html/manual/debug.html

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>Debugging Support</title><meta name="generator" content="DocBook XSL Stylesheets Vsnapshot" /><meta name="keywords" content="C++, debug" /><meta name="keywords" content="ISO C++, library" /><meta name="keywords" content="ISO C++, runtime, library" /><link rel="home" href="../index.html" title="The GNU C++ Library" /><link rel="up" href="using.html" title="Chapter3.Using" /><link rel="prev" href="using_exceptions.html" title="Exceptions" /><link rel="next" href="std_contents.html" title="PartII. Standard Contents" /></head><body><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="3" align="center">Debugging Support</th></tr><tr><td width="20%" align="left"><a accesskey="p" href="using_exceptions.html">Prev</a></td><th width="60%" align="center">Chapter3.Using</th><td width="20%" align="right"><a accesskey="n" href="std_contents.html">Next</a></td></tr></table><hr /></div><div class="section"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a id="manual.intro.using.debug"></a>Debugging Support</h2></div></div></div><p>
  There are numerous things that can be done to improve the ease with
  which C++ binaries are debugged when using the GNU tool chain. Here
  are some of them.
</p><div class="section"><div class="titlepage"><div><div><h3 class="title"><a id="debug.compiler"></a>Using <span class="command"><strong>g++</strong></span></h3></div></div></div><p>
    Compiler flags determine how debug information is transmitted
    between compilation and debug or analysis tools.
  </p><p>
    The default optimizations and debug flags for a libstdc++ build
    are <code class="code">-g -O2</code>. However, both debug and optimization
    flags can be varied to change debugging characteristics. For
    instance, turning off all optimization via the <code class="code">-g -O0
    -fno-inline</code> flags will disable inlining and optimizations,
    and add debugging information, so that stepping through all functions,
    (including inlined constructors and destructors) is possible. In
    addition, <code class="code">-fno-eliminate-unused-debug-types</code> can be
    used when additional debug information, such as nested class info,
    is desired.
</p><p>
  Or, the debug format that the compiler and debugger use to
  communicate information about source constructs can be changed via
  <code class="code">-gdwarf-2</code> or <code class="code">-gstabs</code> flags: some debugging
  formats permit more expressive type and scope information to be
  shown in GDB. Expressiveness can be enhanced by flags like
  <code class="code">-g3</code>. The default debug information for a particular
  platform can be identified via the value set by the
  PREFERRED_DEBUGGING_TYPE macro in the GCC sources.
</p><p>
  Many other options are available: please see <a class="link" href="http://gcc.gnu.org/onlinedocs/gcc/Debugging-Options.html#Debugging%20Options" target="_top">"Options
  for Debugging Your Program"</a> in Using the GNU Compiler
  Collection (GCC) for a complete list.
</p></div><div class="section"><div class="titlepage"><div><div><h3 class="title"><a id="debug.req"></a>Debug Versions of Library Binary Files</h3></div></div></div><p>
  If you would like debug symbols in libstdc++, there are two ways to
  build libstdc++ with debug flags. The first is to create a separate
  debug build by running make from the top-level of a tree
  freshly-configured with
</p><pre class="programlisting">
     --enable-libstdcxx-debug
</pre><p>and perhaps</p><pre class="programlisting">
     --enable-libstdcxx-debug-flags='...'
</pre><p>
  Both the normal build and the debug build will persist, without
  having to specify <code class="code">CXXFLAGS</code>, and the debug library will
  be installed in a separate directory tree, in <code class="code">(prefix)/lib/debug</code>.
  For more information, look at the
  <a class="link" href="configure.html" title="Configure">configuration</a> section.
</p><p>
  A second approach is to use the configuration flags
</p><pre class="programlisting">
     make CXXFLAGS='-g3 -fno-inline -O0' all
</pre><p>
  This quick and dirty approach is often sufficient for quick
  debugging tasks, when you cannot or don't want to recompile your
  application to use the <a class="link" href="debug_mode.html" title="Chapter17.Debug Mode">debug mode</a>.</p></div><div class="section"><div class="titlepage"><div><div><h3 class="title"><a id="debug.memory"></a>Memory Leak Hunting</h3></div></div></div><p>
  On many targets GCC supports AddressSanitizer, a fast memory error detector,
  which is enabled by the <code class="option">-fsanitize=address</code> option.
</p><p>
  There are also various third party memory tracing and debug utilities
  that can be used to provide detailed memory allocation information
  about C++ code. An exhaustive list of tools is not going to be
  attempted, but includes <code class="code">mtrace</code>, <code class="code">valgrind</code>,
  <code class="code">mudflap</code> (no longer supported since GCC 4.9.0), ElectricFence,
  and the non-free commercial product <code class="code">purify</code>.
  In addition, <code class="code">libcwd</code>, jemalloc and TCMalloc have replacements
  for the global <code class="code">new</code> and <code class="code">delete</code> operators
  that can track memory allocation and deallocation and provide useful
  memory statistics.
</p><p>
  For valgrind, there are some specific items to keep in mind. First
  of all, use a version of valgrind that will work with current GNU
  C++ tools: the first that can do this is valgrind 1.0.4, but later
  versions should work better. Second, using an unoptimized build
  might avoid confusing valgrind.
</p><p>
  Third, it may be necessary to force deallocation in other libraries
  as well, namely the "C" library. On GNU/Linux, this can be accomplished
  with the appropriate use of the <code class="code">__cxa_atexit</code> or
  <code class="code">atexit</code> functions.
</p><pre class="programlisting">
   #include &lt;cstdlib&gt;

   extern "C" void __libc_freeres(void);

   void do_something() { }

   int main()
   {
     atexit(__libc_freeres);
     do_something();
     return 0;
   }
</pre><p>or, using <code class="code">__cxa_atexit</code>:</p><pre class="programlisting">
   extern "C" void __libc_freeres(void);
   extern "C" int __cxa_atexit(void (*func) (void *), void *arg, void *d);

   void do_something() { }

   int main()
   {
      extern void* __dso_handle __attribute__ ((__weak__));
      __cxa_atexit((void (*) (void *)) __libc_freeres, NULL,
		   &amp;__dso_handle ? __dso_handle : NULL);
      do_test();
      return 0;
   }
</pre><p>
  Suggested valgrind flags, given the suggestions above about setting
  up the runtime environment, library, and test file, might be:
</p><pre class="programlisting">
   valgrind -v --num-callers=20 --leak-check=yes --leak-resolution=high --show-reachable=yes a.out
</pre><div class="section"><div class="titlepage"><div><div><h4 class="title"><a id="debug.memory.mtalloc"></a>Non-memory leaks in Pool and MT allocators</h4></div></div></div><p>
  There are different kinds of allocation schemes that can be used by
  <code class="code">std::allocator</code>. Prior to GCC 3.4.0 the default was to use
  a pooling allocator, <code class="classname">pool_allocator</code>,
  which is still available as the optional
  <code class="classname">__pool_alloc</code> extension.
  Another optional extension, <code class="classname">__mt_alloc</code>,
  is a high-performance pool allocator.
</p><p>
  In a suspect executable these pooling allocators can give
  the mistaken impression that memory is being leaked,
  when in reality the memory "leak" is a pool being used
  by the library's allocator and is reclaimed after program
  termination.
</p><p>
  If you're using memory debugging tools on a program that uses
  one of these pooling allocators, you can set the environment variable
  <code class="literal">GLIBCXX_FORCE_NEW</code> to keep extraneous pool allocation
  noise from cluttering debug information.
  For more details, see the
  <a class="link" href="mt_allocator.html" title="Chapter19.The mt_allocator">mt allocator</a>
  documentation and look specifically for <code class="code">GLIBCXX_FORCE_NEW</code>.
</p></div></div><div class="section"><div class="titlepage"><div><div><h3 class="title"><a id="debug.races"></a>Data Race Hunting</h3></div></div></div><p>
  All synchronization primitives used in the library internals need to be
  understood by race detectors so that they do not produce false reports.
</p><p>
  Two annotation macros are used to explain low-level synchronization
  to race detectors:
  <code class="code">_GLIBCXX_SYNCHRONIZATION_HAPPENS_BEFORE()</code> and
  <code class="code"> _GLIBCXX_SYNCHRONIZATION_HAPPENS_AFTER()</code>.
  By default, these macros are defined empty -- anyone who wants
  to use a race detector needs to redefine them to call an
  appropriate API.
  Since these macros are empty by default when the library is built,
  redefining them will only affect inline functions and template
  instantiations which are compiled in user code. This allows annotation
  of templates such as <code class="code">shared_ptr</code>, but not code which is
  only instantiated in the library.  Code which is only instantiated in
  the library needs to be recompiled with the annotation macros defined.
  That can be done by rebuilding the entire
  <code class="filename">libstdc++.so</code> file but a simpler
  alternative exists for ELF platforms such as GNU/Linux, because ELF
  symbol interposition allows symbols defined in the shared library to be
  overridden by symbols with the same name that appear earlier in the
  runtime search path. This means you only need to recompile the functions
  that are affected by the annotation macros, which can be done by
  recompiling individual files.
  Annotating <code class="code">std::string</code> and <code class="code">std::wstring</code>
  reference counting can be done by disabling extern templates (by defining
  <code class="code">_GLIBCXX_EXTERN_TEMPLATE=-1</code>) or by rebuilding the
  <code class="filename">src/string-inst.cc</code> file.
  Annotating the remaining atomic operations (at the time of writing these
  are in <code class="code">ios_base::Init::~Init</code>, <code class="code">locale::_Impl</code>,
  <code class="code">locale::facet</code> and <code class="code">thread::_M_start_thread</code>)
  requires rebuilding the relevant source files.
</p><p>
  The approach described above is known to work with the following race
  detection tools:
  <a class="link" href="https://valgrind.org/docs/manual/drd-manual.html" target="_top">
  DRD</a>,
  <a class="link" href="https://valgrind.org/docs/manual/hg-manual.html" target="_top">
  Helgrind</a>, and
  <a class="link" href="https://github.com/google/sanitizers" target="_top">
  ThreadSanitizer</a> (this refers to ThreadSanitizer v1, not the
  new "tsan" feature built-in to GCC itself).
</p><p>
  With DRD, Helgrind and ThreadSanitizer you will need to define
  the macros like this:
</p><pre class="programlisting">
  #define _GLIBCXX_SYNCHRONIZATION_HAPPENS_BEFORE(A) ANNOTATE_HAPPENS_BEFORE(A)
  #define _GLIBCXX_SYNCHRONIZATION_HAPPENS_AFTER(A)  ANNOTATE_HAPPENS_AFTER(A)
</pre><p>
  Refer to the documentation of each particular tool for details.
</p></div><div class="section"><div class="titlepage"><div><div><h3 class="title"><a id="debug.gdb"></a>Using <span class="command"><strong>gdb</strong></span></h3></div></div></div><p>
  </p><p>
  Many options are available for GDB itself: please see <a class="link" href="http://sourceware.org/gdb/current/onlinedocs/gdb/" target="_top">
  "GDB features for C++" </a> in the GDB documentation. Also
  recommended: the other parts of this manual.
</p><p>
  These settings can either be switched on in at the GDB command line,
  or put into a <code class="filename">.gdbinit</code> file to establish default
  debugging characteristics, like so:
</p><pre class="programlisting">
   set print pretty on
   set print object on
   set print static-members on
   set print vtbl on
   set print demangle on
   set demangle-style gnu-v3
</pre><p>
  Starting with version 7.0, GDB includes support for writing
  pretty-printers in Python.  Pretty printers for containers and other
  classes are distributed with GCC from version 4.5.0 and should be installed
  alongside the libstdc++ shared library files and found automatically by
  GDB.
</p><p>
  Depending where libstdc++ is installed, GDB might refuse to auto-load
  the python printers and print a warning instead.
  If this happens the python printers can be enabled by following the
  instructions GDB gives for setting your <code class="code">auto-load safe-path</code>
  in your <code class="filename">.gdbinit</code> configuration file.
</p><p>
  Once loaded, standard library classes that the printers support
  should print in a more human-readable format.  To print the classes
  in the old style, use the <strong class="userinput"><code>/r</code></strong> (raw) switch in the
  print command (i.e., <strong class="userinput"><code>print /r foo</code></strong>).  This will
  print the classes as if the Python pretty-printers were not loaded.
</p><p>
  For additional information on STL support and GDB please visit:
  <a class="link" href="http://sourceware.org/gdb/wiki/STLSupport" target="_top"> "GDB Support
  for STL" </a> in the GDB wiki.  Additionally, in-depth
  documentation and discussion of the pretty printing feature can be
  found in "Pretty Printing" node in the GDB manual.  You can find
  on-line versions of the GDB user manual in GDB's homepage, at
  <a class="link" href="http://sourceware.org/gdb/" target="_top"> "GDB: The GNU Project
  Debugger" </a>.
</p></div><div class="section"><div class="titlepage"><div><div><h3 class="title"><a id="debug.exceptions"></a>Tracking uncaught exceptions</h3></div></div></div><p>
  The <a class="link" href="termination.html#support.termination.verbose" title="Verbose Terminate Handler">verbose
  termination handler</a> gives information about uncaught
  exceptions which kill the program.
</p></div><div class="section"><div class="titlepage"><div><div><h3 class="title"><a id="debug.debug_mode"></a>Debug Mode</h3></div></div></div><p> The <a class="link" href="debug_mode.html" title="Chapter17.Debug Mode">Debug Mode</a>
  has compile and run-time checks for many containers.
  </p></div><div class="section"><div class="titlepage"><div><div><h3 class="title"><a id="debug.compile_time_checks"></a>Compile Time Checking</h3></div></div></div><p> The <a class="link" href="ext_compile_checks.html" title="Chapter16.Compile Time Checks">Compile-Time
  Checks</a> extension has compile-time checks for many algorithms.
  </p></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="using_exceptions.html">Prev</a></td><td width="20%" align="center"><a accesskey="u" href="using.html">Up</a></td><td width="40%" align="right"><a accesskey="n" href="std_contents.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">Exceptions</td><td width="20%" align="center"><a accesskey="h" href="../index.html">Home</a></td><td width="40%" align="right" valign="top">PartII.
    Standard Contents
  </td></tr></table></div></body></html>