w{@code{set osabi}} (@pxref{ABI, ,Configuring the Current ABI}).
34610 @subsection Compatible Architecture
34611 @cindex @code{<compatible>}
34613 This optional field was introduced in @value{GDBN} version 7.0.
34614 Previous versions of @value{GDBN} ignore it.
34616 A @samp{<compatible>} element has this form:
34619 <compatible>@var{arch}</compatible>
34622 @var{arch} is one of the architectures from the set accepted by
34623 @code{set architecture} (@pxref{Targets, ,Specifying a Debugging Target}).
34625 A @samp{<compatible>} element is used to specify that the target
34626 is able to run binaries in some other than the main target architecture
34627 given by the @samp{<architecture>} element. For example, on the
34628 Cell Broadband Engine, the main architecture is @code{powerpc:common}
34629 or @code{powerpc:common64}, but the system is able to run binaries
34630 in the @code{spu} architecture as well. The way to describe this
34631 capability with @samp{<compatible>} is as follows:
34634 <architecture>powerpc:common</architecture>
34635 <compatible>spu</compatible>
34638 @subsection Features
34641 Each @samp{<feature>} describes some logical portion of the target
34642 system. Features are currently used to describe available CPU
34643 registers and the types of their contents. A @samp{<feature>} element
34647 <feature name="@var{name}">
34648 @r{[}@var{type}@dots{}@r{]}
34654 Each feature's name should be unique within the description. The name
34655 of a feature does not matter unless @value{GDBN} has some special
34656 knowledge of the contents of that feature; if it does, the feature
34657 should have its standard name. @xref{Standard Target Features}.
34661 Any register's value is a collection of bits which @value{GDBN} must
34662 interpret. The default interpretation is a two's complement integer,
34663 but other types can be requested by name in the register description.
34664 Some predefined types are provided by @value{GDBN} (@pxref{Predefined
34665 Target Types}), and the description can define additional composite types.
34667 Each type element must have an @samp{id} attribute, which gives
34668 a unique (within the containing @samp{<feature>}) name to the type.
34669 Types must be defined before they are used.
34672 Some targets offer vector registers, which can be treated as arrays
34673 of scalar elements. These types are written as @samp{<vector>} elements,
34674 specifying the array element type, @var{type}, and the number of elements,
34678 <vector id="@var{id}" type="@var{type}" count="@var{count}"/>
34682 If a register's value is usefully viewed in multiple ways, define it
34683 with a union type containing the useful representations. The
34684 @samp{<union>} element contains one or more @samp{<field>} elements,
34685 each of which has a @var{name} and a @var{type}:
34688 <union id="@var{id}">
34689 <field name="@var{name}" type="@var{type}"/>
34695 If a register's value is composed from several separate values, define
34696 it with a structure type. There are two forms of the @samp{<struct>}
34697 element; a @samp{<struct>} element must either contain only bitfields
34698 or contain no bitfields. If the structure contains only bitfields,
34699 its total size in bytes must be specified, each bitfield must have an
34700 explicit start and end, and bitfields are automatically assigned an
34701 integer type. The field's @var{start} should be less than or
34702 equal to its @var{end}, and zero represents the least significant bit.
34705 <struct id="@var{id}" size="@var{size}">
34706 <field name="@var{name}" start="@var{start}" end="@var{end}"/>
34711 If the structure contains no bitfields, then each field has an
34712 explicit type, and no implicit padding is added.
34715 <struct id="@var{id}">
34716 <field name="@var{name}" type="@var{type}"/>
34722 If a register's value is a series of single-bit flags, define it with
34723 a flags type. The @samp{<flags>} element has an explicit @var{size}
34724 and contains one or more @samp{<field>} elements. Each field has a
34725 @var{name}, a @var{start}, and an @var{end}. Only single-bit flags
34729 <flags id="@var{id}" size="@var{size}">
34730 <field name="@var{name}" start="@var{start}" end="@var{end}"/>
34735 @subsection Registers
34738 Each register is represented as an element with this form:
34741 <reg name="@var{name}"
34742 bitsize="@var{size}"
34743 @r{[}regnum="@var{num}"@r{]}
34744 @r{[}save-restore="@var{save-restore}"@r{]}
34745 @r{[}type="@var{type}"@r{]}
34746 @r{[}group="@var{group}"@r{]}/>
34750 The components are as follows:
34755 The register's name; it must be unique within the target description.
34758 The register's size, in bits.
34761 The register's number. If omitted, a register's number is one greater
34762 than that of the previous register (either in the current feature or in
34763 a preceeding feature); the first register in the target description
34764 defaults to zero. This register number is used to read or write
34765 the register; e.g.@: it is used in the remote @code{p} and @code{P}
34766 packets, and registers appear in the @code{g} and @code{G} packets
34767 in order of increasing register number.
34770 Whether the register should be preserved across inferior function
34771 calls; this must be either @code{yes} or @code{no}. The default is
34772 @code{yes}, which is appropriate for most registers except for
34773 some system control registers; this is not related to the target's
34777 The type of the register. @var{type} may be a predefined type, a type
34778 defined in the current feature, or one of the special types @code{int}
34779 and @code{float}. @code{int} is an integer type of the correct size
34780 for @var{bitsize}, and @code{float} is a floating point type (in the
34781 architecture's normal floating point format) of the correct size for
34782 @var{bitsize}. The default is @code{int}.
34785 The register group to which this register belongs. @var{group} must
34786 be either @code{general}, @code{float}, or @code{vector}. If no
34787 @var{group} is specified, @value{GDBN} will not display the register
34788 in @code{info registers}.
34792 @node Predefined Target Types
34793 @section Predefined Target Types
34794 @cindex target descriptions, predefined types
34796 Type definitions in the self-description can build up composite types
34797 from basic building blocks, but can not define fundamental types. Instead,
34798 standard identifiers are provided by @value{GDBN} for the fundamental
34799 types. The currently supported types are:
34808 Signed integer types holding the specified number of bits.
34815 Unsigned integer types holding the specified number of bits.
34819 Pointers to unspecified code and data. The program counter and
34820 any dedicated return address register may be marked as code
34821 pointers; printing a code pointer converts it into a symbolic
34822 address. The stack pointer and any dedicated address registers
34823 may be marked as data pointers.
34826 Single precision IEEE floating point.
34829 Double precision IEEE floating point.
34832 The 12-byte extended precision format used by ARM FPA registers.
34835 The 10-byte extended precision format used by x87 registers.
34838 32bit @sc{eflags} register used by x86.
34841 32bit @sc{mxcsr} register used by x86.
34845 @node Standard Target Features
34846 @section Standard Target Features
34847 @cindex target descriptions, standard features
34849 A target description must contain either no registers or all the
34850 target's registers. If the description contains no registers, then
34851 @value{GDBN} will assume a default register layout, selected based on
34852 the architecture. If the description contains any registers, the
34853 default layout will not be used; the standard registers must be
34854 described in the target description, in such a way that @value{GDBN}
34855 can recognize them.
34857 This is accomplished by giving specific names to feature elements
34858 which contain standard registers. @value{GDBN} will look for features
34859 with those names and verify that they contain the expected registers;
34860 if any known feature is missing required registers, or if any required
34861 feature is missing, @value{GDBN} will reject the target
34862 description. You can add additional registers to any of the
34863 standard features --- @value{GDBN} will display them just as if
34864 they were added to an unrecognized feature.
34866 This section lists the known features and their expected contents.
34867 Sample XML documents for these features are included in the
34868 @value{GDBN} source tree, in the directory @file{gdb/features}.
34870 Names recognized by @value{GDBN} should include the name of the
34871 company or organization which selected the name, and the overall
34872 architecture to which the feature applies; so e.g.@: the feature
34873 containing ARM core registers is named @samp{org.gnu.gdb.arm.core}.
34875 The names of registers are not case sensitive for the purpose
34876 of recognizing standard features, but @value{GDBN} will only display
34877 registers using the capitalization used in the description.
34884 * PowerPC Features::
34889 @subsection ARM Features
34890 @cindex target descriptions, ARM features
34892 The @samp{org.gnu.gdb.arm.core} feature is required for ARM targets.
34893 It should contain registers @samp{r0} through @samp{r13}, @samp{sp},
34894 @samp{lr}, @samp{pc}, and @samp{cpsr}.
34896 The @samp{org.gnu.gdb.arm.fpa} feature is optional. If present, it
34897 should contain registers @samp{f0} through @samp{f7} and @samp{fps}.
34899 The @samp{org.gnu.gdb.xscale.iwmmxt} feature is optional. If present,
34900 it should contain at least registers @samp{wR0} through @samp{wR15} and
34901 @samp{wCGR0} through @samp{wCGR3}. The @samp{wCID}, @samp{wCon},
34902 @samp{wCSSF}, and @samp{wCASF} registers are optional.
34904 The @samp{org.gnu.gdb.arm.vfp} feature is optional. If present, it
34905 should contain at least registers @samp{d0} through @samp{d15}. If
34906 they are present, @samp{d16} through @samp{d31} should also be included.
34907 @value{GDBN} will synthesize the single-precision registers from
34908 halves of the double-precision registers.
34910 The @samp{org.gnu.gdb.arm.neon} feature is optional. It does not
34911 need to contain registers; it instructs @value{GDBN} to display the
34912 VFP double-precision registers as vectors and to synthesize the
34913 quad-precision registers from pairs of double-precision registers.
34914 If this feature is present, @samp{org.gnu.gdb.arm.vfp} must also
34915 be present and include 32 double-precision registers.
34917 @node i386 Features
34918 @subsection i386 Features
34919 @cindex target descriptions, i386 features
34921 The @samp{org.gnu.gdb.i386.core} feature is required for i386/amd64
34922 targets. It should describe the following registers:
34926 @samp{eax} through @samp{edi} plus @samp{eip} for i386
34928 @samp{rax} through @samp{r15} plus @samp{rip} for amd64
34930 @samp{eflags}, @samp{cs}, @samp{ss}, @samp{ds}, @samp{es},
34931 @samp{fs}, @samp{gs}
34933 @samp{st0} through @samp{st7}
34935 @samp{fctrl}, @samp{fstat}, @samp{ftag}, @samp{fiseg}, @samp{fioff},
34936 @samp{foseg}, @samp{fooff} and @samp{fop}
34939 The register sets may be different, depending on the target.
34941 The @samp{org.gnu.gdb.i386.sse} feature is optional. It should
34942 describe registers:
34946 @samp{xmm0} through @samp{xmm7} for i386
34948 @samp{xmm0} through @samp{xmm15} for amd64
34953 The @samp{org.gnu.gdb.i386.avx} feature is optional and requires the
34954 @samp{org.gnu.gdb.i386.sse} feature. It should
34955 describe the upper 128 bits of @sc{ymm} registers:
34959 @samp{ymm0h} through @samp{ymm7h} for i386
34961 @samp{ymm0h} through @samp{ymm15h} for amd64
34965 The @samp{org.gnu.gdb.i386.linux} feature is optional. It should
34966 describe a single register, @samp{orig_eax}.
34968 @node MIPS Features
34969 @subsection MIPS Features
34970 @cindex target descriptions, MIPS features
34972 The @samp{org.gnu.gdb.mips.cpu} feature is required for MIPS targets.
34973 It should contain registers @samp{r0} through @samp{r31}, @samp{lo},
34974 @samp{hi}, and @samp{pc}. They may be 32-bit or 64-bit depending
34977 The @samp{org.gnu.gdb.mips.cp0} feature is also required. It should
34978 contain at least the @samp{status}, @samp{badvaddr}, and @samp{cause}
34979 registers. They may be 32-bit or 64-bit depending on the target.
34981 The @samp{org.gnu.gdb.mips.fpu} feature is currently required, though
34982 it may be optional in a future version of @value{GDBN}. It should
34983 contain registers @samp{f0} through @samp{f31}, @samp{fcsr}, and
34984 @samp{fir}. They may be 32-bit or 64-bit depending on the target.
34986 The @samp{org.gnu.gdb.mips.linux} feature is optional. It should
34987 contain a single register, @samp{restart}, which is used by the
34988 Linux kernel to control restartable syscalls.
34990 @node M68K Features
34991 @subsection M68K Features
34992 @cindex target descriptions, M68K features
34995 @item @samp{org.gnu.gdb.m68k.core}
34996 @itemx @samp{org.gnu.gdb.coldfire.core}
34997 @itemx @samp{org.gnu.gdb.fido.core}
34998 One of those features must be always present.
34999 The feature that is present determines which flavor of m68k is
35000 used. The feature that is present should contain registers
35001 @samp{d0} through @samp{d7}, @samp{a0} through @samp{a5}, @samp{fp},
35002 @samp{sp}, @samp{ps} and @samp{pc}.
35004 @item @samp{org.gnu.gdb.coldfire.fp}
35005 This feature is optional. If present, it should contain registers
35006 @samp{fp0} through @samp{fp7}, @samp{fpcontrol}, @samp{fpstatus} and
35010 @node PowerPC Features
35011 @subsection PowerPC Features
35012 @cindex target descriptions, PowerPC features
35014 The @samp{org.gnu.gdb.power.core} feature is required for PowerPC
35015 targets. It should contain registers @samp{r0} through @samp{r31},
35016 @samp{pc}, @samp{msr}, @samp{cr}, @samp{lr}, @samp{ctr}, and
35017 @samp{xer}. They may be 32-bit or 64-bit depending on the target.
35019 The @samp{org.gnu.gdb.power.fpu} feature is optional. It should
35020 contain registers @samp{f0} through @samp{f31} and @samp{fpscr}.
35022 The @samp{org.gnu.gdb.power.altivec} feature is optional. It should
35023 contain registers @samp{vr0} through @samp{vr31}, @samp{vscr},
35026 The @samp{org.gnu.gdb.power.vsx} feature is optional. It should
35027 contain registers @samp{vs0h} through @samp{vs31h}. @value{GDBN}
35028 will combine these registers with the floating point registers
35029 (@samp{f0} through @samp{f31}) and the altivec registers (@samp{vr0}
35030 through @samp{vr31}) to present the 128-bit wide registers @samp{vs0}
35031 through @samp{vs63}, the set of vector registers for POWER7.
35033 The @samp{org.gnu.gdb.power.spe} feature is optional. It should
35034 contain registers @samp{ev0h} through @samp{ev31h}, @samp{acc}, and
35035 @samp{spefscr}. SPE targets should provide 32-bit registers in
35036 @samp{org.gnu.gdb.power.core} and provide the upper halves in
35037 @samp{ev0h} through @samp{ev31h}. @value{GDBN} will combine
35038 these to present registers @samp{ev0} through @samp{ev31} to the
35041 @node Operating System Information
35042 @appendix Operating System Information
35043 @cindex operating system information
35049 Users of @value{GDBN} often wish to obtain information about the state of
35050 the operating system running on the target---for example the list of
35051 processes, or the list of open files. This section describes the
35052 mechanism that makes it possible. This mechanism is similar to the
35053 target features mechanism (@pxref{Target Descriptions}), but focuses
35054 on a different aspect of target.
35056 Operating system information is retrived from the target via the
35057 remote protocol, using @samp{qXfer} requests (@pxref{qXfer osdata
35058 read}). The object name in the request should be @samp{osdata}, and
35059 the @var{annex} identifies the data to be fetched.
35062 @appendixsection Process list
35063 @cindex operating system information, process list
35065 When requesting the process list, the @var{annex} field in the
35066 @samp{qXfer} request should be @samp{processes}. The returned data is
35067 an XML document. The formal syntax of this document is defined in
35068 @file{gdb/features/osdata.dtd}.
35070 An example document is:
35073 <?xml version="1.0"?>
35074 <!DOCTYPE target SYSTEM "osdata.dtd">
35075 <osdata type="processes">
35077 <column name="pid">1</column>
35078 <column name="user">root</column>
35079 <column name="command">/sbin/init</column>
35080 <column name="cores">1,2,3</column>
35085 Each item should include a column whose name is @samp{pid}. The value
35086 of that column should identify the process on the target. The
35087 @samp{user} and @samp{command} columns are optional, and will be
35088 displayed by @value{GDBN}. The @samp{cores} column, if present,
35089 should contain a comma-separated list of cores that this process
35090 is running on. Target may provide additional columns,
35091 which @value{GDBN} currently ignores.
35105 % I think something like @colophon should be in texinfo. In the
35107 \long\def\colophon{\hbox to0pt{}\vfill
35108 \centerline{The body of this manual is set in}
35109 \centerline{\fontname\tenrm,}
35110 \centerline{with headings in {\bf\fontname\tenbf}}
35111 \centerline{and examples in {\tt\fontname\tentt}.}
35112 \centerline{{\it\fontname\tenit\/},}
35113 \centerline{{\bf\fontname\tenbf}, and}
35114 \centerline{{\sl\fontname\tensl\/}}
35115 \centerline{are used for emphasis.}\vfill}
35117 % Blame: doc@cygnus.com, 1991.