toolchain/share/doc/gdb/Files.html

575 lines
38 KiB
HTML

<html lang="en">
<head>
<title>Files - Debugging with GDB</title>
<meta http-equiv="Content-Type" content="text/html">
<meta name="description" content="Debugging with GDB">
<meta name="generator" content="makeinfo 4.13">
<link title="Top" rel="start" href="index.html#Top">
<link rel="up" href="GDB-Files.html#GDB-Files" title="GDB Files">
<link rel="next" href="File-Caching.html#File-Caching" title="File Caching">
<link href="http://www.gnu.org/software/texinfo/" rel="generator-home" title="Texinfo Homepage">
<!--
Copyright (C) 1988-2019 Free Software Foundation, Inc.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.3 or
any later version published by the Free Software Foundation; with the
Invariant Sections being ``Free Software'' and ``Free Software Needs
Free Documentation'', with the Front-Cover Texts being ``A GNU Manual,''
and with the Back-Cover Texts as in (a) below.
(a) The FSF's Back-Cover Text is: ``You are free to copy and modify
this GNU Manual. Buying copies from GNU Press supports the FSF in
developing GNU and promoting software freedom.''
-->
<meta http-equiv="Content-Style-Type" content="text/css">
<style type="text/css"><!--
pre.display { font-family:inherit }
pre.format { font-family:inherit }
pre.smalldisplay { font-family:inherit; font-size:smaller }
pre.smallformat { font-family:inherit; font-size:smaller }
pre.smallexample { font-size:smaller }
pre.smalllisp { font-size:smaller }
span.sc { font-variant:small-caps }
span.roman { font-family:serif; font-weight:normal; }
span.sansserif { font-family:sans-serif; font-weight:normal; }
--></style>
</head>
<body>
<div class="node">
<a name="Files"></a>
<p>
Next:&nbsp;<a rel="next" accesskey="n" href="File-Caching.html#File-Caching">File Caching</a>,
Up:&nbsp;<a rel="up" accesskey="u" href="GDB-Files.html#GDB-Files">GDB Files</a>
<hr>
</div>
<h3 class="section">18.1 Commands to Specify Files</h3>
<p><a name="index-symbol-table-1202"></a><a name="index-core-dump-file-1203"></a>
You may want to specify executable and core dump file names. The usual
way to do this is at start-up time, using the arguments to
<span class="sc">gdb</span>'s start-up commands (see <a href="Invocation.html#Invocation">Getting In and Out of <span class="sc">gdb</span></a>).
<p>Occasionally it is necessary to change to a different file during a
<span class="sc">gdb</span> session. Or you may run <span class="sc">gdb</span> and forget to
specify a file you want to use. Or you are debugging a remote target
via <code>gdbserver</code> (see <a href="Server.html#Server">file</a>). In these situations the <span class="sc">gdb</span> commands to specify
new files are useful.
<a name="index-executable-file-1204"></a>
<a name="index-file-1205"></a>
<dl><dt><code>file </code><var>filename</var><dd>Use <var>filename</var> 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 <code>run</code> command. If you do not specify a
directory and the file is not found in the <span class="sc">gdb</span> working directory,
<span class="sc">gdb</span> uses the environment variable <code>PATH</code> 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 <span class="sc">gdb</span>
and your program, using the <code>path</code> command.
<p><a name="index-unlinked-object-files-1206"></a><a name="index-patching-object-files-1207"></a>You can load unlinked object <samp><span class="file">.o</span></samp> files into <span class="sc">gdb</span> using
the <code>file</code> command. You will not be able to &ldquo;run&rdquo; an object
file, but you can disassemble functions and inspect variables. Also,
if the underlying BFD functionality supports it, you could use
<kbd>gdb -write</kbd> to patch object files using this technique. Note
that <span class="sc">gdb</span> can neither interpret nor modify relocations in this
case, so branches and some initialized variables will appear to go to
the wrong place. But this feature is still handy from time to time.
<br><dt><code>file</code><dd><code>file</code> with no argument makes <span class="sc">gdb</span> discard any information it
has on both executable file and the symbol table.
<p><a name="index-exec_002dfile-1208"></a><br><dt><code>exec-file </code><span class="roman">[</span> <var>filename</var> <span class="roman">]</span><dd>Specify that the program to be run (but not the symbol table) is found
in <var>filename</var>. <span class="sc">gdb</span> searches the environment variable <code>PATH</code>
if necessary to locate your program. Omitting <var>filename</var> means to
discard information on the executable file.
<p><a name="index-symbol_002dfile-1209"></a><br><dt><code>symbol-file </code><span class="roman">[</span> <var>filename</var> <span class="roman">[</span><code> -o </code><var>offset</var> <span class="roman">]]</span><dd>Read symbol table information from file <var>filename</var>. <code>PATH</code> is
searched when necessary. Use the <code>file</code> command to get both symbol
table and program to run from the same file.
<p>If an optional <var>offset</var> is specified, it is added to the start
address of each section in the symbol file. This is useful if the
program is relocated at runtime, such as the Linux kernel with kASLR
enabled.
<p><code>symbol-file</code> with no argument clears out <span class="sc">gdb</span> information on your
program's symbol table.
<p>The <code>symbol-file</code> command causes <span class="sc">gdb</span> to forget the contents of
some 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
<span class="sc">gdb</span>.
<p><code>symbol-file</code> does not repeat if you press &lt;RET&gt; again after
executing it once.
<p>When <span class="sc">gdb</span> 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 <span class="sc">gnu</span> compiler, or
other compilers that adhere to the local conventions.
Best results are usually obtained from <span class="sc">gnu</span> compilers; for example,
using <span class="sc">gcc</span> you can generate debugging information for
optimized code.
<p>For most kinds of object files, with the exception of old SVR3 systems
using COFF, the <code>symbol-file</code> command does not normally read the
symbol table in full right away. Instead, it scans the symbol table
quickly to find which source files and which symbols are present. The
details are read later, one source file at a time, as they are needed.
<p>The purpose of this two-stage reading strategy is to make <span class="sc">gdb</span>
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 <code>set verbose</code> command can turn these
pauses into messages if desired. See <a href="Messages_002fWarnings.html#Messages_002fWarnings">Optional Warnings and Messages</a>.)
<p>We have not implemented the two-stage strategy for COFF yet. When the
symbol table is stored in COFF format, <code>symbol-file</code> reads the
symbol table data in full right away. Note that &ldquo;stabs-in-COFF&rdquo;
still does the two-stage strategy, since the debug info is actually
in stabs format.
<p><a name="index-readnow-1210"></a><a name="index-reading-symbols-immediately-1211"></a><a name="index-symbols_002c-reading-immediately-1212"></a><br><dt><code>symbol-file </code><span class="roman">[</span><code> -readnow </code><span class="roman">]</span> <var>filename</var><dt><code>file </code><span class="roman">[</span><code> -readnow </code><span class="roman">]</span> <var>filename</var><dd>You can override the <span class="sc">gdb</span> two-stage strategy for reading symbol
tables by using the &lsquo;<samp><span class="samp">-readnow</span></samp>&rsquo; option with any of the commands that
load symbol table information, if you want to be sure <span class="sc">gdb</span> has the
entire symbol table available.
<p><a name="index-g_t_0040code_007b_002dreadnever_007d_002c-option-for-symbol_002dfile-command-1213"></a><a name="index-never-read-symbols-1214"></a><a name="index-symbols_002c-never-read-1215"></a><br><dt><code>symbol-file </code><span class="roman">[</span><code> -readnever </code><span class="roman">]</span> <var>filename</var><dt><code>file </code><span class="roman">[</span><code> -readnever </code><span class="roman">]</span> <var>filename</var><dd>You can instruct <span class="sc">gdb</span> to never read the symbolic information
contained in <var>filename</var> by using the &lsquo;<samp><span class="samp">-readnever</span></samp>&rsquo; option.
See <a href="_002d_002dreadnever.html#g_t_002d_002dreadnever">&ndash;readnever</a>.
<!-- FIXME: for now no mention of directories, since this seems to be in -->
<!-- flux. 13mar1992 status is that in theory GDB would look either in -->
<!-- current dir or in same dir as myprog; but issues like competing -->
<!-- GDB's, or clutter in system dirs, mean that in practice right now -->
<!-- only current dir is used. FFish says maybe a special GDB hierarchy -->
<!-- (eg rooted in val of env var GDBSYMS) could exist for mappable symbol -->
<!-- files. -->
<p><a name="index-core_002dfile-1216"></a><br><dt><code>core-file </code><span class="roman">[</span><var>filename</var><span class="roman">]</span><dt><code>core</code><dd>Specify the whereabouts of a core dump file to be used as the &ldquo;contents
of memory&rdquo;. Traditionally, core files contain only some parts of the
address space of the process that generated them; <span class="sc">gdb</span> can access the
executable file itself for other parts.
<p><code>core-file</code> with no argument specifies that no core file is
to be used.
<p>Note that the core file is ignored when your program is actually running
under <span class="sc">gdb</span>. 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. To do this, use the <code>kill</code> command
(see <a href="Kill-Process.html#Kill-Process">Killing the Child Process</a>).
<p><a name="index-add_002dsymbol_002dfile-1217"></a><a name="index-dynamic-linking-1218"></a><br><dt><code>add-symbol-file </code><var>filename</var> <span class="roman">[</span><code> -readnow </code><span class="roman">|</span><code> -readnever </code><span class="roman">]</span> <span class="roman">[</span><code> -o </code><var>offset</var> <span class="roman">]</span> <span class="roman">[</span> <var>textaddress</var> <span class="roman">]</span> <span class="roman">[</span><code> -s </code><var>section</var> <var>address</var><code> ... </code><span class="roman">]</span><dd>The <code>add-symbol-file</code> command reads additional symbol table
information from the file <var>filename</var>. You would use this command
when <var>filename</var> has been dynamically loaded (by some other means)
into the program that is running. The <var>textaddress</var> parameter gives
the memory address at which the file's text section has been loaded.
You can additionally specify the base address of other sections using
an arbitrary number of &lsquo;<samp><span class="samp">-s </span><var>section</var> <var>address</var></samp>&rsquo; pairs.
If a section is omitted, <span class="sc">gdb</span> will use its default addresses
as found in <var>filename</var>. Any <var>address</var> or <var>textaddress</var>
can be given as an expression.
<p>If an optional <var>offset</var> is specified, it is added to the start
address of each section, except those for which the address was
specified explicitly.
<p>The symbol table of the file <var>filename</var> is added to the symbol table
originally read with the <code>symbol-file</code> command. You can use the
<code>add-symbol-file</code> command any number of times; the new symbol data
thus read is kept in addition to the old.
<p>Changes can be reverted using the command <code>remove-symbol-file</code>.
<p><a name="index-relocatable-object-files_002c-reading-symbols-from-1219"></a><a name="index-object-files_002c-relocatable_002c-reading-symbols-from-1220"></a><a name="index-reading-symbols-from-relocatable-object-files-1221"></a><a name="index-symbols_002c-reading-from-relocatable-object-files-1222"></a><a name="index-g_t_0040file_007b_002eo_007d-files_002c-reading-symbols-from-1223"></a>Although <var>filename</var> is typically a shared library file, an
executable file, or some other object file which has been fully
relocated for loading into a process, you can also load symbolic
information from relocatable <samp><span class="file">.o</span></samp> files, as long as:
<ul>
<li>the file's symbolic information refers only to linker symbols defined in
that file, not to symbols defined by other object files,
<li>every section the file's symbolic information refers to has actually
been loaded into the inferior, as it appears in the file, and
<li>you can determine the address at which every section was loaded, and
provide these to the <code>add-symbol-file</code> command.
</ul>
<p class="noindent">Some embedded operating systems, like Sun Chorus and VxWorks, can load
relocatable files into an already running program; such systems
typically make the requirements above easy to meet. However, it's
important to recognize that many native systems use complex link
procedures (<code>.linkonce</code> section factoring and C<tt>++</tt> constructor table
assembly, for example) that make the requirements difficult to meet. In
general, one cannot assume that using <code>add-symbol-file</code> to read a
relocatable object file's symbolic information will have the same effect
as linking the relocatable object file into the program in the normal
way.
<p><code>add-symbol-file</code> does not repeat if you press &lt;RET&gt; after using it.
<p><a name="index-remove_002dsymbol_002dfile-1224"></a><br><dt><code>remove-symbol-file </code><var>filename</var><br><dt><code>remove-symbol-file -a </code><var>address</var><dd>Remove a symbol file added via the <code>add-symbol-file</code> command. The
file to remove can be identified by its <var>filename</var> or by an <var>address</var>
that lies within the boundaries of this symbol file in memory. Example:
<pre class="smallexample"> (gdb) add-symbol-file /home/user/gdb/mylib.so 0x7ffff7ff9480
add symbol table from file "/home/user/gdb/mylib.so" at
.text_addr = 0x7ffff7ff9480
(y or n) y
Reading symbols from /home/user/gdb/mylib.so...done.
(gdb) remove-symbol-file -a 0x7ffff7ff9480
Remove symbol table from file "/home/user/gdb/mylib.so"? (y or n) y
(gdb)
</pre>
<p><code>remove-symbol-file</code> does not repeat if you press &lt;RET&gt; after using it.
<p><a name="index-add_002dsymbol_002dfile_002dfrom_002dmemory-1225"></a><a name="index-g_t_0040code_007bsyscall-DSO_007d-1226"></a><a name="index-load-symbols-from-memory-1227"></a><br><dt><code>add-symbol-file-from-memory </code><var>address</var><dd>Load symbols from the given <var>address</var> in a dynamically loaded
object file whose image is mapped directly into the inferior's memory.
For example, the Linux kernel maps a <code>syscall DSO</code> into each
process's address space; this DSO provides kernel-specific code for
some system calls. The argument can be any expression whose
evaluation yields the address of the file's shared object file header.
For this command to work, you must have used <code>symbol-file</code> or
<code>exec-file</code> commands in advance.
<p><a name="index-section-1228"></a><br><dt><code>section </code><var>section</var> <var>addr</var><dd>The <code>section</code> command changes the base address of the named
<var>section</var> of the exec file to <var>addr</var>. This can be used if the
exec file does not contain section addresses, (such as in the
<code>a.out</code> format), or when the addresses specified in the file
itself are wrong. Each section must be changed separately. The
<code>info files</code> command, described below, lists all the sections and
their addresses.
<p><a name="index-info-files-1229"></a><a name="index-info-target-1230"></a><br><dt><code>info files</code><dt><code>info target</code><dd><code>info files</code> and <code>info target</code> are synonymous; both print the
current target (see <a href="Targets.html#Targets">Specifying a Debugging Target</a>),
including the names of the executable and core dump files currently in
use by <span class="sc">gdb</span>, and the files from which symbols were loaded. The
command <code>help target</code> lists all possible targets rather than
current ones.
<p><a name="index-maint-info-sections-1231"></a><br><dt><code>maint info sections</code><dd>Another command that can give you extra information about program sections
is <code>maint info sections</code>. In addition to the section information
displayed by <code>info files</code>, this command displays the flags and file
offset of each section in the executable and core dump files. In addition,
<code>maint info sections</code> provides the following command options (which
may be arbitrarily combined):
<dl>
<dt><code>ALLOBJ</code><dd>Display sections for all loaded object files, including shared libraries.
<br><dt><var>sections</var><dd>Display info only for named <var>sections</var>.
<br><dt><var>section-flags</var><dd>Display info only for sections for which <var>section-flags</var> are true.
The section flags that <span class="sc">gdb</span> currently knows about are:
<dl>
<dt><code>ALLOC</code><dd>Section will have space allocated in the process when loaded.
Set for all sections except those containing debug information.
<br><dt><code>LOAD</code><dd>Section will be loaded from the file into the child process memory.
Set for pre-initialized code and data, clear for <code>.bss</code> sections.
<br><dt><code>RELOC</code><dd>Section needs to be relocated before loading.
<br><dt><code>READONLY</code><dd>Section cannot be modified by the child process.
<br><dt><code>CODE</code><dd>Section contains executable code only.
<br><dt><code>DATA</code><dd>Section contains data only (no executable code).
<br><dt><code>ROM</code><dd>Section will reside in ROM.
<br><dt><code>CONSTRUCTOR</code><dd>Section contains data for constructor/destructor lists.
<br><dt><code>HAS_CONTENTS</code><dd>Section is not empty.
<br><dt><code>NEVER_LOAD</code><dd>An instruction to the linker to not output the section.
<br><dt><code>COFF_SHARED_LIBRARY</code><dd>A notification to the linker that the section contains
COFF shared library information.
<br><dt><code>IS_COMMON</code><dd>Section contains common symbols.
</dl>
</dl>
<a name="index-set-trust_002dreadonly_002dsections-1232"></a><a name="index-read_002donly-sections-1233"></a><br><dt><code>set trust-readonly-sections on</code><dd>Tell <span class="sc">gdb</span> that readonly sections in your object file
really are read-only (i.e. that their contents will not change).
In that case, <span class="sc">gdb</span> can fetch values from these sections
out of the object file, rather than from the target program.
For some targets (notably embedded ones), this can be a significant
enhancement to debugging performance.
<p>The default is off.
<br><dt><code>set trust-readonly-sections off</code><dd>Tell <span class="sc">gdb</span> not to trust readonly sections. This means that
the contents of the section might change while the program is running,
and must therefore be fetched from the target when needed.
<br><dt><code>show trust-readonly-sections</code><dd>Show the current setting of trusting readonly sections.
</dl>
<p>All file-specifying commands allow both absolute and relative file names
as arguments. <span class="sc">gdb</span> always converts the file name to an absolute file
name and remembers it that way.
<p><a name="index-shared-libraries-1234"></a><a name="Shared-Libraries"></a><span class="sc">gdb</span> supports <span class="sc">gnu</span>/Linux, MS-Windows, SunOS,
Darwin/Mach-O, SVr4, IBM RS/6000 AIX, QNX Neutrino, FDPIC (FR-V), and
DSBT (TIC6X) shared libraries.
<p>On MS-Windows <span class="sc">gdb</span> must be linked with the Expat library to support
shared libraries. See <a href="Expat.html#Expat">Expat</a>.
<p><span class="sc">gdb</span> automatically loads symbol definitions from shared libraries
when you use the <code>run</code> command, or when you examine a core file.
(Before you issue the <code>run</code> command, <span class="sc">gdb</span> does not understand
references to a function in a shared library, however&mdash;unless you are
debugging a core file).
<!-- FIXME: some @value{GDBN} release may permit some refs to undef -->
<!-- FIXME...symbols-eg in a break cmd-assuming they are from a shared -->
<!-- FIXME...lib; check this from time to time when updating manual -->
<p>There are times, however, when you may wish to not automatically load
symbol definitions from shared libraries, such as when they are
particularly large or there are many of them.
<p>To control the automatic loading of shared library symbols, use the
commands:
<a name="index-set-auto_002dsolib_002dadd-1235"></a>
<dl><dt><code>set auto-solib-add </code><var>mode</var><dd>If <var>mode</var> is <code>on</code>, symbols from all shared object libraries
will be loaded automatically when the inferior begins execution, you
attach to an independently started inferior, or when the dynamic linker
informs <span class="sc">gdb</span> that a new library has been loaded. If <var>mode</var>
is <code>off</code>, symbols must be loaded manually, using the
<code>sharedlibrary</code> command. The default value is <code>on</code>.
<p><a name="index-memory-used-for-symbol-tables-1236"></a>If your program uses lots of shared libraries with debug info that
takes large amounts of memory, you can decrease the <span class="sc">gdb</span>
memory footprint by preventing it from automatically loading the
symbols from shared libraries. To that end, type <kbd>set
auto-solib-add off</kbd> before running the inferior, then load each
library whose debug symbols you do need with <kbd>sharedlibrary
</kbd><var>regexp</var>, where <var>regexp</var> is a regular expression that matches
the libraries whose symbols you want to be loaded.
<p><a name="index-show-auto_002dsolib_002dadd-1237"></a><br><dt><code>show auto-solib-add</code><dd>Display the current autoloading mode.
</dl>
<p><a name="index-load-shared-library-1238"></a>To explicitly load shared library symbols, use the <code>sharedlibrary</code>
command:
<a name="index-info-sharedlibrary-1239"></a>
<a name="index-info-share-1240"></a>
<dl><dt><code>info share </code><var>regex</var><dt><code>info sharedlibrary </code><var>regex</var><dd>Print the names of the shared libraries which are currently loaded
that match <var>regex</var>. If <var>regex</var> is omitted then print
all shared libraries that are loaded.
<p><a name="index-info-dll-1241"></a><br><dt><code>info dll </code><var>regex</var><dd>This is an alias of <code>info sharedlibrary</code>.
<p><a name="index-sharedlibrary-1242"></a><a name="index-share-1243"></a><br><dt><code>sharedlibrary </code><var>regex</var><dt><code>share </code><var>regex</var><dd>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 <code>run</code>. If
<var>regex</var> is omitted all shared libraries required by your program are
loaded.
<br><dt><code>nosharedlibrary</code><dd><a name="index-nosharedlibrary-1244"></a><a name="index-unload-symbols-from-shared-libraries-1245"></a>Unload all shared object library symbols. This discards all symbols
that have been loaded from all shared libraries. Symbols from shared
libraries that were loaded by explicit user requests are not
discarded.
</dl>
<p>Sometimes you may wish that <span class="sc">gdb</span> stops and gives you control
when any of shared library events happen. The best way to do this is
to use <code>catch load</code> and <code>catch unload</code> (see <a href="Set-Catchpoints.html#Set-Catchpoints">Set Catchpoints</a>).
<p><span class="sc">gdb</span> also supports the the <code>set stop-on-solib-events</code>
command for this. This command exists for historical reasons. It is
less useful than setting a catchpoint, because it does not allow for
conditions or commands as a catchpoint does.
<dl>
<dt><code>set stop-on-solib-events</code><dd><a name="index-set-stop_002don_002dsolib_002devents-1246"></a>This command controls whether <span class="sc">gdb</span> should give you control
when the dynamic linker notifies it about some shared library event.
The most common event of interest is loading or unloading of a new
shared library.
<br><dt><code>show stop-on-solib-events</code><dd><a name="index-show-stop_002don_002dsolib_002devents-1247"></a>Show whether <span class="sc">gdb</span> stops and gives you control when shared
library events happen.
</dl>
<p>Shared libraries are also supported in many cross or remote debugging
configurations. <span class="sc">gdb</span> needs to have access to the target's libraries;
this can be accomplished either by providing copies of the libraries
on the host system, or by asking <span class="sc">gdb</span> to automatically retrieve the
libraries from the target. If copies of the target libraries are
provided, they need to be the same as the target libraries, although the
copies on the target can be stripped as long as the copies on the host are
not.
<p><a name="index-where-to-look-for-shared-libraries-1248"></a>For remote debugging, you need to tell <span class="sc">gdb</span> where the target
libraries are, so that it can load the correct copies&mdash;otherwise, it
may try to load the host's libraries. <span class="sc">gdb</span> has two variables
to specify the search directories for target libraries.
<a name="index-prefix-for-executable-and-shared-library-file-names-1249"></a>
<a name="index-system-root_002c-alternate-1250"></a>
<a name="index-set-solib_002dabsolute_002dprefix-1251"></a>
<a name="index-set-sysroot-1252"></a>
<dl><dt><code>set sysroot </code><var>path</var><dd>Use <var>path</var> as the system root for the program being debugged. Any
absolute shared library paths will be prefixed with <var>path</var>; many
runtime loaders store the absolute paths to the shared library in the
target program's memory. When starting processes remotely, and when
attaching to already-running processes (local or remote), their
executable filenames will be prefixed with <var>path</var> if reported to
<span class="sc">gdb</span> as absolute by the operating system. If you use
<code>set sysroot</code> to find executables and shared libraries, they need
to be laid out in the same way that they are on the target, with
e.g. a <samp><span class="file">/bin</span></samp>, <samp><span class="file">/lib</span></samp> and <samp><span class="file">/usr/lib</span></samp> hierarchy under
<var>path</var>.
<p>If <var>path</var> starts with the sequence <samp><span class="file">target:</span></samp> and the target
system is remote then <span class="sc">gdb</span> will retrieve the target binaries
from the remote system. This is only supported when using a remote
target that supports the <code>remote get</code> command (see <a href="File-Transfer.html#File-Transfer">Sending files to a remote system</a>). The part of <var>path</var>
following the initial <samp><span class="file">target:</span></samp> (if present) is used as system
root prefix on the remote file system. If <var>path</var> starts with the
sequence <samp><span class="file">remote:</span></samp> this is converted to the sequence
<samp><span class="file">target:</span></samp> by <code>set sysroot</code><a rel="footnote" href="#fn-1" name="fnd-1"><sup>1</sup></a>. If you want
to specify a local system root using a directory that happens to be
named <samp><span class="file">target:</span></samp> or <samp><span class="file">remote:</span></samp>, you need to use some
equivalent variant of the name like <samp><span class="file">./target:</span></samp>.
<p>For targets with an MS-DOS based filesystem, such as MS-Windows and
SymbianOS, <span class="sc">gdb</span> tries prefixing a few variants of the target
absolute file name with <var>path</var>. But first, on Unix hosts,
<span class="sc">gdb</span> converts all backslash directory separators into forward
slashes, because the backslash is not a directory separator on Unix:
<pre class="smallexample"> c:\foo\bar.dll &rArr; c:/foo/bar.dll
</pre>
<p>Then, <span class="sc">gdb</span> attempts prefixing the target file name with
<var>path</var>, and looks for the resulting file name in the host file
system:
<pre class="smallexample"> c:/foo/bar.dll &rArr; /path/to/sysroot/c:/foo/bar.dll
</pre>
<p>If that does not find the binary, <span class="sc">gdb</span> tries removing
the &lsquo;<samp><span class="samp">:</span></samp>&rsquo; character from the drive spec, both for convenience, and,
for the case of the host file system not supporting file names with
colons:
<pre class="smallexample"> c:/foo/bar.dll &rArr; /path/to/sysroot/c/foo/bar.dll
</pre>
<p>This makes it possible to have a system root that mirrors a target
with more than one drive. E.g., you may want to setup your local
copies of the target system shared libraries like so (note &lsquo;<samp><span class="samp">c</span></samp>&rsquo; vs
&lsquo;<samp><span class="samp">z</span></samp>&rsquo;):
<pre class="smallexample"> <samp><span class="file">/path/to/sysroot/c/sys/bin/foo.dll</span></samp>
<samp><span class="file">/path/to/sysroot/c/sys/bin/bar.dll</span></samp>
<samp><span class="file">/path/to/sysroot/z/sys/bin/bar.dll</span></samp>
</pre>
<p class="noindent">and point the system root at <samp><span class="file">/path/to/sysroot</span></samp>, so that
<span class="sc">gdb</span> can find the correct copies of both
<samp><span class="file">c:\sys\bin\foo.dll</span></samp>, and <samp><span class="file">z:\sys\bin\bar.dll</span></samp>.
<p>If that still does not find the binary, <span class="sc">gdb</span> tries
removing the whole drive spec from the target file name:
<pre class="smallexample"> c:/foo/bar.dll &rArr; /path/to/sysroot/foo/bar.dll
</pre>
<p>This last lookup makes it possible to not care about the drive name,
if you don't want or need to.
<p>The <code>set solib-absolute-prefix</code> command is an alias for <code>set
sysroot</code>.
<p><a name="index-default-system-root-1253"></a><a name="index-g_t_0040samp_007b_002d_002dwith_002dsysroot_007d-1254"></a>You can set the default system root by using the configure-time
&lsquo;<samp><span class="samp">--with-sysroot</span></samp>&rsquo; option. If the system root is inside
<span class="sc">gdb</span>'s configured binary prefix (set with &lsquo;<samp><span class="samp">--prefix</span></samp>&rsquo; or
&lsquo;<samp><span class="samp">--exec-prefix</span></samp>&rsquo;), then the default system root will be updated
automatically if the installed <span class="sc">gdb</span> is moved to a new
location.
<p><a name="index-show-sysroot-1255"></a><br><dt><code>show sysroot</code><dd>Display the current executable and shared library prefix.
<p><a name="index-set-solib_002dsearch_002dpath-1256"></a><br><dt><code>set solib-search-path </code><var>path</var><dd>If this variable is set, <var>path</var> is a colon-separated list of
directories to search for shared libraries. &lsquo;<samp><span class="samp">solib-search-path</span></samp>&rsquo;
is used after &lsquo;<samp><span class="samp">sysroot</span></samp>&rsquo; fails to locate the library, or if the
path to the library is relative instead of absolute. If you want to
use &lsquo;<samp><span class="samp">solib-search-path</span></samp>&rsquo; instead of &lsquo;<samp><span class="samp">sysroot</span></samp>&rsquo;, be sure to set
&lsquo;<samp><span class="samp">sysroot</span></samp>&rsquo; to a nonexistent directory to prevent <span class="sc">gdb</span> from
finding your host's libraries. &lsquo;<samp><span class="samp">sysroot</span></samp>&rsquo; is preferred; setting
it to a nonexistent directory may interfere with automatic loading
of shared library symbols.
<p><a name="index-show-solib_002dsearch_002dpath-1257"></a><br><dt><code>show solib-search-path</code><dd>Display the current shared library search path.
<p><a name="index-DOS-file_002dname-semantics-of-file-names_002e-1258"></a><a name="index-set-target_002dfile_002dsystem_002dkind-_0028unix_007cdos_002dbased_007cauto_0029-1259"></a><a name="index-show-target_002dfile_002dsystem_002dkind-1260"></a><br><dt><code>set target-file-system-kind </code><var>kind</var><dd>Set assumed file system kind for target reported file names.
<p>Shared library file names as reported by the target system may not
make sense as is on the system <span class="sc">gdb</span> is running on. For
example, when remote debugging a target that has MS-DOS based file
system semantics, from a Unix host, the target may be reporting to
<span class="sc">gdb</span> a list of loaded shared libraries with file names such as
<samp><span class="file">c:\Windows\kernel32.dll</span></samp>. On Unix hosts, there's no concept of
drive letters, so the &lsquo;<samp><span class="samp">c:\</span></samp>&rsquo; prefix is not normally understood as
indicating an absolute file name, and neither is the backslash
normally considered a directory separator character. In that case,
the native file system would interpret this whole absolute file name
as a relative file name with no directory components. This would make
it impossible to point <span class="sc">gdb</span> at a copy of the remote target's
shared libraries on the host using <code>set sysroot</code>, and impractical
with <code>set solib-search-path</code>. Setting
<code>target-file-system-kind</code> to <code>dos-based</code> tells <span class="sc">gdb</span>
to interpret such file names similarly to how the target would, and to
map them to file names valid on <span class="sc">gdb</span>'s native file system
semantics. The value of <var>kind</var> can be <code>"auto"</code>, in addition
to one of the supported file system kinds. In that case, <span class="sc">gdb</span>
tries to determine the appropriate file system variant based on the
current target's operating system (see <a href="ABI.html#ABI">Configuring the Current ABI</a>). The supported file system settings are:
<dl>
<dt><code>unix</code><dd>Instruct <span class="sc">gdb</span> to assume the target file system is of Unix
kind. Only file names starting the forward slash (&lsquo;<samp><span class="samp">/</span></samp>&rsquo;) character
are considered absolute, and the directory separator character is also
the forward slash.
<br><dt><code>dos-based</code><dd>Instruct <span class="sc">gdb</span> to assume the target file system is DOS based.
File names starting with either a forward slash, or a drive letter
followed by a colon (e.g., &lsquo;<samp><span class="samp">c:</span></samp>&rsquo;), are considered absolute, and
both the slash (&lsquo;<samp><span class="samp">/</span></samp>&rsquo;) and the backslash (&lsquo;<samp><span class="samp">\\</span></samp>&rsquo;) characters are
considered directory separators.
<br><dt><code>auto</code><dd>Instruct <span class="sc">gdb</span> to use the file system kind associated with the
target operating system (see <a href="ABI.html#ABI">Configuring the Current ABI</a>).
This is the default.
</dl>
</dl>
<p><a name="index-file-name-canonicalization-1261"></a><a name="index-base-name-differences-1262"></a>When processing file names provided by the user, <span class="sc">gdb</span>
frequently needs to compare them to the file names recorded in the
program's debug info. Normally, <span class="sc">gdb</span> compares just the
<dfn>base names</dfn> of the files as strings, which is reasonably fast
even for very large programs. (The base name of a file is the last
portion of its name, after stripping all the leading directories.)
This shortcut in comparison is based upon the assumption that files
cannot have more than one base name. This is usually true, but
references to files that use symlinks or similar filesystem
facilities violate that assumption. If your program records files
using such facilities, or if you provide file names to <span class="sc">gdb</span>
using symlinks etc., you can set <code>basenames-may-differ</code> to
<code>true</code> to instruct <span class="sc">gdb</span> to completely canonicalize each
pair of file names it needs to compare. This will make file-name
comparisons accurate, but at a price of a significant slowdown.
<dl>
<dt><code>set basenames-may-differ</code><dd><a name="index-set-basenames_002dmay_002ddiffer-1263"></a>Set whether a source file may have multiple base names.
<br><dt><code>show basenames-may-differ</code><dd><a name="index-show-basenames_002dmay_002ddiffer-1264"></a>Show whether a source file may have multiple base names.
</dl>
<div class="footnote">
<hr>
<h4>Footnotes</h4><p class="footnote"><small>[<a name="fn-1" href="#fnd-1">1</a>]</small> Historically the
functionality to retrieve binaries from the remote system was
provided by prefixing <var>path</var> with <samp><span class="file">remote:</span></samp></p>
<hr></div>
</body></html>