<page xmlns="http://projectmallard.org/1.0/"
xmlns:its="http://www.w3.org/2005/11/its"
type="topic"
id="memory-management">
<info>
<link type="guide" xref="index#general-guidelines"/>
<credit type="author copyright">
<name>Philip Withnall</name>
<email its:translate="no">philip.withnall@collabora.co.uk</email>
<years>2015</years>
</credit>
<include href="cc-by-sa-3-0.xml" xmlns="http://www.w3.org/2001/XInclude"/>
<desc>Managing memory allocation and deallocation in C</desc>
</info>
<title>Memory Management</title>
<comment>
<p>
Note: GSlice is deliberately not mentioned on this page because we don’t
want to encourage its use. GLib developers are moving towards deprecating
it due to its performance slipping significantly behind the libc
allocator’s on Linux in recent years.
</p>
</comment>
<comment>
<p>
FIXME:
</p>
<list>
<item><p>
g_autoptr() (https://bugzilla.gnome.org/show_bug.cgi?id=743640)
</p></item>
<item><p>
Reference counted memory areas
(https://bugzilla.gnome.org/show_bug.cgi?id=622721)
</p></item>
<item><p>
Integrate
https://tecnocode.co.uk/2013/09/03/const-gchar-vs-gchar-and-other-memory-management-stories/
</p></item>
</list>
</comment>
<p>
The GNOME stack is predominantly written in C, so dynamically allocated
memory has to be managed manually. Through use of GLib convenience APIs,
memory management can be trivial, but programmers always need to keep memory
in mind when writing code.
</p>
<p>
It is assumed that the reader is familiar with the idea of heap allocation
of memory using <code>malloc()</code> and <code>free()</code>, and knows of
the paired GLib equivalents, <code>g_malloc()</code> and
<code>g_free()</code>.
</p>
<synopsis>
<title>Summary</title>
<p>
There are three situations to avoid, in order of descending importance:
</p>
<list type="numbered">
<item><p>Using memory after freeing it (use-after-free).</p></item>
<item><p>Using memory before allocating it.</p></item>
<item><p>Not freeing memory after allocating it (leaking).</p></item>
</list>
<p>
Key principles, in no particular order:
</p>
<list>
<item><p>
Determine and document whether each variable is owned or unowned. They
must never change from one to the other at runtime.
(<link xref="#principles"/>)
</p></item>
<item><p>
Determine and document the ownership transfers at function boundaries.
(<link xref="#principles"/>)
</p></item>
<item><p>
Ensure that each assignment, function call and function return respects
the relevant ownership transfers. (<link xref="#assignments"/>,
<link xref="#function-calls"/>, <link xref="#function-returns"/>)
</p></item>
<item><p>
Use reference counting rather than explicit finalization where possible.
(<link xref="#reference-counting"/>)
</p></item>
<item><p>
Use GLib convenience functions like
<link xref="#g-clear-object"><code>g_clear_object()</code></link> where
possible. (<link xref="#convenience-functions"/>)
</p></item>
<item><p>
Do not split memory management across code paths.
(<link xref="#principles"/>)
</p></item>
<item><p>
Use the single-path cleanup pattern for large or complex functions.
(<link xref="#single-path-cleanup"/>)
</p></item>
<item><p>
Leaks should be checked for using Valgrind or the address sanitizer.
(<link xref="#verification"/>)
</p></item>
</list>
</synopsis>
<section id="principles">
<title>Principles of Memory Management</title>
<p>
The normal approach to memory management is for the programmer to keep
track of which variables point to allocated memory, and to manually free
them when they are no longer needed. This is correct, but can be clarified
by introducing the concept of <em>ownership</em>, which is the piece of
code (such as a function, struct or object) which is responsible for
freeing a piece of allocated memory (an <em>allocation</em>). Each
allocation has exactly one owner; this owner may change as the program
runs, by <em>transferring</em> ownership to another piece of code. Each
variable is <em>owned</em> or <em>unowned</em>, according to whether the
scope containing it is always its owner. Each function parameter and
return type either transfers ownership of the values passed to it, or it
doesn’t. If code which owns some memory doesn’t deallocate that memory,
that’s a memory leak. If code which doesn’t own some memory frees it,
that’s a double-free. Both are bad.
</p>
<p>
By statically calculating which variables are owned, memory
management becomes a simple task of unconditionally freeing the owned
variables before they leave their scope, and <em>not</em> freeing the
unowned variables (see <link xref="#single-path-cleanup"/>). The key
question to answer for all memory is: which code has ownership of this
memory?
</p>
<p>
There is an important restriction here: variables must
<em style="strong">never</em> change from owned to unowned (or vice-versa)
at runtime. This restriction is key to simplifying memory management.
</p>
<p>
For example, consider the functions:
</p>
<code mime="text/x-csrc">gchar *generate_string (const gchar *template);
void print_string (const gchar *str);</code>
<p>
The following code has been annotated to note where the ownership
transfers happen:
</p>
<code mime="text/x-csrc">gchar *my_str = NULL; /* owned */
const gchar *template; /* unowned */
GValue value = G_VALUE_INIT; /* owned */
g_value_init (&value, G_TYPE_STRING);
/* Transfers ownership of a string from the function to the variable. */
template = "XXXXXX";
my_str = generate_string (template);
/* No ownership transfer. */
print_string (my_str);
/* Transfer ownership. We no longer have to free @my_str. */
g_value_take_string (&value, my_str);
/* We still have ownership of @value, so free it before it goes out of scope. */
g_value_unset (&value);</code>
<p>
There are a few points here: Firstly, the ‘owned’ comments by the variable
declarations denote that those variables are owned by the local scope, and
hence need to be freed before they go out of scope. The alternative is
‘unowned’, which means the local scope does <em>not</em> have ownership,
and <em>must not</em> free the variables before going out of scope.
Similarly, ownership <em>must not</em> be transferred to them on
assignment.
</p>
<p>
Secondly, the variable type modifiers reflect whether they transfer
ownership: because <code>my_str</code> is owned by the local scope, it has
type <code>gchar</code>, whereas <code>template</code> is
<code>const</code> to denote it is unowned. Similarly, the
<code>template</code> parameter of <code>generate_string()</code> and the
<code>str</code> parameter of <code>print_string()</code> are
<code>const</code> because no ownership is transferred when those
functions are called. As ownership <em>is</em> transferred for the string
parameter of <code>g_value_take_string()</code>, we can expect its type to
be <code>gchar</code>.
</p>
<p>
(Note that this is not the case for
<link href="https://developer.gnome.org/gobject/stable/gobject-The-Base-Object-Type.html">
<code>GObject</code></link>s and subclasses, which can never be
<code>const</code>. It is only the case for strings and simple
<code>struct</code>s.)
</p>
<p>
Finally, a few libraries use a function naming convention to indicate
ownership transfer, for example using ‘take’ in a function name to
indicate full transfer of parameters, as with
<code>g_value_take_string()</code>. Note that different libraries use
different conventions, as shown below:
</p>
<table shade="rows cols">
<colgroup><col/></colgroup>
<colgroup><col/><col/><col/></colgroup>
<thead>
<tr>
<td><p>Function name</p></td>
<td><p>Convention 1 (standard)</p></td>
<td><p>Convention 2 (alternate)</p></td> <!-- get for everything -->
<td><p>Convention 3 (<cmd>gdbus-codegen</cmd>)</p></td>
</tr>
</thead>
<tbody>
<tr>
<td><p>get</p></td>
<td><p>No transfer</p></td>
<td><p>Any transfer</p></td>
<td><p>Full transfer</p></td>
</tr>
<tr>
<td><p>dup</p></td>
<td><p>Full transfer</p></td>
<td><p>Unused</p></td>
<td><p>Unused</p></td>
</tr>
<tr>
<td><p>peek</p></td>
<td><p>Unused</p></td>
<td><p>Unused</p></td>
<td><p>No transfer</p></td>
</tr>
<tr>
<td><p>set</p></td>
<td><p>No transfer</p></td>
<td><p>No transfer</p></td>
<td><p>No transfer</p></td>
</tr>
<tr>
<td><p>take</p></td>
<td><p>Full transfer</p></td>
<td><p>Unused</p></td>
<td><p>Unused</p></td>
</tr>
<tr>
<td><p>steal</p></td>
<td><p>Full transfer</p></td>
<td><p>Full transfer</p></td>
<td><p>Full transfer</p></td>
</tr>
</tbody>
</table>
<p>
Ideally, all functions have a <code>(transfer)</code>
<link xref="introspection">introspection annotation</link> for all
relevant parameters and the return value. Failing that, here is a set of
guidelines to use to determine whether ownership of a return value is
transferred:
</p>
<steps>
<item><p>
If the type has an introspection <code>(transfer)</code> annotation,
look at that.
</p></item>
<item><p>
Otherwise, if the type is <code>const</code>, there is no transfer.
</p></item>
<item><p>
Otherwise, if the function documentation explicitly specifies the return
value must be freed, there is full or container transfer.
</p></item>
<item><p>
Otherwise, if the function is named ‘dup’, ‘take’ or ‘steal’, there is
full or container transfer.
</p></item>
<item><p>
Otherwise, if the function is named ‘peek’, there is no transfer.
</p></item>
<item><p>
Otherwise, you need to look at the function’s code to determine whether
it intends ownership to be transferred. Then file a bug against the
documentation for that function, and ask for an introspection annotation
to be added.
</p></item>
</steps>
<p>
Given this ownership and transfer infrastructure, the correct approach to
memory allocation can be mechanically determined for each situation. In
each case, the <code>copy()</code> function must be appropriate to the
data type, for example <code>g_strdup()</code> for strings, or
<code>g_object_ref()</code> for GObjects.
</p>
<p>
When thinking about ownership transfer,
<code>malloc()</code>/<code>free()</code> and reference counting are
equivalent: in the former case, a newly allocated piece of heap memory is
transferred; in the latter, a newly incremented reference.
See <link xref="#reference-counting"/>.
</p>
<section id="assignments">
<title>Assignments</title>
<table shade="rows colgroups">
<colgroup><col/></colgroup>
<colgroup><col/><col/></colgroup>
<thead>
<tr>
<td><p>Assignment from/to</p></td>
<td><p>Owned destination</p></td>
<td><p>Unowned destination</p></td>
</tr>
</thead>
<tbody>
<tr>
<td><p>Owned source</p></td>
<td>
<p>
Copy or move the source to the destination.
</p>
<code>owned_dest = copy (owned_src)</code>
<code>owned_dest = owned_src; owned_src = NULL</code>
</td>
<td>
<p>
Pure assignment, assuming the unowned variable is not used after
the owned one is freed.
</p>
<code>unowned_dest = owned_src</code>
</td>
</tr>
<tr>
<td><p>Unowned source</p></td>
<td>
<p>Copy the source to the destination.</p>
<code>owned_dest = copy (unowned_src)</code>
</td>
<td>
<p>Pure assignment.</p>
<code>unowned_dest = unowned_src</code>
</td>
</tr>
</tbody>
</table>
</section>
<section id="function-calls">
<title>Function Calls</title>
<table shade="rows colgroups">
<colgroup><col/></colgroup>
<colgroup><col/><col/></colgroup>
<thead>
<tr>
<td><p>Call from/to</p></td>
<td><p>Transfer full parameter</p></td>
<td><p>Transfer none parameter</p></td>
</tr>
</thead>
<tbody>
<tr>
<td><p>Owned source</p></td>
<td>
<p>
Copy or move the source for the parameter.
</p>
<code>function_call (copy (owned_src))</code>
<code>function_call (owned_src); owned_src = NULL</code>
</td>
<td>
<p>
Pure parameter passing.
</p>
<code>function_call (owned_src)</code>
</td>
</tr>
<tr>
<td><p>Unowned source</p></td>
<td>
<p>Copy the source for the parameter.</p>
<code>function_call (copy (unowned_src))</code>
</td>
<td>
<p>Pure parameter passing.</p>
<code>function_call (unowned_src)</code>
</td>
</tr>
</tbody>
</table>
</section>
<section id="function-returns">
<title>Function Returns</title>
<table shade="rows colgroups">
<colgroup><col/></colgroup>
<colgroup><col/><col/></colgroup>
<thead>
<tr>
<td><p>Return from/to</p></td>
<td><p>Transfer full return</p></td>
<td><p>Transfer none return</p></td>
</tr>
</thead>
<tbody>
<tr>
<td><p>Owned source</p></td>
<td>
<p>
Pure variable return.
</p>
<code>return owned_src</code>
</td>
<td>
<p>
Invalid. The source needs to be freed, so the return value would
use freed memory — a use-after-free error.
</p>
</td>
</tr>
<tr>
<td><p>Unowned source</p></td>
<td>
<p>Copy the source for the return.</p>
<code>return copy (unowned_src)</code>
</td>
<td>
<p>Pure variable passing.</p>
<code>return unowned_src</code>
</td>
</tr>
</tbody>
</table>
</section>
</section>
<section id="documentation">
<title>Documentation</title>
<p>
Documenting the ownership transfer for each function parameter and return,
and the ownership for each variable, is important. While they may be clear
when writing the code, they are not clear a few months later; and may
never be clear to users of an API. They should always be documented.
</p>
<p>
The best way to document ownership transfer is to use the
<link href="https://wiki.gnome.org/Projects/GObjectIntrospection/Annotations#Memory_and_lifecycle_management">
<code>(transfer)</code></link> annotation introduced by
<link xref="introspection">gobject-introspection</link>. Include this in
the API documentation comment for each function parameter and return type.
If a function is not public API, write a documentation comment for it
anyway and include the <code>(transfer)</code> annotations. By doing so,
the introspection tools can also read the annotations and use them to
correctly introspect the API.
</p>
<p>
For example:
</p>
<code mime="text/x-csrc">/**
* g_value_take_string:
* @value: (transfer none): an initialized #GValue
* @str: (transfer full): string to set it to
*
* Function documentation goes here.
*/
/**
* generate_string:
* @template: (transfer none): a template to follow when generating the string
*
* Function documentation goes here.
*
* Returns: (transfer full): a newly generated string
*/</code>
<p>
Ownership for variables can be documented using inline comments. These are
non-standard, and not read by any tools, but can form a convention if used
consistently.
</p>
<code mime="text/x-csrc">GObject *some_owned_object = NULL; /* owned */
GObject *some_unowned_object; /* unowned */</code>
<p>
The documentation for <link xref="#container-types"/> is similarly only a
convention; it includes the type of the contained elements too:
</p>
<code mime="text/x-csrc">GPtrArray/*<owned gchar*>*/ *some_unowned_string_array; /* unowned */
GPtrArray/*<owned gchar*>*/ *some_owned_string_array = NULL; /* owned */
GPtrArray/*<unowned GObject*>*/ *some_owned_object_array = NULL; /* owned */</code>
<p>
Note also that owned variables should always be initialized so that
freeing them is more convenient. See
<link xref="#convenience-functions"/>.
</p>
<p>
Also note that some types, for example basic C types like strings, can
have the <code>const</code> modifier added if they are unowned, to take
advantage of compiler warnings resulting from assigning those variables to
owned variables (which must <em>not</em> use the <code>const</code>
modifier). If so, the <code>/* unowned */</code> comment may be omitted.
</p>
</section>
<section id="reference-counting">
<title>Reference Counting</title>
<p>
As well as conventional <code>malloc()</code>/<code>free()</code>-style
types, GLib has various reference counted types —
<link href="https://developer.gnome.org/gobject/stable/gobject-The-Base-Object-Type.html">
<code>GObject</code></link> being a prime example.
</p>
<p>
The concepts of ownership and transfer apply just as well to reference
counted types as they do to allocated types. A scope <em>owns</em> a
reference counted type if it holds a strong reference to the instance
(for example by calling
<link href="https://developer.gnome.org/gobject/stable/gobject-The-Base-Object-Type.html#g-object-ref">
<code>g_object_ref()</code></link>). An instance can be ‘copied’ by
calling <code>g_object_ref()</code> again. Ownership can be freed with
<link href="https://developer.gnome.org/gobject/stable/gobject-The-Base-Object-Type.html#g-object-unref">
<code>g_object_unref()</code></link> — even though this may not actually
finalize the instance, it frees the current scope’s ownership of that
instance.
</p>
<p>
See <link xref="#g-clear-object"/> for a convenient way of handling
GObject references.
</p>
<p>
There are other reference counted types in GLib, such as
<link href="https://developer.gnome.org/glib/stable/glib-Hash-Tables.html">
<code>GHashTable</code></link> (using
<link href="https://developer.gnome.org/glib/stable/glib-Hash-Tables.html#g-hash-table-ref">
<code>g_hash_table_ref()</code></link> and
<link href="https://developer.gnome.org/glib/stable/glib-Hash-Tables.html#g-hash-table-unref">
<code>g_hash_table_unref()</code></link>), or
<link href="https://developer.gnome.org/glib/stable/glib-GVariant.html">
<code>GVariant</code></link>
(<link href="https://developer.gnome.org/glib/stable/glib-GVariant.html#g-variant-ref">
<code>g_variant_ref()</code></link>,
<link href="https://developer.gnome.org/glib/stable/glib-GVariant.html#g-variant-unref">
<code>g_variant_unref()</code></link>). Some types, like
<code>GHashTable</code>, support both reference counting and explicit
finalization. Reference counting should always be used in preference,
because it allows instances to be easily shared between multiple scopes
(each holding their own reference) without having to allocate multiple
copies of the instance. This saves memory.
</p>
<section id="floating-references">
<title>Floating References</title>
<p>
Classes which are derived from
<link href="https://developer.gnome.org/gobject/stable/gobject-The-Base-Object-Type.html#GInitiallyUnowned"><code>GInitiallyUnowned</code></link>,
as opposed to
<link href="https://developer.gnome.org/gobject/stable/gobject-The-Base-Object-Type.html#GObject-struct"><code>GObject</code></link>
have an initial reference which is <em>floating</em>, meaning that no
code owns the reference. As soon as
<link href="https://developer.gnome.org/gobject/stable/gobject-The-Base-Object-Type.html#g-object-ref-sink"><code>g_object_ref_sink()</code></link>
is called on the object, the floating reference is converted to a strong
reference, and the calling code assumes ownership of the object.
</p>
<p>
Floating references are a convenience for use in C in APIs, such as
GTK+, where large numbers of objects must be created and organized into
a hierarchy. In these cases, calling <code>g_object_unref()</code> to
drop all the strong references would result in a lot of code.
</p>
<example>
<p>
Floating references allow the following code to be simplified:
</p>
<code mime="text/x-csrc" style="invalid">GtkWidget *new_widget;
new_widget = gtk_some_widget_new ();
gtk_container_add (some_container, new_widget);
g_object_unref (new_widget);</code>
<p>
Instead, the following code can be used, with the
<code>GtkContainer</code> assuming ownership of the floating
reference:
</p>
<code mime="text/x-csrc" style="valid">
gtk_container_add (some_container, gtk_some_widget_new ());</code>
</example>
<p>
Floating references are only used by a few APIs — in particular,
<code>GtkWidget</code> and all its subclasses. You must learn which APIs
support it, and which APIs consume floating references, and only use
them together.
</p>
<p>
Note that <code>g_object_ref_sink()</code> is equivalent to
<code>g_object_ref()</code> when called on a non-floating reference,
making <code>gtk_container_add()</code> no different from any other
function in such cases.
</p>
<p>
See the <link href="https://developer.gnome.org/gobject/stable/gobject-The-Base-Object-Type.html#floating-ref">GObject
manual</link> for more information on floating references.
</p>
</section>
</section>
<section id="convenience-functions">
<title>Convenience Functions</title>
<p>
GLib provides various convenience functions for memory management,
especially for GObjects. Three will be covered here, but others exist —
check the GLib API documentation for more. They typically follow similar
naming schemas to these three (using ‘_full’ suffixes, or the verb ‘clear’
in the function name).
</p>
<section id="g-clear-object">
<title><code>g_clear_object()</code></title>
<p>
<link href="https://developer.gnome.org/gobject/stable/gobject-The-Base-Object-Type.html#g-clear-object">
<code>g_clear_object()</code></link> is a version of
<link href="https://developer.gnome.org/gobject/stable/gobject-The-Base-Object-Type.html#g-object-unref">
<code>g_object_unref()</code></link> which unrefs a GObject and then
clears the pointer to it to <code>NULL</code>.
</p>
<p>
This makes it easier to implement code that guarantees a GObject pointer
is always either <code>NULL</code>, or has ownership of a GObject (but
which never points to a GObject it no longer owns).
</p>
<p>
By initialising all owned GObject pointers to <code>NULL</code>, freeing
them at the end of the scope is as simple as calling
<code>g_clear_object()</code> without any checks, as discussed in
<link xref="#single-path-cleanup"/>:
</p>
<code mime="text/x-csrc" style="valid">void
my_function (void)
{
GObject *some_object = NULL; /* owned */
if (rand ())
{
some_object = create_new_object ();
/* do something with the object */
}
g_clear_object (&some_object);
}</code>
</section>
<section id="g-list-free-full">
<title><code>g_list_free_full()</code></title>
<p>
<link href="https://developer.gnome.org/glib/stable/glib-Doubly-Linked-Lists.html#g-list-free-full">
<code>g_list_free_full()</code></link> frees all the elements in a
linked list, <em>and all their data</em>. It is much more convenient
than iterating through the list to free all the elements’ data, then
calling <link href="https://developer.gnome.org/glib/stable/glib-Doubly-Linked-Lists.html#g-list-free">
<code>g_list_free()</code></link> to free the <code>GList</code>
elements themselves.
</p>
</section>
<section id="g-hash-table-new-full">
<title><code>g_hash_table_new_full()</code></title>
<p>
<link href="https://developer.gnome.org/glib/stable/glib-Hash-Tables.html#g-hash-table-new-full">
<code>g_hash_table_new_full()</code></link> is a newer version of
<link href="https://developer.gnome.org/glib/stable/glib-Hash-Tables.html#g-hash-table-new">
<code>g_hash_table_new()</code></link> which allows setting functions to
destroy each key and value in the hash table when they are removed.
These functions are then automatically called for all keys and values
when the hash table is destroyed, or when an entry is removed using
<code>g_hash_table_remove()</code>.
</p>
<p>
Essentially, it simplifies memory management of keys and values to the
question of whether they are present in the hash table. See
<link xref="#container-types"/> for a discussion on ownership of
elements within container types.
</p>
<p>
A similar function exists for <code>GPtrArray</code>:
<link href="https://developer.gnome.org/glib/stable/glib-Pointer-Arrays.html#g-ptr-array-new-with-free-func">
<code>g_ptr_array_new_with_free_func()</code></link>.
</p>
</section>
</section>
<section id="container-types">
<title>Container Types</title>
<p>
When using container types, such as <code>GPtrArray</code> or
<code>GList</code>, an additional level of ownership is introduced: as
well as the ownership of the container instance, each element in the
container is either owned or unowned too. By nesting containers, multiple
levels of ownership must be tracked. Ownership of owned elements belongs
to the container; ownership of the container belongs to the scope it’s in
(which may be another container).
</p>
<p>
A key principle for simplifying this is to ensure that all elements in a
container have the same ownership: they are either all owned, or all
unowned. This happens automatically if the normal
<link xref="#convenience-functions"/> are used for types like
<code>GPtrArray</code> and <code>GHashTable</code>.
</p>
<p>
If elements in a container are <em>owned</em>, adding them to the
container is essentially an ownership transfer. For example, for an array
of strings, if the elements are owned, the definition of
<code>g_ptr_array_add()</code> is effectively:
</p>
<code mime="text/x-csrc">/**
* g_ptr_array_add:
* @array: a #GPtrArray
* @str: (transfer full): string to add
*/
void
g_ptr_array_add (GPtrArray *array,
gchar *str);</code>
<p>
So, for example, constant (unowned) strings must be added to the array
using <code>g_ptr_array_add (array, g_strdup ("constant string"))</code>.
</p>
<p>
Whereas if the elements are unowned, the definition is effectively:
</p>
<code mime="text/x-csrc">/**
* g_ptr_array_add:
* @array: a #GPtrArray
* @str: (transfer none): string to add
*/
void
g_ptr_array_add (GPtrArray *array,
const gchar *str);</code>
<p>
Here, constant strings can be added without copying them:
<code>g_ptr_array_add (array, "constant string")</code>.
</p>
<p>
See <link xref="#documentation"/> for examples of comments to add to
variable definitions to annotate them with the element type and ownership.
</p>
</section>
<section id="single-path-cleanup">
<title>Single-Path Cleanup</title>
<p>
A useful design pattern for more complex functions is to have a single
control path which cleans up (frees) allocations and returns to the
caller. This vastly simplifies tracking of allocations, as it’s no longer
necessary to mentally work out which allocations have been freed on each
code path — all code paths end at the same point, so perform all the frees
then. The benefits of this approach rapidly become greater for larger
functions with more owned local variables; it may not make sense to apply
the pattern to smaller functions.
</p>
<p>
This approach has two requirements:
</p>
<list type="numbered">
<item><p>
The function returns from a single point, and uses <code>goto</code> to
reach that point from other paths.
</p></item>
<item><p>
All owned variables are set to <code>NULL</code> when initialized or
when ownership is transferred away from them.
</p></item>
</list>
<p>
The example below is for a small function (for brevity), but should
illustrate the principles for application of the pattern to larger
functions:
</p>
<listing>
<title>Single-Path Cleanup Example</title>
<desc>
Example of implementing single-path cleanup for a simple function
</desc>
<code mime="text/x-csrc">GObject *
some_function (GError **error)
{
gchar *some_str = NULL; /* owned */
GObject *temp_object = NULL; /* owned */
const gchar *temp_str;
GObject *my_object = NULL; /* owned */
GError *child_error = NULL; /* owned */
temp_object = generate_object ();
temp_str = "example string";
if (rand ())
{
some_str = g_strconcat (temp_str, temp_str, NULL);
}
else
{
some_operation_which_might_fail (&child_error);
if (child_error != NULL)
{
goto done;
}
my_object = generate_wrapped_object (temp_object);
}
done:
/* Here, @some_str is either NULL or a string to be freed, so can be passed to
* g_free() unconditionally.
*
* Similarly, @temp_object is either NULL or an object to be unreffed, so can
* be passed to g_clear_object() unconditionally. */
g_free (some_str);
g_clear_object (&temp_object);
/* The pattern can also be used to ensure that the function always returns
* either an error or a return value (but never both). */
if (child_error != NULL)
{
g_propagate_error (error, child_error);
g_clear_object (&my_object);
}
return my_object;
}</code>
</listing>
</section>
<section id="verification">
<title>Verification</title>
<p>
Memory leaks can be checked for in two ways: static analysis, and runtime
leak checking.
</p>
<p>
Static analysis with tools like
<link xref="tooling#coverity">Coverity</link>, the
<link xref="tooling#clang-static-analyzer">Clang static analyzer</link> or
<link xref="tooling#tartan">Tartan</link> can
catch some leaks, but require knowledge of the ownership transfer of every
function called in the code. Domain-specific static analyzers like Tartan
(which knows about GLib memory allocation and transfer) can perform better
here, but Tartan is quite a young project and still misses things (a low
true positive rate). It is recommended that code be put through a static
analyzer, but the primary tool for detecting leaks should be runtime leak
checking.
</p>
<p>
Runtime leak checking is done using
<link xref="tooling#valgrind">Valgrind</link>, using its
<link xref="tooling#memcheck">memcheck</link> tool. Any leak it detects as
‘definitely losing memory’ should be fixed. Many of the leaks which
‘potentially’ lose memory are not real leaks, and should be added to the
suppression file.
</p>
<p>
If compiling with a recent version of Clang or GCC, the
<link xref="tooling#address-sanitizer">address sanitizer</link> can be
enabled instead, and it will detect memory leaks and overflow problems at
runtime, but without the difficulty of running Valgrind in the right
environment. Note, however, that it is still a young tool, so may fail in
some cases.
</p>
<p>
See <link xref="tooling#valgrind"/> for more information on using
Valgrind.
</p>
</section>
</page>