Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / libxml2 / 7216cfd6622d947695c67b7b430edef8cc0af967 / . / doc / xmlmem.html
blob: 5867c65610bdc11840ba08fbfefa56381c56d58f [file] [log] [blame]
Daniel Veillard43d3f612001-11-10 11:57:23 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/1999/REC-html401-19991224/loose.dtd">
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]2<html>
3<head>
Daniel Veillard7216cfd2002-11-08 15:10:00 +0000[diff] [blame^]4<meta http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
Daniel Veillardc332dab2002-03-29 14:08:27 +0000[diff] [blame]5<link rel="SHORTCUT ICON" href="/favicon.ico">
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]6<style type="text/css"><!--
Daniel Veillard373a4752002-02-21 14:46:29 +0000[diff] [blame]7TD {font-family: Verdana,Arial,Helvetica}
8BODY {font-family: Verdana,Arial,Helvetica; margin-top: 2em; margin-left: 0em; margin-right: 0em}
9H1 {font-family: Verdana,Arial,Helvetica}
10H2 {font-family: Verdana,Arial,Helvetica}
11H3 {font-family: Verdana,Arial,Helvetica}
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]12A:link, A:visited, A:active { text-decoration: underline }
13--></style>
14<title>Memory Management</title>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]15</head>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]16<body bgcolor="#8b7765" text="#000000" link="#000000" vlink="#000000">
17<table border="0" width="100%" cellpadding="5" cellspacing="0" align="center"><tr>
18<td width="180">
Daniel Veillard8f40f1e2002-08-28 21:18:45 +0000[diff] [blame]19<a href="http://www.gnome.org/"><img src="gnome2.png" alt="Gnome2 Logo"></a><a href="http://www.w3.org/Status"><img src="w3c.png" alt="W3C Logo"></a><a href="http://www.redhat.com/"><img src="redhat.gif" alt="Red Hat Logo"></a><div align="left"><a href="http://xmlsoft.org/"><img src="Libxml2-Logo-180x168.gif" alt="Made with Libxml2 Logo"></a></div>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]20</td>
21<td><table border="0" width="90%" cellpadding="2" cellspacing="0" align="center" bgcolor="#000000"><tr><td><table width="100%" border="0" cellspacing="1" cellpadding="3" bgcolor="#fffacd"><tr><td align="center">
22<h1>The XML C library for Gnome</h1>
23<h2>Memory Management</h2>
24</td></tr></table></td></tr></table></td>
25</tr></table>
26<table border="0" cellpadding="4" cellspacing="0" width="100%" align="center"><tr><td bgcolor="#8b7765"><table border="0" cellspacing="0" cellpadding="2" width="100%"><tr>
27<td valign="top" width="200" bgcolor="#8b7765"><table border="0" cellspacing="0" cellpadding="1" width="100%" bgcolor="#000000"><tr><td>
28<table width="100%" border="0" cellspacing="1" cellpadding="3">
29<tr><td colspan="1" bgcolor="#eecfa1" align="center"><center><b>Main Menu</b></center></td></tr>
Daniel Veillard8acca112002-01-21 09:52:27 +0000[diff] [blame]30<tr><td bgcolor="#fffacd"><ul>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]31<li><a href="index.html">Home</a></li>
32<li><a href="intro.html">Introduction</a></li>
33<li><a href="FAQ.html">FAQ</a></li>
34<li><a href="docs.html">Documentation</a></li>
35<li><a href="bugs.html">Reporting bugs and getting help</a></li>
36<li><a href="help.html">How to help</a></li>
37<li><a href="downloads.html">Downloads</a></li>
38<li><a href="news.html">News</a></li>
Daniel Veillard7b602b42002-01-08 13:26:00 +0000[diff] [blame]39<li><a href="XMLinfo.html">XML</a></li>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]40<li><a href="XSLT.html">XSLT</a></li>
Daniel Veillard6dbcaf82002-02-20 14:37:47 +0000[diff] [blame]41<li><a href="python.html">Python and bindings</a></li>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]42<li><a href="architecture.html">libxml architecture</a></li>
43<li><a href="tree.html">The tree output</a></li>
44<li><a href="interface.html">The SAX interface</a></li>
45<li><a href="xmldtd.html">Validation &amp; DTDs</a></li>
46<li><a href="xmlmem.html">Memory Management</a></li>
47<li><a href="encoding.html">Encodings support</a></li>
48<li><a href="xmlio.html">I/O Interfaces</a></li>
49<li><a href="catalog.html">Catalog support</a></li>
50<li><a href="library.html">The parser interfaces</a></li>
51<li><a href="entities.html">Entities or no entities</a></li>
52<li><a href="namespaces.html">Namespaces</a></li>
53<li><a href="upgrade.html">Upgrading 1.x code</a></li>
Daniel Veillard52dcab32001-10-30 12:51:17 +0000[diff] [blame]54<li><a href="threads.html">Thread safety</a></li>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]55<li><a href="DOM.html">DOM Principles</a></li>
56<li><a href="example.html">A real example</a></li>
57<li><a href="contribs.html">Contributions</a></li>
Daniel Veillardfc59c092002-06-05 14:48:26 +0000[diff] [blame]58<li><a href="tutorial/index.html">Tutorial</a></li>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]59<li>
60<a href="xml.html">flat page</a>, <a href="site.xsl">stylesheet</a>
61</li>
62</ul></td></tr>
63</table>
64<table width="100%" border="0" cellspacing="1" cellpadding="3">
Daniel Veillard3bf65be2002-01-23 12:36:34 +0000[diff] [blame]65<tr><td colspan="1" bgcolor="#eecfa1" align="center"><center><b>API Indexes</b></center></td></tr>
Daniel Veillard5ede35e2002-10-01 11:37:35 +0000[diff] [blame]66<tr><td bgcolor="#fffacd">
Daniel Veillard595978c2002-10-09 18:46:35 +0000[diff] [blame]67<form action="search.php" enctype="application/x-www-form-urlencoded" method="GET">
Daniel Veillard5ede35e2002-10-01 11:37:35 +0000[diff] [blame]68<input name="query" type="TEXT" size="20" value=""><input name="submit" type="submit" value="Search ...">
69</form>
70<ul>
Daniel Veillardf8592562002-01-23 17:58:17 +0000[diff] [blame]71<li><a href="APIchunk0.html">Alphabetic</a></li>
Daniel Veillard3bf65be2002-01-23 12:36:34 +0000[diff] [blame]72<li><a href="APIconstructors.html">Constructors</a></li>
73<li><a href="APIfunctions.html">Functions/Types</a></li>
74<li><a href="APIfiles.html">Modules</a></li>
75<li><a href="APIsymbols.html">Symbols</a></li>
Daniel Veillard5ede35e2002-10-01 11:37:35 +0000[diff] [blame]76</ul>
77</td></tr>
Daniel Veillard3bf65be2002-01-23 12:36:34 +0000[diff] [blame]78</table>
79<table width="100%" border="0" cellspacing="1" cellpadding="3">
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]80<tr><td colspan="1" bgcolor="#eecfa1" align="center"><center><b>Related links</b></center></td></tr>
Daniel Veillard8acca112002-01-21 09:52:27 +0000[diff] [blame]81<tr><td bgcolor="#fffacd"><ul>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]82<li><a href="http://mail.gnome.org/archives/xml/">Mail archive</a></li>
83<li><a href="http://xmlsoft.org/XSLT/">XSLT libxslt</a></li>
Daniel Veillard4a859202002-01-08 11:49:22 +0000[diff] [blame]84<li><a href="http://phd.cs.unibo.it/gdome2/">DOM gdome2</a></li>
Daniel Veillard2d347fa2002-03-17 10:34:11 +0000[diff] [blame]85<li><a href="http://www.aleksey.com/xmlsec/">XML-DSig xmlsec</a></li>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]86<li><a href="ftp://xmlsoft.org/">FTP</a></li>
87<li><a href="http://www.fh-frankfurt.de/~igor/projects/libxml/">Windows binaries</a></li>
Daniel Veillarddb9dfd92001-11-26 17:25:02 +0000[diff] [blame]88<li><a href="http://garypennington.net/libxml2/">Solaris binaries</a></li>
Daniel Veillardcb7543b2002-09-09 10:54:06 +0000[diff] [blame]89<li><a href="http://www.zveno.com/open_source/libxml2xslt.html">MacOsX binaries</a></li>
Daniel Veillarde6d8e202002-05-02 06:11:10 +0000[diff] [blame]90<li><a href="http://sourceforge.net/projects/libxml2-pas/">Pascal bindings</a></li>
Daniel Veillard2d347fa2002-03-17 10:34:11 +0000[diff] [blame]91<li><a href="http://bugzilla.gnome.org/buglist.cgi?product=libxml&amp;product=libxml2">Bug Tracker</a></li>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]92</ul></td></tr>
93</table>
94</td></tr></table></td>
95<td valign="top" bgcolor="#8b7765"><table border="0" cellspacing="0" cellpadding="1" width="100%"><tr><td><table border="0" cellspacing="0" cellpadding="1" width="100%" bgcolor="#000000"><tr><td><table border="0" cellpadding="3" cellspacing="1" width="100%"><tr><td bgcolor="#fffacd">
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]96<p>Table of Content:</p>
97<ol>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]98<li><a href="#General3">General overview</a></li>
Daniel Veillard0b28e882002-07-24 23:47:05 +0000[diff] [blame]99 <li><a href="#setting">Setting libxml set of memory routines</a></li>
100 <li><a href="#cleanup">Cleaning up after parsing</a></li>
101 <li><a href="#Debugging">Debugging routines</a></li>
102 <li><a href="#General4">General memory requirements</a></li>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]103</ol>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]104<h3><a name="General3">General overview</a></h3>
105<p>The module <code><a href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlmemory.h</a></code>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]106provides the interfaces to the libxml memory system:</p>
107<ul>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]108<li>libxml does not use the libc memory allocator directly but xmlFree(),
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]109 xmlMalloc() and xmlRealloc()</li>
Daniel Veillard0b28e882002-07-24 23:47:05 +0000[diff] [blame]110 <li>those routines can be reallocated to a specific set of routine, by
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]111 default the libc ones i.e. free(), malloc() and realloc()</li>
Daniel Veillard0b28e882002-07-24 23:47:05 +0000[diff] [blame]112 <li>the xmlmemory.c module includes a set of debugging routine</li>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]113</ul>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]114<h3><a name="setting">Setting libxml set of memory routines</a></h3>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]115<p>It is sometimes useful to not use the default memory allocator, either for
116debugging, analysis or to implement a specific behaviour on memory management
117(like on embedded systems). Two function calls are available to do so:</p>
118<ul>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]119<li>
Daniel Veillardaf43f632002-03-08 15:05:20 +0000[diff] [blame]120<a href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlMemGet
121 ()</a> which return the current set of functions in use by the parser</li>
Daniel Veillard0b28e882002-07-24 23:47:05 +0000[diff] [blame]122 <li>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]123<a href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlMemSetup()</a>
Daniel Veillardaf43f632002-03-08 15:05:20 +0000[diff] [blame]124 which allow to set up a new set of memory allocation functions</li>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]125</ul>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]126<p>Of course a call to xmlMemSetup() should probably be done before calling
127any other libxml routines (unless you are sure your allocations routines are
128compatibles).</p>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]129<h3><a name="cleanup">Cleaning up after parsing</a></h3>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]130<p>Libxml is not stateless, there is a few set of memory structures needing
Daniel Veillard63d83142002-05-20 06:51:05 +0000[diff] [blame]131allocation before the parser is fully functional (some encoding structures
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]132for example). This also mean that once parsing is finished there is a tiny
133amount of memory (a few hundred bytes) which can be recollected if you don't
134reuse the parser immediately:</p>
135<ul>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]136<li>
137<a href="http://xmlsoft.org/html/libxml-parser.html">xmlCleanupParser
Daniel Veillardaf43f632002-03-08 15:05:20 +0000[diff] [blame]138 ()</a> is a centralized routine to free the parsing states. Note that it
139 won't deallocate any produced tree if any (use the xmlFreeDoc() and
140 related routines for this).</li>
Daniel Veillard0b28e882002-07-24 23:47:05 +0000[diff] [blame]141 <li>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]142<a href="http://xmlsoft.org/html/libxml-parser.html">xmlInitParser
Daniel Veillardaf43f632002-03-08 15:05:20 +0000[diff] [blame]143 ()</a> is the dual routine allowing to preallocate the parsing state
144 which can be useful for example to avoid initialization reentrancy
145 problems when using libxml in multithreaded applications</li>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]146</ul>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]147<p>Generally xmlCleanupParser() is safe, if needed the state will be rebuild
148at the next invocation of parser routines, but be careful of the consequences
149in multithreaded applications.</p>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]150<h3><a name="Debugging">Debugging routines</a></h3>
151<p>When configured using --with-mem-debug flag (off by default), libxml uses
Daniel Veillard63d83142002-05-20 06:51:05 +0000[diff] [blame]152a set of memory allocation debugging routines keeping track of all allocated
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]153blocks and the location in the code where the routine was called. A couple of
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]154other debugging routines allow to dump the memory allocated infos to a file
155or call a specific routine when a given block number is allocated:</p>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]156<ul>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]157<li>
Daniel Veillard0b28e882002-07-24 23:47:05 +0000[diff] [blame]158<a href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlMallocLoc()</a>
159 <a href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlReallocLoc()</a>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]160 and <a href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlMemStrdupLoc()</a>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]161 are the memory debugging replacement allocation routines</li>
Daniel Veillard0b28e882002-07-24 23:47:05 +0000[diff] [blame]162 <li>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]163<a href="http://xmlsoft.org/html/libxml-xmlmemory.html">xmlMemoryDump
Daniel Veillardaf43f632002-03-08 15:05:20 +0000[diff] [blame]164 ()</a> dumps all the informations about the allocated memory block lefts
165 in the <code>.memdump</code> file</li>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]166</ul>
Daniel Veillard63d83142002-05-20 06:51:05 +0000[diff] [blame]167<p>When developing libxml memory debug is enabled, the tests programs call
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]168xmlMemoryDump () and the &quot;make test&quot; regression tests will check for any
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]169memory leak during the full regression test sequence, this helps a lot
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]170ensuring that libxml does not leak memory and bullet proof memory
171allocations use (some libc implementations are known to be far too permissive
172resulting in major portability problems!).</p>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]173<p>If the .memdump reports a leak, it displays the allocation function and
174also tries to give some informations about the content and structure of the
175allocated blocks left. This is sufficient in most cases to find the culprit,
Daniel Veillard63d83142002-05-20 06:51:05 +0000[diff] [blame]176but not always. Assuming the allocation problem is reproducible, it is
177possible to find more easily:</p>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]178<ol>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]179<li>write down the block number xxxx not allocated</li>
Daniel Veillard0b28e882002-07-24 23:47:05 +0000[diff] [blame]180 <li>export the environment variable XML_MEM_BREAKPOINT=xxxx , the easiest
Daniel Veillarda8a89fe2002-04-12 21:03:34 +0000[diff] [blame]181 when using GDB is to simply give the command
182 <p><code>set environment XML_MEM_BREAKPOINT xxxx</code></p>
Daniel Veillard0b28e882002-07-24 23:47:05 +0000[diff] [blame]183 <p>before running the program.</p>
184 </li>
185 <li>run the program under a debugger and set a breakpoint on
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]186 xmlMallocBreakpoint() a specific function called when this precise block
187 is allocated</li>
Daniel Veillard0b28e882002-07-24 23:47:05 +0000[diff] [blame]188 <li>when the breakpoint is reached you can then do a fine analysis of the
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]189 allocation an step to see the condition resulting in the missing
190 deallocation.</li>
191</ol>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]192<p>I used to use a commercial tool to debug libxml memory problems but after
193noticing that it was not detecting memory leaks that simple mechanism was
Daniel Veillarda8a89fe2002-04-12 21:03:34 +0000[diff] [blame]194used and proved extremely efficient until now. Lately I have also used <a href="http://developer.kde.org/~sewardj/">valgrind</a> with quite some
195success, it is tied to the i386 architecture since it works by emulating the
196processor and instruction set, it is slow but extremely efficient, i.e. it
197spot memory usage errors in a very precise way.</p>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]198<h3><a name="General4">General memory requirements</a></h3>
199<p>How much libxml memory require ? It's hard to tell in average it depends
200of a number of things:</p>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]201<ul>
Daniel Veillard63d83142002-05-20 06:51:05 +0000[diff] [blame]202<li>the parser itself should work in a fixed amount of memory, except for
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]203 information maintained about the stacks of names and entities locations.
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]204 The I/O and encoding handlers will probably account for a few KBytes.
205 This is true for both the XML and HTML parser (though the HTML parser
206 need more state).</li>
Daniel Veillard0b28e882002-07-24 23:47:05 +0000[diff] [blame]207 <li>If you are generating the DOM tree then memory requirements will grow
Daniel Veillard63d83142002-05-20 06:51:05 +0000[diff] [blame]208 nearly linear with the size of the data. In general for a balanced
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]209 textual document the internal memory requirement is about 4 times the
Daniel Veillard63d83142002-05-20 06:51:05 +0000[diff] [blame]210 size of the UTF8 serialization of this document (example the XML-1.0
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]211 recommendation is a bit more of 150KBytes and takes 650KBytes of main
212 memory when parsed). Validation will add a amount of memory required for
213 maintaining the external Dtd state which should be linear with the
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]214 complexity of the content model defined by the Dtd</li>
Daniel Veillard0b28e882002-07-24 23:47:05 +0000[diff] [blame]215 <li>If you don't care about the advanced features of libxml like
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]216 validation, DOM, XPath or XPointer, but really need to work fixed memory
217 requirements, then the SAX interface should be used.</li>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]218</ul>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]219<p>
Daniel Veillard3f4c40f2002-02-13 09:19:28 +0000[diff] [blame]220<p><a href="bugs.html">Daniel Veillard</a></p>
Daniel Veillardb8cfbd12001-10-25 10:53:28 +0000[diff] [blame]221</td></tr></table></td></tr></table></td></tr></table></td>
222</tr></table></td></tr></table>
Daniel Veillard189446d2000-10-13 10:23:06 +0000[diff] [blame]223</body>
224</html>