defensive-coding-guide/defensive-coding/en-US/CXX-Language.xml
Florian Weimer 8350ea4f61 Use unique XML source file names (#1000781)
This helps Transifex translations.
2013-08-26 15:25:46 +02:00

186 lines
7.1 KiB
XML

<?xml version='1.0' encoding='utf-8' ?>
<!DOCTYPE section PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN" "http://www.oasis-open.org/docbook/xml/4.5/docbookx.dtd" [
]>
<section id="sect-Defensive_Coding-CXX-Language">
<title>The core language</title>
<para>
C++ includes a large subset of the C language. As far as the C
subset is used, the recommendations in <xref
linkend="chap-Defensive_Coding-C"/> apply.
</para>
<section>
<title>Array allocation with <literal>operator new[]</literal></title>
<para>
For very large values of <literal>n</literal>, an expression
like <literal>new T[n]</literal> can return a pointer to a heap
region which is too small. In other words, not all array
elements are actually backed with heap memory reserved to the
array. Current GCC versions generate code that performs a
computation of the form <literal>sizeof(T) * size_t(n) +
cookie_size</literal>, where <literal>cookie_size</literal> is
currently at most 8. This computation can overflow, and
GCC-generated code does not detect this.
</para>
<para>
The <literal>std::vector</literal> template can be used instead
an explicit array allocation. (The GCC implementation detects
overflow internally.)
</para>
<para>
If there is no alternative to <literal>operator new[]</literal>,
code which allocates arrays with a variable length must check
for overflow manually. For the <literal>new T[n]</literal>
example, the size check could be <literal>n || (n > 0 &amp;&amp;
n &gt; (size_t(-1) - 8) / sizeof(T))</literal>. (See <xref
linkend="sect-Defensive_Coding-C-Arithmetic"/>.) If there are
additional dimensions (which must be constants according to the
C++ standard), these should be included as factors in the
divisor.
</para>
<para>
These countermeasures prevent out-of-bounds writes and potential
code execution. Very large memory allocations can still lead to
a denial of service. <xref
linkend="sect-Defensive_Coding-Tasks-Serialization-Decoders"/>
contains suggestions for mitigating this problem when processing
untrusted data.
</para>
<para>
See <xref linkend="sect-Defensive_Coding-C-Allocators-Arrays"/>
for array allocation advice for C-style memory allocation.
</para>
</section>
<section>
<title>Overloading</title>
<para>
Do not overload functions with versions that have different
security characteristics. For instance, do not implement a
function <function>strcat</function> which works on
<type>std::string</type> arguments. Similarly, do not name
methods after such functions.
</para>
</section>
<section>
<title>ABI compatibility and preparing for security updates</title>
<para>
A stable binary interface (ABI) is vastly preferred for security
updates. Without a stable ABI, all reverse dependencies need
recompiling, which can be a lot of work and could even be
impossible in some cases. Ideally, a security update only
updates a single dynamic shared object, and is picked up
automatically after restarting affected processes.
</para>
<para>
Outside of extremely performance-critical code, you should
ensure that a wide range of changes is possible without breaking
ABI. Some very basic guidelines are:
</para>
<itemizedlist>
<listitem>
<para>
Avoid inline functions.
</para>
</listitem>
<listitem>
<para>
Use the pointer-to-implementation idiom.
</para>
</listitem>
<listitem>
<para>
Try to avoid templates. Use them if the increased type
safety provides a benefit to the programmer.
</para>
</listitem>
<listitem>
<para>
Move security-critical code out of templated code, so that
it can be patched in a central place if necessary.
</para>
</listitem>
</itemizedlist>
<para>
The KDE project publishes a document with more extensive
guidelines on ABI-preserving changes to C++ code, <ulink
url="http://techbase.kde.org/Policies/Binary_Compatibility_Issues_With_C++">Policies/Binary
Compatibility Issues With C++</ulink>
(<emphasis>d-pointer</emphasis> refers to the
pointer-to-implementation idiom).
</para>
</section>
<section id="sect-Defensive_Coding-CXX-Language-CXX11">
<title>C++0X and C++11 support</title>
<para>
GCC offers different language compatibility modes:
</para>
<itemizedlist>
<listitem>
<para>
<option>-std=c++98</option> for the original 1998 C++
standard
</para>
</listitem>
<listitem>
<para>
<option>-std=c++03</option> for the 1998 standard with the
changes from the TR1 technical report
</para>
</listitem>
<listitem>
<para>
<option>-std=c++11</option> for the 2011 C++ standard. This
option should not be used.
</para>
</listitem>
<listitem>
<para>
<option>-std=c++0x</option> for several different versions
of C++11 support in development, depending on the GCC
version. This option should not be used.
<!-- There were two incompatibilies before GCC 4.7.2
(std::list and std::pair), but link C++98 and C++11
code is still unsupported, although it currently has
some chance of working by accident. -->
</para>
</listitem>
</itemizedlist>
<para>
For each of these flags, there are variants which also enable
GNU extensions (mostly language features also found in C99 or
C11): <option>-std=gnu++98</option>,
<option>-std=gnu++03</option>, <option>-std=gnu++11</option>.
Again, <option>-std=gnu++11</option> should not be used.
</para>
<para>
If you enable C++11 support, the ABI of the standard C++ library
<literal>libstdc++</literal> will change in subtle ways.
Currently, no C++ libraries are compiled in C++11 mode, so if
you compile your code in C++11 mode, it will be incompatible
with the rest of the system. Unfortunately, this is also the
case if you do not use any C++11 features. Currently, there is
no safe way to enable C++11 mode (except for freestanding
applications).
</para>
<para>
The meaning of C++0X mode changed from GCC release to GCC
release. Earlier versions were still ABI-compatible with C++98
mode, but in the most recent versions, switching to C++0X mode
activates C++11 support, with its compatibility problems.
</para>
<para>
Some C++11 features (or approximations thereof) are available
with TR1 support, that is, with <option>-std=c++03</option> or
<option>-std=gnu++03</option> and in the
<literal>&lt;tr1/*&gt;</literal> header files. This includes
<literal>std::tr1::shared_ptr</literal> (from
<literal>&lt;tr1/memory&gt;</literal>) and
<literal>std::tr1::function</literal> (from
<literal>&lt;tr1/functional&gt;</literal>). For other C++11
features, the Boost C++ library contains replacements.
</para>
</section>
</section>