HP Hewlett Packard gnu source level debugger 5992 4701 User Manual

Debugging with GDB Manual  
The GNU Source-Level Debugger  
HP Part Number: 5992-4701  
Published: February 2009  
Edition: 19  
Table of Contents  
Table of Contents  
3
4
Table of Contents  
Table of Contents  
5
6
Table of Contents  
Table of Contents  
7
8
Table of Contents  
Table of Contents  
9
10  
Table of Contents  
Table of Contents  
11  
12  
Table of Contents  
List of Tables  
13  
Summary of GDB  
The purpose of a debugger such as GDB is to allow you to see what is going on “inside”  
another program while it executes―or what another program was doing at the moment  
it crashed.  
GDB allows you to do the following:  
Load the executable along with any required arguments.  
Stop your program on specified blocks of code.  
Examine your program when it has stopped running due to an error.  
Change things in your program, so you can experiment with correcting the effects  
of one bug and go on to learn about another.  
You can use GDB to debug programs written in C, C++, and Fortran. For more  
information, refer to the “Supported languages” (page 105). For more information on  
supported languages, refer to the “C and C++” (page 106).  
GDB can be used to debug programs written in Fortran, although it may be necessary  
to refer to some variables with a trailing underscore. See “Fortran” (page 112).  
This version of the manual documents WDB, implemented on HP 9000 or HP Integrity  
systems running Release 11.x of the HP-UX operating system. WDB can be used to  
debug code generated by the HP ANSI C, HP ANSI aC++ and HP Fortran compilers  
as well as the GNUC and C++ compilers. It does not support the debugging of Pascal,  
Modula-2 or Chill programs.  
Free Software  
GDB is free software, protected by the GNUGeneral Public License (GPL). The GPL gives  
you the freedom to copy or adapt a licensed program―but every person getting a copy  
also gets with it the freedom to modify that copy (which means that they must get  
access to the source code), and the freedom to distribute further copies. Typical software  
companies use copyrights to limit your freedoms; the Free Software Foundation uses  
the GPL to preserve these freedoms.  
Fundamentally, the General Public License is a license which says that you have these  
freedoms and that you cannot take these freedoms away from anyone else.  
Contributors to GDB  
Richard Stallman was the original author of GDB, and of many other GNUprograms.  
Many others have contributed to its development. This section attempts to credit major  
contributors. One of the virtues of free software is that everyone is free to contribute  
to it; with regret, we cannot actually acknowledge everyone here. The file 'ChangeLog'  
in the GDB distribution approximates a blow-by-blow account.  
Changes much prior to version 2.0 are lost in the mists of time.  
Free Software  
15  
     
Plea: Additions to this section are particularly welcome. If you or your friends (or  
enemies, to be evenhanded) have been unfairly omitted from this list, we would like  
to add your names!  
So that they may not regard their many labors as thankless, we particularly thank those  
who shepherded GDB through major releases: Andrew Cagney (release 5.0); Jim Blandy  
(release 4.18); Jason Molenda (release 4.17); Stan Shebs (release 4.14); Fred Fish (releases  
4.16, 4.15, 4.13, 4.12, 4.11, 4.10, and 4.9); Stu Grossman and John Gilmore (releases 4.8,  
4.7, 4.6, 4.5, and 4.4); John Gilmore (releases 4.3, 4.2, 4.1, 4.0, and 3.9); Jim Kingdon  
(releases 3.5, 3.4, and 3.3); and Randy Smith (releases 3.2, 3.1, and 3.0).  
Richard Stallman, assisted at various times by Peter TerMaat, Chris Hanson, and Richard  
Mlynarik, handled releases through 2.8.  
Michael Tiemann is the author of most of the GNUC++ support in GDB, with significant  
additional contributions from Per Bothner. James Clark wrote the GNUC++ demangler.  
Early work on C++ was by Peter TerMaat (who also did much general update work  
leading to release 3.0).  
GDB 4 uses the BFD subroutine library to examine multiple object-file formats; BFD  
was a joint project of David V. Henkel-Wallace, Rich Pixley, Steve Chamberlain, and  
John Gilmore.  
David Johnson wrote the original COFF support; Pace Willison did the original support  
for encapsulated COFF.  
Brent Benson of Harris Computer Systems contributed DWARF 2 support.  
Adam de Boor and Bradley Davis contributed the ISI Optimum V support. Per Bothner,  
Noboyuki Hikichi, and Alessandro Forin contributed MIPS support. Jean-Daniel Fekete  
contributed Sun 386i support. Chris Hanson improved the HP 9000 support. Noboyuki  
Hikichi and Tomoyuki Hasei contributed Sony/News OS 3 support. David Johnson  
contributed Encore Umax support. Jyrki Kuoppala contributed Altos 3068 support. Jeff  
Law contributed HP PA and SOM support. Keith Packard contributed NS32K support.  
Doug Rabson contributed Acorn Risc Machine support. Bob Rusk contributed Harris  
Nighthawk CX-UX support. Chris Smith contributed Convex support (and Fortran  
debugging). Jonathan Stone contributed Pyramid support. Michael Tiemann contributed  
SPARC support. Tim Tucker contributed support for the Gould NP1 and Gould  
Powernode. Pace Willison contributed Intel 386 support. Jay Vosburgh contributed  
Symmetry support.  
Andreas Schwab contributed M68K Linux support.  
Rich Schaefer and Peter Schauer helped with support of SunOS shared libraries.  
Jay Fenlason and Roland McGrath ensured that GDB and GAS agree about several  
machine instruction sets.  
Patrick Duval, Ted Goldstein, Vikram Koka and Glenn Engel helped develop remote  
debugging. Intel Corporation, Wind River Systems, AMD, and ARM contributed remote  
debugging modules for the i960, VxWorks, A29K UDI, and RDI targets, respectively.  
16  
Brian Fox is the author of the readline libraries providing command-line editing and  
command history.  
Andrew Beers of SUNY Buffalo wrote the language-switching code, the Modula-2  
support, and contributed the Languages chapter of this manual.  
Fred Fish wrote most of the support for Unix System Vr4. He also enhanced the  
command-completion support to cover C++ overloaded symbols.  
Hitachi America, Ltd. sponsored the support for H8/300, H8/500, and Super-H  
processors.  
NEC sponsored the support for the v850, Vr4xxx, and Vr5xxx processors.  
Mitsubishi sponsored the support for D10V, D30V, and M32R/D processors.  
Toshiba sponsored the support for the TX39 Mips processor.  
Matsushita sponsored the support for the MN10200 and MN10300 processors.  
Fujitsu sponsored the support for SPARClite and FR30 processors.  
Kung Hsu, Je Law, and Rick Sladkey added support for hardware watchpoints.  
Michael Snyder added support for tracepoints.  
Stu Grossman wrote gdbserver.  
Jim Kingdon, Peter Schauer, Ian Taylor, and Stu Grossman made nearly innumerable  
bug fixes and cleanups throughout GDB.  
The following people at the Hewlett-Packard Company contributed support for the  
PA-RISC 2.0 architecture, HP-UX 10.20, 10.30, and 11.x (narrow mode), HP's  
implementation of kernel threads, HP's aC++ compiler, and the terminal user interface:  
Ben Krepp, Richard Title, John Bishop, Susan Macchia, Kathy Mann, Satish Pai, India  
Paul, Steve Rehrauer, and Elena Zannoni. Kim Haase, Rosario de la Torre, Alex McKale,  
Michael Coulter, Carl Burch, Bharath Chndramohan, Diwakar Nag, Muthuswami,  
Dennis Handly, Subash Babu and Dipshikha Basu provided HP-specific information  
in this manual.  
Cygnus Solutions has sponsored GDB maintenance and much of its development since  
1991. Cygnus engineers who have worked on GDB full time include Mark Alexander,  
Jim Blandy, Per Bothner, Kevin Buettner, Edith Epstein, Chris Faylor, Fred Fish, Martin  
Hunt, Jim Ingham, John Gilmore, Stu Grossman, Kung Hsu, Jim Kingdon, John Metzler,  
Fernando Nasser, Georey Noer, Dawn Perchik, Rich Pixley, Zdenek Radouch, Keith  
Seitz, Stan Shebs, David Taylor, and Elena Zannoni. In addition, Dave Brolley, Ian  
Carmichael, Steve Chamberlain, Nick Clifton, JT Conklin, Stan Cox, DJ Delorie, Ulrich  
Drepper, Frank Eigler, Doug Evans, Sean Fagan, David Henkel-Wallace, Richard  
Henderson, Jeff Holcomb, Jeff Law, Jim Lemke, Tom Lord, Bob Manson, Michael  
Meissner, Jason Merrill, Catherine Moore, Drew Moseley, Ken Raeburn, Gavin  
Romig-Koch, Rob Savoye, Jamie Smith, Mike Stump, Ian Taylor, Angela Thomas,  
Michael Tiemann, Tom Tromey, Ron Unrau, Jim Wilson, and David Zuhn have made  
contributions both large and small.  
Contributors to GDB  
17  
18  
1 A Sample GDB Session  
This chapter describes the most common GDB commands with the help of an example.  
The following topics are discussed:  
Loading the Executable  
Setting the Display Width  
Setting Breakpoints  
Running the Executable under GDB  
Stepping to the next line  
Stepping into a Subroutine  
Examining the Stack  
Printing Variable Values  
Listing the Source Code  
Setting Variable Values During a Debug Session  
In this sample session, we emphasize user input like this: input, to make it easier to  
pick out from the surrounding output.  
One of the preliminary versions of GNUm4(a generic macro processor) exhibits the  
following bug: sometimes, when we change its quote strings from the default, the  
commands used to capture one macro definition within another stop working. In the  
following short m4session, we define a macro foowhich expands to 0000; we then  
use the m4 built-indefnto define baras the same thing. However, when we  
change the open quote string to <QUOTE>and the close quote string to <UNQUOTE>,  
the same procedure fails to define a new synonym baz:  
$ cd gnu/m4 //change your current directory to the location where the m4 executable is stored.  
$ ./m4 //run the m4 application  
define(foo,0000)  
foo  
0000  
define (bar,defn('foo'))  
bar  
0000  
changequote(<QUOTE>,<UNQUOTE>)  
define(baz,defn(<QUOTE>foo<UNQUOTE>))  
baz  
C-d  
m4: End of input: 0: fatal error: EOF in string  
1.1 Loading the Executable  
Let us use GDB to try to see what is going on.  
1.1 Loading the Executable  
19  
   
$ (gdb) m4  
HP gdb 3.0 for PA-RISC 1.1 or 2.0 (narrow), HP-UX 11.00.  
Copyright 1986 - 2001 Free Software Foundation, Inc.  
Hewlett-Packard Wildebeest 3.0 (based on GDB ) is covered by the  
GNU General Public License. Type "show copying" to see the conditions to  
change it and/or distribute copies. Type "show warranty" for warranty/support.  
GDB reads only enough symbol data to know where to find the rest when needed; as  
a result, the first prompt comes up very quickly.  
1.2 Setting Display width  
We now tell GDB to use a narrower display width than usual, so that examples fit in  
this manual.  
((gdb)) set width 70  
We need to see how the m4 built-inchangequoteworks. Having looked at the  
source, we know the relevant subroutine is m4_changequote, so we set a breakpoint  
there with the GDB break command.  
1.3 Setting Breakpoints  
Here we describe how to set a breakpoint.  
((gdb)) break m4 changequote  
Breakpoint 1 at 0x62f4: file builtin.c, line 879.  
1.4 Running the executable under GDB  
Using the runcommand, we start m4under GDB control. As long as the control does  
not reach the m4_changequotesubroutine, the program runs as usual.  
((gdb)) run  
Starting program: /work/Editorial/gdb/gnu/m4/m4  
define(foo,0000)  
foo  
0000  
To trigger the breakpoint, we call changequote. GDB suspends execution of m4,  
displaying information about the context where it stops.  
changequote(<QUOTE>,<UNQUOTE>)  
Breakpoint 1, m4_changequote (argc=3, argv=0x33c70)  
at builtin.c:879  
879 if (bad_argc(TOKEN_DATA_TEXT(argv[0]),argc,1,3))  
1.5 Stepping to the next line in the source program  
Now we use the command n(next) to advance execution to the next line of the current  
function.  
20  
A Sample GDB Session  
       
((gdb)) n  
882 set_quotes((argc >= 2) ? TOKEN_DATA_TEXT(argv[1])\  
: nil,  
1.6 Stepping into a subroutine  
The set_quoteslooks like a promising subroutine. We can go into it by using the  
command s(step) instead of next. stepgoes to the next line to be executed in any  
subroutine, so it steps into set_quotes.  
((gdb)) s  
set_quotes (lq=0x34c78 "<QUOTE>", rq=0x34c88 "<UNQUOTE>")  
at input.c:530  
530 if (lquote != def_lquote)  
1.7 Examining the Stack  
The display that shows the subroutine where m4is now suspended (and its arguments)  
is called a stack frame display. It shows a summary of the stack. We can use the  
backtracecommand (which can also be spelled bt), to see where we are in the stack  
as a whole: the backtracecommand displays a stack frame for each active subroutine.  
((gdb)) bt  
#0 set_quotes (lq=0x34c78 "<QUOTE>", rq=0x34c88 "<UNQUOTE>")  
at input.c:530  
#1 0x6344 in m4_changequote (argc=3, argv=0x33c70)  
at builtin.c:882  
#2 0x8174 in expand_macro (sym=0x33320) at macro.c:242  
#3 0x7a88 in expand_token (obs=0x0, t=209696, td=0xf7fffa30)  
at macro.c:71  
#4 0x79dc in expand_input () at macro.c:40  
#5 0x2930 in main (argc=0, argv=0xf7fffb20) at m4.c:195  
We step through a few more lines to see what happens. The first two times, we can use  
's'; the next two times we use nto avoid falling into the xstrdupsubroutine.  
((gdb)) s  
0x3b5c 532 if (rquote != def_rquote)  
((gdb)) s  
0x3b80 535 lquote = (lq == nil || *lq == '\0') ? \  
def_lquote : xstrdup(lq);  
((gdb)) n  
536 rquote = (rq == nil || *rq == '\0') ? def_rquote\  
: xstrdup(rq);  
((gdb)) n  
538 len_lquote = strlen(rquote);  
1.8 Printing Variable Values  
The last line displayed looks a little odd in the listing above; we can examine the  
variables lquoteand rquoteto see if they are in fact the new left and right quotes  
we specified. We use the command p(print) to view their values.  
1.6 Stepping into a subroutine  
21  
     
((gdb)) p lquote  
$1 = 0x35d40 "<QUOTE>"  
((gdb)) p rquote  
$2 = 0x35d50 "<UNQUOTE>"  
1.9 Listing Source Code  
lquoteand rquoteare indeed the new left and right quotes. To look at some context,  
we can display ten lines of source surrounding the current line with the l(list)  
command.  
((gdb)) l  
533 xfree(rquote);  
534  
535 lquote = (lq == nil || *lq == '\0') ? def_lquote\  
: xstrdup (lq);  
536 rquote = (rq == nil || *rq == '\0') ? def_rquote\  
: xstrdup (rq);  
537  
538 len_lquote = strlen(rquote);  
539 len_rquote = strlen(lquote);  
540 }  
541  
542 void  
Let us step past the two lines that set len_lquoteand len_rquote, and then examine  
the values of those variables.  
((gdb)) n  
539 len_rquote = strlen(lquote);  
((gdb)) n  
540 }  
((gdb)) p len_lquote  
$3 = 9  
((gdb)) p len_rquote  
$4 = 7  
1.10 Setting Variable Values During a Session  
That certainly looks wrong, assuming len_lquoteand len_rquoteare meant to be  
the lengths of lquoteand rquoterespectively. We can set them to better values using  
the pcommand, since it can print the value of any expression―and that expression  
can include subroutine calls and assignments.  
((gdb)) p len_lquote=strlen(lquote)  
$5 = 7  
((gdb)) p len_rquote=strlen(rquote)  
$6 = 9  
Is that enough to fix the problem of using the new quotes with the m4 built-in  
defn? We can allow m4to continue executing with the c(continue) command, and  
then try the example that caused trouble initially:  
22  
A Sample GDB Session  
   
((gdb)) c  
Continuing.  
define(baz,defn(<QUOTE>foo<UNQUOTE>))  
baz  
0000  
Success! The new quotes now work just as well as the default ones. The problem seems  
to have been just the two typos defining the wrong lengths. We allow m4to exit by  
giving it an EOF as input:  
C-d  
Program exited normally.  
The message `Program exited normally.' is from GDB; it indicates m4has finished  
executing. We can end our GDB session with the GDB quitcommand.  
((gdb)) quit  
1.10 Setting Variable Values During a Session  
23  
24  
2 Getting In and Out of GDB  
This chapter discusses how to start GDB, and exit out of it. The essentials are:  
type '(gdb)' to start GDB.  
type quit or C-d to exit.  
2.1 Invoking GDB  
Invoke GDB by running the program (gdb). Once started, GDB reads commands from  
the terminal until you tell it to exit.  
You can also run (gdb) with a variety of arguments and options, to specify more of  
your debugging environment at the outset.  
The command-line options described here are designed to cover a variety of situations;  
in some environments, some of these options may effectively be unavailable.  
The most usual way to start GDB is with one argument, specifying an executable  
program:  
(gdb) program  
You can also start with both an executable program and a core file specified:  
(gdb) program core  
You can, instead, specify a process ID as a second argument, if you want to debug a  
running process:  
(gdb) program 1234  
would attach GDB to process 1234(unless you also have a file named '1234'; GDB  
does check for a core file first).  
Taking advantage of the second command-line argument requires a fairly complete  
operating system; when you use GDB as a remote debugger attached to a bare board,  
there may not be any notion of “process”, and there is often no way to get a core dump.  
GDB will warn you if it is unable to attach or to read core dumps.  
You can run (gdb) without printing the front material, which describes GDB's  
non-warranty, by specifying -silent:  
gdb -silent  
You can further control how GDB starts up by using command-line options. GDB itself  
can remind you of the options available.  
Type  
(gdb) -help  
to display all available options and briefly describe their use ('(gdb)-h' is a shorter  
equivalent).  
2.1 Invoking GDB  
25  
   
All options and command-line arguments you give are processed in sequential order.  
The order makes a difference when the `-x' option is used.  
2.1.1 Choosing files  
When GDB starts, it reads any arguments other than options as specifying an executable  
file and core file (or process ID). This is the same as if the arguments were specified by  
the '-se' and '-c' options respectively. (GDB reads the first argument that does not  
have an associated option flag as equivalent to the '-se' option followed by that  
argument; and the second argument that does not have an associated option flag, if  
any, as equivalent to the '-c' option followed by that argument.)  
If GDB has not been configured to included core file support, such as for most embedded  
targets, then it will complain about a second argument and ignore it.  
Many options have both long and short forms; both are shown in the following list.  
GDB also recognizes the long forms if you truncate them, so long as enough of the  
option is present to be unambiguous. (If you prefer, you can flag option arguments  
with `--' rather than `-', though we illustrate the more usual convention.)  
-symbols file  
-s file  
Read symbol table from file file.  
-exec file  
-e file  
Use file file as the executable file to execute when  
appropriate, and for examining pure data in  
conjunction with a core dump.  
-se file  
Read symbol table from file file and use it as the  
executable file.  
-core file  
-c file  
Use file file as a core dump to examine.  
-c number  
Connect to process ID number, as with the attach  
command (unless there is a file in core-dump format  
named number, in which case `-c' specifies that file as  
a core dump to read).  
-command file  
-x file  
Execute GDB commands from file file. See “Command  
-directory directory  
-d directory  
Add directory to the path to search for source files.  
26  
Getting In and Out of GDB  
 
-m, -mapped  
Warning: this option depends on operating system facilities  
that are not supported on all systems.  
If memory-mapped files are available on your system  
through the mmapsystem call, you can use this option  
to have GDB write the symbols from your program  
into a reusable file in the current directory. If the  
program you are debugging is called '/tmp/fred', the  
mapped symbol file is '/tmp/fred.syms'. Future GDB  
debugging sessions notice the presence of this file, and  
can quickly map in symbol information from it, rather  
than reading the symbol table from the executable  
program.  
The '.syms' file is specific to the host machine where  
GDB is run. It holds an exact image of the internal GDB  
symbol table. It cannot be shared across multiple host  
platforms.  
-r, -readnow  
Read each symbol file's entire symbol table  
immediately, rather than the default, which is to read  
it incrementally as it is needed. This makes startup  
slower, but makes future operations faster.  
You typically combine the -mappedand -readnowoptions in order to build a '.syms'  
file that contains complete symbol information. (See “Commands to specify files”  
(page 125), for information on '.syms' files.) A simple GDB invocation to do nothing  
but build a '.syms' file for future use is:  
gdb -batch -nx -mapped -readnow programname  
2.1.2 Choosing modes  
You can run GDB in various alternative modes―for example, in batch mode or quiet  
mode.  
-nx, -n  
Do not execute commands found in any initialization  
files (normally called '.gdbinit', or 'gdb.ini' on PCs).  
Normally, GDB executes the commands in these files  
after all the command options and arguments have been  
-quiet, -silent, -q  
“Quiet”. Do not print the introductory and copyright  
messages. These messages are also suppressed in batch  
mode.  
-batch  
Run in batch mode. Exit with status 0after processing  
all the command files specified with '-x' (and all  
commands from initialization files, if not inhibited with  
2.1 Invoking GDB  
27  
 
'-n'). Exit with nonzero status if an error occurs in  
executing the GDB commands in the command files.  
Batch mode may be useful for running GDB as a filter,  
for example to download and run a program on another  
computer; in order to make this more useful, the message  
Program exited normally.  
(which is ordinarily issued whenever a program running  
under GDB control terminates) is not issued when  
running in batch mode.  
-nowindows, -nw  
“No windows”. If GDB comes with a graphical user  
interface (GUI) built in, then this option tells GDB to  
only use the command-line interface. If no GUI is  
available, this option has no effect.  
-windows, -w  
If GDB includes a GUI, then this option requires it to be  
used if possible.  
Run GDB using directory as its working directory, instead  
of the current directory.  
-cd directory  
-dbx  
Support additional dbxcommands, including:  
use  
status(in dbxmode, status has a different  
meaning than in default GDB mode.)  
whereis  
func  
file  
assign  
call  
stop  
-fullname, -f  
GNUEmacs sets this option when it runs GDB as a  
subprocess. It tells GDB to output the full file name and  
line number in a standard, recognizable fashion each  
time a stack frame is displayed (which includes each  
time your program stops). This recognizable format looks  
like two `\032' characters, followed by the file name,  
line number, and character position separated by colons,  
and a newline. The Emacs-to-GDB interface program  
uses the two '\032' characters as a signal to display the  
source code for the frame.  
-epoch  
The Epoch Emacs-GDB interface sets this option when  
it runs GDB as a subprocess. It tells GDB to modify its  
28  
Getting In and Out of GDB  
print routines so as to allow Epoch to display values of  
expressions in a separate window.  
-annotate level  
This option sets the annotation level inside GDB. Its effect  
is identical to using `set annotate level' (see “GDB  
Annotations” (page 297)). Annotation level controls how  
much information does GDB print together with its  
prompt, values of expressions, source lines, and other  
types of output. Level 0 is the normal, level 1 is for use  
when GDB is run as a subprocess of GNUEmacs, level 2  
is the maximum annotation suitable for programs that  
control GDB.  
-async  
Use the asynchronous event loop for the command-line  
1
interface. GDB processes all events, such as user  
keyboard input, via a special event loop. This allows  
GDB to accept and process user commands in parallel  
1
with the debugged process being run , so you do not  
need to wait for control to return to GDB before you type  
the next command.  
NOTE: As of version 5.0, the target side of the  
asynchronous operation is not yet in place, so '-async'  
does not work fully yet.  
When the standard input is connected to a terminal  
device, GDB uses the asynchronous event loop by  
default, unless disabled by the '-noasync' option.  
-noasync  
Disable the asynchronous event loop for the  
command-line interface.  
-baud bps, -b bps  
Set the line speed (baud rate or bits per second) of any  
serial interface used by GDB for remote debugging.  
-tty device, -t device Run using device for your program's standard input and  
output.  
-tui  
Use a Terminal User Interface. For information, use your  
Web browser to read the file 'tui.html', which is  
usually installed in the directory /opt/langtools/  
wdb/docon HP-UX systems. Do not use this option if  
you run GDB from Emacs (see “Using GDB under gnu  
-xdb  
Run in XDB compatibility mode, allowing the use of  
certain XDB commands. For information, see the file  
1. GDB built with DJGPP tools for MS-DOS/MS-Windows supports this mode of operation, but the event  
loop is suspended when the debug target runs.  
2.1 Invoking GDB  
29  
'xdb_trans.html', which is usually installed in the  
directory /opt/langtools/wdb/docon HP-UX  
systems.  
-interpreter interp  
Use the interpreter interp for interface with the  
controlling program or device. This option is meant to  
be set by programs which communicate with GDB using  
it as a back end. For example, '--interpreter=mi'  
causes GDB to use the gdbmi interface (see “The GDB/MI  
-write  
Open the executable and core files for both reading and  
writing. This is equivalent to the 'set write on'  
command inside GDB (see “Patching programs”  
-statistics  
-version  
This option causes GDB to print statistics about time and  
memory usage after it completes each command and  
returns to the prompt.  
This option causes GDB to print its version number and  
no-warranty blurb, and exit.  
-pid  
This option causes GDB to attach to a running process.  
-inline  
This option causes the debugger to start with the inline  
debugging on.  
-src_no_g  
This option is used to set the limited source level  
debugging without compiling.  
2.1.3 Redirecting WDB input and output to a file  
To redirect WDB input and output to a file, use either of these commands to start the  
debugger:  
$ script log1  
$ gdb  
or  
$ gdb | tee log1  
2.2 Quitting GDB  
quit [expression], q To exit GDB, use the quitcommand (abbreviated q), or  
type an end-of-file character (usually C-d). If you do not  
supply expression, GDB will terminate normally; otherwise  
it will terminate using the result of expression as the error  
code.  
An interrupt (often C-c) does not exit from GDB, but rather terminates the action of  
any GDB command that is in progress and returns to GDB command level. It is safe to  
30  
Getting In and Out of GDB  
   
type the interrupt character at any time because GDB does not allow it to take effect  
until a time when it is safe.  
You can use the detachcommand to release an attached process or device.  
2.3 Shell commands  
If you need to execute occasional shell commands during your debugging session,  
there is no need to leave or suspend GDB; you can just use the shellcommand.  
shell command string  
Invoke a standard shell to execute command string. If it  
exists, the environment variable SHELLdetermines  
which shell to run. Otherwise GDB uses the default  
shell ('/bin/sh' on UNIX systems, 'COMMAND.COM' on  
MS-DOS, and so on.).  
The utility makeis often needed in development environments. You do not have to  
use the shellcommand for this purpose in GDB:  
make make-args  
Execute the makeprogram with the specified arguments. This  
is equivalent to 'shell make make-args'.  
2.3 Shell commands  
31  
 
32  
3 GDB Commands  
You can abbreviate a GDB command to the first few letters of the command name, if  
that abbreviation is unambiguous; and you can repeat certain GDB commands by  
typing just RET ). You can also use the TAB key to get GDB to fill out the rest of a  
word in a command (or to show you the alternatives available, if there is more than  
one possibility).  
3.1 Command syntax  
A GDB command is a single line of input. There is no limit on how long it can be.  
It starts with a command name, and is followed by arguments whose meaning  
depends on the command name.  
GDB command names can be truncated if that abbreviation is unambiguous. The  
possible command abbreviations are listed in the documentation for individual  
commands. In some cases, even ambiguous abbreviations are allowed; for example,  
sis specially defined as equivalent to stepeven though there are other commands  
whose names start with s. You can test abbreviations by using them as arguments  
to the help command.  
A blank line as input to GDB (typing just RET) means to repeat the previous  
command. Some commands (for example, run) do not repeat this way. These are  
commands whose unintentional repetition might cause trouble and which you are  
unlikely to want to repeat. The listand xcommands, when you repeat them  
with RET, construct new arguments rather than repeating exactly as typed. This  
permits easy scanning of source or memory.  
GDB can also use RET in another way: to partition lengthy output, in a way similar  
to the common utility more(see “Setting the GDB Screen Size” (page 283)). Since  
it is easy to press one RET too many in this situation, GDB disables command  
repetition after any command that generates this sort of display.  
Any text from a # to the end of the line is a comment; it does nothing. This is useful  
mainly in command files (see “Command files” (page 289)).  
3.2 Command completion  
GDB can fill in the rest of a word in a command for you, if there is only one possibility;  
it can also show you what the valid possibilities are for the next word in a command,  
at any time. This works for GDB commands, GDB subcommands, and the names of  
symbols in your program.  
Press the TAB key whenever you want GDB to fill out the rest of a word. If there is  
only one possibility, GDB fills in the word, and waits for you to finish the command  
(or press RET to enter it). For example, if you type  
((gdb)) info bre TAB  
3.1 Command syntax  
33  
     
GDB fills in the rest of the word 'breakpoints', since that is the only info  
subcommand beginning with 'bre':  
((gdb)) info breakpoints  
You can either press RET at this point, to run the info breakpointscommand, or  
backspace and enter something else, if 'breakpoints' does not look like the command  
you expected. (If you were sure you wanted info breakpointsin the first place,  
you might as well just type RET immediately after 'info bre', to exploit command  
abbreviations rather than command completion.)  
If there is more than one possibility for the next word when you press TAB , GDB  
sounds a bell. You can either supply more characters and try again, or just press TAB  
a second time; GDB displays all the possible completions for that word. For example,  
you might want to set a breakpoint on a subroutine whose name begins with 'make_',  
but when you type b make_TAB GDBjust sounds the bell. Typing TAB again displays  
all the function names in your program that begin with those characters, for example:  
((gdb)) b make_TAB  
GDB sounds bell; press TAB again, to see:  
make_a_section_from_file make_environ  
make_abs_section make_function_type  
make_blockvector make_pointer_type  
make_cleanup make_reference_type  
make_command make_symbol_completion_list  
((gdb)) b make_  
After displaying the available possibilities, GDB copies your partial input ('b make_'  
in the example) so you can finish the command.  
If you just want to see the list of alternatives in the first place, you can press M-? rather  
than pressing TAB twice. M-? means META?. You can type this either by holding down  
a key designated as the META shift on your keyboard (if there is one) while typing ?,  
or as ESC followed by ?.  
Sometimes the string you need, while logically a “word”, may contain parentheses or  
other characters that GDB normally excludes from its notion of a word. To permit word  
completion to work in this situation, you may enclose words in ' (single quote marks)  
in GDB commands.  
The most likely situation where you might need this is in typing the name of a C++  
function. This is because C++ allows function overloading (multiple definitions of the  
same function, distinguished by argument type). For example, when you want to set  
a breakpoint you may need to distinguish whether you mean the version of name that  
takes an int parameter, name(int), or the version that takes a float parameter, name(float).  
To use the word-completion facilities in this situation, type a single quote ' at the  
beginning of the function name. This alerts GDB that it may need to consider more  
information than usual when you press TAB or M-? to request word completion:  
((gdb)) b 'bubble( M-?  
34  
GDB Commands  
bubble(double,double) bubble(int,int)  
((gdb)) b 'bubble(  
In some cases, GDB can tell that completing a name requires using quotes. When this  
happens, GDB inserts the quote for you (while completing as much as it can) if you do  
not type the quote in the first place:  
((gdb)) b bub TAB  
GDB alters your input line to the following, and rings a bell:  
((gdb)) b 'bubble(  
In general, GDB can tell that a quote is needed (and inserts it) if you have not yet started  
typing the argument list when you ask for completion on an overloaded symbol.  
For more information about overloaded functions, see “C++ expressions” (page 109).  
You can use the command set overload-resolution offto disable overload  
3.3 Getting help  
You can always ask GDB itself for information on its commands, using the command  
help.  
help, h  
You can use help(abbreviated h) with no arguments to display  
a short list of named classes of commands:  
((gdb)) help  
List of classes of commands:  
aliases -- Aliases of other commands  
breakpoints -- Making program stop at certain points  
data -- Examining data  
files -- Specifying and examining files  
internals -- Maintenance commands  
obscure -- Obscure features  
running -- Running the program  
stack -- Examining the stack  
status -- Status inquiries  
support -- Support facilities  
tracepoints -- Tracing of program execution without  
stopping the program  
user-defined -- User-defined commands  
Type "help" followed by a class name for a list of  
commands in that class.  
Type "help" followed by command name for full  
documentation.  
Command name abbreviations are allowed if unambiguous.  
((gdb))  
3.3 Getting help  
35  
 
help class  
Using one of the general help classes as an argument, you can  
get a list of the individual commands in that class. For example,  
here is the help display for the class status:  
((gdb)) help status  
Status inquiries.  
List of commands:  
info -- Generic command for showing things  
about the program being debugged  
show -- Generic command for showing things  
about the debugger  
Type "help" followed by command name for full  
documentation.  
Command name abbreviations are allowed if unambiguous.  
((gdb))  
help command  
With a command name as helpargument, GDB displays a short  
paragraph on how to use that command.  
apropos args  
The apropos args command searches through all of the GDB  
commands, and their documentation, for the regular expression  
specified in args. It prints out all matches found. For example:  
apropos reload  
results in:  
set symbol-reloading -- Set dynamic symbol table  
reloading  
multiple times in one run  
show symbol-reloading -- Show dynamic symbol table  
reloading  
multiple times in one run  
complete args  
The complete args command lists all the possible completions  
for the beginning of a command. Use args to specify the  
beginning of the command you want completed. For example:  
complete i  
results in:  
if  
ignore  
info  
inspect  
This is intended for use by GNU Emacs.  
In addition to help, you can use the GDB commands infoand showto inquire about  
the state of your program, or the state of GDB itself. Each command supports many  
topics of inquiry; this manual introduces each of them in the appropriate context. The  
36  
GDB Commands  
listings under infoand under showin the Index point to all the sub-commands. See  
???.  
info This command (abbreviated i) is for describing the state of your program. For  
example, you can list the arguments given to your program with info args,  
list the registers currently in use with info registers, or list the breakpoints  
you have set with info breakpoints. You can get a complete list of the  
infosub-commands with help info.  
set  
You can assign the result of an expression to an environment variable with  
set. For example, you can set the GDB prompt to a $-signwith set prompt  
$.  
show In contrast to info, showis for describing the state of GDB itself. You can  
change most of the things you can show, by using the related command set;  
for example, you can control what number system is used for displays with  
set radix, or simply inquire which is currently in use with show radix.  
To display all the settable parameters and their current values, you can use  
showwith no arguments; you may also use info set. Both commands  
produce the same display.  
Here are three miscellaneous showsubcommands, all of which are exceptional in  
lacking corresponding setcommands:  
show version  
Show what version of GDB is running. You should include this  
information in GDB bug-reports. If multiple versions of GDB are  
in use at your site, you may need to determine which version of  
GDB you are running; as GDB evolves, new commands are  
introduced, and old ones may wither away. Also, many system  
vendors ship variant versions of GDB, and there are variant  
versions of GDB in GNU/Linux distributions as well. The version  
number is the same as the one announced when you start GDB.  
show copying  
show warranty  
Display information about permission for copying GDB.  
Display the GNU “NO WARRANTY” statement, or a warranty,  
if your version of GDB comes with one.  
3.3 Getting help  
37  
38  
4 Running Programs Under GDB  
When you run a program under GDB, you must first generate debugging information  
when you compile it using compiler option cc -g -O.  
You may start GDB with its arguments, if any, in an environment of your choice. If  
you are doing native debugging, you may redirect your program's input and output,  
debug an already running process, or kill a child process.  
4.1 Compiling for debugging  
Following points are noteable while compiling programs for debugging:  
Compile your program with the -g-Ooption to generate debugging information.  
The -g-Ooption is supported by HP ANSI C and HP aC++ compilers and GNU  
gcc compiler.  
Some compilers do not support the -g-Ooptions together.  
The -g-Ooptions do not work on machines with instruction scheduling.  
NOTE: Older versions of the GNU C compiler permitted a variant option '-gg' for  
debugging information. GDB no longer supports this format; if your GNU C compiler  
has this option, do not use it.  
4.2 Starting your program  
run, Use the runcommand to start your program under GDB. You must first specify  
r
the program name (except on VxWorks) with an argument to GDB (see Chapter 2  
(page 25)), or by using the fileor exec-file command (see “Commands  
NOTE: If you are running your program in an execution environment that supports  
processes, run creates an inferior process and makes that process run your program.  
(In environments without processes, run jumps to the start of your program.)  
The execution of a program is affected by the information it receives from the parent  
process. You must provide GDB the information before starting the program. (You can  
change the information after starting your program, but such changes only affect your  
program the next time you start it.) The information that must be passed to GDB can  
be categorized into four categories:  
arguments.  
Specify the arguments to give your program as  
the arguments of the runcommand. If a shell is  
available on your target, the shell is used to pass  
the arguments, so that you may use normal  
conventions (such as wildcard expansion or  
4.1 Compiling for debugging  
39  
     
variable substitution) in describing the  
arguments. On Unix systems, you can control  
which shell is used with the SHELLenvironment  
variable. GDB uses the C shell (/usr/bin/csh).  
environment.  
Your program inherits its environment from  
GDB. However, you can use the GDB commands  
set environmentand unset environment  
to change parts of the environment that affect  
working directory.  
Your program inherits its working directory from  
GDB. You can set the GDB working directory  
with the cdcommand in GDB. See“Working  
standard input and output. Your program as default uses the same device  
for standard input and standard output as GDB  
is using. You can redirect input and output in  
the runcommand line, or you can use the tty  
command to set a different device for your  
40  
Running Programs Under GDB  
WARNING! You can redirect input and output,  
but you cannot use pipes to pass the output of  
the program you are debugging to another  
program; if you attempt this, GDB is likely to  
wind up debugging the wrong program.  
NOTE:  
When you issue the runcommand, your program begins to execute immediately.  
See Chapter 5 (page 51), for discussion of how to arrange for your program to  
stop. Once your program has stopped, you may call functions in your program,  
using the printor callcommands. See Chapter 8 (page 83).  
If the modification time of your symbol file has changed since the last time GDB  
read its symbols, GDB discards its symbol table, and reads it again. When it does  
this, GDB tries to retain your current breakpoints.  
4.3 Arguments To Your Program  
The arguments to your program can be specified by the arguments of the runcommand.  
On HP-UX, they are passed to the C shell (/usr/bin/csh), which expands wildcard  
characters and performs redirection of I/O, and then to your program.  
On non-Unix systems, the program is usually invoked directly by GDB, which emulates  
I/O redirection via the appropriate system calls, and the wildcard characters are  
expanded by the startup code of the program, not by the shell.  
The runcommand used with no arguments uses the same arguments used by the  
previous run, or those set by the set argscommand.  
Following commands are used to pass the argument values to your program:  
set args  
Specify the arguments to be used the next time your program is run. If  
set argshas no arguments, runexecutes your program with no  
arguments. Once you have run your program with arguments, using  
set args before the next runis the only way to run it again without  
arguments.  
show args Show the arguments to give your program when it is started.  
4.4 Program Environment  
The environment consists of a set of environment variables and their values. Environment  
variables conventionally record information such as your user name, your home  
directory, your terminal type, and your search path for programs to run. Usually you  
set up environment variables with the shell and they are inherited by all the other  
4.3 Arguments To Your Program  
41  
   
programs you run. When debugging, it can be useful to try running your program  
with a modified environment without having to start GDB over again.  
show envvar  
List all the environment variables used by GDB.  
show paths  
Display the list of search paths for executables  
(the PATHenvironment variable).  
show environment [varname] Print the value of environment variable varname  
to be given to your program when it starts. If you  
do not supply varname, print the names and  
values of all environment variables to be given  
to your program. You can abbreviate  
environmentas env.  
set environment varname  
[=value]  
Set environment variable varname to value.  
The value changes for your program only, not  
for GDB itself. The value may be any string; the  
values of environment variables are just strings,  
and any interpretation is supplied by your  
program itself. The value parameter is optional;  
if it is eliminated, the variable is set to a null  
value.  
For example, this command:  
set env USER = foo  
tells the debugged program, when subsequently  
run, that its user is named 'foo'. (The spaces  
around '=' are used for clarity here; they are not  
actually required.)  
unset environment varname  
Remove variable varname from the environment  
to be passed to your program. This is different  
from 'set env varname ='; unset  
environmentremoves the variable from the  
environment, rather than assigning it an empty  
value.  
path directory  
Add directory to the front of the PATH  
environment variable (the search path for  
executables), for both GDB and your program.  
You may specify several directory names,  
separated by whitespace or by a  
system-dependent separator character (`:' on  
Unix, `;' on MS-DOS and MS-Windows). If  
directory is already in the path, it is moved  
to the front, so it is searched sooner.  
42  
Running Programs Under GDB  
You can use the string '$cwd' to refer to whatever is the current working directory at  
the time GDB searches the path. If you use '.' instead, it refers to the directory where  
you executed the pathcommand. GDB replaces '.' in the directory argument (with the  
current path) before adding directory to the search path.  
4.5 Working directory  
Each time you start your program with run, it inherits its working directory from the  
current working directory of GDB. The GDB working directory is initially whatever it  
inherited from its parent process (typically the shell), but you can specify a new working  
directory in GDB with the cdcommand.  
The GDB working directory also serves as a default for the commands that specify files  
for GDB to operate on. See “Commands to specify files” (page 125).  
Following commands are used to set the working directory for your program:  
cd directory  
Set the GDB working directory to directory.  
pwd  
Print the GDB working directory.  
4.6 Program Input and Output  
By default, the program you run under GDB does input and output to the same terminal  
that GDB uses. GDB switches the terminal to its own terminal modes to interact with  
you, but it records the terminal modes your program was using and switches back to  
them when you continue running your program.  
Following commands are used for redirecting the input and output:  
info terminal  
Displays information recorded by GDB about the terminal modes  
your program is using.  
tty  
Another way to specify where your program should do input  
and output is with the ttycommand. This command accepts a  
file name as argument, and causes this file to be the default for  
future runcommands. It also resets the controlling terminal for  
the child process, for future runcommands. For example,  
tty /dev/ttyb  
directs that processes started with subsequent runcommands  
default to do input and output on the terminal '/dev/ttyb' and  
have that as their controlling terminal.  
4.5 Working directory  
43  
   
NOTE:  
You can redirect your program input and output using shell redirection with the  
runcommand. For example,  
run > outfile  
starts your program, diverting its output to the file 'outfile'.  
An explicit redirection in runoverrides the ttycommand's effect on the  
input/output device, but not its effect on the controlling terminal.  
When you use the ttycommand or redirect input in the runcommand, only the  
input for your program is affected. The input for GDB still comes from your terminal.  
4.7 Debugging a Running Process  
You can use GDB to debug a running process by specifying the process ID. Following  
commands are used to debug a running process:  
attach process-id  
This command attaches to a running process―one that was  
started outside GDB. (info filesshows your active  
targets.) The command takes as argument a process ID. The  
usual way to find out the process-id of a Unix process is  
with the psutility, or with the 'jobs -l' shell command.  
attachdoes not repeat if you press RET a second time  
after executing the command.  
44  
Running Programs Under GDB  
 
NOTE:  
To use attach, your program must be running in an environment which supports  
processes; for example, attachdoes not work for programs on bare-board targets  
that lack an operating system.  
You must also have permission to send the process a signal.  
When you use attach, the debugger finds the program running in the process  
first by looking in the current working directory, then (if the program is not found)  
by using the source file search path (see “Specifying source directories” (page 79)).  
You can also use the filecommand to load the program. See “Commands to  
GDB stops the process being attached for debugging. You can examine and modify  
an attached process with the GDB commands that are available when you start  
processes with run. You can insert breakpoints; you can step and continue; you  
can modify storage. See “Breakpoints” (page 51). If you want the process to  
continue running, you can use the continuecommand after attaching GDB to  
the process.  
detach When you have finished debugging the attached process, you can use the  
detachcommand to release it from GDB control. The process continues  
its execution after being detached. After the detachcommand, that process  
and GDB become completely independent once more, and you are ready  
to attachanother process or start one with run. detachdoes not repeat  
if you press RET again after executing the command.  
If you exit GDB or use the runcommand while you have an attached process, you kill  
that process. By default, GDB asks for confirmation if you try to do either of these  
things; you can control whether or not you need to confirm by using the set confirm  
NOTE: When GDB attaches to a running program you may get a message saying  
"Attaching to process #nnnnn failed."  
The most likely cause for this message is that you have attached to a process that was  
started across an NFS mount. Versions of the HP-UX kernel before 11.x have a restriction  
that prevents a debugger from attaching to a process started from an NFS mount, unless  
the mount was made non-interruptible with the -nointrflag, see mount(1).  
4.8 Killing the child process  
Following command is used to kill the child process:  
kill Kill the child process in which your program is running under GDB.  
The killcommand is useful if you wish to debug a core dump instead of a running  
process. GDB ignores any core dump file while your program is running.  
4.8 Killing the child process  
45  
 
On some operating systems, a program cannot be executed outside GDB while you  
have breakpoints set on it inside GDB. You can use the killcommand in this situation  
to permit running your program outside the debugger.  
The killcommand is also useful if you wish to recompile and relink your program,  
since on many systems it is impossible to modify an executable file while it is running  
in a process. In this case, when you next type run, GDB notices that the file has changed,  
and reads the symbol table again (while trying to preserve your current breakpoint  
settings).  
4.9 Debugging programs with multiple threads  
In some operating systems, such as HP-UX and Solaris, a single program may have  
more than one thread of execution. The precise semantics of threads differ from one  
operating system to another, but in general the threads of a single program are akin to  
multiple processes―except that they share one address space (that is, they can all  
examine and modify the same variables). On the other hand, each thread has its own  
registers and execution stack, and private memory.  
GDB provides these facilities for debugging multi-thread programs:  
automatic notification of new threads  
thread-specific breakpoints  
WARNING! These facilities are not yet available on every GDB configuration where  
the operating system supports threads. If your GDB does not support threads, these  
commands have no effect. For example, a system without thread support shows no  
output from 'info threads', and always rejects the threadcommand, like this:  
((gdb)) info threads  
((gdb)) thread 1  
Thread ID 1 not known. Use the "info threads" command to  
see the IDs of currently known threads.  
Following commands are used to debug multi-threaded programs:  
'thread threadno', a command to switch among threads  
'info threads', a command to inquire about existing threads  
'thread apply [threadno] [all] args', a command to apply a command  
to a list of threads  
The GDB thread debugging facility allows you to observe all threads while your  
program runs―but whenever GDB takes control, one thread in particular is always  
the focus of debugging. This thread is called the current thread. Debugging commands  
show program information from the perspective of the current thread.  
Whenever GDB detects a new thread in your program, it displays the target system's  
identification for the thread with a message in the form '[New systag]'. systag is a  
46  
Running Programs Under GDB  
 
thread identifier whose form varies depending on the particular system. For example,  
on LynxOS, you might see  
[New process 35 thread 27]  
when GDB notices a new thread. In contrast, on an SGI system, the systag is simply  
something like 'process 368', with no further qualifier.  
For debugging purposes, GDB associates its own thread number―always a single  
integer―with each thread in your program.  
info threads  
Display a summary of all threads currently in your program. GDB  
displays for each thread (in this order):  
1. the thread number assigned by GDB  
2. the target system's thread identifier (systag)  
3. the current stack frame summary for that thread  
An asterisk '*' to the left of the GDB thread number indicates the  
current thread.  
For example,  
((gdb)) info threads  
3 process 35 thread 27 0x34e5 in sigpause ()  
2 process 35 thread 23 0x34e5 in sigpause ()  
* 1 process 35 thread 13 main (argc=1, argv=0x7ffffff8)  
at threadtest.c:68  
On HP-UX systems:  
For debugging purposes, GDB associates its own thread number―a small integer  
assigned in thread-creation order―with each thread in your program.  
Whenever GDB detects a new thread in your program, it displays both GDB's thread  
number and the target system's identification for the thread with a message in the form  
'[New systag]'. systag is a thread identifier whose form varies depending on the  
particular system. For example, on HP-UX, you see  
[New thread 2 (system thread 26594)]  
when GDB notices a new thread.  
On HP-UX systems, you can control the display of thread creation messages. Following  
commands are used to control the display of thread creation:  
set threadverbose on  
Enable the output of informational messages  
regarding thread creation. The default setting is on.  
You can set it to offto stop displaying of messages.  
set threadverbose off  
Disable the output of informational messages  
regarding thread creation. The default setting is on.  
You can set it to onto display messages.  
4.9 Debugging programs with multiple threads  
47  
show threadverbose  
Display whether set threadverboseis onor  
off.  
Here are commands to get more information about threads:  
info threads  
Display a summary of all threads currently in  
your program. GDB displays for each thread (in  
this order):  
1. the thread number assigned by GDB  
2. the target system's thread identifier (systag)  
3. the current stack frame summary for that  
thread  
4. the priority of a thread  
An asterisk '*' to the left of the GDB thread  
number indicates the current thread.  
For example,  
((gdb)) info threads  
* 3 system thread 26607 worker  
(wptr=0x7b09c318 "@") \  
at quicksort.c:137  
2 system thread 26606 0x7b0030d8 in  
__ksleep () \  
from /usr/lib/libc.2  
1 system thread 27905 0x7b003498 in _brk  
() \  
from /usr/lib/libc.2  
thread threadno  
Make thread number threadno the current  
thread. The command argument threadno is  
the internal GDB thread number, as shown in  
the first field of the 'info threads' display.  
GDB responds by displaying the system identifier  
of the thread you selected, and its current stack  
frame summary:  
((gdb)) thread 2  
[Switching to thread 2 (system thread  
26594)]  
0x34e5 in sigpause ()  
As with the '[New ...]' message, the form of  
the text after 'Switching to' depends on your  
system's conventions for identifying threads.  
thread apply [threadno]  
[all] args  
The thread applycommand allows you to  
apply a command to one or more threads. Specify  
the numbers of the threads that you want affected  
with the command argument threadno.  
48  
Running Programs Under GDB  
threadno is the internal GDB thread number,  
as shown in the first field of the 'info threads'  
display. To apply a command to all threads, use  
thread apply allargs.  
Whenever GDB stops your program, due to a breakpoint or a signal, it automatically  
selects the thread where that breakpoint or signal happened. GDB alerts you to the  
context switch with a message of the form '[Switching to systag]' to identify the  
thread.  
about how GDB behaves when you stop and start programs with multiple threads.  
See “Killing the child process” (page 45), for information about watchpoints in programs  
with multiple threads.  
NOTE: On HP-UX 11.x, debugging a multi-thread process can cause a deadlock if  
the process is waiting for an NFS-server response. A thread can be stopped while asleep  
in this state, and NFS holds a lock on the rnode while asleep.  
To prevent the thread from being interrupted while holding the rnode lock, make the  
NFS mount non-interruptible with the '-nointr' flag. See mount(1).  
4.10 Debugging programs with multiple processes  
On most systems, GDB has no special support for debugging programs which create  
additional processes using the forkfunction. When a program forks, GDB will continue  
to debug the parent process and the child process will run unimpeded. If you have set  
a breakpoint in any code which the child then executes, the child will get a SIGTRAP  
signal which (unless it catches the signal) will cause it to terminate.  
However, if you want to debug the child process there is a workaround which isn't too  
painful. Put a call to sleepin the code which the child process executes after the fork.  
It may be useful to sleep only if a certain environment variable is set, or a certain file  
exists, so that the delay need not occur when you do not want to run GDB on the child.  
While the child is sleeping, use the psprogram to get its process ID. Then tell GDB (a  
new invocation of GDB if you are also debugging the parent process) to attach to the  
child process (see “Debugging a Running Process” (page 44)). From that point on you  
can debug the child process just like any other process which you attached to.  
On HP-UX (11.x and later only), GDB provides support for debugging programs that  
create additional processes using the forkor vforkfunction.  
By default, when a program forks, GDB will continue to debug the parent process and  
the child process will run unimpeded.  
If you want to follow the child process instead of the parent process, use the command  
set follow-fork-mode.  
4.10 Debugging programs with multiple processes  
49  
 
set follow-fork-mode mode  
Set the debugger response to a program call of  
forkor vfork. A call to forkor vforkcreates  
a new process. The mode can be:  
parent The original process is debugged after  
a fork. The child process runs  
unimpeded. This is the default.  
child  
The new process is debugged after a  
fork. The parent process runs  
unimpeded.  
show follow-fork-mode  
Display the current debugger response to a fork  
or vforkcall.  
If you ask to debug a child process and a vforkis followed by an exec, GDB executes  
the new target up to the first breakpoint in the new target. If you have a breakpoint set  
on mainin your original program, the breakpoint will also be set on the child process's  
main.  
When a child process is spawned by vfork, you cannot debug the child or parent until  
an execcall completes.  
If you issue a runcommand to GDB after an execcall executes, the new target restarts.  
To restart the parent process, use the filecommand with the parent executable name  
as its argument.  
You can use the catchcommand to make GDB stop whenever a fork, vfork, or  
execcall is made. See “Setting catchpoints” (page 56).  
50  
Running Programs Under GDB  
5 Stopping and Continuing  
The principal purpose of a debugger is to let you stop your program before it terminates  
abnormally or runs into trouble, so that you can investigate and determine the reason.  
Inside GDB, your program can stop for several reasons, such as a signal, a breakpoint,  
or reaching a new line after a GDB command such as step. You can then examine and  
change variables, set new breakpoints or remove old ones, and then continue execution.  
Usually, the messages shown by GDB provide information on the status of your  
program―but you can also explicitly request this information at any time.  
info program  
Display information about the status of your program: whether it  
is running or not, what process it is, and why it stopped.  
5.1 Breakpoints  
A breakpoint makes your program stop whenever a certain point in the program is  
reached. For each breakpoint, you can add conditions to control in finer detail whether  
your program stops. You can set breakpoints with the breakcommand and its variants.  
(see “Setting breakpoints” (page 52)) You can stop your program by line number,  
function name or an address in the program.  
You can arrange to have values from your program displayed automatically whenever  
GDB stops at a breakpoint. See Automatic display” (page 89).  
In HP-UX, SunOS 4.x, SVR4, and Alpha OSF/1 configurations, you can set breakpoints  
in shared libraries before the executable is run. See “Debugging support for shared  
A catchpoint is another special breakpoint that stops your program when a certain kind  
of event occurs, such as the throwing of a C++ exception or the loading of a library. As  
with watchpoints, you use a different command to set a catchpoint (see “Setting  
catchpoints” (page 56)), but apart from that, you can manage a catchpoint like any  
other breakpoint. (To stop when your program receives a signal, use the handle  
command; see “Signals” (page 67).)  
GDB assigns a number to each breakpoint, watchpoint, or catchpoint when you create  
it; these numbers are successive integers starting with one. In many of the commands  
for controlling various features of breakpoints you use the breakpoint number to say  
which breakpoint you want to change. Each breakpoint may be enabled or disabled;  
if disabled, it has no effect on your program until you enable it again.  
Some GDB commands accept a range of breakpoints on which to operate. A breakpoint  
range is either a single breakpoint number, like '5', or two such numbers, in increasing  
order, separated by a hyphen, like '5-7'. When a breakpoint range is given to a  
command, all breakpoint in that range are operated on.  
5.1 Breakpoints  
51  
   
5.1.1 Setting breakpoints  
Breakpoints are set with the breakcommand (abbreviated b). The debugger  
convenience variable '$bpnum' records the number of the breakpoint you have set most  
recently; see “Convenience variables” (page 96), for a discussion of what you can do  
with convenience variables.  
You have several ways to say where the breakpoint should go.  
break function  
Set a breakpoint at entry to function function. When  
using source languages that permit overloading  
of symbols, such as C++, function may refer to more  
than one possible place to break. See “Breakpoint  
menus” (page 62), for a discussion of that  
situation.  
break +offset, break  
-offset  
Set a breakpoint some number of lines forward or  
back from the position at which execution stopped  
in the currently selected stack frame. (See “Stack  
frames” (page 71), for a description of stack  
frames.)  
break linenum  
Set a breakpoint at line linenum in the current  
source file. The current source file is the last file  
whose source text was printed. The breakpoint  
will stop your program just before it executes any  
of the code on that line.  
break filename:linenum  
Set a breakpoint at line linenum in source file  
filename.  
break filename:function  
Set a breakpoint at entry to function function  
found in file filename. Specifying a fie name as  
well as a function name is superfluous except  
when multiple files contain similarly named  
functions.  
break *address  
Set a breakpoint at address address. You can use  
this to set breakpoints in parts of your program  
which do not have debugging information or  
source files.  
break  
When called without any arguments, breaksets  
a breakpoint at the next instruction to be executed  
in the selected stack frame (see Chapter 6  
(page 71)). In any selected frame but the  
innermost, this makes your program stop as soon  
as control returns to that frame. This is similar to  
the effect of a finishcommand in the frame  
inside the selected frame―except that finish  
52  
Stopping and Continuing  
 
does not leave an active breakpoint. If you use  
breakwithout an argument in the innermost  
frame, GDB stops the next time it reaches the  
current location; this may be useful inside loops.  
GDB normally ignores breakpoints when it  
resumes execution, until at least one instruction  
has been executed. If it did not do this, you would  
be unable to proceed past a breakpoint without  
first disabling the breakpoint. This rule applies  
whether or not the breakpoint already existed  
when your program stopped.  
break ... if cond  
Set a breakpoint with condition cond; evaluate  
the expression cond each time the breakpoint is  
reached, and stop only if the value is  
nonzero―that is, if cond evaluates as true. '...'  
stands for one of the possible arguments described  
above (or no argument) specifying where to break.  
information on breakpoint conditions.  
tbreak args  
Set a breakpoint enabled only for one stop. args  
are the same as for the breakcommand, and the  
breakpoint is set in the same way, but the  
breakpoint is automatically deleted after the first  
time your program stops there. See “Disabling  
hbreak args  
Set a hardware-assisted breakpoint. args are the  
same as for the breakcommand and the  
breakpoint is set in the same way, but the  
breakpoint requires hardware support and some  
target hardware may not have this support. The  
main purpose of this is EPROM/ROM code  
debugging, so you can set a breakpoint at an  
instruction without changing the instruction. This  
can be used with the new trap-generation provided  
by SPARClite DSU and some x86-based targets.  
These targets will generate traps when a program  
accesses some data or instruction address that is  
assigned to the debug registers. However, the  
hardware breakpoint registers can take a limited  
number of breakpoints. For example, on the DSU,  
only two data breakpoints can be set at a time, and  
GDB will reject this command if more than two  
5.1 Breakpoints  
53  
are used. Delete or disable unused hardware  
breakpoints before setting new ones (see  
thbreak args  
Set a hardware-assisted breakpoint enabled only  
for one stop. args are the same as for the hbreak  
command and the breakpoint is set in the same  
way. However, like the tbreakcommand, the  
breakpoint is automatically deleted after the first  
time your program stops there. Also, like the  
hbreakcommand, the breakpoint requires  
hardware support and some target hardware may  
not have this support. See “Disabling breakpoints”  
rbreak regex  
Set breakpoints on all functions matching the  
regular expression regex. This command sets an  
unconditional breakpoint on all matches, printing  
a list of all breakpoints it set. Once these  
breakpoints are set, they are treated just like the  
breakpoints set with the breakcommand. You  
can delete them, disable them, or make them  
conditional the same way as any other breakpoint.  
The syntax of the regular expression is the  
standard one used with tools like 'grep'. Note that  
this is different from the syntax used by shells, so  
for instance foo*matches all functions that  
include an fofollowed by zero or more os. There  
is an implicit .* leading and trailing the regular  
expression you supply, so to match only functions  
that begin with foo, use ^foo.  
When debugging C++ programs, rbreakis useful  
for setting breakpoints on overloaded functions  
that are not members of any special classes.  
info breakpoints [n], info Print a table of all breakpoints, watchpoints, and  
catchpoints set and not deleted, with the following  
columns for each breakpoint:  
break [n], info watchpoints  
[n]  
Breakpoint Numbers,  
Type  
Breakpoint,  
watchpoint, or  
catchpoint.  
Disposition  
Whether the  
breakpoint is  
54  
Stopping and Continuing  
marked to be  
disabled or deleted  
when hit.  
Enabled or Disabled  
Enabled  
breakpoints are  
marked with 'y'. 'n'  
marks breakpoints  
that are not  
enabled.  
Address  
What  
Where the  
breakpoint is in  
your program, as a  
memory address.  
Where the  
breakpoint is in the  
source for your  
program, as a file  
and line number.  
If a breakpoint is conditional, info breakshows  
the condition on the line following the affected  
breakpoint; breakpoint commands, if any, are  
listed after that.  
info break with a breakpoint number n as  
argument lists only that breakpoint. The  
convenience variable $_and the default  
examining-address for the x command are set to  
the address of the last breakpoint listed (see  
info break displays a count of the number of  
times the breakpoint has been hit. This is especially  
useful in conjunction with the ignorecommand.  
You can ignore a large number of breakpoint hits,  
look at the breakpoint info to see how many times  
the breakpoint was hit, and then run again,  
ignoring one less than that number. This will get  
you quickly to the last hit of that breakpoint.  
GDB allows you to set any number of breakpoints at the same place in your program.  
There is nothing silly or meaningless about this. When the breakpoints are conditional,  
this is even useful (see “Break conditions” (page 59).  
5.1 Breakpoints  
55  
GDB itself sometimes sets breakpoints in your program for special purposes, such as  
proper handling of longjmp(in C programs). These internal breakpoints are assigned  
negative numbers, starting with -1; 'info breakpoints' does not display them.  
You can see these breakpoints with the GDB maintenance command 'maint info  
breakpoints'.  
maint info breakpoints  
Using the same format as 'info breakpoints',  
display both the breakpoints you have set explicitly,  
and those GDB is using for internal purposes.  
Internal breakpoints are shown with negative  
breakpoint numbers. The type column identifies  
what kind of breakpoint is shown:  
breakpoint  
watchpoint  
longjmp  
Normal, explicitly set  
breakpoint.  
Normal, explicitly set  
watchpoint.  
Internal breakpoint, used to  
handle correctly stepping  
through longjmpcalls.  
longjmp resume  
until  
Internal breakpoint at the  
target of a longjmp.  
Temporary internal  
breakpoint used by the GDB  
untilcommand.  
finish  
Temporary internal  
breakpoint used by the GDB  
finishcommand.  
shlib events  
Shared library events.  
5.1.2 Setting catchpoints  
You can use catchpoints to cause the debugger to stop for certain kinds of program  
events, such as C++ exceptions or the loading of a shared library. Use the catch  
command to set a catchpoint.  
catch event  
Stop when event occurs. event can be any of the following:  
throw  
catch  
exec  
The throwing of a C++ exception.  
The catching of a C++ exception.  
A call to exec. This is currently only  
available for HP-UX.  
fork  
A call to fork. This is currently only  
available for HP-UX.  
56  
Stopping and Continuing  
 
vfork  
A call to vfork. This is currently only  
available for HP-UX.  
load, load  
libname  
The dynamic loading of any shared library,  
or the loading of the library libname. This  
is currently only available for HP-UX.  
unload, unload  
libname  
The unloading of any dynamically loaded  
shared library, or the unloading of the  
library libname. This is currently only  
available for HP-UX.  
tcatch event  
Set a catchpoint that is enabled only for one stop. The catchpoint  
is automatically deleted after the first time the event is caught.  
Use the info breakcommand to list the current catchpoints.  
There are currently some limitations to C++ exception handling (catch throwand  
catch catch) in GDB:  
If you call a function interactively, GDB normally returns control to you when the  
function has finished executing. If the call raises an exception, however, the call  
may bypass the mechanism that returns control to you and cause your program  
either to abort or to simply continue running until it hits a breakpoint, catches a  
signal that GDB is listening for, or exits. This is the case even if you set a catchpoint  
for the exception; catchpoints on exceptions are disabled within interactive calls.  
You cannot raise an exception interactively.  
You cannot install an exception handler interactively.  
Sometimes catchis not the best way to debug exception handling: if you need to know  
exactly where an exception is raised, it is better to stop before the exception handler is  
called, since that way you can see the stack before any unwinding takes place. If you  
set a breakpoint in an exception handler instead, it may not be easy to find out where  
the exception was raised.  
To stop just before an exception handler is called, you need some knowledge of the  
implementation. In the case of GNU C++, exceptions are raised by calling a library  
function named _ _raise_exceptionwhich has the following ANSI C interface:  
/* addr is where the exception identifier is stored.  
id is the exception identifier. */  
void __raise_exception (void **addr, void *id);  
To make the debugger catch all exceptions before any stack unwinding takes place, set  
a breakpoint on __raise_exception(see “Breakpoints” (page 51)).  
With a conditional breakpoint (see “Break conditions” (page 59)) that depends on the  
value of id, you can stop your program when a specific exception is raised. You can  
use multiple conditional breakpoints to stop your program when any of a number of  
exceptions are raised.  
5.1 Breakpoints  
57  
5.1.3 Deleting breakpoints  
It is often necessary to eliminate a breakpoint, watchpoint, or catchpoint once it has  
done its job and you no longer want your program to stop there. This is called deleting  
the breakpoint. A breakpoint that has been deleted no longer exists; it is forgotten.  
With the clearcommand you can delete breakpoints according to where they are in  
your program. With the deletecommand you can delete individual breakpoints,  
watchpoints, or catchpoints by specifying their breakpoint numbers.  
It is not necessary to delete a breakpoint to proceed past it. GDB automatically ignores  
breakpoints on the first instruction to be executed when you continue execution without  
changing the execution address.  
clear  
Delete any breakpoints at the next instruction to  
be executed in the selected stack frame (see  
innermost frame is selected, this is a good way  
to delete a breakpoint where your program just  
stopped.  
clear function, clear  
filename:function  
Delete any breakpoints set at entry to the function  
function.  
clear linenum, clear  
filename:linenum  
Delete any breakpoints set at or within the code  
of the specified line.  
delete [breakpoints]  
[range...]  
Delete the breakpoints, watchpoints, or  
catchpoints of the breakpoint ranges specified as  
arguments. If no argument is specified, delete all  
breakpoints (GDB asks confirmation, unless you  
have set confirm off). You can abbreviate  
this command as d.  
5.1.4 Disabling breakpoints  
Rather than deleting a breakpoint, watchpoint, or catchpoint, you might prefer to disable  
it. This makes the breakpoint inoperative as if it had been deleted, but remembers the  
information on the breakpoint so that you can enable it again later.  
You disable and enable breakpoints, watchpoints, and catchpoints with the enable  
and disablecommands, optionally specifying one or more breakpoint numbers as  
arguments. Use info break or info watchto print a list of breakpoints,  
watchpoints, and catchpoints if you do not know which numbers to use.  
A breakpoint, watchpoint, or catchpoint can have any of four different states of  
enablement:  
Enabled. The breakpoint stops your program. A breakpoint set with the break  
command starts out in this state.  
Disabled. The breakpoint has no effect on your program.  
58  
Stopping and Continuing  
   
Enabled once. The breakpoint stops your program, but then becomes disabled.  
Enabled for deletion. The breakpoint stops your program, but immediately after  
it does so it is deleted permanently. A breakpoint set with the tbreakcommand  
starts out in this state.  
You can use the following commands to enable or disable breakpoints, watchpoints,  
and catchpoints:  
disable [breakpoints]  
[range...]  
Disable the specified breakpoints―or all  
breakpoints, if none are listed. A disabled  
breakpoint has no effect but is not forgotten. All  
options such as ignore-counts, conditions, and  
commands are remembered in case the  
breakpoint is enabled again later. You may  
abbreviate disableas dis.  
enable [breakpoints]  
[range...]  
Enable the specified breakpoints (or all defined  
breakpoints). They become effective once again  
in stopping your program.  
enable [breakpoints] once  
range...  
Enable the specified breakpoints temporarily.  
GDB disables any of these breakpoints  
immediately after stopping your program.  
enable [breakpoints] delete Enable the specified breakpoints to work once,  
then die. GDB deletes any of these breakpoints  
as soon as your program stops there.  
range...  
Except for a breakpoint set with tbreak(see “Setting breakpoints” (page 52)),  
breakpoints that you set are initially enabled; subsequently, they become disabled or  
enabled only when you use one of the commands above. (The command untilcan  
set and delete a breakpoint of its own, but it does not change the state of your other  
5.1.5 Break conditions  
The simplest sort of breakpoint breaks every time your program reaches a specified  
place. You can also specify a condition for a breakpoint. A condition is just a Boolean  
expression in your programming language (see “Expressions” (page 83)). A breakpoint  
with a condition evaluates the expression each time your program reaches it, and your  
program stops only if the condition is true.  
This is the converse of using assertions for program validation; in that situation, you  
want to stop when the assertion is violated―that is, when the condition is false. In C,  
if you want to test an assertion expressed by the condition assert, you should set the  
condition '! assert' on the appropriate breakpoint.  
Conditions are also accepted for watchpoints; you may not need them, since a  
watchpoint is inspecting the value of an expression anyhow―but it might be simpler,  
5.1 Breakpoints  
59  
 
say, to just set a watchpoint on a variable name, and specify a condition that tests  
whether the new value is an interesting one.  
Break conditions can have side effects, and may even call functions in your program.  
This can be useful, for example, to activate functions that log program progress, or to  
use your own print functions to format special data structures. The effects are completely  
predictable unless there is another enabled breakpoint at the same address. (In that  
case, GDB might see the other breakpoint first and stop your program without checking  
the condition of this one.) Note that breakpoint commands are usually more convenient  
and flexible than break conditions for the purpose of performing side effects when a  
breakpoint is reached (see “Breakpoint command lists” (page 61)).  
Break conditions can be specified when a breakpoint is set, by using 'if' in the arguments  
to the breakcommand. See “Setting breakpoints” (page 52). They can also be changed  
at any time with the conditioncommand.  
You can also use the if keyword with the watchcommand. The catchcommand does  
not recognize the if keyword; conditionis the only way to impose a further condition  
on a catchpoint.  
condition bnum expression  
Specify expression as the break condition for  
breakpoint, watchpoint, or catchpoint number  
bnum. After you set a condition, breakpoint bnum  
stops your program only if the value of  
expression is true (nonzero, in C). When you  
use condition, GDB checks expression  
immediately for syntactic correctness, and to  
determine whether symbols in it have referents  
in the context of your breakpoint. If expression  
uses symbols not referenced in the context of the  
breakpoint, GDB prints an error message:  
No symbol "foo" in current context.  
GDB does not actually evaluate expression at the  
time the conditioncommand (or a command  
that sets a breakpoint with a condition, like break  
if ...) is given, however. See “Expressions”  
condition bnum  
Remove the condition from breakpoint number  
bnum. It becomes an ordinary unconditional  
breakpoint.  
A special case of a breakpoint condition is to stop only when the breakpoint has been  
reached a certain number of times. This is so useful that there is a special way to do it,  
using the ignore count of the breakpoint. Every breakpoint has an ignore count,  
which is an integer. Most of the time, the ignore count is zero, and therefore has no  
effect. But if your program reaches a breakpoint whose ignore count is positive, then  
60  
Stopping and Continuing  
instead of stopping, it just decrements the ignore count by one and continues. As a  
result, if the ignore count value is n, the breakpoint does not stop the next n times your  
program reaches it.  
ignore bnum count  
Set the ignore count of breakpoint number bnum to count.  
The next count times the breakpoint is reached, your  
program's execution does not stop; other than to decrement  
the ignore count, GDB takes no action.  
To make the breakpoint stop the next time it is reached,  
specify a count of zero.  
When you use continueto resume execution of your  
program from a breakpoint, you can specify an ignore count  
directly as an argument to continue, rather than using  
If a breakpoint has a positive ignore count and a condition,  
the condition is not checked. Once the ignore count reaches  
zero, GDB resumes checking the condition.  
You could achieve the effect of the ignore count with a  
condition such as '$foo-- <= 0' using a debugger  
convenience variable that is decremented each time. See  
Ignore counts apply to breakpoints, watchpoints, and catchpoints.  
5.1.6 Breakpoint command lists  
You can give any breakpoint (or watchpoint or catchpoint) a series of commands to  
execute when your program stops due to that breakpoint. For example, you might  
want to print the values of certain expressions, or enable other breakpoints.  
commands [bnum], ...  
command-list ..., end  
Specify a list of commands for breakpoint number  
bnum. The commands themselves appear on the  
following lines. Type a line containing just endto  
terminate the commands.  
To remove all commands from a breakpoint, type  
commandsand follow it immediately with end; that  
is, give no commands.  
With no bnum argument, commandsrefers to the last  
breakpoint, watchpoint, or catchpoint set (not to the  
breakpoint most recently encountered).  
Pressing RET as a means of repeating the last GDB command is disabled within a  
command-list.  
5.1 Breakpoints  
61  
 
You can use breakpoint commands to start your program up again. Simply use the  
continuecommand, or step, or any other command that resumes execution.  
Any other commands in the command list, after a command that resumes execution,  
are ignored. This is because any time you resume execution (even with a simple next  
or step), you may encounter another breakpoint―which could have its own command  
list, leading to ambiguities about which list to execute.  
If the first command you specify in a command list is silent, the usual message about  
stopping at a breakpoint is not printed. This may be desirable for breakpoints that are  
to print a specific message and then continue. If none of the remaining commands print  
anything, you see no sign that the breakpoint was reached. silentis meaningful only  
at the beginning of a breakpoint command list.  
The commands echo, output, and printfallow you to print precisely controlled  
output, and are often useful in silent breakpoints. See “Commands for controlled  
For example, here is how you could use breakpoint commands to print the value of  
x at entry to foowhenever xis positive.  
break foo if x>0  
commands  
silent  
printf "x is %d\n",x  
cont  
end  
One application for breakpoint commands is to compensate for one bug so you can  
test for another. Put a breakpoint just after the erroneous line of code, give it a condition  
to detect the case in which something erroneous has been done, and give it commands  
to assign correct values to any variables that need them. End with the continue  
command so that your program does not stop, and start with the silentcommand  
so that no output is produced. Here is an example:  
break 403  
commands  
silent  
set x = y + 4  
cont  
end  
5.1.7 Breakpoint menus  
Some programming languages (notably C++) permit a single function name to be  
defined several times, for application in different contexts. This is called overloading.  
When a function name is overloaded, 'break function' is not enough to tell GDB  
where you want a breakpoint. If you realize this is a problem, you can use something  
like 'break function(types)' to specify which particular version of the function  
you want. Otherwise, GDB offers you a menu of numbered choices for different possible  
breakpoints, and waits for your selection with the prompt '>'. The first two options are  
62  
Stopping and Continuing  
 
always '[0] cancel' and '[1] all'. Typing 1 sets a breakpoint at each definition of  
function, and typing 0 aborts the breakcommand without setting any new breakpoints.  
For example, the following session excerpt shows an attempt to set a breakpoint at the  
overloaded symbol String::after. We choose three particular definitions of that function  
name:  
((gdb)) b String::after  
[0] cancel  
[1] all  
[2] file:String.cc; line number:867  
[3] file:String.cc; line number:860  
[4] file:String.cc; line number:875  
[5] file:String.cc; line number:853  
[6] file:String.cc; line number:846  
[7] file:String.cc; line number:735  
> 2 4 6  
Breakpoint 1 at 0xb26c: file String.cc, line 867.  
Breakpoint 2 at 0xb344: file String.cc, line 875.  
Breakpoint 3 at 0xafcc: file String.cc, line 846.  
Multiple breakpoints were set.  
Use the "delete" command to delete unwanted  
breakpoints.  
((gdb))  
5.1.8 “Cannot insert breakpoints”  
Under some operating systems, breakpoints cannot be used in a program if any other  
process is running that program. In this situation, attempting to run or continue a  
program with a breakpoint causes GDB to print an error message:  
Cannot insert breakpoints.  
The same program may be running in another process.  
When this happens, you have three ways to proceed:  
1. Remove or disable the breakpoints, then continue.  
2. Suspend GDB, and copy the file containing your program to a new name. Resume  
GDB and use the exec-filecommand to specify that GDB should run your  
program under that name. Then start your program again.  
3. Relink your program so that the text segment is nonsharable, using the linker  
option '-N'. The operating system limitation may not apply to nonsharable  
executables.  
A similar message can be printed if you request too many active hardware-assisted  
breakpoints and watchpoints:  
Stopped; cannot insert breakpoints.  
You may have requested too many hardware breakpoints and watchpoints.  
This message is printed when you attempt to resume the program, since only then  
GDB knows exactly how many hardware breakpoints and watchpoints it needs to  
insert.  
5.1 Breakpoints  
63  
 
When this message is printed, you need to disable or remove some of the  
hardware-assisted breakpoints and watchpoints, and then continue.  
5.2 Continuing and stepping  
Continuing means resuming program execution until your program completes normally.  
In contrast, stepping means executing just one more “step” of your program, where  
“step” may mean either one line of source code, or one machine instruction (depending  
on what particular command you use). Either when continuing or when stepping, your  
program may stop even sooner, due to a breakpoint or a signal. (If it stops due to a  
signal, you may want to use handle, or use 'signal 0' to resume execution. See  
continue [ignore-count], c Resume program execution, at the address where  
your program last stopped; any breakpoints set at  
that address are bypassed. The optional argument  
ignore-count allows you to specify a further  
number of times to ignore a breakpoint at this  
location; its effect is like that of ignore(see “Break  
[ignore-count], fg  
[ignore-count]  
The argument ignore-count is meaningful only  
when your program stopped due to a breakpoint.  
At other times, the argument to continueis  
ignored.  
The synonyms cand fg (for foreground, as the  
debugged program is deemed to be the foreground  
program) are provided purely for convenience,  
and have exactly the same behavior as continue.  
To resume execution at a different place, you can use return(see “Returning from a  
function” (page 121)) to go back to the calling function; or jump(see “Continuing at a  
different address” (page 120)) to go to an arbitrary location in your program.  
A typical technique for using stepping is to set a breakpoint (see “Breakpoints”  
(page 51)) at the beginning of the function or the section of your program where a  
problem is believed to lie, run your program until it stops at that breakpoint, and then  
step through the suspect area, examining the variables that are interesting, until you  
see the problem happen.  
64  
Stopping and Continuing  
 
step  
Continue running your program until control reaches a different  
source line, then stop it and return control to GDB. This  
command is abbreviated s.  
WARNING! If you use the stepcommand while control is  
within a function that was compiled without debugging  
information, execution proceeds until control reaches a function  
that does have debugging information. Likewise, it will not step  
into a function which is compiled without debugging  
information. To stepthrough functions without debugging  
information, use the stepicommand, described below.  
The stepcommand only stops at the first instruction of a source  
line. This prevents the multiple stops that could otherwise occur  
in switch statements, for loops, and so on. stepcontinues to  
stop if a function that has debugging information is called within  
the line. In other words, stepsteps inside any functions called  
within the line.  
Also, the stepcommand only enters a function if there is line  
number information for the function. Otherwise it acts like the  
nextcommand. This avoids problems when using cc -gl  
on MIPS machines. Previously, step entered subroutines if there  
was any debugging information about the routine.  
step count  
Continue running as in step, but do so count times. If a  
breakpoint is reached, or a signal not related to stepping occurs  
before count steps, stepping stops right away.  
next [count]  
Continue to the next source line in the current (innermost) stack  
frame. This is similar to step, but function calls that appear  
within the line of code are executed without stopping. Execution  
stops when control reaches a different line of code at the original  
stack level that was executing when you gave the next  
command. This command is abbreviated n.  
An argument count is a repeat count, as for step.  
The nextcommand only stops at the first instruction of a source  
line. This prevents multiple stops that could otherwise occur in  
switch statements, for loops, and so on.  
finish  
Continue running until just after function in the selected stack  
frame returns. Print the returned value (if any).  
Contrast this with the returncommand (see “Returning from  
5.2 Continuing and stepping  
65  
until, u  
Continue running until a source line past the current line, in the  
current stack frame, is reached. This command is used to avoid  
single stepping through a loop more than once. It is like the  
nextcommand, except that when untilencounters a jump,  
it automatically continues execution until the program counter  
is greater than the address of the jump.  
This means that when you reach the end of a loop after single  
stepping though it, untilmakes your program continue  
execution until it exits the loop. In contrast, a nextcommand  
at the end of a loop simply steps back to the beginning of the  
loop, which forces you to step through the next iteration.  
untilalways stops your program if it attempts to exit the  
current stack frame.  
untilmay produce somewhat counterintuitive results if the  
order of machine code does not match the order of the source  
lines. For example, in the following excerpt from a debugging  
session, the f(frame) command shows that execution is stopped  
at line 206; yet when we use until, we get to line 195:  
((gdb)) f  
#0 main (argc=4, argv=0xf7fffae8) at m4.c:206  
206 expand_input();  
((gdb)) until  
195 for ( ; argc > 0; NEXTARG) {  
This happened because, for execution efficiency, the compiler  
had generated code for the loop closure test at the end, rather  
than the start, of the loop―even though the test in a C for-loop  
is written before the body of the loop. The untilcommand  
appeared to step back to the beginning of the loop when it  
advanced to this expression; however, it has not really gone to  
an earlier statement―not in terms of the actual machine code.  
untilwith no argument works by means of single instruction  
stepping, and hence is slower than untilwith an argument.  
until location, Continue running your program until either the specified  
location is reached, or the current stack frame returns. location  
is any of the forms of argument acceptable to break(see  
“Setting breakpoints” (page 52)). This form of the command  
uses breakpoints, and hence is quicker than untilwithout an  
argument.  
u location  
66  
Stopping and Continuing  
stepi, stepi arg, Execute one machine instruction, then stop and return to the  
debugger.  
si  
It is often useful to do 'display/i $pc' when stepping by  
machine instructions. This makes GDB automatically display  
the next instruction to be executed, each time your program  
An argument is a repeat count, as in step.  
nexti, nexti arg, Execute one machine instruction, but if it is a function call,  
proceed until the function returns.  
ni  
An argument is a repeat count, as in next.  
5.3 Signals  
A signal is an asynchronous event that can happen in a program. The operating system  
defines the possible kinds of signals, and gives each kind a name and a number. For  
example, in Unix SIGINTis the signal a program gets when you type an interrupt  
character (often C- c); SIGSEGVis the signal a program gets from referencing a place  
in memory far away from all the areas in use; SIGALRMoccurs when the alarm clock  
timer goes off (which happens only if your program has requested an alarm).  
Some signals, including SIGALRM, are a normal part of the functioning of your program.  
Others, such as SIGSEGV, indicate errors; these signals are fatal (they kill your program  
immediately) if the program has not specified in advance some other way to handle  
the signal. SIGINTdoes not indicate an error in your program, but it is normally fatal  
so it can carry out the purpose of the interrupt: to kill the program.  
GDB has the ability to detect any occurrence of a signal in your program. You can tell  
GDB in advance what to do for each kind of signal.  
Normally, GDB is set up to ignore non-erroneous signals like SIGALRM(so as not to  
interfere with their role in the functioning of your program) but to stop your program  
immediately whenever an error signal happens. You can change these settings with  
the handlecommand.  
5.3 Signals  
67  
 
NOTE: Use caution if you disable all signals from certain processes. Disabling  
'SIGTRAP' in your program may cause your program to hang.  
HP-UX uses 'SIGTRAP' to communicate with the debugger. If you disable all signals  
from certain processes so that signals will be delivered to the right process, your  
program may hang when you try to debug it. This behavior occurs because if you  
disable 'SIGTRAP', the debugger no longer receives notification of events such as  
breakpoint hits and loading or unloading of shared libraries.  
To prevent this problem:  
Make certain you set this flag:  
(gdb) set complain-if-sigtrap-disabled on  
Also make certain the following warning was not emitted by the debugger before your  
program hung:  
Warning: Thread %d (in process %d) has disabled SIGTRAPs.  
Debugging this thread is probably impossible.  
If you do not want to see this message again, use:  
"set complain-if-sigtrap-disabled 0"  
info signals, info handle  
Print a table of all the kinds of signals and how  
GDB has been told to handle each one. You can  
use this to see the signal numbers of all the  
defined types of signals.  
info handleis an alias for info signals.  
handle signal keywords...  
Change the way GDB handles signal signal.  
signal can be the number of a signal or its name  
(with or without the 'SIG' at the beginning). The  
keywords say what change to make.  
The keywords allowed by the handlecommand can be abbreviated. Their full names  
are:  
nostop  
GDB should not stop your program when this signal happens. It may still  
print a message telling you that the signal has come in.  
stop  
GDB should stop your program when this signal happens. This implies  
the print keyword as well.  
print  
GDB should print a message when this signal happens.  
noprint GDB should not mention the occurrence of the signal at all. This implies  
the nostop keyword as well.  
68  
Stopping and Continuing  
pass  
GDB should allow your program to see this signal; your program can  
handle the signal, or else it may terminate if the signal is fatal and not  
handled.  
nopass  
GDB should not allow your program to see this signal.  
When a signal stops your program, the signal is not visible to the program until you  
continue. Your program sees the signal then, if passis in effect for the signal in question  
at that time. In other words, after GDB reports a signal, you can use the handle  
command with passor nopassto control whether your program sees that signal  
when you continue.  
You can also use the signalcommand to prevent your program from seeing a signal,  
or cause it to see a signal it normally would not see, or to give it any signal at any time.  
For example, if your program stopped due to some sort of memory reference error,  
you might store correct values into the erroneous variables and continue, hoping to  
see more execution; but your program would probably terminate immediately as a  
result of the fatal signal once it saw the signal. To prevent this, you can continue with  
5.4 Stopping and starting multi-thread programs  
When your program has multiple threads (see “Debugging programs with multiple  
threads” (page 46)), you can choose whether to set breakpoints on all threads, or on a  
particular thread.  
break linespec thread  
threadno, break linespec  
thread threadno if ...  
linespec specifies source lines; there are several  
ways of writing them, but the effect is always to  
specify some source line.  
Use the qualifier 'thread threadno' with a  
breakpoint command to specify that you only  
want GDB to stop the program when a particular  
thread reaches this breakpoint. threadno is one  
of the numeric thread identifiers assigned by  
GDB, shown in the first column of the 'info  
threads' display.  
If you do not specify 'thread threadno' when  
you set a breakpoint, the breakpoint applies to  
all threads of your program.  
You can use the threadqualifier on conditional  
breakpoints as well; in this case, place 'thread  
threadno' before the breakpoint condition, like  
this:  
((gdb)) break frik.c:13 thread 28 if  
bartab > lim  
5.4 Stopping and starting multi-thread programs  
69  
 
Whenever your program stops under GDB for any reason, all threads of execution stop,  
not just the current thread. This allows you to examine the overall state of the program,  
including switching between threads, without worrying that things may change  
underfoot.  
Conversely, whenever you restart the program, all threads start executing. This is true  
even when single-stepping with commands like stepor next.  
In particular, GDB cannot single-step all threads in lockstep. Since thread scheduling  
is up to your debugging target's operating system (not controlled by GDB), other threads  
may execute more than one statement while the current thread completes a single step.  
Moreover, in general other threads stop in the middle of a statement, rather than at a  
clean statement boundary, when the program stops.  
You might even find your program stopped in another thread after continuing or even  
single-stepping. This happens whenever some other thread runs into a breakpoint, a  
signal, or an exception before the first thread completes whatever you requested.  
On some OSes, you can lock the OS scheduler and thus allow only a single thread to  
run.  
set scheduler-locking mode Set the scheduler locking mode. If it is off, then  
there is no locking and any thread may run at  
any time. If on, then only the current thread may  
run when the inferior is resumed. The step  
mode optimizes for single-stepping. It stops other  
threads from “seizing the prompt” by  
preempting the current thread while you are  
stepping. Other threads will only rarely (or  
never) get a chance to run when you step. They  
are more likely to run when you 'next' over a  
function call, and they are completely free to run  
when you use commands like 'continue',  
'until', or 'finish'. However, unless another  
thread hits a breakpoint during its timeslice, they  
will never steal the GDB prompt away from the  
thread that you are debugging.  
show scheduler-locking  
Display the current scheduler locking mode.  
70  
Stopping and Continuing  
6 Examining the Stack  
When your program has stopped, the first thing you need to know is where it stopped  
and how it got there.  
Each time your program performs a function call, information about the call is generated.  
The information includes the location of the call in your program, the arguments of  
the call, and the local variables of the function being called. The information is saved  
in a block of data called a stack frame. The stack frames are allocated in a region of  
memory called the call stack.  
The GDB commands for examining the stack allow you to view all of this information.  
6.1 Stack frames  
The call stack is divided up into contiguous pieces called stack frames, or frames for short;  
each frame is the data associated with one call to one function. The frame contains the  
arguments given to the function, the local variables, and the address at which the  
function is executing.  
When your program is started, the stack has only one frame, that of the function main.  
This is called the initial frame or the outermost frame. Each time a function is called, a  
new frame is made. Each time a function returns, the frame for that function invocation  
is eliminated. If a function is recursive, there can be many frames for the same function.  
The frame for the function in which execution is actually occurring is called the innermost  
frame. This is the most recently created of all the stack frames that still exist.  
Inside your program, stack frames are identified by their addresses. A stack frame  
consists of many bytes, each of which has its own address; each kind of computer has  
a convention for choosing one byte whose address serves as the address of the frame.  
Usually this address is kept in a register called the frame pointer register while execution  
is going on in that frame.  
GDB assigns numbers to all existing stack frames, starting with zero for the innermost  
frame, one for the frame that called it, and so on upward. These numbers do not really  
exist in your program; they are assigned by GDB to give you a way of designating stack  
frames in GDB commands.  
One of the stack frames is selected by GDB and the GDB commands refer implicitly to  
the selected frame. In particular, whenever you ask GDB for the value of a variable in  
your program, the value is found in the selected frame. There are special GDB  
commands to select whichever frame you are interested in. See “Selecting a frame”  
When your program stops, GDB automatically selects the currently executing frame  
and describes it briefly, similar to the framecommand (see “Information about a  
6.1 Stack frames  
71  
   
6.2 Stacks Without frames  
Some compilers provide a way to compile functions so that they operate without stack  
frames. (For example, the gcc option  
'-fomit-frame-pointer'  
generates functions without a frame.) This is occasionally done with heavily used  
library functions to save the frame setup time. GDB has limited facilities for dealing  
with these function invocations. If the innermost function invocation has no stack frame,  
GDB nevertheless regards it as though it had a separate frame, which is numbered zero  
as usual, allowing correct tracing of the function call chain. However, GDB has no  
provision for frameless functions elsewhere in the stack.  
6.3 Commands for Examining the Stack  
The following commands are used for examining the stack:  
frame args  
Select and print a stack frame. With no argument, prints the selected  
stack frame. An argument specifies the frame to select. It can be a  
stack frame number or the address of the frame. With argument,  
nothing is printed if input is coming from a command file or a  
user-defined command.  
select-frame  
The select-framecommand allows you to move from one stack  
frame to another without printing the frame. This is the silent  
version of frame.  
6.4 Backtraces  
A backtrace is a report of the active stack frames instantiated by the execution of a  
program. It shows one line per frame, for all the active frames, starting with the currently  
executing frame (frame zero), followed by its caller (frame one), and on up the stack.  
The following commands are used for backtrace:  
backtrace, bt  
Print a backtrace of the entire stack: one line per  
frame for all frames in the stack.  
You can stop the backtrace at any time by typing  
the system interrupt character, normally C-c.  
backtrace n, bt n  
Similar, but print only the innermost n frames.  
backtrace -n, bt -n  
backtrace-other-thread  
Similar, but print only the outermost n frames.  
Print backtrace of all stack frames for a thread with  
stack pointer SP and program counter PC. This  
command is useful in cases where the debugger  
does not support a user thread package fully.  
72  
Examining the Stack  
     
The names whereand info stack(abbreviated info s) are additional aliases for  
backtrace.  
Each line in the backtrace shows the frame number and the function name. The program  
counter value is also shown―unless you use set print address off. The backtrace  
also shows the source file name and line number, as well as the arguments to the  
function. The program counter value is omitted if it is at the beginning of the code for  
that line number.  
Here is an example of a backtrace. It was made with the command 'bt 3', so it shows  
the innermost three frames.  
#0 m4_traceon (obs=0x24eb0, argc=1, argv=0x2b8c8)  
at builtin.c:993  
#1 0x6e38 in expand_macro (sym=0x2b600) at macro.c:242  
#2 0x6840 in expand_token (obs=0x0, t=177664, td=0xf7fffb08)  
at macro.c:71  
(More stack frames follow...)  
The display for frame zero does not begin with a program counter value, indicating  
that your program has stopped at the beginning of the code for line 993 of builtin.c.  
6.5 Selecting a frame  
Most commands for examining the stack and other data in your program work on  
whichever stack frame is selected at the moment.  
The following commands are used for selecting a stack frame; all of them finish by  
printing a brief description of the stack frame selected.  
frame n, f n Select frame number n. Recall that frame zero is the innermost  
(currently executing) frame, frame one is the frame that called the  
innermost one, and so on. The highest-numbered frame is the one for  
main.  
frame addr, Select the frame at address addr. This is useful mainly if the chaining  
f addr  
of stack frames has been damaged by a bug, making it impossible for  
GDB to assign numbers properly to all frames. In addition, this can  
be useful when your program has multiple stacks and switches  
between them.  
6.5 Selecting a frame  
73  
 
NOTE:  
On the SPARC architecture, frameneeds two addresses to select  
an arbitrary frame: a frame pointer and a stack pointer.  
On the MIPS and Alpha architecture, it needs two addresses: a  
stack pointer and a program counter.  
On the 29k architecture, it needs three addresses: a register stack  
pointer, a program counter, and a memory stack pointer.  
up n  
Move n frames up the stack. For positive numbers n, this advances  
toward the outermost frame, to higher frame numbers, to frames that  
have existed longer. n defaults to one.  
down n  
Move n frames down the stack. For positive numbers n, this advances  
toward the innermost frame, to lower frame numbers, to frames that  
were created more recently. n defaults to one. You may abbreviate  
downas do.  
All of these commands end by printing two lines of output describing the frame. The  
first line shows the frame number, the function name, the arguments, and the source  
file and line number of execution in that frame. The second line shows the text of that  
source line.  
For example:  
((gdb)) up  
#1 0x22f0 in main (argc=1, argv=0xf7fffbf4, env=0xf7fffbfc)  
at env.c:10  
10 read_input_file (argv[i]);  
After such a printout, the listcommand with no arguments prints ten lines centered  
on the point of execution in the frame. See “Printing source lines” (page 77).  
up-silently n,  
down-silently n  
These two commands are variants of upand down,  
respectively; they differ in that they do their work silently,  
without causing display of the new frame. They are intended  
primarily for use in GDB command scripts, where the output  
might be unnecessary and distracting.  
6.6 Information about a frame  
The following commands are used to print information about the selected stack frame:  
frame, f  
When used without any argument, this command does not  
change which frame is selected, but prints a brief description  
of the currently selected stack frame. It can be abbreviated f.  
With an argument, this command is used to select a stack  
74  
Examining the Stack  
 
info frame, info  
f
This command prints a verbose description of the selected  
stack frame, including:  
the address of the frame  
the address of the next frame down (called by this frame)  
the address of the next frame up (caller of this frame)  
the language in which the source code corresponding to  
this frame is written  
the address of the frame's arguments  
the address of the frame's local variables  
the program counter saved in it (the address of execution  
in the caller frame)  
which registers were saved in the frame  
The verbose description is useful when something has gone  
wrong that has made the stack format fail to fit the usual  
conventions.  
info frame addr,  
info f addr  
Print a verbose description of the frame at address addr,  
without selecting that frame. The selected frame remains  
unchanged by this command. This requires the same kind of  
address (more than one for some architectures) that you specify  
in the framecommand. See “Selecting a frame” (page 73).  
info args  
Print the arguments of the selected frame, each on a separate  
line.  
info locals  
Print the local variables of the selected frame, each on a  
separate line. These are all variables (declared either static or  
automatic) accessible at the point of execution of the selected  
frame.  
info catch  
Print a list of all the exception handlers that are active in the  
current stack frame at the current point of execution. To see  
other exception handlers, visit the associated frame (using the  
up, down, or framecommands); then type info catch. See  
6.6 Information about a frame  
75  
76  
7 Examining Source Files  
GDB can print parts of the source code of your program, since the debugging  
information recorded in the program tells GDB what source files were used to build  
it. When your program stops, GDB spontaneously prints the line where it stopped.  
Likewise, when you select a stack frame (see “Selecting a frame” (page 73)), GDB prints  
the line where execution in that frame has stopped. You can print other portions of  
source files by explicit command.  
You can invoke GDB from its GNU Emacs interface to view the source code see  
7.1 Printing source lines  
To print lines from a source file, use the listcommand (abbreviated l). By default,  
ten lines are printed. There are several ways to specify what part of the file you want  
to print.  
The following forms of the listcommand are used:  
list linenum  
Prints lines centered around line number linenum in the current  
source file.  
list function  
Prints lines centered around the beginning of function function.  
list  
Prints more lines. The listcommand prints lines following the  
lines printed by a previously executed listcommand. If the  
command prior to executing a listjust printed the stack frame,  
then the listcommand only prints the lines around that line.  
list-  
Prints lines just before the lines last printed.  
By default, GDB prints ten source lines with any of these forms of the listcommand.  
The number of lines printed by GDB can be set by the set listsizecommand. The  
following two forms are supported:  
set listsize count  
Makes the listcommand display count source lines  
(unless the listargument explicitly specifies some other  
number).  
show listsize  
Displays the number of lines that listprints.  
Repeating a listcommand with RET discards the argument, so it is equivalent to  
typing just list. This is more useful than listing the same lines again. An exception  
is made for an argument of '-'; that argument is preserved in repetition so that each  
repetition moves up in the source file.  
In general, the listcommand expects you to supply zero, one or two linespecs.  
Linespecs specify source lines. There are several ways of writing them, but the most  
common way is to specify some source line.  
7.1 Printing source lines  
77  
   
The following arguments can be given to the listcommand:  
list linespec  
Print lines centered around the line specified by linespec.  
list first,last  
Print lines from first to last. Both arguments must be  
linespecs.  
list, last  
list first,  
list +  
Print lines ending with last.  
Print lines starting with first.  
Print lines just after the lines last printed.  
Print lines just before the lines last printed.  
As described in the preceding table.  
list -  
list  
A single source line can be specified in the following ways:  
number  
Specifies line number of the current source file. When a  
listcommand has two linespecs, this refers to the same  
source file as the first linespec.  
+offset  
Specifies the line offset lines after the last line printed.  
When used as the second linespec in a listcommand that  
has two, this specifies the line offset lines down from the  
first linespec.  
-offset  
Specifies the line offset lines before the last line printed.  
filename:number  
function  
Specifies line number in the source file filename.  
Specifies the line that begins the body of the function  
function. For example: in C, this is the line with the open  
brace.  
filename:function  
Specifies the line of the open-brace that begins the body of  
the function function in the file filename. You only  
need the file name with a function name to avoid ambiguity  
when there are identically named functions in different  
source files.  
*address  
Specifies the line containing the program address address.  
address may be any expression.  
7.2 Searching source files  
There are two commands for searching through the current source file for a regular  
expression.  
forward-search regexp,  
search regexp  
The command 'forward-search regexp' checks  
each line, starting with one of the following in the  
last line listed, for a match of the regexp. It lists the  
line that is found. You can use the synonym 'search  
regexp' or abbreviate the command name as fo.  
78  
Examining Source Files  
 
reverse-search regexp  
The command 'reverse-search regexp' checks  
each line, starting with the one before the last line  
listed and going backward, for a match for the  
regexp. It lists the line(s) that is found. You can  
abbreviate this command as rev.  
7.3 Specifying source directories  
Executable programs sometimes do not record the directories of the source files from  
which they were compiled. Even when they do, the directories can be moved between  
the compilation and your debugging session. GDB has a list of directories to search for  
source files; this is called the source path. Each time GDB looks for a source file, it tries  
all the directories in the list, in the order they are present in the list, until it finds a file  
with the desired name. Note that the executable search path is not used for this purpose.  
Neither is the current working directory, unless it happens to be in the source path.  
If GDB cannot find a source file in the source path, and the object program records a  
directory, GDB tries that directory too. If the source path is empty, and there is no  
record of the compilation directory, GDB looks in the current directory as a last resort.  
Whenever you reset or rearrange the source path, GDB clears out any information it  
has cached about where the source files are located and where each line is in the  
respective file.  
When you start GDB, its source path includes only 'cdir' and 'cwd', in that order.  
To add other directories, you can use the directorycommand.  
directory dirname ...,  
dir dirname ...  
Add directory dirname to the front of the source  
path. Several directory names may be given to this  
command, separated by ':' (';' on MS-DOS and  
MS-Windows, where ':' usually appears as part of  
absolute file names) or a whitespace. You can specify  
a directory that is already in the source path; this  
moves it forward, so GDB searches it sooner.  
You can use the string '$cdir' to refer to the  
compilation directory (if one is recorded), and '$cwd'  
to refer to the current working directory. '$cwd' is  
not the same as '.'. The former tracks the current  
working directory as it changes during your GDB  
session, while the latter is immediately expanded to  
the current directory at the time you add an entry to  
the source path.  
directory  
Reset the source path to empty again. This requires  
confirmation from the user.  
7.3 Specifying source directories  
79  
 
show directories  
Print the source path and display the directories it  
contains.  
If your source path is cluttered with directories that are no longer of interest, GDB can  
end up detecting the wrong version of the source. To correct this situation, follow these  
steps:  
1. Use directorywith no arguments to reset the source path to empty.  
2. Use directorywith suitable arguments to reinstall the directories you want in  
the source path. You can add all the directories in one command.  
7.4 Source and machine code  
You can use the command info lineto map source lines to program addresses (and  
vice versa), and the command disassembleto display a range of addresses as machine  
instructions. When run under GNU Emacs mode, the info linecommand causes  
the arrow to point to the line specified. Also, info lineprints addresses in symbolic  
form as well as hex.  
info line linespec  
Print the starting and ending addresses of the compiled  
code for source line linespec. You can specify source  
lines in any of the ways understood by the listcommand  
For example, we can use info lineto discover the location of the object code for the  
first line of function.  
m4_changequote:  
((gdb)) info line m4_changequote  
Line 895 of "builtin.c" starts at pc 0x634c and ends at 0x6350.  
We can also inquire (using *addr as the form for linespec) what source line covers  
a particular address. For example,  
((gdb)) info line *0x63ff  
Line 926 of "builtin.c" starts at pc 0x63e4 and ends at 0x6404.  
After info line, the default address for the xcommand is changed to the starting  
address of the line, so that 'x/i' is sufficient to begin examining the machine code (see  
Section 8.5 (page 87)). Also, this address is saved as the value of the convenience  
variable $_(see Section 8.9 (page 96)).  
disassemble  
This specialized command dumps a range of memory as machine  
instructions. The default memory range is the function surrounding  
the program counter of the selected frame. A single argument to this  
command is a program counter value; GDB dumps the function  
surrounding this value. Two arguments specify a range of addresses  
(first inclusive, second exclusive) to dump.  
80  
Examining Source Files  
 
The following example shows the disassembly of a range of addresses of HP PA-RISC  
2.0 code:  
((gdb)) disas 0x32c4 0x32e4  
Dump of assembler code from 0x32c4 to 0x32e4:  
0x32c4 <main+204>: addil 0,dp  
0x32c8 <main+208>: ldw 0x22c(sr0,r1),r26  
0x32cc <main+212>: ldil 0x3000,r31  
0x32d0 <main+216>: ble 0x3f8(sr4,r31)  
0x32d4 <main+220>: ldo 0(r31),rp  
0x32d8 <main+224>: addil -0x800,dp  
0x32dc <main+228>: ldo 0x588(r1),r26  
0x32e0 <main+232>: ldil 0x3000,r31  
End of assembler dump.  
Some architectures have more than one commonly-used set of instruction mnemonics  
or other syntax.  
set disassembly-flavor  
instruction-set  
Select the instruction set to use when  
disassembling the program via the  
disassembleor x/icommands.  
Currently this command is only defined for the  
Intel x86 family. You can set instruction-set  
to either intelor att. The default is att, the  
AT&T flavor used by default by Unix assemblers  
for x86-based targets.  
7.4 Source and machine code  
81  
82  
8 Examining Data  
The usual way to examine data in your program is with the printcommand  
(abbreviated p), or its synonym inspect. It evaluates and prints the value of an  
expression of the language your program is written in (see Chapter 9 (page 101)).  
The following forms of printcommand are supported:  
print expr,  
print /f expr  
expr is an expression (in the source language). By default, the  
value of expr is printed in a format appropriate to its data type;  
you can choose a different format by specifying '/f', where f is  
a letter specifying the format; see “Output formats” (page 86).  
print, print /f If you omit expr, GDB displays the last value again (from the  
value history; see “Value history” (page 95)). This allows you to  
conveniently inspect the same value in an alternative format.  
A more low-level way of examining data is with the xcommand. It examines data in  
memory at a specified address and prints it in a specified format. See “Examining  
If you are interested in information about types, or about how the fields of a struct or  
a class are declared, use the ptype exp command rather than print. See Chapter 10  
8.1 Expressions  
printand many other GDB commands accept an expression and compute its value.  
Any kind of constant, variable or operator defined by the programming language you  
are using is valid in an expression in GDB. This includes conditional expressions,  
function calls, casts, and string constants. It unfortunately does not include symbols  
defined by preprocessor #definecommands.  
GDB supports array constants in expressions input by the user. The syntax is  
{element, element. . .}. For example, you can use the command print {1,  
2, 3}to build up an array in memory that calls mallocin the target program.  
Because C is so widespread, most of the expressions shown in examples in this manual  
are in C. See Chapter 9 (page 101), for information on how to use expressions in other  
languages.  
In this section, we discuss operators that you can use in GDB expressions regardless  
of your programming language.  
Casts are supported in all languages, not just in C, because it is so useful to cast a  
number into a pointer in order to examine a structure at that address in memory.  
GDB supports these operators, in addition to those common to programming languages:  
@
'@' is a binary operator for treating parts of memory as arrays. Refer  
to See Artificial arrays” (page 85), for more information.  
8.1 Expressions  
83  
   
::  
'::' allows you to specify a variable in terms of the file or function  
where it is defined. See “Program variables” (page 84).  
{type} addr  
Refers to an object of type type stored at address addr in memory.  
addr may be any expression whose value is an integer or pointer  
(but parentheses are required around binary operators, just as in a  
cast). This construct is allowed regardless of what kind of data is  
normally supposed to reside at addr.  
8.2 Program variables  
The most common kind of expression to use is the name of a variable in your program.  
Variables in expressions are understood in the selected stack frame (see “Selecting a  
frame” (page 73); they must be either:  
global (or file-static)  
or  
visible according to the scope rules of the programming language from the point  
of execution in that frame  
This means that in the function  
foo (a)  
int a;  
{
bar (a);  
{
int b = test ();  
bar (b);  
}
}
you can examine and use the variable awhenever your program is executing within  
the function foo, but you can only use or examine the variable bwhile your program  
is executing inside the block where bis declared.  
However, you can refer to a variable or function whose scope is a single source file  
even if the current execution point is not in this file. But it is possible to have more than  
one such variable or function with the same name (in different source files). If that  
happens, referring to that name has unpredictable effects. If you wish, you can specify  
a static variable in a particular function or file, using the colon-colon notation:  
file::variable  
function::variable  
Here file or function is the name of the context for the static variable. In the case  
of file names, you can use quotes to make sure GDB parses the file name as a single  
word. For example, to print a global value of xdefined in 'f2.c':  
((gdb)) p 'f2.c'::x  
84  
Examining Data  
 
This use of '::' is very rarely in conflict with the very similar use of the same notation  
in C++. GDB also supports use of the C++ scope resolution operator in GDB expressions.  
WARNING! Occasionally, a local variable may appear to have the wrong value at  
certain points in a function just after entry to a new scope, and just before exit.  
You may see this problem when you are stepping by machine instructions. This is  
because, on most machines, it takes more than one instruction to set up a stack frame  
(including local variable definitions); if you are stepping by machine instructions,  
variables may appear to have the wrong values until the stack frame is completely  
built. On exit, it usually also takes more than one machine instruction to destroy a stack  
frame; after you begin stepping through that group of instructions, local variable  
definitions may be gone.  
This may also happen when the compiler does significant optimizations. To be sure of  
always seeing accurate values, turn o all optimization when compiling.  
Another possible effect of compiler optimizations is to optimize unused variables out  
of existence, or assign variables to registers (as opposed to memory addresses).  
Depending on the support for such cases offered by the debug info format used by the  
compiler, GDB might not be able to display values for such local variables. If that  
happens, GDB will print a message like this:  
No symbol "foo" in current context.  
To solve such problems, either recompile without optimizations, or use a different  
debug info format, if the compiler supports several such formats. For example, GCC,  
the GNU C/C++ compiler usually supports the '-gstabs' option. The '-gstabs'  
produces debug information in a format that is superior to formats such as COFF. You  
may be able to use DWARF-2('-gdwarf-2'), which is also an effective form for debug  
8.3 Artificial arrays  
It is often useful to print out several successive objects of the same type in memory; a  
section of an array, or an array of dynamically determined size for which only a pointer  
exists in the program.  
You can do this by referring to a contiguous span of memory as an artificial array, using  
the binary operator '@'. The left operand of '@' should be the first element of the desired  
array and be an individual object. The right operand should be the desired length of  
the array. The result is an array value whose elements are all of the type of the left  
argument. The first element is actually the left argument; the second element comes  
from bytes of memory immediately following those that hold the first element, and so  
on. Here is an example. If a program says  
int *array = (int *) malloc (len * sizeof (int));  
you can print the contents of arraywith  
8.3 Artificial arrays  
85  
 
p *array@len  
The left operand of '@' must reside in memory. Array values made with '@' in this way  
behave just like other arrays in terms of subscripting, and are coerced to pointers when  
used in expressions. Artificial arrays most often appear in expressions via the value  
history (see “Value history” (page 95)), after printing one out.  
Another way to create an artificial array is to use a cast. This re-interprets a value as if  
it were an array. The value need not be in memory:  
((gdb)) p/x (short[2])0x12345678  
$1 = {0x1234, 0x5678}  
As a convenience, if you leave the array length out (as in '(type[])value'), GDB  
calculates the size to fill the value (as 'sizeof(value)/sizeof(type)':  
((gdb)) p/x (short[])0x12345678  
$2 = {0x1234, 0x5678}  
Sometimes the artificial array mechanism is not quite enough; in moderately complex  
data structures, the elements of interest may not actually be adjacent―for example, if  
you are interested in the values of pointers in an array. One useful work-around in this  
situation is to use a convenience variable (See “Convenience variables” (page 96)) as  
a counter in an expression that prints the first interesting value, and then repeat that  
expression via RET. For instance, suppose you have an array dtabof pointers to  
structures, and you are interested in the values of a field fvin each structure. Here is  
an example of what you might type:  
set $i = 0  
p dtab[$i++]->fv  
<RET>  
<RET>  
...  
8.4 Output formats  
By default, GDB prints a value according to its data type. Sometimes this is not what  
you want. For example, you might want to print a number in hex, or a pointer in  
decimal. Or you might want to view data in memory at a certain address as a character  
string or as an instruction. To do these things, specify an output format when you print  
a value.  
The simplest use of output formats is to say how to print a value already computed.  
This is done by starting the arguments of the printcommand with a slash and a format  
letter. The format letters supported are:  
x
d
u
o
Regard the bits of the value as an integer, and print the integer in hexadecimal.  
Print as integer in signed decimal.  
Print as integer in unsigned decimal.  
Print as integer in octal.  
86  
Examining Data  
 
2
2
t
a
Print as integer in binary. The letter 't' stands for “two” .  
Print as an address, both absolute in hexadecimal and as an offset from the nearest  
preceding symbol. You can use this format used to discover where (in what function)  
an unknown address is located:  
((gdb)) p/a 0x54320  
$3 = 0x54320 <_initialize_vx+396>  
c
f
Regard as an integer and print it as a character constant.  
Regard the bits of the value as a floating point number and print using typical  
floating point syntax.  
For example, to print the program counter in hex (see “Registers” (page 98)), type  
p/x $pc  
Note that no space is required before the slash; this is because command names in GDB  
cannot contain a slash.  
To reprint the last value in the value history with a different format, you can use the  
printcommand with just a format and no expression. For example, 'p/x' reprints the  
last value in hex.  
8.5 Examining memory  
You can use the command x(for “examine”) to examine memory in any of several  
formats, independent of your program data types.  
x/nfu addr, Use the xcommand to examine memory.  
x addr, x  
[n],[ f], and [u] are all optional parameters that specify how much memory to display  
and how to format it; addr is an expression giving the address where you want to start  
displaying memory. If you use defaults for nfu, you need not type the slash '/'. Several  
commands set convenient defaults for addr.  
[n], the repeat count  
The repeat count is a decimal integer and the  
default is 1. It specifies how much memory  
(counting by units u) to display.  
[f], the display format  
The display format is one of the formats used by  
print, 's' (null-terminated string), or 'i'  
(machine instruction). The default is 'x'  
(hexadecimal) initially. The default changes each  
time you use either xor print.  
[u], the unit size  
The unit size is any of  
b
Bytes.  
h
Halfwords (two bytes).  
2. 'b' cannot be used because these format letters are also used with the x command, where 'b' stands for  
8.5 Examining memory  
87  
 
w
g
Words (four bytes). This is the initial default.  
Giant words (eight bytes).  
Each time you specify a unit size with x, that  
size becomes the default unit the next time  
you use x. (For the 's' and 'i' formats, the  
unit size is ignored and is normally not  
written.)  
addr, starting display address  
addr is the address where you want GDB to  
begin displaying memory. The expression need  
not have a pointer value (though it may); it is  
always interpreted as an integer address of a byte  
of memory. Refer to See “Expressions” (page 83),  
for more information on expressions. The default  
for addr is usually just after the last address  
examined―but several other commands also set  
the default address: info breakpoints(to the  
address of the last breakpoint listed), info line  
(to the starting address of a line), and print(if  
you use it to display a value from memory).  
For example, 'x/3uh 0x54320' is a request to display three halfwords (h) of memory,  
formatted as unsigned decimal integers ('u'), starting at address 0x54320. 'x/4xw $sp'  
prints the four words ('w') of memory above the stack pointer (here, '$sp'; see “Registers”  
(page 98)) in hexadecimal ('x').  
Since the letters indicating unit sizes are all distinct from the letters specifying output  
formats, you do not have to remember whether unit size or format comes first; either  
order works. The output specifications '4xw' and '4wx' mean exactly the same thing.  
(However, the count n must come first; 'wx4' does not work.)  
Even though the unit size u is ignored for the formats 's' and 'i', you might still want  
to use a count n; for example, '3i' specifies that you want to see three machine  
instructions, including any operands. The command disassemblegives an alternative  
way of inspecting machine instructions; see “Source and machine code” (page 80).  
All the defaults for the arguments to xare designed to make it easy to continue scanning  
memory with minimal specifications each time you use x. For example, after you have  
inspected three machine instructions with 'x/3i addr', you can inspect the next seven  
with just 'x/7'. If you use RET to repeat the xcommand, the repeat count n is used  
again; the other arguments default as for successive uses of x.  
The addresses and contents printed by the xcommand are not saved in the value  
history because there is often too much of them and they would get in the way. Instead,  
GDB makes these values available for subsequent use in expressions as values of the  
convenience variables $_and $__. After an xcommand, the last address examined is  
88  
Examining Data  
available for use in expressions in the convenience variable $_. The contents of that  
address, as examined, are available in the convenience variable $__.  
If the xcommand has a repeat count, the address and contents saved are from the last  
memory unit printed; this is not the same as the last address printed if several units  
were printed on the last line of output.  
8.6 Automatic display  
If you find that you want to print the value of an expression frequently (to see how it  
changes), you might want to add it to the automatic display list so that GDB prints its  
value each time your program stops. Each expression added to the list is given a number  
to identify it; to remove an expression from the list, you specify that number. The  
automatic display looks like this:  
2: foo = 38  
3: bar[5] = (struct hack *) 0x3804  
This display shows item numbers, expressions, and their current values. As with  
displays you request manually using xor print, you can specify the output format  
you prefer; in fact, displaydecides whether to use printor xdepending on how  
elaborate your format specification is―it uses xif you specify a unit size, or one of the  
two formats ('i' and 's') that are only supported by x; otherwise it uses print.  
display expr  
Add the expression expr to the list of expressions to display  
each time your program stops. See “Expressions” (page 83).  
displaydoes not repeat if you press RET again after using  
it.  
display/fmt expr  
For fmt specifying only a display format and not a size or  
count, add the expression expr to the auto-display list but  
arrange to display it each time in the specified format fmt.  
display/fmt addr  
For fmt 'i' or 's', or including a unit-size or a number of  
units, add the expression addr as a memory address to be  
examined each time your program stops. Examining means  
in effect doing 'x/fmt addr'. See “Examining memory”  
For example, `display/i $pc' can be helpful, to view the machine instruction about  
to be executed each time execution stops (`$pc' is a common name for the program  
counter; see “Registers” (page 98)).  
undisplay dnums..., delete  
display dnums...  
Remove item numbers dnums from the list of  
expressions to display. undisplaydoes not  
repeat if you press RET after using it. (Otherwise  
you would just get the error 'No display number  
...'.)  
8.6 Automatic display  
89  
 
disable display dnums...  
Disable the display of item numbers dnums. A  
disabled display item is not printed  
automatically, but is not forgotten. It may be  
enabled again later.  
enable display dnums...  
Enable display of item numbers dnums. It  
becomes effective once again in auto display of  
its expression, until you specify otherwise.  
display  
Display the current values of the expressions on  
the list, just as is done when your program stops.  
info display  
Print the list of expressions previously set up to  
display automatically, each one with its item  
number, but without showing the values. This  
includes disabled expressions, which are marked  
as such. It also includes expressions which would  
not be displayed right now because they refer to  
automatic variables not currently available.  
If a display expression refers to local variables, then it does not make sense outside the  
lexical context for which it was set up. Such an expression is disabled when execution  
enters a context where one of its variables is not defined. For example, if you give the  
command display last_charwhile inside a function with an argument last_char,  
GDB displays this argument while your program continues to stop inside that function.  
When it stops elsewhere, where there is no variable last_char, the display is disabled  
automatically. The next time your program stops where last_charis meaningful,  
you can enable the display expression again.  
8.7 Print settings  
GDB provides the following ways to control how arrays, structures, and symbols are  
printed.  
These settings are useful for debugging programs in any language:  
set print address, set  
print address on  
GDB prints memory addresses showing the location  
of stack traces, structure values, pointer values,  
breakpoints, and so forth, even when it also displays  
the contents of those addresses. The default is on. For  
example, this is what a stack frame display looks like  
with set print address on:  
((gdb)) f  
#0 set_quotes (lq=0x34c78 "<<", rq=0x34c88  
">>")  
at input.c:530  
530 if (lquote != def_lquote)  
90  
Examining Data  
 
set print address off  
Do not print addresses when displaying their  
contents. For example, this is the same stack frame  
displayed with set print address off:  
((gdb)) set print addr off  
((gdb)) f  
#0 set_quotes (lq="<<", rq=">>") at  
input.c:530  
530 if (lquote != def_lquote)  
You can use 'set print address off' to eliminate  
all machine dependent displays from the GDB  
interface. For example, with print address off,  
you should get the same text for backtraces on all  
machines―whether or not they involve pointer  
arguments.  
show print address  
Show whether or not addresses are to be printed.  
When GDB prints a symbolic address, it normally prints the closest previous symbol  
plus an offset. If that symbol does not uniquely identify the address (for example, it is  
a name whose scope is a single source file), you may need to clarify it. One way to do  
this is with info line. For example 'info line *0x4537'. Alternately, you can  
set GDB to print the source file and the line number when it prints a symbolic address:  
set print symbol-filename  
on  
Tell GDB to print the source file name and line  
number of a symbol in the symbolic form of an  
address.  
set print symbol-filename  
off  
Do not print source file name and line number  
of a symbol. This is the default.  
show print symbol-filename Show whether or not GDB will print the source  
file name and line number of a symbol in the  
symbolic form of an address.  
Another situation where it is helpful to show symbol filenames and line numbers is  
when disassembling code. GDB shows you the line number and source file that  
corresponds to each instruction.  
Also, you may wish to see the symbolic form only if the address being printed is  
reasonably close to the closest earlier symbol:  
set print  
max-symbolic-offset  
max-offset  
Tell GDB to only display the symbolic form of  
an address if the offset between the closest  
symbol and the address is less than  
max-offset. The default is 0, which tells GDB  
to always print the symbolic form of an address  
if any symbol precedes it.  
show print  
max-symbolic-offset  
Ask how large the maximum offset is that GDB  
prints in a symbolic address.  
8.7 Print settings  
91  
If you have a pointer and you are not sure where it points, try 'set print  
symbol-filename on'. Then you can determine the name and source file location  
of the variable where it points, using 'p/a pointer'. This interprets the address in  
symbolic form. For example, here GDB shows that a variable pttpoints at another  
variable t, defined in 'hi2.c':  
((gdb)) set print symbol-filename on  
((gdb)) p/a ptt  
$4 = 0xe008 <t in hi2.c>  
WARNING! For pointers that point to a local variable, 'p/a' does not show the symbol  
name and filename of the referent, even with the appropriate set printoptions  
turned on.  
Other settings to control how different kinds of objects are printed:  
set print array, set print  
array on  
Pretty print arrays. This format is more  
convenient to read, but uses more space. The  
default is off.  
set print array off  
show print array  
Return to compressed format for arrays.  
Show whether compressed or pretty format is  
selected for displaying arrays.  
set print elements  
number-of-elements  
Set a limit on how many elements of an array  
GDB will print. If GDB is printing a large array,  
it stops printing after it has printed the number  
of elements set by the set print elements  
command. This limit also applies to the display  
of strings. When GDB starts, this limit is set to  
200. Setting number-of-elements to zero  
means that the printing is unlimited.  
show print elements  
set print null-stop  
Display the number of elements of a large array  
that GDB will print. If the number is 0, then the  
printing is unlimited.  
Cause GDB to stop printing the characters of an  
array when the first NULLis encountered. This  
is useful when large arrays actually contain only  
short strings. The default is off.  
set print pretty on  
Cause GDB to print structures in an indented  
format with one member per line, like this:  
$1 = {  
next = 0x0,  
flags = {  
sweet = 1,  
sour = 1  
92  
Examining Data  
},  
meat = 0x54 "Pork"  
}
set print pretty off  
Cause GDB to print structures in a compact  
format, like this:  
$1 = {next = 0x0, flags = {sweet = 1,  
sour = 1}, \  
meat = 0x54 "Pork"}  
This is the default format.  
show print pretty  
Show which format GDB is using to print  
structures.  
set print sevenbit-strings Print using only seven-bit characters; if this  
option is set, GDB displays any eight-bit  
characters (in strings or character values) using  
the notation \nnn. This setting is best if you are  
working in English (ASCII) and you use the  
high-order bit of characters as a marker or “meta”  
bit.  
on  
set print sevenbit-strings Print full eight-bit characters. This allows the use  
of more international character sets, and is the  
default.  
off  
show print sevenbit-strings Show whether or not GDB is printing only  
seven-bit characters.  
set print union on  
set print union off  
show print union  
Tell GDB to print unions which are contained in  
structures. This is the default setting.  
Tell GDB not to print unions which are contained  
in structures.  
Ask GDB whether or not it will print unions  
which are contained in structures.  
For example, given the declarations  
typedef enum {Tree, Bug} Species;  
typedef enum {Big_tree, Acorn, Seedling}  
Tree_forms;  
typedef enum {Caterpillar, Cocoon,  
Butterfly}  
Bug_forms;  
struct thing {  
Species it;  
union {  
Tree_forms tree;  
Bug_forms bug;  
8.7 Print settings  
93  
} form;  
};  
struct thing foo = {Tree, {Acorn}};  
with set print union onin effect 'p foo'  
would print  
$1 = {it = Tree, form = {tree = Acorn,  
bug = Cocoon}}  
and with set print union offin effect it  
would print  
$1 = {it = Tree, form = {...}}  
These settings are of interest when debugging C++ programs:  
set print demangle, set  
print demangle on  
Print C++ names in their source form rather than  
in the encoded (“mangled”) form passed to the  
assembler and linker for type-safe linkage. The  
default is on.  
show print demangle  
Show whether C++ names are printed in mangled  
or demangled form.  
set print asm-demangle, set Print C++ names in their source form rather than  
their mangled form, even in assembler code  
printouts such as instruction disassemblies. The  
default is off.  
print asm-demangle on  
show print asm-demangle  
Show whether C++ names in assembly listings  
are printed in mangled or demangled form.  
set demangle-style style  
Choose among several encoding schemes used  
by different compilers to represent C++ names.  
On HP-UX, WDB automatically chooses the  
appropriate style.  
The choices for style currently supported are:  
auto  
gnu  
hp  
Allow GDB to choose a decoding style  
by inspecting your program.  
Decode based on the GNU C++  
compiler (g++) encoding algorithm.  
Decode based on the HP ANSI C++  
(aCC) encoding algorithm.  
lucid Decode based on the Lucid C++  
compiler (lcc) encoding algorithm.  
94  
Examining Data  
arm  
Decode using the algorithm in the C++  
Annotated Reference Manual.  
WARNING! This setting alone is not  
sufficient to allow debugging cfront  
generated executables. GDB would  
require further enhancement to permit  
that.  
If you omit style, you will see a list  
of possible formats.  
show demangle-style  
Display the encoding style currently in use for  
decoding C++ symbols.  
set print object, set print When displaying a pointer to an object, identify  
the actual (derived) type of the object rather than  
the declared type, using the virtual function  
table.  
object on  
set print object off  
show print object  
Display only the declared type of objects, without  
reference to the virtual function table. This is the  
default setting.  
Show whether actual or declared object types are  
displayed.  
set print static-members,  
set print static-members  
on  
Print static members when displaying a C++  
object. The default is on.  
set print static-members  
off  
Do not print static members when displaying a  
C++ object.  
show print static-members  
Show whether C++ static members are printed,  
or not.  
set print vtbl, set print  
vtbl on  
Pretty print C++ virtual function tables. The  
default is off. (The vtblcommands do not work  
on programs compiled with the HP ANSI C++  
compiler (aCC).)  
set print vtbl off  
show print vtbl  
Do not pretty print C++ virtual function tables.  
Show whether C++ virtual function tables are  
pretty printed, or not.  
8.8 Value history  
Values printed by the printcommand are saved in the GDB value history. This allows  
you to refer to them in other expressions. Values are kept until the symbol table is  
8.8 Value history  
95  
 
re-read or discarded (for example with the fileor symbol-filecommands). When  
the symbol table changes, the value history is discarded, since the values may contain  
pointers back to the types defined in the symbol table.  
The values printed are given history numbers by which you can refer to them. These are  
a range of integers starting with one. printshows you the history number assigned  
to a value by printing '$num = ' before the value; here num is the history number.  
To refer to any previous value, use '$' followed by the history number of the value. The  
way printlabels its output is designed to remind you of this. Just $ refers to the most  
recent value in the history, and $$ refers to the value before that. $$n refers to the nth  
value from the end; $$2 is the value just prior to $$, $$1 is equivalent to $$, and $$0 is  
equivalent to $.  
For example, suppose you have just printed a pointer to a structure and want to see  
the contents of the structure. It suffices to type  
p *$  
If you have a chain of structures where the component nextpoints to the next one,  
you can print the contents of the next one with this:  
p *$.next  
You can print successive links in the chain by repeating this command using the RET  
key.  
Note that the history records values, not expressions. If the value of xis 4 and you type  
these commands:  
print x  
set x=5  
then the value recorded in the value history by the printcommand remains 4 even  
though the value of xhas changed.  
show values  
Print the last ten values in the value history, with their item  
numbers. This is like 'p $$9' repeated ten times, except that show  
valuesdoes not change the history.  
show values n  
Print ten history values centered on history item number n.  
show values +  
Print ten history values following the values last printed. If no  
more values are available, show values +produces no display.  
Pressing RET to repeat show values n has exactly the same effect as 'show values  
+'.  
8.9 Convenience variables  
GDB provides convenience variables that you can use within GDB to hold on to a value  
and refer to it later. These variables exist entirely within GDB. They are not part of your  
program, and setting a convenience variable has no direct effect on further execution  
of your program. That is why you can use them freely.  
96  
Examining Data  
 
Convenience variables are prefixed with '$'. Any name preceded by '$' can be used for  
a convenience variable, unless it is one of the predefined machine-specific register  
names (see “Registers” (page 98)). (Value history references, in contrast, are numbers  
preceded by '$'. See “Value history” (page 95).)  
You can save a value in a convenience variable with an assignment expression, just as  
you would set a variable in your program. For example:  
set $foo = *object_ptr  
would save in $foothe value contained in the object pointed to by object_ptr.  
Using a convenience variable for the first time creates it, but its value is voiduntil you  
assign a new value. You can alter the value with another assignment at any time.  
Convenience variables have no fixed types. You can assign a convenience variable any  
type of value, including structures and arrays, even if that variable already has a value  
of a different type. The convenience variable, when used as an expression, has the type  
of its current value.  
show convenience  
Print a list of convenience variables used so far, and their  
values. Abbreviated show conv.  
A convenient variable can be used as a counter to be incremented or a pointer to be  
advanced. For example, to print a field from successive elements of an array of  
structures:  
set $i = 0  
print bar[$i++]->contents  
Repeat that command by typing RET.  
Some convenience variables are created automatically by GDB and assigned values.  
$_  
The variable $_is automatically set by the xcommand to the last  
address examined (see “Examining memory” (page 87)). Other  
commands which provide a default address for xto examine, also set  
$_to that address. These commands include info lineand info  
breakpoint. The type of $_is void *except when set by the x  
command, in which case it is a pointer to the type of $__.  
$__  
The variable $__is automatically set by the xcommand to the value  
found in the last address examined. Its type is chosen to match the  
format in which the data was printed.  
$_exitcode  
The variable $_exitcodeis automatically set to the exit code when  
the program being debugged terminates.  
On HP-UX systems, if you refer to a function or variable name that begins with a dollar  
sign, GDB searches for a user or system name first, before it searches for a convenience  
variable.  
8.9 Convenience variables  
97  
8.10 Registers  
You can refer to machine register contents, in expressions, as variables with names  
starting with '$'. The names of registers are different for each machine. Use info  
registersto view the names used on your machine.  
info registers  
Print the names and values of all registers except  
floating-point registers (in the selected stack  
frame).  
info all-registers  
Print the names and values of all registers,  
including floating-point registers.  
info registers regname ... Print the relativized value of each specified  
register regname. As discussed in detail below,  
register values are normally relative to the  
selected stack frame. regname may be any  
register name valid on the machine you are  
using, with or without the initial '$'.  
GDB has four standard register names that are available (in expressions) on most  
machines―whenever they do not conflict with an architecture's canonical mnemonics  
for registers. The register names $pcand $spare used for the program counter register  
and the stack pointer. $fpis used for a register that contains a pointer to the current  
stack frame, and $psis used for a register that contains the processor status. For  
example, you could print the program counter in hex with  
p/x $pc  
or print the instruction to be executed next with  
x/i $pc  
3
3
or add four to the stack pointer with with  
set $sp += 4  
Whenever possible, these four standard register names are available on your machine  
even though the machine has different canonical mnemonics, so long as there is no  
conflict. The info registerscommand shows the canonical names. For example,  
on the SPARC, info registersdisplays the processor status register as $psrbut  
you can also refer to it as $ps; and on x86-based machines $psis an alias for the EFLAGS  
register.  
GDB always considers the contents of an ordinary register as an integer when the  
register is examined in this way. Some machines have special registers which can hold  
nothing but floating point; these registers are considered to have floating point values.  
There is no way to refer to the contents of an ordinary register as floating point value  
(although you can print it as a floating point value with 'print/f $regname').  
3. This is a way of removing one word from the stack, on machines where stacks grow downward in memory  
(most machines, nowadays). This assumes that the innermost stack frame is selected; setting $spis not  
allowed when other stack frames are selected. To pop entire frames off the stack, regardless of machine  
architecture, use return; see “Returning from a function” (page 121).  
98  
Examining Data  
 
Some registers have distinct raw and virtual data formats. This means that the data  
format in which the register contents are saved by the operating system is not the same  
one that your program normally sees. For example, the registers of the 68881 floating  
point coprocessor are always saved in “extended” (raw) format, but all C programs  
expect to work with “double” (virtual) format. In such cases, GDB normally works  
with the virtual format only (the format that makes sense for your program), but the  
info registerscommand prints the data in both formats.  
Normally, register values are relative to the selected stack frame (see “Selecting a frame”  
(page 73)). This means that you get the value that the register would contain if all stack  
frames farther in were exited and their saved registers restored. In order to see the true  
contents of hardware registers, you must select the innermost frame (with 'frame 0').  
However, GDB must deduce where registers are saved, from the machine code generated  
by your compiler. If some registers are not saved, or if GDB is unable to locate the saved  
registers, the selected stack frame makes no difference.  
8.11 Printing Floating Point Values  
You can print the values of floating-point registers in different formats.  
To print both single and double-precision values:  
(gdb) info reg $fr5  
fr5 (single precision) 10.1444092  
fr5  
To get the bit pattern, try the following macro:  
define pbits  
set *((float *) $sp)=$arg0  
p/x *((int *) $sp)  
end  
This is what the macro produces:  
(gdb) pbits $fr6  
$1 = 0x4082852d  
8.12 Floating point hardware  
Depending on the configuration, GDB may be able to give you more information about  
the status of the floating point hardware.  
info float  
Display hardware-dependent information about the floating point  
unit. The exact contents and layout vary depending on the floating  
point chip. Currently, 'info float' is supported on the ARM and  
x86 machines.  
8.11 Printing Floating Point Values  
99  
   
100  
9 Using GDB with Different Languages  
Although programming languages generally have common aspects, they are rarely  
expressed in the same manner. For instance, in ANSI C, dereferencing a pointer pis  
accomplished by *p, but in Modula-2, it is accomplished by p^. Values can also be  
represented (and displayed) differently. Hex numbers in C appear as '0x1ae', while  
in Modula-2 they appear as '1AEH'.  
Language-specific information is built into GDB for some languages, allowing you to  
express operations like the above in the native language of your program, and allowing  
GDB to output values in a manner consistent with the syntax of the native language.  
The language you use to build expressions is called the working language.  
9.1 Switching between source languages  
There are two ways to control the working language. You can have GDB set it  
automatically, or you can select it manually. You can use the set languagecommand  
for either purpose. On startup, GDB sets the default language automatically. The  
working language is used to determine how expressions are interpreted, how values  
are printed, and so on.  
In addition to the working language, every source file that GDB knows about has its  
own working language. For some object file formats, the compiler might indicate which  
language a particular source file is in. However, most of the time GDB infers the  
language from the name of the file. The language of a source file controls whether C++  
names are demangled―this way backtracecan show each frame appropriately for  
its own language. There is no way to set the language of a source file from within GDB,  
but you can set the language associated with a filename extension. See “Displaying the  
This is a common problem when you use a program, such as cfrontor f2c, that  
generates C but is written in another language. In that case, make the program use  
#linedirectives in its C output; that way GDB will know the correct language of the  
source code of the original program, and will display that source code, not the generated  
C code.  
9.1.1 List of filename extensions and languages  
If a source file name ends in one of the following extensions, then GDB infers that its  
language is the one indicated.  
'.c'  
C source file  
'.C',  
C++ source file  
'.cc',  
'.cp',  
'.cpp',  
9.1 Switching between source languages 101  
     
'.cxx',  
'.c++'  
'.f', '.F', Fortran source file. GDB does not distinguish between Fortran 77 and Fortran  
90 files.  
'.f90'  
'.s', '.S' Assembler source file. This actually behaves almost like C, but GDB does  
not skip over function prologues when stepping.  
In addition, you may set the language associated with a filename extension. See  
9.1.2 Setting the working language  
If you allow GDB to set the language automatically, expressions are interpreted the  
same way in your debugging session and your program.  
If you wish, you may set the language manually. To do this, issue the command 'set  
language lang', where langis the name of a language, such as c. For a list of the  
supported languages, type 'set language'.  
Setting the language manually prevents GDB from updating the working language  
automatically. This can lead to confusion if you try to debug a program when the  
working language is not the same as the source language, when an expression is  
acceptable to both languages―but means different things. For instance, if the current  
source file was written in C, and GDB was parsing Modula-2, a command such as:  
print a = b + c  
might not have the effect you intended. In C, this means to add band cand place the  
result in a. The result printed would be the value of a. In Modula-2, this means to  
compare ato the result of b+c, yielding a BOOLEAN value.  
9.1.3 Having GDB infer the source language  
To have GDB set the working language automatically, use 'set language local'  
or 'set language auto'. GDB then infers the working language. That is, when your  
program stops in a frame (usually by encountering a breakpoint), GDB sets the working  
language to the language recorded for the function in that frame. If the language for a  
frame is unknown (that is, if the function or block corresponding to the frame was  
defined in a source file that does not have a recognized extension), the current working  
language is not changed, and GDB issues a warning.  
This may not seem necessary for most programs, which are written entirely in one  
source language. However, program modules and libraries written in one source  
language can be used by a main program written in a different source language. Using  
'set language auto' in this case frees you from having to set the working language  
manually.  
102 Using GDB with Different Languages  
   
9.2 Displaying the language  
The following commands help you find out which language is the working language,  
and also what language source files were written in.  
show language  
info frame  
Display the current working language. This is the language you  
can use with commands such as printto build and compute  
expressions that may involve variables in your program.  
Display the source language for this frame. This language becomes  
the working language if you use an identifier from this frame.  
See “Information about a frame” (page 74), to identify the other  
information listed here.  
info source  
Display the source language of this source file. Refer to See  
Chapter 10 (page 115), to identify the other information listed here.  
In unusual circumstances, you may have source files with extensions not in the standard  
list. You can then set the extension associated with a language explicitly:  
set extension-language .ext Set source files with extension .extto be  
language  
assumed to be in the source language language.  
However, this is not valid on Unix systems.  
info extensions  
List all the filename extensions and the associated  
languages. Not valid on Unix systems.  
9.3 Type and range checking  
Some languages are designed to guard you against making seemingly common errors  
through a series of compile and run-time checks. These include checking the type of  
arguments to functions and operators, and making sure mathematical overflows are  
caught at run time. Checks such as these help to ensure the correctness of the program  
once it has been compiled by eliminating type mismatches, and providing active checks  
for range errors when your program is running.  
GDB can check for conditions like the above if you wish. Although GDB does not check  
the statements in your program, it can check expressions entered directly into GDB for  
evaluation via the printcommand, for example. As with the working language, GDB  
can also decide whether or not to check automatically based on your source language.  
See “Supported languages” (page 105), for the default settings of supported languages.  
9.3.1 An overview of type checking  
Some languages are strongly typed, meaning that the arguments to operators and  
functions have to be of the correct type, otherwise an error occurs. These checks prevent  
type mismatch errors from causing run-time problems. For example,  
1 + 2 3  
but  
9.2 Displaying the language 103  
     
error 1 + 2.3  
The second example fails because the CARDINAL1 is not type-compatible with the  
REAL2.3.  
For the expressions you use in GDB commands, you can tell the GDB type checker to  
skip checking; to treat any mismatches as errors and abandon the expression; or to only  
issue warnings when type mismatches occur, and evaluate the expression anyway.  
When you choose the last of these, GDB evaluates expressions like the second example  
above, but also issues a warning.  
Even if you turn type checking off, there may be other reasons related to type that  
prevent GDB from evaluating an expression. For instance, GDB does not know how  
to add an intand a struct foo. These particular type errors have nothing to do  
with the language in use, and usually arise from expressions, such as the one described  
above, which make little sense to evaluate anyway.  
Each language defines to what degree it is strict about type. For instance C requires  
the arguments to arithmetical operators to be numbers. In C, enumerated types and  
pointers can be represented as numbers, so that they are valid arguments to  
mathematical operators. See “Supported languages” (page 105), for further details on  
specific languages.  
GDB provides some additional commands for controlling the type checker:  
set check type auto  
Set type checking on or off based on the current working  
language. See “Supported languages” (page 105), for the  
default settings for each language.  
set check type on, set Set type checking on or off, overriding the default setting  
for the current working language. Issue a warning if the  
setting does not match the language default. If any type  
mismatches occur in evaluating an expression while type  
checking is on, GDB prints a message and aborts  
evaluation of the expression.  
check type off  
set check type warn  
Cause the type checker to issue warnings, but to always  
attempt to evaluate the expression. Evaluating the  
expression may still be impossible for other reasons. For  
example, GDB cannot add numbers and structures.  
show type  
Show the current setting of the type checker, and  
whether or not GDB is setting it automatically.  
9.3.2 An overview of range checking  
In some languages it is an error to exceed the bounds of a type; this is enforced with  
run-time checks. Such range checking is meant to ensure program correctness by making  
sure computations do not overflow, or indices on an array element access do not exceed  
the bounds of the array.  
104 Using GDB with Different Languages  
 
For expressions you use in GDB commands, you can tell GDB to treat range errors in  
one of three ways: ignore them, always treat them as errors and abandon the expression,  
or issue warnings but evaluate the expression anyway.  
A range error can result from numerical overflow, from exceeding an array index  
bound, or when you type a constant that is not a member of any type. Some languages,  
however, do not treat overflows as an error. In many implementations of C,  
mathematical overflow causes the result to “wrap around” to lower values―for example,  
if mis the largest integer value, and sis the smallest, then  
m + 1 s  
This, too, is specific to individual languages, and in some cases specific to individual  
compilers or machines. Refer to See “Supported languages” (page 105), for further  
details on specific languages.  
GDB provides the following additional commands for controlling the range checker:  
set check range auto  
Set range checking on or off based on the current  
working language. See “Supported languages”  
(page 105), for the default settings for each language.  
set check range on, set Set range checking on or off, overriding the default  
setting for the current working language. A warning  
is issued if the setting does not match the default  
language. If a range error occurs and range checking  
is on, then a message is printed and evaluation of the  
expression is aborted.  
check range off  
set check range warn  
Output messages when the GDB range checker detects  
a range error, but attempt to evaluate the expression  
anyway. Evaluating the expression may still be  
impossible for other reasons, such as accessing memory  
that the process does not own (a typical example from  
many Unix systems).  
show range  
Show the current setting of the range checker, and  
whether or not it is being set automatically by GDB.  
9.4 Supported languages  
GDB supports C, C++, and Fortran. Refer to for specific information about Fortran.  
Some GDB features may be used in expressions regardless of the language you use:  
the GDB @and ::operators, and the '{type}addr' construct (see “Expressions”  
(page 83)) can be used with the constructs of any supported language.  
The following section discusses GDB support for each source language. These sections  
are not meant to be language tutorials or references, but serve only as a reference guide  
to what the GDB expression parser accepts, and what input and output formats should  
look like for different languages.  
9.4 Supported languages 105  
 
9.4.1 C and C++  
Since C and C++ are so closely related, many features of GDB apply to both languages.  
Whenever this is the case, we discuss those languages together.  
The C++ debugging facilities are jointly implemented by the C++ compiler and GDB.  
Therefore, to debug your C++ code effectively, you must compile your C++ programs  
with a supported C++ compiler, such as GNU g++, or the HP ANSI C++ compiler (aCC).  
For best results when using GNU C++, use the stabs debugging format. You can select  
that format explicitly with the g++ command-line options '-gstabs' or '-gstabs+'.  
Refer to section “Options for Debugging Your Program or GNU CC” in Using GNU  
CC, for more information.  
9.4.1.1 C and C++ operators  
Operators must be defined on values of specific types. For instance, + is defined on  
numbers, but not on structures. Operators are often defined on groups of types.  
For the purposes of C and C++, the following definitions hold:  
Integral types include intwith any of its storage-class specifiers; char; enum; and,  
for C++, bool.  
Floating-point types include float, double, and long double(if supported by  
the target platform).  
Pointer types include all types defined as (type *).  
Scalar types include all of the above.  
The following operators are supported. They are listed here in order of increasing  
precedence:  
,
The comma or sequencing operator. Expressions in a  
comma-separated list are evaluated from left to right, with the  
result of the entire expression being the last expression evaluated.  
=
Assignment. The value of an assignment expression is the value  
assigned. Defined on scalar types.  
op=  
Used in an expression of the form a op= b, and translated to a = a  
op b. op= and = have the same precedence. op is any one of the  
operators |, ^, &, <<, >>, +, -, *, /, %.  
?:  
The ternary operator. a ? b : c can be thought of as: if a then b else  
c. a should be of an integral type.  
||  
&&  
|
Logical OR. Defined on integral types.  
Logical AND. Defined on integral types.  
Bitwise OR. Defined on integral types.  
Bitwise exclusive-OR. Defined on integral types.  
Bitwise AND. Defined on integral types.  
^
&
106 Using GDB with Different Languages  
   
==, !=  
Equality and inequality. Defined on scalar types. The value of these  
expressions is 0 for false and non-zero for true.  
<, >, <=, >=  
Less than, greater than, less than or equal, greater than or equal.  
Defined on scalar types. The value of these expressions is 0 for false  
and non-zero for true.  
<<, >>  
@
left shift, and right shift. Defined on integral types.  
The GDB “artificial array” operator (see “Expressions” (page 83)).  
+, -  
Addition and subtraction. Defined on integral types, floating-point  
types and pointer types.  
*, /, %  
++, --  
Multiplication, division, and modulus. Multiplication and division  
are defined on integral and floating-point types. Modulus is defined  
on integral types.  
Increment and decrement. When appearing before a variable, the  
operation is performed before the variable is used in an expression;  
when appearing after it, the value of the variable is used before  
the operation takes place.  
*
&
Pointer dereferencing. Defined on pointer types. Same precedence  
as ++.  
Address operator. Defined on variables. Same precedence as ++.  
For debugging C++, GDB implements a use of '&' beyond what is  
allowed in the C++ language itself: you can use '&(&ref)' (or, if  
you prefer, simply '&&ref') to examine the address where a C++  
reference variable (declared with '&ref') is stored.  
-
Negative. Defined on integral and floating-point types. Same  
precedence as ++.  
!
Logical negation. Defined on integral types. Same precedence as  
++.  
~
Bitwise complement operator. Defined on integral types. Same  
precedence as ++.  
., ->  
Structure member, and pointer-to-structure member. For  
convenience, GDB regards the two as equivalent, choosing whether  
to dereference a pointer based on the stored type information.  
Defined on structand uniondata.  
.*, ->*  
[]  
Dereferences pointers to members.  
Array indexing. a[i]is defined as *(a+i). Same precedence as  
->.  
()  
Function parameter list. Same precedence as ->.  
9.4 Supported languages 107  
::  
::  
C++ scope resolution operator. Defined on struct, union, and  
classtypes.  
Double colons also represent the GDB scope operator (see  
“Expressions” (page 83)). Same precedence as ::, above.  
If an operator is redefined in the user code, GDB usually attempts to invoke the  
redefined version instead of using the original meaning.  
9.4.1.2 C and C++ constants  
GDB allows you to express the constants of C and C++ in the following ways:  
Integer constants are a sequence of digits. Octal constants are specified by a leading  
'0' (that is zero), and hexadecimal constants by a leading '0x' or '0X'. Constants can  
also end with a letter 'l', specifying that the constant should be treated as a long  
value.  
Floating point constants are a sequence of digits, followed by a decimal point,  
followed by a sequence of digits, and optionally followed by an exponent. An  
exponent is of the form: 'e[[+]|-]nnn', where nnnis a sequence of digits. The  
'+' is optional for positive exponents. A floating-point constant may also end with  
a letter 'f' or 'F', specifying that the constant should be treated as being of the  
float(as opposed to the default double) type; or with a letter 'l' or `L', which  
specifies a long doubleconstant.  
Enumerated constants consist of enumerated identifiers, or their integral  
equivalents.  
Character constants are a single character surrounded by single quotes ('), or a  
number or the ordinal value of the corresponding character (usually its ASCII  
value). Within quotes, the single character may be represented by a letter or by  
escape sequences, which are of the form '\nnn', where nnnis the octal representation  
of the character's ordinal value; or of the form '\x', where 'x' is a predefined special  
character―for example, '\n' for newline.  
String constants are a sequence of character constants surrounded by double quotes  
("). Any valid character constant (as described above) may appear. Double quotes  
within the string must be preceded by a backslash, so for instance '"a\"b'c"' is  
a string of five characters.  
Pointer constants are an integral value. You can also write pointers to constants  
using the C operator '&'.  
Array constants are comma-separated lists surrounded by braces '{' and '}'; for  
example, '{1,2,3}' is a three-element array of integers, '{{1,2}, {3,4},  
{5,6}}' is a three-by-two array, and '{&"hi", &"there", &"fred"}' is a  
three-element array of pointers.  
108 Using GDB with Different Languages  
 
9.4.1.3 C++ expressions  
GDB expression handling can interpret most C++ expressions.  
WARNING! GDB can only debug C++ code if you use the proper compiler. Typically,  
C++ debugging depends on the use of additional debugging information in the symbol  
table, and thus requires special support. In particular, if your compiler generates a.out,  
MIPS ECOFF, RS/6000 XCOFF, or ELF with stabs extensions to the symbol table, these  
facilities are all available. (With GNU CC, you can use the '-gstabs' option to request  
stabs debugging extensions explicitly.) Where the object code format is standard COFF  
or DWARF in ELF, on the other hand, most of the C++ support in GDB does not work.  
1. Member function calls are allowed; you can use expressions like  
count = aml->GetOriginal(x, y)  
2. While a member function is active (in the selected stack frame), your expressions  
have the same namespace available as the member function; that is, GDB allows  
implicit references to the class instance pointer thisfollowing the same rules as  
C++.  
3. You can call overloaded functions; GDB resolves the function call to the right  
definition, with some restrictions. GDB does not perform overload resolution  
involving user-defined type conversions, calls to constructors, or instantiations of  
templates that do not exist in the program. It also cannot handle ellipsis argument  
lists or default arguments.  
It does perform integral conversions and promotions, floating-point promotions,  
arithmetic conversions, pointer conversions, conversions of class objects to base  
classes, and standard conversions such as those of functions or arrays to pointers;  
it requires an exact match on the number of function arguments.  
Overload resolution is always performed, unless you have specified set  
overload-resolution off. See “GDB features for C++” (page 111).  
You must specify set overload-resolution offin order to use an explicit  
function signature to call an overloaded function, as in  
p 'foo(char,int)'('x', 13)  
The GDB command-completion facility can simplify this. Refer to “Command  
4. GDB understands variables declared as C++ references; you can use them in  
expressions just as you do in C++ source―they are automatically dereferenced.  
In the parameter list shown when GDB displays a frame, the values of reference  
variables are not displayed (unlike other variables); this avoids clutter, since  
9.4 Supported languages 109  
 
references are often used for large structures. The address of a reference variable  
is always shown, unless you have specified 'set print address off'.  
5. GDB supports the C++ name resolution operator ::―your expressions can use it  
just as expressions in your program do. Since one scope may be defined in another,  
you can use ::repeatedly if necessary, for example in an expression like  
'scope1::scope2::name'. GDB also allows resolving name scope by reference  
to source files, in both C and C++ debugging (see “Program variables” (page 84)).  
In addition, when used with the HP aC++ compiler, GDB supports calling virtual  
functions correctly, printing out virtual bases of objects, calling functions in a base  
subobject, casting objects, and invoking user-defined operators.  
NOTE: GDB cannot display debugging information for classes or functions defined  
in a shared library that is not compiled for debugging (with the -g0 option). GDB  
displays the function with the message <no data fields>.  
For example, after 'd3' is created by the following line:  
`RWCollectableDate d3(15,5,2001);'  
printing the variable or class returns:  
(gdb) p d3  
$3 = {<No data fields>}  
(gdb) ptype RWCollectableDate  
type = class RWCollectableDate {  
<no data fields>  
9.4.1.4 C and C++ defaults  
If you allow GDB to set type and range checking automatically, they both default to  
offwhenever the working language changes to C or C++. This happens regardless of  
whether you or GDB selects the working language.  
If you allow GDB to set the language automatically, it recognizes source files whose  
names end with '.c', '.C', or '.cc', and so on, and when GDB enters code compiled  
from one of these files, it sets the working language to C or C++. Refer to See “Having  
9.4.1.5 C and C++ type and range checks  
By default, when GDB parses C or C++ expressions, type checking is not used. However,  
if you turn type checking on, GDB considers two variable types equivalent if:  
The two variables are structured and have the same structure, union, or enumerated  
tag.  
The two variables have the same type name, or types that have been declared  
equivalent through typedef.  
110 Using GDB with Different Languages  
   
Range checking, if turned on, is done on mathematical operations. Array indices are  
not checked, since they are often used to index a pointer that is not itself an array.  
9.4.1.6 GDB and C  
The set print unionand show print unioncommands apply to the union  
type. When set to 'on', any unionthat is inside a structor classis also printed.  
Otherwise, it appears as '{...}'.  
The @operator aids in the debugging of dynamic arrays, formed with pointers and a  
memory allocation function. See “Expressions” (page 83).  
9.4.1.7 GDB features for C++  
Some GDB commands are particularly useful with C++, and some are designed  
specifically for use with C++. Here is a summary:  
breakpoint menus  
When you want a breakpoint in a function whose  
name is overloaded, GDB breakpoint menus help  
you specify which function definition you want.  
rbreak regex  
Setting breakpoints using regular expressions is  
helpful for setting breakpoints on overloaded  
functions that are not members of any special  
catch throw, catch catch  
Debug C++ exception handling using these  
ptype typename  
Print inheritance relationships as well as other  
information for type. typename. See Chapter 10  
set print demangle, show  
print demangle, set print  
asm-demangle, show print  
asm-demangle  
Control whether C++ symbols display in their  
source form, both when displaying code as C++  
source and when displaying disassemblies. See  
set print object, show print Choose whether to print derived (actual) or  
declared types of objects. See “Print settings”  
object  
set print vtbl, show print  
vtbl  
Control the format for printing virtual function  
tables. See “Print settings” (page 90). (The vtbl  
commands do not work on programs compiled  
with the HP ANSI C++ compiler (aCC).)  
set overload-resolution on Enable overload resolution for C++ expression  
evaluation. The default is on. For overloaded  
functions, GDB evaluates the arguments and  
9.4 Supported languages 111  
   
searches for a function whose signature matches  
the argument types, using the standard C++  
conversion rules (see “C++ expressions”  
(page 109), for details). If it cannot find a match,  
it emits a message.  
set overload-resolution off Disable overload resolution for C++ expression  
evaluation. For overloaded functions that are not  
class member functions, GDB chooses the first  
function of the specified name that it finds in the  
symbol table, whether or not its arguments are  
of the correct type. For overloaded functions that  
are class member functions, GDB searches for a  
function whose signature exactly matches the  
argument types.  
show overload-resolution  
Display current overload resolution setting for  
C++ expression evaluation.  
Overloaded symbolnames  
You can specify a particular definition of an  
overloaded symbol, using the same notation that  
is used to declare such symbols in C++: type  
symbol(types)rather than just symbol. You  
can also use the GDB command-line word  
completion facilities to list the available choices,  
or to finish the type list for you. See “Command  
completion” (page 33), for details on how to do  
this.  
9.4.2 Fortran  
You can use WDB to debug programs written in Fortran. WDB does not distinguish  
between Fortran 77 and Fortran 90 files.  
WDB provides the following command to control case sensitivity:  
case-sensitive [on | off]  
The default for Fortran is off while for other  
languages the default is on.  
Other supported features are:  
Fortran 90 pointers  
Structures and unions  
Calling functions with integer, logical, real, complex arguments  
Intrinsic support  
9.4.2.1 Fortran types  
Fortran types supported:  
112 Using GDB with Different Languages  
   
integer*1, integer*2,  
integer*4, integer*8  
logical*1, logical*2,  
logical*4, logical*8 byte,  
real*4, real*8, real*16  
complex*8, complex*16  
character*len,  
allocatable  
assumed-size  
assumed-shape  
adjustable  
automatic  
explicit-shape  
Array elements are displayed in column-major  
order. Use () for array member access (for  
example, arr(i) instead of arr[i]). Use set print  
elementsto control the number of elements  
printed out when specifying a whole array. The  
default is 200 elements or the number of elements  
of the array, whichever is smaller.  
character*(*) [len is a  
user supplied length]  
arrays  
9.4.2.2 Fortran operators  
The following Fortran operators are listed here in the order of increasing precedence:  
=
Assignment  
*, -, *, /  
+, -  
Binary operators  
Unary operators  
Exponentiation  
Equal  
Not equal, or concatenation  
Less than  
Less than or equal to  
Greater than  
Greater than or equal to  
Concatenation  
**  
.EQ., =  
.NE., /=  
.LT., <  
.LE., <=  
.GT., >  
.GE., >=  
//  
.NOT.  
Logical negation  
Logical AND  
.AND.  
.OR.  
Logical OR  
.EQV.  
Logical equivalence  
Logical non-equivalence  
.NEQV., .XOR.  
Logical constants are represented as .TRUE. or .FALSE.  
GDB includes support for viewing Fortran common blocks.  
info common  
Lists common blocks visible in the current frame.  
9.4 Supported languages 113  
 
info common  
<common_block_name>  
Lists values of variables in the named common  
block.  
Fortran entry points are supported.  
You can set a break point specifying an entry point name.  
9.4.2.3 Fortran special issues  
Fortran allows main to be a non-main procedure; therefore, to set a breakpoint in the  
main program, use break _MAIN_or break <program_name>.  
Do not use break mainunless it is the name of a non-main procedure.  
114 Using GDB with Different Languages  
 
10 Examining the Symbol Table  
The commands described in this chapter allow you to inquire about the symbols (names  
of variables, functions, and types) defined in your program. This information is inherent  
in the text of your program and does not change as your program executes. GDB finds  
it in your program's symbol table, in the file indicated when you started GDB (see  
“Choosing files” (page 26)), or by one of the file-management commands (see  
Occasionally, you may need to refer to symbols that contain unusual characters, which  
GDB ordinarily treats as word delimiters. The most frequent case is in referring to static  
variables in other source files (see “Program variables” (page 84)). File names are  
recorded in object files as debugging symbols, but GDB would ordinarily parse a typical  
file name, like 'foo.c', as the three words 'foo' '.' 'c'. To allow GDB to recognize  
'foo.c' as a single symbol, enclose it in single quotes; for example,  
p 'foo.c'::x  
looks up the value of xin the scope of the file 'foo.c'.  
info address symbol  
Describe where the data for symbolis stored.  
For a register variable, this says which register  
it is kept in. For a non-register local variable, this  
prints the stack-frame offset at which the variable  
is always stored.  
Note the contrast with 'print &symbol', which  
does not work at all for a register variable, and  
for a stack local variable prints the exact address  
of the current instantiation of the variable.  
whatis expr  
Print the data type of expression expr. expris  
not actually evaluated, and any side-effecting  
operations (such as assignments or function calls)  
inside it do not take place. See “Expressions”  
whatis  
Print the data type of $, the last value in the value  
history.  
ptype typename  
Print a description of data type typename.  
typenamemay be the name of a type, or for C  
code it may have the form 'class class-name',  
'struct struct-tag', 'union union-tag'  
or 'enum enum-tag'.  
ptype expr, ptype  
Print a description of the type of expression  
expr. ptypediffers from whatisby printing a  
115  
 
detailed description, instead of just the name of  
the type.  
For example, for this variable declaration:  
struct complex {double real; double  
imag;} v;  
the two commands give this output:  
((gdb)) whatis v  
type = struct complex  
((gdb)) ptype v  
type = struct complex {  
double real;  
double imag;  
}
As with whatis, using ptypewithout an  
argument refers to the type of $, the last value  
in the value history.  
info types regexp, info  
types  
Print a brief description of all types whose names  
match regexp(or all types in your program, if  
you supply no argument). Each complete  
typename is matched as though it were a  
complete line; thus, 'i type value' gives  
information on all types in your program whose  
names include the string value, but 'i type  
^value$' gives information only on types whose  
complete name is value.  
This command differs from ptypein two ways:  
first, like whatis, it does not print a detailed  
description; second, it lists all source files where  
a type is defined.  
info source  
info sources  
Show the name of the current source file―that  
is, the source file for the function containing the  
current point of execution―and the language it  
was written in.  
Print the names of all source files in your  
program for which there is debugging  
information, organized into two lists: files whose  
symbols have already been read, and files whose  
symbols will be read when needed.  
info functions  
Print the names and data types of all defined  
functions.  
116 Examining the Symbol Table  
info functions regexp  
Print the names and data types of all defined  
functions whose names contain a match for  
regular expression regexp. Thus, 'info fun  
step' finds all functions whose names include  
step; 'info fun ^step' finds those whose  
names start with step.  
info variables  
Print the names and data types of all variables  
that are declared outside of functions (that is,  
excluding local variables).  
info variables regexp  
Print the names and data types of all variables  
(except for local variables) whose names contain  
a match for regular expression regexp.  
Some systems allow individual object files that  
make up your program to be replaced without  
stopping and restarting your program. For  
example, in VxWorks you can simply recompile  
a defective object file and keep on running. If you  
are running on one of these systems, you can  
allow GDB to reload the symbols for  
automatically relinked modules:  
set  
Replace symbol definitions  
for the corresponding source  
file when an object file with  
a particular name is seen  
again.  
symbol-reloading  
on  
set  
Do not replace symbol  
definitions when  
symbol-reloading  
off  
encountering object files of  
the same name more than  
once. This is the default  
state; if you are not running  
on a system that permits  
automatic relinking of  
modules, you should leave  
symbol-reloading off,  
since otherwise GDB may  
discard symbols when  
linking large programs, that  
may contain several  
modules (from different  
directories or libraries) with  
the same name.  
117  
show  
Show the current onor off  
symbol-reloading setting.  
set opaque-type-resolution Tell GDB to resolve opaque types. An opaque  
on  
type is a type declared as a pointer to a struct,  
class, or union―for example, struct  
MyType *―that is used in one source file  
although the full declaration of struct MyType  
is in another source file. The default is on.  
A change in the setting of this subcommand will  
not take effect until the next time symbols for a  
file are loaded.  
set opaque-type-resolution Tell GDB not to resolve opaque types. In this  
case, the type is printed as follows:  
off  
{<no data fields>}  
show opaque-type-resolution Show whether opaque types are resolved or not.  
maint print symbols  
Write a dump of debugging symbol data into the  
file filename. These commands are used to  
debug the GDB symbol-reading code. Only  
symbols with debugging data are included. If  
you use 'maint print symbols', GDB  
includes all the symbols for which it has already  
collected full details: that is, filenamereflects  
symbols for only those files whose symbols GDB  
has read. You can use the command info  
sourcesto find out which files these are. If you  
use 'maint print psymbols' instead, the  
dump shows information about symbols that  
GDB only knows partially―that is, symbols  
defined in files that GDB has skimmed, but not  
yet read completely. Finally, 'maint print  
msymbols' dumps just the minimal symbol  
information required for each object file from  
which GDB has read some symbols. See  
filename, maint print  
psymbols filename, maint  
print msymbols filename  
discussion of how GDB reads symbols (in the  
description of symbol-file).  
118 Examining the Symbol Table  
11 Altering Execution  
Once you think you have found an error in your program, you might want to find out  
for certain whether correcting the apparent error would lead to correct results in the  
rest of the run. You can find the answer by experiment, using the GDB features for  
altering execution of the program.  
For example, you can store new values into variables or memory locations, give your  
program a signal, restart it at a different address, or even return prematurely from a  
function.  
11.1 Assignment to variables  
To alter the value of a variable, evaluate an assignment expression. See “Expressions”  
(page 83). For example,  
print x=4  
stores the value 4 into the variable x, and then prints the value of the assignment  
expression (which is 4). See Chapter 9 (page 101), for more information on operators in  
supported languages.  
If you are not interested in seeing the value of the assignment, use the setcommand  
instead of the printcommand. setis really the same as printexcept that the  
expression's value is not printed and is not put in the value history (see “Value history”  
(page 95)). The expression is evaluated only for its effects.  
The setcommand has a number of subcommands that conflict with the names of  
program variables. The set variablecommand is a better alternative for setting  
program variables. The following two examples illustrate the same:  
Example 1  
((gdb)) whatis width  
type = double  
((gdb)) p width  
$4 = 13  
((gdb)) set width=47  
Invalid syntax in expression.  
The invalid expression, of course, is '=47'. In order to actually set the program's  
variable width, use  
Example 2  
((gdb)) set var width=47  
if your program has a variable g, you run into problems if you try to set a new  
value with just 'set g=4', because GDB has the command set gnutarget,  
abbreviated set g:  
((gdb)) whatis g  
type = double  
11.1 Assignment to variables 119  
   
((gdb)) p g  
$1 = 1  
((gdb)) set g=4  
((gdb)) p g  
$2 = 1  
((gdb)) r  
The program being debugged has been started already.  
Start it from the beginning? (y or n) y  
Starting program: /home/smith/cc_progs/a.out  
"/home/smith/cc_progs/a.out": can't open to read symbols:  
Invalid bfd target.  
((gdb)) show g  
The current BFD target is "=4".  
The steps shown above sets the gnutargetto an invalid value in place of the program  
variable g.  
In order to set the variable g, use  
((gdb)) set var g=4  
GDB allows more implicit conversions in assignments than C; you can freely store an  
integer value into a pointer variable or vice versa, and you can convert any structure  
to any other structure that is the same length or shorter.  
To store values into arbitrary places in memory, use the '{...}' construct to generate  
a value of specified type at a specified address (see“Expressions” (page 83)). For  
example, {int}0x83040refers to memory location 0x83040as an integer (which  
implies a certain size and representation in memory), and  
set {int}0x83040 = 4  
stores the value 4 into that memory location.  
11.2 Continuing at a different address  
Ordinarily, when you continue your program, you do so at the place where it stopped,  
with the continuecommand. You can continue at a selected address using one of the  
following commands:  
jump linespec  
Resume execution at line linespec. Execution stops again  
immediately if there is a breakpoint there. See “Printing source  
lines” (page 77), for a description of the different forms of  
linespec. It is common practice to use the tbreakcommand  
in conjunction with jump. See “Breakpoints” (page 51).  
The jumpcommand does not change the current stack frame, the  
stack pointer, the contents of any memory location or any register  
other than the program counter. If line linespecis in a different  
function from the one currently executing, the results may be  
bizarre if the two functions expect different patterns of arguments  
120 Altering Execution  
 
or of local variables. For this reason, the jumpcommand requests  
confirmation if the specified line is not in the function currently  
executing. However, even bizarre results are predictable if you  
are well acquainted with the machine-language code of your  
program.  
jump *address  
Resume execution at the instruction at address address.  
On many systems, you can get much the same effect as the jumpcommand by storing  
a new value into the register $pc. This does not start the execution of your program  
at the specified address, instead only changes the program counter.  
For example,  
set $pc = 0x485  
makes the next continuecommand or stepping command execute at address 0x485,  
rather than at the address where your program stopped. See “Continuing and stepping”  
The most common occasion to use the jumpcommand is to back up―perhaps with  
more breakpoints set―over a portion of a program that has already executed, in order  
to examine its execution in more detail.  
11.3 Giving your program a signal  
You can use the following command to send signals to your program:  
signal signal  
Resume execution where your program stopped, but immediately  
give it the signal signal. signalcan be the name or the number  
of a signal. For example, on many systems signal 2and signal  
SIGINTare both ways of sending an interrupt signal.  
Alternatively, if signalis zero, continue execution without  
giving a signal. This is useful when your program stopped on  
account of a signal and would ordinary see the signal when  
resumed with the continuecommand; 'signal 0' causes it to  
resume without a signal.  
signaldoes not repeat when you press RET a second time after  
executing the command.  
Invoking the signalcommand is not the same as invoking the killutility from the  
shell. Sending a signal with killcauses GDB to decide what to do with the signal  
depending on the signal handling tables (see “Signals” (page 67)). The signal  
command passes the signal directly to your program.  
11.4 Returning from a function  
You can use the following command to return from a function:  
11.3 Giving your program a signal 121  
   
return, return  
expression  
You can cancel execution of a function call with the return  
command. If you give an expressionargument, its value  
is used as the return value from the function value.  
When you use return, GDB discards the selected stack frame (and all frames within  
it). You can think of this as making the discarded frame return prematurely. If you  
wish to specify a value to be returned, give that value as the argument to return.  
This pops the selected stack frame (see “Selecting a frame” (page 73)), and any other  
frames inside of it, leaving its caller as the innermost remaining frame. That frame  
becomes selected. The specified value is stored in the registers used for returning values  
of functions.  
The returncommand does not resume execution; it leaves the program stopped in  
the state that would exist if the function had just returned. In contrast, the finish  
command (see “Continuing and stepping” (page 64)) resumes execution until the  
selected stack frame returns naturally.  
11.5 Calling program functions  
call expr Evaluate the expression exprwithout displaying voidreturned values.  
You can use this variant of the printcommand if you want to execute a function from  
your program, but without cluttering the output with voidreturned values. If the  
result is not void, it is printed and saved in the value history.  
For the A29K, a user-controlled variable call_scratch_addressspecifies the location  
of a scratch area to be used when GDB calls a function in the target. This is necessary  
because the usual method of putting the scratch area on the stack does not work in  
systems that have separate instruction and data spaces.  
11.6 Patching programs  
By default, GDB opens the file containing the executable code of your program (or the  
corefile) as read-only. This prevents accidental alteration to machine code; and it also  
prevents you from intentionally patching your program binary.  
If you would like to be able to patch the binary, you can specify that explicitly with the  
set writecommand. For example, you might want to turn on internal debugging  
flags, or even to make emergency repairs.  
set write on,  
set write off  
If you specify 'set write on', GDB opens executable and core  
files for both reading and writing; if you specify 'set write  
off' (the default), GDB opens them as read-only.  
If you have already loaded a file, you must load it again (using  
the exec-fileor core-filecommand) after changing set  
write, for your new setting to take effect.  
122 Altering Execution  
   
show write  
Display whether executable files and core files are opened for  
writing as well as reading.  
11.6 Patching programs 123  
124  
12 GDB Files  
GDB needs to know the file name of the program to be debugged, both in order to read  
its symbol table and in order to start your program. To debug a core dump of a previous  
run, you must also tell GDB the name of the core dump file.  
12.1 Commands to specify files  
You can specify executable and core dump file names as arguments to the GDB start-up  
Occasionally it is necessary to change to a different file during a GDB session. In these  
situations the GDB commands to specify new files are useful.  
file filename  
Use filename as the program to be debugged. It  
is read for its symbols and for the contents of  
pure memory. It is also the program executed  
when you use the run command. If you do not  
specify a directory and the file is not found in  
the GDB working directory, GDB uses the  
environment variable PATH as a list of  
directories to search, just as the shell does when  
looking for a program to run. You can change  
the value of this variable, for both GDB and your  
program, using the path command.  
On systems with memory-mapped files, an  
auxiliary file named filename.symsmay hold  
symbol table information for filename. If so, GDB  
maps in the symbol table from `filename.syms',  
starting up more quickly. See the descriptions of  
the file options -mappedand -readnow  
(available on the command line, and with the  
commands file, symbol-file, or  
add-symbol-file, described below) for more  
information.  
file  
file with no argument makes GDB discard any  
information it has on both executable file and the  
symbol table.  
exec-file[ filename]  
Specify that the program to be run (but not the  
symbol table) is found in file- name. GDB  
searches the environment variable PATH if  
necessary to locate your program. Omitting  
12.1 Commands to specify files 125  
   
filename means to discard information on the  
executable file.  
symbol-file[ filename ]  
Read symbol table information from file filename.  
PATH is searched when necessary. Use the file  
command to get both symbol table and program  
to run from the same file.  
symbol-file with no argument clears out GDB  
information on the symbol table of your program.  
The symbol-file command causes GDB to forget  
the contents of its convenience variables, the  
value history, and all breakpoints and  
auto-display expressions. This is because they  
may contain pointers to the internal data  
recording symbols and data types, which are part  
of the old symbol table data being discarded  
inside GDB.  
symbol-file does not repeat if you press RET  
again after executing it once. When GDB is  
configured for a particular environment, it  
understands debugging information in whatever  
format is the standard generated for that  
environment; you may use either a gnu compiler,  
or other compilers that adhere to the local  
conventions.  
For most kinds of object files, the symbol-file  
command does not normally read the symbol  
table in full right away. Instead, it scans the  
symbol table quickly to nd which source files  
and which symbols are present. The details are  
read later, one source file at a time, as they are  
needed.  
The purpose of this two-stage reading strategy  
is to make GDB start up faster. For the most part,  
it is invisible except for occasional pauses while  
the symbol table details for a particular source  
file are being read. (The set verbosecommand  
can turn these pauses into messages if desired.  
symbol-filefilename[  
-readnow] [ -mapped], file  
You can override the GDB two-stage strategy for  
reading symbol tables by using the `-readnow'  
126 GDB Files  
filename[ -readnow ] [  
-mapped]  
option with any of the commands that load  
symbol table information, if you want to be sure  
GDB has the entire symbol table available.  
If memory-mapped files are available on your  
system through the mmapsystem call, you can  
use another option, `-mapped', to cause GDB to  
write the symbols for your program into a  
reusable file. Future GDB debugging sessions  
map in symbol information from this auxiliary  
symbol file (if the program has not changed),  
rather than spending time reading the symbol  
table from the executable program. Using the  
`-mapped' option has the same effect as starting  
GDB with the `-mapped' command-line option.  
You can use both options together, to make sure  
the auxiliary symbol file has all the symbol  
information for your program.  
The auxiliary symbol file for a program called  
myprogis called myprog.syms. Once this file  
exists (so long as it is newer than the  
corresponding executable), GDB always attempts  
to use it when you debug myprog; no special  
options or commands are needed.  
The `.syms' file is specific to the host machine  
where you run GDB. It holds an exact image of  
the internal GDB symbol table. It cannot be  
shared across multiple host platforms.  
core-file[ filename]  
Specify the whereabouts of a core dump file to  
be used as the contents of memory.  
Traditionally, core files contain only some parts  
of the address space of the process that generated  
them; GDB can access the executable file itself  
for other parts.  
core-file with no argument specifies that no core  
file is to be used.  
Note that the core file is ignored when your  
program is actually running under GDB. So, if  
you have been running your program and you  
wish to debug a core file instead, you must kill  
the subprocess in which the program is running.  
12.1 Commands to specify files 127  
To do this, use the kill command (see “Killing  
add-symbol-filefilename  
address, add-symbol-file  
filenameaddress[ -readnow]  
[ -mapped], add-symbol-file  
filenameaddress  
data_address bss_address,  
add-symbol-filefilename  
-section address  
The add-symbol-file command reads additional  
symbol table information from the file filename.  
You would use this command when filename  
has been dynamically loaded (by some other  
means) into the program that is running.  
addressshould be the memory address at  
which the file has been loaded; GDB cannot  
figure this out for itself. You can specify up to  
three addresses, in which case they are taken to  
be the addresses of the text, data, and bss  
segments respectively. For complicated cases,  
you can specify an arbitrary number of  
-ssectionaddress pairs, to give an explicit  
section name and base address for that section.  
You can specify any address as an expression.  
The symbol table of the file filename is added to  
the symbol table originally read with the  
symbol-file command. You can use the  
add-symbol-file command any number of times;  
the new symbol data thus read keeps adding to  
the old. To discard all old symbol data instead,  
use the symbol-file command without any  
arguments.  
add-symbol-file does not repeat if you press RET  
after using it.  
You can use the `-mapped' and `-readnow'  
options just as with the symbol- file command,  
to change how GDB manages the symbol table  
information for filename.  
section  
The sectioncommand changes the base  
address of section SECTION of the exec file to  
ADDR. This can be used if the exec file does not  
contain section addresses, (such as in the a.out  
format), or when the addresses specified in the  
file itself are wrong. Each section must be  
changed separately. The infofiles command,  
described below, lists all the sections and their  
addresses.  
128 GDB Files  
infofiles, infotarget  
info filesand info targetare  
synonymous; both commands print the current  
target (see Chapter 13 (page 133)), including the  
names of the executable and core dump files  
currently in use by GDB, and the files from which  
symbols were loaded. Both the commands list  
all possible targets rather than the current targets.  
All file-specifying commands allow both absolute and relative file names as arguments.  
GDB always converts the file name to an absolute file name and remembers it that way.  
GDB automatically loads symbol definitions from shared libraries when you use the  
run command, or when you examine a core file. (Before you issue the run command,  
GDB does not understand references to a function in a shared library, however ―  
unless you are debugging a core file).  
On HP-UX, if the program loads a library explicitly, GDB automatically loads the  
symbols at the time of the shl_load call. See “Breakpoints” (page 51), for more  
information.  
infoshare, info  
sharedlibrary  
Print the names of the shared libraries which are  
currently loaded.  
sharedlibrary regex, share Load shared object library symbols for files matching a  
Unix regular expression. As with files loaded  
automatically, it only loads shared libraries required by  
your program for a core file or after typing run. If regex  
is omitted all shared libraries required by your program  
are loaded.  
regex  
On HP-UX systems, GDB detects the loading of a shared library and automatically  
reads in symbols from the newly loaded library, up to a threshold that is initially set  
but that you can modify if you wish.  
Beyond that threshold, symbols from shared libraries must be explicitly loaded. To  
load these symbols, use the command sharedlibraryfilename. The base address  
of the shared library is determined automatically by GDB and need not be specified.  
To display or set the threshold, use the commands:  
set auto-solib-add  
threshold  
Set the autoloading size threshold, in megabytes.  
If thresholdis nonzero, symbols from all  
shared object libraries will be loaded  
automatically when the inferior begins execution  
or when the dynamic linker informs GDB that a  
new library has been loaded, until the symbol  
table of the program and libraries exceeds this  
threshold. Otherwise, symbols must be loaded  
manually, using the sharedlibrarycommand.  
The default threshold is 100 megabytes.  
12.1 Commands to specify files 129  
show auto-solib-add  
Display the current autoloading size threshold,  
in megabytes.  
12.2 Specifying shared library locations  
On HP-UX, when the shared libraries your program uses are in a different directory  
than the path specified in the source or object files, specify the correct files to use with  
one of two environment variables.  
`GDB_SHLIB_PATH'  
Set this variable to a colon-separated list of directory path  
names where the desired shared libraries reside. GDB  
searches specified list of directories for shared libraries before  
searching the default system directories.  
`GDB_SHLIB_ROOT'  
Set this variable to point to the root of the library in which  
the desired libraries reside.  
NOTE: If you set both the `GDB_SHLIB_PATH'and `GDB_SHLIB_ROOT'  
environment variables, the `GDB_SHLIB_PATH'behavior overrides  
`GDB_SHLIB_ROOT'.  
These environment variables are useful when you are analyzing core files on a system  
other than the one that produced the core file.  
For example, if you want GDB to search for libraries in /home/debugger/liband  
/tmp/libbefore searching the default system directories for libraries, you can use  
this setting:  
GDB_SHLIB_PATH=/home/debugger/lib:/tmp/lib  
With this setting, GDB searches the directories in the order specified until it finds a  
library with the correct name.  
In this example, if GDB encounters a library by the name of /usr/lib/libsubs.sl,  
GDB searches first for /home/debugger/lib/libsubs.sland then for /tmp/lib/  
libsubs.sl. If neither of these exists, then GDB searches the default system directories  
and finds /usr/lib/libsubs.sl.  
In most cases, GDB_SHLIB_PATHallows more flexibility than GDB_SHLIB_ROOT'  
because it allows you to specify more than one path. However, there are some cases in  
which you may want to choose to use GDB_SHLIB_ROOT.  
For example, if you have more than one shared library with the same name but different  
path names, you may want to use GDB_SHLIB_ROOTbecause GDB searches for libraries  
based on the full path name.  
Note that GDB_SHLIB_PATHmay not give you the results you expect because GDB  
searches for libraries that match only the name, regardless of the path, and always  
accepts the first library that matches the name.  
130 GDB Files  
 
For example, if you want to use /tmp/usr/lib/libsubs.sland /tmp/usr/share/  
lib/libsubs.sl, you can set GDB_SHLIB_ROOT' to /tmp. Now whenever GDB  
encounters a library with the name /usr/lib/libsubs.sland /usr/share/lib/  
libsubs.sl, GDB looks at /tmp/usr/lib/libsubs.sland /tmp/usr/share/  
lib/libsubs.slrespectively.  
12.3 Errors reading symbol files  
While reading a symbol file, GDB occasionally encounters problems, such as symbol  
types it does not recognize, or known bugs in compiler output. By default, GDB does  
not notify you of such problems, since they are relatively common and primarily of  
interest to people debugging compilers. If you are interested in seeing information  
about ill-constructed symbol tables, you can either ask GDB to print only one message  
about each such type of problem, no matter how many times the problem occurs; or  
you can ask GDB to print more messages, to see how many times the problems occur,  
with the set complaints command (see“Optional warnings and messages” (page 284)).  
The messages currently printed, and their meanings, include:  
inner block not inside outer block in The symbol information shows where symbol  
symbol  
scopes begin and end (such as at the start of a  
function or a block of statements). This error  
indicates that an inner scope block is not fully  
contained in its outer scope blocks.  
GDB circumvents the problem by treating the  
inner block as if it had the same scope as the  
outer block. In the error message, symbol may be  
shown as "(don't know)"if the outer block  
is not a function.  
block at address out of order  
The symbol information for symbol scope blocks  
should occur in order of in- creasing addresses.  
This error indicates that it does not do so.  
GDB does not circumvent this problem, and has  
trouble locating symbols in the source file whose  
symbols it is reading. (You can often determine  
what source file is affected by specifying set  
bad block start address patched  
The symbol information for a symbol scope block  
has a start address smaller than the address of  
12.3 Errors reading symbol files 131  
 
the preceding source line. This is known to occur  
in the SunOS 4.1.1 (and earlier) C compiler.  
GDB circumvents the problem by treating the  
symbol scope block as starting on the previous  
source line.  
bad string table offset in symbol n  
Symbol number n contains a pointer into the  
string table which is larger than the size of the  
string table.  
GDB circumvents the problem by considering  
the symbol to have the name foo, which may  
cause other problems if many symbols end up  
with this name.  
unknown symbol type 0xnn  
The symbol information contains new data types  
that GDB does not yet know how to read. 0xnn  
is the symbol type of the uncomprehended  
information, in hexadecimal.  
GDB circumvents the error by ignoring this  
symbol information. This usually allows you to  
debug your program, though certain symbols  
are not accessible. If you encounter such a  
problem and feel like debugging it, you can  
debug (gdb) with itself, breakpoint on  
complain, then go up to the function  
read_dbx_ symtaband examine *bufpto see  
the symbol.  
stub type has NULL name  
GDB could not nd the full definition for a struct  
or class.  
const/volatile indicator missing (ok The symbol information for a C++ member  
if using g++ v1.x), got...  
function is missing some information that recent  
versions of the compiler should have output for  
it.  
info mismatch between compiler  
and debugger  
GDB could not parse a type specification output  
by the compiler.  
132 GDB Files  
13 Specifying a Debugging Target  
A targetis the execution environment occupied by your program.  
Often, GDB runs in the same host environment as your program; in that case, the  
debugging target is specified as a side effect when you use the fileor corecommands.  
For HP-UX specific information, see undefined [HP-UX Targets], page undefined.  
When you need more flexibility, for example, running GDB on a physically separate  
host, or controlling a standalone system over a serial port or a realtime system over a  
TCP/IP connection you can use the targetcommand to specify one of the target types  
13.1 Active targets  
There are three classes of targets: processes, core files, and executable files. GDB can  
work concurrently on up to three active targets, one in each class. This allows you, for  
example, to start a process and inspect its activity without abandoning your work on  
a core file.  
For example, if you execute `gdb a.out', then the executable file a.out is the only active  
target. If you designate a core file as well presumably from a prior run that crashed  
and coredumped, then GDB has two active targets and uses them in tandem, looking  
first in the corefile target, then in the executable file, to satisfy requests for memory  
addresses. (Typically, these two classes of target are complementary, since core files  
contain only the contents of the program read-write memory, variables, machine status  
etc. While the executable files contain only the program text and initialized data.)  
When you type run, your executable file becomes an active process target as well. When  
a process target is active, all GDB commands requesting memory addresses refer to  
that target; addresses in an active core file or executable file target are obscured while  
the process target is active.  
Use the core-file and exec-file commands to select a new core file or executable target  
(see “Commands to specify files” (page 125)). To specify as a target a process that is  
already running, use the attach command (see “Debugging a Running Process”  
13.2 Commands for managing targets  
target type parameters  
Connects the GDB host environment to a target  
machine or process. A target is typically a protocol  
for talking to debugging facilities. You use the  
argument type to specify the type or protocol of the  
target machine.  
Further parameters are interpreted by the target  
protocol, but typically include things like device  
13.1 Active targets 133  
     
names or host names to connect with, process  
numbers, and baud rates.  
The targetcommand does not repeat if you press  
RET again after executing the command.  
help target  
Displays the names of all targets available. To  
display targets currently selected, use either info  
target or info files (see “Commands to specify files”  
help targetname  
Describe a particular target, including any  
parameters necessary to select it.  
set gnutarget args  
GDB uses its own library BFD to read your files.  
GDB knows whether it is reading an executable, a  
core, or a .ofile; however, you can specify the file  
format with the set gnutargetcommand. Unlike  
most targetcommands, with gnutargetthe  
targetrefers to a program, not a machine.  
Warning: To specify a file format with set gnutarget,  
you must know the actual BFD name.  
show gnutarget  
Use the show gnutargetcommand to display  
what file format gnutargetis set to read. If you  
have not set gnutarget, GDB will determine the  
file format for each file automatically, and show  
gnutargetdisplays `The current BDF target is  
"auto"'.  
Here are some common targets (available, or not, depending on the GDB configuration):  
target execprogram  
target corefilename  
target remotedev  
An executable file. target exec programis the  
same as exec-file program.  
A core dump file. target core filenameis the  
same as core-file filename.  
Remote serial target in GDB-specific protocol. The  
argument dev specifies what serial device to use for the  
connection (for example, /dev/ttya). targetremote  
supports the loadcommand. This is only useful if you  
have some other way of getting the stub to the target  
system, and you can put it somewhere in memory  
where it will not get clobbered by the download.  
134 Specifying a Debugging Target  
target sim  
Builtin CPU simulator. GDB includes simulators for  
most architectures. In general,the following commands  
work:  
target sim  
load  
run  
However, you cannot assume that a specific memory  
map, device drivers, or even basic I/O is available,  
although some simulators do provide these.  
Some configurations may include these targets as well:  
target nromdev  
NetROM ROM emulator. This  
target only supports  
downloading.  
Different targets are available on different configurations of GDB; your configuration  
may have more or fewer targets.  
Many remote targets require you to download the executable code once you have  
successfully established a connection.  
loadfilename  
Depending on what remote debugging facilities are configured  
into GDB, the load command may be available. Where it exists,  
it is meant to make filename (an executable) available for  
debugging on the remote system|by downloading, or dynamic  
linking, for example. loadalso records the filenamesymbol  
table in GDB, like the add-symbol-filecommand.  
If your GDB does not have a loadcommand, attempting to  
execute it gets the error message "You can't do that when your  
target is ...".  
The file is loaded at whatever address is specified in the  
executable. For some object file formats, you can specify the load  
address when you link the program; for other formats, like a.out,  
the object file format specifies a fixed address. loaddoes not  
repeat if you press RET again after using it.  
loaddoes not repeat if you press RET again after using it.  
13.3 Choosing target byte order  
Some types of processors, such as the MIPS, PowerPC, and Hitachi SH, offer the ability  
to run either big-endian or little-endian byte orders. Usually the executable or symbol  
will include a bit to designate the endian-ness, and you will not need to worry about  
which to use. However, you can adjust the processor byte order manually using one  
of the following commands:  
13.3 Choosing target byte order 135  
 
set endianbig  
set endianlittle  
set endianauto  
Instruct GDB to assume the target is big-endian.  
Instruct GDB to assume the target is little-endian.  
Instruct GDB to use the byte order associated with the  
executable.  
show endian  
Display GDB's current idea of the target byte order.  
Note that these commands merely adjust interpretation of symbolic data on the host,  
and that they have absolutely no effect on the target system.  
136 Specifying a Debugging Target  
14 HP-UX Configuration-Specific Information  
While nearly all GDB commands are available for all native and cross versions of the  
debugger, there are some exceptions. This chapter describes features, commands and,  
enhancements available only on HP-UX.  
14.1 Summary of HP Enhancements to GDB  
WDB provides the following features in addition to the standard GDB features:  
Support for debugging memory problems.  
Support for heap usage reporting  
The min-heap-size <num> option for set heap-check command reports  
the heap allocations that exceed the specified number <num>of bytes based on  
the cumulative number of bytes that are allocated at each call-site, which is inclusive  
of multiple calls to mallocat a particular call site.  
Heap checking commands info heap high-memand set heap-check  
high-mem-count X_number.  
Commands which, exit, info heap processand, info heap arena.  
WDB supports the +check compiler option on Integrity systems to invoke batch  
RTC, to determine runtime memory problems.  
Enhanced batch RTC support, for better reporting, and options have replaced old  
ones.  
RTC heap corruption checks for calls to strcpy(), memset(), and memcpy()  
have been added.  
Support for memory checking analysis for user defined memory management  
routines.  
High water mark records the number of times the brk()value changes.  
Heap analysis on programs with pending signals using the info leakcommand.  
Commands info module ADDRESS, show envvars, and info corruption.  
Command line option (-pid or -p) to attach to an existing process.  
Support for debugging kernel threads and user threads.  
Support for enabling and disabling threads.  
Thread debugging commands set thread-check on/off, info thread  
[thread-id], info mutex [mutex- id], info condvar [condvar-id],  
and info rwlock [rwlock- id].  
Enhanced thread debugging options for set thread-checkcommand such as  
recursive-relock, num-waiter, stack-util,  
thread-exit-no-join-detach, thread-exit-own-mutex,  
cv-wait-no-mx, cv-multiple-mxs, mixed-sched- policy, and  
unlock-not-own.  
14.1 Summary of HP Enhancements to GDB 137  
   
Serial debugging of a parent and child process.  
Support for Parallel Processing limited to pthreadparallelism, but not the compiler  
generated parallelism, for example, with directives.  
Implementation of ask mode for set follow-fork-mode.  
Support for setting breakpoints using shared library name.  
Support for core file commands packcore, unpackcore, getcore, dumpcore  
and info rtti address.  
On PA-RISC systems, sanity check for core files dumped by hardware generated  
signals can be performed. HP WDB can detect and warn the user about certain  
cases of corrupted core files.  
Inline support is on by default on Integrity systems. On PA-RISC, the inline support  
is still off by default.  
For PA 64-bit applications, WDB can step into shared library bind-on-reference  
calls. This support is available for PA 32-bit as well.  
Interception of synchronous signals used by sigwait(), sigwaitinfo()and  
sigtimedwait()functions. These signals are displayed by WDB just like  
asynchronous signals but are always passed to the debugger whether [nopass]  
is set or not.  
Support for debugging hardware watchpoints and shared libraries.  
For PA 64-bit applications, WDB can step into shared library bind-on-reference  
calls. This support is available for PA 32-bit as well.  
Implementation of -mapsharedoption to suppress mapping all the shared libraries  
in a process private.  
Support for deferred breakpoints on dlopenedand shl_loadedlibraries with  
stripped main program.  
Additional support for procedural breakpoints.  
C99 variable arrays implemented on Integrity systems. This support is available  
on PA-RISC as well.  
The show envvarscommand can be used to print information on environment  
variables supported by WDB.  
Support for handling __fpregdata type.  
Support for making changes to the program when debugging without having to  
re- compile or re-link it.  
Hardware breakpoints on Integrity systems.  
Support for debugging PA-RISC Applications on Integrity systems.  
Unwinding Java stack frames on Integrity systems. The 64-bit version of gdb can  
unwind through Java stack frames using the shared library in the Java product.  
The 64-bit library is part of the JDK 1.4.2.10 and JDK 1.5.0.03 products.  
Enhanced nextiand stepicommands. The WDB nextiand stepicommands,  
prints the assembly instruction along with the next source line.  
138 HP-UX Configuration-Specific Information  
Enhanced info symbol ADDRESScommand. The info symbol ADDRESS  
command has been enhanced to search for a symbol at the given address.  
Previously, the info symbolcommand could be used only to search the global  
namespace.  
Enhanced Java debugging support.  
Support for debugging C++ programs running under Linux Runtime Environment.  
Support for stop in/at dbx commands.  
Support for GNU GDB logging commands.  
Support for persistent display of expressions across re-runs. The user does not  
need to re-initiate the display settings for every run.  
Viewing wide character and wide-character strings of type wchar_t using the  
type printcommand.  
Support for debugging of executables with method and expressions involving  
covariant return types  
Support for commands, catch throw, catch catch, and info catch, for  
debugging exception handlers in C++ on Integrity systems, with the aCC compiler  
A.06.00 and later.  
Support for the steplastcommand for C and C++. However, steplast  
command is not supported.  
Support for dumping an array into an ASCII file.  
Support for the Fortran array slices.  
Source level debugging of Fortran applications that contain nested routines.  
Support for making command line calls in a stripped executable.  
Support for a terminal user interface that provides view of source while debugging  
at the WDB command line.  
support for debugging 32-bit and 64-bit PA-RISC programs as well as 64-bit Itanium  
programs .  
Support for assembly-level debugging.  
Support for a subset of xdbcommands, enabled with -xdbmode.  
Support for Java/C/aCC++ stack unwinding with Java SDK version 1.3.1.02 or later  
for HP-UX.  
Visual Interface for HP WDB -tuimode supports output logging  
Command line calls for 64-bit PA-RISC applications that are not linked with end.o.  
Command watch_target.  
Command line option set display-full-paththat displays the full pathname  
of the source file name while printing the frame information.  
Command line option set dereference [on |off]when off, WDB does  
not dereference char *variables by default.  
Support for show macro, macro expand to view, and expand macro  
expressions.  
14.1 Summary of HP Enhancements to GDB 139  
Support for evaluating macros.  
Support for printing the execution path entries in the current frame, or thread.  
New Command for Searching a Pattern in the Memory Address Space  
New Option to Limit the Number of Frames Unwound  
Updated Procedure to Process the Initialization File  
Enhancement to the info sharedCommand  
Enhancement to the info targetCommand  
Support for automatic debugging of shared libraries  
Attaching a PA-RISC debugger to a PA-RISC process  
Improvement in the performance of memory debugging for the stringoption  
NOTE: For new commands see HP WDB Release Notes available at http://  
14.2 HP-UX dependencies  
14.2.1 Linker Dependencies  
Several features available in WDB depend on specific versions of the linker or the  
compiler.  
Linker patch required for +objdebug  
For releases prior to HP-UX 11i v2 (for IA) and HP-UX 11i v1 (for PA-RISC), you  
must install the latest linker patch to generate object modules that enable faster  
linking and smaller executable file sizes for large applications. See your compiler  
release notes for more details.  
Support for debugging incrementally linked 64-bit programs  
This feature requires linker version B.11.18 or later on HP-UX 11i v1.  
Support to automatically preload librtc.slwith chatr +mem_checkoption.  
This feature requires linker version B.11.61 and later on HP 9000 systems, and  
linker version B.12.46 and later on Integrity systems. +mem_checkoption is used  
only for memory debugging.  
Support to automatically preload librtc.slwith chatr +rtcoption.  
This feature requires linker version B.11.66 and later on HP 9000 systems, and  
linker version B.12.51 and later on Integrity systems. However, +mem_check  
option is retained with the latest Linker version.  
14.2.2 Dependent Standard Library Routines for Run Time Checking  
The Run Time Checking feature (Interactive and Batch Mode) of WDB cannot be used  
with applications that re-define or over-ride the default system-supplied versions of  
140 HP-UX Configuration-Specific Information  
     
the standard library routines under libc.soand libdld.so. The following standard  
libraries are dependencies for runtime checking:  
abort()  
unlink()  
atoi()  
chdir()  
uwx_register_callbacks()  
lseek()  
dlsym()  
open()  
strstr()  
strcat()  
ctime()  
sprintf()  
strcmp()  
printf()  
dlclose()  
memchr()  
strrchr()  
dlgetname()  
dlget()  
pthread_self()  
putenv()  
shmctl()  
strchr()  
rand()  
clock_gettime()  
strlen()  
dlhook()  
dlmodinfo()  
environ()  
getenv()  
strlen()  
execl()  
uwx_self_lookupip()  
shl_get()  
shl_unload()  
shl_findsym ()  
strtok_r()  
time()  
uwx_get_reg()  
shl_get_r()  
perror()  
exit()  
fclose()  
fork()  
strdup()  
fopen()  
uwx_init()  
uwx_self_copyin()  
creat()  
uwx_step()  
fprintf()  
fscanf()  
sscanf()  
strcasecmp()  
getcwd()  
getpagesize()  
getpid()  
srand()  
write()  
uwx_self_init_context()  
pthread_getschedparam()  
uwx_self_init_info()  
uwx_register_alloc_cb()  
U_STACK_TRACE()  
close()  
strchr()  
The runtime checking (of dynamic memory, libraries, and pthreads) in the debugger  
relies on the semantic and standard behavior of these library routines. Run Time  
Checking results in unexpected and unpredictable behavior when used with applications  
that substitute or re-define these library routines.  
Before enabling the Run Time Checking feature in WDB, use the nm(1) command to  
determine if the application or the dependent libraries in the application redefine or  
substitute these standard library routines.  
14.2 HP-UX dependencies 141  
14.3 Supported Platforms and Modes  
Supported Platforms  
HP WDB supports source-level debugging of programs written in HP C, HP aC++,  
and Fortran 90 on Integrity systems running on HP-UX 11i v2 or later and PA-  
RISC systems running HP-UX 11i v1 and later.  
Support for assembly-level debugging  
HP WDB provides support for assembly-level debugging.  
Support for automatic loading of debug information  
Debug information is automatically loaded from modules when an application is  
com- piled with the +objdebugoption.  
Support for debugging PA-RISC programs on Itanium-based systems  
You can debug PA-RISC applications and core files on Itanium-based systems.  
When you start HP WDB, if the debug target is a PA-RISC binary program, the  
debugger automatically loads PA-RISC WDB. The PA-RISC version of HP WDB  
is provided as part of the HP-UX operating system.  
Support for debugging large core files (> 2GB)  
HP WDB supports debugging of core files with sizes more than 2 GB.  
Support co-variant type  
HP WDB can step into a co-variant function. The compiler-generated function  
called thunks, which is used internally by the compiler to support co-variant return  
type, is not shown when you do a backtrace or switch from one frame to another  
frame. Similarly, using a finishor returncommand at a co-variant callee  
function directly returns the control back to the caller of thunks.  
New attach command line options and handling (-pid or -p)  
HP WDB accepts -pidor -pfollowed by a process ID to attach a running process  
to the debugger.  
142 HP-UX Configuration-Specific Information  
 
NOTE: HP WDB cannot be attached to a process that is traced by tools which use  
ttrace, such as Caliper, adb, and tusc. The debugger displays the following error message  
on attempting to attach to such a process:  
Attaching to process <pid> failed.  
Hint: Check if this process is already being traced by another gdb or  
other trace tools like caliper and tusc.  
Hint: Check whether program is on an NFS-mounted file-system.  
If so, you will need to mount the file system with the "nointr" option  
with mount(1) or make a local copy of the program to resolve this problem.  
14.4 HP-UX targets  
On HP-UX systems, GDB has been configured to support debugging of processes  
running on the PA-RISC and Itanium architectures. This means that the only possible  
targets are:  
An executable that has been compiled and linked to run on HP-UX. This includes  
binaries that have been marked as SHMEM_MAGIC.  
A live HP-UX process, either started by WDB (with the runcommand) or started  
outside of WDB and attached to (with the attachcommand).  
A core file generated by an HP-UX process that previously aborted execution.  
GDB on HP-UX has not been configured to support remote debugging, or to support  
programs running on other platforms.  
WDB can only debug C++ programs compiled with HP aC++, the ANSI-compatible  
C++ compiler.  
14.5 Support for Alternate root  
HP WDB supports alternate root functionality, which is helpful when you do not want  
to use the system-installed HP WDB or its components.  
The environment variable WDB_ROOTspecifies the alternate root for HP WDB. You  
must specify a structure similar to the default /opt/langtoolsused for HP WDB.  
You can use the environment variable GDB_ROOTto specify an alternate root for GDB.  
If you specify both WDB_ROOTand GDB_ROOT, the value for GDB_ROOTis ignored.  
HP WDB supports these environment variables to override the location of different  
component of HP WDB.  
Defined Variable /opt/langtools/ GDB_ROOT  
$GDB_SERVER  
bin  
WDB Location  
GDB location  
/opt/langtools/ n/a  
n/a  
lib  
WDB_ROOT  
$GDB_ROOT/bin LIBRTC_SERVER  
14.4 HP-UX targets 143  
   
librtc.sllocation $WDB_ROOT/bin /opt/langtools/ n/a  
bin  
None  
$WDB_ROOT/bin GDB_SERVER  
n/a  
/opt/langtools/ $WDB_ROOT/lib  
bin  
n/a  
$LIBRTC_SERVER  
NOTE: If you define WDB_ROOTor GDB_ROOTbut do not create the correct directory  
structure below it, the debugger may fail.  
14.6 Specifying object file directories  
GDB enables automatic loading of debug information from object modules when an  
application is compiled with the +objdebugoption.  
GDB uses the full path name to the object module files and searches the same directories  
for source files.  
Although this behavior is transparent, you can control when and how object files are  
loaded with three commands:  
objectdirpath  
Specifies a colon (:) separated list of directories in which  
GDB searches for object files. These directories are added  
to the beginning of the existing objectdirpath. If you  
specify a directory that is already in the objectdirpath,  
the specified directory is moved up in the objectdir  
path so that it is searched earlier.  
GDB recognizes two special directory names: $cdir,  
which refers to the compilation directory (if available) and  
$cwd, which tracks GDB's current working directory.  
objectloadfile.c  
Causes GDB to load the debug information for file.c  
immediately. The default is to load debug information  
from object modules on demand.  
objectretryfile.c  
Forces GDB to retry loading an object file if GDB  
encounters a file error while reading an object module.  
File errors that might cause this include incorrect  
permissions, file not found, or if the objectdirpath  
changes. By default, GDB does not try to read an object  
file after an error.  
pathmap  
Enables you to define a list of substitution rules to be  
applied to path names to identify object files and the  
corresponding source files. The pathmapcommand,  
144 HP-UX Configuration-Specific Information  
 
however, may not find source files if the object files are  
not available.  
This minimizes or eliminates the need to specify multiple  
objectdircommands when object files are moved from  
the compilation directories or when compilation  
directories are mounted over NFS.  
To use this feature, the program must be compiled with  
the +objdebugoption. For information on how pathmap  
works type help pathmapat the HP WDB prompt.  
If the debugger cannot find the source files, perform the following steps:  
1. Make certain the files were compiled with the -gswitch. Type info sources to nd  
the list of files that the debugger knows were compiled with -g.  
2. Make certain that the debugger can find the source file. Type show dirto find  
the list of directories the debugger uses to search for source files and type set  
dirto change that path.  
On HP-UX, the debug information does not contain the full path name to the source  
file, only the relative path name that was recorded at compile time. Consequently,  
you may need several dircommands for a complex application with multiple  
source directories. One way to do this is to place them in a .gdbinitfile placed  
in the directory used to debug the application.  
A sample of the .gdbinitfile might look like the following:  
dir /home/fred/appx/system  
dir /home/fred/appx/display  
dir /home/fred/appx/actor  
dir /home/fred/appx/actor/sys  
...  
When you compile the program with the +objdebugoption, the debugger may  
nd the source files without using the dircommand. This happens because the  
debugger stores the full path name to the object files and searches for source files  
in the same directories.  
14.7 Fix and continue debugging  
Fix and continue enables you to see the result of changes you make to a program you  
are debugging without having to re-compile and re-link the entire program.  
For example, you can edit a function and use the fixcommand, which automatically  
re-compiles the code, links it into a shared library, and continues execution of the  
program, without leaving the debugger.  
With Fix and Continue, you can experiment with various ways of fixing problems until  
you are satisfied with the correction, before you exit the debugger.  
14.7 Fix and continue debugging 145  
 
The advantages include:  
You do not have to recompile and relink the entire program.  
You do not have to reload the program into the debugger.  
You can resume execution of the program from the x location.  
You can speed up the development cycle.  
NOTE: Fix and Continue is only supported with the most recent versions of HP C  
and HP aC++ on PA-RISC systems.  
In command-line mode, you use the editcommand before invoking the fixcommand.  
The editcommand has the following syntax:  
edit file1 file2  
where  
file represents one or more source files for the current executable. If you do not specify  
a file name, WDB edits the currently open source file.  
When you edit a file with the editcommand and save the changes, the original source  
file contains the changes, even if you do not use the fixcommand to recompile the  
program in the debugger.  
14.7.1 Fix and Continue compiler dependencies  
Fix and Continue is supported only for PA-RISC on HP-UX 11.x with these compilers:  
HP C/ANSI C A.11.01.20, or later  
HP aC++ A.03.25, or later  
HP Fortran 90 2.4, or later  
14.7.2 Fix and Continue restrictions  
Fix and Continue has the following restrictions and behaviors:  
You cannot recompile code that has been optimized.  
You cannot add, delete, or reorder the local variables and parameters in a function  
currently active on the stack.  
If you x a routine in a file that contains function pointers, those function pointers  
become invalid and will likely cause the program to receive a SIGSEGV error if  
the pointers are used.  
You cannot change the type of a local variable, file static, global variable, or  
parameter of a function.  
You cannot add any function calls that increase the size of the parameter area.  
You cannot change a local or file static or global variable to be a register variable,  
and vice-versa.  
146 HP-UX Configuration-Specific Information  
   
You cannot add an alloca()function to a frame that did not previously use  
alloca().  
New structure fields can be added at the end of a structure object, not in the middle  
of a structure. New fields are only accessible by the modified files. Old structure  
fields remain intact; no swapping of them is permitted.  
If the redefined function is in the call stack but not on the top of the call stack, the  
modified code will not be executed when the execution resumes.  
The modified function will be executed when it is called next time, or a rerun.  
Breakpoints in the original source file are moved to the modified fie. Breakpoints  
in the already instantiated functions on the call stack in the original file are lost.  
If you change the name of a function and there was a breakpoint set to the old  
function, WDB does not move the breakpoint to the new function. The old  
breakpoint is still valid.  
If the number of lines of the modified file is different from that of the original file,  
the placement of breakpoints may not be correct.  
When the program resumes, the program counter is moved to the beginning of  
the same line in the modified function. The program counter may be at the wrong  
line.  
14.7.3 Using Fix and Continue  
When WDB recompiles a fixed source file, it uses the same compiler and the same  
options that were used to create the original executable. If the compiler generates any  
syntax errors or it encounters any of the restrictions, WDB does not patch the changes  
into the executable image being debugged.  
After you successfully recompile the changes, WDB uses the fixed version of the code  
when you use any of the execution commands such as step, run, or continue.  
When you use the editcommand, WDB then monitors any edited source files for  
additional changes. After you enter the initial fixcommand, WDB checks for additional  
saved changes to a source file each time you enter a program execution command. If  
a saved source file has been changed, WDB asks if you want to fix the changed source,  
allowing you to apply repeated fixes without explicitly entering the fixcommand.  
The Fix and Continue facility enables you to make the following changes:  
Change existing function definitions.  
Disable, reenable, save, and delete redefinitions  
Adding global and file static variables.  
Add new structure fields to the end of a structure type object.  
Set breakpoints in and single-step within redefined code.  
14.7 Fix and continue debugging 147  
 
NOTE: You must rebuild the program after you use the fixcommand because the  
changes you make are temporarily patched into the executable image. The changes are  
lost if you load a different executable and are not re ected in the original executable  
when you exit the debugger.  
14.7.4 Example Fix and Continue session  
This example shows how you can make and test changes to a function without leaving  
the debugger session.  
Here is a short sample C program with an error:  
int sum (num) int num;  
{
int j, total = 0;  
for (j = 0; j <= num; j++)  
total += num;  
}
main()  
{
int num = 10;  
printf("The sum from 1 to %d is = %d\n", num, sum(num));  
}
1. Compile the program.  
cc sum.c -g -o mysum  
/usr/ccs/bin/ld: (Warning) At least one PA 2.0 object file  
(sum.o) was detected.  
The linked output may not run on a PA 1.x system.  
2. Run the program.  
./mysum  
The sum from 1 to 10 is = 0  
This result is obviously wrong. We need to debug the program.  
3. Run the debugger:  
gdb mysum  
HP gdb 3.0 for PA-RISC 1.1 or 2.0 (narrow), HP-UX 11.00.  
Copyright 1986 - 2001 Free Software Foundation, Inc.  
Hewlett-Packard Wildebeest 3.0 (based on GDB ) is covered by the  
GNU General Public License. Type "show copying" to see the  
conditions to change it and/or distribute copies. Type  
"show warranty" for warranty/support.  
If the TERMenvironment variable is not set to hpterm, start the debugger and set  
the terminal type for editing in WDB with this command (kshshell):  
TERM=hpterm gdb mysum  
148 HP-UX Configuration-Specific Information  
 
The problem might be that there is no return for the num function. You can correct  
this without leaving the debugger.  
4. Set a break point at main:  
(gdb) b main  
Breakpoint 1 at 0x23f8: file sum.c, line 11.  
5. Run the program:  
(gdb) run  
Starting program: /tmp/hmc/mysum  
Breakpoint 1, main () at sum.c:11  
11 int num = 10;  
6. When the program stops at the break point, use the editcommand to make  
changes to the source file.  
Because you are going to edit the current file, you do not need to specify a source  
file name.  
(gdb) edit  
The editcommand opens a new terminal session using your environment variable  
settings for terminal and editor. The debugger automatically loads the source file.  
7. Make the necessary changes. In this case, add:  
return total;  
to the function named num.  
8. Save the edited source file and exit the editor. This saves the changes in the actual  
source file for the program.  
9. Use the fixcommand to recompile the program to see the results of the changes:  
(gdb) fix  
Compiling /dev/src/sum.c...  
Linking...  
Applying code changes to sum.c.  
Fix succeeded.  
The fixcommand creates a new executable that includes the changes you made  
to the source file.  
The debugger automatically uses the new executable and picks up the debugging  
session where you stopped before using the editcommand.  
For example, you can continue stepping through the program and you will nd the  
new return total; statement in the source view. You can print the value of  
total, and see that the result is 110.  
14.7 Fix and continue debugging 149  
10. When you finish with the debugging session, you can exit the debugger normally:  
(gdb) q  
The following modules in /dev/src/mysum have been fixed:  
/dev/src/sum.c  
Remember to remake the program.  
The debugger message lists the source files that you have changed during the  
debugging session.  
NOTE: You must rebuild the program after you use the fixcommand because  
the changes you make are temporarily patched into the executable image. The  
changes are lost when you exit the debugger or you load a different executable.  
14.8 Inline Support  
HP WDB enables you to debug inline functions in applications compiled with -goption.  
To enable inline debugging in HP 9000 systems, the applications must be compiled  
with the +inline_debugoption (introduced in the A.03.65 and later versions of the  
HP aC++ compiler). In Integrity systems, the applications that are compiled with -g  
option support inline debugging by default and require no additional options. Compiler  
versions A.06.02 and later support the inline debugging feature in Integrity systems.  
14.8.1 Inline Debugging in HP 9000 Systems  
To debug inline functions in HP 9000 systems, complete the following steps:  
1. Compile the source files with the +inline_debugoption.  
For example:  
/opt/aCC/bin/aCC -g +inline_debug test.c  
2. Inline debugging is enabled by default. To explicitly enable or disable inline de-  
bugging, complete either of the following steps before loading the application to  
the debugger:  
$ gdb --inline=<on/off> a.out  
or  
(gdb) set inline-debug <on/off>  
150 HP-UX Configuration-Specific Information  
   
3. You can use the following commands for debugging inline functions in HP 9000  
systems:  
step  
next  
list  
backtrace  
frame <n>  
info locals  
info args  
The following commands are not available for debugging inline functions in HP  
9000 systems:  
breakpoint  
info frame  
disassembly  
NOTE: Inline debugging commands are not available for inlined template  
functions and inlined functions which are defined after the call site.  
14.8.2 Inline Debugging in Integrity Systems  
In Integrity systems, applications that are compiled with -g option support inline  
debugging by default. Compiler versions A.06.02 and later support the inline debugging  
feature in Integrity systems and require no additional options.  
WDB 5.6 and later versions enable you to set and modify breakpoints in inline functions  
for programs compiled with optimization level less than +O2. The breakpoint features  
for inline functions are introduced as additional options in the set inline-debug  
command.  
You can toggle the options for inline debugging by entering either of the following  
commands:  
(gdb) set inline-debug <option>  
or  
$ gdb --inline= <option>  
The following options available for the set inline-debugcommand:  
on  
off  
inline_bp_all  
inline_bp_individual  
The set inline-debugon command enables the inline debugging feature without  
the inline breakpoints options in Integrity systems. This command is enabled by default.  
14.8 Inline Support 151  
 
The set inline-debug offcommand disables the inline debugging feature. You  
can disable inline debugging by entering this command before attaching the debugger  
to the application.  
The set inline-debug inline_bp_allcommand enables you to set and modify  
break- points on all instances of a particular inline function. It also enables the inline  
debugging feature. A single instance of the specified inline function is displayed as a  
representative in- stance for all the instances of the specified inline function. This creates  
a single-breakpoint illusion for multiple instances of the inline function. You can set  
and modify breakpoints on all the instances of the inline functions by setting and  
modifying breakpoints on the displayed instance of the inline function. You must enter  
this command before attaching the debugger to the application.  
The set inline-debug inline_bp_individualcommand enables you to set  
and modify breakpoints on a specific instance of an inline function. It also enables the  
inline debugging feature. All instances of the inline function are displayed separately  
with individual breakpoint occurrences. You can set or delete individual breakpoints  
on a specific instance of an inline function without modifying the breakpoints on other  
instances of the inline function. You must enter this command before attaching the  
debugger to the application.  
Limitations:  
The inline breakpoint features are not available for programs that are com- piled  
with +O2 optimization level and above.  
The inline breakpoint features can degrade performance of the application that is  
being debugged. You can explicitly disable the breakpoint features when the  
features are not required and continue to use other inline debugging features, such  
as stepand next.  
14.8.2.1 Debugging Inline Functions in Integrity Systems  
To debug inline functions in Integrity systems, complete the following steps:  
1. The application must be compiled with the -g option for inline debugging. No  
additional options are required. For example:  
/opt/aCC/bin/aCC -g test.c  
2. Inline debugging without the breakpoint feature is enabled by default. You can  
modify the inline debugging settings by toggling the options for the set inline-  
debugcommand.  
To enable inline debugging without inline breakpoint support, enter either  
of the following commands:  
(gdb) set inline-debug on  
or  
$ gdb --inline = on  
To set and modify breakpoints collectively on all instances of inline functions  
and enable inline debugging, enter either of the following commands:  
152 HP-UX Configuration-Specific Information  
 
(gdb) set inline-debug inline_bp_all  
or  
$ gdb --inline = inline_bp_all  
To set and modify individual breakpoints on specific instances of inline  
functions and enable inline debugging, enter either of the following commands  
be- fore debugging the application:  
(gdb) set inline-debug inline_bp_individual  
or  
$ gdb --inline = inline_bp_individual  
To disable inline debugging, enter either of the following commands be- fore  
debugging the application:  
(gdb) set inline-debug off  
or  
$ gdb --inline= off  
3. You can use the following commands for debugging inline functions in Integrity  
systems:  
step  
next  
list  
backtrace  
frame <n>  
info locals  
info args  
breakpoint  
The following commands are not available for debugging inline functions in  
Integrity systems:  
info frame  
disassembly  
14.9 Debugging Macros  
HP WDB 5.7 and later versions of the debugger enable you to display and evaluate  
macro definitions for programs running on Integrity systems. This feature is available  
only for compiler versions A.06.15 and later.  
14.9.1 Viewing and Evaluating Macro Definitions  
HP WDB 5.7 and later versions of the debugger provide the following support for de-  
bugging macros:  
Displaying Macro Definitions  
Displaying Macro Definitions HP WDB provides the following commands to  
display macro definitions:  
14.9 Debugging Macros 153  
   
- show macro [macro-name] or info macro [macro-name]  
Displays the macro definition, source file name, and the line number. For example:  
(gdb) info macro VAR2  
Defined at scope.c:21  
#define VAR2 201  
- macro expand [macro-name]  
Expands the macro and the parameters in the macro. If there are any parameters  
in the macro, they are substituted in the macro definition when the definition is  
displayed.  
For example:  
#define YY 6  
#define MAC (67 + YY)  
...  
$ gdb  
...  
(gdb) macro expand MAC  
expands to: (67 + 6)  
Evaluating Macros  
HP WDB enables you to evaluate a macro and display the output. You can evaluate  
the macro by using the commonly used gdb commands for evaluating and  
displaying expressions, such as print. HP WDB supports the evaluation of macros  
with variables, constants, complex algebraic expressions involving variables, nested  
macros, and function calls. HP WDB does not support the evaluation of macros  
with multiple statements in the macro definitions, or the evaluation of macros  
with stringifying and pasting tokens in the macro definitions.  
14.9.1.1 Compiler Options to Enable Macro Debugging  
To enable macro debugging, the program must be compiled with the  
+macro_debug=[all|none|ref]compiler option.  
Additionally, the program must be compiled with one of the -goptions (-g, -g0, or  
-g1) to enable macro debugging. For example:  
cc -g +macro_debug=all -o sample sample.c  
The following options are available for the +macro_debug compiler option:  
all  
To view and evaluate all the macro expressions in the program, you must  
compile the program with +macro_debug=all. This option can cause a  
significant increase in object file size.  
ref  
To view and evaluate only the reference macros in the program, you must  
compile the program with +macro_debug=ref. This is the default for -g,  
-g0, or -g1.  
none To disable macro debugging, you must compile the program with +macro_  
debug=none  
154 HP-UX Configuration-Specific Information  
 
The macro debugging features are supported for +objdebugand +noobjdebug  
compiler options.  
14.9.2 Examples for Macro Debugging  
The following example illustrates the use of the macro debugging:  
Sample Program:  
$ cat scope.c  
1
2 #include <stdio.h>  
3
4 #define USED1 100  
5 #define USED2 200  
6 #define UNUSED1 0  
7 #define UNUSED2 0  
8 #define DUMMY1  
9 #define DUMMY2  
10  
11 int  
12 main ()  
13 {  
14  
15 int val = USED1;  
16  
17 #undef UNUSED1  
18 #undef USED2  
19 #undef USED1  
20 #define USED1 101  
21 #define USED2 201  
22  
23 val = USED1 + USED2;  
24  
25 #undef USED1  
26 #undef UNUSED2  
27 #undef USED2  
28 #define USED1 102  
29  
30 val = USED1;  
31  
32 return 0;  
33 }  
Sample Debugging Session  
The following debugging session illustrates macro debugging when the program  
is com- piled with +macro_debug=alloption:  
$ cc -g +macro_debug=all -o sc scope.c  
$ gdb sc  
HP gdb for HP Itanium (32 or 64 bit) and target HP-UX 11.2x.  
Copyright 1986 - 2001 Free Software Foundation, Inc.  
...  
14.9 Debugging Macros 155  
 
(gdb) b 13  
Breakpoint 1 at 0x40007d0:0: file scope.c, line 13 from sc.  
(gdb) b 23  
Breakpoint 2 at 0x40007d0:2: file scope.c, line 23 from sc.  
(gdb) b 30  
Breakpoint 3 at 0x40007e0:0: file scope.c, line 30 from sc.  
(gdb) r  
Starting program: sc  
Breakpoint 1, main () at scope.c:13  
13 {  
(gdb) print USED1  
100  
(gdb) print USED1+10  
110  
(gdb) info macro USED1  
Defined at scope.c:4  
#define USED1 100  
(gdb) info macro USED2  
Defined at scope.c:5  
#define USED2 200  
(gdb) c  
Continuing.  
Breakpoint 2, main () at scope.c:23  
23 val = USED1 + USED2;  
(gdb) info macro USED1  
Defined at scope.c:20  
#define USED1 101  
(gdb) info macro USED2  
Defined at scope.c:21  
#define USED2 201  
(gdb) c  
Continuing.  
Breakpoint 3, main () at scope.c:30  
30 val = USED1;  
(gdb) info macro USED1  
Defined at scope.c:28  
#define USED1 102  
(gdb) info macro USED2  
The macro `USED2' has no definition in the current scope.  
(gdb)  
The following debugging session illustrates macro debugging when the program  
is com- piled with +macro_debug=referencedoption:  
$ cc -g +macro_debug=referenced -o sc1 scope.c  
$ gdb sc1  
HP gdb for HP Itanium (32 or 64 bit) and target HP-UX 11.2x.  
Copyright 1986 - 2001 Free Software Foundation, Inc.  
...  
(gdb) b 13  
Breakpoint 1 at 0x40007d0:0: file scope.c, line 13 from sc1.  
(gdb) b 23  
156 HP-UX Configuration-Specific Information  
Breakpoint 2 at 0x40007d0:2: file scope.c, line 23 from sc1.  
(gdb) b 30  
Breakpoint 3 at 0x40007e0:0: file scope.c, line 30 from sc1.  
(gdb) r  
Starting program: sc1  
Breakpoint 1, main () at scope.c:13  
13 {  
(gdb) print USED1  
100  
(gdb) print USED1+10  
110  
(gdb) info macro USED1  
Defined at scope.c:4  
#define USED1 100  
(gdb) info macro USED2  
The macro `USED2' has no definition in the current scope.  
(gdb) c  
Continuing.  
Breakpoint 2, main () at scope.c:23  
23 val = USED1 + USED2;  
(gdb) info macro USED1  
Defined at scope.c:20  
#define USED1 101  
(gdb) info macro USED2  
Defined at scope.c:21  
#define USED2 201  
(gdb) c  
Continuing.  
Breakpoint 3, main () at scope.c:30  
30 val = USED1;  
(gdb) info macro USED1  
Defined at scope.c:28  
#define USED1 102  
(gdb) info macro USED2  
The macro `USED2' has no definition in the current scope.  
(gdb) q  
The program is running. Exit anyway? (y or n) y.  
14.10 Debugging Memory Problems  
You can use WDB to nd leaks, profile heap usage and detect other heap-related errors  
in HP C, HP aC++, and HP Fortran programs written for HP-UX 11.x systems. (Both  
32-bit and 64-bit programs are supported.)  
On HP-UX 11.x, the memory debugging features of WDB work with both  
single-threaded and multi-threaded programs that use POSIX threads.  
14.10 Debugging Memory Problems 157  
 
For more information on memory debugging with WDB, see the Debugging Dynamic  
Memory Usage Errors Using HP WDB whitepaper at the HP WDB Documentation  
webpage at: http://www.hp.com/go/wdb.  
14.10.1 When to suspect a memory leak  
You should suspect a memory leak in the code when you notice that the system is  
running out of swap space or running slower, or both.  
Applications or non-kernel code (including daemons) that have memory leaks can  
eventually use up all swap space. You can run top(1)to verify whether the process  
data space (SIZE, RES) is growing more than you expect.  
If the system is running out of swap space, programs will fail with out-of-memory  
(ENOMEM) errors or SIGBUSsignals. In addition, the system might run slower and  
slower until it comes to a stop; all processes requiring swap to continue running will  
wait for it indefinitely. GDB allows you to catch out-of-memory conditions through  
runtime memory checking. Use the command catch nomemto detect out-of-memory  
conditions. GDB will stop whenever malloc returns NULLand allows you to look at  
the current context.  
14.10.2 Memory debugging restrictions  
Programs with these attributes are not supported:  
CMA or DCE threaded programs on 11.x (32-bit and 64-bit)  
Memory checking features. These features work only in programs that directly or  
indirectly call malloc, realloc, free, mmap, or munmap from the standard C library  
libc.sl.  
Programs that link the archive version of the standard C library, libc.a, or the  
core library, libcl.a, on HP-UX 11.x  
NOTE: Linker with version number B.11.19 or higher is required for debugging  
memory problems.  
From HP WDB 5.7 onwards, the archive version of the run time check library,  
librtc.a, is not available. You must use the shared version of the library,  
librtc.[sl|so], instead.  
14.10.3 Memory Debugging Methodologies  
WDB enables you to debug memory problems in applications written for HP-UX 11.x  
or later using C, aC++, FORTRAN 77, and Fortran 90.  
WDB provides several commands that help expose memory-related problems.  
158 HP-UX Configuration-Specific Information  
     
HP WDB offers the following memory-debugging capabilities:  
Reports memory leaks  
Reports heap allocation profile  
Stops program execution if bad writes occur with string operations such as strcpy  
and memcpy  
Stops program execution when freeing unallocated or de-allocated blocks  
Stops program execution when freeing a block if bad writes occur outside block  
boundary  
Stops program execution conditionally based on whether a specified block address  
is allocated or de-allocated  
Scrambles previous memory contents at malloc()and free()calls  
Simulates and detects out-of-memory event errors  
Detects dangling pointers and dangling blocks  
Detects in-blockcorruption of freed blocks  
Specifies the amount of guard bytes for every block of allocated memory  
Displays the run time type information for C++ polymorphic objects  
You can use any of the following methods to identify memory problems:  
14.10.4 Debugging Memory in Interactive Mode  
This section describes the various commands which help in debugging memory  
problems when the debugger is used in the interactive mode.  
14.10.4.1 Commands for interactive memory debugging  
To debug memory problems, use these commands:  
set heap-check [on|off]  
This toggles the capability for detection of leaks,  
heap profiles, bounds checking, and checking for  
double free.  
info heap  
Displays a heap report, listing information such  
as the start of heap, end of heap, heap size, heap  
allocations, size of blocks, and number of  
instances. The report shows heap usage at the  
point you use the info heapcommand. The  
report does not show allocations that have  
already been freed. For example, if you make  
several allocations, free them all, and then you  
14.10 Debugging Memory Problems 159  
   
use info heap, the result does not show any  
allocations.  
info heapfilename  
Writes heap report output to the specified file.  
info heapidnumber  
Produces detailed information on the specified  
heap allocation including the allocation call stack.  
show heap-check  
Displays all current settings for memory  
checking.  
set heap-check interval <  
nn >  
This command starts incremental heap growth  
profile. All allocations prior to the execution of  
this command are ignored from report. If  
incremental heap growth profile is already on,  
executing this command will reset the counters  
and start a fresh collection. Interval is specified  
in seconds.  
set heap-check repeat < nn This allows user to specify the number of  
intervals GDB should collect the incremental  
heap growth. The default value is 100. Every  
repeat of the interval tracks heap allocation  
during that interval.  
>
Example:  
< gdb > set heap-check interval 10  
< gdb > set heap-check repeat 100  
Here WDB will create 100 incremental apart heap  
profiles which are 10 seconds apart.  
set heap-check reset  
GDB stores incremental heap growth data in an  
internal file. During one session data is appended  
to this file. If the session is very long, it is possible  
that this file may become very large. This  
command discards the data existing in the file  
and creates a new data file. Once this command  
is executed, the user cannot see the old data.  
info heap-interval < file  
name >  
This command creates the report of heap growth.  
The data for each interval has the start and end  
time of the interval. If file name is mentioned a  
detailed report is written in the file.  
set heap-check leaks [on | Controls WDB memory leak checking.  
off]  
info leaks  
Displays a leak report, listing information such  
as the leaks, size of blocks, and number of  
instances.  
160 HP-UX Configuration-Specific Information  
info leaksfilename  
Writes the complete leak report output to the  
specified file.  
info leakleaknumber  
Produces detailed information on the specified  
leak including the allocation call stack.  
set heap-check block-size  
num-bytes  
Instructs WDB to stop the program whenever it  
tries to allocate a block larger than num-bytes in  
size.  
set heap-check heap-size  
num-size  
Instructs WDB to stop the program whenever it  
tries to increase the program heap by at least  
num-bytes.  
min-heap-size  
This option reports the heap allocations that  
exceed the specified number <num> of bytes  
based on the cumulative number of bytes that  
are allocated at each call-site, which is inclusive  
of multiple calls to mallocat a particular call  
site.  
set heap-check watch  
address  
Stops the program whenever a block at a  
specified address is allocated or deal- located.  
set heap-check null-check  
[N | random]  
Force mallocto return NULLafter N or random  
number of invocations of malloc. Use set  
heap-check random-range Nand set  
heap-check seed-value Nto define the seed  
value and range for random number calculation.  
Useful for simulating out-of-memory situations.  
set heap-check  
null-check-size N  
Force mallocto return NULLafter N bytes have  
been allocated by the program.  
catch nomem  
Allows the user to get control of an  
out-of-memory event. The user can step through  
once the nomemevent is caught to see whether  
the out-of-memory event is handled correctly.  
set heap-check free [on |  
off]  
When set to on, forces WDB to stop the program  
when it detects a call to free()with an  
improper argument, a call to realloc()that  
does not point to a valid currently allocated heap  
block, or a call to free a block of memory, which  
has been corrupted. Refer set heap-check  
boundsfor details on detecting memory  
corruption.  
set heap-check string [on  
| off]  
Toggles validation of calls to strcpy, strncpy,  
memcpy, memccpy, memset, memmove, bzero,  
14.10 Debugging Memory Problems 161  
and, bcopy. WDB 5.6 and later versions of the  
debugger also validates calls to strcatand  
strncat.  
NOTE: HP WDB 6.0 and later versions of the  
debugger improves performance of memory  
debugging when the stringoption is set for  
32-bit applications on HP-UX 11i v2 and later on  
Itanium systems. However, the performance  
degrades for 64-bit applications on HP-UX 11i  
v2 and later on Itanium systems.  
set heap-check bounds [on  
| off]  
Allocates extra space at the beginning and end  
of a heap block during allocation and fills it with  
a specific pattern. When blocks are freed, WDB  
verifies whether these patterns are intact. If they  
are corrupted, an under flow or over- flow must  
have occurred and WDB reports the problem.  
This option increases the program's memory  
requirements.  
set heap-check scramble [on Scrambles a memory block and overwrites it with  
a specific pattern when it is allocated or  
deallocated. This change to the memory contents  
increases the chance that erroneous behaviors,  
such as attempting to access space that is freed  
or depending on initial values of malloc()  
blocks, cause the program to fail.  
| off]  
info dangling  
Displays a list of all the dangling pointers and  
dangling blocks that are potential sources of  
memory corruption( may have false positives).  
info corruption  
Checks for corruption in the currently allocated  
heap blocks.In addition, it lists the potential  
in-blockcorruptions in all the freed blocks.  
set heap-check  
min-leak-sizenum  
Collects a stack trace only when the size of the  
leak exceeds the number of bytes you specify for  
this value. Larger values improve run-time  
performance. The default value is zero (0) bytes.  
set heap-check frame-count Controls the depth of the call stack collected.  
Larger values increase run time. The default  
value is four (4) stack frames.  
num  
set heap-check header-size Sets the Header guard for each block of the  
allocated memory. The default number of bytes  
num of bytes  
162 HP-UX Configuration-Specific Information  
for the footer is 16 bytes if this option is not used.  
If the user specifies a value less than 16 for the  
number of bytes, the debugger ignores it and  
takes the default of 16 bytes. If the user specifies  
more than 16 bytes, then the debugger considers  
the largest and closest 16 byte integral from the  
user-specified value.  
Example:  
If the user specifies 60 bytes, the debugger takes  
it as 48 bytes. If the user specifies 65, the  
debugger considers 64 bytes.  
set heap-check footer-size Sets the Footer guard for each block of the  
allocated memory. The default number of bytes  
for the footer is one byte if this option is not used.  
num of bytes  
14.10.4.2 Example for interactive debugging session  
This example describes checking a program running on HP-UX 11.x using linker version  
B.11.19 or later:  
1. Link the program with /usr/lib/libc.slinstead of libc.a.  
2. Run the debugger and load the program:  
> gdb a.out  
3. Turn on leak checking:  
(gdb) set heap-check leaks on  
4. Set one or more breakpoints in the code where you want to examine cumulative  
leaks:  
(gdb) b myfunction  
5. Run the program in the debugger:  
(gdb) run  
6. Use the infocommand to show list of memory leaks:  
(gdb) info leaks  
Scanning for memory leaks...done  
2439 bytes leaked in 25 blocks  
No. Total bytes Blocks Address Function  
0 1234 1 0x40419710 foo()  
1 333 1 0x40410bf8 main()  
2 245 8 0x40410838 strdup()  
[...]  
14.10 Debugging Memory Problems 163  
 
The debugger assigns each leak a numeric identifier.  
7. To display a stack trace for a specific leak, use the info leakcommand and  
specify the number from the list associated with a leak:  
(gdb) infoleak 2  
245 bytes leaked in 8 blocks (10.05% of all bytes leaked)  
These range in size from 26 to 36 bytes and are allocated  
in strdup ()  
in link_the_list () at test.c:55  
in main () at test.c:13  
in _start ()  
14.10.5 Debugging Memory in Batch Mode  
HP WDB supports batch mode memory leak detection, also called batch Run Time  
Checking (Batch RTC). Most of the memory debugging features supported in interactive  
mode are also supported in batch mode.  
NOTE: The batch mode commands may not always work when invoked through a  
shell script.  
14.10.5.1 Setting Configuration Options for Batch Mode  
You can specify the batch mode configuration through a configuration file called  
rtcconfig. The configuration file supports these variables:  
check_free=on|off(or) set  
heap-check free <on/off>  
Enables detection of multiple incorrect free of  
memory  
check_heap|heap=on|off (or) Enables heap profiling.  
set heap-check <on/off>  
check_leaks|leaks=on|off  
(or) set heap-check leaks  
<on/off>  
Enables leak detection.  
check_string=on|off (or)  
set heap-check string  
<on/off>  
Enables detection for writing out of boundary  
for strcpy, strncpy, memcpy, memccpy,  
memset, memmove, bzero, bcopy.  
check_bounds|bounds=on|off Enables checking of bounds corruption.  
(or) set heap-check bounds  
<on/off>  
files=<file1:file2:...|fileN> Specifies the executables for which memory leak  
detection is enabled. If files option is not  
specified, after setting BATCH_RTC=on, RTC will  
instrument ALL executables.  
164 HP-UX Configuration-Specific Information  
   
frame_count=no_frames (or) Sets the number of frames to be printed for leak  
context.  
set heap-check frame-count  
<no_frames>  
min_heap_size=block_size  
(or) set heap-check  
min-heap-size <block_size>  
Sets the minimum block size to use for heap  
reporting.  
min_leak_size=block_size  
(or) set heap-check  
min-leak-size <block_size>  
Sets the minimum block size to use for leak  
detection.  
output_dir=output_data_dir Specifies the name of the output data directory.  
scramble_blocks=on|off (or) Enables block scrambling.  
set heap-check scramble  
<on/off>  
Batch mode leak detection stops the application at the end, when libraries are being  
unloaded, and invokes HP WDB to print the leak or heap data.  
NOTE: It is incorrect usage to use spaces before or after the '=' symbol in the batch  
mode configuration options in the configuration file, rtcconfig. Additionally, it is  
incorrect usage to use spaces before the batch mode configuration options.  
For example:  
Correct Usage:  
$ cat rtcconfig  
check_leaks=on  
check_heap=on  
files=batchrtc4  
$
Incorrect Usage:  
$ cat rtcconfig  
check_leaks=on  
check_heap = on  
files=batchrtc4  
$
Steps for Batch Mode Memory Debugging  
To use batch memory debugging, complete the following steps:  
1. Compile the source files.  
2. Create an rtcconfigfile in the current directory.  
3. Dene an environment variable BATCH_RTC. If you are using the Korn or Posix  
shell, enter the following command at the HP-UX prompt:  
export BATCH_RTC=on  
14.10 Debugging Memory Problems 165  
4. Complete one of the following steps to preload the librtcruntime library:  
Set the target application to preload librtc by using the +rtcoption for the  
chatrcommand. In addition to automatically loading the librtclibrary,  
the +rtcoption for the chatrcommand also maps the shared libraries as  
private. To enable or disable the target application to preload the librtc runtime  
library, enter the following command at the HP-UX prompt:  
$ chatr +rtc <enable|disable> <executable>  
NOTE: The chatr+rtcoption preloads the librtc runtime library from the  
following default paths:  
– For 32 bit IPF applications,  
/opt/langtools/lib/hpux32/librtc.so  
– For 64 bit IPF applications,  
/opt/langtools/lib/hpux64/librtc.so  
– For 32 bit PA applications,  
opt/langtools/lib/librtc.sl  
– For 64 bit PA applications,  
/opt/langtools/lib/pa20_64/librtc.sl  
To preload the librtc runtime library from a path that is different from the  
default paths, you must use the LD_PRELOADenvironment variable.  
Instead of automatically preloading librtc and mapping the shared libraries,  
you can explicitly preload the required librtc library after mapping the shared  
libraries private.  
In the case of HP 9000 systems, you must explicitly map the share libraries as  
private by using the +dbg enableoption for the chatrcommand, as follows:  
$ chatr +dbg enable ./<executable>  
(This step is not required on Integrity systems.)  
To explicitly preload the librtcruntime library and start the target  
application, enter one of the following commands:  
– For 32-bit IPF applications,  
LD_PRELOAD=/opt/langtools/lib/hpux32/librtc.so <executable>  
– For 64-bit IPF applications,  
LD_PRELOAD=/opt/langtools/lib/hpux64/librtc.so <executable>  
– For 32-bit PA applications,  
166 HP-UX Configuration-Specific Information  
LD_PRELOAD=/opt/langtools/lib/librtc.sl <executable>  
– For 64-bit IPF applications,  
LD_PRELOAD=/opt/langtools/lib/pa20_64/librtc.sl <executable>  
If LD_PRELOADand chatr +rtcare used to preload the librtc runtime library  
, the librtcruntime library is loaded from the path specified by  
LD_PRELOAD.  
NOTE: Batch Mode RTC displays one of the following errors and causes the  
program to temporarily hang if the version of WDB and librtc.[sl|so]  
do not match, or if WDB is not available on the system:  
"/opt/langtools/bin/gdb: unrecognized option `-brtc'  
Use `/opt/langtools/bin/gdb --help' for a complete list of options."  
OR  
"execl failed. Cannot print RTC info: No such file or directory"  
This error does not occur under normal usage where WDB or  
librtc.[sl|so] is used from the default location at /opt/langtools/  
...However, this error occurs if GDB_SERVERand/or LIBRTC_SERVERare  
set to a mismatched version of WDB or librtc.[sl|so]respectively.  
5. At the end of the run, output data file is created in output_data_dir, if defined  
in rtcconfig, or the current directory. HP WDB creates output data file for each  
run. It creates a separate file for leak detection and heap information. The naming  
convention for output files is as follows:  
<file_name>.<pid>.<suffix>  
where, <pid>is the process id and the value for <sux>can be either leaks, heap,  
or mem.  
NOTE: During operations such as system(3s)and popen(), which invoke a  
new shell, librtc.sl|somust not be loaded to the invoked shell. You must use  
LD_PRELOAD_ONCE, instead of LD_PRELOAD, to exclusively load the  
librtc.sl|sofile to the calling process only. Following is the syntax for using  
LD_PRELOAD_ONCE:  
LD_PRELOAD_ONCE= /opt/langtools/lib/librtc.sl  
14.10.5.2 Environment variable setting for Batch mode debugging  
Batch mode memory leak detection uses the following environment variables:  
GDBRTC_CONFIGspecifies the location of rtcconfiguration file. If this option is  
not specified, the configuration file is assumed to be in the current location, and  
14.10 Debugging Memory Problems 167  
 
has the filename rtcconfig. If user prefers to set this option, it must include the  
filename.  
Incorrect usage:  
export GDBRTC_CONFIG=./  
export GDBRTC_CONFIG=/tmp  
Correct usage:  
export GDBRTC_CONFIG=/tmp/yet_another_config  
export GDBRTC_CONFIG=/tmp/rtcconfig  
BATCH_RTCenables or disables batch memory leak detection.  
GDB_SERVERis used to override the default path from where the gdb executable  
is used to provide the information on memory leak. By default, /opt/langtools/  
bin/gdbis used to print the output. This can be overridden by setting  
GDB_SERVERappropriately.  
RTC_MALLOC_CONFIGis used to override the default configand rtcconfig  
file settings. This variable can be set as follows:  
export RTC_MALLOC_CONFIG=config_string1[;config_strings].  
The config_strings are separated by a semi-colon(;).  
The following config_strings options exist for RTC_MALLOC_CONFIG:  
abort_on_bounds=[01]  
abort_on_bad_free=[01]  
abort_on_nomem=[01]  
Aborts execution when heap objects  
bounds check fail, value is 1, and the  
environment variable RTC_NO_ABORT  
is not set.  
Aborts execution when free or realloc is  
trying to free a heap object which is not  
valid, value is 1, and environment  
variable RTC_NO_ABORTis not set.  
Aborts execution when out of memory  
if value is 1, and environment variable  
RTC_NO_ABORTis not set.  
leak_logfile=stderr[+]filename The log file for batch mode must be  
specified.  
stderr: error message goes to stderr  
[+]filename: error message goes to  
filename, + means output is appended to  
the file.  
168 HP-UX Configuration-Specific Information  
mem_logfile=stderr[+]filename  
heap_logfile=stderr[+]filename  
Specify config_stringsfor +check=mallocon Itanium or WDB memory  
check batch mode on Integrity systems.  
RTC_PROCESS_GDBINITis an optional environment variable used to enable  
processing of the .gdbinitfile. You can use the .gdbinitfile to specify path  
settings such as dir, objectdir, and pathmap to set the path of the source and object  
files in case the source or object paths are different than the current directory, so  
that the generated RTC reports display the symbol names and line numbers  
correctly. This feature is optionally enabled only when the RTC_PROCESS_GDBINIT  
environment variable is set to 1.  
There are limitations on whatcommands in the .gdbinitfile. If there are erroneous  
commands in the .gdbinitfile, the batch RTC session can possibly hang and not  
produce the expected RTC reports. Following are some examples:  
1. If the 'q' (quit) command is used, the session would hang and finally terminate  
after approximately 10 mins, and not generate any RTC reports. It would print  
the error message "Broken synchronization between child/parent process".  
2. If any gdb command that would take more processing time is used, this would  
interfere with the assumptions of RTC and the session may hang and print  
the error message "Broken synchronization between child/parent process".  
14.10.5.3 Example for Batch Mode RTC  
This section illustrates examples of batch mode RTC on HP 9000 and Integrity Systems.  
Example for Batch Mode RTC of a PA-RISC 32 bit Executable  
1. Compile the source files.  
2. The rtcconfigfile must contain entries, such as the following:  
check_heap=on  
check_leaks=on  
check_free=on  
files=executable_name  
output_dir=/tmp/results  
3. Set the following environment variables as follows:  
export BATCH_RTC=on  
14.10 Debugging Memory Problems 169  
 
4. Complete one of the following steps:  
– Map the shared libraries privately using chatr, as follows:  
chatr +dbg enable <executable>  
On HP-UX 11i v3 Integrity systems, WDB enables automatic debugging  
of shared libraries without them being mapped private while attaching to  
a running program. For enabling automatic debugging of shared libraries,  
you must install the kernel patches PHKL_38651 and PHKL_38778.  
Explicitly preload the librtcruntime library and start the target  
application, as follows:  
LD_PRELOAD=/opt/langtools/lib//librtc.sl <executable>  
<arguments>  
OR  
– Set the target application to preload librtcby using the +rtcoption for  
the chatrcommand. In addition to automatically loading the librtc  
library, the +rtcoption for the chatrcommand also maps the shared  
libraries as private.  
To set the target application to preload the librtcruntime library, enter  
the following command at the HP-UX prompt:  
$ chatr +rtc enable <executable>  
Example for Batch Mode RTC on Integrity Systems: Example-1  
1. Compile the source files.  
2. The rtcconfigfile must contain entries such as the following:  
check_heap=on  
check_leaks=on  
check_free=on  
files=executable_name  
output_dir=/tmp/results  
3. Set the following environment variables as follows:  
export BATCH_RTC=on  
4. Complete one of the following steps:  
– Preload the librtcruntime library, as follows:  
PRELOAD=/opt/langtools/lib/hpux32/librtc.so <executable>  
<arguments if any>  
OR  
– Preload the librtcruntime library, as follows:  
$ chatr +rtc enable <executable>  
Example from Integrity Systems: Example-2  
1. Compile the source files.  
170 HP-UX Configuration-Specific Information  
2. The rtcconfig file should contain entries such as the following:  
check_heap=on  
check_leaks=on  
check_free=on  
files=exec1:exec2:exec3  
output_dir=/tmp/results  
3. Set the following environment variables as follows:  
export BATCH_RTC=on  
4. Complete one of the following steps:  
— Use the +rtcoption for the chatrcommand on each of the required  
executable files that must be instrumented, as follows:  
$ chatr +rtc enable exec1 exec2 exec3  
OR  
— Preload the librtclibrary as follows:  
export LD_PRELOAD /opt/langtools/lib/hpux32/librtc.so  
5. Run the program as follows:  
./exec1  
Suppose exec1eventually spawns exec2, exec3, exec4, exec5, only  
exec1, exec2, exec3will be instrumented based on the settings in the  
rtcconfigfile.  
14.10.6 Debugging Memory Interactively After Attaching to a Running Process  
HP WDB accepts -pidor -pfollowed by a process ID to attach a running process to  
the debugger.  
To use Batch RTC after attaching GDB to a running process, complete the following  
steps:  
14.10 Debugging Memory Problems 171  
 
1. Complete one of the following steps to preload the librtc runtime library:  
Set the target application to preload librtcby using the +rtcoption for the  
chatrcommand.In addition to automatically loading the librtc library, the  
+rtcoption for the chatrcommand also maps the shared libraries as private.  
To enable or disable the target application to preload the librtcruntime  
library, enter the following command at the HP-UX prompt:  
$ chatr +rtc <enable|disable> <executable>  
NOTE: To attach and nd leaks for PA-32 applications from the startup, the  
environment variable RTC_INIT should be set to on in addition to preloading  
the librtc.[sl|so]library before starting the application, as follows:  
$ RTC_INIT=on <executable>  
OR  
Instead of automatically preloading librtcand mapping the shared libraries,  
you can explicitly preload the required librtc library.  
In the case of HP 9000 systems, you must explicitly map the share libraries as  
private by using the +dbgenable option for the chatrcommand, as follows:  
$ chatr +dbg enable ./<executable>  
(This step is not required on Integrity systems.)  
To explicitly preload the librtcruntime library and start the target  
application, enter one of the following commands:  
For 32 bit IPF applications:  
LD_PRELOAD=/opt/langtools/lib/hpux32/librtc.so <executable>  
For 64 bit IPF applications:  
LD_PRELOAD=/opt/langtools/lib/hpux64/librtc.so <executable>  
For 32 bit PA applications:  
LD_PRELOAD=/opt/langtools/lib/librtc.sl <executable>  
For 64 bit PA applications:  
LD_PRELOAD=/opt/langtools/lib/pa20_64/librtc.sl <executable>  
NOTE: To attach and nd leaks for PA-32 applications from the startup, the  
environment variable RTC_INITshould be set to onin addition to preloading  
the librtc.[sl|so]library before starting the application, as follows:  
$ LD_PRELOAD=/opt/langtools/lib/librtc.sl RTC_INIT=on <application>  
If RTC_INITis turned on, librtc starts recording heap information for PA32  
process, by default. Hence, you must set this environment variable only when  
it is required, and you must not export this environment variable for shell.  
172 HP-UX Configuration-Specific Information  
2. Run the program.  
3. Start a debugging session as follows:  
gdb -leaks <executable-name> <process-id>  
4. Use info heapand info leakscommands to obtain a memory analysis report  
of the application.  
NOTE: From HP WDB 5.7 onwards, the archive version of the run time check library,  
librtc.a, is not available. You must use the shared version of the library,  
librtc.[sl|so], instead.  
14.10.7 Configuring memory debugging settings  
The following configuration settings are supported to control the level of details of  
information required to be displayed when debugging memory leaks.  
14.10.7.1 Specifying the stack depth  
Memory debugging reduces the performance of an application by 20-40% because of  
stack unwinding. To provide a clear profile of every allocation in the program, the  
debugger collects the stack trace information for every allocation in the debugged  
application. Reducing the stack depth (the number of stack frames that the debugger  
collects for each allocation) reduces the performance degradation.  
The set heap-check frame-count command enables you to control the depth  
of the stack frames that are collected by WDB for each allocation. By default, four stack  
frames are displayed from the allocating call stack.  
To set the depth of the stack frames that is collected by WDB, enter the following  
command at the gdb prompt:  
$ set heap-check frame-count [num]  
The stack depth, [num], is the number of stack frames that WDB collects for each  
allocation.  
You can specify a higher value for [num] to view more stack frames for each reported  
allocation. However, the performance of the application is reduced because of the  
increased stack depth.  
14.10.7.2 Specifying minimum leak size  
WDB enables you to specify the minimum leak size for stack trace collection to improve  
the program's performance.  
Stack trace collection slows down the program because it occurs on every allocation  
call. Therefore, if the program is allocation-intensive, WDB can spend a substantial  
amount of time collecting stack traces.  
You can improve performance by using the command:  
14.10 Debugging Memory Problems 173  
     
set heap-check min-leak-size num  
For example, if you use,  
set heap-check min-leak-size 100  
WDB does not collect stack traces for allocations smaller than 100 bytes. HP WDB still  
reports leaks smaller than this size, but does not include a stack trace.  
14.10.7.3 Specifying minimum block size  
The min-heap-sizeoption reports the heap allocations that exceed the specified  
number <num> of bytes based on the cumulative number of bytes that are allocated at  
each call-site, which is inclusive of multiple calls to malloc at a particular call site. For  
example:  
set heap-check min-heap-size 100  
When the option min-heap-sizeis set to 100, GDB reports all the cumulative block  
allocations that 100 bytes at each call-site.  
14.10.8 Scenarios in memory debugging  
14.10.8.1 Stop when freeing unallocated or deallocated blocks  
WDB enables you to locate improper calls to free()and realloc(), with the  
command:  
In interactive debugging mode: set heap-check free [on | off].  
In batch mode debugging: check_free [on | off].  
With this setting on, whenever the program calls free()or realloc()WDB inspects  
the parameters to verify that they are pointing to valid currently allocated heap blocks.  
If WDB detects an erroneous call to free(), it stops the program and reports this  
condition. You can then look at the stack trace to understand where and how the  
problem occurred.  
14.10.8.2 Stop when freeing a block if bad writes occurred outside block boundary  
WDB enables you to locate problems caused by heap block over flow or underflow  
with the command  
In Interactive debugging mode: set heap-check bounds [on | off]  
In batch mode debugging: check_bounds [on | off]  
When bounds checking is turned on, WDB allocates extra space at the beginning and  
end of a block during allocation and fills it up with a specific pattern. When blocks are  
freed, WDB verifies whether these patterns are intact. If they are corrupted, an under  
flow or over flow must have occurred and WDB reports the problem. If you want to  
nd corruption at any time, use the info corruptioncommand.  
The info corruptioncommand can detect memory corruption in an application.  
That is, it reports all the memory blocks that have over-writes and under-writes.  
174 HP-UX Configuration-Specific Information  
       
Syntax:  
info corruption [<file name>]  
The run time memory checking must be enabled before using the info corruption  
command to detect memory corruption. The corruption information is written to a file  
specified in the .filename argument if provided. Otherwise, it is printed to the  
stdout.  
NOTE: Turning on bounds checking increases the program's memory requirements  
because the extra guard bytes must be allocated at the beginning and end of each block.  
14.10.8.3 Stop when a specified block address is allocated or deallocated  
To stop a program whenever a block at a specified address is allocated or deallocated,  
use the command:  
set heap-check watch address  
You can use this to debug situations such as multiple free() calls to the same block.  
Limitation : This is not supported in batch mode debugging.  
14.10.8.4 Scramble previous memory contents at malloc/free calls  
WDB enables you to potentially expose latent memory access defects with the command:  
In Interactive debugging mode: set heap-check scramble [on | off]  
In batch mode debugging: scramble_blocks [on | off].  
When this setting is turned on, any time a memory block is allocated or deallocated,  
WDB scrambles the space and overwrites it with a specific pattern.  
This change to the memory contents increases the chance that erroneous behaviors will  
cause the program to fail. Examples of such behavior include attempting to access space  
that is freed or depending on initial values of malloc()blocks.  
You can now look at the stack trace to understand where and how the problem occurred.  
NOTE: Turning on scrambling slows down the program slightly, because at every  
malloc()and free()call, the space involved must be overwritten.  
14.10.8.5 Detect dangling pointers and dangling blocks  
A pointer is a Dangling pointer if the block of memory it points to, has been freed by  
the application. The block is called Dangling Block.  
The same freed block could be subsequently allocated to the application in response  
to another memory allocation request. In this scenario, if the application incorrectly  
tries to write into the freed memory block using the dangling pointer, it could result  
in incorrect or an undefined program behavior, as the new owner or function owning  
the same allocated block would nd different values in the heap block.  
14.10 Debugging Memory Problems 175  
     
NOTE: Software literature names this concept as premature free or Reading/writing freed  
memory using a pointer.  
WDB tracks the dangling pointers and dangling blocks using a modified version of  
Garbage collection. The enabler for doing this is by retaining all the freed blocks  
internally within RTC without actually freeing it as long as possible. It displays all the  
potential pointers to the freed dangling blocks, in the application data space.  
The pointers are potential because the pointers need not be actual pointers and could  
be a datum value and hence there are chances of false positives in the dangling report.  
NOTE: WDB tries to help as much as possible to detect if these pointers are of type  
datum or real pointers. In a -g compiled binary, WDB performs a look-up on a symbol  
table to nd the symbol name and type to nd the symbol name of the potential pointer  
and if its of pointer type, then the corresponding dangling block is really dangling(not  
a false positive).  
WDB turns on these checks, only when you specify set heap-check  
retain-freed- blocks on.  
14.10.8.6 Detect in-block corruption of freed blocks  
HP WDB detects all the attempts of a program to write to the freed dangling blocks  
using dangling pointers. We detect such in-block corruptions and are reported as part  
of the existing info corruptioncommand output.  
14.10.8.7 Specify the amount of guard bytes for every block of allocated memory  
HP WDB enables you to programmatically control the size of guard bytes for every  
block of the allocated memory. You can use these guard bytes to spot very rare and  
non-trivial boundary (buffer over-run and buffer under-run) corruptions. This again  
is available optionally when the user specifies set heap-check  
retain-freed-blocks <on>.  
14.10.9 Comparison of Memory Debugging Commands in Interactive Mode and Batch  
Mode  
HP WDB 5.6 and later versions provide consistency in format for the batch mode options  
and the interactive mode commands.  
The following table lists the memory debugging commands available in batch and  
interactive mode:  
176 HP-UX Configuration-Specific Information  
     
Table 14-1 Memory Debugging Commands in Interactive and Batch Mode  
Command Description  
Interactive mode  
Batch mode  
Toggles heap profiling and detection set heap-check [on check_heap= [on | off] (or)  
of leaks, bounds, and double free  
| off]  
set heap-check [on | off]  
Toggle the leak detection capability set heap-check leaks [on check_leaks =[on | off](or)  
|off]  
set heap-check leaks [on  
|off]  
Toggle validation of calls to strcpy, set heap-check  
check_string= [on |off] (or)  
set heap-check string [on |  
off]  
strncpy, memcpy, memccpy,  
memset, memmove, bzero, and  
bcopy  
string [on | off]  
Toggle validation of calls to free() set heap-check free check_free = [on |off](or)  
[on |off]  
set heap-check free [on |  
off]  
Specify whether freed blocks should set heap-check  
scramble_blocks=[on| off](or)  
scramble [on |off] set heap-check scramble [on  
|off]  
be scrambled  
Toggle bounds check on heap blocks set heap-check  
check_bounds = [on | off]  
(or) set heap-check bounds  
[on| off]  
bounds [on| off]  
Specify the minimum size of a block set heap-check  
min_heap_size = <num>(or) set  
min-heap-size <num> heap-check min-heap-size  
<num>  
for stack trace collection. GDB will  
report blocks of size greater than or  
equal to <num>at each call-site.  
Specify the minimum size of a block set heap-check  
min_leak_size = <num>(or) set  
min-leak-size <num> heap-check min-leak-size  
<num>  
for stack trace collection. GDB report  
leaks of blocks, smaller than this  
value. However, no stack trace is  
provided.  
Specify the depth of call stack to be set heap-check  
captured  
frame_count = <num>(or) set  
heap-check frame-count <num>  
frame-count <num>  
Instruct GDB to stop, whenever a set heap-check watch Not supported in batch mode  
block at the given address gets  
allocated or deallocated.  
address  
Instructs mallocto return null after set heap-check  
<num>invocations of malloc. null-check <num>  
Not supported in batch mode  
Not supported in batch mode  
Instructs mallocto return null after set heap-check  
<num>bytes have been allocated by null-check-size  
malloc.  
<size>  
Specifies the seed value to be used set heap-check  
Not supported in batch mode  
for generating random  
seed-value <num>  
null-check-count.  
14.10 Debugging Memory Problems 177  
 
Table 14-1 Memory Debugging Commands in Interactive and Batch Mode (continued)  
Command Description  
Interactive mode  
Batch mode  
Specifies the random range to be  
used by random-range.  
set heap-check  
random-range <num>  
Not supported in batch mode  
Specifies the time interval to be used set heap-check  
Not supported in batch mode  
Not supported in batch mode  
for incremental memory profile.  
interval <num>  
Perform incremental profile for  
set heap-check  
<num>interval periods where each repeat <num>  
period duration is defined by set  
heap-check interval command. The  
default value is 100.  
NULL pointer return by memory catch nomem  
allocators; used with set  
heap-check on, with/without  
null-checkenabled.  
Not supported in batch mode  
NOTE: The files=<executable-name>and output_dir=<path>options are  
exclusive to batch mode debugging.  
14.10.10 Heap Profiling  
The heap profile is useful for identifying how memory is being used by the program.  
You can use WDB to profile an application's current heap usage.  
14.10.10.1 Commands for heap profiling  
info heap  
Displays a heap report, listing information such  
as the start of the heap, end of the heap, heap  
allocations, size of blocks, and number of  
instances. The report shows heap usage at the  
point you use the info heapcommand.  
The report does not show allocations that have  
already been freed. For example, if you make  
several allocations, free them all, and then you  
use info heap, the result does not show any  
allocations.  
info heapfilename  
Writes heap report output to the specified file.  
info heapidnumber  
Produces detailed information on the specified  
heap allocation including the allocation call stack.  
set heap-check frame-count Controls the depth of the call stack collected.  
Larger values increase run time. The default  
value is four (4) stack frames.  
num  
178 HP-UX Configuration-Specific Information  
   
show heap-check  
Displays all current settings for memory  
checking.  
14.10.10.2 info heap arena  
The info heap arenacommand enables the user to view high level memory usage  
details of each arena. The info heap arenais not supported in batch mode. This  
command is available only for applications running on 11i v3 or later.  
14.10.10.3 info heap arena [0 |1|2|..]blocks stacks  
Displays the memory profile report for block level and overall memory usage with  
stack trace where applicable. This command is available only for applications running  
on 11i v3 or later.  
14.10.10.4 info moduleADDRESS  
The info modulecommand identifies load modules, and determines whether it lies  
in the text or data region for a given address. This command is available only for  
applications running on 11i v3 or later.  
Syntax:  
info module ADDRESS  
14.10.10.5 info heap process  
The info heap process command enables the user to view a high level memory usage  
report of a process. This command is available only for applications running on 11i v3  
or later.  
14.10.10.6 Example for heap profiling  
This example shows how to use this feature on HP-UX 11.x:  
1. If the linker version is earlier than B.11.19, link with /opt/langtools/lib/  
pa20_ 64/librtc.slfor PA-64 programs. For a 32-bit program, you must link  
with /opt/langtools/lib/librtc.sl.  
If the dynamic linker version is B.11.19 or later, skip this step because WDB  
automatically loads the librtc.sl library.  
2. Turn on profiling with the set heap-check on command:  
(gdb) set heap-check on  
3. Set a breakpoint:  
(gdb) b myfunction  
14.10 Debugging Memory Problems 179  
         
4. When the program is stopped at a breakpoint, use the info heapcommand:  
(gdb) info heap  
Analyzing heap ...done  
Actual Heap Usage:  
Heap Start = 0x40408000  
Heap End = 0x4041a900  
Heap Size = 76288 bytes  
Outstanding Allocations:  
41558 bytes allocated in 28 blocks  
No. Total bytes Blocks Address Function  
0 34567 1 0x40411000 foo()  
1 4096 1 0x7bd63000 bar()  
2 1234 1 0x40419710 baz()  
3 245 8 0x404108b0 boo()  
[...]  
The display shows the currently allocated heap blocks. Any blocks that have been  
allocated and already freed are not listed.  
To look at a specific allocation, specify the allocation number with the info heap  
command:  
(gdb) info heap 1  
4096 bytes at 0x7bd63000 (9.86% of all bytes allocated)  
in bar () at test.c:108  
in main () at test.c:17  
in _start ()  
in $START$ ()  
14.10.11 Memory Checking Analysis for User Defined Memory Management Routines  
Many user applications have their own memory management routines. These custom  
allocator routines are either user defined or sometimes wrappers to the default memory  
management routines. Memory checking features has been extended for user defined  
memory management routines. Memory leak, profile of heap memory or memory  
errors can be obtained for user defined memory management routines.  
Limitations:  
This feature routes to default memory routines instead of calling user defined  
memory management routines to obtain memory analysis.  
This feature is not supported in batch mode and attach mode debugging.  
14.10.12 Commands to track the change in data segment value  
The high water mark records the number of times the brk()value changes. The  
following commands are supported:  
info heap high-mem  
Displays the number of times brk()value  
changes for a given run.  
180 HP-UX Configuration-Specific Information  
   
set heap-check high-mem-count  
X_number  
Stops when brk()value has moved X_number  
of times.  
Limitations:  
This feature assumes that an application has a deterministic memory allocation  
pattern from one run to another.  
The high_memfeature is not supported in batch mode debugging.  
14.11 Thread Debugging Support  
HP WDB provides thread-debugging support for kernel, user, and MxN threads. You  
can exclusively disable or enable specific thread execution. Advanced thread debugging  
support in HP WDB enables you to view information on pthread primitives and detect  
certain thread-related conditions.  
NOTE: WDB supports pthreadparallelism, but it does not support  
compiler-generated parallelism such as parallelism with directives.  
14.11.1 Support for Enabling and Disabling Specific Threads  
When you suspect that a specific thread is causing a problem while debugging a multi-  
threaded application, you can suspend the execution of other threads in the application  
and debug the doubtful thread exclusively. HP WDB 3.2 and later versions provide  
the following commands to disable and enable specific thread execution:  
thread disable  
This command prevents the specified threads from running  
until they are enabled again using the thread enable  
command.  
thread enable  
This command enables the specified thread to run when you  
issue the continueor stepcommand. By default, all threads  
are in the enabled state. You can use the thread enable  
command to reactivate a disabled thread.  
Consider the following example:  
(gdb) info threadsystem thread 4189 0x7f666da8  
in __pthread_create_system+0x3d8 () from /usr/lib/libpthread.1  
2 system thread 4188 worker (wptr=0x40004640 ") at quicksort.c:135  
1 system thread 4184 0x7f66f728 in _lwp_create+0x10 () from /usr/lib/libpthread.1  
To disable a thread, execute the following command:  
(gdb) thread disable 1  
warning: ATTENTION!! Disabling threads may result in  
deadlocks in the program.Disabling thread 1  
To enable a thread, execute the following command:  
(gdb) thread enable 1  
Enabling thread 1  
14.11 Thread Debugging Support 181  
   
14.11.2 Backtrace Support for Thread Debugging  
The following commands are available as backtrace support for thread debugging:  
bt  
The btcommand provides the stack trace of the  
current thread that is being executed or the thread  
that accepts the signal in case of a core file.  
thread apply all bt  
You can use the thread apply all btcommand  
to display the backtrace of all threads. The bt  
command only provides the stack trace of the  
current thread under execution.  
backtrace_other_thread  
The backtrace_other_threadcommand prints  
the backtrace of all stack frames for a thread with  
stack pointer SP, program counter PC and address  
of gr32in the backing store BSP. This command  
enables you to view the stack trace when the stack  
is corrupted. When using this command, you must  
ensure that the SP, PC, and BSP values are valid.  
The syntax for the backtrace_other_thread  
command is as follows:  
backtrace_other_thread SP PC BSP  
14.11.3 Advanced Thread Debugging Support  
Advanced thread debugging support is available for multi-threaded applications  
running on HP-UX 11iv2, or HP-UX 11iv3.  
HP WDB 5.5 and later versions provide advanced thread debugging features to display  
extended information on the state of pthreadprimitives such as mutexes, read-write  
locks and conditional variables.  
HP WDB 5.6 and later versions provide advanced thread-debugging options to detect  
the following thread-related conditions:  
The thread attempts to acquire a non-recursive mutex that it currently holds.  
The thread attempts to unlock a mutex or a read-write lock that it has not ac-  
quired.  
The thread waits (blocked) on a mutex or read-write lock that is held by a thread  
with a different scheduling policy.  
Different threads non-concurrently wait on the same condition variable, but with  
different associated mutexes.  
The thread terminates execution without unlocking the associated mutexes or  
read-write locks.  
The thread waits on a condition variable for which the associated mutex is not  
locked.  
182 HP-UX Configuration-Specific Information  
   
The thread terminates execution, and the resources associated with the terminated  
thread continues to exist in the application because the thread has not been joined  
or detached.  
The thread uses more than the specified percentage of the stack allocated to the  
thread.  
The number of threads waiting on any pthread object exceeds the specified thresh-  
old number.  
14.11.3.1 Pre-requisites for Advanced Thread Debugging  
The following pre-requisites apply for advanced thread debugging:  
The advanced thread debugging features are supported only on HP-UX 11i v2  
and later on both PA-RISC and Integrity systems.  
The tracing pthreadlibrary is required for advanced thread-debugging. The  
pthreadtracer library is available by default on systems running on HP-UX 11i  
v2 or later. You must install HP WDB 5.5 or later versions of the debugger to  
support enhanced thread debugging. The installation scripts for HP WDB 5.5 and  
later versions of the debugger automatically add links at /opt/langtools/lib/  
to replace the standard libpthreadlibrary with libpthread tracer library  
at run time.  
The thread debugging feature in HP WDB is dependent on the availability of the  
dynamic Linker Version B.11.19.  
HP WDB uses librtc.slto enable thread debugging support. If the debugger  
is installed in a directory other than the default /opt/langtools/bindirectory,  
you must use the environment variable, LIBRTC_SERVER, to export the path of  
the appropriate version of librtc.sl.  
HP WDB does not support debugging of programs that link with the archive  
version of the standard C library libc.aor the core library libcl.a. The  
programs must be linked with libc.sl.  
The advanced thread debugging commands work only if set thread-check  
is set to on.  
For PA-RISC 32-bit applications, the dynamic library path look-up must be enabled  
for advanced thread debugging. To enable dynamic library path look-up for  
advanced thread debugging, enter the following command at HP-UX prompt:  
# chatr +s enable <PA32-bitApp>  
This command automatically enables dynamic library path look-up. No additional  
environmental variables are required to be set.  
14.11.3.2 Enabling and Disabling Advanced Thread Debugging Features  
HP WDB 5.6 and later versions of the debugger provide advanced thread debugging  
features for debugging multi-threaded applications running on 11i v2, or 11i v3. These  
14.11 Thread Debugging Support 183  
   
features are available as options to the set thread-checkcommand. The syntax for  
the set thread-checkcommand is as follows:  
set thread check { [on|off]| [option] [on|off] |[option] [num]}  
The set thread-check [on|off] command enables or disables advanced thread  
debugging. This feature is off by default. The set thread-check [on|off]  
command must be enabled prior to running the application under the debugger, to  
force the underlying runtime system to collect information on pthreadprimitives.  
The advanced thread debugging features can be enabled only if the set  
thread-check [on]command is enabled. The following advanced thread debugging  
options are available for the set thread-checkcommand:  
recursive-relock [on|off]  
This set thread-check recursive-relock  
[on|off]command checks if a thread has  
attempted to acquire a non-recursive mutex that  
it currently holds. Re-locking a non-recursive  
mutex results in a deadlock. At run-time, the  
debugger keeps track of each mutex in the  
application and the thread that currently holds  
each mutex. When a thread attempts to acquire  
a lock on a non-recursive mutex, the debugger  
checks if the thread currently holds the lock  
object for the mutex. The debugger transfers the  
execution control to the user and prints a  
warning message when this condition is detected.  
unlock-not-own [on|off]  
The set thread-check unlock-not-own  
[on|off]command checks if a thread has  
attempted to unlock a mutex or a read-write lock  
that it has not acquired. The debugger transfers  
the execution control to the user and prints a  
warning message when this condition is detected.  
mixed-sched-policy [on|off] The set thread-check  
mixed-sched-policy [on|off]command  
checks if a thread is waiting (blocked) on a mutex  
or a read-write lock that is held by a thread with  
a different scheduling policy. This is not an  
application error and does not always result in  
deadlock. However, it degrades the application  
performance. The debugger transfers the  
execution control to the user and prints a  
warning message when this condition is detected.  
cv-multiple-mxs [on|off]  
The set thread-check cv-multiple-mxs  
[on|off]command checks if an application  
184 HP-UX Configuration-Specific Information  
uses the same condition variable in multiple calls,  
by different threads to pthread_cond_wait()  
or pthread_cond_timedwait(), but specifies  
different mutexes. The debugger transfers the  
execution control to the user and prints a  
warning message when this condition is detected.  
All threads that concurrently wait on any single  
condition variable must specify the same  
associated mutex. For example,the pthread  
implementation does not allow thread 1to  
wait on condition variable Aby specifying  
mutex A, while thread 2waits on the same  
condition variable Aby specifying mutex  
B. This returns an EINVALerror to the  
application. In contrast to this obvious EINVAL  
error, the cv-multiple-mxsoption detects the  
less-obvious and non-concurrent use of multiple  
mutexes with the same condition variable by  
different threads. This option detects potential  
EINVALerrors that exist as a result of different  
timings or execution paths.  
This option is used to detect unintentional use  
of multiple mutexes with the same condition  
variable by different threads. In the case of  
applications that use a dynamic pool of mutexes,  
the cv-multiple-mxs option is not required  
because this usage is normal and expected  
application behavior.  
cv-wait-no-mx [on|off]  
The set thread-check cv-wait-no-mx  
[on|off]checks if the associated mutex of a  
condition variable is locked when the thread calls  
the pthread_cond_wait()routine. The  
debugger transfers the execution control to the  
user and prints a warning message when this  
condition is detected. This check is not a POSIX.1  
standard requirement for the  
pthread_cond_wait()routine. It is an  
additional check provided by WDB.  
thread-exit-own-mutex  
[on|off]  
The set thread-check  
thread-exit-own-mutex [on|off]  
command checks if any thread has terminated  
execution without unlocking the mutexes or  
14.11 Thread Debugging Support 185  
read-write locks that are associated with the  
thread. The debugger transfers the execution  
control to the user and prints a warning message  
when this condition is detected. This situation  
can result in deadlocks if other threads are  
waiting to acquire the locked mutexes or  
read-write locks.  
thread-exit-no-join-detach The set thread-check  
[on|off]  
thread-exit-no-join-detach [on|off]  
command checks if a thread has terminated  
execution (either successfully or because of an  
exception or a cancel) without joining or  
detaching the thread. The debugger transfers the  
execution control to the user and prints a  
warning message when this condition is detected.  
The resources associated with a terminated  
thread continue to exist in the application if the  
thread is not joined or detached. The thread must  
be explicitly joined or detached, or it must be  
created with the detach attribute. When an  
application repeatedly creates threads without  
a join or detach operation, the application leaks  
resources. This may result in application failure.  
stack-util [num]  
The set thread-check stack-util [num]  
command checks if any thread has used more  
than the specified percentage [num]of the stack  
allocation. The debugger transfers the execution  
control to the user and prints a warning message  
when this condition is detected. You can use this  
option to verify if there is a margin of safety in  
stack utilization by the threads. The application  
must ensure that the thread stack size is sufficient  
for all the operations of the thread. Each thread  
is assigned a stack allocation when it is created.  
If the stack allocation is not specified for a thread,  
the default value is used. The stack allocation  
cannot be modified after a thread is created. If a  
thread attempts to use more stack space than the  
allocated stack space, it results in a stack  
overflow. Stack over flow can result in memory  
access violations, bus errors, or segmentation  
faults.  
186 HP-UX Configuration-Specific Information  
num-waiters [num]  
The set thread-check num-waiters  
[num]command checks if the number of threads  
waiting on any pthread object exceeds the  
specified threshold number [num]. The debugger  
transfers the execution control to the user and  
prints a warning message when this condition is  
detected.  
14.11.3.3 Commands to view information on pthread primitives  
WDB 5.5 and later versions of the debugger display extended information on pthread  
primitives for multi-threaded applications running on HP-UX 11i v2, or 11i v3. This  
feature is available only if the set thread-check [on|off]command is enabled.  
The following commands enable you to view extended information on threads, mutexes,  
read-write locks and conditional variables in multi-threaded applications:  
info thread [thread-id]  
info mutex [mutex-id]  
info condvar [condvar-id]  
The info thread [thread-id]command  
displays a list of known threads. If a thread-id is  
provided, the command displays extended  
information on the specified thread.  
The info mutex [mutex-id]command  
displays a list of known mutexes. If a mutex-id  
is provided, the command displays extended  
information on the specified mutex.  
The info condvar [condvar-id]command  
displays a list of known condition variables. If  
condvar-id is provided, the command displays  
extended information on the specified condition  
variable.  
info rwlock [rwlock-id]  
The info rwlock [rwlock-id]command  
displays a list of known read-write locks. If  
rwlock-id is provided, the command displays  
extended information on the specified read-write  
lock.  
14.11.4 Debugging Threads Interactively After Attaching to a Process  
HP WDB provides support to attach a running process to the debugger. To use thread  
debugging commands after attaching GDB to a running process, complete the following  
steps:  
14.11 Thread Debugging Support 187  
   
1. Set LD_LIBRARY_PATH to include the appropriate directory, by entering one of  
the following commands:  
For 32 bit IPF applications,  
LD_LIBRARY_PATH=/opt/langtools/wdb/lib/hpux32  
For 64 bit IPF applications,  
LD_LIBRARY_PATH=/opt/langtools/wdb/lib/hpux64  
For 32 bit PA applications,  
LD_LIBRARY_PATH=/opt/langtools/wdb/lib  
For 64-bit PA applications,  
LD_LIBRARY_PATH=/opt/langtools/wdb/lib/pa20_64  
2. Complete one of the following steps to preload the librtcruntime library:  
Set the target application to preload librtc by using the +rtcoption for the  
chatrcommand. In addition to automatically loading the librtclibrary,  
the +rtcoption for the chatrcommand also maps the shared libraries as  
private. To enable or disable the target application to preload the librtc  
runtime library, enter the following command at the HP-UX prompt:  
$ chatr +rtc <enable|disable> <executable>  
NOTE: The chatr+rtcoption preloads the librtcruntime library from  
the following default paths:  
— For 32 bit IPF applications,  
/opt/langtools/lib/hpux32/librtc.so  
— For 64 bit IPF applications,  
/opt/langtools/lib/hpux64/librtc.so  
— For 32 bit PA applications,  
opt/langtools/lib/librtc.sl  
— For 64-bit PA applications,  
/opt/langtools/lib/pa20_64/librtc.sl  
To preload the librtcruntime library from a path that is different from the  
default paths, you must use the LD_PRELOADenvironment variable.  
OR  
Instead of automatically preloading librtcand mapping the shared libraries,  
you can explicitly preload the required librtclibrary after mapping the  
shared libraries private.  
In the case of HP 9000 systems, you must explicitly map the share libraries as  
private by using the +dbg enableoption for the chatrcommand, as follows:  
188 HP-UX Configuration-Specific Information  
$ chatr +dbg enable ./<executable>  
(This step is not required on Integrity systems.)  
To explicitly preload the librtc runtime library and start the target application,  
enter one of the following commands:  
— For 32 bit IPF applications,  
LD_PRELOAD=/opt/langtools/lib/hpux32/librtc.so <executable>  
— For 64 bit IPF applications,  
LD_PRELOAD=/opt/langtools/lib/hpux64/librtc.so <executable>  
— For 32 bit PA applications,  
LD_PRELOAD=/opt/langtools/lib/librtc.sl <executable>  
— For 64-bit PA applications,  
LD_PRELOAD=/opt/langtools/lib/pa20_64/librtc.sl <executable>  
If LD_PRELOADand chatr+rtcare used to preload the librtc runtime library  
, the librtcruntime library is loaded from the path specified by  
LD_PRELOAD.  
3. Complete one of the following steps:  
Attach the debugger to the required process and enable thread debugging,  
as follows:  
gdb -thread -p <pid>  
or  
gdb -thread <executable> <pid>  
Alternately, you can attach the process to the debugger and consequently  
invoke thread debugging, as follows:  
$ gdb <executable> <pid>  
...  
(gdb)set thread-check on  
14.11.5 Thread Debugging in Batch Mode  
HP WDB supports batch mode of debugging threads for HP-UX 11iv2 and later, on  
Integrity systems and on HP-UX 11i v3 in PA-RISC systems for 64 bit applications. The  
debugger provides a log file with the list of thread-related errors that occur in the  
application.  
In batch mode, the debugger detects the all the thread-conditions that are detected  
during an interactive debugging session.  
The debugger reports extended information such as variable address, name, id and  
other specifications related to the involved pthread objects. In addition, it displays the  
stack trace of the executing thread at the point of error.  
14.11 Thread Debugging Support 189  
 
NOTE: Use the set frame-countsetting in the rtconfigfile to control the depth  
of the stack trace file. This command controls the depth of the call stack collected. Larger  
values increase the run time.  
14.11.5.1 Pre-requisites for Batch mode of Thread Debugging  
The various prerequisites for Batch mode of Thread Debugging are as follows:  
The thread-debugging feature in HP WDB is dependent on the availability of the  
dynamic linker version B.11.19.  
Advanced thread-debugging feature requires the pthread tracer library which is  
available by default on systems running HP-UX 11i v2 or later.  
Steps to debug threads in batch mode  
1. Compile the source files.  
Set the LD_LIBRARY_PATH environment variable, based on the platform as  
follows:  
For 32 bit IPF applications, set  
LD_LIBRARY_PATH=/opt/langtools/wdb/lib/hpux32  
For 64 bit IPF applications, set  
LD_LIBRARY_PATH=/opt/langtools/wdb/lib/hpux64  
For 64-bit PA applications, set  
LD_LIBRARY_PATH=/opt/langtools/wdb/lib/pa20_64  
2. Map the share libraries as private for HP 9000 systems using the following  
command:  
$ chatr +dbg enable ./<executable>  
NOTE: This step is not applicable for Integrity systems.  
3. Create a configuration file, rtcconfigto specify the various thread conditions  
that you want the debugger to detect.  
190 HP-UX Configuration-Specific Information  
 
NOTE: The configuration file contains lines of the following form:  
set thread-check [on|off] | [option] [on|off] | [option] [num]  
And/Or  
set frame-count [num]  
And/Or  
files=<name of the executable on which the thread checking is to be  
done>  
4. Set the environment variable BATCH_RTCto on as:  
Set BATCH_RTC=on  
14.11 Thread Debugging Support 191  
5. Complete one of the following steps to preload the librtcruntime library:  
Set the target application to preload librtcby using the +rtcoption for the  
chatrcommand. In addition to automatically loading the librtclibrary,  
the +rtcoption for the chatrcommand also maps the shared libraries as  
private.  
To enable or disable the target application to preload the librtcruntime  
library, enter the following command at the HP-UX prompt:  
$ chatr +rtc <enable|disable> <executable>  
NOTE: The chatr+rtcoption preloads the librtcruntime library from  
the following default paths:  
— For 32 bit IPF applications,  
/opt/langtools/lib/hpux32/librtc.so  
— For 64 bit IPF applications,  
/opt/langtools/lib/hpux64/librtc.so  
— For 64-bit PA applications,  
/opt/langtools/lib/pa20_64/librtc.sl  
To preload the librtc runtime library from a path that is different from the  
default paths, you must use the LD_PRELOADenvironment variable.  
OR  
Instead of automatically preloading librtcand mapping the shared libraries,  
you can explicitly preload the required librtclibrary after mapping the  
shared libraries private.  
In the case of HP 9000 systems, you must explicitly map the share libraries as  
private by using the +dbg enable option for the chatrcommand, as follows:  
$ chatr +dbg enable ./<executable>  
(This step is not required on Integrity systems.)  
To explicitly preload the librtc runtime library and start the target application,  
enter one of the following commands:  
— For 32 bit IPF applications,  
LD_PRELOAD=/opt/langtools/lib/hpux32/librtc.so <executable>  
— For 64 bit IPF applications,  
LD_PRELOAD=/opt/langtools/lib/hpux64/librtc.so <executable>  
— For 64-bit PA applications,  
LD_PRELOAD=/opt/langtools/lib/pa20_64/librtc.sl <executable>  
192 HP-UX Configuration-Specific Information  
If LD_PRELOADand chatr+rtcare used to preload the librtcruntime  
library , the librtcruntime library is loaded from the path specified by  
LD_PRELOAD.  
If HP WDB detects any thread error condition during the application run, the error  
log is output to a file in the current working directory. The output file has the  
following naming convention:  
<executable name>.<pid>.threads  
where,  
<pid>is the process id.  
14.11.5.2 Limitations in Batch mode of thread debugging  
The feature does not obtain the thread-error information in batch mode for forked  
process in a multiprocessing application. However, if the librtc.sllibrary is  
preloaded, the debugger obtains the thread-error information in the batch mode for  
exec-edapplication.  
You cannot specify an alternate output directory for the thread-error log. The thread-  
error log file is output into the current working directory only.  
HP WDB cannot execute both batch mode thread check and batch mode heap check  
together. If the rtcconfigfile has both entries, then batch heap check overrides the  
batch thread check.  
14.11.6 Thread Debugging in +checkMode  
A new compiler option +check=threadenables batch mode thread debugging features  
of HP WDB.  
NOTE: This feature is available only for compiler versions A.06.20 and later.  
It is a convenient way of launching the batch mode advanced thread checking features  
without setting any other environment variables at runtime. In other words, batch  
mode thread checking has two modes of invocation. The first method is to use the run-  
time environment variables LD_LIBRARY_PATH, LD_PRELOAD, and BATCH_RTCon  
existing precompiled applications. The second method is to use the +check=thread  
option at the compile time.  
+check=threadmust only be used with multithreaded programs. It is not enabled  
by +check=all. This functionality requires HP WDB 5.9 or later.  
The default configuration used by +check=threadoption is as follows:  
thread-check=1;recursive-relock=1;unlock-not-own=1;  
mix-sched-policy=1;cv-multiple-mxs=1;cv-wait-no-mx=1;  
thread-exit-own-mutex=1;thread-exit-no-join-detach=1;stack-util=80;  
num-waiters=0;frame_count=4;output_dir=.;  
14.11 Thread Debugging Support 193  
   
Behavior of the +check=threadoption can be changed by users by providing their  
own rtcconfigfile. The user specified rtcconfigfile can be in the current directory  
or in a directory specified by the GDBRTC_CONFIGenvironment variable.  
If any thread error condition is detected during the application run, the error log will  
be output to a file in the current working directory. The output file will have the  
following naming convention:  
<executable name>.<pid>.threads,  
where <pid>is the process identifier.  
14.11.7 Known issues with Thread Debugging for Interactive and Batch mode  
Issue 1:  
During the execution of advanced thread checking for applications that fork, in the  
interactive mode, the following message appears if the GDB follows the child:  
Pthread analysis file missing!  
This error message appears because the thread-error information for the forked process  
is not available.  
However, if the forked process exec()s another binary, the thread-error information  
is available for the exec-ed binary.  
Issue 2  
In both interactive and batch modes, if the applications exceed their thread stack  
utilization, the following error message appears:  
Error accessing memory address  
This occurs when GDB attempts a command line call on an already overflowing thread  
stack.  
14.12 Debugging MPI Programs  
You can attach the debugger to Message Passing Interface (MPI) programs for  
debugging. You must set the one of the following environment variables before you  
launch the MPI process:  
set MPI_FLAGS= egdb for invoking GDB  
or  
set MPI_FLAGS= ewdb for invoking WDB  
For more information, see the mpidebug(1)and mpienv(1)manpages.  
Attaching the debugger to an MPI process (or to any other process that has not been  
compiled for debugging) can result in the following warning:  
warning: reading 'r3' register: No data  
194 HP-UX Configuration-Specific Information  
   
14.13 Debugging multiple processes ( programs with forkand vfork  
calls)  
14.13.1 Ask mode for set follow-fork-mode  
The ask command prompts the user to select between parent and child as the debugger  
response to a program call of fork/vfork. Based on the user selection, the parent or  
the child process is debugged.  
For example,  
(gdb) set follow-fork-mode ask (gdb) show follow-fork-mode  
The debugger response to a program call to forkor vforkis ask.  
(gdb) run Starting program: sample [New process 4941] Select follow-fork-  
mode: [0] parent [1] child  
14.13.2 Serial mode for set follow-fork-mode  
The option serial, for the follow-fork-modecommand, enables debugging of a  
parent and child process within a debugger session. During a debug session, if the  
parent process forks a child, the debugger starts debugging the child process. When  
the child process exits, the debugger switches back to the parent process. The  
follow-fork-modewill work only if there is a wait() call by the parent process. This  
feature is enabled by setting the follow-fork-modeflag to serial, as specified in the  
following example:  
(gdb) set follow-fork-mode serial  
The follow-fork-modeis not supported under following conditions:  
MxN threaded programs  
Parent process is 32-bit and child process is 64-bit and vice versa. For the follow-  
fork-modeto work both parent and child process must be of the same mode.  
14.13.3 Support for showing unwind info  
The maint info unwindcommand prints the unwind information for the regions  
unwinded at the given address expression. Usage:  
maint info unwind exp  
where, expis an address expression.  
For example,  
(gdb) maint info unwind $pc  
modsched:  
0x4000930 .. 0x4000a20, end_prologue@0x4000970  
Info block version:0x0, flags:0x0, length:4 * 4 == 16  
0x40172b20: (0c) R1prologue rlen=12  
0x40172b21: (e8) P7preds_when t=11  
0x40172b23: (b1) P3preds_gr gr=41  
0x40172b25: (ea) P7lc_when t=7  
14.13 Debugging multiple processes ( programs with fork and vfork calls) 195  
       
0x40172b27: (b2) P3lc_gr gr=40  
0x40172b29: (61) R3body rlen=33  
0x40172b2b: (81) B1label_state label=1  
0x40172b2c: (c0) B2epilogue t=44  
0x40172b2e: (00) R1prologue rlen=0  
0x40172b2f: (00) R1prologue rlen=0  
14.13.4 Printing CFM and PFS registers  
On Integrity systems, HP WDB prints Current Frame Marker (CFM) and Previous  
Frame State (PFS) ar64registers in two different formats:  
raw values  
special formats identifying the size of rotating registers, frame and locals.  
For example,  
ar64: 0xc00000000000050c (sor:0, sol:10, sof:12)  
cfm: 0x800000000000450a (sor:1, sol:10, sof:10)  
14.14 Command to Search for a Pattern in the Memory Address Space  
The HP WDB findcommand searches for a pattern in the given memory address  
range for both live and corefile debugging. Following is the syntax for the find  
command:  
find [/size-char] [/max-count] start-address, end-address, expr1  
[, expr2 ...]  
find [/size-char] [/max-count] start-address, +length, expr1 [,  
expr2 ...]  
where:  
/size-char  
Specifies the size of the pattern. It can be b, h, w, or gfor  
8-, 16-, 32-, 64-bit values, respectively. It is applicable only  
for hexadecimal patterns.  
If /size-charis not specified, the size is obtained from  
the type of the expression in the current programming  
language.  
This is an optional parameter.  
/max-count  
Specifies the maximum number of matching patterns to  
be displayed. This is an optional parameter.  
start-address  
end-address  
Specifies the start address in the given memory address  
range. This parameter is mandatory.  
Specifies the end address in the given memory address  
range. This parameter is mandatory.  
196 HP-UX Configuration-Specific Information  
   
+length  
Specifies the length of the memory address range. This  
parameter is mandatory when end-addressis not  
specified.  
expr1, expr2, ....  
Specifies a variable or pattern to be searched. The pattern  
can be a variable, hexadecimal character, a character, or  
a string value. The character pattern must be enclosed  
within single quotes. The string pattern must be enclosed  
within double quotes. This parameter is mandatory.  
The $numfoundvariable contains the number of patterns matched within the given  
memory address range.  
14.14 Command to Search for a Pattern in the Memory Address Space 197  
Example 14-1 Sample Output for the findcommand  
$ cat example.c  
#include <stdio.h>  
#include <stdlib.h>  
int main()  
{
char *str;  
str = (char *) malloc (15);  
strcpy(str,"hihi-hikh");  
return 0;  
}
(gdb) find &str[0], &str[15], "hi"  
0x400123e0  
0x400123e2  
0x400123e5  
3 patterns found.  
(gdb) find/2 &str[0], &str[15], "hi"  
0x400123e0  
0x400123e2  
2 patterns found.  
(gdb) find/2b &str[0], &str[15], 0x68  
0x400123e0  
0x400123e2  
2 patterns found.  
(gdb) find/2b &str[0], +10, 0x68  
0x400123e0  
0x400123e2  
2 patterns found.  
(gdb) find/2b &str[0], +10, 0x68, 0x69  
0x400123e0  
0x400123e2  
2 patterns found.  
(gdb) find &str[0], &str[15], "hi", "hi"  
0x400123e0  
1 pattern found.  
The following examples provide sample usages of the different parameters in the find  
command:  
Using the start address (start-address), end address (end-address), and  
patterns (expr1, expr2)  
(gdb) find &a[0], &a[10], "el",'l'  
2 patterns found.  
198 HP-UX Configuration-Specific Information  
 
where:  
&a[0]  
&a[10]  
“el”, 'l'  
Specifies the start address of the memory address range.  
Specifies the end address of the memory address range.  
Specifies the pattern.  
Using the start address (start-address), length (+length) parameter, and a  
pattern (expr1)  
find &str[0], +11, "hihi"  
&str[0] Specifies the starting address.  
+11  
Specifies the length of the memory address range, starting from  
&str[0].  
"hihi"  
Specifies the pattern (expr1).  
Using the /max-countparameter  
(gdb) find /1 &int8_search_buf[0], +sizeof(int8_search_buf), 'a', 'a', 'a'  
where:  
/1  
Specifies the findcommand to display only  
one matching pattern.  
&int8_search_buf[0]  
+sizeof(int8_search_buf)  
'a', 'a', 'a'  
Specifies the starting address.  
Specifies the ending address.  
Specifies the pattern (expr1, expr2,  
expr3).  
Using the /size-charparameter  
(gdb) find /b &int8_search_buf[0], &int8_search_buf[0]+sizeof(int8_search_buf),  
0x61, 0x61, 0x61, 0x61  
where:  
/b  
Specifies that the size of the pattern is 8 bits.  
Specifies the starting address.  
&int8_search_buf[0]  
&int8_search_buf[0]  
+sizeof(int8_search_buf)  
Specifies the ending address.  
0x61, 0x61, 0x61, 0x61  
Specifies the pattern (expr1, expr2, expr3,  
exp4).  
14.14 Command to Search for a Pattern in the Memory Address Space 199  
NOTE: Following are different ways of representing the /size-charand  
/max-countparameters:  
/1b  
/b1  
/b /1  
/1 /b  
where:  
1
Specifies that findmust display 1 matching pattern.  
b
Specifies that the size of the pattern is 8 bits.  
14.15 Debugging Core Files  
14.15.1 Generating core files with packcore/unpackcore/getcore  
The contents of a core file can be viewed only on a system which has all the shared  
libraries that were in use on the system on which the core file was generated. If you  
want to view the content of the core file on a system which does not have the shared  
libraries, you have to set the environment variables GDB_SHLIB_PATHor GDB_SHLIB_  
ROOTto make it search for the desired libraries. The commands packcore, unpackcore,  
and core simplify the process of examining the contents of a core file on a system other  
than the one in which it was generated.  
The packcorecommand is used on the system which generated the core file. When  
you are examining the core file on the original system, you can execute packcoreto  
make a packcore.tar.Zfile. This is a compressed tar file which contains the core  
file, the executable, and all the shared libraries that were used by the program when  
it generated the core file. The core file is removed after it is added to the  
packcore.tar.Z file.  
The packcorecommand has one optional argument, basename, which can be used  
instead of packcoreto make packcore.tar.Z.  
The packcore.tar.Zfile can be copied to a different system and the gdb command  
unpackcoreunpacks the packcore.tar.Zfile in the current directory, creating a  
new packcore directory. After unpacking the packcore file, the unpackcorecommand  
invokes getcoreto load the executable and the core file from the packcore directory,  
and sets GDB_SHLIB_PATH to the modules directory in the packcore directory. The  
modules directory holds all of the shared libraries that were used when the core file  
was generated.  
The unpackcorecommand has two optional arguments. The first defaults to  
packcore.tar.Zand is the name of the packcore file to be unpacked. The second  
argument is given if the core file is too large to fit in the packcore file. It is the path to  
200 HP-UX Configuration-Specific Information  
   
the core file to be used if the packcore directory does not contain a core file. If used,  
this second argument causes a symbolic link to be created in the packcore directory in  
place of the missing core file.  
The getcorecommand can be used to examine a packcore directory which was  
previously created by unpackcore. It takes one optional argument, the name of the  
packcore directory, which defaults to packcore.  
14.15.2 Support for the info targetCommand  
The info targetcommand is enhanced to display the memory segments for the  
corefile. The output also displays other details, such as system name, node name,  
operating system release name, license level, and machine model, for the core file. The  
enhancement is available for HP-UX 11i v2 and later versions of Itanium systems.  
Following is the sample output for the enhanced info targetcommand:  
(gdb) file info_target  
Reading symbols from info_target...done.  
(gdb) core core  
Core was generated by `info_target'.  
Program terminated with signal 6, Aborted.  
#0 0xc00000000032c5b0:0 in kill+0x30 () from /usr/lib/hpux64/libc.so.1  
(gdb) info target  
Operating System Information:  
sysname : HP-UX  
nodename : test1  
release : B.11.23  
version : U (128-user, 256-user or Unlimited-user system)  
machine : ia64  
idnumber : 0735747273  
Symbols from "/tmp/info_target".  
Local core dump file:  
`/tmp/core',  
file type elf64-big.  
0x6000000000000000 - 0x6000000000010000 is segment4 (PT_HP_CORE_LOADABLE)  
0x9ffffffffb7d9000 - 0x9ffffffffb7da000 is segment5 (PT_HP_CORE_MMF)  
0x9ffffffffb7da000 - 0x9ffffffffb7dc000 is segment6 (PT_HP_CORE_MMF)  
0x9ffffffffb7dc000 - 0x9ffffffffb7e0000 is segment7 (PT_HP_CORE_MMF)  
0x9ffffffffb7e0000 - 0x9ffffffffb7e8000 is segment8 (PT_HP_CORE_MMF)  
0x9ffffffffb7e8000 - 0x9ffffffffb7f3000 is segment9 (PT_HP_CORE_MMF)  
0x9ffffffffb7f3000 - 0x9ffffffffb7f4000 is segment10 (PT_HP_CORE_MMF)  
0x9ffffffffb7f4000 - 0x9ffffffffb7f6000 is segment11 (PT_HP_CORE_MMF)  
0x9ffffffffb7f6000 - 0x9ffffffffb7fa000 is segment12 (PT_HP_CORE_MMF)  
0x9ffffffffb7fa000 - 0x9ffffffffb7fc000 is segment13 (PT_HP_CORE_MMF)  
0x9ffffffffb7fc000 - 0x9ffffffffb7fe000 is segment14 (PT_HP_CORE_MMF)  
0x9ffffffffb7fe000 - 0x9ffffffffb7ff000 is segment15 (PT_HP_CORE_MMF)  
0x9ffffffffb7ff000 - 0x9ffffffffb800000 is segment16 (PT_HP_CORE_STACK - RSE)  
0x9fffffffffffa000 - 0xa000000000000000 is segment17 (PT_HP_CORE_STACK)  
0x6000000000000058 - 0x600000000000d000 is heap segment  
0x9fffffff7b7f6270 - 0x9fffffff7b7f62e8 is .note.hpux_options in /usr/lib/hpux64/dld.so  
0x9fffffff7b7f62e8 - 0x9fffffff7b7f6458 is .dynamic in /usr/lib/hpux64/dld.so  
............  
...........  
14.15 Debugging Core Files 201  
 
NOTE: Limitations for the enhanced info targetcommand on corefile are as  
follows:  
The enhanced info targetcommand is not supported when the  
expanded_node_host_names kernel parameter is set. It is supported only for the  
default utsname.  
The heap segment listed in the memory segment does not contain mmapped  
memory for the given core file.  
14.15.3 Support for the dumpcorecommand  
HP WDB provides the command, dumpcore to generate a core image file for a process  
running under the debugger during execution.  
The dumpcore command does not require any argument. It saves the core image for  
the current process being debugged in the file named core.<pid>, where <pid> is the  
process ID number.  
To dump the core for a live process, you must pass the following commands:  
(gdb) run Starting program: sample Breakpoint 3, main () at sample.c:1010 b  
= foo(a); (gdb) dumpcore Dumping core to the core file core.24091(gdb)  
When starting from the HP WDB command line:  
(gdb) file sample  
Reading  
symbols from sample...done  
(gdb) set live-core 1  
(gdb) core-file core.24091  
Core was generated by 'sample'.#0 main () at sample.c:1010 b = foo(a);  
(gdb) backtrace#0 main () at sample.c:10  
(gdb)  
When starting from the shell prompt:  
% gdb --lcore a.out core.<pid>  
For example:  
% ./gdb --lcore sample core.24091  
HP gdb...  
Type "show warranty" for warranty/support....  
Core was generated by 'sample'.#0 main () at sample.c:10  
(gdb)  
14.15.3.1 Enhancements to the dumpcorecommand  
HP WDB provides an option for the dumpcore command, to specify a <core-  
filename>, to generate a core image file of a process running under the debugger in  
the middle of execution and saves it in the file named <core-filename>.  
The dumpcorecommand with no arguments saves the core image for the current  
process being debugged in the file named core.<pid>, where pidis the process ID  
number.  
202 HP-UX Configuration-Specific Information  
   
To analyze this core file with HP WDB on HP-UX 11i version 2, you must do the  
following:  
When starting from HP WDB command line:  
(gdb) core-file [core.pid | core-filename]  
When starting from shell prompt:  
$ gdb -core a.out [core.pid | core-filename]  
14.15.4 Support for display of run time type information  
HP WDB enables you to view the run time type information for C++ polymorphic  
object.  
info rtti address  
This command displays run time type information for C++ polymorphic object. The  
input to this command is addressof the C++ polymorphic object. GDB displays the  
demangled type name as output.  
NOTE: This command is supported only on Integrity systems.  
Sample Output  
(gdb) info rtti <address>  
RTTI: <run time type/class name>  
14.16 Printing the Execution Path Entries for the Current Frame or Thread  
HP WDB 5.7 and later versions of the debugger enable you to print the execution path  
entries in the current frame, or the current thread for programs running on Integrity  
systems. This feature enables the display of the execution path taken across branched  
modules. The first instruction in each block associated with the executed branch is  
displayed.  
This feature is supported only for compiler versions A.06.15 and later.  
HP WDB supports the following commands to print the execution path entries in the  
current frame, or in the current thread:  
info exec-path [start_index] [end_index] (aliased to info  
ep)  
Lists all the local execution path entries in the current frame. The [start_index]  
and [end_index]indicate the range of table indexes (execution path entries)  
that must be displayed.  
If [end_index]is not specified, the debugger displays the complete table of  
execution path entries, starting from [start_index].  
If [start_index]and [end_index]are not specified, the complete table of  
execution path entries is displayed.  
14.16 Printing the Execution Path Entries for the Current Frame or Thread 203  
   
For example,  
(gdb) i ep 4 10  
info exec-path summary  
Prints the summary information about all the local execution path entries in the  
current frame. This command displays the total number of branches for the frame,  
the number of branches executed in this frame in the last iteration, and the last  
executed branch number.  
info global-exec-path [start_index] [end_index](aliased to  
info gep)  
Lists all the global execution path entries for the current thread.  
The [start_index]and [end_index]indicate the range of table indexes  
(execution path entries) that must be displayed.  
If [end_index]is not specified, the debugger displays the complete table of  
execution path entries, starting from [start_index].  
If [start_index]and [end_index]are not specified, the complete table of  
execution path entries is displayed.  
info global-exec-path summary  
Prints the summary information about all the global execution path entries in the  
current frame. This command displays the total number of global execution path  
entries that can be stored, the number of global execution path entries in this frame  
in the last iteration, and the last executed global execution path number.  
exec-path [up][down][path_index] (aliased to ep)  
Enables you to select, print, and navigate through the execution path entries. When  
no arguments are specified, it prints the selected execution path entry. You can  
specify the argument as an execution path index from the info exec-pathor  
the info global-exec-pathcommands. Alternately, you can use the up or down  
command to navigate through the execution path entries.  
14.16.1 Compiler Dependencies for Printing the Execution Path Entries  
The +pathtracecompiler option provides a mechanism to record program execution  
control flow into global path tables, local path tables, or both. This saved information  
enables the debugger to print the execution path entries in the current thread or frame.  
To print the execution path entries in the current thread or frame for programs running  
on Integrity systems, you can set the required sub-options for the +pathtracecompiler  
option.  
You must set the following +pathtracecompiler option to enable the debugger to  
print the execution path entries:  
+pathtrace= [<global|global_fixed_size>:<local>]  
204 HP-UX Configuration-Specific Information  
 
For more information on this feature, see the following example.  
14.16.2 Example Illustrating Execution Path Recovery  
The following example illustrates the use of the execution path recovery feature in HP  
WDB:  
Sample Program:  
$cat execpath.c  
#include <stdio.h>  
#include <string.h>  
#include <stdlib.h>  
int main()  
{
int a = 3, b = 0, c = 4;  
if (a)  
printf("Value of a greater than 0\n");  
if (b)  
printf("Value of b greater than 0\n");  
if (c)  
printf("Value of c greater than 0\n");  
printf("All condition checking done\n");  
return 0;  
}
Sample Debugging Session:  
$cc +pathtrace -g execpath.c  
$gdb a.out  
HP gdb ...  
Type "show warranty" for warranty/support.  
...  
(gdb) b main  
Breakpoint 1 at 0x4000a60:0: file execpath.c, line 7 from a.out.  
(gdb) r  
Starting program: a.out  
Breakpoint 1, main () at execpath.c:7  
7 int a = 3, b = 0, c = 4;  
(gdb) n  
9 if (a)  
(gdb) i ep  
Local execution path table for main():  
empty  
(gdb) i gep  
Global execution path table:  
empty  
(gdb) n  
10 printf("Value of a greater than 0\n");  
(gdb) n  
Value of a greater than 0  
12 if (b)  
14.16 Printing the Execution Path Entries for the Current Frame or Thread 205  
 
(gdb) i ep  
Local execution path table for main():  
0 0x4000a80:2 (execpath.c:10)  
(gdb) i gep  
Global execution path table:  
G0 0x4000a80:2 main (execpath.c:10)  
(gdb) n  
15 if (c)  
(gdb) i ep  
Local execution path table for main():  
0 0x4000a80:2 (execpath.c:10)  
(gdb) i gep  
Global execution path table:  
G0 0x4000a80:2 main (execpath.c:10)  
(gdb) n  
16 printf("Value of c greater than 0\n");  
(gdb) n  
Value of c greater than 0  
18 printf("All condition checking done\n");  
(gdb) i ep  
Local execution path table for main():  
0 0x4000a80:2 (execpath.c:10)  
2 0x4000bd0:2 (execpath.c:16)  
(gdb) i ep summary  
Summary for local execution path table for main()  
Size: 3 \*Total Number of Branch Paths in Current Function  
Effective entries: 2 \*Number of Branches executed till this instant  
Current entry: 2 \* Last executed branch number  
(gdb) i gep  
Global execution path table:  
G0 0x4000a80:2 main (execpath.c:10)  
G1 0x4000bd0:2 main (execpath.c:16)  
(gdb) i gep summary  
Summary for global execution path table  
Size: 65536 \*Maximum execution path entries to be stored  
Effective entries: 2 \*Number of global execution path entries  
Current entry: 1 \*The last Global Path ID executed  
(gdb)  
14.17 Command to Unwind Beyond 10000 Frames  
The number of frames to be unwound by default is set to 10000 so that HP WDB does  
not run out-of-memory. To enable unwinding beyond 10000 frames, HP WDB supports  
the set unwind-all-framescommand at the gdbprompt.  
Following is the syntax for the unwind-all-framescommand:  
set unwind-all-frames [on|off]  
where:  
on  
Enables WDB to unwind beyond 10000 frames.  
off Disables the limitation on the number of frames.  
206 HP-UX Configuration-Specific Information  
 
14.18 Invoking GDB Before a Program Aborts  
This -crashdebugoption enables GDB to monitor the execution of a process or a  
program. It invokes GDB when the program execution is about to abort. Once the  
debugger is invoked, you can debug the application using the normal debugger  
commands. This option enables you to debug a live process instead of getting a core  
dump if the program is about to abort.  
You can examine the state of the process, make changes to the state and continue  
program execution, force a core dump, or terminate execution. It enables you to control  
program execution under the debugger if the program is about to abort. You can load  
a new process or attach to a running process for monitoring.  
To monitor a new process, enter the following command:  
gdb -crashdebug <command> <options>  
To monitor a running process, attach to the process using the following command:  
gdb -crashdebug -pid <pid>  
14.19 Aborting a Command Line Call  
When a command line call is issued and it is interrupted by a breakpoint or a signal  
before completing the program execution, the abort command enables the user to abort  
the command line call without allowing the signal to modify the state of the debugged  
process.  
When a signal interrupts program execution, it can modify the process state of the de-  
bugged program and result in an abrupt termination of the program (due to addressing  
errors from a call that is not a part of the source program). In such cases, the abort  
command is particularly useful in exiting the command line call without modifying  
the process state of the debugged program.  
The following example illustrates the use of the abort command:  
(gdb) break main  
Breakpoint 1 at 0x2c74: file .../address_error.c, line 41.  
(gdb) run  
Starting program: ./address_error  
Breakpoint 1, main () at ./address_error.c:41  
41 fun (count, count*1.1);  
(gdb) p fun(10, 1.1)  
Program received signal SIGBUS, Bus error  
si_code: 0 - BUS_UNKNOWN - Unknown Error.  
0x2c38 in fun (i=10, f=0) at ./address_error.c:37  
37 count = *p;  
The program being debugged was signaled while in a function called from GDB.  
GDB remains in the frame where the signal was received.  
To change this behavior use "set unwindonsignal on"  
Evaluation of the expression containing the function (fun) will be abandoned.  
(gdb) bt  
#0 0x2c38 in fun (i=10, f=0) at ../address_error.c:37  
#1 0x1920 in _sr4export+0x8 ()  
#2 <function called from gdb>  
#3 0x2c74 in main () at ./address_error.c:40  
(gdb) abort  
14.18 Invoking GDB Before a Program Aborts 207  
   
Abort gdb command line call? (y or n) y  
#0 main () at ./address_error.c:41  
41 fun (count, count*1.1);  
(gdb) bt  
#0 main () at ./address_error.c:41  
(gdb) quit  
The program is running. Exit anyway? (y or n) y  
14.20 Instruction Level Stepping  
During instruction level stepping with nexti and stepi, WDB prints the assembly  
instruction along with the next source line.  
(gdb) stepi 0x101530:0 st4 [r9]=r34 1337 args.argc = argc;  
It also prints DOC line information, which includes actual line number and the column  
number, when debugging a binary with -g -O.  
(gdb) stepi ;;; [8, 1] 0x4000820:1 nop.m 0x0  
GDB cannot step into a function with no debug information. GDB cannot do a next  
over a line when there is not debug information. However, the continuecommand  
works in such situations.  
14.21 Enhanced support for watchpoints and breakpoints  
14.21.1 Deferred watchpoints  
When you try to set a watchpoint in an expression, HP WDB places a deferred watch-  
point if HP WDB cannot evaluate the expression. The watchpoint is automatically  
enabled whenever the expression can be evaluated during the programs execution.  
This is especially useful when placing the watchpoints on unallocated addresses.  
14.21.2 Hardware watchpoints  
HP WDB provides support for hardware watchpoints on HP-UX 11.x.  
14.21.3 Hardware breakpoints  
The hbreakcommand sets hardware assisted breakpoints.  
hbreak args  
The arguments (args) is same as that for the break command and the breakpoint is set  
in the same way. However, the breakpoint uses hardware assisted breakpoint registers.  
There are only two hardware breakpoints that can be set on Integrity systems. This is  
useful in ROM code debugging and shared library debugging for libraries, including  
dld, that are not loaded private.  
The normal breakpoints are converted to a hardware breakpoint when WDB cannot  
set a normal breakpoint in the shared library.  
208 HP-UX Configuration-Specific Information  
         
14.21.3.1 Setting breakpoints in unstripped shared library  
GDB will not be able to put breakpoints using symbolic names(of the symbols not in  
export list) or line numbers in the stripped modules.  
GDB will be able to place breakpoints using symbol names in the unstripped shared  
libraries loaded into the stripped executable.  
14.21.4 Support for procedural breakpoints  
HP WDB enables you to set breakpoints at the beginning (first executable line) of every  
function that can be debugged. In addition, you can specify a set of commands to be  
performed when the breakpoint is reached. These breakpoints work like procedural  
breakpoints in the xdb debugger.  
The breakpoint commands are rbpand rdp.  
rbp  
Sets breakpoints at the first executable statement in all the functions that can be  
debugged, including any shared libraries that are already loaded. The rbp  
command sets breakpoints in all the functions, which can be debugged, in all the  
source files. After you set these breakpoints, you can manage them like any  
standard breakpoints. You can delete them, disable them, or make them conditional.  
Each time you use the rbpcommand, HP WDB adds an additional breakpoint at  
the beginning of each function that performs the commands you specify, if any.  
rdp  
Deletes all the breakpoints set by the rbpcommand.  
This example shows how to set a breakpoint at the start of each procedure that displays  
information at the breakpoint:  
(gdb) file a.out  
Reading symbols from a.out...done.  
(gdb) rbp  
Breakpoints set from 170 to 211  
Type commands to execute when the breakpoint is hit (one command per line).  
End with a line saying just "end".  
>info break  
>end  
(gdb)  
14.21.5 Support for template breakpoints  
With HP WDB 5.0, you can set breakpoints on all instantiations of the template class  
by just specifying the template name with member function name.  
For example:  
(gdb) break ::  
It is not necessary to specify the instantiation type.  
14.21 Enhanced support for watchpoints and breakpoints 209  
     
Setting a breakpoint on a template method with multiple instantiations displays a menu  
showing all instantiations and the user can choose to set breakpoints on all or any one  
or none.  
For example,  
(gdb) file test  
Reading symbols from test...done.  
(gdb) b MyClass::MyMember  
[0] cancel  
[1] all  
[2] MyClass::MyMember(int, int) at test.C:14  
[3] MyClass::MyMember(int, float) at test.C:14  
[4] MyClass::MyMember(int, double) at test.C:14  
14.22 Debugging support for shared libraries  
On HP-UX, shared libraries are special. Until the library is loaded, GDB does not know  
the names of symbols. However, GDB gives you two ways to set breakpoints in shared  
libraries:  
deferred breakpoints  
catch load command  
14.22.1 Using shared library as main program  
If the main program is in a shared library and you try to load it as follows:  
(gdb) symbol-file main.sl  
Load new symbol table from "main.sl"? (y or n) y  
Reading symbols from main.sl  
done.  
Things don't appear to work.  
This command is not the correct thing to do. This command assumes that main.slis  
loaded at its link time address. This is not true for shared libraries.  
Do not use symbol-file with shared libraries.  
Instead, what you should do is to use the deferred breakpoint feature to set breakpoints  
on any functions necessary before the program starts running.  
(gdb) b main  
Breakpoint 1 (deferred) at "main" ("main" was not found).  
Breakpoint deferred until a shared library containing "main" is loaded.  
(gdb) r  
Once the program has started running, it will hit the breakpoint. In addition, the de-  
bugger will then already know about the sources for main, since it gets this information  
when the shared library is loaded.  
210 HP-UX Configuration-Specific Information  
   
14.22.2 Setting Deferred Breakpoints in Shared Library  
On HP-UX, GDB automatically loads symbol definitions from shared libraries when  
you use the run command, or when you examine a core file. (Before you issue the run  
command, GDB does not understand references to a function in a shared library| unless  
you are debugging a core file.)  
When you specify a breakpoint using a name that GDB does not recognize, the debugger  
warns you with a message that it is setting a deferred breakpoint on the name you  
specified. If any shared library is loaded with a matching name, then GDB sets the  
breakpoint.  
For example, if you type:  
`break foo'  
the debugger does not know whether foois a misspelled name or whether it is the  
name of a routine that has not yet been loaded from a shared library. The debugger  
displays a warning message that it is setting a deferred breakpoint on foo. If any shared  
library is loaded that contains a foo, then GDB sets the breakpoint.  
If this is not what you want (for example, if the name was mistyped), then you can  
delete the breakpoint.  
14.22.3 Using catch load  
The `catch load <libname>' command causes the debugger to stop when the  
particular library is loaded. This gives you a chance to set breakpoints before routines  
are executed.  
14.22.4 Privately mapping shared libraries  
In cases where you attach to a running program and you try to set a breakpoint in a  
shared library, GDB may generate the following message:  
The shared libraries were not privately mapped; setting a breakpoint  
in a shared library will not work until you rerun the program.  
GDB generates this message because the debugger sets breakpoints by replacing an  
instruction with a BREAK instruction. The debugger cannot set a breakpoint in a shared  
library because doing so can affect other processes on the system in addition to the  
process being debugged.  
14.22 Debugging support for shared libraries 211  
     
To set the breakpoint you must kill the program and then rerun it so that the dynamic  
linker maps a copy of the shared library privately. There are two ways to run the  
program:  
Rerun the program under GDB to have the debugger cause dld to map all shared  
libraries as private, enabling breakpoint debugging.  
On PA-RISC systems, use the following command on an executable:  
/opt/langtools/bin/pxdb -s on executable-name  
The pxdb -son command marks the executable so that dldmaps shared libraries  
as private when the program starts up.  
You can verify the status of the shared library with this command:  
/opt/langtools/bin/pxdb -s status executable-name  
On both PA-RISC and IA64 systems, use the following command on an executable:  
chatr +dbg enable executable-name  
This is similar to the pxdb command described above wherein it directs the dldto load  
the shared libraries as private when the program starts up.  
On HP-UX 11i v3 Integrity systems, WDB enables automatic debugging of shared  
libraries without them being mapped private while attaching to a running program.  
For enabling automatic debugging of shared libraries, you must install the kernel  
patches PHKL_38651 and PHKL_38778.  
14.22.5 Selectively Mapping Shared Libraries As Private  
The -mapsharedoption suppresses mapping all shared libraries in a process private.  
This option enables new functions in the dynamic loader (patch PHSSS_33110 or later)  
to designate individual shared libraries for debugging. By default, HP WDB instructs  
the shared library dynamic loader, dld.sl(5), to map shared libraries in a process  
private, regardless of whether the chatrcommand is run for a particular shlib with  
+dbgor not.  
The -mapsharedoption is used to save virtual memory for debugging applications  
with large amounts of code in shared libraries on machines with simultaneous debug  
sessions. The chatr +dbgoption, and the _HP_DLDOPTSenvironment variable are  
used to identify shared libraries for debugging. The -mapsharedoption ensures that  
the text segments of all other shared libraries is shared across the system. The shared  
libraries are not mapped private and cannot have breakpoints set in them.  
The set mapshared oncommand can be used to change modes from the (gdb)  
prompt.  
(gdb) set mapshared on  
The set mapshared offcommand can be used to load shared libraries after the  
current point is mapped private.  
(gdb) set mapshared off  
212 HP-UX Configuration-Specific Information  
 
The -mapsharedoption is implemented on both PA-RISC and Itanium platforms in  
HP WDB 5.2. This option is provided in the WDB GUI and HP WDB. The default  
behavior does not change if the -mapsharedoption for all shared libraries in processes  
started under the debugger, or dynamically loaded after an attach, are mapped private.  
14.22.6 Setting breakpoints in shared library  
Breakpoints can be set on functions in a shared library, by specifying the library name  
with the function name.  
(gdb) b libutil.sl:fun Breakpoint 1 at 0x79a86228: file simple.c,  
line 13 from /CLO/Components/WDB/build/hppa1.1-hp-hpux11.00/gdb/  
simple_shared/lib.sl  
14.22.7 Enhancement to the info sharedCommand  
HP WDB 6.0 and later versions provide the enhanced info sharedcommand to  
display the global pointer (gp) values along with text start, text end, data start, and  
data end values of the shared library. Following is a sample output of the info shared  
command with the gpvalues:  
tstart  
tend  
dstart  
dend  
gp  
/usr/lib/hpux32/dld.so  
0x2000000065858000 0x20000000658fb080 0x2000000065853000 0x2000000065857a38 0x2000000065855720  
/usr/lib/hpux32/libstd_v2.so.1  
0x20000000656ab000 0x200000006584bfc0 0x200000006569d000 0x20000000656aac28 0x20000000656a6d78  
14.23 Debugging support for Decimal Floating Point data type  
HP WDB versions 5.9 and above enable you to print and evaluate decimal floating  
point data types for programs running on the September 2008 release of HP-UX 11i  
v3, Integrity systems. This feature is available only for compiler versions A.06.20 and  
later.  
The various features supported for decimal floating point data types are as follows:  
Printing Decimal Floating point data types  
— Printing Decimal floating point constant  
— Printing Decimal floating point variable  
Handling Decimal Floating Point data types  
Evaluating Decimal Floating Point data types  
— Printing Type of Decimal floating point data type.  
14.23.1 Printing Decimal Floating point data types  
HP WDB versions 5.9 and above enable you to print a decimal floating point data type  
constant or variable. It handles and prints decimal floating point constant or variable  
when you use common GDB commands such as stack trace and command-line calls.  
14.23 Debugging support for Decimal Floating Point data type 213  
       
14.23.1.1 Printing Decimal floating point constant  
(gdb) print <num><df/dd/dl/DF/DD/DL>  
df, DF - _Decimal32  
dd, DD - _Decimal64  
dl, DL - _Decimal128  
This prints the decimal floating point constant based on the data type.  
14.23.1.2 Printing Decimal floating point variable  
(gdb) print/<fmt> <var-name>  
<fmt> = df, dd, dl  
<var-name> - variable name  
This prints the decimal floating point variable. If you specify format <fmt>then it  
prints the variable based on its data type.  
14.23.2 Printing NaT Registers  
On Integrity systems, following is the command to print the NaT bit corresponding to  
a register:  
print $natr<reg_nbr>  
where, reg_nbrrepresents the NaT bit corresponding to a register.  
Example 14-2 Sample Commands to Print NaT Registers  
Following are sample commands to print NaT registers:  
(gdb) p $natr32  
$1 = 0 '\000'  
(gdb) p $natr33  
$2 = 0 '\000'  
(gdb) p $natr39  
$3 = 0 '\000'  
14.23.3 Handling Decimal Floating Point Data types  
GDB supports decimal floating point values in command line call of functions that  
contain decimal floating point arguments and which return decimal floating point  
arguments.  
(gdb) print func1(1.2dd)  
14.23.4 Evaluating Decimal Floating Point data types  
HP WDB versions 5.9 and above enable you to evaluate the decimal floating point  
variable and displays the output. Use the commonly used GDB commands for evaluating  
and displaying expressions, such as print to evaluate the decimal floating point variable.  
214 HP-UX Configuration-Specific Information  
           
HP WDB supports:  
Evaluation of expressions with decimal floating point constants and variables.  
Setting or assignment of decimal floating point constant or variable.  
Arithmetic operations such as addition, subtraction, multiplication, division, and  
negation with decimal floating point constants or variables.  
Comparison operations such as ==,!=, >, >=, <, <= with decimal floating point  
constants or variables.  
Conversion between data types during assignment, arithmetic and comparison  
operations and while printing values in specified format.  
Assignment of Decimal floating point value to variable:  
(gdb) print <variable> = <dfp-const>/<dfp-var>  
<dfp-const> = decimal floating point constant  
<dfp-var> = decimal floating point variable  
This assigns decimal floating point value to the variable according to its data type.  
Arithmetic Operations:  
(gdb) print a(op1)b  
(gdb) print (op2)a  
op1 = addition, subtraction, multiplication and division.  
op2 = negation  
This performs arithmetic operation with decimal floating point data types.  
Comparison Operations:  
(gdb) p <dfp-val> (op) <dfp-val>  
where,  
op = ==, !=, >, >=, <, <=  
If expression contains comparison operation, then GDB compares the decimal floating  
point data types accordingly.  
(gdb) p 1.2dd == 1.2dd  
(gdb) 2.4  
14.23.4.1 Printing type of Decimal Floating Point variable  
(gdb) ptype <dfp-const>/<dfp-var>  
This prints the type of the Decimal floating point variable or constant.  
(gdb) ptype 1.22dd  
type = _Decimal64  
14.23 Debugging support for Decimal Floating Point data type 215  
 
Conversion of types:  
GDB handles conversion of data types during assignment, printing, and arithmetic  
and comparison operation.  
(gdb) p 1.2df +1.2dd  
This converts double data type (1.2) to _Decimal64data type and performs addition  
operation with _Decimal64data type (1.2dd) and prints value in _Decimal64type.  
HP WDB handles exceptions such as overflow, infinity and division by zero for Decimal  
Floating Point data type.  
(gdb) print 1.2dd / 0  
(gdb) inf  
HP WDB handles finite, infinite and NaN(Not a Number) values of decimal floating  
point data type.  
NOTE: HP WDB will not support:  
Command line calls of intrinsic functions(Mathematical functions like cos, sin, log  
etc)  
Decimal Floating Point data type support for Fortran and C++  
HP-UX 11iv2 Integrity and HP 9000 systems  
14.24 Additional Support for binary floating point data type  
14.24.1 Support for Binary Floating Point constants f, l  
If the binary floating point constant contains the letter f or l then HP WDB recognizes  
it as float or long double binary floating point constant. Generally a floating point  
constant without f or l is considered as double binary floating point constant.  
(gdb) p <num><f/l>  
This prints the binary floating point constant depending upon its data type.  
14.24.2 Support Binary Floating Point variables with format specifier  
HP WDB versions 5.9 and later provide the following format specifiers for binary  
floating point variables, which print the variables in the specified binary floating point  
format:  
(gdb) p/f <var-name>  
This prints the binary floating point value as oat.  
(gdb) p/db <var-name>  
This prints the binary floating point value as double.  
(gdb) p/l <var-name>  
This prints the binary floating point value as long double.  
216 HP-UX Configuration-Specific Information  
     
14.25 Language support  
14.25.1 Enhanced Java Debugging Support  
HP WDB shows stack traces of mixed Java, C, and C++ programs. It supports unwinding  
across Java frames and provides an effective way to examine stack traces containing  
mixed language frames (Java and C/C++) of both live Java processes and core files. This  
has been implemented by adding subcommands for Java VM debugging to gdb. The  
stack trace functionality requires Java SDK version 1.3.1.02 or later versions, for HP-UX.  
To nd the availability of Java SDK version 1.3.1.02 or later, go to the HP web site for  
Java, http://www.hp.com/go/java. Java stack unwind and the gdb Java subcommands  
features are available in gdb version 4.5 and later versions. From gdb version 5.3 and  
later versions, it requires SDK 1.4.2.10 and later versions, or JDK 1.5.0.03 and later  
versions in order to use the Java VM debugging features.  
In order to use this functionality, the GDB_JAVA_UNWINDLIBenvironment variable  
must be set to the path name of the Java unwind library. This environment variable  
must be set for normal java debugging and java corefile debugging. The default location  
of the Java unwind library on various systems is shown following. The examples are  
for SDK 1.4; if you are using JDK 1.5, substitute /opt/java1.5for /opt/java1.4.  
For 32 bit IPF applications,  
export GDB_JAVA_UNWINDLIB = /opt/java1.4/jre/lib/IA64N/server/libjunwind.so  
For 64 bit IPF applications,  
export GDB_JAVA_UNWINDLIB = /opt/java1.4/jre/lib/IA64W/server/libjunwind.so  
For 32 bit PA applications,  
export GDB_JAVA_UNWINDLIB = /opt/java1.4/jre/lib/PA_RISC/server/libjunwind.sl  
For 64-bit PA applications,  
export GDB_JAVA_UNWINDLIB = /opt/java1.4/jre/lib/PA_RISC2.0/server/libjunwind.sl  
export GDB_JAVA_UNWINDLIB = /opt/java1.4/jre/lib/PA_RISC2.0W/server/libjunwind.sl  
If the SDK is installed in a location other than the default, substitute the non-default  
location for /opt/java1.4in the previous commands.  
14.25.1.1 Java Stack Unwind Features  
The Java stack unwind features are useful for troubleshooting problems in the Java  
VM. Following is a list of the Java stack unwind features:  
View mixed language frames information, including Java frames and C/C++ native  
frames, in a gdb backtrace.  
Distinguish various Java frame types including interpreted, compiled, and adapter  
frames.  
View Java method name, signature, and class package name for Java method  
frames.  
14.25 Language support 217  
     
Additional stack unwind features are available starting with SDK 1.4.2. These features  
fall into three categories: Java stack unwind enhancements, Java heap support, and  
Java threads support.  
These additional features are available as part of the Java stack unwind enhancements:  
View Java compiled frame inlined methods.  
View Java interpreted or compiled frame specific information.  
View Java interpreted or compiled frame arguments and local variables.  
Disassemble Java method bytecodes.  
Print out the Java unwind table.  
These additional features are available as part of Java heap support:  
View Java heap parameters.  
Dump Java object.  
Print Java heap histogram.  
Find all the instances of a given Java class.  
Find all the references to a given object in the Java heap.  
Find out the object OOP (object-oriented pointer) of the given field address.  
These additional features are available as part of Java threads support:  
View Java threads state information.  
View current Java thread information.  
View Java interpreted frame monitors information.  
14.25.1.2 gdb Subcommands for Java VM Debugging  
Commands for Examining Java Virtual Machine(JVM) internals  
To view the gdb commands that support Java VM debugging, type help java at the  
gdb prompt.  
(gdb) help java  
Java and JVM debugging commands.  
List of java subcommands:  
java args -- Show the current or specified Java frame arguments info  
java bytecodes -- Disassemble the given Java method's bytecodes  
java heap-histogram -- Show the Java heap object histogram  
java instances -- Find all the instances of the given klassOop in the Java heap  
java jvm-state -- Show Java virtual machine's current internal states  
java locals -- Show the current or specified Java frame locals info  
java mutex-info -- Print out details of the static mutexes  
java object -- Print out the given Java object's fields info  
java oop -- Find the Java object oop of the given Java heap address  
java references -- Find all the references to the given Java object in the Java heap  
java unwind-info -- Show the unwind info of the code where the given pc is located  
java unwind-table -- Print out the dynamically generated Java Unwind Table  
Type help javafollowed by java subcommand name for full documentation.  
Command name abbreviations are allowed if unambiguous.  
218 HP-UX Configuration-Specific Information  
 
Java VM Debugging Commands  
The following commands have been added to enhance Java debugging support:  
brbacktrace  
Prints backtrace of mixed Java and native frames. Standard  
backtrace command of GDB has been enhanced to work with mixed  
Java and native stack frames.  
info frame  
Prints Java frame specific information for a Java frame. Standard  
frame command of GDB has been enhanced to interpret a Java  
stack frame.  
info threads  
thread  
Prints state information for Java threads.  
Prints detailed state information for the current Java thread.  
Java subcommands  
The following Java subcommands have been added:  
java args <frame-number>  
Prints the current or specified Java frame  
arguments information  
java bytecodes <methodOop>  
java heap-histogram  
Disassembles the given Java method bytecode  
Displays the Java heap object histogram  
java instances <klassOop>  
Locates the instances of the given klassOop in  
the Java heap.  
java jvm-state  
java locals  
Prints current status of JVM internal states  
Prints the current or specified Java frame locals  
information  
java mutex-info  
Prints details of static mutexes  
java object <object-ptr>  
Prints the given Java object field information  
java oop <Java_heap_address> Locates Java object oop of the given Java heap  
address  
java references <oop>  
java unwind-info <pc>  
java unwind-table  
Locates references to the given Java object in the  
Java heap  
Prints unwind information of the code where the  
PC is located  
Prints the dynamically generated Java Unwind  
Table  
Type help javafollowed by the subcommand name for full documentation. Command  
name abbreviations are allowed if they are unambiguous.  
14.25 Language support 219  
14.25.1.3 Java corefile debugging support  
HP WDB shows stack traces of mixed Java, C, and C++ programs for java corefile.  
GDB_JAVA_UNWINDLIBenvironment variable must be set to the path name of the Java  
unwind library as explained above.  
Following are examples that illustrate the gdb command-line options for invoking gdb  
on a core file:  
1. Invoke gdb on a core file generated when running a 32-bit Java application on an  
Integrity system with /opt/java1.4/bin/java:  
$ gdb /opt/java1.4/bin/IA64N/java core.java  
2. Invoke gdb on a core file generated when running a 64-bit Java application on an  
Integrity system with /opt/java1.4/bin/java -d64:  
$ gdb /opt/java1.4/bin/IA64W/java core.java  
3. Invoke gdb on a core file generated when running a 32-bit Java application on  
PA-RISC using /opt/java1.4/bin/java:  
$ gdb /opt/java1.4/bin/PA_RISC2.0/java core.java  
4. Invoke gdb on a core file generated when running a 64-bit Java application on  
PA-RISC using /opt/java1.4/bin/java:  
$ gdb /opt/java1.4/bin/PA_RISC2.0W/java core.java  
When debugging a core file, it is good practice to rename the file from core to another  
name to avoid accidentally overwriting it.  
If the Java and system libraries used by the failed application reside in non-standard  
locations, then the GDB_SHLIB_PATHenvironment variable must be set to specify the  
location of the libraries.  
14.25.1.4 Java attach mode debugging support  
HP WDB supports java debugging in attach mode also. GDB_JAVA_UNWINDLIB  
environment variable must be set to the path name of the Java unwind library. From  
gdb version 5.6 and later versions, GDB JAVA UNWINDLIBenvironment variable need  
not be set to the path name of the Java unwind library. HP WDB uses the  
libjunwind.slspecified by the Java Virtual Machine.  
The following examples illustrate how to invoke gdb on a hung process:  
1. Determine the process id:  
$ ps -u user1 | grep java  
23989 pts/9 8:52 java  
220 HP-UX Configuration-Specific Information  
   
2. Attach gdb to the running process:  
$ gdb -p 23989  
HP gdb 5.0 for HP Itanium (32 or 64 bit) and target HP-UX 11.2x.  
Copyright 1986 - 2001 Free Software Foundation, Inc.  
Hewlett-Packard Wildebeest 5.0 (based on GDB) is covered by the  
GNU General Public License.Type "show copying" to see the con-  
ditions to  
change it and/or distribute copies. Type "show warranty" for  
warranty/support.  
Reading symbols from /opt/java1.4/bin/IA64N/java...  
(no debugging symbols found)...done.  
Attaching to program: /opt/java1.4/bin/IA64N/java, process 23989  
(no debugging symbols found)...  
Reading symbols from /usr/lib/hpux32/libpthread.so.1...  
(no debugging symbols found)...done.  
Reading symbols from /usr/lib/hpux32/libdl.so.1...  
...  
NOTE: If the version of gdb on the system is older than version 4.5, it will be necessary  
to specify the full path of the Java executable in order to use the gdb subcommands.  
For example: gdb /opt/java1.4/bin/PA_RISC2.0/java p 23989  
14.25.2 Enhanced support for C++ templates  
This version of HP WDB includes these features to support C++ templates:  
Setting breakpoints in template class functions and template functions without  
having to specify details about the instantiation.  
The ptypecommand shows any one of the class instantiations.  
A option -vin ptype command will now display the field offset and size information  
of a struct/union/classin addition to the default type information.  
Syntax:  
(gdb) ptype -v struct info  
type = struct info /* off 0 bits, len 512 bits */  
int i;  
/* off 0 bits, len 32 bits */  
char a[20];  
/* off 32 bits, len 160 bits */  
struct details d;  
/* off 192 bits, len 256 bits */  
int b : 2;  
/* off 448 bits, len 2 bits */  
int c : 3;  
/* off 450 bits, len 3 bits */  
< filler >  
14.25 Language support 221  
 
/* off 453 bits, len 27 bits */  
float f;  
/* off 480 bits, len 32 bits */  
14.25.3 Support for _ _fpregdata type on IPF  
WDB internally converts __fpregdata type to long double data type to evaluate an  
expression or to print the value of the expression. Since long double data type has only  
15 bit exponent as opposed to 17 bit exponent of __fpreg, some precision is lost when  
the exponent is larger than that can t in 15 bits.  
14.25.4 Support for _Complex variables in HP C  
HP C on Itanium systems supports a _Complexdata type built from any of the floating  
point types.  
A _Complexnumber holds a pair of floating point numbers; the first is the “real part”  
and the second is the “imaginary part”.  
Here are examples of declarations and initializations using _Complexnumbers:  
float _Complex glob_float_complex;  
double _Complex glob_double_complex = 6;  
long double _Complex glob_long_double_complex = _Imaginary_I;  
__float80 _Complex glob_float80_complex = 8 + 9 * _Imaginary_I;  
_Imaginary_I is a keyword which represents the square root of -1.  
The debugger has limited support for _Complexvariables. No arithmetic operations  
are allowed with _Complexnumbers. A _Complexnumber may be cast or assigned  
to any numeric data type and vice versa.  
A _Complexvariable can be initialized with an expression of the form:  
A + B * _Imaginary_I  
where, A and B are ordinary numeric expressions, perhaps in parentheses.  
This is also the format in which the debugger displays a Complex value.  
Imaginary values cannot be assigned to variables because there is no imaginary data  
type. You can take a normal number and multiply it by an imaginary number and get  
another imaginary number. You can take a normal number and add it to an imaginary  
number to get a complex number.  
Complex numbers cannot be used in arithmetic expressions in the debugger.  
For more information of _Complex type, refer to the HP C/ANSI C documentation.  
14.25.5 Support for debugging namespaces  
This release of HP WDB provides full support for debugging namespaces.  
222 HP-UX Configuration-Specific Information  
     
You do not need to use fully qualified names to access symbols within a namespace.  
The debugger can compile a list of namespaces active in the scope that you are in and,  
when possible, choose an appropriate symbol.  
The debugger recognizes using declarations, using directives, namespace aliases, nested  
namespaces, and unqualified lookup within a namespace. It also recognizes using  
directives and aliases, and using declarations within namespaces.  
When the debugger has a list of possible resolutions for a given symbol, it displays a  
menu that shows all names fully qualified whenever namespaces are involved. You  
can choose the appropriate symbol from the list.  
For example, if you stop the debugger in a function that contains an int i using directive  
for a namespace such as:  
using namespace A::AB::ABC::ABCD  
You can use the command print i and if the only possible resolution for i is  
A::AB::ABC::ABCD::i the debugger prints out the name of the symbol and its value. If,  
however, a global i exists, the debugger displays a menu from which to choose:  
(1) i  
(2) A::AB::ABC::ABCD::i  
>
Setting breakpoints on functions works in the same way.  
The debugger also allows semi-qualified names. For example, if you stop in a function  
in namespace B, which is nested in namespace A, and namespace A has an int i, you  
can use print B::i to display the value of A::B::i.  
To disable namespace support, use the command:  
(gdb) set namespaces-enabled off  
14.25.6 Command for evaluating the address of an expression  
The watch_targetcommand takes an expression as an argument, evaluates it to an  
address, and watches the target of that address.  
For example:  
(gdb) watch_target current_frame  
This is equivalent to executing:  
(gdb) print current_frame $1 = (struct frame_info *) 0x7fdf78 (gdb) watch  
*(struct frame_info *) 0x7fdf78  
14.26 Viewing Wide Character Strings  
HP WDB printcommand can print wide characters and wide-character strings of the  
type wchar_t. The user must use the /Woption of the print command to print wide  
characters and wide-character strings.  
print /W <wide-char-symbol-name>  
14.26 Viewing Wide Character Strings 223  
   
14.27 Support for output logging  
The Visual Interface for HP WDB terminal user interface (TUI) mode supports the  
command, log logfile_name, that saves the content of a session to the specified log  
file.  
When you use the logcommand, the debugger saves a snapshot of the current session,  
from the start of the session to the point where you issued the logcommand. Each  
time you use the logcommand, HP WDB overwrites the specified log file with a new  
snapshot from the start of the session.  
To run the Visual Interface for HP WDB, use the following command:  
$vdb -tui  
To redirect HP WDB output to a log file named mylogfile, use the logcommand in  
the following manner:  
(gdb) log mylogfile  
The Visual Interface for HP WDB stores the log file, mylogfile, in the current directory.  
To view the log file from Visual Interface for HP WDB, start a shell process and use  
the following command:  
(gdb) shell vi mylogfile  
14.27.1 Support for dumping array in an ASCII file  
HP WDBN supports dumping an array into an ASCII file.  
The array elements are stored in Array format of Matrix Market in a predefined  
(column-major order for Fortran arrays) order. The objective is to provide a simple  
mechanism to facilitate the exchange of matrix data and to enable easier parsing of the  
array elements. For common file formats, see  
http://math.nist.gov/MatrixMarket/formats.html.  
To dump an array, ARRAY, to a file named DUMPFILE, use the following command:  
(gdb) dump2file ARRAY DUMPFILE  
The entries of ARRAY are dumped into an ASCII file named DUMPFILEin the array  
format. The file is created in the current working directory. The content of the file has  
the following format:  
%%ArrayBrowsing matrix array ARRAY  
% A 5x5 matrix  
5 5  
0
2
4
6
8
2
..  
..  
224 HP-UX Configuration-Specific Information  
   
where, ARRAYis the name of the array, and its size is 5x5.  
The first two lines are comments about this file and the array. The third line denotes  
the array coordinates. From the fourth line, the elements of the array are listed. Note:  
This feature is not supported for the Fortran array slices.  
14.27.2 Support for Fortran array slices  
HP WDB prints Fortran array slices if you specify a range of elements by the Fortran  
90 array section syntax. For instance, for an array Xdeclared by REAL, DIMENSION(-  
1:1, 2:10) :: X, you could print all five even-numbered elements of the row with  
the first dimension equal to 0 by typing the WDB command print X(0,2:10:2).  
14.27.3 Displaying enumerators  
You can display the union of several enumeration elements by specifying a value if the  
elements of the enumeration type are the powers of 2 and the given value is a sum of  
any given combination of the enumeration elements.  
For example, assume you have an enumerated type named color in the program, with  
these elements: RED=0, ORANGE=1, YELLOW=2, GREEN=8, and BLUE=16. If you  
use the command printf 3, the debugger displays ORANGE|YELLOW, the elements  
corresponding to 1 and 2. If you print 5, you will get the value, 5, because it does not  
form the sum of any combination in the set. However, if you wanted to print 25, you  
will get Orange|Green|Blue.  
Values that do not form the sum of any combination of the elements will be displayed  
as integers while the values that form the sum of any combination of the elements will  
be printed as unions.  
14.27.4 Support for debugging typedefs  
When you have a typedef class as a template parameter, you can set a breakpoint on  
a member function by using the command:  
break Class<typedef_classB>::memfunc  
14.27.5 Support for steplast command for C and C++  
Typically, if a function call has arguments that make further function calls, executing  
a simple stepcommand in GDB steps into the argument evaluation call. HP WDB  
includes the steplastcommand, which helps to step into a function, and not into  
the calls for evaluating the arguments. However, the steplastcommand is not  
available on Integrity systems. The following example illustrates how GDB behaves  
when you execute the steplastcommand:  
(gdb) 16 foo (bar ()); ---> bar() will return 10 (gdb) steplast foo (x=10) at  
foo.c:4 4 int k = 10;  
If the steplastcommand is not meaningful for the current line, GDB displays the  
following error message:  
14.27 Support for output logging 225  
       
"Steplast is not meaningful for the current line."  
For example,  
(gdb) 4 int k = 10; (gdb) sl ---> alias to "steplast" command error: Steplast  
is not meaningful for the current line  
To execute the steplastcommand in C++ compiled applications, you must compile  
the application using the HP aC++ version A.03.50 or later with the -g0 option.  
In C++, the steplastcommand is helpful while debugging heavy templated functions,  
because it directly steps into the call, thus skipping the constructor calls, if any. This  
behavior is unlike the stepcommand that steps into the constructor itself.  
Consider the following example:  
void call_me ( string s ) ... (gdb)  
10  
call_me ( "hello" );  
(gdb) steplast call_me (s=static npos = 4294967295,  
static nullref = ref_hdr = mutex_= dummy1 = 0x7f4f79e0, dummy2 = 2136325568,  
refs_ = 2136327612,  
capacity_ = 2136327468, nchars_ = 2136327464, eos_char = 64 '@',  
alloc_ = <No data fields>,  
value_allocator = alloc_ = 0x7f7f133c,  
data_ = 0x40003a64 "hello") at str.C:55  
printf ("Will just print the value of \n");  
If there are multiple top-level calls, the steplastcommand enables you to step into  
each top-level call. For example, for the following line, the steplastcommand takes  
you to the first top-level call, (foo()):  
foo(bar()) + bar(foo());  
Debug foo(), use the finishcommand to exit from the first top-level call, (foo()),  
execute the steplastcommand to step into the next top-level call, (bar()). The  
following example illustrates the use of steplast command:  
(gdb)10 foo( bar() ) + bar( foo() ) (gdb) sl Use the steplast (sl) command to step  
14.28 Getting information from a non-debug executable  
You can get some information about the arguments passed to the functions displayed  
in the stack trace in a non-debug, optimized executable.  
When GDB has no debug information; it does not know where the arguments are  
located or even the type of the arguments. GDB cannot infer this in an optimized,  
non-debug executable.  
However, for integer arguments you can nd the first few parameters for the top-of-  
stack frame by looking at the registers. On PA-RISC systems, the first parameter will  
be in $r26, the second in $r25, and so on. On IPF systems, the first few parameters will  
be in $gr32 and $gr33.  
226 HP-UX Configuration-Specific Information  
 
14.29 Debugging optimized code  
HP WDB supports debugging of optimized code (compiled with both -g and -O) for  
HP aC++, HP ANSI C and HP WDB for HP Itanium.  
The following commands evaluate the name of a function and hence are affected by  
the optimization level of the program being debugged (in particular, due to inlining):  
break  
call  
clear  
disassem  
list  
The following commands evaluate an expression referring to variables in the user pro-  
gram and hence, are affected by the optimization level of the program being debugged:  
break  
call  
cond  
jump  
return  
print  
set <var>  
watch  
whatis x  
NOTE: The breakand callcommands involve evaluation of both the name of a  
function and an expression.  
The following commands are also affected by the optimization level of the program  
being debugged:  
backtrace  
display  
down  
finish  
frame  
info *  
next  
step  
tbreak  
rbreak  
up  
14.29 Debugging optimized code 227  
 
The following commands are not affected by the optimization level of the program  
being debugged:  
attach  
catch  
commands  
continue  
core  
delete  
define  
detach  
disable  
enable *  
file  
forw  
handle *  
help *  
ignore  
kill  
load  
nexti  
path  
quit  
rev  
run  
set args, set env, set <param>  
show args, show <param>  
signal  
source  
stepi  
symbol  
target  
tty  
undisplay  
unset env  
until  
228 HP-UX Configuration-Specific Information  
14.29.1 Debugging Optimized Code at Various Optimization Levels  
The following sections describe debugging optimized code support at each optimization  
level.  
14.29.1.1 +O0and +O1  
At +O1level, optimizations that affect the user visible state of a program are avoided.  
Line numbers are accurately associated with each machine instruction. Global or local  
variables may be examined, except for unused variables (which may be eliminated).  
New values may be assigned to a global and a local variable (set <var> =  
<expression>) when stepping by line (step/next/break <line>). However,  
while stepping by instruction (stepi/nexti) at optimization level +O1, assign a value to  
a variable only if stopped at the very first instruction. This is a must as local  
optimizations are performed within a statement.  
Backtrace commands (backtrace) may be used to display the current nest of function  
calls, including for calls that are inlined. Note that even at +O1, C++ methods that are  
defined within a class and Fortran arithmetic statement functions are implicitly inlinable  
and are inlined. Other functions are not inlined, regardless of the inline pragmas or  
keywords.  
14.29.1.2 +O2/+O3/+O4/-ipo  
Stepping by line number (step/next) and running to a breakpoint(break) moves the  
state of a program forward. However, the program execution does not necessarily stop  
at the given line.  
You can set breakpoints (break) at the entry to a routine that is not inlined and examine  
the values of parameters when the program execution stops at the entry of a routine.  
The local variables can be examined within a function. However, the values of the local  
variables may not be available at all code locations in the function. Assignment of new  
values to formal parameters or local variables is NOT supported in code compiled with  
optimization above +O1.  
Optimization of code results in the reordering of the instructions and the source line-  
numbers. Hence, the value of the variable, which is printed by the debugger may not  
correspond to the reported source code location. The debugger may print the value of  
the variable at a source code location either before or after the reported source code  
location.(If the printed value is not current with respect to the current source line, the  
printed value will be the immediately previous or immediately later value for the  
variable.)  
Backtrace commands (backtrace) can be used to display the current nest of function  
calls, including calls that are inlined. When stopped within the code for an inlined call,  
the parameters and the local variables of the inlined routine are not reported or available.  
The disassemcommand does not work for functions that have no code (because all  
calls to these functions are inlined or these functions are not called at all). HP WDB 5.7  
14.29 Debugging optimized code 229  
     
and later versions provide support to prevent the debugged program from stopping  
at instructions that are predicated false. The program execution can be stopped by a  
software breakpoint, a hardware breakpoint, or an asynchronous signal. In the case of  
optimizations such as if-conversion, the predicated false instructions indicate that an  
alternate source path is executed. Hence, stopping the program at a predicated false  
instruction results in the misleading conclusion that the path corresponding to the  
predicated false instruction is executed. To prevent this ambiguity, HP WDB does not  
stop at predicated false instructions.  
The predicated false instructions are equated to NOPs(No OPeration), because these  
instructions do not modify the processor state. The exception to this rule is the use of  
certain instructions, such as wtop, wexit, and frcpa, which modify the processor  
state even when predicated false. In such cases, the debugger stops at the instructions  
irrespective of the predicate value of the instructions. Assembly and low-level  
programmers, who require the old behavior of the debugger to stop at the instructions  
irrespective of the predicate value of these instructions, can explicitly turn o this feature.  
To explicitly turn o this feature, enter the following command at the gdb prompt:  
(gdb) set no-predication-handling  
The following limitations apply when debugging optimized code:  
Support for high-level loop transformations such as modulo-scheduled loops, or  
LNO-optimized loop nests is limited. (This limited support includes all loop  
optimizations that are enabled at +O3and above, and some loop optimizations at  
+O2or -O.)  
Debug support for local aggregates and arrays is limited.  
Complete debug support for inlined subroutines is not available.  
Values that are not at the current code location will be reported as being  
unavailable, even if these values can be computed from some other values that  
are available.  
Step operations may include occasional "backwards" steps, because of the re-  
ordered code during optimization.  
The program stops at asynchronous signal stops even if the reported instruction  
is predicated false.  
Complete support is available for debugging at the assembly language level. Stepping  
by instructions (stepi/nexti) steps as expected and reports the associated source line  
numbers for each instruction.  
NOTE: The -ipocompilation implies the +noobjdebugoption because the -ipo  
object files do not store executable code or debug info.  
14.30 Debugging with ARIES  
The ARIES fast interpreter emulates a complete set of non-privileged PA-RISC  
instructions with no user intervention. During interpretation, it monitors the applications  
230 HP-UX Configuration-Specific Information  
 
execution pattern and translates only the frequently executed code into native Itanium  
(R)code at runtime.  
14.30.1 Debugging the application using GDB under ARIES  
ARIES supports debugging of HP 9000 HP-UX applications on HP-UX 11i Integrity  
servers using the HP 9000 HP-UX GDB.  
Both the GDB and the application run under ARIES.  
No change in GDB user interface (including WDB GUI).  
Negligible loss of performance in interactive mode.  
The HP 9000 GDB is included by default as part of the Integrity HP-UX WDB/GDB  
package.  
All GDB commands work just like they would on an HP 9000 server.  
Use the following steps to debug HP 9000 HP-UX applications on HP-UX 11i Integrity  
servers using GDB:  
1. Set the environment variable PA_DEBUG to 1.  
2. Set the environment variable SHELL to point to an HP 9000 shell, copied from an  
HP 9000 HP-UX system from /usr/bin path.  
3. Add /usr/ccs/binto the PATH environment variable.  
4. Invoke GDB as:  
gdb PA-RISC_executable  
After the debugging is finished, perform the following steps:  
1. Unset the environment variable PA_DEBUG.  
2. Restore the original value of the SHELL environment variable.  
3. The rest of the debugging operations are the same as that on the HP 9000 HP-UX  
platform.  
NOTE: Make sure that the user has write permission on /tmpdirectory and that  
there is enough space to create a temporary file of one page size as obtained by  
sysconf(_SC_PAGE_SIZE)system call.  
14.30.1.1 Limitations of GDB Support under ARIES  
No support for debuggers other than HP 9000 HP-UX GDB for debugging HP  
9000 applications under ARIES on HP-UX 11i Integrity servers.  
No support for old GDB versions (of HP-UX 10.20 and earlier). However, debugging  
of HP-UX 10.20 applications using a HP-UX 11.0 (and newer) HP 9000 GDB is  
supported.  
HP 9000 GDB behaves differently for child processes created using fork()and  
vfork()system calls. ARIES emulates fork()and vfork()system calls  
14.30 Debugging with ARIES 231  
   
identically. The exact behavior shown by HP 9000 GDB under ARIES may differ  
from that on a HP 9000 HP-UX server.  
If the debugged process is blocking in a system call, any attempt to get to the GDB  
command prompt by pressing ctrl-C does not work. The process needs to be killed  
from a different shell.  
ARIES does not provide true emulation of MxN threads, and thus does not support  
debugging of HP 9000 applications that are linked with pthreads library and create  
threads in MxN model.  
NOTE: The HP 9000 applications linked with MxN pthreads library are emulated  
under ARIES as traditional 1x1 threads, and thus can only be debugged under  
ARIES as any other non-MxN multi-threaded application.  
ARIES supports debugging of 32-bit and 64-bit HP 9000 HP-UX applications using  
32-bit HP 9000 HP-UX GDB. 64-bit HP 9000 HP-UX gdb is not supported under  
ARIES.  
14.30.2 Attaching GDB to an already running emulated process  
You can attach GDB to an already running HP 9000 application process on HP-UX 11i  
for HP Integrity servers to debug the application. After a successful attach, all of the  
GDB commands work in exactly the same manner as they do on an HP 9000 HP-UX  
server.  
Use the steps below to attach GDB to HP 9000 application process under ARIES on  
HP-UX 11i Integrity servers:  
Perform the same preparatory steps as required for debugging an HP 9000 HP-UX  
application using GDB under ARIES on HP-UX 11i Integrity server.  
Invoke GDB as follows:  
gdb <pa_process_name> <pa_process_name>  
You can also invoke GDB through any of the attach modes supported by GDB  
using the -pid <pa_pid>option or by issuing the attach <pa_pid>command  
at the gdbprompt  
After the debugging, the process can continue or abort, as specified by the user. This  
feature is especially useful when triaging problems in an environment with a large  
number of processes and a mix of Integrity native and HP 9000 processes.  
14.30.3 Detecting memory leaks using GDB under ARIES  
Applications cannot leak memory under ARIES unless they do so on HP 9000 servers.  
HP GDB can be used to detect memory leaks of HP 9000 applications under ARIES.  
Refer to HP GDB documentation for more details.  
232 HP-UX Configuration-Specific Information  
   
14.31 Visual Interface for WDB  
WDB includes an HP-supported Visual Interface for WDB with both graphical and  
terminal modes. The interface is based on Vim 5.7 and WDB. This interface replaces  
the -tuimode on Itanium-based systems.  
When you use the interface you are actually using vim, which is a vi-compatible editor.  
With the interface you can use vi commands to browse in the WDB display.  
Most of Visual Interface for WDB functionality is also available for emacs users. Visual  
Interface for WDB does not require knowledge of vi commands.  
Visual Interface for WDB identifies you as an emacs user by looking at the environment  
variable `$EDITOR'. If this variable has a value that matches emacs, or gmacs, or xemacs,  
then Visual Interface for WDB starts in emacs mode automatically.  
NOTE: If the program expects unbuffered input or uses curses, termcap, or terminfo,  
or otherwise transmits escape or control sequences to the terminal, you must use one  
of the following methods to run Visual Interface for WDB:  
Start the process in one terminal and attach to it with Visual Interface for WDB.  
Use the ttycommand at the debugger prompt so the program's input and output  
are directed to another terminal.  
NOTE: If the underlying GDB terminates abnormally when you are using Visual  
Interface for WDB, do not close the Visual Interface for WDB window. Wait for a minute  
or two. Visual Interface for WDB captures the stack trace and the debugging session  
details and sends you an email. You can then forward this to HP when you report the  
problem. This is helpful to HP in reconstructing the crash scenario.  
14.31.1 Starting and stopping Visual Interface for WDB  
You can use Visual Interface for WDB in either of two modes:  
X-window-based graphical interface: Supports mouse and keyboard commands.  
Terminal interface: Supports keyboard commands only.  
Visual Interface for WDB accepts the same command line arguments as GDB so you  
can add options to the startup command. See the man page for GDB for the list of  
arguments.  
To start Visual Interface for WDB in graphical mode with mouse support, run  
Visual Interface for WDB with the command:  
/opt/langtools/bin/vdb  
To start Visual Interface for WDB in terminal user interface mode, run Visual  
Interface for WDB with the command:  
/opt/langtools/bin/vdb -tui  
To stop Visual Interface for WDB, type quiton the WDB command line:  
14.31 Visual Interface for WDB 233  
   
(wdb) quit  
14.31.2 Navigating the Visual Interface for WDB display  
The Visual Interface for WDB window consists of two areas:  
Source pane at the top  
Debugger pane at the bottom  
You can use the arrow and pagination keys on the keyboard to move the cursor:  
Pagination keys move the cursor in the source window, at the top, above the status  
line.  
Holding the shift key down while using the pagination keys moves the cursor in  
the debugger window.  
The up and down arrow keys move the cursor in the source window.  
Holding the shift key down while using the up and down arrow keys move the  
cursor in the debugger window.  
The left and right arrow keys move the cursor in the debugger window.  
Two rows of labeled softkeys at the bottom of the display give you quick access  
to common commands.  
Visual Interface for WDB GUI display  
33 #include <stdlib.h>  
34 #include "Deck.h"  
35 #include "Player.h"  
36 #include "House.h"  
37  
38 int main ()  
39 {  
*> 40 srand ((int) time(0));  
41  
42 Deck theDeck;  
43 Player thePlayer (100);  
44 House theHouse (16);  
45  
46 theHouse.Instructions();  
47 }  
File: BlackJack.C  
Function: main Line: 40  
Pc: 0x3ea0  
(wdb) b main  
Breakpoint 1 at 0x3ea0: file BlackJack.C, line 40.  
(wdb) run  
Starting program: /work/wdb/blackjack/blackjack  
Breakpoint 1, main () at BlackJack.C:40  
40 srand ((int) time(0));  
(wdb)  
234 HP-UX Configuration-Specific Information  
 
Run  
Faq  
Resume Stop Up Visual  
Interface for  
WDB  
Finish Print Type List  
Stop  
Next Down  
Prompt Print* Edit Credits  
You can click the softkey or press a function key on the keyboard to invoke the  
command.  
The function keys F1 through F8 correspond to the bottom row of softkeys. The function  
keys F9 and up correspond to the top row.  
14.31.3 Specifying foreground and background colors  
To change the foreground and background colors, update the `.Xdefaults' file in the  
home directory. The resources are the same as for `hpterm'.  
Here is a sample entry:  
HPterm*foreground: white  
HPterm*background: rgb:68/90/C4  
14.31.4 Using the X-window graphical interface  
To start Visual Interface for WDB in graphical mode with mouse support, run Visual  
Interface for WDB with the command:  
/opt/langtools/bin/vdb  
Visual Interface for WDB opens an `hpterm' window, ignoring the value of the TERM  
environment variable, for debugging a program.  
With a mouse you can do the following:  
Left-click the line number to insert or remove breakpoints.  
Left-click an identifier to select the identifier as an operand for the Print, Print*,  
Type, and List softkeys.  
Where necessary, manually select an expression by dragging the cursor over it.  
Right-click the line number to activate a pop-up menu with several useful  
commands.  
Right-click an identifier to automatically select it and use the selection as an operand  
for the pop-up window that appears.  
Right-click an empty region for a third pop-up menu with several useful actions.  
Left-click the command softkeys at the bottom of Visual Interface for WDB window.  
Click the middle button to paste the selection.  
Drag the status bar with the mouse to resize the debugger window relative to the  
source window.  
14.31 Visual Interface for WDB 235  
   
14.31.5 Using the TUI mode  
To start Visual Interface for WDB in terminal user interface (TUI) mode, run Visual  
Interface for WDB with the command:  
/opt/langtools/bin/vdb -tui  
This mode works well with hptermand xtermand fairly well with dttermand VT100  
(telnet) terminals.  
NOTE: A defect in dttermmay truncate the display of lines that do not t within the  
window. To work around this defect, refresh the display with CTRL-L or widen the  
terminal window so source lines do not wrap.  
If you use xtermand dtterm, update the `.Xdefaults' file with keyboard translations  
to get the shifted arrows and shifted paging keys to work.  
For xterm, use the following:  
XTerm*vt100.translations: #override \  
Shift <Key>Prior: string(0x2) \n \  
Shift <Key>Next: string(0x6) \n \  
Shift <Key>Up: string(0x5) \n \  
Shift <Key>Down: string(0x19) \n \  
Shift <Key>Left: string(0x1b) string(i) \n \  
Shift <Key>Right: string(0x1b) string(la)  
For dterm, use the following:  
*DtTerm*Translations: #override\n \  
Shift <Key>osfPageUp: string(0x2) \n \  
Shift <Key>osfPageDown: string(0x6) \n \  
Shift <Key>osfUp: string(0x5) \n \  
Shift <Key>osfDown: string(0x19) \n \  
Shift <Key>osfLeft: string(0x1b) string(i) \n \  
Shift <Key>osfRight: string(0x1b) string(la)  
Mouse operations are not supported in the -tuimode. Also the paging and shift keys  
do not work well with VT100.  
14.31.6 Changing the size of the source or debugger pane  
1. Escape to vicommand mode first.  
2. Drag the status bar using the mouse.  
If you are using -tuimode, use these commands to change the size of the current  
window:  
CTRL-W +  
to increase  
CTRL-W -  
to decrease  
In Visual Interface for WDB, the current window is usually the debugger window.  
236 HP-UX Configuration-Specific Information  
   
14.31.7 Using commands to browse through source files  
browse. For example, CTRL-B, CTRL-F, CTRL-D, CTRL-Uare useful for browsing the  
debugger window. These commands work whether or not you escape to `vi' mode.  
These `vim' commands require you to escape to `vi' mode. For example:  
`/'  
Search forward  
`?'  
`n' `N'  
`%'  
Search backward  
Repeat search  
Match braces  
`[[', `]]'  
`:line number'  
Skip to the next procedure  
Go to any line number  
All these commands require you to escape to `vi' command mode first. When you are  
done, type a for append or i for insert or other `vi' commands to return to text insertion  
mode.  
Or you can simply click the Prompt softkey.  
14.31.8 Loading source files  
Escape to `vi' command mode and use the :e command to load a source file.  
:e filename  
When the source files are located in multiple directories, you can simply specify the  
base name alone as long as file names are unique and the appropriate dir commands  
have been executed.  
Pressing the Prompt softkey takes you to the command prompt and also updates the  
source window so that the cursor remains where the program is stopped.  
14.31.9 Editing source files  
To edit a file, kill the process then click the Edit button. If you do not kill the process,  
the source file and binaries can get out of sync.  
14.31.10 Editing the command line and command-line history  
Visual Interface for WDB preserves the entire session's transactions so you can browse  
through these at any time.  
To edit the command line, press ESC to enter vi mode and then use vi commands. You  
can recall previous commands in history by using [jk^P^N]. Complete command  
lines using TAB.  
14.31.11 Saving the contents of a debugging session to a file  
You can save the contents of the current debugging session, including debugger  
input/output and program input/output to a file.  
14.31 Visual Interface for WDB 237  
         
To save a session to a file:  
1. Right-click an empty region of the source or debugger pane.  
2. Choose "Save Session to vdb.pid" from the pop-up menu.  
The debugger writes the input and output to a file whose name ends in the pid of the  
debugger. If you save the session more than once, the new transactions are appended  
to the file.  
14.32 Support for ddd  
GDB works with ddd, the free GDB GUI shell available at http://mumm.ibr.cs.tu-bs.de/.  
While this is not formally supported by Hewlett-Packard, these two do work together.  
Note however if you have dddissues, you'll need to report them to the dddsupport  
channel.  
14.33 Support for XDB commands  
HP WDB provides support for a subset of XDB commands, enabled with the -xdb  
option.  
14.33.1 stop in/at dbxcommands  
The commands <stop in function/address>and <stop at line>are  
equivalent of dbx <break function/address /line>command. WDB supports  
the <stop in/at>command in non-dbx mode.  
For example:  
$ gdb a.out (gdb) stop in main Breakpoint 1 at 0x2a34: file list.c, line 18  
from /tmp/a.out (gdb) stop at 25 Breakpoint 2 at 0x2a6c: file list.c, line  
25 from /tmp/a.out (gdb)  
14.34 GNU GDB Logging Commands  
The following commands control GDB logging activities:  
set logging file: Set the current log file  
set logging off: Set logging off  
set logging on: Set logging on  
set logging overwrite[on|off]: Set whether logging overwrites or appends  
to the log file.  
set logging redirect [on|off]: Set the logging output mode  
14.35 Support for command line calls in a stripped executable  
HP WDB enables you to perform command line calls in a stripped executable.  
238 HP-UX Configuration-Specific Information  
         
14.35.1 Support for command line calls in a stripped executable on PA-RISC systems  
In WDB, to perform command line calls in a shared library without the help of dynamic  
linker (using end.o), you must execute the following command:  
chatr -B immediate <executable>  
In addition, modify all the calls to shl_load()to specify BIND_IMMEDIATE.  
To perform command line calls after attaching a running process to GDB, you must  
execute one of the following commands:  
/opt/langtools/bin/pxdb -s on <executable>  
chatr +dbg enable <executable>  
14.35.2 Additional support for command line calls in a stripped executable  
HP WDB enables you to perform command line calls in a stripped executable. The  
various scenarios in which you can make command line calls in a stripped executable  
are as follows:  
14.35.2.1 For 32-bit applications:  
To perform command line calls in a shared library, without the help of dynamic linker  
(using end.o), you must perform the following operations:  
Execute the chatr -B immediate <executable>command.  
Modify all the calls to shl_load()to specify BIND_IMMEDIATE, if any.  
To perform command line calls after attaching GDB to a running process, without the  
help of dynamic linker (using end.o), you must do the following for the program:  
Execute the chatr -B immediate <executable>command  
Modify all the calls to shl_load()to specify BIND_IMMEDIATE, if any.  
Execute the /opt/langtools/bin/pxdb -s on <executable>or chatr  
+dbg enable <executable>command.  
On HP-UX 11i v3 Integrity systems, WDB enables automatic debugging of shared  
libraries without them being mapped private while attaching to a running program.  
For enabling automatic debugging of shared libraries, you must install the kernel  
patches PHKL_38651 and PHKL_38778.  
To avoid changing of the run-time binding behavior of a program to BIND_IMMEDIATE,  
to perform command line call, do the following:  
Use the linker option, +ea, to export symbols from an object file.  
Install the linker patch, PHSS 28870 (for 11.0) or PHSS 28871 (for 11i).  
Execute the following commands:  
cc -c file.c cc file.o -Wl,+ea,/opt/langtools/lib/end.o -s  
14.35 Support for command line calls in a stripped executable 239  
     
14.35.2.2 For 64-bit applications  
To perform command line calls in a stripped executable, linked with end.o, you need  
to do the following:  
In the +stdlink mode, GDB supports this feature without any changes. You must  
export the __wdb_call_dummysymbol as shown in the next line.  
In the +compatlink mode, execute the following command:  
cc +DD64-g file.c -Wl,+ee,__wdb_call_dummy -s  
14.35.3 Support for debugging stripped binaries  
HP WDB provides limited support for debugging stripped binaries.  
14.35.3.1 Printing of locals and globals in a stripped module  
GDB will not be able to print the locals and statics declared in a module which has  
been stripped. GDB will be able to print the exported symbols since exported symbols  
are not stripped with strip command (they stay in .dynsym).  
GDB will be able to access the globals or locals defined in other unstripped shared  
libraries loaded into the stripped executable when you are in the right scope.  
14.35.3.2 Backtrace on stripped frames  
GDB should be able to backtrace properly stripped frames. Arguments will not be  
displayed (as in the case of non -gbinary). If it is a fully archived stripped binary,  
function names will not be displayed (but PCs will be).  
14.35.3.3 Command line calls to non-stripped library  
Command line calls to the functions (exported symbols) in the stripped binary work  
fine. Command line calls to the non-stripped library work normally regardless where  
the process is stopped.  
14.35.3.4 Setting breakpoints in unstripped shared library  
GDB will not be able to put breakpoints using symbolic names(of the symbols not in  
export list) or line numbers in the stripped modules.  
GDB will be able to place breakpoints using symbol names in the unstripped shared  
libraries loaded into the stripped executable.  
240 HP-UX Configuration-Specific Information  
           
14.36 Displaying the current block scope information  
The whichcommand takes a symbol as an argument and prints the information on a  
given symbol. It prints the following information:  
current block scope addresses  
line information of the definition of the symbol  
filename in which the definition of the symbol occurs  
The whichcommand does not work for global and type symbols since they do not  
contain line information.  
Syntax:  
which <symbol>  
For example :  
(gdb) which i Line 4 of "example.c" block starts at address 0x29a8 <main> and ends at 0x29e4  
<main+0x3c>  
14.37 Linux support  
Linux Runtime Environment (LRE) on HP-UX Itanium enables users to execute Intel  
Itanium Linux applications on HP-UX. HP WDB provides a prototype for LRE  
debugging, which allows you to debug applications ported from Linux that run under  
LRE. This provides a minimal debugging capability for LRE.  
14.36 Displaying the current block scope information 241  
   
242  
15 The HP-UX Terminal User Interface  
By default, GDB runs in line mode. For users who prefer an interface similar (though  
not identical) to that of the XDB debugger, HP provides a terminal user interface (TUI),  
which appears when you invoke the gdb command with the -tui option.  
Use the -xdboption to enable the use of a number of XDB commands. See the “XDB  
15.1 Starting the TUI  
Invoke the debugger using a command like the following:  
gdb -xdb -tui a.out  
These examples use the default terminal screen size of 24 by 80 characters. Following  
is a sample terminal screen window:  
|----------------------------------------------------------------------|  
|30  
|31  
|32  
|33  
|34  
|35  
|36  
|37  
|38  
|39  
|40  
|41  
|42  
{
|
|
|
|
|
|
|
|
|
|
|
|
|
/* Try two test cases. */  
print_average (my_list, first, last);  
print_average (my_list, first, last - 3);  
}
|----------------------------------------------------------------------|  
File: average.c Procedure: ?? Line: ?? pc: ??  
Wildebeest is free software, covered by the GNU General Public License, and  
you are welcome to change it and/or distribute copies of it under certain  
conditions. Type "show copying" to see the conditions. There is  
absolutely no warranty for Wildebeest. Type "show warranty" for details.  
---Type <return> to continue, or q <return> to quit---  
Wildebeest was built for PA-RISC 1.1 or 2.0 (narrow), HP-UX 11.00.  
..  
(gdb)  
The terminal window is divided into two panes: a Source pane at the top and a  
Command pane at the bottom. In the middle is a locator barthat shows the current  
file, procedure, line, and program counter (PC) address, when they are known to the  
debugger.  
When you set a breakpoint on the main program by issuing the command  
b main  
an asterisk (*) appears opposite the first executable line of the program.  
15.1 Starting the TUI 243  
   
When you execute the program up to the first breakpoint by issuing the command  
run  
a right angle bracket (>) points to the current location. So after you issue those  
commands, the window looks something like this:  
|----------------------------------------------------------------------|  
|27  
|28  
|29  
|30  
|31  
}
|
|
|
|
|
|
|
|
|
|
|
|
|
int main(void)  
{
/* Try two test cases. */  
print_average (my_list, first, last);  
print_average (my_list, first, last - 3);  
*>|32  
|33  
|34  
|35  
|36  
|37  
|38  
|39  
}
|----------------------------------------------------------------------|  
File: average.c  
..  
Procedure: main  
Line: 32  
pc: 0x3524  
(gdb) b main  
Breakpoint 1 at 0x3524: file average.c, line 32.  
(gdb) run  
Starting program: /home/work/wdb/a.out  
Breakpoint 1, main () at average.c:32  
(gdb)  
15.2 Automatically running a program at startup  
WDB does not start running the target executable at startup as do `xdb' and HP DDE.  
This makes it easy to set break points before the target program's main function.  
To make WDB automatically start running the target program add these lines to your  
startup file, .gdbinit:  
break main  
run  
15.3 Screen Layouts  
The TUI supports four panes within the terminal window, in various combinations:  
Command  
Source  
Disassembly  
Register  
244 The HP-UX Terminal User Interface  
   
The Command pane is always present. The possible configurations of the other panes  
are:  
Source  
Disassembly  
Source/Disassembly  
Disassembly/Register  
Source/Register  
The layoutcommand (abbreviated la) enables you to change from one window  
configuration to another.  
NOTE: You can abbreviate any command to its shortest unambiguous form.  
15.3.1 Source pane  
The Source pane, Figure 1, appears by default when you invoke the debugger. You can  
also make it appear by issuing the command  
la src  
15.3.2 Disassembly pane  
The Disassembly pane appears when you issue the command  
la asm  
The pane looks like this:  
|----------------------------------------------------------------------|  
|;;;  
print_average (my_list, first, last);  
|
|
|
|
|
|
|
|
|
|
|
|
|
*>|0x3524 <main+8> addil L'-0x800,%dp,%r1  
|0x3528 <main+12>  
|0x352c <main+16>  
|0x3530 <main+20>  
|0x3534 <main+24>  
|0x3538 <main+28>  
|0x353c <main+32>  
ldo 0x730(%r1),%r26  
ldi 9,%r24  
ldi 0,%r25  
ldil L'0x3000,%r31  
be,l 0x498(%sr4,%r31)  
copy %r31,%rp  
|;;; print_average (my_list, first, last - 3);  
|0x3540 <main+36>  
|0x3544 <main+40>  
|0x3548 <main+44>  
|0x354c <main+48>  
addil L'-0x800,%dp,%r1  
ldo 0x730(%r1),%r26  
ldi 6,%r24  
ldi 0,%r25  
|----------------------------------------------------------------------|  
File: average.c  
(gdb) b main  
Procedure: main  
Line: 32  
pc: 0x3524  
Breakpoint 1 at 0x3524: file average.c, line 32.  
(gdb) r  
Starting program: /home/work/wdb/a.out  
Breakpoint 1, main () at average.c:32  
(gdb) la asm  
(gdb)  
15.3 Screen Layouts 245  
   
15.3.3 Source/Disassembly pane  
The Source/Disassembly pane appears when you issue the command:  
la split  
You can also reach this pane from the Source pane with the XDB command:  
td  
The window looks like this:  
:......................................................................:  
*>:32  
:33  
print_average (my_list, first, last);  
print_average (my_list, first, last - 3);  
:
:
:
:
:
:
:34  
:35  
:36  
:37  
}
:......................................................................:  
|;;;  
print_average (my_list, first, last);  
|
|
|
|
|
|
*>|0x3524 <main+8> addil L'-0x800,%dp,%r1  
|0x3528 <main+12>  
|0x352c <main+16>  
|0x3530 <main+20>  
|0x3534 <main+24>  
ldo 0x730(%r1),%r26  
ldi 9,%r24  
ldi 0,%r25  
ldil L'0x3000,%r31  
|----------------------------------------------------------------------|  
File: average.c Procedure: main Line: 32 pc: 0x3524  
Breakpoint 1 at 0x3524: file average.c, line 32.  
(gdb) r  
Starting program: /home/work/wdb/a.out  
Breakpoint 1, main () at average.c:32  
(gdb) la asm  
(gdb) la split  
(gdb)  
15.3.4 Disassembly/Register pane  
The Disassembly/Register pane appears when you issue the command  
la regs  
when the current pane is the Source/Disassembly pane. By default, the debugger  
displays the general registers.  
The window looks like this:  
:.........................................................................:  
:flags 29000041  
:r3 7f7f0000  
:r6 7f7f06fc  
:r9 40006b10  
:r12 1  
r1 51a800  
r4 1  
r7 7f7f0800  
r10 0  
r13 0  
r16 40003fb8  
rp 7f6ce597  
r5 7f7f06f4  
r8 7f7f0800  
r11 40004b78  
r14 0  
:
:
:
:
:
:
:r15 0  
r17 4  
:.........................................................................:  
246 The HP-UX Terminal User Interface  
   
|;;;  
print_average (my_list, first, last);  
|
|
|
|
|
|
*> |0x3524 <main+8> addil L'-0x800,%dp,%r1  
|0x3528 <main+12>  
|0x352c <main+16>  
|0x3530 <main+20>  
|0x3534 <main+24>  
ldo 0x730(%r1),%r26  
ldi 9,%r24  
ldi 0,%r25  
ldil L'0x3000,%r31  
|----------------------------------------------------------------------|  
File: average.c  
(gdb) r  
Procedure: main  
Line: 32  
pc: 0x3524  
Starting program: /home/work/wdb/a.out  
Breakpoint 1, main () at average.c:32  
(gdb) la asm  
(gdb) la split  
(gdb) la regs  
(gdb)  
15.3.5 Source/Register pane  
The Source/Register pane appears when you issue the command:  
la regs  
when the current pane is the Source pane.  
The screen looks like this:  
:.........................................................................:  
:flags 29000041  
:r3 7f7f0000  
:r6 7f7f06fc  
:r9 40006b10  
:r12 1  
r1 51a800  
r4 1  
r7 7f7f0800  
r10 0  
r13 0  
r16 40003fb8  
rp 7f6ce597  
r5 7f7f06f4  
r8 7f7f0800  
r11 40004b78  
r14 0  
:
:
:
:
:
:
:r15 0  
r17 4  
:.........................................................................:  
*>|32  
|33  
print_average (my_list, first, last);  
print_average (my_list, first, last - 3);  
|
|
|
|
|
|
|34  
|35  
|36  
|37  
}
|----------------------------------------------------------------------|  
File: average.c  
Procedure: main  
Line: 32  
pc: 0x3524  
Breakpoint 1, main () at average.c:32  
(gdb) la asm  
(gdb) la split  
(gdb) la regs  
(gdb) la src  
(gdb) la regs  
(gdb)  
15.3 Screen Layouts 247  
 
15.4 Cycling through the panes  
Use the commands  
la next  
and  
la prev  
to move from one pane to another without specifying a window name. If you specify  
la nextrepeatedly, the order the debugger uses is:  
Source (src)  
Disassembly (asm)  
Source/Disassembly (split)  
Source/Register  
Disassembly/Register  
If you invoked the gdb command with the -xdboption as well as the -tuioption,  
you can also use the following commands:  
td Toggle between Source and Disassembly/Register panes.  
td Toggle split-screen mode.  
15.5 Changing pane focus  
The command pane always has keyboard focus, so that you can enter debugger  
commands. If there is only one other pane (Source or Disassembly), the other pane has  
the logical focus, so that you can scroll within that pane by using the arrow keys or the  
Page Up and Page Down keys (on some keyboards these are Prev and Next).  
NOTE: In the command pane, the scrolling behavior only works for an `hpterm' and  
not for an `xterm' or `dtterm'.  
If you are in split-screen mode, you may want to change the logical focus of the pane.  
To do so, use the command focus  
{win_name | prev | next}  
where win name can be src, asm, regs, or cmd.  
Remember, if you change the focus to a pane other than the command pane, you need  
to use focus cmdto switch back to the command pane to enter or scroll through  
commands.  
For example, with the sequence of commands just issued, you are in split-screen mode  
with the focus in the Source pane.  
The pane with logical focus has a border constructed from "|" and "-".  
A pane that does not have logical focus has a border constructed from ":"vand ".":  
248 The HP-UX Terminal User Interface  
   
:.........................................................................:  
:flags 29000041  
:r3 7f7f0000  
:r6 7f7f06fc  
:r9 40006b10  
:r12 1  
r1 51a800  
r4 1  
r7 7f7f0800  
r10 0  
r13 0  
r16 40003fb8  
rp 7f6ce597  
r5 7f7f06f4  
r8 7f7f0800  
r11 40004b78  
r14 0  
:
:
:
:
:
:
:r15 0  
r17 4  
:.........................................................................:  
*>|32  
|33  
print_average (my_list, first, last);  
print_average (my_list, first, last - 3);  
|
|
|
|
|
|
|34  
|35  
|36  
|37  
}
|----------------------------------------------------------------------|  
File: average.c Procedure: main Line: 32 pc: 0x3524  
Breakpoint 1, main () at average.c:32  
(gdb) la asm  
(gdb) la split  
(gdb) la regs  
(gdb) la src  
(gdb) la regs  
(gdb)  
By default, the Source pane can scroll. To change the focus so that you can scroll in the  
Register pane, use the focuscommand (abbreviated focor fs):  
fs regs  
or  
foc next  
If you then use the Page Down key to scroll in the Register pane, the window looks  
like this:  
|-------------------------------------------------------------------------|  
|flags 29000041  
|r3 7f7f0000  
|r6 7f7f06fc  
|r9 40006b10  
|r12 1  
r1 51a800  
r4 1  
r7 7f7f0800  
r10 0  
r13 0  
r16 40003fb8  
rp 7f6ce597  
r5 7f7f06f4  
r8 7f7f0800  
r11 40004b78  
r14 0  
|
|
|
|
|
|
|r15 0  
r17 4  
|-------------------------------------------------------------------------|  
*>:32  
:33  
print_average (my_list, first, last);  
print_average (my_list, first, last - 3);  
:
:
:
:
:
:
:34  
:35  
:36  
:37  
}
:.......................................................................:  
File: average.c  
(gdb) la asm  
Procedure: main  
Line: 32  
pc: 0x3524  
15.5 Changing pane focus 249  
(gdb) la split  
(gdb) la regs  
(gdb) la src  
(gdb) la regs  
(gdb) foc next  
Focus set to REGS window.  
(gdb)  
15.6 Scrolling panes  
To scroll within a pane, you can use the arrow keys or the Page Up and Page Down  
keys (on some keyboards these are Prev and Next). You can also use the following  
commands:  
{+ | -} [num_lines]  
[win_name]  
Vertically scroll the pane forward (+) or backward  
(-). + or - with no arguments scrolls the pane  
forward or backward one page. Use num_lines  
to specify how many lines to scroll the pane. Use  
win_name to specify a pane other than the one  
with logical focus.  
{< | >} [num_char]  
[win_name]  
Horizontally scroll the pane left (<) or right (>)  
the specified number of characters. If you do not  
specify num_char, the pane is scrolled one  
character.  
Note that a space is required between the +, -, <, or > and the number.  
To scroll the command pane, use the scroll bars on the terminal pane.  
15.7 Changing the register display  
To look at the floating-point or special registers instead of the general registers, and  
then to return to the general registers, you can use the following XDB commands:  
fr, display $fregs  
sr, display $sregs  
gr, display $gregs  
Display the floating-point registers.  
Display the special registers.  
Display the general registers.  
For example, if you use the frcommand, the window looks like this:  
|-------------------------------------------------------------------------|  
|flags 29000041  
|r3 7f7f0000  
|r6 7f7f06fc  
|r9 40006b10  
|r12 1  
r1 51a800  
r4 1  
r7 7f7f0800  
r10 0  
r13 0  
r16 40003fb8  
rp 7f6ce597  
r5 7f7f06f4  
r8 7f7f0800  
r11 40004b78  
r14 0  
|
|
|
|
|
|
|r15 0  
r17 4  
:......................................................................:  
:30  
:31  
{
:
:
/* Try two test cases. */  
250 The HP-UX Terminal User Interface  
   
*>:32  
:33  
print_average (my_list, first, last);  
print_average (my_list, first, last - 3);  
:
:
:
:
:34  
:35  
}
:......................................................................:  
File: average.c  
(gdb) la regs  
(gdb) la src  
Procedure: main  
Line: 32  
pc: 0x3524  
(gdb) la regs  
(gdb) foc next  
Focus set to REGS window.  
(gdb) fr  
#0 main () at average.c:32  
(gdb)  
The default floating-point register display is single-precision. To change the register  
display to double-precision and then back again, use the XDB toggle float  
command:  
toggle $fregs  
The window looks like this:  
|-------------------------------------------------------------------------|  
|fpsr 0  
|fpe2 0  
|fpe4 0  
|fpe6 0  
|fr4  
fpe1 0  
fpe3 0  
fpe5 0  
fpe7 0  
fr4R  
|
|
|
|
|
|
0
0
|fr5  
1.0000000000000011  
fr5R  
7.00649232e-45  
|-------------------------------------------------------------------------|  
*>:32  
:33  
print_average (my_list, first, last);  
print_average (my_list, first, last - 3);  
:
:
:
:
:
:
:34  
:35  
:36  
:37  
}
:......................................................................:  
File: average.c  
(gdb) la regs  
(gdb) la src  
Procedure: main  
Line: 32  
pc: 0x3524  
(gdb) la regs  
(gdb) foc next  
Focus set to REGS window.  
(gdb) fr  
(gdb) tf  
(gdb)  
15.8 Changing the pane size  
To specify a new height for a pane or to increase or decrease the current height, use  
the winheightcommand (abbreviated winhor wh).  
The syntax is:  
15.8 Changing the pane size 251  
 
winheight [win_name] [+ | -] num_lines  
If you omit win_name, the pane with logical focus is resized. When you increase the  
height of a pane, the height of the Command pane is decreased by the same amount,  
and vice versa. The height of any other panes remains unchanged.  
For example, the command  
wh src +3  
increases the size of the source pane, and decreases the size of the command pane, by  
3 lines.  
To nd out the current sizes of all panes, use the info win command. For example, if you  
have a split-screen layout, the command output might be as follows:  
(gdb) i win  
SRC (8 lines)  
REGS (8 lines)  
CMD (8 lines)  
If you use the mouse or window menus to resize the terminal window during a  
debugging session, the window remains the same size it was when you started. To  
change the window size, you must exit the debugger and restart it.  
15.9 Refreshing and updating the window  
If the screen display is disrupted for some reason, use the refreshcommand (ref) to  
restore the windows to their previous state:  
ref  
If you use stack-navigation commands such as up, down, and frameto change your  
source location, and you want to return the display to the current point of execution,  
use the update command (upd):  
upd  
252 The HP-UX Terminal User Interface  
 
16 XDB to WDB Transition Guide  
This transition aid is designed for XDB users who are learning WDB, an HP-supported  
version of the industry-standard GDB debugger. Select one of these lists for a table that  
shows WDB equivalents for many common XDB commands and other features.  
Invoke WDB with the command gdb -tuito obtain a terminal user interface (TUI)  
similar to that provided by XDB. Commands marked "(with -tui)" are valid when  
you use the -tuioption.  
Invoke WDB with the command gdb -xdbto turn on XDB compatibility mode, which  
enables you to use many XDB commands as synonyms for GDB commands. Commands  
marked "(with -xdb)" are valid when you use the -xdboption.  
You may use both -xdband -tuiat the same time. Some commands are valid only  
when you use both options.  
For a tutorial introduction to WDB, refer to the Getting Started with WDB.  
16.1 By-function lists of XDB commands and HP WDB equivalents  
16.1 By-function lists of XDB commands and HP WDB equivalents 253  
   
16.1.1 Invocation commands  
By default, HP WDB runs in line mode. To run it with a terminal user interface similar  
to that of XDB, use the -tuioption.  
The following table lists the XDB and the equivalent WDB commands for invoking the  
terminal user interface:  
Table 16-1 Invocation commands  
XDB Command  
xdb program  
WDB Equivalent  
Meaning  
gdb -xdb program, gdb -xdb Debug program  
-tui program  
xdb program corefile  
xdb -d dir  
gdb -xdb program -c  
corefile  
Debug core file  
gdb -xdb -d dir  
Specify alternate directory to  
search for source files  
xdb -P pid program  
gdb -xdb program pid  
Attach to running pro- gram at  
invocation  
xdb -i  
xdb -o  
(after starting) run < file  
Specify input to target program  
(after starting) run > file  
Specify output from target  
program  
16.1.2 Window mode commands  
The following commands are TUI mode or XDB compatibility mode commands. They  
are available when you invoke WDB by using the -tuior -xdbor both options.  
Table 16-2 Window mode commands  
XDB Command  
{+ | -}r  
WDB Equivalent  
Meaning  
{+ | -}r (with -xdb  
Scroll floating-point registers  
-tui), {+ | -} data (with forward or backward (src, cmd,  
-tui), gdb -xdb -tui  
program  
and asmare also valid window  
names)  
fr  
fr (with -xdb -tui),  
display $fregs (with  
-tui)  
Display floating-point registers  
254 XDB to WDB Transition Guide  
       
Table 16-2 Window mode commands (continued)  
XDB Command  
WDB Equivalent  
Meaning  
gr  
sr  
gr (with -xdb -tui),  
display $regs (with -tui)  
Display general registers  
sr (with -xdb -tui),  
display $sregs (with  
-tui)  
Display special registers  
td  
tf  
td (with -xdb -tui)  
Toggle disassembly mode  
tf (with -xdb -tui),  
toggle $fregs (with -tui)  
Toggle float register display  
precision  
ts  
u
ts (with -xdb -tui)  
Toggle split-screen mode  
u (with -xdb -tui),  
update (with -tui)  
Update screen to current execution  
point  
V
U (with -xdb -tui),  
refresh (with -tui)  
Refresh all windows  
w number  
w number (with -xdb  
-tui), winheight src  
number (with -tui)  
Set size of source window  
16.1.3 File viewing commands  
The following table lists the XDB and the equivalent WDB commands for viewing  
source files:  
Table 16-3 File viewing commands  
XDB Command  
WDB Equivalent  
Meaning  
{+ | -}[number]  
{+ | -}[ number] (with  
-tui; note that a space  
is required between + or  
- and the number)  
Move view location forward or  
backward in source file number  
lines  
/[string]  
/[string] (with -xdb),  
search regexp, forw  
regexp  
Search source forward for [last]  
string  
?[string]  
D "dir"  
L
?[string] (with -xdb),  
rev regexp  
Search source backward for [last]  
string  
D "dir" (with -xdb), dir Add a directory search path for  
source files  
pathname  
L (with -xdb)  
Show current viewing location or  
current point of execution  
16.1 By-function lists of XDB commands and HP WDB equivalents 255  
   
Table 16-3 File viewing commands (continued)  
XDB Command  
WDB Equivalent  
Meaning  
ld  
lf  
ld (with -xdb), show  
directories  
List source directory search path  
(list all directories)  
lf (with -xdb), info  
sources  
List all source files  
lf [string]  
No equivalent  
fo or rev  
fo or rev  
List matching files  
n
N
Repeat previous search  
Repeat previous search in opposite  
direction  
v
v (with -xdb), list  
Show one source window forward  
from current  
v location  
va address  
va label  
v location (with -xdb), View source at locationin  
list location  
source window  
va address (with -xdb), View addressin disassembly  
disas address  
window  
va label (with -xdb),  
disas label  
View labelin disassembly  
window (labelis a location)  
va label + offset  
va label + offset (with View label+ offset in  
-xdb), disas label +  
disassembly window  
offset  
16.1.4 Source directory mapping commands  
Use the D or dir command to add new directories to be searched for source files. See  
GDB does not provide a source directory mapping capability and therefore does not  
have any equivalent of the apm, dpm, and lpmcommands.  
16.1.5 Data Viewing and modification commands  
There are many info commands in addition to those shown here. Use help info to get  
a list.  
The following table lists the XDB and equivalent WDB commands for viewing and  
modifying the program data:  
256 XDB to WDB Transition Guide  
   
Table 16-4 Data viewing and modification commands  
XDB Command  
WDB Equivalent  
Meaning  
l
l (with -xdb), info args Move view location forward or  
backward in source file number  
lines  
followed by info locals  
lc [string]  
lg [string]  
ll [string]  
lc [string] (with -xdb), Search source forward for [last]  
string  
info common string  
lg [string] (with -xdb), Search source backward for [last]  
string  
info variables [string]  
info functions [string], Add a directory search path for  
source files  
info variables [string],  
maint print msymbols file  
lm  
show user  
Show current viewing location or  
current point of execution  
lm string  
show user string  
List source directory search path  
(list all directories)  
lo [[class]::][string]  
lp  
info func  
[[class]::][string]  
List all source files  
info functions  
Show current scope, list program  
blocks, list names (symbols)  
lp [[class]::]string  
info func  
[[class]::]string info  
addr [[class]::]string  
List all (or matching) procedures  
lr  
lr (with -xdb), info  
all-reg  
List all registers  
lr [string]  
ls [string]  
mm  
lr string (with -xdb),  
info reg string  
List matching registers  
No equivalent  
List all (or matching) special  
variables  
info sharedlibrary  
No equivalent  
Show memory map of all loaded  
shared libraries  
mm string  
p expr[\format  
Show memory map of matching  
loaded shared libraries  
p[/format expr[Note: The  
count and size portions of  
formats are not allowed in the p  
(print)command. They are  
allowed in the xcommand  
(examine memory).]  
Print value using the specified  
format  
16.1 By-function lists of XDB commands and HP WDB equivalents 257  
 
Table 16-4 Data viewing and modification commands (continued)  
XDB Command  
p expr?format  
WDB Equivalent  
Meaning  
p/format &amp;expr  
Print address using specified for-  
mat  
p class::  
No equivalent  
show language  
Print static members of class  
Inquire what language is used  
p $lang  
p {+ | -}[\format  
Use x/format command to  
obtain initial value, then use x  
with no argument to obtain value  
of next memory location. To  
obtain value of previous memory  
location, use "x $_ - 1".  
Print value of next/previous  
memory location using format  
pq expr  
set expr, set var expr  
Evaluate using the specified  
format  
pq expr?format  
No equivalent  
Determine address using specified  
format  
pq class::  
No equivalent  
No equivalent  
Evaluate static members of class  
pq {+ | -}[\format  
Evaluate next/previous memory  
location using format  
16.1.6 Stack viewing commands  
The GDB concept of the top and bottom of the stack is the opposite of XDB, so the XDB  
up is GDBdown.  
The following table lists the XDBand equivalent WDBcommands for viewing the stack  
contents:  
Table 16-5 Stack viewing commands  
XDB Command  
WDB Equivalent  
Meaning  
down  
up  
View procedure one level nearer  
outermost frame of stack (higher  
number)  
down number  
t [depth]  
T [depth]  
top  
up number  
View procedure number levels  
nearer outermost frame of stack  
t [depth] (with -xdb), bt Print stack trace to depth  
[depth]  
T [depth] (with -xdb),  
bt full [depth]  
Print stack trace and show local  
vars  
frame 0  
View procedure at innermost  
frame of stack  
258 XDB to WDB Transition Guide  
   
Table 16-5 Stack viewing commands (continued)  
XDB Command  
WDB Equivalent  
Meaning  
up  
down  
View procedure one level nearer  
innermost frame of stack (lower  
number)  
up number  
down number  
View procedure number levels  
nearer innermost frame of stack  
V [depth]  
V [depth] (with -xdb),  
frame [depth]  
Display text for current active  
procedure or at specified depth on  
stack  
16.1.7 Status-viewing command  
Type the showcommand with no arguments to get a list of current debugger settings.  
Table 16-6 Status viewing commands  
XDB Command  
WDB Equivalent  
Meaning  
I
info(many kinds), show(many Display state of debugger and  
kinds)  
program  
16.1.8 Job control commands  
The following table lists the XDB and equivalent WDB commands for controlling  
program execution:  
Table 16-7 Job control commands  
XDB Command  
WDB Equivalent  
c, continue  
Meaning  
c
Continue from breakpoint,  
ignoring any pending signal  
c location  
until location  
Continue from breakpoint,  
ignoring any pending signal, set  
temporary breakpoint at  
location  
C
c, continue  
Continue, allowing any pending  
signal  
C [location]  
until location  
Continue, allowing any pending  
signal, set temporary breakpoint  
at location  
g line  
g line (with -xdb), go  
line, tb line followed by  
jump line  
Go to linein current procedure  
g #label  
No equivalent  
Go to labelin current procedure  
16.1 By-function lists of XDB commands and HP WDB equivalents 259  
       
Table 16-7 Job control commands (continued)  
XDB Command  
g {+ | -}lines  
WDB Equivalent  
Meaning  
g {+ | -}lines (with  
-xdb), go {+ | -}lines, tb  
{+ | -}linesfollowed by  
jump {+ | -}lines  
Go forward or back given # lines  
g {+ | -}  
g {+ | -} (with -xdb), go Go forward or back 1 line  
{+ | -}1, tb {+ | -}1  
followed by jump {+ | -}1  
k
k
Detach and terminate target  
r [arguments]  
r [arguments]  
Run with last arguments [or with  
new arguments]  
R
s
R (with -xdb), r  
Rerun with no arguments  
s, si  
Single step (into procedures) (si:  
step by instruction)  
s number  
s number, si number  
Single step number steps (into  
procedures) (si: step by  
instruction)  
S
S (with -xdb), n, ni  
Step over (ni: step over by  
instruction)  
S number  
S number (with -xdb), n  
number, ninumber  
Step over by numberstatements  
or instructions (ni: step over by  
instruction)  
16.2 Overall breakpoint commands  
The following table lists the XDB and equivalent WDB commands for setting additional  
breakpoints:  
Table 16-8 Overall breakpoint commands  
XDB Command  
WDB Equivalent  
lb (with -xdb), ib  
No equivalent  
Meaning  
List breakpoints  
Toggle overall breakpoint state  
lb  
tb  
16.2.1 Auxiliary breakpoint commands  
The following table lists the XDB and equivalent WDB auxiliary breakpoint related  
commands:  
260 XDB to WDB Transition Guide  
     
Table 16-9 Auxillary breakpoint commands  
XDB Command  
any_string  
WDB Equivalent  
p "any string"  
Meaning  
Print any string  
if expr {cmds} [{cmds}] if expr cmds [else cmds] Conditionally execute cmds  
end  
Q
Q (with -xdb), silent(must Quiet breakpoints  
be first command in a commands  
list)  
16.2.2 Breakpoint creation commands  
The GDB equivalent of the countand cmdsarguments is to use the commandsbnum  
command to set an ignore count and/or to specify commands to be executed for that  
breakpoint.  
For C++ programs, you can use the regular-expression breakpoint command rbreak  
to set breakpoints on all the member functions of a class or on overloaded functions  
outside a class.  
The following table lists the XDB and equivalent WDB commands for creating  
breakpoints:  
Table 16-10 Breakpoint creation commands  
XDB Command  
WDB Equivalent  
Meaning  
b loc  
b loc  
Set a breakpoint at the specified  
location.  
b
b
Set a breakpoint at the current line.  
ba address  
bb [depth]  
bi expr.proc  
bi -c expr  
ba address (with -xdb), b Set breakpoint at a code address.  
*address  
No equivalent (use bproc)  
Set breakpoint at procedure  
beginning.  
b class::proc cond bnum Set an instance breakpoint at the  
first executable line of expr.proc.  
(this == expr)  
No Equivalent  
Set an instance breakpoint at first  
executable line. No equivalent of  
all non-static member functions of  
the instance of a class (no base  
classes).  
bi -C expr  
No equivalent  
Set an instance breakpoint at first  
executable line of all non- static  
member functions of the instance's  
class (base classes included).  
16.2 Overall breakpoint commands 261  
     
Table 16-10 Breakpoint creation commands (continued)  
XDB Command  
bpc -c class  
WDB Equivalent  
rb ^class::*  
Meaning  
Set a class breakpoint at first  
executable line of all member  
functions of the instance's class (no  
base classes).  
bpc -C class  
Use rb ^class::*for base  
classes also  
Set a class breakpoint at first ex-  
ecutableUse rb ^class::* for base  
classes also line of all member  
functions of the class (base classes  
included).  
bpo proc  
rbproc  
Set breakpoints on overloaded  
functions outside a class.  
bpo class::proc  
bt [depth]  
b class::proc  
No equivalent  
Set breakpoints on overloaded  
functions in a class.  
Set trace breakpoint at procedure  
at specified depth on program  
stack.  
bt proc  
b proccommands bnum  
finish c end  
Set trace breakpoint at proc.  
bu [depth]  
bu [depth] (with -xdb).  
The finishcommand is  
equivalent to the sequence bu,  
c, db(to continue out of the  
current routine).  
Set up-level breakpoint.  
bx [depth]  
bx [depth] (with -xdb)  
Set a breakpoint at procedure exit.  
16.2.3 Breakpoint status commands  
The following table lists the XDB and equivalent WDB commands for changing the  
breakpoint status:  
Table 16-11 Overall breakpoint commands  
XDB Command  
ab number  
WDB Equivalent  
enable number  
Meaning  
Activate suspended breakpoint of  
the given number  
ab *  
enable  
Activate all suspended break-  
points  
ab @shared- library  
No equivalent  
Activate suspended breakpoints  
in named shared library  
262 XDB to WDB Transition Guide  
   
Table 16-11 Overall breakpoint commands (continued)  
XDB Command  
bc number expr  
WDB Equivalent  
Meaning  
bc number expr(with -xdb), Set a breakpoint count  
ignorenumber expr(within a  
commands list)  
db  
clear  
Delete breakpoint at current line  
db number  
delete number  
Delete breakpoint of the given  
number  
db *  
delete  
Delete all breakpoints  
sb number  
disable number  
Suspend breakpoint of the given  
number  
sb *  
disable  
Suspend all breakpoints  
sb @shared- library  
No equivalent  
Suspend breakpoints in named  
shared library  
16.2.4 All-procedures breakpoint commands  
GDB does not provide the ability to set breakpoints on all procedures with a single  
command. Therefore, it does not have any equivalent of the following commands:  
bp  
bpt  
bpx  
dp  
Dpt  
Dpx  
16.2.5 Global breakpoint commands  
The following table lists the XDB and equivalent WDB commands for setting global  
breakpoints:  
Table 16-12 Global breakpoint commands  
XDB Command  
abc cmds  
WDB Equivalent  
Meaning  
No exact equivalent, but  
Set or delete cmdsto execute at  
display expris equivalent to every stop  
abc print expr  
dbc  
undisplay  
Stop displaying values at each stop  
16.2 Overall breakpoint commands 263  
     
16.2.6 Assertion control commands  
GDB does not provide the ability to trace by instruction. Watchpoints, however, provide  
similar functionality to xdbassertions.  
For example, watchpoints can be:  
Enabled (corresponds to aa)  
Disabled (corresponds to da)  
Listed (corresponds to info watch)  
Added (corresponds to x)  
WDB does not have explicit equivalents for the following commands:  
a
aa  
da  
la  
sa  
ta  
x
16.2.7 Record and playback commands  
Use the source command to read commands from a file. GDB does not provide a  
recording capability like XDB's, but you can use the set history savecommand  
to record all GDB commands in the file ./.gdb_history(similar to the $HOME/  
.xdbhistfile). The history file is not saved until the end of your debugging session.  
To change the name of the history file, use set history filename.  
To stop recording, use set history save off.  
To display the current history status, use show history. For an equivalent of the  
XDB record-all facility, pipe the output of the gdbcommand to the tee(1)command.  
For example:  
gdb a.out | tee mylogfile  
This solution works with the default line-mode user interface, not with the terminal  
user interface.  
The following table lists the XDB and the equivalent WDB commands for handling  
macros:  
16.2.8 Macro facility commands  
Use the show user or help user-defined command to obtain a list of all user-defined  
commands.  
264 XDB to WDB Transition Guide  
     
Table 16-13 Macro facility commands  
XDB Command  
WDB Equivalent  
Meaning  
def name replacement-text def name[GDB prompts for  
Define a user-defined command  
commands]  
tm  
No equivalent  
Toggle the macro substitution  
mechanism  
undef name  
undef *  
def name[follow with empty  
Remove the macro definition for  
name  
command list]  
No equivalent  
Remove all macro definitions  
16.2.9 Signal control commands  
The following table lists the XDB and equivalent WDB commands for signal control:  
Table 16-14 Signal control commands  
XDB Command  
WDB Equivalent  
Meaning  
lz  
lz(with -xdb), info signals List signal handling  
z number s  
z number i  
z number s(with -xdb),  
handle numberstop, handle  
number nostop  
Toggle stop flag for signal number  
z number i(with -xdb),  
handle numbernopass,  
handle number pass  
Toggle ignore flag for signal  
number  
z number r  
z number r(with -xdb),  
handle number print,  
handle number noprint  
Toggle report flag for signal  
number  
z number Q  
z number Q(with -xdb),  
handle number noprint  
Do not print the new state of the  
signal  
16.2.10 Miscellaneous commands  
Some of the additional XDB and the equivalent WDB commands are discussed below:  
Table 16-15 Miscellaneous commands  
XDB Command  
WDB Equivalent  
Meaning  
Return  
Return  
Return  
Repeat previous command  
Repeat previous command  
~
;
No equivalent (one command per Separate commands in command  
line in command list)  
list  
16.2 Overall breakpoint commands 265  
         
Table 16-15 Miscellaneous commands (continued)  
XDB Command  
! cmd line  
WDB Equivalent  
Meaning  
! cmd line(with -xdb), she Invoke a shell  
cmd line  
{cmd list}  
commands [number]... end  
Execute command list (group  
commands)  
Control-C  
# [text]  
am  
Control-C  
Interrupt the program  
A comment  
# [text]  
am(with -xdb), set height  
num  
Activate more (turn on pagination)  
f ["printf-style- fmt"] No equivalent  
Set address printing format  
Help  
h
h
M[{t | c} [expr[;  
expr...]]]  
No equivalent  
Print object or corefile map  
q
q
Quit debugger  
sm  
sm(with -xdb), set height 0 Suspend more (turn o pagination)  
ss file  
No equivalent  
Save (breakpoint, macro, assertion)  
state  
tc  
No equivalent  
Toggle case sensitivity in searches  
16.3 XDB data formats and HP WDB equivalents  
The format of the printcommand is different in XDB and GDB:  
XDB: p expr\fmt  
GDB: p/fmt expr  
Use the set print prettycommand to obtain a structured display similar to the  
default XDB display.  
The following table lists the XDB and equivalent WDB commands for setting data  
display formats:  
Table 16-16 Data format commands  
XDB Command  
WDB Equivalent  
Meaning  
Byte in decimal  
b
d
d
c
B (1)  
c
Byte in decimal  
Character  
C (1)  
c
Wide character  
266 XDB to WDB Transition Guide  
   
Table 16-16 Data format commands (continued)  
XDB Command  
WDB Equivalent  
Meaning  
Decimal integer  
d
d
d
D (1)  
Long decimal integer  
e
No equivalent  
No equivalent  
No equivalent  
No equivalent  
f
e Floating-point notation as float  
e Foating-point notation as double  
f floating-point notation as float  
f floating-point notation as double  
g floating-point notation as float  
g floating-point notation as double  
Machine instruction (disassembly)  
Formatted structure display  
E (1)  
f
F (1)  
g
G (1)  
f
i
Use x/i command  
No equivalent  
No equivalent  
k
K (1)  
Formatted structure display with  
base classes  
n
print  
Normal (default) format, based on  
type  
o
o
o
a
Expression in octal as integer  
O (1)  
P
Expression in octal as long integer  
Print name of procedure  
containing address  
s
No equivalent  
No equivalent  
whatis, ptype  
ptype  
String  
S
Formatted structure display  
Show type of the expression  
t
T (1)  
Show type of expression, including  
base class information  
u
u
u
Expression in unsigned decimal  
format  
U (1)  
Expression in long unsigned  
decimal format  
w
No equivalent  
Wide character string  
W (1)  
x
No equivalent  
Address of wide character string  
Print in hexadecimal  
x
x
X (1)  
Print in long hexadecimal  
16.3 XDB data formats and HP WDB equivalents 267  
Table 16-16 Data format commands (continued)  
XDB Command  
WDB Equivalent  
Meaning  
Print in binary  
Print in long binary  
z
t
t
Z (1)  
(1) HP WDB will display data in the size appropriate for the data. It will not extend  
the length displayed in response to one of the uppercase formchars (for example, O,  
D, F).  
16.4 XDB location syntax and HP WDB equivalents  
The following command lists the XDB and the equivalent WDB commands for locating  
source lines:  
Table 16-17 Macro facility commands  
XDB location syntax  
WDB Equivalent  
Meaning  
line  
line  
Source line and code address  
Source line and code address  
Procedure name  
file[:line]  
proc  
file[:line]  
proc  
[file:]proc[:proc[...]][:line] No equivalent  
Source line and code address  
Source line and code address  
Source line and code address  
Source line and code address  
Source line and code address  
Code address  
[file:]proc[:proc[...]][:#label] No equivalent  
[class]::proc  
[class]::proc  
[class]::proc[:line]  
[class]::proc[#label]  
proc#line  
No equivalent  
No equivalent  
No equivalent  
No equivalent  
No equivalent  
No equivalent  
[class]::proc#line  
#label  
Code address  
Source line and code address  
name@shared-library  
Address of name in shared library  
shared-library  
16.5 XDB special language operators and HP WDB equivalents  
The following table lists the XDB and the equivalent WDB commands for language  
operators:  
268 XDB to WDB Transition Guide  
     
Table 16-18 Special language operators  
XDB Language Operator  
$addr  
WDB Equivalent  
Meaning  
Depends on language  
No equivalent  
Unary operator, address of object  
$in  
Unary Boolean operator, execution  
in procedure  
$sizeof  
sizeof  
Unary operator, size of object  
16.6 XDB special variables and HP WDB equivalents  
GDB does not provide special variables of the kind that XDB has, but you can use show  
and set to display and modify many debugger settings.  
Table 16-19 Special variables  
XDB Special Variable  
$cplusplus  
WDB Equivalent  
No equivalent  
Meaning  
C++ feature control flags  
$depth  
No equivalent  
No equivalent  
No equivalent  
show language  
Default stack depth for local  
variables  
$fpa  
Treat FPA sequence as one  
instruction  
$fpa_reg  
$lang  
Address register for FPA  
sequences  
Current language for expression  
evaluation  
$line  
No equivalent  
No equivalent  
Current source line number  
$malloc  
Debugger memory allocation  
(bytes)  
$print  
No equivalent  
Display mode for character data  
Hardware registers  
$regname  
$result  
$regname  
Use $n(value history number  
assigned to the desired result)  
Return value of last command line  
procedure call  
$signal  
$step  
No equivalent  
No equivalent  
Current child procedure signal  
number  
Number of instructions debugger  
will step in non-debuggable  
procedures before free-running  
$var  
$var  
Dene or use special variable  
(convenience variable)  
16.6 XDB special variables and HP WDB equivalents 269  
     
16.7 XDB variable identifiers and HP WDB equivalents  
Table 16-20 Variable Identifiers  
XDB Variable Identifier  
WDB Equivalent  
Meaning  
Search for var  
var  
class::  
var  
class::  
Search class for var(bug: not yet)  
[[class]::]proc:[class::]var proc::var  
Search proc for var(static  
variables only)  
[[class]::]proc:depth:[class::] No equivalent  
Search proc for depth on stack  
. (dot)  
Empty string; for example, p is  
the equivalent of p .  
Shorthand for last thing you  
looked at  
:var or ::var  
::varto distinguish a global from Search for global variable only  
a local variable with same name  
16.8 Alphabetical lists of XDB commands and HP WDB equivalents  
16.8.1 A  
Table 16-21 A  
XDB Command  
Equivalent WDB Command  
No equivalent  
a [cmds]  
aa number  
aa *  
No equivalent  
No equivalent  
enable number  
enable  
ab number  
ab *  
ab @shared-library  
No equivalent  
270 XDB to WDB Transition Guide  
         
Table 16-21 A (continued)  
XDB Command  
Equivalent WDB Command  
abc cmds  
No exact equivalent, but display expris equivalent  
to abc print expr  
am  
am (with -xdb), set height num  
No equivalent  
apm oldpath [newpath]  
apm "" [newpath]  
No equivalent  
16.8.2 B  
Table 16-22 B  
XDB Command  
Equivalent WDB Command  
b loc  
b loc  
b
b
ba address  
bb [depth]  
ba address(with -xdb), b *address  
No equivalent (use b proc)  
bc number expr  
bc number expr(with -xdb), ignore number expr  
(within a commands list)  
bi expr.proc  
bi -c expr  
bi -C expr  
bp  
b class::proccond bnum (this == expr)  
No equivalent  
No equivalent  
No equivalent  
bp cmds  
No equivalent  
bpc -c class  
bpc -C class  
bpo proc  
bpo class::proc  
bpt  
rb ^class::*  
Use rb ^class::* for base classesalso  
rb proc  
b class::proc  
No equivalent  
bpt cmds  
bpx  
No equivalent  
No equivalent  
bpx cmds  
bt proc  
No equivalent  
b proc commands bnum finish c end  
16.8 Alphabetical lists of XDB commands and HP WDB equivalents 271  
   
Table 16-22 B (continued)  
XDB Command  
Equivalent WDB Command  
bu [depth]  
bu [depth] (with -xdb). The finish command is  
equivalent to the sequence bu, c, db (to continue  
out of the current routine).  
bx [depth]  
bx [depth] (with -xdb)  
16.8.3 C through D  
Table 16-23 C through D  
XDB Command  
Equivalent WDB Command  
c, continue  
c
c location  
until location  
C
c, continue  
C location  
until location  
D "dir"  
D "dir" (with -xdb), dir pathname  
No equivalent  
No equivalent  
clear  
da number  
da *  
db  
db number  
delete number  
delete  
db *  
dbc  
undisplay  
def name replacement-text  
def name [GDB prompts for commands]  
up  
down  
down number  
dp  
up number  
No equivalent  
No equivalent  
No equivalent  
No equivalent  
No equivalent  
dpm index  
dpm *  
Dpt  
Dpx  
272 XDB to WDB Transition Guide  
   
16.8.4 F through K  
Table 16-24 F through K  
XDB Command  
Equivalent WDB Command  
No equivalent  
f ["printf-style-fmt"]  
fr  
fr (with -xdb -tui), display $fregs (with -tui)  
g line  
g line (with -xdb), go line, tb line followed by jump  
line  
g #label  
No equivalent  
g {+ | -}lines  
g {+ | -}lines (with -xdb), go {+ | -}lines tb {+ | -}lines  
followed by jump {+ | -}lines  
g {+ | -}  
g {+ | -} (with -xdb), go {+ | -}1, tb {+ | -}1 followed  
by jump {+ | -}1  
gr  
gr (with -xdb -tui), display $regs (with -tui)  
h
h
if expr {cmds} [{cmds}]  
if expr cmds [else cmds] end  
I
k
info (many kinds), show (many kinds)  
k
16.8.5 L  
Table 16-25 L  
XDB Command  
Equivalent WDB Command  
l (with -xdb), info args followed by info locals  
L (with -xdb)  
l
L
la  
No equivalent  
lb  
lb (with -xdb), i b  
lc [string]  
ld  
lc [string] (with -xdb), info common string  
ld (with -xdb), show directories  
lf (with -xdb), info sources  
No equivalent  
lf  
lf [string]  
lg [string]  
ll [string]  
lg [string] (with -xdb), info variables [string]  
info functions [string], info variables [string], maint  
print msymbolsfile  
16.8 Alphabetical lists of XDB commands and HP WDB equivalents 273  
       
Table 16-25 L (continued)  
XDB Command  
Equivalent WDB Command  
show user [string]  
lm [string]  
lo [[class]::][string]  
info func [[class]::][string]  
info functions  
lp  
lp [[class]::]string  
info func [[class]::]string info addr [[class]::]string  
No equivalent  
lpm  
lr  
lr (with -xdb), info all-reg  
lr string (with -xdb), info reg string  
No equivalent  
lr string  
ls [string]  
lz  
lz (with -xdb), info signals  
16.8.6 M through P  
Table 16-26 M through P  
XDB Command  
Equivalent WDB Command  
No equivalent  
M[{t | c} [expr[; expr...]]]  
mm  
info sharedlibrary  
No equivalent  
mm string  
N
fo or rev  
n
fo or rev  
p expr[\format]  
p[/format] expr [Note: The count and size portions  
of formats are not allowed in the p (print) command.  
They are allowed in the x command (examine  
memory).]  
p expr?format  
p class::  
p/format &amp;expr  
No equivalent  
p $lang  
show language  
p {+ | -}[\format  
Use x/format command to obtain initial value, then  
use x with no argument to obtain value of next  
memory location. To obtain value of previous  
memory location, use "x $_ - 1".  
pq expr  
set expr, set var expr  
No equivalent  
pq expr?format  
274 XDB to WDB Transition Guide  
   
Table 16-26 M through P (continued)  
XDB Command  
pq class::  
Equivalent WDB Command  
No equivalent  
No equivalent  
pq [+ | -][\format  
16.8.7 Q through S  
Table 16-27 Q through S  
XDB Command  
Equivalent WDB Command  
q
q
Q
Q (with -xdb), silent (must be first command in a  
commands list)  
r [arguments]  
r [arguments]  
R
R (with -xdb), r  
s
s, si  
s number  
s number, si number  
S (with -xdb), n, ni  
S number (with -xdb), n number, ninumber  
No equivalent  
S
S number  
sa number  
sa *  
No equivalent  
sb number  
disable number  
sb *  
disable  
sb @shared-library  
No equivalent  
sm  
sm (with -xdb), set height 0  
sr (with -xdb -tui), display $sregs (with -tui)  
No equivalent  
sr  
ss file  
16.8.8 T  
Table 16-28 T  
XDB Command  
Equivalent WDB Command  
t [depth] (with -xdb), bt [depth]  
t [depth]  
T [depth]  
T [depth] (with -xdb), bt full [depth]  
16.8 Alphabetical lists of XDB commands and HP WDB equivalents 275  
       
Table 16-28 T (continued)  
XDB Command  
Equivalent WDB Command  
No equivalent  
ta  
tb  
No equivalent  
tc  
No equivalent  
td  
td (with -xdb -tui)  
tf (with -xdb -tui), toggle $fregs (with -tui)  
No equivalent  
tf  
tm  
top  
tr [@]  
ts  
frame 0  
No equivalent  
ts (with -xdb -tui)  
16.8.9 U through Z  
Table 16-29 U through Z  
XDB Command  
Equivalent WDB Command  
u (with -xdb -tui), update (with -tui)  
U (with -xdb -tui), refresh (with -tui)  
def name [follow with empty command list]  
No equivalent  
u
U
undef name  
undef *  
up  
down  
up number  
v
down number  
v (with -xdb), list  
v location  
V [depth]  
va address  
va label  
va label + offset  
w number  
v location (with -xdb), list location  
V [depth] (with -xdb), frame [depth]  
va address (with -xdb), disas address  
va label (with -xdb), disas label  
va label + offset (with -xdb), disas label + offset  
w number (with -xdb -tui), winheight src number  
(with - tui)  
x [expr]  
No equivalent  
xdb program  
gdb -xdb program, gdb -xdb -tui program  
276 XDB to WDB Transition Guide  
   
Table 16-29 U through Z (continued)  
XDB Command  
xdb program corefile  
xdb -d dir  
Equivalent WDB Command  
gdb -xdb program -c corefile  
gdb -xdb -d dir  
xdb -P pid program  
xdb -i  
gdb -xdb program pid  
(after starting) run < file  
(after starting) run > file  
xdb -o  
z number s  
z number s (with -xdb), handle number stop, handle  
number nostop  
z number i  
z number r  
z number Q  
z number i (with -xdb), handle number nopass,  
handle number pass  
z number r (with -xdb), handle number print,  
handle number noprint  
z number Q (with -xdb), handle number noprint  
16.8.10 Symbols  
Table 16-30 Symbols  
XDB Command  
Equivalent WDB Command  
line  
line  
file[:line]  
file[:line]  
proc  
proc  
[file:]proc[:proc[...]][:line]  
[file:]proc[:proc[...]][:#label]  
[class]::proc  
No equivalent  
No equivalent  
[class]::proc  
No equivalent  
[class]::proc[:line]  
[class]::proc[#label]  
proc#line  
No equivalent  
No equivalent  
name@shared-library  
var  
No equivalent  
var  
class::var  
class::var (bug: not yet)  
proc::var (static variables only)  
No equivalent  
[[class]::]proc:[class::]var  
[[class]::]proc:depth:[class::]var  
16.8 Alphabetical lists of XDB commands and HP WDB equivalents 277  
   
Table 16-30 Symbols (continued)  
XDB Command  
Return  
Equivalent WDB Command  
Return  
"any string"  
. (dot)  
p "any string"  
Empty string; for example, p is the equivalent  
of p .  
-
Return  
{+ | -}r  
{+ | -}r (with -xdb -tui), {+ | -} data  
(with -tui)  
{+ | -}[number]  
{+ | -}[ number](with -tui; note that a space  
is required between + or - and the number)  
/[string]  
/[string] (with -xdb), search regexp,  
forw regexp  
?[string]  
?[string] (with -xdb), rev regexp  
;
No equivalent (one command per line in command  
list)  
:var or ::var  
::var  
! cmd_line  
{cmd_list}  
<file  
<<file  
>
! cmd line (with -xdb), she cmd line  
commands [number]... end  
source file  
No equivalent  
No equivalent  
>file  
>c  
No equivalent  
No equivalent  
>f  
No equivalent  
>t  
No equivalent  
>@[c | f | t]  
>@file  
>>  
No equivalent  
No equivalent  
No equivalent  
>>file  
>>@  
No equivalent  
No equivalent  
>>@file  
No equivalent  
278 XDB to WDB Transition Guide  
Table 16-30 Symbols (continued)  
XDB Command  
Control-C  
Equivalent WDB Command  
Control-C  
# [text]  
# [text]  
#label  
No equivalent  
16.8 Alphabetical lists of XDB commands and HP WDB equivalents 279  
280  
17 Controlling GDB  
You can alter the way GDB interacts with you by using the set command. For commands  
controlling how GDB displays data, see “Print settings” (page 90). Other settings are  
described here.  
17.1 Setting the GDB Prompt  
GDB indicates its readiness to read a command by printing a string called the prompt.  
This string is normally `((gdb))'. You can change the prompt string with the set prompt  
command. For instance, when debugging GDB with GDB, it is useful to change the  
prompt in one of the GDB sessions so that you can always tell which one you are talking  
to.  
NOTE: set prompt does not add a space for you after the prompt you set. This allows  
you to set a prompt which ends in a space or a prompt that does not.  
set prompt newprompt  
Directs GDB to use newpromptas its prompt string  
henceforth.  
show prompt  
Prints a line of the form: “Gdb's prompt is:  
your-prompt”  
17.2 Setting Command Editing Options in GDB  
GDB reads its input commands via the readline interface. This gnu library provides  
consistent behavior for programs which provide a command line interface to the user.  
Ad- vantages are gnu Emacs-style or vi-style inline editing of commands, csh-like  
history substitution, and a storage and recall of command history across debugging  
sessions.  
You may control the behavior of command line editing in GDB with the command set.  
set editing, set editing Enable command line editing (enabled by default).  
on  
set editing off  
show editing  
Disable command line editing.  
Show whether command line editing is enabled.  
17.3 Setting Command History Feature in GDB  
GDB can keep track of the commands you type during your debugging sessions, so  
that you can be certain of precisely what happened. Use these commands to manage  
the GDB command history facility.  
To make command history understand your vi key bindings you need to create a ~/  
.inputrcfile with the following contents:  
17.1 Setting the GDB Prompt 281  
       
set editing-mode vi  
The readline interface uses the .inputrcfile to  
control the settings.  
set history filename fname  
Set the name of the GDB command history file  
to fname. This is the file where GDB reads an  
initial command history list, and where it writes  
the command history from this session when it  
exits. You can access this list through history  
expansion or through the history command  
editing characters listed below. This file defaults  
to the value of the environment variable  
GDBHISTFILE, or to ./.gdb_history(./  
_gdb_historyon MS-DOS) if this variable is  
not set.  
set history save, set history save on Record command history in a file, whose name  
may be specified with the set history  
filenamecommand. By default, this option is  
disabled.  
set history save off  
Stop recording command history in a file.  
set history size size  
Set the number of commands which GDB keeps  
in its history list. This defaults to the value of the  
environment variable HISTSIZE, or to 256 if this  
variable is not set.  
History expansion assigns special meaning to the character !.  
Since !is also the logical not operator in C, history expansion is o by default. If you  
decide to enable history expansion with the set history expansion on command, you  
may sometimes need to follow !(when it is used as logical not, in an expression) with  
a space or a tab to prevent it from being expanded. The readline history facilities do  
not attempt substitution on the strings !=and !(, even when history expansion is  
enabled.  
The commands to control history expansion are:  
set history expansion on, set history Enable history expansion. History expansion is  
expansion  
o by default.  
set history expansion off  
Disable history expansion.  
The readline code comes with more complete  
documentation of editing and history expansion  
features. Users unfamiliar with GNUEmacs or vi  
may wish to read it.  
show history, show history filename, These commands display the state of the GDB  
show history save, show history  
size, show history expansion  
history parameters. show historyby itself  
displays all four states.  
282 Controlling GDB  
show commands  
show commands n  
show commands +  
Display the last ten commands in the command  
history.  
Print ten commands centered on command  
number n.  
Print ten commands just after the commands last  
printed.  
17.4 Setting the GDB Screen Size  
Certain commands to GDB may produce large amounts of information output to the  
screen. To help you read all of it, GDB pauses and asks you for input at the end of each  
page of output. Type RET when you want to continue the output, or qto discard the  
remaining output. Also, the screen width setting determines when to wrap lines of  
output. Depending on what is being printed, GDB tries to break the line at a readable  
place, rather than simply letting it over flow onto the following line.  
Normally GDB knows the size of the screen from the terminal driver software. For  
example, on Unix, GDB uses the termcap data base together with the value of the TERM  
environment variable and the sttyrows and sttycols settings. If this is not correct,  
you can override it with the set heightand set widthcommands:  
set height lpp, show These setcommands specify a screen height of lpplines and  
height se, set width a screen width of cplcharacters. The associated show  
cpl, show width  
commands display the current settings.  
If you specify a height of zero lines, GDB does not pause during  
output no matter how long the output is. This is useful if output  
is to a file or to an editor buffer.  
Likewise, you can specify set width 0to prevent GDB from  
wrapping its output.  
17.5 Supported Number Formats  
You can always enter numbers in octal, decimal, or hexadecimal in GDB by the usual  
conventions: octal numbers begin with `0', decimal numbers end with `.', and  
hexadecimal numbers begin with `0x'. Numbers that begin with none of these are, by  
default, entered in base 10; likewise, the default display for numbers|when no particular  
format is specified| is base 10. You can change the default base for both input and  
output with the set radix command.  
set input-radix base  
Set the default base for numeric input. Supported  
choices for base are decimal 8, 10, or 16. basemust  
itself be specified either unambiguously or using the  
current default radix; for example, any of  
set radix 012  
set radix 10  
17.4 Setting the GDB Screen Size 283  
   
set radix 0xa  
sets the base to decimal. On the other hand, set  
radix 10leaves the radix unchanged no matter  
what it was.  
set output-radix base  
Set the default base for numeric display. Supported  
choices for baseare decimal 8, 10, or 16. basemust  
itself be specified either unambiguously or using the  
current default radix.  
show input-radix  
show output-radix  
Display the current default base for numeric input.  
Display the current default base for numeric display.  
17.6 Optional warnings and messages  
By default, GDB is silent about its inner workings. If you are running on a slow machine,  
you may want to use the set verbosecommand. This makes GDB tell you when it  
does a lengthy internal operation, so you will not think it has crashed.  
Currently, the messages controlled by set verbose are those which announce that the  
symbol table for a source file is being read; see symbol-file in “Commands to specify  
set verbose on  
set verbose off  
show verbose  
Enables GDB output of certain informational messages.  
Disables GDB output of certain informational messages.  
Displays whether set verbose is on or o.  
By default, if GDB encounters bugs in the symbol table of an object file, it is silent; but  
if you are debugging a compiler, you may nd this information useful (see “Specifying  
set complaints limit  
Permits GDB to output limitcomplaints about each  
type of unusual symbols before becoming silent about  
the problem. Set limitto zero to suppress all  
complaints; set it to a large number to prevent  
complaints from being suppressed.  
show complaints  
Displays how many symbol complaints GDB is  
permitted to produce.  
By default, GDB is cautious, and asks the user to confirm on certain commands. For  
example, if you try to run a program which is already running:  
((gdb)) run  
The program being debugged has been started already.  
Start it from the beginning? (y or n)  
If you are willing to unflinchingly face the consequences of your own commands, you  
can disable this feature:  
set confirm off  
Disables confirmation requests.  
284 Controlling GDB  
 
set confirm on  
show confirm  
Enables confirmation requests (the default).  
Displays state of confirmation requests.  
17.7 Optional messages about internal happenings  
set debug arch  
Turns on or off display of gdbarchdebugging info.  
The default is off  
show debug arch  
set debug event  
Displays the current state of displaying gdbarch  
debugging information.  
Turns on or o display of GDB event debugging  
information. The default is o.  
show debug event  
set debug expression  
show debug expression  
set debug overload  
Displays the current state of displaying GDB event  
debugging info.  
Turns on or off display of GDB expression debugging  
information. The default is o.  
Displays the current state of displaying GDB  
expression debugging info.  
Turns on or o display of GDB C++ overload  
debugging info. This includes info such as ranking  
of functions, etc. The default is o.  
show debug overload  
set debug remote  
Displays the current state of displaying GDB C++  
overload debugging info.  
Turns on or o display of reports on all packets sent  
back and forth across the serial line to the remote  
machine. The info is printed on the GDB standard  
output stream. The default is o.  
show debug remote  
set debug serial  
Displays the state of display of remote packets.  
Turns on or o display of GDB serial debugging info.  
The default is o.  
show debug serial  
set debug target  
Displays the current state of displaying GDB serial  
debugging info.  
Turns on or o display of GDB target debugging info.  
This info includes what is going on at the target level  
of GDB, as it happens. The default is o.  
show debug target  
set debug varobj  
Displays the current state of displaying GDB target  
debugging info.  
Turns on or o display of GDB variable object  
debugging info. The default is o.  
17.7 Optional messages about internal happenings 285  
 
show debug varobj  
Displays the current state of displaying GDB variable  
object debugging info.  
286 Controlling GDB  
18 Canned Sequences of Commands  
In addition to breakpoint commands (see “Breakpoint command lists” (page 61)), GDB  
provides the following two ways to store sequence of commands for execution as a  
unit:  
user-defined commands  
command files  
18.1 User-defined commands  
A user-defined command is a sequence of GDB commands to which you assign a new  
name as a command. This is done with the definecommand. User commands may  
accept up to 10 arguments separated by whitespace. Arguments are accessed within  
the user command via $arg0. . . $arg9. The following example illustrates the use  
of canned sequence of commands:  
define adder  
print $arg0 + $arg1 + $arg2  
To execute the command use:  
adder 1 2 3  
This defines the command adder, which prints the sum of its three arguments. Note  
the arguments are text substitutions, so they may reference variables, use complex  
expressions, or even perform further functions calls.  
The following constructs can be used to create canned sequence of commands:  
define commandname  
Define a command named commandname. If there is  
already a command by that name, you are asked to  
confirm that you want to redefine it. The definition  
of the command is made up of other GDB command  
lines, which are given following the define  
command. The end of these commands is marked by  
a line containing end.  
if  
Takes a single argument, which is an expression to  
evaluate. It is followed by a series of commands that  
are executed only if the expression is true (nonzero).  
The ifclause can be followed by an optional else  
clause. You can add a list of commands to the else  
clause which get executed only if the expression is  
false.  
while  
The syntax is similar to if:the command takes a  
single argument, which is an expression to evaluate,  
and must be followed by the commands to execute,  
one per line, terminated by an end. The commands  
18.1 User-defined commands 287  
   
are executed repeatedly as long as the expression  
evaluates to true.  
document commandname  
Document the user-defined command commandname,  
so that it can be accessed by help. The command  
commandnamemust already be defined. This  
command reads lines of documentation just as define  
reads the lines of the command definition, ending  
with end. After the document command is finished,  
help on command commandnamedisplays the  
documentation you have written.  
You may use the document command again to change  
the documentation of a command. Redefining the  
command with define does not change the  
documentation.  
help user-defined  
List all user-defined commands, with the first line of  
the documentation (if any) for each.  
show user, show user  
commandname  
Display the GDB commands used to define  
commandname(but not its documentation). If no  
commandnameis given, display the definitions for  
all user- defined commands.  
When user-defined commands are executed, the commands of the definition are not  
printed. An error in any command stops execution of the user-defined command.  
If used interactively, commands that would ask for confirmation proceed without  
asking when used inside a user-defined command. Many GDB commands that normally  
print messages to say what they are doing omit the messages when used in a  
user-defined command.  
18.2 User-defined command hooks  
You may define hooks, which are a special kind of user-defined command. Whenever  
you run the command foo, if the user-defined command hook-fooexists, it is executed  
(with no arguments) before that command.  
In addition, a pseudo-command, stopexists. Defining (hook-stop) makes the  
associated commands execute every time execution stops in your program: before  
breakpoint commands are run, displays are printed, or the stack frame is printed.  
For example, to ignore SIGALRMsignals while single-stepping, and treat them normally  
during normal execution, you could define:  
define hook-stop  
handle SIGALRM nopass  
end  
define hook-run  
handle SIGALRM pass  
288 Canned Sequences of Commands  
 
end  
define hook-continue  
handle SIGLARM pass  
end  
You can define a hook for any single-word command in GDB, and not for command  
aliases; Also you should define a hook for the basic command name, for example,  
backtracerather than bt. If an error occurs during the execution of your hook,  
execution of GDB commands stops and GDB issues a prompt (before the command  
that you actually typed had a chance to run).  
If you try to define a hook which does not match any known command, GDB issues a  
warning from the definecommand.  
18.3 Command files  
A command file for GDB is a file of lines that are GDB commands. Comments (lines  
starting with #) may also be included. An empty line in a command file does nothing;  
it does not mean to repeat the last command, as it would from the terminal.  
When you start GDB, it executes commands from its initfiles. These are files named  
.gdbiniton UNIX. During startup, GDB does the following:  
4
1. Read the init file (if any) in your home directory.  
2. Process command-line options and operands.  
3. Read the init file (if any) in the current working directory,  
4. Read the init file (if any) in the /tmpdirectory if the GDB_PROCESS_TMP_GDBINIT  
environment variable is set.  
5. Read command files specified by the `-x' option.  
The init file in your home directory can set options (such as `set complaints') that affect  
subsequent processing of command line options and operands. Init files are not executed  
if you use the `-nx' option (see “Choosing modes” (page 27)).  
It can be useful to create a `.gdbinit' file in the directory where you are debugging an  
application. This file will set the actions that apply for this application.  
For example, one might add lines like:  
dir /usr/src/path/to/source/files  
to add source directories or:  
break fatal  
to set breakpoints on fatal error routines or diagnostic routines.  
On some configurations of GDB, the initfile is known by a different name (these are  
typically environments where a specialized form of GDB may need to coexist with  
other forms, hence a different name for the specialized version's initfile). These are  
the environments with special initfile names:  
18.3 Command files 289  
 
VxWorks (Wind River Systems real-time OS): .vxgdbinit  
OS68K (Enea Data Systems real-time OS): .os68gdbinit  
ES-1800 (Ericsson Telecom AB M68000 emulator): .esgdbinit  
You can also request the execution of a command file with the sourcecommand:  
source filename Execute the command file filename.  
The lines in a command file are executed sequentially. They are not printed as they are  
executed. An error in any command terminates execution of the command file.  
Commands that would ask for confirmation if used interactively proceed without  
asking when used in a command file. Many GDB commands that normally print  
messages to say what they are doing omit the messages when called from command  
4
files.  
18.4 Commands for controlled output  
During the execution of a command file or a user-defined command, normal GDB  
output is suppressed; the only output that appears is what is explicitly printed by the  
commands in the definition. This section describes three commands useful for generating  
exactly the output you want.  
echo text  
Print text. Nonprinting characters can be  
included in textusing C escape sequences, such  
as \n to print a newline. No newline is printed  
unless you specify one. In addition to the  
standard C escape sequences, a backslash  
followed by a space stands for a space. This is  
useful for displaying a string with spaces at the  
beginning or the end, since leading and trailing  
spaces are otherwise trimmed from all  
arguments. To print and foo =, use the  
command echo \and foo = \.  
A backslash at the end of text can be used, as in  
C, to continue the command onto subsequent  
lines. For example,  
echo This is some text\n\  
which is continued\n\  
onto several lines.\n  
produces the same output as  
echo This is some text\n  
echo which is continued\n  
echo onto several lines.\n  
output expression  
Print the value of expression and nothing but  
that value: no newlines, no `$nn = '. The value is  
4. On DOS/Windows systems, the home directory is the one pointed to by the HOME environment variable.  
290 Canned Sequences of Commands  
 
not entered in the value history either. See  
output/fmt expression  
Print the value of expression in format fmt. You  
can use the same formats as for print. See  
printf string, expressions...  
Print the values of the expressionsunder the  
control of string. The expressionsare  
separated by commas and may be either numbers  
or pointers. Their values are printed as specified  
by string, exactly as if your program were to  
execute the C subroutine  
printf (string, expressions...);  
For example, you can print two values in hex like  
this:  
printf "foo, bar-foo = 0x%x, 0x%x\n",  
foo, bar-foo  
The only backslash-escape sequences that you  
can use in the format string are the simple ones  
that consist of backslash followed by a letter.  
18.4 Commands for controlled output 291  
292  
19 Using GDB under gnu Emacs  
A special interface allows you to use gnu Emacs to view (and edit) the source files for  
the program you are debugging with GDB.  
To use this interface, use the command M-x gdb in Emacs. Give the executable file you  
want to debug as an argument. This command starts GDB as a subprocess of Emacs,  
with input and output through a newly created Emacs buffer.  
Using GDB under Emacs is just like using GDB normally except for two things:  
All terminalinput and output goes through the Emacs buffer.  
This applies both to GDB commands and their output, and to the input and output  
done by the program you are debugging.  
This is useful because it means that you can copy the text of previous commands  
and input them again; you can even use parts of the output in this way.  
All the facilities of Emacs' Shell mode are available for interacting with your  
program. In particular, you can send signals the usual way|for example, C-c  
C-cfor an interrupt, C-c C-zfor a stop.  
GDB displays source code through Emacs.  
Each time GDB displays a stack frame, Emacs automatically finds the source file  
for that frame and puts an arrow (`=>') at the left margin of the current line. Emacs  
uses a separate buffer for source display, and splits the screen to show both your  
GDB session and the source.  
Explicit GDB list or search commands still produce output as usual, but you  
probably have no reason to use them from Emacs.  
Warning: If the directory where your program resides is not your current di-  
rectory, it can be easy to confuse Emacs about the location of the source files,  
in which case the auxiliary display buffer does not appear to show your source.  
GDB can nd programs by searching your environment's PATH variable, so the  
GDB input and output session proceeds normally; but Emacs does not get  
enough information back from GDB to locate the source files in this situation.  
To avoid this problem, either start GDB mode from the directory where your  
program resides, or specify an absolute file name when prompted for the M-x  
gdb argument.  
A similar confusion can result if you use the GDB file command to switch to  
debugging a program in some other location, from an existing GDB buffer in  
Emacs.  
By default, M-x gdb calls the program called `gdb'. If you need to call GDB by a  
different name (for example, if you keep several configurations around, with  
different names) you can set the Emacs variable gdb-command-name; for example,  
(setq gdb-command-name "mygdb")  
(preceded by M-:or ESC:, or typed in the *scratch* buffer, or in your .emacs  
file) makes Emacs call the program named mygdbinstead.  
293  
 
In the GDB I/O buffer, you can use these special Emacs commands in addition to the  
standard Shell mode commands:  
C-h m  
Describe the features of Emacs' GDB Mode.  
M-s  
Execute to another source line, like the GDB step command; also  
update the display window to show the current file and location.  
M-n  
Execute to next source line in this function, skipping all function  
calls, like the GDB nextcommand. Then update the display  
window to show the current file and location.  
M-i  
Execute one instruction, like the GDB stepicommand; update  
display window accordingly.  
M-x gdb-nexti  
C-c C-f  
Execute to next instruction, using the GDB nexticommand;  
update display window accordingly.  
Execute until exit from the selected stack frame, like the GDB  
finishcommand.  
M-c  
Continue execution of your program, like the GDB continue  
command.  
WARNING! In Emacs v19, this command is C-c C-p.  
M-u  
Go up the number of frames indicated by the numeric argument  
(see section "Numeric Arguments" in The gnu Emacs Manual), like  
the GDB upcommand.  
WARNING! In Emacs v19, this command is C-c C-u.  
M-d  
Go down the number of frames indicated by the numeric  
argument, like the GDB downcommand.  
WARNING! In Emacs v19, this command is C-c C-d.  
C-x &  
Read the number where the cursor is positioned, and insert it at  
the end of the GDB I/O buffer. For example, if you wish to  
disassemble code around an address that was displayed earlier,  
type disassemble; then move the cursor to the address display,  
and pick up the argument for disassemble by typing C-x &.  
You can customize this further by defining elements of the list  
gdb-print- command; once it is defined, you can format or  
otherwise process numbers picked up by C-x &before they are  
inserted. A numeric argument to C-x &indicates that you wish  
special formatting, and also acts as an index to pick an element  
of the list. If the list element is a string, the number to be inserted  
294 Using GDB under gnu Emacs  
is formatted using the Emacs function format; otherwise the  
number is passed as an argument to the corresponding list  
element.  
In any source file, the Emacs command C-x SPC (gdb-break)tells GDB to set a  
break- point on the source line point is on.  
If you accidentally delete the source-display buffer, an easy way to get it back is to type  
the command fin the GDB buffer, to request a frame display; when you run under  
Emacs, this recreates the source buffer if necessary to show you the context of the  
current frame.  
The source files displayed in Emacs are in ordinary Emacs buffers which are visiting  
the source files in the usual way. You can edit the files with these buffers if you wish;  
but keep in mind that GDB communicates with Emacs in terms of line numbers. If you  
add or delete lines from the text, the line numbers that GDB knows cease to correspond  
properly with the code.  
295  
296  
20 GDB Annotations  
This chapter describes annotations in GDB. Annotations are designed to interface GDB  
to graphical user interfaces or other similar programs which want to interact with GDB  
at a relatively high level.  
20.1 What is an annotation?  
To produce annotations, start GDB with the --annotate=2option.  
Annotations start with a newline character, two control-zcharacters, and the name  
of the annotation. If there is no additional information associated with this annotation,  
the name of the annotation is followed immediately by a newline. If there is additional  
information, the name of the annotation is followed by a space, the additional  
information, and a newline. The additional information cannot contain newline  
characters.  
Any output not beginning with a newline and two control-z characters denotes  
literal output from GDB. Currently there is no need for GDB to output a newline  
followed by two control-zcharacters, but if there was such a need, the annotations  
could be extended with an `escape' annotation which means those three characters as  
output.  
A simple example of starting up GDB with annotations is:  
$ gdb --annotate=2  
GNU GDB 5.0  
Copyright 2000 Free Software Foundation, Inc.  
GDB is free software, covered by the GNU General Public License,  
and you are welcome to change it and/or distribute copies of it  
under certain conditions.  
Type "show copying" to see the conditions.  
There is absolutely no warranty for GDB. Type "show warranty"  
for details.  
This GDB was configured as "sparc-sun-sunos4.1.3"  
^Z^Zpre-prompt  
(gdb)  
^Z^Zprompt  
quit  
^Z^Zpost-prompt  
$
Here quitis input to GDB; the rest is output from GDB. The three lines beginning  
`^Z^Z' (where `^Z' denotes a control-zcharacter) are annotations; the rest is output  
from GDB.  
20.2 The server prefix  
To issue a command to GDB without affecting certain aspects of the state which is seen  
by users, prefix it with server. This means that this command will not affect the  
20.1 What is an annotation? 297  
     
command history, nor will it affect GDB's notion of which command to repeat if RET  
is pressed on a line by itself.  
The server prefix does not affect the recording of values into the value history; to print  
a value without recording it into the value history, use the outputcommand instead  
of the printcommand.  
20.3 Values  
When a value is printed in various contexts, GDB uses annotations to delimit the value  
from the surrounding text.  
If a value is printed using print and added to the value history, the annotation looks  
like:  
^Z^Zvalue-history-begin history-number value-flags  
history-string  
^Z^Zvalue-history-value  
the-value  
^Z^Zvalue-history-end  
where history-numberis the number it is getting in the value history,  
history-stringis a string, such as $5 = , which introduces the value to the user,  
the-valueis the output corresponding to the value itself, and value- flagsis `*'  
for a value which can be dereferenced and `-' for a value which cannot.  
If the value is not added to the value history (it is an invalid float or it is printed with  
the outputcommand), the annotation is similar:  
^Z^Zvalue-begin value-flags  
the-value  
^Z^Zvalue-end  
When GDB prints an argument to a function (for example, in the output from the  
backtracecommand), it annotates it as follows:  
^Z^Zarg-begin  
argument-name  
^Z^Zarg-name-end  
separator-string  
^Z^Zarg-value value-flags  
the-value  
^Z^Zarg-end  
where argument-nameis the name of the argument, separator-stringis text  
which separates the name from the value for the user's benefit (such as =), and  
value-flagsand the-valuehave the same meanings as in a  
value-history-beginannotation.  
When printing a structure, GDB annotates it as follows:  
^Z^Zfield-begin value-flags  
field-name  
^Z^Zfield-name-end  
separator-string  
298 GDB Annotations  
 
^Z^Zfield-value  
the-value  
^Z^Zfield-end  
where field-nameis the name of the field, separator-stringis text which  
separates the name from the value for the user's benefit (such as `='), and value-flags  
and the-valuehave the same meanings as in a value-history-beginannotation.  
When printing an array, GDB annotates it as follows:  
^Z^Zarray-section-begin array-index value-flags  
where array-indexis the index of the first element being annotated and  
value-flagshas the same meaning as in a value-history-beginannotation.  
This is followed by any number of elements, where the element can be either a single  
element or a repeated element as shown in the examples below:  
`,' whitespace ; omitted for the first element  
the-value  
^Z^Zelt  
`,' whitespace ; omitted for the first element  
the-value  
^Z^Zelt-rep number-of-repititions  
repetition-string  
^Z^Zelt-rep-end  
In both cases, the-valueis the output for the value of the element and whitespace  
can contain spaces, tabs, and newlines. In the repeated case, number-of-repititons  
is the number of consecutive array elements which contain that value, and  
repetition-stringis a string which is designed to convey to the user that repetition  
is being depicted.  
Once all the array elements have been output, the array annotation is ended with:  
^Z^Zarray-section-end  
20.4 Frames  
Whenever GDB prints a frame, it annotates it. For example, this applies to frames  
printed when GDB stops, output from commands such as backtrace or up, etc.  
The frame annotation begins with:  
^Z^Zframe-begin level address  
level-string  
where levelis the number of the frame (0 is the innermost frame, and other frames  
have positive numbers), addressis the address of the code executing in that frame,  
and level-stringis a string designed to convey the level to the user. addressis  
in the form `0x' followed by one or more lowercase hex digits (note that this does not  
depend on the language). The frame ends with:  
^Z^Zframe-end  
Between these annotations is the main body of the frame, which can consist of:  
20.4 Frames 299  
 
^Z^Zfunction-call  
function-call-string  
where function-call-stringis text designed to convey to the user that this  
frame is associated with a function call made by GDB to a function in the program  
being debugged.  
^Z^Zsignal-handler-caller  
signal-handler-caller-string  
where signal-handler-caller-stringis text designed to convey to the user  
that this frame is associated with whatever mechanism is used by this operating  
system to call a signal handler (it is the frame which calls the signal handler, not  
the frame for the signal handler itself).  
A normal frame.  
This can optionally (depending on whether this is thought of as interesting  
information for the user to see) begin with  
^Z^Zframe-address  
address  
^Z^Zframe-address-end  
separator-string  
where addressis the address executing in the frame (the same address as in the  
frame-beginannotation, but printed in a form which is intended for user  
consumption|in particular, the syntax varies depending on the language), and  
separator-stringis a string intended to separate this address from what  
follows for the user's benefit.  
Then comes  
^Z^Zframe-function-name  
function-name  
^Z^Zframe-args  
arguments  
where function-nameis the name of the function executing in the frame, or `??'  
if not known, and argumentsare the arguments to the frame, with parentheses  
around them (each argument is annotated individually as well, see “Values”  
If source information is available, a reference to it is then printed:  
^Z^Zframe-source-begin  
source-intro-string  
^Z^Zframe-source-file  
filename  
^Z^Zframe-source-file-end  
:
^Z^Zframe-source-line  
line-number  
^Z^Zframe-source-end  
300 GDB Annotations  
where source-intro-stringseparates for the user's benefit the reference from  
the text which precedes it, filenameis the name of the source file, and  
line-numberis the line number within that file (the first line is line 1).  
If GDB prints some information about where the frame is from (which library,  
which load segment, etc.; currently only done on the RS/6000), it is annotated with  
^Z^Zframe-where  
information  
Then, if source is to be actually displayed for this frame (for example, this is not  
true for output from the backtrace command), then a source annotation (see  
“Displaying source” (page 304)) is displayed. Unlike most annotations, this is output  
instead of the normal text which would be output, not in addition.  
20.5 Displays  
When GDB is told to display something using the displaycommand, the results of  
the display are annotated:  
^Z^Zdisplay-begin  
number  
^Z^Zdisplay-number-end  
number-separator  
^Z^Zdisplay-format  
format  
^Z^Zdisplay-expression  
expression  
^Z^Zdisplay-expression-end  
expression-separator  
^Z^Zdisplay-value  
value  
^Z^Zdisplay-end  
where numberis the number of the display, number-separatoris intended to  
separate the number from what follows for the user, formatincludes information  
such as the size, format, or other informationabout how the value is being displayed,  
expressionis the expression being displayed, expression-separatoris intended  
to separate the expression from the text that follows for the user,and valueis the actual  
value being displayed.  
20.6 Annotation for GDB input  
When GDB prompts for input, it annotates this fact so it is possible to know when to  
send output, when the output from a given command is over, etc.  
Different kinds of input each have a different input type. Each input type has three  
annotations: a preannotation, which denotes the beginning of any prompt which is  
being output, a plain annotation, which denotes the end of the prompt, and then a post-  
annotation which denotes the end of any echo which may (or may not) be associated  
with the input. For example, the prompt input type features the following annotations:  
20.5 Displays 301  
   
^Z^Zpre-prompt  
^Z^Zprompt  
^Z^Zpost-prompt  
The input types are:  
prompt  
When GDB is prompting for a command (the main GDB  
prompt).  
commands  
When GDB prompts for a set of commands, like in the  
commands command. The annotations are repeated for  
each command which is input.  
overload-choice  
query  
When GDB wants the user to select between various  
overloaded functions.  
When GDB wants the user to confirm a potentially  
dangerous operation.  
prompt-for-continue  
When GDB is asking the user to press return to continue.  
Note: Don't expect this to work well; instead use set  
height 0to disable prompting. This is because the  
counting of lines is buggy in the presence of annotations.  
20.7 Errors  
^Z^Zquit  
This annotation occurs right before GDB responds to an interrupt.  
^Z^Zerror  
This annotation occurs right before GDB responds to an error.  
Quit and error annotations indicate that any annotations which GDB was in the middle  
of may end abruptly. For example, if a value-history-beginannotation is followed  
by a error, one cannot expect to receive the matching value-history-end. One  
cannot expect not to receive it either; however, an error annotation does not necessarily  
mean that GDB is immediately returning all the way to the top level.  
A quit or error annotation may be preceded by:  
^Z^Zerror-begin  
Any output between that and the quit or error annotation is the error message.  
Warning messages are not yet annotated.  
20.8 Information on breakpoints  
The output from the info breakpoints command is annotated as follows:  
^Z^Zbreakpoints-headers  
header-entry  
^Z^Zbreakpoints-table  
302 GDB Annotations  
   
where header-entryhas the same syntax as an entry (see below) but instead of  
containing data, it contains strings which are intended to convey the meaning of each  
field to the user. This is followed by any number of entries. If a field does not apply  
for this entry, it is omitted. Fields may contain trailing whitespace. Each entry consists  
of:  
^Z^Zrecord  
^Z^Zfield 0  
number  
^Z^Zfield 1  
type  
^Z^Zfield 2  
disposition  
^Z^Zfield 3  
enable  
^Z^Zfield 4  
address  
^Z^Zfield 5  
what  
^Z^Zfield 6  
frame  
^Z^Zfield 7  
condition  
^Z^Zfield 8  
ignore-count  
^Z^Zfield 9  
commands  
Note that addressis intended for user consumption|the syntax varies depending on  
the language.  
The output ends with:  
^Z^Zbreakpoints-table-end  
20.9 Invalidation notices  
The following annotations say that certain pieces of state may have changed:  
^Z^Zframes-invalid  
The frames (for example, output from the  
backtrace command) may have changed.  
^Z^Zbreakpoints-invalid  
The breakpoints may have changed. For example,  
the user just added or deleted a breakpoint.  
20.10 Running the program  
When the program starts executing due to a GDB command such as stepor continue,  
^Z^Zstarting  
is output. When the program stops,  
^Z^Zstopped  
20.9 Invalidation notices 303  
   
is output. Before the stoppedannotation, a variety of annotations describe how the  
program stopped.  
^Z^Zexited exit-status  
The program exited, and exit-statusis the exit  
status (zero for successful exit, otherwise nonzero).  
^Z^Zsignalled  
The program exited with a signal. After the  
^Z^Zsignalled, the annotation continues:  
intro-text  
^Z^Zsignal-name  
name  
^Z^Zsignal-name-end  
middle-text  
^Z^Zsignal-string  
string  
^Z^Zsignal-string-end  
end-text  
where nameis the name of the signal, such as  
SIGILL or SIGSEGV, and stringis the explanation  
of the signal, such as Illegal Instructionor  
Segmentation fault. intro-text,  
middle-text, and end-textare for the user's  
benefit and have no particular format.  
^Z^Zsignal  
The syntax of this annotation is just like signalled,  
but GDB is just saying that the program received  
the signal, not that it was terminated with it.  
^Z^Zbreakpoint number  
The program hit breakpoint number number.  
^Z^Zwatchpoint number  
The program hit watchpoint number number.  
20.11 Displaying source  
The following annotation is used instead of displaying source code:  
^Z^Zsource filename:line:character:middle:addr  
where filenameis an absolute file name indicating which source file, lineis the line  
number within that file (where 1 is the first line in the file), characteris the character  
position within the file (where 0 is the first character in the file, for most debug formats  
this will necessarily point to the beginning of a line), middleis `middle' if addris in  
the middle of the line, or `beg' if addris at the beginning of the line, and addris the  
address in the target program associated with the source which is being displayed.  
addris in the form `0x' followed by one or more lowercase hex digits (note that this  
does not depend on the language).  
304 GDB Annotations  
 
20.12 Annotations We Might Want in the Future  
target-invalid  
the target might have changed (registers, heap contents, or execution status). For  
performance, we might eventually want to hit registers-invalidand  
all-registers-invalidwith greater precision  
systematic annotation for set/show parameters (including invalidation notices).  
similarly, `info' returns a list of candidates for invalidation notices.  
20.12 Annotations We Might Want in the Future 305  
 
306  
21 The GDB/MIInterface  
Function and purpose  
GDB/MIis a line based machine oriented text interface to GDB. It is specifically intended  
to support the development of systems which use the debugger as just one small  
component of a larger system.  
This chapter is a specification of the GDB/MIinterface. It is written in the form of a  
reference manual.  
Notation and terminology  
This chapter uses the following notation:  
|separates two alternatives.  
[ something] indicates that somethingis optional: it may or may not be given.  
( group)* means that groupinside the parentheses may repeat zero or more  
times.  
( group)+ means that groupinside the parentheses may repeat one or more times.  
string” means a literal string.  
Acknowledgments  
In alphabetic order: Andrew Cagney, Fernando Nasser, Stan Shebs and Elena Zannoni.  
21.1 GDB/MICommand Syntax  
21.1.1 GDB/MIInput syntax  
command  
cli-command| mi-command  
cli-command  
[ token] cli-command nl, where cli-commandis  
any existing GDB CLI command.  
mi-command  
[ token] "-" operation( " " option)* [ " --" ] ( "  
" parameter)* nl  
token  
"any sequence of digits"  
option  
"-" parameter[ " " parameter]  
non-blank-sequence| c-string  
any of the operations described in this chapter  
anything, provided it does not contain special  
characters such as “-”, nl, """ and of course “”  
parameter  
operation  
non-blank-sequence  
c-string  
nl  
""" seven-bit-iso-c-string-content"""  
CR | CR-LF  
21.1 GDB/MI Command Syntax 307  
     
Notes:  
The CLI commands are still handled by the MIinterpreter; their output is described  
below.  
The token, when present, is passed back when the command finishes.  
Some MIcommands accept optional arguments as part of the parameter list. Each  
option is identified by a leading `-' (dash) and may be followed by an optional  
argument parameter. Options occur first in the parameter list and can be delimited  
from normal parameters using `--' (this is useful when some parameters begin  
with a dash).  
Pragmatics:  
We want easy access to the existing CLI syntax (for debugging).  
We want it to be easy to spot a MIoperation.  
21.1.2 GDB/MIOutput syntax  
The output from GDB/MIconsists of zero or more out-of-band records followed,  
optionally, by a single result record. This result record is for the most recent command.  
The sequence of output records is terminated by '(gdb)'.  
If an input command was prefixed with a token then the corresponding output for that  
command will also be prefixed by that same token.  
output  
( out-of-band-record )* [  
result-record ] "(gdb)" nl  
result-record  
[ token ] "^" result-class ( ","  
result )* nl  
out-of-band-record →  
async-record | stream-record  
async-record  
exec-async-output |  
status-async-output |  
notify-async-output  
exec-async-output  
status-async-output  
notify-async-output  
async-output  
[ token ] "*" async-output  
[ token ] "+" async-output  
[ token ] "=" async-output  
async-class ( "," result )* nl  
result-class  
"done" | "running" | "connected" |  
"error" | "exit"  
async-class  
"stopped" | others(where otherswill be  
added depending on the needs―this is still in  
development).  
result  
variable "=" value  
variable  
string  
308 The GDB/MIInterface  
 
value  
const  
tuple  
list  
const | tuple | list  
c-string  
"{}" | "{" result ( "," result )* "}"  
"[]" | "[" value ( "," value )* "]"  
| "[" result ( "," result )* "]"  
stream-record  
console-stream-output |  
target-stream-output |  
log-stream-output  
console-stream-output  
"~" c-string  
target-stream-output  
"@" c-string  
log-stream-output  
"&" c-string  
nl  
CR | CR-LF  
token  
any sequence of digits.  
Notes:  
All output sequences end in a single line containing a period.  
The tokenis from the corresponding request. If an execution command is  
interrupted by the '-exec-interrupt' command, the tokenassociated with the  
'*stopped' message is the one of the original execution command, not the one of  
the interrupt command.  
status-async-outputcontains on-going status information about the progress  
of a slow operation. It can be discarded. All status output is prefixed by '+'.  
exec-async-outputcontains asynchronous state change on the target (stopped,  
started, disappeared). All async output is prefixed by '*'.  
notify-async-outputcontains supplementary information that the client  
should handle (for example, a new breakpoint information). All notify output is  
prefixed by '='.  
console-stream-outputis output that should be displayed as is in the console.  
It is the textual response to a CLI command. All the console output is prefixed by  
'~'.  
target-stream-outputis the output produced by the target program. All the  
target output is prefixed by '@'.  
log-stream-outputis output text coming from GDB's internals, for instance  
messages that should be displayed as part of an error log. All the log output is  
prefixed by '&'.  
New GDB/MIcommands should only output lists containing values.  
See GDB/MIstream records” (page 311), for more details about the various output  
records.  
21.1 GDB/MI Command Syntax 309  
21.1.3 Simple examples of GDB/MIinteraction  
This subsection presents several simple examples of interaction using the GDB/MI  
interface. In these examples, '->' means that the following line is passed to GDB/MIas  
input, while '<-' means the output received from GDB/MI.  
Evaluate expression  
Here is an example to evaluate an expression:  
-> -data-evaluate-expression 2+3  
<- (gdb)  
<- ^done,value="5"  
<- (gdb)  
and later:  
<- *stop,reason="stop",address="0x123",source="a.c:123"  
<- (gdb)  
Simple CLI command  
Here is an example of a simple CLI command being passed through GDB/MIand on  
to the CLI.  
-> print 1+2  
<- &"print 1+2\n"  
<- ~"$1 = 3\n"  
<- ^done  
<- (gdb)  
A bad command  
Here is what happens if you pass a bad command:  
-> -rubbish  
<- ^error,msg="Undefined MI command: rubbish"  
<- (gdb)  
21.2 GDB/MIcompatibility with CLI  
To help users get familiar with GDB CLI, GDB/MIaccepts existing CLI commands. As  
specified by the syntax, such commands can be directly entered into the GDB/MI  
interface and GDB will respond.  
This mechanism is provided as an aid to developers of GDB/MIclients and not as a  
reliable interface into the CLI. Since the command is being interpreted in an environment  
that assumes GDB/MIbehaviour, the exact output of such commands is likely to end  
up being an un-supported hybrid of GDB/MIand CLI output.  
310 The GDB/MIInterface  
   
21.3 GDB/MIoutput records  
21.3.1 GDB/MIresult records  
In addition to a number of out-of-band notifications, the response to a GDB/MIcommand  
includes one of the following result indications:  
"^done" [ "," results ]  
The synchronous operation was successful,  
resultsare the return values.  
"^running"  
The asynchronous operation was successfully  
started. The target is running.  
"^error" "," c-string  
The operation failed. The c-stringcontains the  
corresponding error message.  
21.3.2 GDB/MIstream records  
GDB internally maintains a number of output streams: the console, the target, and the  
log. The output intended for each of these streams is funneled through the gdb/mi  
interface using stream records.  
Each stream record begins with a unique prefix character which identifies its stream (see  
GDB/MIOutput syntax” (page 308)). In addition to the prefix, each stream record  
contains a string-output. This is either raw text (with an implicit new line) or a  
quoted C string (which does not contain an implicit new line).  
"~" string-output  
The console output stream contains text that should be  
displayed in the CLI console window. It contains the textual  
responses to CLI commands.  
"@" string-output  
The target output stream contains any textual output from  
the running target.  
"&" string-output  
The log stream contains internal debugging messages being  
produced by GDB.  
21.3.3 GDB/MIout-of-band records  
Out-of-band records are used to notify the GDB/MIclient of additional changes that  
have occurred. Those changes can either be a consequence of GDB/MI(for example, a  
breakpoint modified) or a result of target activity (for example, target stopped).  
The following is a preliminary list of possible out-of-band records.  
"*" "stop"  
21.4 GDB/MIcommand description format  
The remaining sections describe blocks of commands. Each block of commands is laid  
out in a fashion similar to this section.  
21.3 GDB/MI output records 311  
         
Note the line breaks shown in the examples are here only for readability. They do not  
appear in the real output. Also note that the commands with a non-available example  
(N.A.) are not yet implemented.  
Motivation  
The motivation for this collection of commands.  
Introduction  
A brief introduction to this collection of commands as a whole.  
Commands  
For each command in the block, the following is described:  
Synopsis  
-command args...  
GDB command  
The corresponding GDB CLI command.  
Result  
Out-of-band  
Notes  
Example  
21.5 GDB/MIbreakpoint table commands  
This section documents GDB/MIcommands for manipulating breakpoints.  
The -break-afterCommand  
Synopsis  
-break-after number count  
The breakpoint number numberis not in effect until it has been hit counttimes. To  
see how this is reflected in the output of the '-break-list' command, see the  
description of the '-break-list' command below.  
GDB command  
The corresponding GDB command is 'ignore'.  
312 The GDB/MIInterface  
 
Example  
(gdb)  
-break-insert main  
^done,bkpt=number="1",type="breakpoint",disp="keep",enabled="y",  
addr="0x000100d0",func="main",file="hello.c",line="5",times="0"(gdb)  
-break-after 1 3  
~
^done  
(gdb)  
-break-list  
^done,BreakpointTable={nr_rows="1",nr_cols="6",  
hdr=[{width="3",alignment="-1",col_name="number",colhdr="Num"},  
{width="14",alignment="-1",col_name="type",colhdr="Type"},  
{width="4",alignment="-1",col_name="disp",colhdr="Disp"},  
{width="3",alignment="-1",col_name="enabled",colhdr="Enb"},  
{width="10",alignment="-1",col_name="addr",colhdr="Address"},  
{width="40",alignment="2",col_name="what",colhdr="What"}],  
body=[bkpt={number="1",type="breakpoint",disp="keep",enabled="y",  
addr="0x000100d0",func="main",file="hello.c",line="5",times="0",  
ignore="3"}]}  
(gdb)  
The -break-conditioncommand  
Synopsis  
-break-condition number expr  
Breakpoint numberwill stop the program only if the condition in expris true. The  
condition becomes part of the '-break-list' output (see the description of the  
'-break-list' command below).  
GDB command  
The corresponding GDB command is 'condition'.  
Example  
(gdb)  
-break-condition 1 1  
^done  
(gdb)  
-break-list  
^done,BreakpointTable={nr_rows="1",nr_cols="6",  
hdr=[{width="3",alignment="-1",col_name="number",colhdr="Num"},  
{width="14",alignment="-1",col_name="type",colhdr="Type"},  
{width="4",alignment="-1",col_name="disp",colhdr="Disp"},  
{width="3",alignment="-1",col_name="enabled",colhdr="Enb"},  
{width="10",alignment="-1",col_name="addr",colhdr="Address"},  
{width="40",alignment="2",col_name="what",colhdr="What"}],  
body=[bkpt={number="1",type="breakpoint",disp="keep",enabled="y",  
21.5 GDB/MI breakpoint table commands 313  
addr="0x000100d0",func="main",file="hello.c",line="5",cond="1",  
times="0",ignore="3"}]}  
(gdb)  
The -break-deletecommand  
Synopsis  
-break-delete ( breakpoint )+  
Delete the breakpoint(s) whose number(s) are specified in the argument list. This is  
obviously reflected in the breakpoint list.  
GDB command  
The corresponding GDB command is 'delete'.  
Example  
(gdb)  
-break-delete 1  
^done  
(gdb)  
-break-list  
^done,BreakpointTable={nr_rows="0",nr_cols="6",  
hdr=[{width="3",alignment="-1",col_name="number",colhdr="Num"},  
{width="14",alignment="-1",col_name="type",colhdr="Type"},  
{width="4",alignment="-1",col_name="disp",colhdr="Disp"},  
{width="3",alignment="-1",col_name="enabled",colhdr="Enb"},  
{width="10",alignment="-1",col_name="addr",colhdr="Address"},  
{width="40",alignment="2",col_name="what",colhdr="What"}],  
body=[]}  
(gdb)  
The -break-disablecommand  
Synopsis  
-break-disable ( breakpoint )+  
Disable the named breakpoint(s). The field 'enabled' in the break list is now set to  
'n' for the named breakpoint(s).  
GDB command  
The corresponding GDB command is 'disable'.  
Example  
(gdb)  
-break-disable 2  
^done  
(gdb)  
314 The GDB/MIInterface  
-break-list  
^done,BreakpointTable={nr_rows="1",nr_cols="6",  
hdr=[{width="3",alignment="-1",col_name="number",colhdr="Num"},  
{width="14",alignment="-1",col_name="type",colhdr="Type"},  
{width="4",alignment="-1",col_name="disp",colhdr="Disp"},  
{width="3",alignment="-1",col_name="enabled",colhdr="Enb"},  
{width="10",alignment="-1",col_name="addr",colhdr="Address"},  
{width="40",alignment="2",col_name="what",colhdr="What"}],  
body=[bkpt={number="2",type="breakpoint",disp="keep",enabled="n",  
addr="0x000100d0",func="main",file="hello.c",line="5",times="0"}]}  
(gdb)  
The -break-enablecommand  
Synopsis  
-break-enable ( breakpoint )+  
Enable (previously disabled) breakpoint(s).  
GDB command  
The corresponding GDB command is 'enable'.  
Example  
(gdb)  
-break-enable 2  
^done  
(gdb)  
-break-list  
^done,BreakpointTable={nr_rows="1",nr_cols="6",  
hdr=[{width="3",alignment="-1",col_name="number",colhdr="Num"},  
{width="14",alignment="-1",col_name="type",colhdr="Type"},  
{width="4",alignment="-1",col_name="disp",colhdr="Disp"},  
{width="3",alignment="-1",col_name="enabled",colhdr="Enb"},  
{width="10",alignment="-1",col_name="addr",colhdr="Address"},  
{width="40",alignment="2",col_name="what",colhdr="What"}],  
body=[bkpt={number="2",type="breakpoint",disp="keep",enabled="y",  
addr="0x000100d0",func="main",file="hello.c",line="5",times="0"}]}  
(gdb)  
The -break-infoCommand  
Synopsis  
-break-info breakpoint  
Get information about a single breakpoint.  
GDB command  
The corresponding GDB command is 'info break breakpoint'.  
21.5 GDB/MI breakpoint table commands 315  
Example  
N.A.  
The -break-insertcommand  
Synopsis  
-break-insert [ -t ] [ -h ] [ -r ]  
[ -c condition ] [ -i ignore-count ]  
[ -p thread ] [ line | addr ]  
If specified, line, can be one of:  
function  
filename:linenum  
filename:function  
*address  
The possible optional parameters of this command are:  
'-t'  
Insert a temporary breakpoint.  
'[-h]'  
Insert a hardware breakpoint.  
'[-c condition]'  
'[-i ignore-count]'  
`-r'  
Make the breakpoint conditional on condition.  
Initialize the ignore-count.  
Insert a regular breakpoint in all the functions whose  
names match the given regular expression. Other flags  
are not applicable to regular expression.  
Result  
The result is in the form:  
^done,bkptno="number",func="funcname",  
file="filename",line="lineno"  
where numberis the GDB number for this breakpoint, funcnameis the name of the  
function where the breakpoint was inserted, filenameis the name of the source file  
which contains this function, and linenois the source line number within that file.  
NOTE: This format is open to change.  
GDB command  
The corresponding GDB commands are 'break', 'tbreak', 'hbreak', 'thbreak', and  
'rbreak'.  
Example  
(gdb)  
-break-insert main  
316 The GDB/MIInterface  
^done,bkpt=number="1",type="breakpoint",disp="keep",enabled="y",addr="0x0001072c",  
file="recursive2.c",line="4",times="0"  
(gdb)  
-break-insert -t foo  
^done,bkpt=number="2",type="breakpoint",disp="keep",enabled="y",addr="0x00010774",  
file="recursive2.c",line="11",times="0"(gdb)  
-break-list  
^done,BreakpointTable={nr_rows="2",nr_cols="6",  
hdr=[{width="3",alignment="-1",col_name="number",colhdr="Num"},  
{width="14",alignment="-1",col_name="type",colhdr="Type"},  
{width="4",alignment="-1",col_name="disp",colhdr="Disp"},  
{width="3",alignment="-1",col_name="enabled",colhdr="Enb"},  
{width="10",alignment="-1",col_name="addr",colhdr="Address"},  
{width="40",alignment="2",col_name="what",colhdr="What"}],  
body=[bkpt={number="1",type="breakpoint",disp="keep",enabled="y",  
addr="0x0001072c", func="main",file="recursive2.c",line="4",times="0"},  
bkpt={number="2",type="breakpoint",disp="del",enabled="y",  
addr="0x00010774",func="foo",file="recursive2.c",line="11",times="0"}]}  
(gdb)  
-break-insert -r foo.*  
~int foo(int, int);  
^done,bkpt={number="3",addr="0x00010774",file="recursive2.c",line="11"}  
(gdb)  
The -break-listcommand  
Synopsis  
-break-list  
Displays the list of inserted breakpoints, showing the following fields:  
'Number'  
number of the breakpoint  
'Type'  
type of the breakpoint: 'breakpoint' or 'watchpoint'  
'Disposition'  
should the breakpoint be deleted or disabled when it is hit: 'keep'  
or 'nokeep'  
'Enabled'  
'Address'  
'What'  
is the breakpoint enabled or no: 'y' or 'n'  
memory location at which the breakpoint is set  
logical location of the breakpoint, expressed by function name,  
file name, line number  
'Times'  
number of times the breakpoint has been hit  
If there are no breakpoints or watchpoints, the BreakpointTable bodyfield is an  
empty list.  
GDB command  
The corresponding GDB command is 'info break'.  
Example  
(gdb)  
-break-list  
^done,BreakpointTable={nr_rows="2",nr_cols="6",  
21.5 GDB/MI breakpoint table commands 317  
hdr=[{width="3",alignment="-1",col_name="number",colhdr="Num"},  
{width="14",alignment="-1",col_name="type",colhdr="Type"},  
{width="4",alignment="-1",col_name="disp",colhdr="Disp"},  
{width="3",alignment="-1",col_name="enabled",colhdr="Enb"},  
{width="10",alignment="-1",col_name="addr",colhdr="Address"},  
{width="40",alignment="2",col_name="what",colhdr="What"}],  
body=[bkpt={number="1",type="breakpoint",disp="keep",enabled="y",  
addr="0x000100d0",func="main",file="hello.c",line="5",times="0"},  
bkpt={number="2",type="breakpoint",disp="keep",enabled="y",  
addr="0x00010114",func="foo",file="hello.c",line="13",times="0"}]}  
(gdb)  
Here is an example of the result when there are no breakpoints:  
(gdb)  
-break-list  
^done,BreakpointTable={nr_rows="0",nr_cols="6",  
hdr=[{width="3",alignment="-1",col_name="number",colhdr="Num"},  
{width="14",alignment="-1",col_name="type",colhdr="Type"},  
{width="4",alignment="-1",col_name="disp",colhdr="Disp"},  
{width="3",alignment="-1",col_name="enabled",colhdr="Enb"},  
{width="10",alignment="-1",col_name="addr",colhdr="Address"},  
{width="40",alignment="2",col_name="what",colhdr="What"}],  
body=[]}  
(gdb)  
The -break-watchcommand  
Synopsis  
-break-watch [ -a | -r ]  
Create a watchpoint. With the '-a' option it will create an access watchpoint, that is a  
watchpoint that triggers either on a read from or on a write to the memory location.  
With the '-r' option, the watchpoint created is a read watchpoint, that is it will trigger  
only when the memory location is accessed for reading. Without either of the options,  
the watchpoint created is a regular watchpoint, that is it will trigger when the memory  
location is accessed for writing. See “Enhanced support for watchpoints and  
Note that '-break-list' will report a single list of watchpoints and breakpoints  
inserted.  
GDB command  
The corresponding GDB commands are 'watch', 'awatch', and 'rwatch'.  
Example  
Setting a watchpoint on a variable in the mainfunction:  
(gdb)  
-break-watch i  
^done,wpt=number="2",exp="i"  
318 The GDB/MIInterface  
(gdb)  
-exec-continue  
^running  
(gdb)  
*stopped,reason="watchpoint-trigger",wpt=number="2",exp="i",value=old="0",new="7"  
,thread-id="1",frame=addr="0x000029c4",func="main",args=[],file="hello.c",line="8"  
(gdb)  
Setting a watchpoint on a variable local to a function. GDB will stop the program  
execution twice: first for the variable changing value, then for the watchpoint going  
out of scope.  
(gdb)  
-break-watch j  
^done,wpt=number="2",exp="j"  
(gdb)  
-exec-continue  
^running  
(gdb)  
*stopped,reason="watchpoint-trigger",wpt=number="2",exp="j",value=old="0",new="17",  
thread-id="1",frame=addr="0x000029bc",func="call",args=[],file="hello.c",line="10"  
(gdb)  
-exec-continue  
^running  
(gdb)  
*stopped,reason="watchpoint-scope",wpnum="2",thread-id="1",frame=addr="0x000029ec",  
func="main",args=[],file="hello.c",line="18"  
(gdb)  
Listing breakpoints and watchpoints, at different points in the program execution. Note  
that once the watchpoint goes out of scope, it is deleted.  
-break-watch j  
^done,wpt=number="2",exp="j"  
(gdb)  
-break-list  
^done,BreakpointTable=nr_rows="2",nr_cols="6",hdr=[width="3",  
alignment="-1",col_name="number",colhdr="Num",width="14",alignment="-1",  
col_name="type",colhdr"Type",width="4",alignment="-1",col_name="disp",  
colhdr="Disp",width="3",alinment="-1",col_name="enabled",colhdr="Enb",  
width="10",alignment="-1",col_name"addr",colhdr="Address",  
width="40",alignment="2",col_name="what",colhdr="What],body=[bkpt=number="1",  
type="breakpoint",disp="keep",enabled="y",addr="0x00029b4",func="call",  
file="hello.c",line="9",times="1",bkpt=number="2",type="wathpoint",disp="keep",  
enabled="y",addr="",what="j",times="0"]  
(gdb)  
-exec-continue  
^running  
(gdb)  
*stopped,reason="watchpoint-trigger",wpt=number="2",exp="j",value=old="0",ne="17",  
thread-id="1",frame=addr="0x000029bc",func="call",args=[],file="hello.c,line="10"  
(gdb)  
-break-list  
^done,BreakpointTable=nr_rows="2",nr_cols="6",hdr=[width="3",alignment="-1",  
col_name="number",colhdr="Num",width="14",alignment="-1",col_name="type",colhdr"Type",  
width="4",alignment="-1",col_name="disp",colhdr="Disp",width="3",alinment="-1",  
col_name="enabled",colhdr="Enb",width="10",alignment="-1",col_name"addr",colhdr="Address",  
width="40",alignment="2",col_name="what",colhdr="What],body=[bkpt=number="1",  
type="breakpoint",disp="keep",enabled="y",addr="0x00029b4",func="call",file="hello.c",  
line="9",times="1",bkpt=number="2",type="wathpoint",disp="keep",enabled="y",addr="",  
what="j",times="1"]  
(gdb)  
-exec-continue  
^running  
(gdb)  
*stopped,reason="watchpoint-scope",wpnum="2",thread-id="1",frame=addr="0x000029ec",  
func="main",args=[],file="hello.c",line="18"  
(gdb)  
-break-list  
21.5 GDB/MI breakpoint table commands 319  
^done,BreakpointTable=nr_rows="1",nr_cols="6",hdr=[width="3",alignment="-1",  
col_name="number",colhdr="Num",width="14",alignment="-1",col_name="type",colhdr="Type",  
width="4",alignment="-1",col_name="disp",colhdr="Disp",width="3",alignment="-1",  
col_name="enabled",colhdr="Enb",width="10",alignment="-1",col_name="addr",  
colhdr="Address",width="40",alignment="2",col_name="what",colhdr="What"],  
body=[bkpt=number="1",type="breakpoint",disp="keep",enabled="y",addr="0x000029b4",  
func="call",file="hello.c",line="9",times="1"]  
(gdb)  
21.6 GDB/MIData manipulation  
This section describes the GDB/MIcommands that manipulate data: examine memory  
and registers, evaluate expressions, and so on.  
The -data-disassemblecommand  
Synopsis  
-data-disassemble  
[ -s start-addr -e end-addr ]  
| [ -f filename -l linenum [ -n lines ] ]  
-- mode  
Where:  
'start-addr'  
'end-addr'  
'filename'  
'linenum'  
'lines'  
is the beginning address (or $pc)  
is the end address  
is the name of the file to disassemble  
is the line number to disassemble around  
is the number of disassembly lines to be produced. If it is -1, the  
whole function will be disassembled, in case no end-addris  
specified. If end-addris specified as a non-zero value, and lines  
is lower than the number of disassembly lines between  
start-addrand end-addr, only lineslines are displayed; if  
linesis higher than the number of lines between start-addr  
and end-addr, only the lines up to end-addrare displayed.  
'mode'  
is either 0 (meaning only disassembly) or 1 (meaning mixed source  
and disassembly).  
Result  
The output for each instruction is composed of four fields:  
Address  
Func-name  
Offset  
Instruction  
Note that whatever included in the instruction field, is not manipulated directly by  
GDB/MI, that is it is not possible to adjust its format.  
320 The GDB/MIInterface  
 
GDB command  
There is no direct mapping from this command to the CLI.  
Example  
Disassemble from the current value of $pcto $pc + 20:  
(gdb)  
-data-disassemble -s $pc -e "$pc + 20" -- 0  
^done,  
asm_insns=[  
{address="0x000107c0",func-name="main",offset="4",  
inst="mov 2, %o0"},  
{address="0x000107c4",func-name="main",offset="8",  
inst="sethi %hi(0x11800), %o2"},  
{address="0x000107c8",func-name="main",offset="12",  
inst="or %o2, 0x140, %o1\t! 0x11940 <_lib_version+8>"},  
{address="0x000107cc",func-name="main",offset="16",  
inst="sethi %hi(0x11800), %o2"},  
{address="0x000107d0",func-name="main",offset="20",  
inst="or %o2, 0x168, %o4\t! 0x11968 <_lib_version+48>"}]  
(gdb)  
Disassemble the whole mainfunction. Line 32 is part of main.  
-data-disassemble -f basics.c -l 32 -- 0  
^done,asm_insns=[  
{address="0x000107bc",func-name="main",offset="0",  
inst="save %sp, -112, %sp"},  
{address="0x000107c0",func-name="main",offset="4",  
inst="mov 2, %o0"},  
{address="0x000107c4",func-name="main",offset="8",  
inst="sethi %hi(0x11800), %o2"},  
[...]  
{address="0x0001081c",func-name="main",offset="96",inst="ret "},  
{address="0x00010820",func-name="main",offset="100",inst="restore "}]  
(gdb)  
Disassemble 3 instructions from the start of main:  
(gdb)  
-data-disassemble -f basics.c -l 32 -n 3 -- 0  
^done,asm_insns=[  
{address="0x000107bc",func-name="main",offset="0",  
inst="save %sp, -112, %sp"},  
{address="0x000107c0",func-name="main",offset="4",  
inst="mov 2, %o0"},  
{address="0x000107c4",func-name="main",offset="8",  
inst="sethi %hi(0x11800), %o2"}]  
(gdb)  
Disassemble 3 instructions from the start of mainin mixed mode:  
(gdb)  
-data-disassemble -f basics.c -l 32 -n 3 -- 1  
21.6 GDB/MI Data manipulation 321  
^done,asm_insns=[  
src_and_asm_line={line="31",  
file="/kwikemart/marge/ezannoni/flathead-dev/devo/gdb/ \  
testsuite/gdb.mi/basics.c",line_asm_insn=[  
{address="0x000107bc",func-name="main",offset="0",  
inst="save %sp, -112, %sp"}]},  
src_and_asm_line={line="32",  
file="/kwikemart/marge/ezannoni/flathead-dev/devo/gdb/ \  
testsuite/gdb.mi/basics.c",line_asm_insn=[  
{address="0x000107c0",func-name="main",offset="4",  
inst="mov 2, %o0"},  
{address="0x000107c4",func-name="main",offset="8",  
inst="sethi %hi(0x11800), %o2"}]}]  
(gdb)  
The -data-evaluate-expressioncommand  
Synopsis  
-data-evaluate-expression expr  
Evaluate expras an expression. The expression could contain an inferior function call.  
The function call will execute synchronously. If the expression contains spaces, it must  
be enclosed in double quotes.  
GDB command  
The corresponding GDB commands are 'print', 'output', and 'call'. In gdbtkonly,  
there is a corresponding 'gdb_eval' command.  
Example  
In the following example, the numbers that precede the commands are the tokens  
described in GDB/MICommand Syntax” (page 307). Notice how GDB/MIreturns the  
same tokens in its output.  
211-data-evaluate-expression A  
211^done,value="1"  
(gdb)  
311-data-evaluate-expression &A  
311^done,value="0xefffeb7c"  
(gdb)  
411-data-evaluate-expression A+3  
411^done,value="4"  
(gdb)  
511-data-evaluate-expression "A + 3"  
511^done,value="4"  
(gdb)  
322 The GDB/MIInterface  
The -data-list-changed-registersCommand  
Synopsis  
-data-list-changed-registers  
Display a list of the registers that have changed.  
GDB command  
GDB does not have a direct analog for this command; gdbtkhas the corresponding  
command 'gdb_changed_register_list'.  
Example  
On a PPC MBX board:  
(gdb)  
-exec-continue  
^running  
(gdb)  
*stopped,reason="breakpoint-hit",bkptno="1",frame={func="main",  
args=[],file="try.c",line="5"}  
(gdb)  
-data-list-changed-registers  
^done,changed-registers=["0","1","2","4","5","6","7","8","9",  
"10","11","13","14","15","16","17","18","19","20","21","22","23",  
"24","25","26","27","28","30","31","64","65","66","67","69"]  
(gdb)  
The -data-list-register-namescommand  
Synopsis  
-data-list-register-names [ ( regno )+ ]  
Show a list of register names for the current target. If no arguments are given, it shows  
a list of the names of all the registers. If integer numbers are given as arguments, it will  
print a list of the names of the registers corresponding to the arguments. To ensure  
consistency between a register name and its number, the output list may include empty  
register names.  
GDB command  
GDB does not have a command which corresponds to  
'-data-list-register-names'. In gdbtkthere is a corresponding command  
'gdb_regnames'.  
Example  
For the PPC MBX board:  
21.6 GDB/MI Data manipulation 323  
(gdb)  
-data-list-register-names  
^done,register-names=["r0","r1","r2","r3","r4","r5","r6","r7",  
"r8","r9","r10","r11","r12","r13","r14","r15","r16","r17","r18",  
"r19","r20","r21","r22","r23","r24","r25","r26","r27","r28","r29",  
"r30","r31","f0","f1","f2","f3","f4","f5","f6","f7","f8","f9",  
"f10","f11","f12","f13","f14","f15","f16","f17","f18","f19","f20",  
"f21","f22","f23","f24","f25","f26","f27","f28","f29","f30","f31",  
"", "pc","ps","cr","lr","ctr","xer"]  
(gdb)  
-data-list-register-names 1 2 3  
^done,register-names=["r1","r2","r3"]  
(gdb)  
The -data-list-register-valuescommand  
Synopsis  
-data-list-register-values fmt [ ( regno )*]  
Display the registers contents. fmtis the format according to which the registers'  
contents are to be returned, followed by an optional list of numbers specifying the  
registers to display. A missing list of numbers indicates that the contents of all the  
registers must be returned.  
Allowed formats for fmtare:  
x
o
t
d
r
N
Hexadecimal  
Octal  
Binary  
Decimal  
Raw  
Natural  
GDB command  
The corresponding GDB commands are 'info reg', 'info all-reg', and (in gdbtk)  
'gdb_fetch_registers'.  
Example  
For a PPC MBX board (note: line breaks are for readability only, they do not appear in  
the actual output):  
(gdb)  
-data-list-register-values r 64 65  
^done,register-values=[{number="64",value="0xfe00a300"},  
{number="65",value="0x00029002"}]  
(gdb)  
-data-list-register-values x  
324 The GDB/MIInterface  
^done,register-values=[{number="0",value="0xfe0043c8"},  
{number="1",value="0x3fff88"},{number="2",value="0xfffffffe"},  
{number="3",value="0x0"},{number="4",value="0xa"},  
{number="5",value="0x3fff68"},{number="6",value="0x3fff58"},  
{number="7",value="0xfe011e98"},{number="8",value="0x2"},  
{number="9",value="0xfa202820"},{number="10",value="0xfa202808"},  
{number="11",value="0x1"},{number="12",value="0x0"},  
{number="13",value="0x4544"},{number="14",value="0xffdfffff"},  
{number="15",value="0xffffffff"},{number="16",value="0xfffffeff"},  
{number="17",value="0xefffffed"},{number="18",value="0xfffffffe"},  
{number="19",value="0xffffffff"},{number="20",value="0xffffffff"},  
{number="21",value="0xffffffff"},{number="22",value="0xfffffff7"},  
{number="23",value="0xffffffff"},{number="24",value="0xffffffff"},  
{number="25",value="0xffffffff"},{number="26",value="0xfffffffb"},  
{number="27",value="0xffffffff"},{number="28",value="0xf7bfffff"},  
{number="29",value="0x0"},{number="30",value="0xfe010000"},  
{number="31",value="0x0"},{number="32",value="0x0"},  
{number="33",value="0x0"},{number="34",value="0x0"},  
{number="35",value="0x0"},{number="36",value="0x0"},  
{number="37",value="0x0"},{number="38",value="0x0"},  
{number="39",value="0x0"},{number="40",value="0x0"},  
{number="41",value="0x0"},{number="42",value="0x0"},  
{number="43",value="0x0"},{number="44",value="0x0"},  
{number="45",value="0x0"},{number="46",value="0x0"},  
{number="47",value="0x0"},{number="48",value="0x0"},  
{number="49",value="0x0"},{number="50",value="0x0"},  
{number="51",value="0x0"},{number="52",value="0x0"},  
{number="53",value="0x0"},{number="54",value="0x0"},  
{number="55",value="0x0"},{number="56",value="0x0"},  
{number="57",value="0x0"},{number="58",value="0x0"},  
{number="59",value="0x0"},{number="60",value="0x0"},  
{number="61",value="0x0"},{number="62",value="0x0"},  
{number="63",value="0x0"},{number="64",value="0xfe00a300"},  
{number="65",value="0x29002"},{number="66",value="0x202f04b5"},  
{number="67",value="0xfe0043b0"},{number="68",value="0xfe00b3e4"},  
{number="69",value="0x20002b03"}]  
(gdb)  
The -data-read-memorycommand  
Synopsis  
-data-read-memory [ -o byte-offset ]  
address word-format word-size  
nr-rows nr-cols [ aschar ]  
where:  
'address'  
An expression specifying the address of the first memory word  
to be read. Complex expressions containing embedded white  
space should be quoted using the C convention.  
21.6 GDB/MI Data manipulation 325  
'word-format'  
The format to be used to print the memory words. The notation  
is the same as for GDB printcommand (see“Output formats”  
'word-size'  
'nr-rows'  
'nr-cols'  
'aschar'  
The size of each memory word in bytes.  
The number of rows in the output table.  
The number of columns in the output table.  
If present, indicates that each row should include an ASCII dump.  
The value of ascharis used as a padding character when a byte  
is not a member of the printable ASCII character set (printable  
ASCII characters are those whose code is between 32 and 126,  
inclusively).  
'byte-offset'  
An offset to add to the addressbefore fetching memory.  
This command displays memory contents as a table of nr-rowsby nr-colswords,  
each word being word-sizebytes. In total, nr-rows * nr-cols * word-size  
bytesare read (returned as 'total-bytes'). Should less than the requested number  
of bytes be returned by the target, the missing words are identified using 'N/A'. The  
number of bytes read from the target is returned in 'nr-bytes' and the starting address  
used to read memory in 'addr'.  
The address of the next/previous row or page is available in 'next-row' and  
'prev-row', 'next-page' and 'prev-page'.  
GDB command  
The corresponding GDB command is 'x'. gdbtkhas 'gdb_get_mem' memory read  
command.  
Example  
Read six bytes of memory starting at bytes+6 but then offset by -6 bytes. Format as  
three rows of two columns. One byte per word. Display each word in hex.  
(gdb)  
9-data-read-memory -o -6 -- bytes+6 x 1 3 2  
9^done,addr="0x00001390",nr-bytes="6",total-bytes="6",  
next-row="0x00001396",prev-row="0x0000138e",next-page="0x00001396",  
prev-page="0x0000138a",memory=[  
{addr="0x00001390",data=["0x00","0x01"]},  
{addr="0x00001392",data=["0x02","0x03"]},  
{addr="0x00001394",data=["0x04","0x05"]}]  
(gdb)  
Read two bytes of memory starting at address shorts+ 64 and display as a single  
word formatted in decimal.  
(gdb)  
5-data-read-memory shorts+64 d 2 1 1  
5^done,addr="0x00001510",nr-bytes="2",total-bytes="2",  
326 The GDB/MIInterface  
next-row="0x00001512",prev-row="0x0000150e",  
next-page="0x00001512",prev-page="0x0000150e",memory=[  
{addr="0x00001510",data=["128"]}]  
(gdb)  
Read thirty two bytes of memory starting at bytes+16 and format as eight rows of  
four columns. Include a string encoding with 'x' used as the non-printable character.  
(gdb)  
4-data-read-memory bytes+16 x 1 8 4 x  
4^done,addr="0x000013a0",nr-bytes="32",total-bytes="32",  
next-row="0x000013c0",prev-row="0x0000139c",  
next-page="0x000013c0",prev-page="0x00001380",memory=[  
{addr="0x000013a0",data=["0x10","0x11","0x12","0x13"],ascii="xxxx"},  
{addr="0x000013a4",data=["0x14","0x15","0x16","0x17"],ascii="xxxx"},  
{addr="0x000013a8",data=["0x18","0x19","0x1a","0x1b"],ascii="xxxx"},  
{addr="0x000013ac",data=["0x1c","0x1d","0x1e","0x1f"],ascii="xxxx"},  
{addr="0x000013b0",data=["0x20","0x21","0x22","0x23"],ascii=" !\"#"},  
{addr="0x000013b4",data=["0x24","0x25","0x26","0x27"],ascii="$%&'"},  
{addr="0x000013b8",data=["0x28","0x29","0x2a","0x2b"],ascii="()*+"},  
{addr="0x000013bc",data=["0x2c","0x2d","0x2e","0x2f"],ascii=",-./"}]  
(gdb)  
The -display-deletecommand  
Synopsis  
-display-delete number  
Delete the display number.  
GDB command  
The corresponding GDB command is 'delete display'.  
Example  
N.A.  
The -display-disableCommand  
Synopsis  
-display-disable number  
Disable display number.  
GDB command  
The corresponding GDB command is 'disable display'.  
Example  
N.A.  
21.6 GDB/MI Data manipulation 327  
The -display-enablecommand  
Synopsis  
-display-enable number  
Enable display number.  
GDB command  
The corresponding GDB command is 'enable display'.  
Example  
N.A.  
The -display-insertCommand  
Synopsis  
-display-insert expression  
Display expressionevery time the program stops.  
GDB command  
The corresponding GDB command is 'display'.  
Example  
N.A.  
The -display-listcommand  
Synopsis  
-display-list  
List the displays. Do not show the current values.  
GDB command  
The corresponding GDB command is 'info display'.  
Example  
N.A.  
The -environment-cdcommand  
Synopsis  
-environment-cd pathdir  
328 The GDB/MIInterface  
Set the GDB working directory.  
GDB command  
The corresponding GDB command is 'cd'.  
Example  
(gdb)  
-environment-cd /kwikemart/marge/ezannoni/flathead-dev/devo/gdb  
^done  
(gdb)  
The -environment-directorycommand  
Synopsis  
-environment-directory pathdir  
Add directory pathdirto the beginning of search path for source files.  
GDB command  
The corresponding GDB command is 'dir'.  
Example  
(gdb)  
-environment-directory /kwikemart/marge/ezannoni/flathead-dev/devo/gdb  
^done  
(gdb)  
The -environment-pathcommand  
Synopsis  
-environment-path ( pathdir )+  
Add directories pathdirto beginning of search path for object files.  
GDB command  
The corresponding GDB command is 'path'.  
Example  
(gdb)  
-environment-path /kwikemart/marge/ezannoni/flathead-dev/ppc-eabi/gdb  
^done  
(gdb)  
21.6 GDB/MI Data manipulation 329  
The -environment-pwdcommand  
Synopsis  
-environment-pwd  
Show the current working directory.  
GDB command  
The corresponding GDB command is 'pwd'.  
Example  
(gdb)  
-environment-pwd  
~Working directory /kwikemart/marge/ezannoni/flathead-dev/devo/gdb.  
^done  
(gdb)  
21.7 GDB/MIprogram control  
Program termination As a result of execution, the inferior program can run to  
completion, if it does not encounter any breakpoints. In this case the output will include  
an exit code, if the program has exited exceptionally.  
Examples Program exited normally:  
(gdb)  
-exec-run  
^running  
(gdb)  
x = 55  
*stopped,reason="exited-normally"  
(gdb)  
Program exited exceptionally:  
(gdb)  
-exec-run  
^running  
(gdb)  
x = 55  
*stopped,reason="exited",exit-code="01"  
(gdb)  
Another way the program can terminate is if it receives a signal such as SIGINT. In this  
case, GDB/MIdisplays this:  
(gdb)  
*stopped,reason="exited-signalled",signal-name="SIGINT",  
signal-meaning="Interrupt"  
330 The GDB/MIInterface  
 
The -exec-abortcommand  
Synopsis  
-exec-abort  
Kill the inferior running program.  
GDB command  
The corresponding GDB command is 'kill'.  
Example  
N.A.  
The -exec-argumentscommand  
Synopsis  
-exec-arguments args  
Set the inferior program arguments, to be used in the next '-exec-run'.  
GDB command  
The corresponding GDB command is 'set args'.  
Example  
Do not have one around.  
The -exec-continuecommand  
Synopsis  
-exec-continue  
Asynchronous command. Resumes the execution of the inferior program until a  
breakpoint is encountered, or until the inferior exits.  
GDB command  
The corresponding GDB is 'continue'.  
Example  
-exec-continue  
^running  
(gdb)  
@Hello world  
*stopped,reason="breakpoint-hit",bkptno="2",thread-id="1",frame=addr="0x000029d8",  
func="foo",args=[],file="hello.c",line="16"file="hello.c",line="13"}  
(gdb)  
21.7 GDB/MI program control 331  
The -exec-finishcommand  
Synopsis  
-exec-finish  
Asynchronous command. Resumes the execution of the inferior program until the  
current function is exited. Displays the results returned by the function.  
GDB command  
The corresponding GDB command is 'finish'.  
Example  
Function returning void.  
-exec-finish  
^running  
(gdb)  
@hello from foo  
*stopped,reason="function-finished",thread-id="1",frame=addr="0x000029ec",  
func="main",args=[],file="hello.c",line="7file="hello.c",line="7"}  
(gdb)  
Function returning other than void. The name of the internal GDB variable storing  
the result is printed, together with the value itself.  
-exec-finish  
^running  
(gdb)  
*stopped,reason="function-finished",thread-id="1",  
frame=addr="0x000107b0",func="foo",  
args=[name="a"],name="b",  
file="recursive2.c",line="14"},  
gdb-result-var="$1",return-value="0"  
(gdb)  
The -exec-interruptcommand  
Synopsis  
-exec-interrupt  
Asynchronous command. Interrupts the background execution of the target. Note how  
the token associated with the stop message is the one for the execution command that  
has been interrupted. The token for the interrupt itself only appears in the '^done'  
output. If the user is trying to interrupt a non-running program, an error message will  
be printed.  
GDB command  
The corresponding GDB command is 'interrupt'.  
332 The GDB/MIInterface  
Example  
(gdb)  
111-exec-continue  
111^running  
(gdb)  
222-exec-interrupt  
222^done  
(gdb)  
111*stopped,signal-name="SIGINT",signal-meaning="Interrupt",  
frame={addr="0x00010140",func="foo",args=[],file="try.c",line="13"}  
(gdb)  
(gdb)  
-exec-interrupt  
^error,msg="mi_cmd_exec_interrupt: Inferior not executing."  
(gdb)  
The -exec-nextcommand  
Synopsis  
-exec-next  
Asynchronous command. Resumes execution of the inferior program, stopping when  
the beginning of the next source line is reached.  
GDB command  
The corresponding GDB command is 'next'.  
Example  
-exec-next  
^running  
(gdb)  
*stopped,reason="end-stepping-range",thread-id="1",frame=addr="0x00002a10",  
func="main",args=[],file="hello.c",line="24"(gdb)  
The -exec-next-instructioncommand  
Synopsis  
-exec-next-instruction  
Asynchronous command. Executes one machine instruction. If the instruction is a  
function, call continues until the function returns. If the program stops at an instruction  
in the middle of a source line, the address will be printed as well.  
GDB command  
The corresponding GDB command is 'nexti'.  
21.7 GDB/MI program control 333  
Example  
(gdb)  
-exec-next-instruction  
^running  
(gdb)  
*stopped,reason="end-stepping-range",thread-  
id="1",frame=addr="0x00002a14",func="main",args=[],file="hello.c",line="24"  
(gdb)  
The -exec-returncommand  
Synopsis  
-exec-return  
Makes current function return immediately. Does not execute the inferior. Displays  
the new current frame.  
GDB command  
The corresponding GDB command is return'.  
Example  
(gdb)  
-break-insert call1  
^done,bkpt=number="1",type="breakpoint",disp="keep",enabled="y",addr="0x000029ac",  
func="call1",file="hello.c",line="9",times="0"  
(gdb)  
-exec-run  
^running  
(gdb)  
~"3"  
*stopped,reason="breakpoint-hit",bkptno="1",thread-id="1",frame=addr="0x000029ac",  
func="call1",args=[name="a"],file="hello.c",line="9"  
(gdb)  
-exec-return  
~"2"  
~"3"  
^done,frame=level="0 ",addr="0x000029e8",func="call",args=[name="a",name="b"],  
file="hello.c",line="17"  
(gdb)  
The -exec-runcommand  
Synopsis  
-exec-run  
Asynchronous command. Starts execution of the inferior from the beginning. The  
inferior executes until either a breakpoint is encountered or the program exits.  
GDB command  
The corresponding GDB command is 'run'.  
334 The GDB/MIInterface  
Example  
(gdb)  
-break-insert main  
^done,bkpt=number="1",type="breakpoint",disp="keep",enabled="y",addr="0x00002a08",  
func="main",file="hello.c",line="23",times="0"  
(gdb)  
-exec-run  
^running  
(gdb)  
*stopped,reason="breakpoint-hit",bkptno="1",thread-id="1",frame=addr="0x00002a08",  
func="main",args=[],file="hello.c",line="23"  
(gdb)  
The -exec-show-arguments command  
Synopsis  
-exec-show-arguments  
Prints the arguments of the program.  
GDB command  
The corresponding GDB command is 'show args'.  
Example  
N.A.  
The -exec-stepCommand  
Synopsis  
-exec-step  
Asynchronous command. Resumes execution of the inferior program, stopping when  
the beginning of the next source line is reached, if the next source line is not a function  
call. If it is, stop at the first instruction of the called function.  
GDB command  
The corresponding GDB command is 'step'.  
Example  
Stepping into a function:  
-exec-step  
^running  
(gdb)  
~"2"  
~"3"  
*stopped,reason="end-stepping-range",thread-id="1",frame=addr="0x000029d0",  
func="call",args=[name="a",name="b"],file="hello.c",line="15"  
(gdb)  
21.7 GDB/MI program control 335  
Regular stepping:  
-exec-step  
^running  
(gdb)  
~"2"  
~"3"  
*stopped,reason="end-stepping-range",thread-id="1",frame=addr="0x000029d8",  
func="call",args=[name="a",name="b"],file="hello.c",line="16"  
(gdb)  
The -exec-step-instructioncommand  
Synopsis  
-exec-step-instruction  
Asynchronous command. Resumes the inferior which executes one machine instruction.  
The output, once GDB has stopped, will vary depending on whether we have stopped  
in the middle of a source line or not. In the former case, the address at which the  
program stopped will be printed as well.  
GDB command  
The corresponding GDB command is 'stepi'.  
Example  
(gdb)  
-exec-step-instruction  
^running  
(gdb)  
~"2"  
~"3"  
*stopped,reason="end-stepping-range",thread-id="1",frame=addr="0x000029dc",  
func="call",args=[name="a",name="b"],file="hello.c",line="16"  
(gdb)  
-exec-step-instruction  
^running  
(gdb)  
*stopped,reason="end-stepping-range",  
frame={addr="0x000100f4",func="foo",args=[],file="try.c",line="10"}  
(gdb)  
The -exec-untilcommand  
Synopsis  
-exec-until [ location ]  
Asynchronous command. Executes the inferior until the locationspecified in the  
argument is reached. If there is no argument, the inferior executes until a source line  
336 The GDB/MIInterface  
greater than the current one is reached. The reason for stopping in this case will be  
'location-reached'.  
GDB command  
The corresponding GDB command is 'until'.  
Example  
(gdb)  
-exec-until recursive2.c:6  
^running  
(gdb)  
x = 55  
*stopped,reason="location-reached",thread-id="1",frame=addr="0x00002a24",func="main",args=[],  
file="recursive2.c",line="6"  
(gdb)  
The -file-exec-and-symbolscommand  
Synopsis  
-file-exec-and-symbols file  
Specify the executable file to be debugged. This file is the one from which the symbol  
table is also read. If no file is specified, the command clears the executable and symbol  
information. If breakpoints are set when using this command with no arguments, GDB  
will produce error messages. Otherwise, no output is produced, except a completion  
notification.  
GDB command  
The corresponding GDB command is 'file'.  
Example  
(gdb)  
-file-exec-and-symbols /kwikemart/marge/ezannoni/TRUNK/mbx/hello.mbx  
^done  
(gdb)  
The -file-exec-filecommand  
Synopsis  
-file-exec-file file  
Specify the executable file to be debugged. Unlike '-file-exec-and-symbols', the  
symbol table is not read from this file. If used without argument, GDB clears the  
information about the executable file. No output is produced, except a completion  
notification.  
21.7 GDB/MI program control 337  
GDB command  
The corresponding GDB command is 'exec-file'.  
Example  
(gdb)  
-file-exec-file /kwikemart/marge/ezannoni/TRUNK/mbx/hello.mbx  
^done  
(gdb)  
The -file-list-exec-sectionscommand  
Synopsis  
-file-list-exec-sections  
List the sections of the current executable file.  
GDB command  
The GDB command 'info file' shows, among the rest, the same information as this  
command. gdbtkhas a corresponding command 'gdb_load_info'.  
Example  
N.A.  
The -file-list-exec-source-filescommand  
Synopsis  
-file-list-exec-source-files  
List the source files for the current executable.  
GDB command  
There is no GDB command which directly corresponds to this one. gdbtkhas an  
analogous command 'gdb_listfiles'.  
Example  
N.A.  
The -file-list-shared-librariescommand  
Synopsis  
-file-list-shared-libraries  
List the shared libraries in the program.  
338 The GDB/MIInterface  
GDB command  
The corresponding GDB command is info shared.  
Example  
N.A.  
The -file-list-symbol-filescommand  
Synopsis  
-file-list-symbol-files  
List symbol files.  
GDB command  
The corresponding GDB command is 'info file' (part of it).  
Example  
N.A.  
The -file-symbol-filecommand  
Synopsis  
-file-symbol-file file  
Read symbol table info from the specified fileargument. When used without  
arguments, clears GDB's symbol table info. No output is produced, except for a  
completion notification.  
GDB command  
The corresponding GDB command is 'symbol-file'.  
Example  
(gdb)  
-file-symbol-file /kwikemart/marge/ezannoni/TRUNK/mbx/hello.mbx  
^done  
(gdb)  
21.8 Miscellaneous GDB commands in GDB/MI  
The -gdb-exitcommand  
Synopsis  
-gdb-exit  
21.8 Miscellaneous GDB commands in GDB/MI 339  
 
Exit GDB immediately.  
GDB command  
Approximately corresponds to 'quit'.  
Example  
(gdb)  
-gdb-exit  
The -gdb-setcommand  
Synopsis  
-gdb-set  
Set an internal GDB variable.  
GDB command  
The corresponding GDB command is 'set'.  
Example  
(gdb)  
-gdb-set $foo=3  
^done  
(gdb)  
The -gdb-showcommand  
Synopsis  
-gdb-show  
Show the current value of a GDB variable.  
GDB command  
The corresponding GDB command is 'show'.  
Example  
(gdb)  
-gdb-show annotate  
^done,value="0"  
(gdb)  
The -gdb-versioncommand  
Synopsis  
-gdb-version  
340 The GDB/MIInterface  
Show version information for GDB. Used mostly in testing.  
GDB command  
The corresponding GDB command is 'show version'.  
Example  
(gdb)  
-gdb-version  
~GNU gdb 5.2.1  
~Copyright 2000 Free Software Foundation, Inc.  
~GDB is free software, covered by the GNU General Public License, and  
~you are welcome to change it and/or distribute copies of it under  
~ certain conditions.  
~Type "show copying" to see the conditions.  
~There is absolutely no warranty for GDB. Type "show warranty" for  
~ details.  
~This GDB was configured as  
"--host=sparc-sun-solaris2.5.1 --target=ppc-eabi".  
^done  
(gdb)  
21.9 GDB/MIStack Manipulation Commands  
The -stack-info-framecommand  
Synopsis  
-stack-info-frame  
Get info on the current frame.  
GDB command  
The corresponding GDB command is 'info frame' or 'frame' (without arguments).  
Example  
N.A.  
The -stack-info-depthcommand  
Synopsis  
-stack-info-depth [ max-depth ]  
Return the depth of the stack. If the integer argument max-depthis specified, do not  
count beyond max-depthframes.  
GDB command  
There is no equivalent GDB command.  
21.9 GDB/MI Stack Manipulation Commands 341  
 
Example  
For a stack with frame levels 0 through 11:  
(gdb)  
-stack-info-depth  
^done,depth="12"  
(gdb)  
-stack-info-depth 4  
^done,depth="4"  
(gdb)  
-stack-info-depth 12  
^done,depth="12"  
(gdb)  
-stack-info-depth 11  
^done,depth="11"  
(gdb)  
-stack-info-depth 13  
^done,depth="12"  
(gdb)  
The -stack-list-arguments command  
Synopsis  
-stack-list-arguments show-values  
[ low-frame high-frame ]  
Display a list of the arguments for the frames between low-frameand high-frame  
(inclusive). If low-frameand high-frameare not provided, list the arguments for  
the whole call stack.  
The show-valuesargument must have a value of 0 or 1. A value of 0 means that only  
the names of the arguments are listed, a value of 1 means that both names and values  
of the arguments are printed.  
GDB command  
GDB does not have an equivalent command. gdbtkhas a 'gdb_get_args' command  
which partially overlaps with the functionality of '-stack-list-arguments'.  
Example  
(gdb)  
-stack-list-frames  
^done,  
stack=[  
frame={level="0 ",addr="0x00010734",func="callee4",  
file="../../../devo/gdb/testsuite/gdb.mi/basics.c",line="8"},  
frame={level="1 ",addr="0x0001076c",func="callee3",  
file="../../../devo/gdb/testsuite/gdb.mi/basics.c",line="17"},  
frame={level="2 ",addr="0x0001078c",func="callee2",  
file="../../../devo/gdb/testsuite/gdb.mi/basics.c",line="22"},  
frame={level="3 ",addr="0x000107b4",func="callee1",  
342 The GDB/MIInterface  
file="../../../devo/gdb/testsuite/gdb.mi/basics.c",line="27"},  
frame={level="4 ",addr="0x000107e0",func="main",  
file="../../../devo/gdb/testsuite/gdb.mi/basics.c",line="32"}]  
(gdb)  
-stack-list-arguments 0  
^done,  
stack-args=[  
frame={level="0",args=[]},  
frame={level="1",args=[name="strarg"]},  
frame={level="2",args=[name="intarg",name="strarg"]},  
frame={level="3",args=[name="intarg",name="strarg",name="fltarg"]},  
frame={level="4",args=[]}]  
(gdb)  
-stack-list-arguments 1  
^done,  
stack-args=[  
frame={level="0",args=[]},  
frame={level="1",  
args=[{name="strarg",value="0x11940 \"A string argument.\""}]},  
frame={level="2",args=[  
{name="intarg",value="2"},  
{name="strarg",value="0x11940 \"A string argument.\""}]},  
{frame={level="3",args=[  
{name="intarg",value="2"},  
{name="strarg",value="0x11940 \"A string argument.\""},  
{name="fltarg",value="3.5"}]},  
frame={level="4",args=[]}]  
(gdb)  
-stack-list-arguments 0 2 2  
^done,stack-args=[frame={level="2",args=[name="intarg",name="strarg"]}]  
(gdb)  
-stack-list-arguments 1 2 2  
^done,stack-args=[frame={level="2",  
args=[{name="intarg",value="2"},  
{name="strarg",value="0x11940 \"A string argument.\""}]}]  
(gdb)  
The -stack-list-framescommand  
Synopsis  
-stack-list-frames [ low-frame high-frame ]  
List the frames currently on the stack. For each frame it displays the following info:  
'level'  
The frame number, 0 being the topmost frame, that is the innermost  
function.  
'addr'  
'func'  
'file'  
'line'  
The $pc value for that frame.  
Function name.  
File name of the source file where the function lives.  
Line number corresponding to the $pc.  
If invoked without arguments, this command prints a backtrace for the whole stack.  
If given two integer arguments, it shows the frames whose levels are between the two  
21.9 GDB/MI Stack Manipulation Commands 343  
arguments (inclusive). If the two arguments are equal, it shows the single frame at the  
corresponding level.  
GDB command  
The corresponding GDB commands are 'backtrace' and 'where'.  
Example  
Full stack backtrace:  
(gdb)  
-stack-list-frames  
^done,stack=  
[frame={level="0 ",addr="0x0001076c",func="foo",  
file="recursive2.c",line="11"},  
frame={level="1 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"},  
frame={level="2 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"},  
frame={level="3 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"},  
frame={level="4 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"},  
frame={level="5 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"},  
frame={level="6 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"},  
frame={level="7 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"},  
frame={level="8 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"},  
frame={level="9 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"},  
frame={level="10",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"},  
frame={level="11",addr="0x00010738",func="main",  
file="recursive2.c",line="4"}]  
(gdb)  
Show frames between low_frameand high_frame:  
(gdb)  
-stack-list-frames 3 5  
^done,stack=  
[frame={level="3 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"},  
frame={level="4 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"},  
frame={level="5 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"}]  
(gdb)  
Show a single frame:  
344 The GDB/MIInterface  
(gdb)  
-stack-list-frames 3 3  
^done,stack=  
[frame={level="3 ",addr="0x000107a4",func="foo",  
file="recursive2.c",line="14"}]  
(gdb)  
The -stack-list-localscommand  
Synopsis  
-stack-list-locals print-values  
Display the local variable names for the current frame. With an argument of 0 prints  
only the names of the variables, with argument of 1 prints also their values.  
GDB command  
'info locals' in GDB, 'gdb_get_locals' in gdbtk.  
Example  
(gdb)  
-stack-list-locals 0  
^done,locals=[name="A",name="B",name="C"]  
(gdb)  
-stack-list-locals 1  
^done,locals=[{name="A",value="1"},{name="B",value="2"},  
{name="C",value="3"}]  
(gdb)  
The -stack-select-framecommand  
Synopsis  
-stack-select-frame framenum  
Change the current frame. Select a different frame framenumon the stack.  
GDB command  
The corresponding GDB commands are 'frame', 'up', 'down', 'select-frame',  
'up-silent', and 'down-silent'.  
Example  
(gdb)  
-stack-select-frame 2  
^done  
(gdb)  
21.9 GDB/MI Stack Manipulation Commands 345  
21.10 GDB/MISymbol query commands  
The -symbol-info-addresscommand  
Synopsis  
-symbol-info-address symbol  
Describe where symbolis stored.  
GDB command  
The corresponding GDB command is 'info address'.  
Example  
N.A.  
The -symbol-info-filecommand  
Synopsis  
-symbol-info-file  
Show the file for the symbol.  
GDB command  
There is no equivalent GDB command. gdbtkhas 'gdb_find_file'.  
Example  
N.A.  
The -symbol-info-functioncommand  
Synopsis  
-symbol-info-function  
Show which function the symbol lives in.  
GDB command  
'gdb_get_function' in gdbtk.  
Example  
N.A.  
346 The GDB/MIInterface  
 
The -symbol-info-linecommand  
Synopsis  
-symbol-info-line  
Show the core addresses of the code for a source line.  
GDB command  
The corresponding GDB command is 'info line'. gdbtkhas the 'gdb_get_line'  
and 'gdb_get_file' commands.  
Example  
N.A.  
The -symbol-info-symbolcommand  
Synopsis  
-symbol-info-symbol addr  
Describe what symbol is at location addr.  
GDB command  
The corresponding GDB command is 'info symbol'.  
Example  
N.A.  
The -symbol-list-functionscommand  
Synopsis  
-symbol-list-functions  
List the functions in the executable.  
GDB command  
'info functions' in GDB, 'gdb_listfunc' and 'gdb_search' in gdbtk.  
Example  
N.A.  
The -symbol-list-typescommand  
Synopsis  
-symbol-list-types  
21.10 GDB/MI Symbol query commands 347  
List all the type names.  
GDB command  
The corresponding commands are 'info types' in GDB, 'gdb_search' in gdbtk.  
Example  
N.A.  
The -symbol-list-variablescommand  
Synopsis  
-symbol-list-variables  
List all the global and static variable names.  
GDB command  
'info variables' in GDB, 'gdb_search' in gdbtk.  
Example  
N.A.  
The -symbol-locatecommand  
Synopsis  
-symbol-locate  
GDB command  
'gdb_loc' in gdbtk.  
Example  
N.A.  
The -symbol-typecommand  
Synopsis  
-symbol-type variable  
Show type of variable.  
GDB command  
The corresponding GDB command is 'ptype', gdbtkhas 'gdb_obj_variable'.  
348 The GDB/MIInterface  
Example  
N.A.  
21.11 GDB/MITarget Manipulation Commands  
The -target-attachcommand  
Synopsis  
-target-attach pid | file  
Attach to a process pidor a file fileoutside of GDB.  
GDB command  
The corresponding GDB command is 'attach'.  
Example  
N.A.  
The -target-compare-sectionscommand  
Synopsis  
-target-compare-sections [ section ]  
Compare data of section sectionon target to the exec file. Without the argument, all  
sections are compared.  
GDB command  
The GDB equivalent is 'compare-sections'.  
Example  
N.A.  
The -target-detachcommand  
Synopsis  
-target-detach  
Disconnect from the remote target. There is no output.  
GDB command  
The corresponding GDB command is 'detach'.  
21.11 GDB/MI Target Manipulation Commands 349  
 
Example  
(gdb)  
-target-detach  
^done  
(gdb)  
The -target-downloadcommand  
Synopsis  
-target-download  
Loads the executable onto the remote target. It prints out an update message every half  
second, which includes the fields:  
'section'  
The name of the section.  
'section-sent'  
'section-size'  
'total-sent'  
The size of what has been sent so far for that section.  
The size of the section.  
The total size of what was sent so far (the current and the  
previous sections).  
'total-size'  
The size of the overall executable to download.  
Each message is sent as status record (see GDB/MIOutput syntax” (page 308).  
In addition, it prints the name and size of the sections, as they are downloaded. These  
messages include the following fields:  
'section'  
The name of the section.  
'section-size'  
'total-size'  
The size of the section.  
The size of the overall executable to download.  
At the end, a summary is printed.  
GDB command  
The corresponding GDB command is 'load'.  
Example  
NOTE: Each status message appears on a single line. Here the messages have been  
broken down so that they can fit onto a page.  
(gdb)  
-target-download  
+download,{section=".text",section-size="6668",total-size="9880"}  
+download,{section=".text",section-sent="512",section-size="6668",  
total-sent="512",total-size="9880"}  
+download,{section=".text",section-sent="1024",section-size="6668",  
total-sent="1024",total-size="9880"}  
350 The GDB/MIInterface  
+download,{section=".text",section-sent="1536",section-size="6668",  
total-sent="1536",total-size="9880"}  
+download,{section=".text",section-sent="2048",section-size="6668",  
total-sent="2048",total-size="9880"}  
+download,{section=".text",section-sent="2560",section-size="6668",  
total-sent="2560",total-size="9880"}  
+download,{section=".text",section-sent="3072",section-size="6668",  
total-sent="3072",total-size="9880"}  
+download,{section=".text",section-sent="3584",section-size="6668",  
total-sent="3584",total-size="9880"}  
+download,{section=".text",section-sent="4096",section-size="6668",  
total-sent="4096",total-size="9880"}  
+download,{section=".text",section-sent="4608",section-size="6668",  
total-sent="4608",total-size="9880"}  
+download,{section=".text",section-sent="5120",section-size="6668",  
total-sent="5120",total-size="9880"}  
+download,{section=".text",section-sent="5632",section-size="6668",  
total-sent="5632",total-size="9880"}  
+download,{section=".text",section-sent="6144",section-size="6668",  
total-sent="6144",total-size="9880"}  
+download,{section=".text",section-sent="6656",section-size="6668",  
total-sent="6656",total-size="9880"}  
+download,{section=".init",section-size="28",total-size="9880"}  
+download,{section=".fini",section-size="28",total-size="9880"}  
+download,{section=".data",section-size="3156",total-size="9880"}  
+download,{section=".data",section-sent="512",section-size="3156",  
total-sent="7236",total-size="9880"}  
+download,{section=".data",section-sent="1024",section-size="3156",  
total-sent="7748",total-size="9880"}  
+download,{section=".data",section-sent="1536",section-size="3156",  
total-sent="8260",total-size="9880"}  
+download,{section=".data",section-sent="2048",section-size="3156",  
total-sent="8772",total-size="9880"}  
+download,{section=".data",section-sent="2560",section-size="3156",  
total-sent="9284",total-size="9880"}  
+download,{section=".data",section-sent="3072",section-size="3156",  
total-sent="9796",total-size="9880"}  
^done,address="0x10004",load-size="9880",transfer-rate="6586",  
write-rate="429"  
(gdb)  
The -target-exec-statuscommand  
Synopsis  
-target-exec-status  
Provide information on the state of the target (whether it is running or not, for instance).  
GDB command  
There is no equivalent GDB command.  
21.11 GDB/MI Target Manipulation Commands 351  
Example  
N.A.  
The -target-list-available-targetscommand  
Synopsis  
-target-list-available-targets  
List the possible targets to connect to.  
GDB command  
The corresponding GDB command is 'help target'.  
Example  
N.A.  
The -target-list-current-targetscommand  
Synopsis  
-target-list-current-targets  
Describe the current target.  
GDB command  
The corresponding information is printed by 'info file' (among other things).  
Example  
N.A.  
The -target-list-parameterscommand  
Synopsis  
-target-list-parameters  
GDB command  
No equivalent.  
Example  
N.A.  
352 The GDB/MIInterface  
The -target-selectcommand  
Synopsis  
-target-select type parameters ...  
Connect GDB to the remote target. This command takes two args:  
'type'  
The type of target, for instance 'async', 'remote', and so on.  
'parameters'  
Device names, host names and the like. See “Commands for  
managing targets” (page 133), for more details.  
The output is a connection notification, followed by the address at which the target  
program is, in the following form:  
^connected,addr="address",func="function name",  
args=[arg list]  
GDB command  
The corresponding GDB command is 'target'.  
Example  
(gdb)  
-target-select async /dev/ttya  
^connected,addr="0xfe00a300",func="??",args=[]  
(gdb)  
21.12 GDB/MIthread commands  
The -thread-infocommand  
Synopsis  
-thread-info  
GDB command  
No equivalent.  
Example  
N.A.  
The -thread-list-all-threadscommand  
Synopsis  
-thread-list-all-threads  
21.12 GDB/MI thread commands 353  
 
GDB command  
The equivalent GDB command is 'info threads'.  
Example  
N.A.  
The -thread-list-idscommand  
Synopsis  
-thread-list-ids  
Produces a list of the currently known GDB thread ids. At the end of the list it also  
prints the total number of such threads.  
GDB command  
Part of 'info threads' supplies the same information.  
Example  
No threads present, besides the main process:  
(gdb)  
-thread-list-ids  
^done,thread-ids={},number-of-threads="0"  
(gdb)  
Several threads:  
(gdb)  
-thread-list-ids  
^done,thread-ids={thread-id="3",thread-id="2",thread-id="1"},  
number-of-threads="3"  
(gdb)  
The -thread-selectcommand  
Synopsis  
-thread-select threadnum  
Make threadnumthe current thread. It prints the number of the new current thread,  
and the topmost frame for that thread.  
GDB command  
The corresponding GDB command is 'thread'.  
Example  
(gdb)  
-exec-next  
354 The GDB/MIInterface  
^running  
(gdb)  
~"0x7f7f0aec"  
*stopped,reason="end-stepping-range",thread-id="2",frame=addr="0x00002ca4",func  
="printme",args=[name="ip"],file="multithread.c",line="9"  
(gdb)  
-thread-list-ids  
^done,thread-ids=thread-id="2",thread-id="1",number-of-threads="2"  
(gdb)  
-thread-select 1  
^done,new-thread-id="1",frame=level="0 ",addr="0x7ad47d70",func="_lwp_create","  
+0x10",args=[],from="/usr/lib/libpthread.1"  
(gdb)  
21.13 GDB/MItracepoint commands  
The tracepoint commands are not yet implemented.  
21.14 GDB/MIvariable objects  
Motivation for variable objects in GDB/MI For the implementation of a variable  
debugger window (locals, watched expressions, and so on.), we are proposing the  
adaptation of the existing code used by Insight.  
The two main reasons for that are:  
1. It has been proven in practice (it is already on its second generation).  
2. It will shorten development time (needless to say how important it is now).  
The original interface was designed to be used by Tcl code, so it was slightly changed  
so it could be used through GDB/MI. This section describes the GDB/MIoperations that  
will be available and gives some hints about their use.  
NOTE: In addition to the set of operations described here, we expect the GUI  
implementation of a variable window to require, at least, the following operations:  
-gdb-show output-radix  
-stack-list-arguments  
-stack-list-locals  
-stack-select-frame  
Introduction to variable objects in GDB/MI The basic idea behind variable objects is  
the creation of a named object to represent a variable, an expression, a memory location  
or even a CPU register. For each object created, a set of operations is available for  
examining or changing its properties.  
Furthermore, complex data types, such as C structures, are represented in a tree format.  
For instance, the structtype variable is the root and the children will represent the  
struct members. If a child is itself of a complex type, it will also have children of its  
own. Appropriate language differences are handled for C, C++ and Java.  
When returning the actual values of the objects, this facility allows for the individual  
selection of the display format used in the result creation. It can be chosen among:  
21.13 GDB/MI tracepoint commands 355  
   
binary, decimal, hexadecimal, octal, and natural. Natural refers to a default format  
automatically chosen based on the variable type (like decimal for an int, hex for  
pointers, and so on.).  
The following is the complete set of GDB/MIoperations defined to access this  
functionality:  
Table 21-1 GDB/MIOperations  
Operation  
Description  
-var-create  
create a variable object  
-var-delete  
delete the variable object and its children  
set the display format of this variable  
show the display format of this variable  
tells how many children this object has  
return a list of the object children  
show the type of this variable object  
print what this variable object represents  
is this variable editable? does it exist here?  
get the value of this variable  
-var-set-format  
-var-show-format  
-var-info-num-children  
-var-list-children  
-var-info-type  
-var-info-expression  
-var-show-attributes  
-var-evaluate-expression  
-var-assign  
set the value of this variable  
-var-update  
update the variable and its children  
In the next subsection we describe each operation in detail and suggest how it can be  
used.  
Description and use of operations on variable objects  
The -var-createcommand  
Synopsis  
-var-create {name | "-"}  
{frame-addr | "*"} expression  
This operation creates a variable object, which allows the monitoring of a variable, the  
result of an expression, a memory cell or a CPU register.  
The nameparameter is the string by which the object can be referenced. It must be  
unique. If '-' is specified, the varobj system will generate a string “varNNNNNN”  
automatically. It will be unique provided that one does not specify nameon that format.  
The command fails if a duplicate name is found.  
356 The GDB/MIInterface  
 
The frame under which the expression should be evaluated can be specified by  
frame-addr. A '*' indicates that the current frame should be used.  
expressionis any expression valid on the current language set (must not begin with  
a '*'), or one of the following:  
'*addr', where addris the address of a memory cell  
'*addr-addr' ― a memory address range (TBD)  
'$regname' ― a CPU register name  
Result  
This operation returns the name, number of children and the type of the object created.  
Type is returned as a string as the ones generated by the GDB CLI:  
name="name",numchild="N",type="type"  
The -var-deletecommand  
Synopsis  
-var-delete name  
Deletes a previously created variable object and all of its children.  
Returns an error if the object nameis not found.  
The -var-set-formatcommand  
Synopsis  
-var-set-format name format-spec  
Sets the output format for the value of the object nameto be format-spec.  
The syntax for the format-specis as follows:  
format-spec  
{binary | decimal | hexadecimal | octal | natural}  
The -var-show-formatcommand  
Synopsis  
-var-show-format name  
Returns the format used to display the value of the object name.  
format  
format-spec  
The -var-info-num-childrencommand  
Synopsis  
-var-info-num-children name  
21.14 GDB/MI variable objects 357  
Returns the number of children of a variable object name:  
numchild=n  
The -var-list-childrencommand  
Synopsis  
-var-list-children name  
Returns a list of the children of the specified variable object:  
numchild=n,children={{name=name,  
numchild=n,type=type},(repeats N times)}  
The -var-info-typecommand  
Synopsis  
-var-info-type name  
Returns the type of the specified variable name. The type is returned as a string in the  
same format as it is output by the GDB CLI:  
type=typename  
The -var-info-expressioncommand  
Synopsis  
-var-info-expression name  
Returns what is represented by the variable object name:  
lang=lang-spec,exp=expression  
where lang-specis {"C" | "C++" | "Java"}.  
The -var-show-attributescommand  
Synopsis  
-var-show-attributes name  
List attributes of the specified variable object name:  
status=attr [ ( ,attr )* ]  
where attris { { editable | noneditable } | TBD }.  
The -var-evaluate-expressioncommand  
Synopsis  
-var-evaluate-expression name  
Evaluates the expression that is represented by the specified variable object and returns  
its value as a string in the current format specified for the object:  
358 The GDB/MIInterface  
value=value  
The -var-assigncommand  
Synopsis  
-var-assign name expression  
Assigns the value of expressionto the variable object specified by name. The object  
must be 'editable'.  
The -var-updatecommand  
Synopsis  
-var-update {name | "*"}  
Update the value of the variable object nameby evaluating its expression after fetching  
all the new values from memory or registers. A '*' causes all existing variable objects  
to be updated.  
21.14 GDB/MI variable objects 359  
360  
22 Reporting Bugs in GDB  
Your bug reports play an essential role in making GDB reliable.  
Reporting a bug may help you by bringing a solution to your problem, or it may not.  
But in any case the principal function of a bug report is to help the entire community  
by making the next version of GDB work better. Bug reports are your contribution to  
the maintenance of GDB.  
In order for a bug report to serve its purpose, you must include the information that  
enables us to x the bug.  
22.1 Have you found a bug?  
If you are not sure whether you have found a bug, here are some guidelines:  
If the debugger gets a fatal signal, for any input whatever, that is a GDB bug.  
Reliable debuggers never crash.  
If GDB produces an error message for valid input, that is a bug. (Note that if you're  
cross debugging, the problem may also be somewhere in the connection to the  
target.)  
If GDB does not produce an error message for invalid input, that is a bug. However,  
you should note that your idea of \invalid input" might be our idea of \an  
extension" or \support for traditional practice".  
If you are an experienced user of debugging tools, your suggestions for  
improvement of GDB are welcome in any case.  
22.2 How to report bugs  
If you obtained GDB (Hewlett-Packard Wildebeest (based on GDB 5.0-hpwdb-  
20000516)) as part of your HP ANSI C, HP ANSI C++, or HP Fortran compiler kit, report  
problems to your HP Support Representative.  
If you obtained GDB (Hewlett-Packard Wildebeest (based on GDB 5.0-hpwdb-  
20000516)) from the Hewlett-Packard Web site, report problems to your HP Support  
Representative. Support is covered under the support contract for your HP compiler.  
The fundamental principle of reporting bugs usefully is this: “report all the facts”. If  
you are not sure whether to state a fact or leave it out, state it!  
Often people omit facts because they think they know what causes the problem and  
assume that some details do not matter. Thus, you might assume that the name of the  
variable you use in an example does not matter. Well, probably it does not, but one  
cannot be sure. Perhaps the bug is a stray memory reference which happens to fetch  
from the location where that name is stored in memory; perhaps, if the name were  
different, the contents of that location would fool the debugger into doing the right  
thing despite the bug. Play it safe and give a specific, complete example. That is the  
easiest thing for you to do, and the most helpful.  
22.1 Have you found a bug? 361  
     
Keep in mind that the purpose of a bug report is to enable us to x the bug. It may be  
that the bug has been reported previously, but neither you nor we can know that unless  
your bug report is complete and self-contained.  
Sometimes people give a few sketchy facts and ask, \Does this ring a bell?" Those bug  
reports are useless, and we urge everyone to refuse to respond to them except to chide  
the sender to report bugs properly.  
To enable us to x the bug, you should include all these things:  
The version of GDB. GDB announces it if you start with no arguments; you can  
also print it at any time using show version.  
Without this, we will not know whether there is any point in looking for the bug  
in the current version of GDB.  
The type of machine you are using, and the operating system name and version  
number.  
What compiler (and its version) was used to compile the program you are  
debugging| e.g. \HP92453-01 A.10.32.03 HP C Compiler". Use the whatcommand  
with the pathname of the compile command (`what /opt/ansic/bin/cc', for example)  
to obtain this information.  
The command arguments you gave the compiler to compile your example and  
observe the bug. For example, did you use `-O'? To guarantee you will not omit  
something important, list them all. A copy of the Makefile (or the output from  
make) is sufficient.  
If we were to try to guess the arguments, we would probably guess wrong and  
then we might not encounter the bug.  
A complete input script, and all necessary source files, that will reproduce the bug  
A description of what behavior you observe that you believe is incorrect. For  
example, \It gets a fatal signal."  
Of course, if the bug is that GDB gets a fatal signal, then we will certainly notice  
it. But if the bug is incorrect output, we might not notice unless it is glaringly  
wrong. You might as well not give us a chance to make a mistake.  
Even if the problem you experience is a fatal signal, you should still say so explicitly.  
Suppose something strange is going on, such as, your copy of GDB is out of synch,  
or you have encountered a bug in the C library on your system. (This has  
happened!) Your copy might crash and ours would not. If you told us to expect a  
crash, then when ours fails to crash, we would know that the bug was not  
happening for us. If you had not told us to expect a crash, then we would not be  
able to draw any conclusion from our observations.  
Here are some things that are not necessary:  
362 Reporting Bugs in GDB  
A description of the envelope of the bug.  
Often people who encounter a bug spend a lot of time investigating which changes  
to the input file will make the bug go away and which changes will not affect it.  
This is often time consuming and not very useful, because the way we will nd the  
bug is by running a single example under the debugger with breakpoints, not by  
pure deduction from a series of examples. We recommend that you save your time  
for something else.  
Of course, if you can nd a simpler example to report instead of the original one,  
that is a convenience for us. Errors in the output will be easier to spot, running  
under the debugger will take less time, and so on.  
However, simplification is not vital; if you do not want to do this, report the bug  
anyway and send us the entire test case you used.  
A patch for the bug.  
A patch for the bug does help us if it is a good one. But do not omit the necessary  
information, such as the test case, on the assumption that a patch is all we need.  
We might see problems with your patch and decide to x the problem another way,  
or we might not understand it at all.  
Sometimes with a program as complicated as GDB it is very hard to construct an  
example that will make the program follow a certain path through the code. If you  
do not send us the example, we will not be able to construct one, so we will not  
be able to verify that the bug is fixed.  
And if we cannot understand what bug you are trying to x, or why your patch  
should be an improvement, we will not install it. A test case will help us to  
understand.  
A guess about what the bug is or what it depends on.  
Such guesses are usually wrong. Even we cannot guess right about such things  
without first using the debugger to nd the facts.  
22.2 How to report bugs 363  
364  
A Installing GDB  
If you obtain GDB (WDB) as part of the HP ANSI C, HP ANSI C++ Developer's Kit for  
HP-UX Release 11.x, or HP Fortran, you do not have to take any special action to build  
or install GDB.  
If you obtain GDB (WDB) from an HP web site, you may download either an swinstall  
package or a source tree, or both.  
Most customers will want to install the GDB binary that is part of the swinstall package.  
To do so, use a command of the form:  
/usr/sbin/swinstall -s package-name WDB  
Alternatively, it is possible to build GDB from the source distribution. If you want to  
modify the debugger sources to tailor GDB to your needs, you may wish to do this.  
The source distribution consists of a tarfile containing the source tree rooted at  
gdb-4.17/.... The instructions that follow describe how to build a `gdb' executable  
from this source tree. HP believes that these instructions apply to the WDB source tree  
that it distributes. However, HP does not explicitly support building a `gdb' for any  
non-HP platform from the WDB source tree. It may work, but HP has not tested it for  
any platforms other than those described in the WDB Release Notes.  
You can nd additional information specific to Hewlett-Packard in the  
`README.HP.WDB' file at the root of the source tree.  
GDB comes with a configure script that automates the process of preparing GDB for  
5
installation; you can then use make to build the gdb program.  
The GDB distribution includes all the source code you need for GDB in a single  
directory, whose name is usually composed by appending the version number to `gdb'.  
For example, the GDB version gdb-199991101 distribution is in the `gdb-gdb-199991101'  
directory. That directory contains:  
gdb-gdb-199991101/configure (and script for configuring GDB and all its supporting  
supporting files)  
libraries  
gdb-gdb-199991101/gdb  
gdb-gdb-199991101/bfd  
gdb-gdb-199991101/include  
gdb-gdb-199991101/libiberty  
gdb-gdb-199991101/opcodes  
the source specific to GDB itself  
source for the Binary File Descriptor library  
gnu include files  
source for the `-liberty' free software library  
source for the library of opcode tables and  
disassemblers  
gdb-gdb-199991101/readline  
5
source for the gnu command-line interface  
gdb-gdb-199991101/glob  
source for the gnu filename pattern-matching  
subroutine  
5. If you have a more recent version of GDB than gdb-199991101, look at the READMEfile in the sources; we  
may have improved the installation procedures since publishing this manual.  
365  
 
gdb-gdb-199991101/mmalloc  
source for the gnu memory-mapped malloc  
package  
The simplest way to configure and build GDB is to run configure from the  
gdb-version-numbersource directory, which in this example is the  
gdb-gdb-199991101directory.  
First switch to the gdb-version-numbersource directory if you are not already in  
it; then run configure. Pass the identifier for the platform on which GDB will run as an  
argument.  
For example:  
cd gdb-gdb-199991101  
./configure host  
make  
where host is an identifier such as sun4'or decstation, that identifies the platform  
where GDB will run. (You can often leave off host; configuretries to guess the  
correct value by examining your system.)  
Running configurehostand then running make builds the `bfd', `readline',  
`mmalloc', and `libiberty' libraries, then gdb itself. The configured source files, and  
the binaries, are left in the corresponding source directories.  
configureis a Bourne-shell (/bin/sh) script; if your system does not recognize this  
automatically when you run a different shell, you may need to run sh on it explicitly:  
sh configure host  
If you run configure from a directory that contains source directories for multiple  
libraries or programs, such as the gdb-gdb-199991101source directory for version  
gdb-199991101, configurecreates configuration files for every directory level  
underneath (unless you tell it not to, with the --norecursionoption).  
You can run the configurescript from any of the subordinate directories in the GDB  
distribution if you only want to configure that subdirectory, but be sure to specify a  
pathto it.  
For example, with version gdb-199991101, type the following to configure only the  
bfd subdirectory:  
cd gdb-gdb-199991101/bfd  
../configure host  
You can install (gdb) anywhere; it has no hardwired paths. However, you should make  
sure that the shell on your path (named by the SHELLenvironment variable) is publicly  
readable. Remember that GDB uses the shell to start your program|some systems  
refuse to let GDB debug child processes whose programs are not readable.  
A.1 Compiling GDB in another directory  
If you want to run GDB versions for several host or target machines, you need a different  
gdb compiled for each combination of host and target. configureis designed to make  
this easy by allowing you to generate each configuration in a separate subdirectory,  
rather than in the source directory. If your makeprogram handles the VPATHfeature  
366 Installing GDB  
 
(GNU makedoes), running makein each of these directories builds the gdb program  
specified there.  
To build gdb in a separate directory, run configurewith the --srcdiroption to  
specify where to find the source. (You also need to specify a pathto find configure  
itself from your working directory. If the path to configure would be the same as the  
argument to --srcdir, you can leave out the --srcdiroption; it is assumed.)  
For example, with version gdb-199991101, you can build GDB in a separate directory  
for a Sun 4 like this:  
cd gdb-gdb-199991101  
mkdir ../gdb-sun4  
cd ../gdb-sun4  
../gdb-gdb-199991101/configure sun4  
make  
When configure builds a configuration using a remote source directory, it creates a tree  
for the binaries with the same structure (and using the same names) as the tree under  
the source directory. In the example, you'd nd the Sun 4 library `libiberty.a' in the  
directory gdb-sun4/libiberty, and GDB itself in `gdb-sun4/gdb.  
One popular reason to build several GDB configurations in separate directories is to  
configure GDB for cross-compiling (where GDB runs on one machine — the host —  
while debugging programs that run on another machine — the target). You specify a  
cross-debugging target by giving the --target=targetoption to configure.  
When you run maketo build a program or library, you must run it in a configured  
directory — whatever directory you were in when you called configure(or one of  
its subdirectories).  
The Makefilethat configuregenerates in each source directory also runs recursively.  
If you type makein a source directory such as gdb-gdb-199991101 (or in a separate  
configured directory configured with `--srcdir=dirname/gdb-gdb-199991101),  
you will build all the required libraries, and then build GDB.  
When you have multiple hosts or targets configured in separate directories, you can  
run makeon them in parallel (for example, if they are NFS-mounted on each of the  
hosts); they will not interfere with each other.  
A.2 Specifying names for hosts and targets  
The specifications used for hosts and targets in the configurescript are based on a  
three-part naming scheme, but some short predefined aliases are also supported. The  
full naming scheme encodes three pieces of information in the following pattern:  
architecture-vendor-os  
For example, you can use the alias sun4 as a host argument, or as the value for target  
in a --target=targetoption. The equivalent full name is sparc-sun-sunos4.  
The configurescript accompanying GDB does not provide any query facility to list  
all supported host and target names or aliases. configurecalls the Bourne shell script  
config.subto map abbreviations to full names; you can read the script, if you wish,  
or you can use it to test your guesses on abbreviations|for example:  
A.2 Specifying names for hosts and targets 367  
 
% sh config.sub i386-linux  
i386-pc-linux-gnu  
% sh config.sub alpha-linux  
alpha-unknown-linux-gnu  
% sh config.sub hp9k700  
hppa1.1-hp-hpux  
% sh config.sub sun4  
sparc-sun-sunos4.1.1  
% sh config.sub sun3  
m68k-sun-sunos4.1.1  
% sh config.sub i986v  
Invalid configuration `i986v': machine `i986v' not recognized  
config.subis also distributed in the GDB source directory (gdb-gdb-199991101,  
for version gdb-199991101).  
A.3 configureoptions  
Here is a summary of the configureoptions and arguments that are most often useful  
for building GDB. configurealso has several other options not listed here. See Info  
file configure.info, or the node What Configure Does, for a full explanation of  
configure.  
configure [--help]  
[--prefix=dir]  
[--exec-prefix=dir]  
[--srcdir=dirname]  
[--norecursion] [--rm]  
[--target=target]  
host  
You may introduce options with a single -rather than --if you prefer; but you may  
abbreviate option names if you use --.  
--help  
Display a quick summary of how to invoke configure.  
--prefix=dir  
Configure the source to install programs and files under  
directory `dir.  
--exec-prefix=dir  
Configure the source to install programs under directory  
dir.  
--srcdir=dirname  
Warning: using this option requires gnu make, or another  
make that implements the VPATHfeature. Use this option  
to make configurations in directories separate from the GDB  
source directories. Among other things, you can use this to  
build (or maintain) several configurations simultaneously,  
in separate directories. configurewrites configuration  
specific files in the current directory, but arranges for them  
to use the source in the directory dirname. configure  
creates directories under the working directory in parallel  
to the source directories below dirname.  
368 Installing GDB  
 
--norecursion  
--target=target  
Configure only the directory level where configureis  
executed; do not propagate configuration to subdirectories.  
Configure GDB for cross-debugging programs running on  
the specified target. Without this option, GDB is configured  
to debug programs that run on the same machine (host) as  
GDB itself. There is no convenient way to generate a list of  
all available targets.  
host ...  
Configure GDB to run on the specified host. There is no  
convenient way to generate a list of all available hosts.  
There are many other options available as well, but they are generally needed for special  
purposes only.  
A.3 configure options 369  

Hotpoint E600 User Manual
GE General Electric Cooktop PP942 User Manual
GE DECT 6 0 28112 User Manual
Electrolux EKD60760 User Manual
Cypress NoBL CY7C1372DV25 User Manual
Cypress CY7C1515JV18 User Manual
Cuisinart PG 26757 User Manual
Corsair AirFlow Pro User Manual
Black Box Computer Hardware IC181C User Manual
Atmel AT85DVK 07 User Manual