docs/manual/dso.html

	dso.html revision 9c756be3f06fb4cc275a81297605786f91905332
214N/A<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
214N/A<HTML><HEAD>
935N/A<TITLE>Apache 1.3 Dynamic Shared Object (DSO) support</TITLE>
214N/A</HEAD>
214N/A
919N/A<!-- Background white, links blue (unvisited), navy (visited), red (active) -->
919N/A<BODY
919N/A BGCOLOR="#FFFFFF"
919N/A TEXT="#000000"
919N/A LINK="#0000FF"
919N/A VLINK="#000080"
919N/A ALINK="#FF0000"
919N/A>
919N/A<BLOCKQUOTE>
919N/A<!--#include virtual="header.html" -->
919N/A
919N/A<DIV ALIGN=CENTER>
919N/A
919N/A<H1>
919N/AApache 1.3<BR>
919N/ADynamic Shared Object (DSO)<BR>
919N/ASupport
214N/A</H1>
214N/A
493N/A<ADDRESS>Originally written by<BR>
493N/ARalf S. Engelschall &lt;rse@apache.org&gt, April 1998</ADDRESS>
1003N/A
1003N/A</DIV>
1003N/A
1003N/A<H3>Background</H3>
214N/A
493N/A<P>On modern Unix derivatives there exists a nifty mechanism usually called
214N/Adynamic linking/loading of <EM>Dynamic Shared Objects</EM> (DSO) which
493N/Aprovides a way to build a piece of program code in a special format for
493N/Aloading it at run-time into the address space of an executable program.
935N/A
935N/A<P>This loading can usually be done in two ways: Automatically by a system
935N/Aprogram called <CODE>ld.so</CODE> when an executable program is started or
810N/Amanually from within the executing program via a programmatic system interface
810N/Ato the Unix loader through the system calls <CODE>dlopen()/dlsym()</CODE>.
810N/A
493N/A<P>In the first way the DSO's are usually called <EM>shared libraries</EM> or
493N/A<EM>DSO libraries</EM> and named <CODE>libfoo.so</CODE> or
493N/A<CODE>libfoo.so.1.2</CODE>.  They reside in a system directory (usually
910N/A<CODE>/usr/lib</CODE>) and the link to the executable program is established
493N/Aat build-time by specifying <CODE>-lfoo</CODE> to the linker command. This
214N/Ahard-codes library references into the executable program file so that at
214N/Astart-time the Unix loader is able to locate <CODE>libfoo.so</CODE> in
493N/A<CODE>/usr/lib</CODE>, in paths hard-coded via linker-options like
214N/A<CODE>-R</CODE> or in paths configured via the environment variable
214N/A<CODE>LD_LIBRARY_PATH</CODE>. It then resolves any (yet unresolved) symbols in
935N/Athe executable program which are available in the DSO.
935N/A
935N/A<P>Symbols in the executable program are usually not referenced by the DSO
935N/A(because it's a reusable library of general code) and hence no further
935N/Aresolving has to be done. The executable program has no need to do anything on
935N/Aits own to use the symbols from the DSO because the complete resolving is done
935N/Aby the Unix loader. (In fact, the code to invoke <CODE>ld.so</CODE> is part of
935N/Athe run-time startup code which is linked into every executable program which
935N/Ahas been bound non-static). The advantage of dynamic loading of common library
935N/Acode is obvious: the library code needs to be stored only once, in a system
935N/Alibrary like <CODE>libc.so</CODE>, saving disk space for every program.
935N/A
935N/A<P>In the second way the DSO's are usually called <EM>shared objects</EM> or
935N/A<EM>DSO files</EM> and can be named with an arbitrary extension (although the
935N/Acanonical name is <CODE>foo.so</CODE>). These files usually stay inside a
935N/Aprogram-specific directory and there is no automatically established link to
935N/Athe executable program where they are used. Instead the executable program
935N/Amanually loads the DSO at run-time into its address space via
935N/A<CODE>dlopen()</CODE>. At this time no resolving of symbols from the DSO for
935N/Athe executable program is done. But instead the Unix loader automatically
935N/Aresolves any (yet unresolved) symbols in the DSO from the set of symbols
935N/Aexported by the executable program and its already loaded DSO libraries
935N/A(especially all symbols from the ubiquitous <CODE>libc.so</CODE>).  This way
935N/Athe DSO gets knowledge of the executable program's symbol set as if it had
935N/Abeen statically linked with it in the first place.
935N/A
935N/A<P>Finally, to take advantage of the DSO's API the executable program has to
935N/Aresolve particular symbols from the DSO via <CODE>dlsym()</CODE> for later use
935N/Ainside dispatch tables <EM>etc.</EM> In other words: The executable program has to
214N/Amanually resolve every symbol it needs to be able to use it.  The advantage of
910N/Asuch a mechanism is that optional program parts need not be loaded (and thus
214N/Ado not spend memory) until they are needed by the program in question. When
935N/Arequired, these program parts can be loaded dynamically to extend the base
program's functionality.

<P>Although this DSO mechanism sounds straightforward there is at least one
difficult step here: The resolving of symbols from the executable program for
the DSO when using a DSO to extend a program (the second way). Why? Because
"reverse resolving" DSO symbols from the executable program's symbol set is
against the library design (where the library has no knowledge about the
programs it is used by) and is neither available under all platforms nor
standardized. In practice the executable program's global symbols are often
not re-exported and thus not available for use in a DSO.  Finding a way to
force the linker to export all global symbols is the main problem one has to
solve when using DSO for extending a program at run-time.

<H3>Practical Usage</H3>

<P>The shared library approach is the typical one, because it is what the DSO
mechanism was designed for, hence it is used for nearly all types of libraries
the operating system provides. On the other hand using shared objects for
extending a program is not used by a lot of programs.

<P>As of 1998 there are only a few software packages available which use the
DSO mechanism to actually extend their functionality at run-time: Perl 5 (via
its XS mechanism and the DynaLoader module), Netscape Server, <EM>etc.</EM>  Starting
with version 1.3, Apache joined the crew, because Apache already uses a module
concept to extend its functionality and internally uses a dispatch-list-based
approach to link external modules into the Apache core functionality. So,
Apache is really predestined for using DSO to load its modules at run-time.

<P>As of Apache 1.3, the configuration system supports two optional features
for taking advantage of the modular DSO approach: compilation of the Apache
core program into a DSO library for shared usage and compilation of the
Apache modules into DSO files for explicit loading at run-time.

<H3>Implementation</H3>

<P>The DSO support for loading individual Apache modules is based on a module
named <A HREF="mod/mod_so.html"><CODE>mod_so.c</CODE></A> which has to be
statically compiled into the Apache core. It is the only module besides
<CODE>http_core.c</CODE> which cannot be put into a DSO itself
(bootstrapping!). Practically all other distributed Apache modules then can
then be placed into a DSO by individually enabling the DSO build for them via
<CODE>configure</CODE>'s <CODE>--enable-shared</CODE> option (see top-level
<CODE>INSTALL</CODE> file) or by changing the <CODE>AddModule</CODE> command
in your <CODE>src/Configuration</CODE> into a <CODE>SharedModule</CODE>
command (see <CODE>src/INSTALL</CODE> file).  After a module is compiled into
a DSO named <CODE>mod_foo.so</CODE> you can use <A
HREF="mod/mod_so.html"><CODE>mod_so</CODE></A>'s <A
HREF="mod/mod_so.html#loadmodule"><CODE>LoadModule</CODE></A> command in your
<CODE>httpd.conf</CODE> file to load this module at server startup or restart.

<P>To simplify this creation of DSO files for Apache modules (especially for
third-party modules) a new support program named <CODE>apxs</CODE> (<EM>APache
eXtenSion</EM>) is available. It can be used to build DSO based modules
<EM>outside of</EM> the Apache source tree. The idea is simple: When
installing Apache the <CODE>configure</CODE>'s <CODE>make install</CODE>
procedure installs the Apache C header files and puts the platform-dependent
compiler and linker flags for building DSO files into the <CODE>apxs</CODE>
program.  This way the user can use <CODE>apxs</CODE> to compile his Apache
module sources without the Apache distribution source tree and without having
to fiddle with the platform-dependent compiler and linker flags for DSO
support.

<P>To place the complete Apache core program into a DSO library (only required
on some of the supported platforms to force the linker to export the apache
core symbols -- a prerequisite for the DSO modularization) the rule
<CODE>SHARED_CORE</CODE> has to be enabled via <CODE>configure</CODE>'s
<CODE>--enable-rule=SHARED_CORE</CODE> option (see top-level
<CODE>INSTALL</CODE> file) or by changing the <CODE>Rule</CODE> command in
your <CODE>Configuration</CODE> file to <CODE>Rule SHARED_CORE=yes</CODE> (see
<CODE>src/INSTALL</CODE> file). The Apache core code is then placed into a DSO
library named <CODE>libhttpd.so</CODE>. Because one cannot link a DSO against
static libraries on all platforms, an additional executable program named
<CODE>libhttpd.ep</CODE> is created which both binds this static code and
provides a stub for the <CODE>main()</CODE> function.  Finally the
<CODE>httpd</CODE> executable program itself is replaced by a bootstrapping
code which automatically makes sure the Unix loader is able to load and start
<CODE>libhttpd.ep</CODE> by providing the <CODE>LD_LIBRARY_PATH</CODE> to
<CODE>libhttpd.so</CODE>.

<H3>Supported Platforms</H3>

<P>Apache's <CODE>src/Configure</CODE> script currently has only limited but
adequate built-in knowledge on how to compile DSO files, because as already
mentioned this is heavily platform-dependent. Nevertheless all major Unix
platforms are supported.  The definitive current state (May 1999) is this:

<P>
<UL>
<LI>Out-of-the-box supported platforms:<BR>
(actually tested versions in parenthesis)

<PRE>
o  FreeBSD            (2.1.5, 2.2.x, 3.x, 4.x)
o  OpenBSD            (2.x)
o  NetBSD             (1.3.1)
o  BSDI               (3.x, 4.x)
o  Linux              (Debian/1.3.1, RedHat/4.2)
o  Solaris            (2.4, 2.5, 2.6, 2.7)
o  SunOS              (4.1.3)
o  Digital UNIX       (4.0)
o  IRIX               (5.3, 6.2)
o  HP/UX              (10.20)
o  UnixWare           (2.01, 2.1.2)
o  SCO                (5.0.4)
o  AIX                (3.2, 4.1.5, 4.2, 4.3)
o  ReliantUNIX/SINIX  (5.43)
o  SVR4               (-)
o  Mac OS X Server    (1.0)
o  Mac OS             (10.0 preview 1)
o  OpenStep/Mach      (4.2)
o  DGUX               (??)
o  NetWare            (5.1)
</PRE>

<P>
<LI> Explicitly unsupported platforms:

<PRE>
o  Ultrix             (no dlopen-style interface under this platform)
</PRE>

</UL>

<H3>Usage Summary</H3>

<P>To give you an overview of the DSO features of Apache 1.3, here is a short
and concise summary:

<OL>

<LI>Placing the Apache core code (all the stuff which usually forms the
<CODE>httpd</CODE> binary) into a DSO <CODE>libhttpd.so</CODE>, an executable
program <CODE>libhttpd.ep</CODE> and a bootstrapping executable program
<CODE>httpd</CODE> (Notice: this is only required on some of the supported
platforms to force the linker to export the Apache core symbols, which in turn
is a prerequisite for the DSO modularization):

<P>
<UL>
<LI>Build and install via <CODE>configure</CODE> (preferred):
<TABLE BGCOLOR="#f0f0f0" CELLPADDING=10><TR><TD>
<PRE>
$ /configure --prefix=/path/to/install
              --enable-rule=SHARED_CORE ...
$ make install
</PRE>
</TD></TR></TABLE>

<LI>Build and install manually:
<TABLE BGCOLOR="#f0f0f0" CELLPADDING=10><TR><TD>
<PRE>
- Edit src/Configuration:
  &lt;&lt; Rule SHARED_CORE=default
  &gt;&gt; Rule SHARED_CORE=yes
  &lt;&lt; EXTRA_CFLAGS=
  &gt;&gt; EXTRA_CFLAGS= -DSHARED_CORE_DIR=\"/path/to/install/modules\"
$ make
$ cp src/libhttpd.so* /path/to/install/modules/
$ cp src/libhttpd.ep  /path/to/install/modules/
$ cp src/httpd        /path/to/install/bin/
</PRE>
</TD></TR></TABLE>
</UL>

<LI>Build and install a <EM>distributed</EM> Apache module, say
<CODE>mod_foo.c</CODE>, into its own DSO <CODE>mod_foo.so</CODE>:

<P>
<UL>
<LI>Build and install via <CODE>configure</CODE> (preferred):
<TABLE BGCOLOR="#f0f0f0" CELLPADDING=10><TR><TD>
<PRE>
$ /configure --prefix=/path/to/install
        --enable-shared=foo
$ make install
</PRE>
</TD></TR></TABLE>

<LI>Build and install manually:
<TABLE BGCOLOR="#f0f0f0" CELLPADDING=10><TR><TD>
<PRE>
- Edit src/Configuration:
  &lt;&lt; AddModule    modules/xxxx/mod_foo.o
  &gt;&gt; SharedModule modules/xxxx/mod_foo.so
$ make
$ cp src/xxxx/mod_foo.so /path/to/install/modules
- Edit /path/to/install/etc/httpd.conf
  &gt;&gt; LoadModule foo_module /path/to/install/modules/mod_foo.so
</PRE>
</TD></TR></TABLE>
</UL>

<LI>Build and install a <EM>third-party</EM> Apache module, say
<CODE>mod_foo.c</CODE>, into its own DSO <CODE>mod_foo.so</CODE>

<P>
<UL>
<LI>Build and install via <CODE>configure</CODE> (preferred):
<TABLE BGCOLOR="#f0f0f0" CELLPADDING=10><TR><TD>
<PRE>
$ /configure --add-module=/path/to/3rdparty/mod_foo.c
        --enable-shared=foo
$ make install
</PRE>
</TD></TR></TABLE>

<LI>Build and install manually:
<TABLE BGCOLOR="#f0f0f0" CELLPADDING=10><TR><TD>
<PRE>
$ cp /path/to/3rdparty/mod_foo.c /path/to/apache-1.3/src/modules/extra/
- Edit src/Configuration:
  &gt;&gt; SharedModule modules/extra/mod_foo.so
$ make
$ cp src/xxxx/mod_foo.so /path/to/install/modules
- Edit /path/to/install/etc/httpd.conf
  &gt;&gt; LoadModule foo_module /path/to/install/modules/mod_foo.so
</PRE>
</TD></TR></TABLE>
</UL>

<P>
<LI>Build and install a <EM>third-party</EM> Apache module, say
<CODE>mod_foo.c</CODE>, into its own DSO <CODE>mod_foo.so</CODE> <EM>outside
of</EM> the Apache source tree:

<P>
<UL>
<LI>Build and install via <CODE>apxs</CODE>:
<TABLE BGCOLOR="#f0f0f0" CELLPADDING=10><TR><TD>
<PRE>
$ cd /path/to/3rdparty
$ apxs -c mod_foo.c
$ apxs -i -a -n foo mod_foo.so
</PRE>
</TD></TR></TABLE>
</UL>

</OL>

<H3>Advantages & Disadvantages</H3>

<P>The above DSO based features of Apache 1.3 have the following advantages:

<UL>
<LI> The server package is more flexible at run-time because the actual server
     process can be assembled at run-time via <A
     HREF="mod/mod_so.html#loadmodule"><CODE>LoadModule</CODE></A>
     <CODE>httpd.conf</CODE> configuration commands instead of
     <CODE>Configuration</CODE> <CODE>AddModule</CODE> commands at build-time.
     For instance this way one is able to run different server instances
     (standard &amp; SSL version, minimalistic &amp; powered up version
     [mod_perl, PHP3], <EM>etc.</EM>) with only one Apache installation.
<P>
<LI> The server package can be easily extended with third-party modules even
     after installation. This is at least a great benefit for vendor package
     maintainers who can create a Apache core package and additional packages
     containing extensions like PHP3, mod_perl, mod_fastcgi, <EM>etc.</EM>
<P>
<LI> Easier Apache module prototyping because with the DSO/<CODE>apxs</CODE>
     pair you can both work outside the Apache source tree and only need an
     <CODE>apxs -i</CODE> command followed by an <CODE>apachectl
     restart</CODE> to bring a new version of your currently developed module
     into the running Apache server.
</UL>

<P>DSO has the following disadvantages:

<UL>
<LI> The DSO mechanism cannot be used on every platform because not all
     operating systems support dynamic loading of code into the address space
     of a program.
<P>
<LI> The server is approximately 20% slower at startup time because of the
     symbol resolving overhead the Unix loader now has to do.
<P>
<LI> The server is approximately 5% slower at execution time under some
     platforms because position independent code (PIC) sometimes needs
     complicated assembler tricks for relative addressing which are not
     necessarily as fast as absolute addressing.
<P>
<LI> Because DSO modules cannot be linked against other DSO-based libraries
     (<CODE>ld -lfoo</CODE>) on all platforms (for instance a.out-based
     platforms usually don't provide this functionality while ELF-based
     platforms do) you cannot use the DSO mechanism for all types of modules.
     Or in other words, modules compiled as DSO files are restricted to only
     use symbols from the Apache core, from the C library (<CODE>libc</CODE>)
     and all other dynamic or static libraries used by the Apache core, or
     from static library archives (<CODE>libfoo.a</CODE>) containing position
     independent code. The only chances to use other code is to either make
     sure the Apache core itself already contains a reference to it, loading
     the code yourself via <CODE>dlopen()</CODE> or enabling the
     <CODE>SHARED_CHAIN</CODE> rule while building Apache when your platform
     supports linking DSO files against DSO libraries.
<P>
<LI> Under some platforms (many SVR4 systems) there is no way to force the
     linker to export all global symbols for use in DSO's when linking the
     Apache httpd executable program. But without the visibility of the Apache
     core symbols no standard Apache module could be used as a DSO. The only
     chance here is to use the <CODE>SHARED_CORE</CODE> feature because this
     way the global symbols are forced to be exported. As a consequence the
     Apache <CODE>src/Configure</CODE> script automatically enforces
     <CODE>SHARED_CORE</CODE> on these platforms when DSO features are used in
     the <CODE>Configuration</CODE> file or on the configure command line.
</UL>

<!--#include virtual="footer.html" -->
</BLOCKQUOTE>
</BODY>
</HTML>