<?xml version="1.0"?> <!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN" "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd"> <refentry> <refmeta> <refentrytitle>talloc</refentrytitle> <manvolnum>3</manvolnum> </refmeta> <refnamediv> <refname>talloc</refname> <refpurpose>hierarchical reference counted memory pool system with destructors</refpurpose> </refnamediv> <refsynopsisdiv> <synopsis>#include <talloc/talloc.h></synopsis> </refsynopsisdiv> <refsect1><title>DESCRIPTION</title> <para> If you are used to talloc from Samba3 then please read this carefully, as talloc has changed a lot. </para> <para> The new talloc is a hierarchical, reference counted memory pool system with destructors. Quite a mouthful really, but not too bad once you get used to it. </para> <para> Perhaps the biggest change from Samba3 is that there is no distinction between a "talloc context" and a "talloc pointer". Any pointer returned from talloc() is itself a valid talloc context. This means you can do this: </para> <programlisting> struct foo *X = talloc(mem_ctx, struct foo); X->name = talloc_strdup(X, "foo"); </programlisting> <para> and the pointer <literal role="code">X->name</literal> would be a "child" of the talloc context <literal role="code">X</literal> which is itself a child of <literal role="code">mem_ctx</literal>. So if you do <literal role="code">talloc_free(mem_ctx)</literal> then it is all destroyed, whereas if you do <literal role="code">talloc_free(X)</literal> then just <literal role="code">X</literal> and <literal role="code">X->name</literal> are destroyed, and if you do <literal role="code">talloc_free(X->name)</literal> then just the name element of <literal role="code">X</literal> is destroyed. </para> <para> If you think about this, then what this effectively gives you is an n-ary tree, where you can free any part of the tree with talloc_free(). </para> <para> If you find this confusing, then I suggest you run the <literal role="code">testsuite</literal> program to watch talloc in action. You may also like to add your own tests to <literal role="code">testsuite.c</literal> to clarify how some particular situation is handled. </para> </refsect1> <refsect1><title>TALLOC API</title> <para> The following is a complete guide to the talloc API. Read it all at least twice. </para> <refsect2><title>(type *)talloc(const void *ctx, type);</title> <para> The talloc() macro is the core of the talloc library. It takes a memory <emphasis role="italic">ctx</emphasis> and a <emphasis role="italic">type</emphasis>, and returns a pointer to a new area of memory of the given <emphasis role="italic">type</emphasis>. </para> <para> The returned pointer is itself a talloc context, so you can use it as the <emphasis role="italic">ctx</emphasis> argument to more calls to talloc() if you wish. </para> <para> The returned pointer is a "child" of the supplied context. This means that if you talloc_free() the <emphasis role="italic">ctx</emphasis> then the new child disappears as well. Alternatively you can free just the child. </para> <para> The <emphasis role="italic">ctx</emphasis> argument to talloc() can be NULL, in which case a new top level context is created. </para> </refsect2> <refsect2><title>void *talloc_size(const void *ctx, size_t size);</title> <para> The function talloc_size() should be used when you don't have a convenient type to pass to talloc(). Unlike talloc(), it is not type safe (as it returns a void *), so you are on your own for type checking. </para> </refsect2> <refsect2><title>int talloc_free(void *ptr);</title> <para> The talloc_free() function frees a piece of talloc memory, and all its children. You can call talloc_free() on any pointer returned by talloc(). </para> <para> The return value of talloc_free() indicates success or failure, with 0 returned for success and -1 for failure. The only possible failure condition is if <emphasis role="italic">ptr</emphasis> had a destructor attached to it and the destructor returned -1. See <link linkend="talloc_set_destructor"><quote>talloc_set_destructor()</quote></link> for details on destructors. </para> <para> If this pointer has an additional parent when talloc_free() is called then the memory is not actually released, but instead the most recently established parent is destroyed. See <link linkend="talloc_reference"><quote>talloc_reference()</quote></link> for details on establishing additional parents. </para> <para> For more control on which parent is removed, see <link linkend="talloc_unlink"><quote>talloc_unlink()</quote></link>. </para> <para> talloc_free() operates recursively on its children. </para> </refsect2> <refsect2 id="talloc_reference"><title>void *talloc_reference(const void *ctx, const void *ptr);</title> <para> The talloc_reference() function makes <emphasis role="italic">ctx</emphasis> an additional parent of <emphasis role="italic">ptr</emphasis>. </para> <para> The return value of talloc_reference() is always the original pointer <emphasis role="italic">ptr</emphasis>, unless talloc ran out of memory in creating the reference in which case it will return NULL (each additional reference consumes around 48 bytes of memory on intel x86 platforms). </para> <para> If <emphasis role="italic">ptr</emphasis> is NULL, then the function is a no-op, and simply returns NULL. </para> <para> After creating a reference you can free it in one of the following ways: </para> <para> <itemizedlist> <listitem> <para> you can talloc_free() any parent of the original pointer. That will reduce the number of parents of this pointer by 1, and will cause this pointer to be freed if it runs out of parents. </para> </listitem> <listitem> <para> you can talloc_free() the pointer itself. That will destroy the most recently established parent to the pointer and leave the pointer as a child of its current parent. </para> </listitem> </itemizedlist> </para> <para> For more control on which parent to remove, see <link linkend="talloc_unlink"><quote>talloc_unlink()</quote></link>. </para> </refsect2> <refsect2 id="talloc_unlink"><title>int talloc_unlink(const void *ctx, const void *ptr);</title> <para> The talloc_unlink() function removes a specific parent from <emphasis role="italic">ptr</emphasis>. The <emphasis role="italic">ctx</emphasis> passed must either be a context used in talloc_reference() with this pointer, or must be a direct parent of ptr. </para> <para> Note that if the parent has already been removed using talloc_free() then this function will fail and will return -1. Likewise, if <emphasis role="italic">ptr</emphasis> is NULL, then the function will make no modifications and return -1. </para> <para> Usually you can just use talloc_free() instead of talloc_unlink(), but sometimes it is useful to have the additional control on which parent is removed. </para> </refsect2> <refsect2 id="talloc_set_destructor"><title>void talloc_set_destructor(const void *ptr, int (*destructor)(void *));</title> <para> The function talloc_set_destructor() sets the <emphasis role="italic">destructor</emphasis> for the pointer <emphasis role="italic">ptr</emphasis>. A <emphasis role="italic">destructor</emphasis> is a function that is called when the memory used by a pointer is about to be released. The destructor receives <emphasis role="italic">ptr</emphasis> as an argument, and should return 0 for success and -1 for failure. </para> <para> The <emphasis role="italic">destructor</emphasis> can do anything it wants to, including freeing other pieces of memory. A common use for destructors is to clean up operating system resources (such as open file descriptors) contained in the structure the destructor is placed on. </para> <para> You can only place one destructor on a pointer. If you need more than one destructor then you can create a zero-length child of the pointer and place an additional destructor on that. </para> <para> To remove a destructor call talloc_set_destructor() with NULL for the destructor. </para> <para> If your destructor attempts to talloc_free() the pointer that it is the destructor for then talloc_free() will return -1 and the free will be ignored. This would be a pointless operation anyway, as the destructor is only called when the memory is just about to go away. </para> </refsect2> <refsect2><title>void talloc_increase_ref_count(const void *<emphasis role="italic">ptr</emphasis>);</title> <para> The talloc_increase_ref_count(<emphasis role="italic">ptr</emphasis>) function is exactly equivalent to: </para> <programlisting>talloc_reference(NULL, ptr);</programlisting> <para> You can use either syntax, depending on which you think is clearer in your code. </para> </refsect2> <refsect2 id="talloc_set_name"><title>void talloc_set_name(const void *ptr, const char *fmt, ...);</title> <para> Each talloc pointer has a "name". The name is used principally for debugging purposes, although it is also possible to set and get the name on a pointer in as a way of "marking" pointers in your code. </para> <para> The main use for names on pointer is for "talloc reports". See <link linkend="talloc_report"><quote>talloc_report()</quote></link> and <link linkend="talloc_report_full"><quote>talloc_report_full()</quote></link> for details. Also see <link linkend="talloc_enable_leak_report"><quote>talloc_enable_leak_report()</quote></link> and <link linkend="talloc_enable_leak_report_full"><quote>talloc_enable_leak_report_full()</quote></link>. </para> <para> The talloc_set_name() function allocates memory as a child of the pointer. It is logically equivalent to: </para> <programlisting>talloc_set_name_const(ptr, talloc_asprintf(ptr, fmt, ...));</programlisting> <para> Note that multiple calls to talloc_set_name() will allocate more memory without releasing the name. All of the memory is released when the ptr is freed using talloc_free(). </para> </refsect2> <refsect2><title>void talloc_set_name_const(const void *<emphasis role="italic">ptr</emphasis>, const char *<emphasis role="italic">name</emphasis>);</title> <para> The function talloc_set_name_const() is just like talloc_set_name(), but it takes a string constant, and is much faster. It is extensively used by the "auto naming" macros, such as talloc_p(). </para> <para> This function does not allocate any memory. It just copies the supplied pointer into the internal representation of the talloc ptr. This means you must not pass a <emphasis role="italic">name</emphasis> pointer to memory that will disappear before <emphasis role="italic">ptr</emphasis> is freed with talloc_free(). </para> </refsect2> <refsect2><title>void *talloc_named(const void *<emphasis role="italic">ctx</emphasis>, size_t <emphasis role="italic">size</emphasis>, const char *<emphasis role="italic">fmt</emphasis>, ...);</title> <para> The talloc_named() function creates a named talloc pointer. It is equivalent to: </para> <programlisting>ptr = talloc_size(ctx, size); talloc_set_name(ptr, fmt, ....);</programlisting> </refsect2> <refsect2><title>void *talloc_named_const(const void *<emphasis role="italic">ctx</emphasis>, size_t <emphasis role="italic">size</emphasis>, const char *<emphasis role="italic">name</emphasis>);</title> <para> This is equivalent to: </para> <programlisting>ptr = talloc_size(ctx, size); talloc_set_name_const(ptr, name);</programlisting> </refsect2> <refsect2><title>const char *talloc_get_name(const void *<emphasis role="italic">ptr</emphasis>);</title> <para> This returns the current name for the given talloc pointer, <emphasis role="italic">ptr</emphasis>. See <link linkend="talloc_set_name"><quote>talloc_set_name()</quote></link> for details. </para> </refsect2> <refsect2><title>void *talloc_init(const char *<emphasis role="italic">fmt</emphasis>, ...);</title> <para> This function creates a zero length named talloc context as a top level context. It is equivalent to: </para> <programlisting>talloc_named(NULL, 0, fmt, ...);</programlisting> </refsect2> <refsect2><title>void *talloc_new(void *<emphasis role="italic">ctx</emphasis>);</title> <para> This is a utility macro that creates a new memory context hanging off an exiting context, automatically naming it "talloc_new: __location__" where __location__ is the source line it is called from. It is particularly useful for creating a new temporary working context. </para> </refsect2> <refsect2><title>(<emphasis role="italic">type</emphasis> *)talloc_realloc(const void *<emphasis role="italic">ctx</emphasis>, void *<emphasis role="italic">ptr</emphasis>, <emphasis role="italic">type</emphasis>, <emphasis role="italic">count</emphasis>);</title> <para> The talloc_realloc() macro changes the size of a talloc pointer. It has the following equivalences: </para> <programlisting>talloc_realloc(ctx, NULL, type, 1) ==> talloc(ctx, type); talloc_realloc(ctx, ptr, type, 0) ==> talloc_free(ptr);</programlisting> <para> The <emphasis role="italic">ctx</emphasis> argument is only used if <emphasis role="italic">ptr</emphasis> is not NULL, otherwise it is ignored. </para> <para> talloc_realloc() returns the new pointer, or NULL on failure. The call will fail either due to a lack of memory, or because the pointer has more than one parent (see <link linkend="talloc_reference"><quote>talloc_reference()</quote></link>). </para> </refsect2> <refsect2><title>void *talloc_realloc_size(const void *ctx, void *ptr, size_t size);</title> <para> the talloc_realloc_size() function is useful when the type is not known so the type-safe talloc_realloc() cannot be used. </para> </refsect2> <refsect2><title>void *talloc_steal(const void *<emphasis role="italic">new_ctx</emphasis>, const void *<emphasis role="italic">ptr</emphasis>);</title> <para> The talloc_steal() function changes the parent context of a talloc pointer. It is typically used when the context that the pointer is currently a child of is going to be freed and you wish to keep the memory for a longer time. </para> <para> The talloc_steal() function returns the pointer that you pass it. It does not have any failure modes. </para> <para> NOTE: It is possible to produce loops in the parent/child relationship if you are not careful with talloc_steal(). No guarantees are provided as to your sanity or the safety of your data if you do this. </para> </refsect2> <refsect2><title>off_t talloc_total_size(const void *<emphasis role="italic">ptr</emphasis>);</title> <para> The talloc_total_size() function returns the total size in bytes used by this pointer and all child pointers. Mostly useful for debugging. </para> <para> Passing NULL is allowed, but it will only give a meaningful result if talloc_enable_leak_report() or talloc_enable_leak_report_full() has been called. </para> </refsect2> <refsect2><title>off_t talloc_total_blocks(const void *<emphasis role="italic">ptr</emphasis>);</title> <para> The talloc_total_blocks() function returns the total memory block count used by this pointer and all child pointers. Mostly useful for debugging. </para> <para> Passing NULL is allowed, but it will only give a meaningful result if talloc_enable_leak_report() or talloc_enable_leak_report_full() has been called. </para> </refsect2> <refsect2 id="talloc_report"><title>void talloc_report(const void *ptr, FILE *f);</title> <para> The talloc_report() function prints a summary report of all memory used by <emphasis role="italic">ptr</emphasis>. One line of report is printed for each immediate child of ptr, showing the total memory and number of blocks used by that child. </para> <para> You can pass NULL for the pointer, in which case a report is printed for the top level memory context, but only if talloc_enable_leak_report() or talloc_enable_leak_report_full() has been called. </para> </refsect2> <refsect2 id="talloc_report_full"><title>void talloc_report_full(const void *<emphasis role="italic">ptr</emphasis>, FILE *<emphasis role="italic">f</emphasis>);</title> <para> This provides a more detailed report than talloc_report(). It will recursively print the entire tree of memory referenced by the pointer. References in the tree are shown by giving the name of the pointer that is referenced. </para> <para> You can pass NULL for the pointer, in which case a report is printed for the top level memory context, but only if talloc_enable_leak_report() or talloc_enable_leak_report_full() has been called. </para> </refsect2> <refsect2 id="talloc_enable_leak_report"><title>void talloc_enable_leak_report(void);</title> <para> This enables calling of talloc_report(NULL, stderr) when the program exits. In Samba4 this is enabled by using the --leak-report command line option. </para> <para> For it to be useful, this function must be called before any other talloc function as it establishes a "null context" that acts as the top of the tree. If you don't call this function first then passing NULL to talloc_report() or talloc_report_full() won't give you the full tree printout. </para> <para> Here is a typical talloc report: </para> <screen format="linespecific">talloc report on 'null_context' (total 267 bytes in 15 blocks) libcli/auth/spnego_parse.c:55 contains 31 bytes in 2 blocks libcli/auth/spnego_parse.c:55 contains 31 bytes in 2 blocks iconv(UTF8,CP850) contains 42 bytes in 2 blocks libcli/auth/spnego_parse.c:55 contains 31 bytes in 2 blocks iconv(CP850,UTF8) contains 42 bytes in 2 blocks iconv(UTF8,UTF-16LE) contains 45 bytes in 2 blocks iconv(UTF-16LE,UTF8) contains 45 bytes in 2 blocks </screen> </refsect2> <refsect2 id="talloc_enable_leak_report_full"><title>void talloc_enable_leak_report_full(void);</title> <para> This enables calling of talloc_report_full(NULL, stderr) when the program exits. In Samba4 this is enabled by using the --leak-report-full command line option. </para> <para> For it to be useful, this function must be called before any other talloc function as it establishes a "null context" that acts as the top of the tree. If you don't call this function first then passing NULL to talloc_report() or talloc_report_full() won't give you the full tree printout. </para> <para> Here is a typical full report: </para> <screen format="linespecific">full talloc report on 'root' (total 18 bytes in 8 blocks) p1 contains 18 bytes in 7 blocks (ref 0) r1 contains 13 bytes in 2 blocks (ref 0) reference to: p2 p2 contains 1 bytes in 1 blocks (ref 1) x3 contains 1 bytes in 1 blocks (ref 0) x2 contains 1 bytes in 1 blocks (ref 0) x1 contains 1 bytes in 1 blocks (ref 0) </screen> </refsect2> <refsect2><title>(<emphasis role="italic">type</emphasis> *)talloc_zero(const void *<emphasis role="italic">ctx</emphasis>, <emphasis role="italic">type</emphasis>);</title> <para> The talloc_zero() macro is equivalent to: </para> <programlisting>ptr = talloc(ctx, type); if (ptr) memset(ptr, 0, sizeof(type));</programlisting> </refsect2> <refsect2><title>void *talloc_zero_size(const void *<emphasis role="italic">ctx</emphasis>, size_t <emphasis role="italic">size</emphasis>)</title> <para> The talloc_zero_size() function is useful when you don't have a known type. </para> </refsect2> <refsect2><title>void *talloc_memdup(const void *<emphasis role="italic">ctx</emphasis>, const void *<emphasis role="italic">p</emphasis>, size_t size);</title> <para> The talloc_memdup() function is equivalent to: </para> <programlisting>ptr = talloc_size(ctx, size); if (ptr) memcpy(ptr, p, size);</programlisting> </refsect2> <refsect2><title>char *talloc_strdup(const void *<emphasis role="italic">ctx</emphasis>, const char *<emphasis role="italic">p</emphasis>);</title> <para> The talloc_strdup() function is equivalent to: </para> <programlisting>ptr = talloc_size(ctx, strlen(p)+1); if (ptr) memcpy(ptr, p, strlen(p)+1);</programlisting> <para> This function sets the name of the new pointer to the passed string. This is equivalent to: </para> <programlisting>talloc_set_name_const(ptr, ptr)</programlisting> </refsect2> <refsect2><title>char *talloc_strndup(const void *<emphasis role="italic">t</emphasis>, const char *<emphasis role="italic">p</emphasis>, size_t <emphasis role="italic">n</emphasis>);</title> <para> The talloc_strndup() function is the talloc equivalent of the C library function strndup(3). </para> <para> This function sets the name of the new pointer to the passed string. This is equivalent to: </para> <programlisting>talloc_set_name_const(ptr, ptr)</programlisting> </refsect2> <refsect2><title>char *talloc_vasprintf(const void *<emphasis role="italic">t</emphasis>, const char *<emphasis role="italic">fmt</emphasis>, va_list <emphasis role="italic">ap</emphasis>);</title> <para> The talloc_vasprintf() function is the talloc equivalent of the C library function vasprintf(3). </para> </refsect2> <refsect2><title>char *talloc_asprintf(const void *<emphasis role="italic">t</emphasis>, const char *<emphasis role="italic">fmt</emphasis>, ...);</title> <para> The talloc_asprintf() function is the talloc equivalent of the C library function asprintf(3). </para> <para> This function sets the name of the new pointer to the passed string. This is equivalent to: </para> <programlisting>talloc_set_name_const(ptr, ptr)</programlisting> </refsect2> <refsect2><title>char *talloc_asprintf_append(char *s, const char *fmt, ...);</title> <para> The talloc_asprintf_append() function appends the given formatted string to the given string. </para> </refsect2> <refsect2><title>(type *)talloc_array(const void *ctx, type, uint_t count);</title> <para> The talloc_array() macro is equivalent to: </para> <programlisting>(type *)talloc_size(ctx, sizeof(type) * count);</programlisting> <para> except that it provides integer overflow protection for the multiply, returning NULL if the multiply overflows. </para> </refsect2> <refsect2><title>void *talloc_array_size(const void *ctx, size_t size, uint_t count);</title> <para> The talloc_array_size() function is useful when the type is not known. It operates in the same way as talloc_array(), but takes a size instead of a type. </para> </refsect2> <refsect2><title>void *talloc_realloc_fn(const void *ctx, void *ptr, size_t size)</title> <para> This is a non-macro version of talloc_realloc(), which is useful as libraries sometimes want a realloc function pointer. A realloc(3) implementation encapsulates the functionality of malloc(3), free(3) and realloc(3) in one call, which is why it is useful to be able to pass around a single function pointer. </para> </refsect2> <refsect2><title>void *talloc_autofree_context(void);</title> <para> This is a handy utility function that returns a talloc context which will be automatically freed on program exit. This can be used to reduce the noise in memory leak reports. </para> </refsect2> <refsect2><title>void *talloc_check_name(const void *ptr, const char *name);</title> <para> This function checks if a pointer has the specified <emphasis role="italic">name</emphasis>. If it does then the pointer is returned. It it doesn't then NULL is returned. </para> </refsect2> <refsect2><title>(type *)talloc_get_type(const void *ptr, type);</title> <para> This macro allows you to do type checking on talloc pointers. It is particularly useful for void* private pointers. It is equivalent to this: </para> <programlisting>(type *)talloc_check_name(ptr, #type)</programlisting> </refsect2> <refsect2><title>talloc_set_type(const void *ptr, type);</title> <para> This macro allows you to force the name of a pointer to be a particular <emphasis>type</emphasis>. This can be used in conjunction with talloc_get_type() to do type checking on void* pointers. </para> <para> It is equivalent to this: </para> <programlisting>talloc_set_name_const(ptr, #type)</programlisting> </refsect2> </refsect1> <refsect1><title>PERFORMANCE</title> <para> All the additional features of talloc(3) over malloc(3) do come at a price. We have a simple performance test in Samba4 that measures talloc() versus malloc() performance, and it seems that talloc() is about 10% slower than malloc() on my x86 Debian Linux box. For Samba, the great reduction in code complexity that we get by using talloc makes this worthwhile, especially as the total overhead of talloc/malloc in Samba is already quite small. </para> </refsect1> <refsect1><title>SEE ALSO</title> <para> malloc(3), strndup(3), vasprintf(3), asprintf(3), <ulink url="http://talloc.samba.org/"/> </para> </refsect1> <refsect1><title>COPYRIGHT/LICENSE</title> <para> Copyright (C) Andrew Tridgell 2004 </para> <para> This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. </para> <para> This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. </para> <para> You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. </para> </refsect1> </refentry>