path: root/zh_TW.UTF-8/books/developers-handbook/sockets/chapter.xml

<?xml version="1.0" encoding="utf-8"?>
<!--
     The FreeBSD Documentation Project

     $FreeBSD$
-->
<chapter xmlns="http://docbook.org/ns/docbook" xmlns:xlink="http://www.w3.org/1999/xlink" version="5.0" xml:id="sockets">
  <info><title>Sockets</title>
    <authorgroup>
      <author><personname><firstname>G. Adam</firstname><surname>Stanislav</surname></personname><contrib>Contributed by </contrib></author>
    </authorgroup>
  </info>

  

  <sect1 xml:id="sockets-synopsis">
    <title>Synopsis</title>

    <para><acronym>BSD</acronym> sockets take interprocess
      communications to a new level. It is no longer necessary for the
      communicating processes to run on the same machine. They still
      <emphasis>can</emphasis>, but they do not have to.</para>

    <para>Not only do these processes not have to run on the same
      machine, they do not have to run under the same operating
      system. Thanks to <acronym>BSD</acronym> sockets, your FreeBSD
      software can smoothly cooperate with a program running on a
      &macintosh;, another one running on a &sun; workstation, yet another
      one running under &windows; 2000, all connected with an
      Ethernet-based local area network.</para>

    <para>But your software can equally well cooperate with processes
      running in another building, or on another continent, inside a
      submarine, or a space shuttle.</para>

    <para>It can also cooperate with processes that are not part of a
      computer (at least not in the strict sense of the word), but of
      such devices as printers, digital cameras, medical equipment.
      Just about anything capable of digital communications.</para>

  </sect1>

  <sect1 xml:id="sockets-diversity">
    <title>Networking and Diversity</title>

    <para>We have already hinted on the <emphasis>diversity</emphasis>
      of networking. Many different systems have to talk to each
      other. And they have to speak the same language. They also have
      to <emphasis>understand</emphasis> the same language the same
      way.</para>

    <para>People often think that <emphasis>body language</emphasis>
      is universal. But it is not. Back in my early teens, my father
      took me to Bulgaria. We were sitting at a table in a park in
      Sofia, when a vendor approached us trying to sell us some
      roasted almonds.</para>

    <para>I had not learned much Bulgarian by then, so, instead of
      saying no, I shook my head from side to side, the
      <quote>universal</quote> body language for
      <emphasis>no</emphasis>. The vendor quickly started serving us
      some almonds.</para>

    <para>I then remembered I had been told that in Bulgaria shaking
      your head sideways meant <emphasis>yes</emphasis>. Quickly, I
      started nodding my head up and down. The vendor noticed, took
      his almonds, and walked away. To an uninformed observer, I did
      not change the body language: I continued using the language of
      shaking and nodding my head. What changed was the
      <emphasis>meaning</emphasis> of the body language. At first, the
      vendor and I interpreted the same language as having completely
      different meaning. I had to adjust my own interpretation of that
      language so the vendor would understand.</para>

    <para>It is the same with computers: The same symbols may have
      different, even outright opposite meaning. Therefore, for
      two computers to understand each other, they must not only
      agree on the same <emphasis>language</emphasis>, but on the
      same <emphasis>interpretation</emphasis> of the language.
      </para>
  </sect1>

  <sect1 xml:id="sockets-protocols">
    <title>Protocols</title>

    <para>While various programming languages tend to have complex
      syntax and use a number of multi-letter reserved words (which
      makes them easy for the human programmer to understand), the
      languages of data communications tend to be very terse. Instead
      of multi-byte words, they often use individual
      <emphasis>bits</emphasis>.  There is a very convincing reason
      for it: While data travels <emphasis>inside</emphasis> your
      computer at speeds approaching the speed of light, it often
      travels considerably slower between two computers.</para>

    <para>Because the languages used in data communications are so
      terse, we usually refer to them as
      <emphasis>protocols</emphasis> rather than languages.</para>

    <para>As data travels from one computer to another, it always uses
      more than one protocol. These protocols are
      <emphasis>layered</emphasis>.  The data can be compared to the
      inside of an onion: You have to peel off several layers of
      <quote>skin</quote> to get to the data.  This is best
      illustrated with a picture:</para>

    <mediaobject>
      <imageobject>
        <imagedata fileref="sockets/layers"/>
      </imageobject>

      <textobject>
        <literallayout class="monospaced">+----------------+
|    Ethernet    |
|+--------------+|
||      IP      ||
||+------------+||
|||     TCP    |||
|||+----------+|||
||||   HTTP   ||||
||||+--------+||||
|||||   PNG  |||||
|||||+------+|||||
|||||| Data ||||||
|||||+------+|||||
||||+--------+||||
|||+----------+|||
||+------------+||
|+--------------+|
+----------------+</literallayout>
      </textobject>

      <textobject>
        <phrase>Protocol Layers</phrase>
      </textobject>
    </mediaobject>

    <para>In this example, we are trying to get an image from a web
      page we are connected to via an Ethernet.</para>

    <para>The image consists of raw data, which is simply a sequence
      of <acronym>RGB</acronym> values that our software can process,
      i.e., convert into an image and display on our monitor.</para>

    <para>Alas, our software has no way of knowing how the raw data is
      organized: Is it a sequence of <acronym>RGB</acronym> values, or
      a sequence of grayscale intensities, or perhaps of
      <acronym>CMYK</acronym> encoded colors? Is the data represented
      by 8-bit quanta, or are they 16 bits in size, or perhaps 4 bits?
      How many rows and columns does the image consist of? Should
      certain pixels be transparent?</para>

    <para>I think you get the picture...</para>

    <para>To inform our software how to handle the raw data, it is
      encoded as a <acronym>PNG</acronym> file. It could be a
      <acronym>GIF</acronym>, or a <acronym>JPEG</acronym>, but it is
      a <acronym>PNG</acronym>.</para>

    <para>And <acronym>PNG</acronym> is a protocol.</para>

    <para>At this point, I can hear some of you yelling,
      <emphasis><quote>No, it is not! It is a file
      format!</quote></emphasis></para>

    <para>Well, of course it is a file format. But from the
      perspective of data communications, a file format is a protocol:
      The file structure is a <emphasis>language</emphasis>, a terse
      one at that, communicating to our <emphasis>process</emphasis>
      how the data is organized. Ergo, it is a
      <emphasis>protocol</emphasis>.</para>

    <para>Alas, if all we received was the <acronym>PNG</acronym>
      file, our software would be facing a serious problem: How is it
      supposed to know the data is representing an image, as opposed
      to some text, or perhaps a sound, or what not? Secondly, how is
      it supposed to know the image is in the <acronym>PNG</acronym>
      format as opposed to <acronym>GIF</acronym>, or
      <acronym>JPEG</acronym>, or some other image format?</para>

    <para>To obtain that information, we are using another protocol:
      <acronym>HTTP</acronym>. This protocol can tell us exactly that
      the data represents an image, and that it uses the
      <acronym>PNG</acronym> protocol. It can also tell us some other
      things, but let us stay focused on protocol layers here.
      </para>

    <para>So, now we have some data wrapped in the <acronym>PNG</acronym>
      protocol, wrapped in the <acronym>HTTP</acronym> protocol.
      How did we get it from the server?</para>

    <para>By using <acronym>TCP/IP</acronym> over Ethernet, that is
      how.  Indeed, that is three more protocols. Instead of
      continuing inside out, I am now going to talk about Ethernet,
      simply because it is easier to explain the rest that way.</para>

    <para>Ethernet is an interesting system of connecting computers in
      a <emphasis>local area network</emphasis>
      (<acronym>LAN</acronym>).  Each computer has a <emphasis>network
      interface card</emphasis> (<acronym>NIC</acronym>), which has a
      unique 48-bit <acronym>ID</acronym> called its
      <emphasis>address</emphasis>. No two Ethernet
      <acronym>NIC</acronym>s in the world have the same address.
      </para>

    <para>These <acronym>NIC</acronym>s are all connected with each
      other. Whenever one computer wants to communicate with another
      in the same Ethernet <acronym>LAN</acronym>, it sends a message
      over the network. Every <acronym>NIC</acronym> sees the
      message. But as part of the Ethernet
      <emphasis>protocol</emphasis>, the data contains the address of
      the destination <acronym>NIC</acronym> (among other things). So,
      only one of all the network interface cards will pay attention
      to it, the rest will ignore it.</para>

    <para>But not all computers are connected to the same
      network. Just because we have received the data over our
      Ethernet does not mean it originated in our own local area
      network. It could have come to us from some other network (which
      may not even be Ethernet based) connected with our own network
      via the Internet.</para>

    <para>All data is transferred over the Internet using
      <acronym>IP</acronym>, which stands for <emphasis>Internet
      Protocol</emphasis>. Its basic role is to let us know where in
      the world the data has arrived from, and where it is supposed to
      go to. It does not <emphasis>guarantee</emphasis> we will
      receive the data, only that we will know where it came from
      <emphasis>if</emphasis> we do receive it.</para>

    <para>Even if we do receive the data, <acronym>IP</acronym> does
      not guarantee we will receive various chunks of data in the same
      order the other computer has sent it to us. So, we can receive
      the center of our image before we receive the upper left corner
      and after the lower right, for example.</para>

    <para>It is <acronym>TCP</acronym> (<emphasis>Transmission Control
      Protocol</emphasis>) that asks the sender to resend any lost
      data and that places it all into the proper order.</para>

    <para>All in all, it took <emphasis>five</emphasis> different
      protocols for one computer to communicate to another what an
      image looks like.  We received the data wrapped into the
      <acronym>PNG</acronym> protocol, which was wrapped into the
      <acronym>HTTP</acronym> protocol, which was wrapped into the
      <acronym>TCP</acronym> protocol, which was wrapped into the
      <acronym>IP</acronym> protocol, which was wrapped into the
      <acronym>Ethernet</acronym> protocol.</para>

    <para>Oh, and by the way, there probably were several other
      protocols involved somewhere on the way. For example, if our
      <acronym>LAN</acronym> was connected to the Internet through a
      dial-up call, it used the <acronym>PPP</acronym> protocol over
      the modem which used one (or several) of the various modem
      protocols, et cetera, et cetera, et cetera...</para>

    <para>As a developer you should be asking by now,
      <emphasis><quote>How am I supposed to handle it
      all?</quote></emphasis></para>

    <para>Luckily for you, you are <emphasis>not</emphasis> supposed
      to handle it all. You <emphasis>are</emphasis> supposed to
      handle some of it, but not all of it. Specifically, you need not
      worry about the physical connection (in our case Ethernet and
      possibly <acronym>PPP</acronym>, etc). Nor do you need to handle
      the Internet Protocol, or the Transmission Control
      Protocol.</para>

    <para>In other words, you do not have to do anything to receive
      the data from the other computer. Well, you do have to
      <emphasis>ask</emphasis> for it, but that is almost as simple as
      opening a file.</para>

    <para>Once you have received the data, it is up to you to figure
      out what to do with it. In our case, you would need to
      understand the <acronym>HTTP</acronym> protocol and the
      <acronym>PNG</acronym> file structure.</para>

    <para>To use an analogy, all the internetworking protocols become
      a gray area: Not so much because we do not understand how it
      works, but because we are no longer concerned about it. The
      sockets interface takes care of this gray area for us:</para>

    <mediaobject>
      <imageobject>
        <imagedata fileref="sockets/slayers"/>
      </imageobject>

      <textobject>
        <literallayout class="monospaced">+----------------+
|xxxxEthernetxxxx|
|+--------------+|
||xxxxxxIPxxxxxx||
||+------------+||
|||xxxxxTCPxxxx|||
|||+----------+|||
||||   HTTP   ||||
||||+--------+||||
|||||   PNG  |||||
|||||+------+|||||
|||||| Data ||||||
|||||+------+|||||
||||+--------+||||
|||+----------+|||
||+------------+||
|+--------------+|
+----------------+</literallayout>
      </textobject>

      <textobject>
        <phrase>Sockets Covered Protocol Layers</phrase>
      </textobject>
    </mediaobject>

    <para>We only need to understand any protocols that tell us how to
      <emphasis>interpret the data</emphasis>, not how to
      <emphasis>receive</emphasis> it from another process, nor how to
      <emphasis>send</emphasis> it to another process.</para>

  </sect1>

  <sect1 xml:id="sockets-model">
    <title>The Sockets Model</title>

    <para><acronym>BSD</acronym> sockets are built on the basic &unix;
      model: <emphasis>Everything is a file.</emphasis> In our
      example, then, sockets would let us receive an <emphasis>HTTP
      file</emphasis>, so to speak. It would then be up to us to
      extract the <emphasis><acronym>PNG</acronym> file</emphasis>
      from it.
      </para>

    <para>Because of the complexity of internetworking, we cannot just
      use the <function role="syscall">open</function> system call, or
      the <function>open()</function> C function. Instead, we need to
      take several steps to <quote>opening</quote> a socket.</para>

    <para>Once we do, however, we can start treating the
      <emphasis>socket</emphasis> the same way we treat any
      <emphasis>file descriptor</emphasis>: We can
      <function>read</function> from it, <function>write</function> to
      it, <function>pipe</function> it, and, eventually,
      <function>close</function> it.</para>

  </sect1>

  <sect1 xml:id="sockets-essential-functions">
    <title>Essential Socket Functions</title>

    <para>While FreeBSD offers different functions to work with
      sockets, we only <emphasis>need</emphasis> four to
      <quote>open</quote> a socket.  And in some cases we only need
      two.</para>

    <sect2 xml:id="sockets-client-server">
      <title>The Client-Server Difference</title>

      <para>Typically, one of the ends of a socket-based data
        communication is a <emphasis>server</emphasis>, the other is a
        <emphasis>client</emphasis>.</para>

      <sect3 xml:id="sockets-common-elements">
        <title>The Common Elements</title>

        <sect4 xml:id="sockets-socket">
	  <title><function>socket</function></title>

	  <para>The one function used by both, clients and servers, is
	    &man.socket.2;. It is declared this way:</para>

<programlisting>
int socket(int domain, int type, int protocol);
</programlisting>

          <para>The return value is of the same type as that of
	    <function>open</function>, an integer. FreeBSD allocates
	    its value from the same pool as that of file handles.
	    That is what allows sockets to be treated the same way as
	    files.</para>

	  <para>The <varname>domain</varname> argument tells the
	    system what <emphasis>protocol family</emphasis> you want
	    it to use. Many of them exist, some are vendor specific,
	    others are very common. They are declared in
	    <filename>sys/socket.h</filename>.</para>

	  <para>Use <constant>PF_INET</constant> for
	    <acronym>UDP</acronym>, <acronym>TCP</acronym> and other
	    Internet protocols (<acronym>IP</acronym>v4).</para>

	  <para>Five values are defined for the
	    <varname>type</varname> argument, again, in
	    <filename>sys/socket.h</filename>.  All of them start with
	    <quote><constant>SOCK_</constant></quote>.  The most
	    common one is <constant>SOCK_STREAM</constant>, which
	    tells the system you are asking for a <emphasis>reliable
	    stream delivery service</emphasis> (which is
	    <acronym>TCP</acronym> when used with
	    <constant>PF_INET</constant>).</para>

	  <para>If you asked for <constant>SOCK_DGRAM</constant>, you
	    would be requesting a <emphasis>connectionless datagram
	    delivery service</emphasis> (in our case,
	    <acronym>UDP</acronym>).</para>

	  <para>If you wanted to be in charge of the low-level
	    protocols (such as <acronym>IP</acronym>), or even network
	    interfaces (e.g., the Ethernet), you would need to specify
	    <constant>SOCK_RAW</constant>.</para>

	  <para>Finally, the <varname>protocol</varname> argument
	    depends on the previous two arguments, and is not always
	    meaningful.  In that case, use <constant>0</constant> for
	    its value.</para>

	  <note xml:id="sockets-unconnected">
	    <title>The Unconnected Socket</title>

	    <para>Nowhere, in the <function>socket</function> function
	      have we specified to what other system we should be
	      connected.  Our newly created socket remains
	      <emphasis>unconnected</emphasis>.</para>

	    <para>This is on purpose: To use a telephone analogy, we
	      have just attached a modem to the phone line. We have
	      neither told the modem to make a call, nor to answer if
	      the phone rings.</para>
	  </note>

	</sect4>

        <sect4 xml:id="sockets-sockaddr">
	  <title><varname>sockaddr</varname></title>

	  <para>Various functions of the sockets family expect the
 	    address of (or pointer to, to use C terminology) a small
 	    area of the memory. The various C declarations in the
 	    <filename>sys/socket.h</filename> refer to it as
 	    <varname>struct sockaddr</varname>. This structure is
 	    declared in the same file:</para>

<programlisting>
/*
 * Structure used by kernel to store most
 * addresses.
 */
struct sockaddr {
	unsigned char	sa_len;		/* total length */
	sa_family_t	sa_family;	/* address family */
	char		sa_data[14];	/* actually longer; address value */
};
#define	SOCK_MAXADDRLEN	255		/* longest possible addresses */
</programlisting>

          <para>Please note the <emphasis>vagueness</emphasis> with
	    which the <varname>sa_data</varname> field is declared,
	    just as an array of <constant>14</constant> bytes, with
	    the comment hinting there can be more than
	    <constant>14</constant> of them.</para>

	  <para>This vagueness is quite deliberate. Sockets is a very
	    powerful interface. While most people perhaps think of it
	    as nothing more than the Internet interface&mdash;and most
	    applications probably use it for that
	    nowadays&mdash;sockets can be used for just about
	    <emphasis>any</emphasis> kind of interprocess
	    communications, of which the Internet (or, more precisely,
	    <acronym>IP</acronym>) is only one.</para>

	  <para>The <filename>sys/socket.h</filename> refers to the
	    various types of protocols sockets will handle as
	    <emphasis>address families</emphasis>, and lists them
	    right before the definition of
	    <varname>sockaddr</varname>:</para>

<programlisting>
/*
 * Address families.
 */
#define	AF_UNSPEC	0		/* unspecified */
#define	AF_LOCAL	1		/* local to host (pipes, portals) */
#define	AF_UNIX		AF_LOCAL	/* backward compatibility */
#define	AF_INET		2		/* internetwork: UDP, TCP, etc. */
#define	AF_IMPLINK	3		/* arpanet imp addresses */
#define	AF_PUP		4		/* pup protocols: e.g. BSP */
#define	AF_CHAOS	5		/* mit CHAOS protocols */
#define	AF_NS		6		/* XEROX NS protocols */
#define	AF_ISO		7		/* ISO protocols */
#define	AF_OSI		AF_ISO
#define	AF_ECMA		8		/* European computer manufacturers */
#define	AF_DATAKIT	9		/* datakit protocols */
#define	AF_CCITT	10		/* CCITT protocols, X.25 etc */
#define	AF_SNA		11		/* IBM SNA */
#define AF_DECnet	12		/* DECnet */
#define AF_DLI		13		/* DEC Direct data link interface */
#define AF_LAT		14		/* LAT */
#define	AF_HYLINK	15		/* NSC Hyperchannel */
#define	AF_APPLETALK	16		/* Apple Talk */
#define	AF_ROUTE	17		/* Internal Routing Protocol */
#define	AF_LINK		18		/* Link layer interface */
#define	pseudo_AF_XTP	19		/* eXpress Transfer Protocol (no AF) */
#define	AF_COIP		20		/* connection-oriented IP, aka ST II */
#define	AF_CNT		21		/* Computer Network Technology */
#define pseudo_AF_RTIP	22		/* Help Identify RTIP packets */
#define	AF_IPX		23		/* Novell Internet Protocol */
#define	AF_SIP		24		/* Simple Internet Protocol */
#define	pseudo_AF_PIP	25		/* Help Identify PIP packets */
#define	AF_ISDN		26		/* Integrated Services Digital Network*/
#define	AF_E164		AF_ISDN		/* CCITT E.164 recommendation */
#define	pseudo_AF_KEY	27		/* Internal key-management function */
#define	AF_INET6	28		/* IPv6 */
#define	AF_NATM		29		/* native ATM access */
#define	AF_ATM		30		/* ATM */
#define pseudo_AF_HDRCMPLT 31		/* Used by BPF to not rewrite headers
					 * in interface output routine
					 */
#define	AF_NETGRAPH	32		/* Netgraph sockets */
#define	AF_SLOW		33		/* 802.3ad slow protocol */
#define	AF_SCLUSTER	34		/* Sitara cluster protocol */
#define	AF_ARP		35
#define	AF_BLUETOOTH	36		/* Bluetooth sockets */
#define	AF_MAX		37

</programlisting>

          <para>The one used for <acronym>IP</acronym> is
	    <symbol>AF_INET</symbol>.  It is a symbol for the constant
	    <constant>2</constant>.</para>

	  <para>It is the <emphasis>address family</emphasis> listed
	    in the <varname>sa_family</varname> field of
	    <varname>sockaddr</varname> that decides how exactly the
	    vaguely named bytes of <varname>sa_data</varname> will be
	    used.</para>

	  <para>Specifically, whenever the <emphasis>address
	    family</emphasis> is <symbol>AF_INET</symbol>, we can use
	    <varname>struct sockaddr_in</varname> found in
	    <filename>netinet/in.h</filename>, wherever
	    <varname>sockaddr</varname> is expected:</para>

<programlisting>
/*
 * Socket address, internet style.
 */
struct sockaddr_in {
	uint8_t		sin_len;
	sa_family_t	sin_family;
	in_port_t	sin_port;
	struct	in_addr sin_addr;
	char	sin_zero[8];
};
</programlisting>

          <para>We can visualize its organization this way:</para>

          <mediaobject>
           <imageobject>
             <imagedata fileref="sockets/sain"/>
           </imageobject>

           <textobject>
             <literallayout class="monospaced">        0        1        2       3
   +--------+--------+-----------------+
 0 |    0   | Family |       Port      |
   +--------+--------+-----------------+
 4 |             IP Address            |
   +-----------------------------------+
 8 |                 0                 |
   +-----------------------------------+
12 |                 0                 |
   +-----------------------------------+</literallayout>
           </textobject>

           <textobject>
             <phrase>sockaddr_in</phrase>
           </textobject>
         </mediaobject>

          <para>The three important fields are
	    <varname>sin_family</varname>, which is byte 1 of the
	    structure, <varname>sin_port</varname>, a 16-bit value
	    found in bytes 2 and 3, and <varname>sin_addr</varname>, a
	    32-bit integer representation of the <acronym>IP</acronym>
	    address, stored in bytes 4-7.</para>

	  <para>Now, let us try to fill it out. Let us assume we are
	    trying to write a client for the
	    <emphasis>daytime</emphasis> protocol, which simply states
	    that its server will write a text string representing the
	    current date and time to port 13. We want to use
	    <acronym>TCP/IP</acronym>, so we need to specify
	    <constant>AF_INET</constant> in the address family
	    field. <constant>AF_INET</constant> is defined as
	    <constant>2</constant>. Let us use the
	    <acronym>IP</acronym> address of <systemitem class="ipaddress">192.43.244.18</systemitem>, which is the time
	    server of US federal government (<systemitem class="fqdomainname">time.nist.gov</systemitem>).</para>

          <mediaobject>
            <imageobject>
              <imagedata fileref="sockets/sainfill"/>
            </imageobject>

            <textobject>
              <literallayout class="monospaced">        0        1        2       3
   +--------+--------+-----------------+
 0 |    0   |   2    |        13       |
   +-----------------+-----------------+
 4 |           192.43.244.18           |
   +-----------------------------------+
 8 |                 0                 |
   +-----------------------------------+
12 |                 0                 |
   +-----------------------------------+</literallayout>
            </textobject>

            <textobject>
              <phrase>Specific example of sockaddr_in</phrase>
            </textobject>
          </mediaobject>

          <para>By the way the <varname>sin_addr</varname> field is
 	    declared as being of the <varname>struct in_addr</varname>
 	    type, which is defined in
 	    <filename>netinet/in.h</filename>:</para>

<programlisting>
/*
 * Internet address (a structure for historical reasons)
 */
struct in_addr {
	in_addr_t s_addr;
};
</programlisting>

          <para>In addition, <varname>in_addr_t</varname> is a 32-bit
            integer.</para>

	  <para>The <systemitem class="ipaddress">192.43.244.18</systemitem> is
	    just a convenient notation of expressing a 32-bit integer
	    by listing all of its 8-bit bytes, starting with the
	    <emphasis>most significant</emphasis> one.</para>

          <para>So far, we have viewed <varname>sockaddr</varname> as
	    an abstraction.  Our computer does not store
	    <varname>short</varname> integers as a single 16-bit
	    entity, but as a sequence of 2 bytes. Similarly, it stores
	    32-bit integers as a sequence of 4 bytes.</para>

          <para>Suppose we coded something like this:</para>

<programlisting>
	sa.sin_family      = AF_INET;
	sa.sin_port        = 13;
	sa.sin_addr.s_addr = (((((192 &lt;&lt; 8) | 43) &lt;&lt; 8) | 244) &lt;&lt; 8) | 18;
</programlisting>

          <para>What would the result look like?</para>

	  <para>Well, that depends, of course. On a &pentium;, or other
	    x86, based computer, it would look like this:</para>

          <mediaobject>
            <imageobject>
              <imagedata fileref="sockets/sainlsb"/>
            </imageobject>

            <textobject>
              <literallayout class="monospaced">        0        1        2       3
   +--------+--------+--------+--------+
 0 |    0   |   2    |   13   |   0    |
   +--------+--------+--------+--------+
 4 |   18   |  244   |   43   |  192   |
   +-----------------------------------+
 8 |                 0                 |
   +-----------------------------------+
12 |                 0                 |
   +-----------------------------------+</literallayout>
            </textobject>

            <textobject>
              <phrase>sockaddr_in on an Intel system</phrase>
            </textobject>
          </mediaobject>

          <para>On a different system, it might look like this:
	    </para>

          <mediaobject>
            <imageobject>
              <imagedata fileref="sockets/sainmsb"/>
            </imageobject>

            <textobject>
              <literallayout class="monospaced">        0        1        2       3
   +--------+--------+--------+--------+
 0 |    0   |   2    |    0   |   13   |
   +--------+--------+--------+--------+
 4 |   192  |   43   |   244  |   18   |
   +-----------------------------------+
 8 |                 0                 |
   +-----------------------------------+
12 |                 0                 |
   +-----------------------------------+</literallayout>
            </textobject>

            <textobject>
              <phrase>sockaddr_in on an MSB system</phrase>
            </textobject>
          </mediaobject>

	  <para>And on a PDP it might look different yet. But the
	    above two are the most common ways in use today.</para>

	  <para>Ordinarily, wanting to write portable code,
	    programmers pretend that these differences do not
	    exist. And they get away with it (except when they code in
	    assembly language). Alas, you cannot get away with it that
	    easily when coding for sockets.</para>

	  <para>Why?</para>

	  <para>Because when communicating with another computer, you
	    usually do not know whether it stores data <emphasis>most
	    significant byte</emphasis> (<acronym>MSB</acronym>) or
	    <emphasis>least significant byte</emphasis>
	    (<acronym>LSB</acronym>) first.</para>

	  <para>You might be wondering, <emphasis><quote>So, will
	    sockets not handle it for me?</quote></emphasis></para>

	  <para>It will not.</para>

	  <para>While that answer may surprise you at first, remember
 	    that the general sockets interface only understands the
 	    <varname>sa_len</varname> and <varname>sa_family</varname>
 	    fields of the <varname>sockaddr</varname> structure. You
 	    do not have to worry about the byte order there (of
 	    course, on FreeBSD <varname>sa_family</varname> is only 1
 	    byte anyway, but many other &unix; systems do not have
 	    <varname>sa_len</varname> and use 2 bytes for
 	    <varname>sa_family</varname>, and expect the data in
 	    whatever order is native to the computer).</para>

	  <para>But the rest of the data is just
	    <varname>sa_data[14]</varname> as far as sockets
	    goes. Depending on the <emphasis>address
	    family</emphasis>, sockets just forwards that data to its
	    destination.</para>

	  <para>Indeed, when we enter a port number, it is because we
	    want the other computer to know what service we are asking
	    for. And, when we are the server, we read the port number
	    so we know what service the other computer is expecting
	    from us. Either way, sockets only has to forward the port
	    number as data. It does not interpret it in any way.</para>

	  <para>Similarly, we enter the <acronym>IP</acronym> address
	    to tell everyone on the way where to send our data
	    to. Sockets, again, only forwards it as data.</para>

	  <para>That is why, we (the <emphasis>programmers</emphasis>,
	    not the <emphasis>sockets</emphasis>) have to distinguish
	    between the byte order used by our computer and a
	    conventional byte order to send the data in to the other
	    computer.</para>

	  <para>We will call the byte order our computer uses the
	    <emphasis>host byte order</emphasis>, or just the
	    <emphasis>host order</emphasis>.</para>

	  <para>There is a convention of sending the multi-byte data
	    over <acronym>IP</acronym>
	    <emphasis><acronym>MSB</acronym> first</emphasis>.  This,
	    we will refer to as the <emphasis>network byte
	    order</emphasis>, or simply the <emphasis>network
	    order</emphasis>.</para>

	  <para>Now, if we compiled the above code for an Intel based
	    computer, our <emphasis>host byte order</emphasis> would
	    produce:</para>

	  <mediaobject>
	    <imageobject>
              <imagedata fileref="sockets/sainlsb"/>
            </imageobject>

            <textobject>
              <literallayout class="monospaced">        0        1        2       3
   +--------+--------+--------+--------+
 0 |    0   |   2    |   13   |   0    |
   +--------+--------+--------+--------+
 4 |   18   |  244   |   43   |  192   |
   +-----------------------------------+
 8 |                 0                 |
   +-----------------------------------+
12 |                 0                 |
   +-----------------------------------+</literallayout>
            </textobject>

            <textobject>
              <phrase>Host byte order on an Intel system</phrase>
            </textobject>
          </mediaobject>

          <para>But the <emphasis>network byte order</emphasis>
 	    requires that we store the data <acronym>MSB</acronym>
 	    first:</para>

          <mediaobject>
            <imageobject>
              <imagedata fileref="sockets/sainmsb"/>
            </imageobject>

            <textobject>
              <literallayout class="monospaced">        0        1        2       3
   +--------+--------+--------+--------+
 0 |    0   |   2    |    0   |   13   |
   +--------+--------+--------+--------+
 4 |   192  |   43   |   244  |   18   |
   +-----------------------------------+
 8 |                 0                 |
   +-----------------------------------+
12 |                 0                 |
   +-----------------------------------+</literallayout>
            </textobject>

            <textobject>
              <phrase>Network byte order</phrase>
            </textobject>
          </mediaobject>

          <para>Unfortunately, our <emphasis>host order</emphasis> is
	    the exact opposite of the <emphasis>network
	    order</emphasis>.</para>

	  <para>We have several ways of dealing with it. One would be
	    to <emphasis>reverse</emphasis> the values in our code:
	    </para>

<programlisting>
	sa.sin_family      = AF_INET;
	sa.sin_port        = 13 &lt;&lt; 8;
	sa.sin_addr.s_addr = (((((18 &lt;&lt; 8) | 244) &lt;&lt; 8) | 43) &lt;&lt; 8) | 192;
</programlisting>

           <para>This will <emphasis>trick</emphasis> our compiler
	     into storing the data in the <emphasis>network byte
	     order</emphasis>. In some cases, this is exactly the way
	     to do it (e.g., when programming in assembly
	     language). In most cases, however, it can cause a
	     problem.</para>

	   <para>Suppose, you wrote a sockets-based program in C. You
	     know it is going to run on a &pentium;, so you enter all
	     your constants in reverse and force them to the
	     <emphasis>network byte order</emphasis>. It works
	     well.</para>

           <para>Then, some day, your trusted old &pentium; becomes a
	     rusty old &pentium;.  You replace it with a system whose
	     <emphasis>host order</emphasis> is the same as the
	     <emphasis>network order</emphasis>. You need to recompile
	     all your software. All of your software continues to
	     perform well, except the one program you wrote.</para>

	   <para>You have since forgotten that you had forced all of
	     your constants to the opposite of the <emphasis>host
	     order</emphasis>. You spend some quality time tearing out
	     your hair, calling the names of all gods you ever heard
	     of (and some you made up), hitting your monitor with a
	     nerf bat, and performing all the other traditional
	     ceremonies of trying to figure out why something that has
	     worked so well is suddenly not working at all.</para>

	   <para>Eventually, you figure it out, say a couple of swear
	     words, and start rewriting your code.</para>

	   <para>Luckily, you are not the first one to face the
	     problem. Someone else has created the &man.htons.3; and
	     &man.htonl.3; C functions to convert a
	     <varname>short</varname> and <varname>long</varname>
	     respectively from the <emphasis>host byte
	     order</emphasis> to the <emphasis>network byte
	     order</emphasis>, and the &man.ntohs.3; and &man.ntohl.3;
	     C functions to go the other way.</para>

	   <para>On <emphasis><acronym>MSB</acronym>-first</emphasis>
	     systems these functions do nothing. On
	     <emphasis><acronym>LSB</acronym>-first</emphasis> systems
	     they convert values to the proper order.</para>

	   <para>So, regardless of what system your software is
	     compiled on, your data will end up in the correct order
	     if you use these functions.</para>

	 </sect4>

      </sect3>

      <sect3 xml:id="sockets-client-functions">
        <title>Client Functions</title>

        <para>Typically, the client initiates the connection to the
	  server. The client knows which server it is about to call:
	  It knows its <acronym>IP</acronym> address, and it knows the
	  <emphasis>port</emphasis> the server resides at. It is akin
	  to you picking up the phone and dialing the number (the
	  <emphasis>address</emphasis>), then, after someone answers,
	  asking for the person in charge of wingdings (the
	  <emphasis>port</emphasis>).</para>

        <sect4 xml:id="sockets-connect">
          <title><function>connect</function></title>

          <para>Once a client has created a socket, it needs to
	    connect it to a specific port on a remote system. It uses
	    &man.connect.2;:</para>

<programlisting>
int connect(int s, const struct sockaddr *name, socklen_t namelen);
</programlisting>

          <para>The <varname>s</varname> argument is the socket, i.e.,
	    the value returned by the <function>socket</function>
	    function. The <varname>name</varname> is a pointer to
	    <varname>sockaddr</varname>, the structure we have talked
	    about extensively. Finally, <varname>namelen</varname>
	    informs the system how many bytes are in our
	    <varname>sockaddr</varname> structure.</para>

          <para>If <function>connect</function> is successful, it
	    returns <constant>0</constant>. Otherwise it returns
	    <constant>-1</constant> and stores the error code in
	    <varname>errno</varname>.</para>

          <para>There are many reasons why
	    <function>connect</function> may fail.  For example, with
	    an attempt to an Internet connection, the
	    <acronym>IP</acronym> address may not exist, or it may be
	    down, or just too busy, or it may not have a server
	    listening at the specified port. Or it may outright
	    <emphasis>refuse</emphasis> any request for specific
	    code.</para>

	</sect4>

        <sect4 xml:id="sockets-first-client">
          <title>Our First Client</title>

	  <para>We now know enough to write a very simple client, one
	    that will get current time from <systemitem class="ipaddress">192.43.244.18</systemitem> and print it to
	    <filename>stdout</filename>.</para>

<programlisting>
/*
 * daytime.c
 *
 * Programmed by G. Adam Stanislav
 */
#include &lt;stdio.h&gt;
#include &lt;sys/types.h&gt;
#include &lt;sys/socket.h&gt;
#include &lt;netinet/in.h&gt;

int main() {
  register int s;
  register int bytes;
  struct sockaddr_in sa;
  char buffer[BUFSIZ+1];

  if ((s = socket(PF_INET, SOCK_STREAM, 0)) &lt; 0) {
    perror("socket");
    return 1;
  }

  bzero(&amp;sa, sizeof sa);

  sa.sin_family = AF_INET;
  sa.sin_port = htons(13);
  sa.sin_addr.s_addr = htonl((((((192 &lt;&lt; 8) | 43) &lt;&lt; 8) | 244) &lt;&lt; 8) | 18);
  if (connect(s, (struct sockaddr *)&amp;sa, sizeof sa) &lt; 0) {
    perror("connect");
    close(s);
    return 2;
  }

  while ((bytes = read(s, buffer, BUFSIZ)) &gt; 0)
    write(1, buffer, bytes);

  close(s);
  return 0;
}
</programlisting>

          <para>Go ahead, enter it in your editor, save it as
	    <filename>daytime.c</filename>, then compile and run
	    it:</para>

<screen>&prompt.user; <userinput>cc -O3 -o daytime daytime.c</userinput>
&prompt.user; <userinput>./daytime</userinput>

52079 01-06-19 02:29:25 50 0 1 543.9 UTC(NIST) *
&prompt.user;</screen>

          <para>In this case, the date was June 19, 2001, the time was
	    02:29:25 <acronym>UTC</acronym>. Naturally, your results
	    will vary.</para>

	</sect4>

      </sect3>

      <sect3 xml:id="sockets-server-functions">
        <title>Server Functions</title>

        <para>The typical server does not initiate the
	  connection. Instead, it waits for a client to call it and
	  request services. It does not know when the client will
	  call, nor how many clients will call. It may be just sitting
	  there, waiting patiently, one moment, The next moment, it
	  can find itself swamped with requests from a number of
	  clients, all calling in at the same time.</para>

	<para>The sockets interface offers three basic functions to
	  handle this.</para>

        <sect4 xml:id="sockets-bind">
          <title><function>bind</function></title>

          <para>Ports are like extensions to a phone line: After you
	    dial a number, you dial the extension to get to a specific
	    person or department.</para>

	  <para>There are 65535 <acronym>IP</acronym> ports, but a
	    server usually processes requests that come in on only one
	    of them. It is like telling the phone room operator that
	    we are now at work and available to answer the phone at a
	    specific extension. We use &man.bind.2; to tell sockets
	    which port we want to serve.</para>

<programlisting>
int bind(int s, const struct sockaddr *addr, socklen_t addrlen);
</programlisting>

          <para>Beside specifying the port in <varname>addr</varname>,
	    the server may include its <acronym>IP</acronym>
	    address. However, it can just use the symbolic constant
	    <symbol>INADDR_ANY</symbol> to indicate it will serve all
	    requests to the specified port regardless of what its
	    <acronym>IP</acronym> address is. This symbol, along with
	    several similar ones, is declared in
	    <filename>netinet/in.h</filename></para>

<programlisting>
#define	INADDR_ANY		(u_int32_t)0x00000000
</programlisting>

          <para>Suppose we were writing a server for the
	    <emphasis>daytime</emphasis> protocol over
	    <acronym>TCP</acronym>/<acronym>IP</acronym>. Recall that
	    it uses port 13. Our <varname>sockaddr_in</varname>
	    structure would look like this:</para>

          <mediaobject>
            <imageobject>
              <imagedata fileref="sockets/sainserv"/>
            </imageobject>

            <textobject>
              <literallayout class="monospaced">        0        1        2       3
   +--------+--------+--------+--------+
 0 |    0   |   2    |    0   |   13   |
   +--------+--------+--------+--------+
 4 |                 0                 |
   +-----------------------------------+
 8 |                 0                 |
   +-----------------------------------+
12 |                 0                 |
   +-----------------------------------+</literallayout>
            </textobject>

            <textobject>
              <phrase>Example Server sockaddr_in</phrase>
            </textobject>
          </mediaobject>
        </sect4>

        <sect4 xml:id="sockets-listen">
          <title><function>listen</function></title>

          <para>To continue our office phone analogy, after you have
	    told the phone central operator what extension you will be
	    at, you now walk into your office, and make sure your own
	    phone is plugged in and the ringer is turned on. Plus, you
	    make sure your call waiting is activated, so you can hear
	    the phone ring even while you are talking to someone.</para>

          <para>The server ensures all of that with the &man.listen.2;
            function.</para>

<programlisting>
int listen(int s, int backlog);
</programlisting>

          <para>In here, the <varname>backlog</varname> variable tells
	    sockets how many incoming requests to accept while you are
	    busy processing the last request. In other words, it
	    determines the maximum size of the queue of pending
	    connections.</para>

        </sect4>

        <sect4 xml:id="sockets-accept">
          <title><function>accept</function></title>

          <para>After you hear the phone ringing, you accept the call
	    by answering the call. You have now established a
	    connection with your client. This connection remains
	    active until either you or your client hang up.</para>

	  <para>The server accepts the connection by using the
  	    &man.accept.2; function.</para>

<programlisting>
int accept(int s, struct sockaddr *addr, socklen_t *addrlen);
</programlisting>

          <para>Note that this time <varname>addrlen</varname> is a
            pointer. This is necessary because in this case it is the
            socket that fills out <varname>addr</varname>, the
            <varname>sockaddr_in</varname> structure.</para>

	  <para>The return value is an integer. Indeed, the
	    <function>accept</function> returns a <emphasis>new
	    socket</emphasis>. You will use this new socket to
	    communicate with the client.</para>

          <para>What happens to the old socket? It continues to listen
	    for more requests (remember the <varname>backlog</varname>
	    variable we passed to <function>listen</function>?) until
	    we <function>close</function> it.</para>

	  <para>Now, the new socket is meant only for
	    communications. It is fully connected. We cannot pass it
	    to <function>listen</function> again, trying to accept
	    additional connections.</para>

        </sect4>

        <sect4 xml:id="sockets-first-server">
          <title>Our First Server</title>

	  <para>Our first server will be somewhat more complex than
	    our first client was: Not only do we have more sockets
	    functions to use, but we need to write it as a
	    daemon.</para>

	  <para>This is best achieved by creating a <emphasis>child
	    process</emphasis> after binding the port. The main
	    process then exits and returns control to the
	    <application>shell</application> (or whatever program
	    invoked it).</para>

	  <para>The child calls <function>listen</function>, then
	    starts an endless loop, which accepts a connection, serves
	    it, and eventually closes its socket.</para>

<programlisting>
/*
 * daytimed - a port 13 server
 *
 * Programmed by G. Adam Stanislav
 * June 19, 2001
 */
#include &lt;stdio.h&gt;
#include &lt;time.h&gt;
#include &lt;unistd.h&gt;
#include &lt;sys/types.h&gt;
#include &lt;sys/socket.h&gt;
#include &lt;netinet/in.h&gt;

#define BACKLOG 4

int main() {
    register int s, c;
    int b;
    struct sockaddr_in sa;
    time_t t;
    struct tm *tm;
    FILE *client;

    if ((s = socket(PF_INET, SOCK_STREAM, 0)) &lt; 0) {
        perror("socket");
        return 1;
    }

    bzero(&amp;sa, sizeof sa);

    sa.sin_family = AF_INET;
    sa.sin_port   = htons(13);

    if (INADDR_ANY)
        sa.sin_addr.s_addr = htonl(INADDR_ANY);

    if (bind(s, (struct sockaddr *)&amp;sa, sizeof sa) &lt; 0) {
        perror("bind");
        return 2;
    }

    switch (fork()) {
        case -1:
            perror("fork");
            return 3;
            break;
        default:
            close(s);
            return 0;
            break;
        case 0:
            break;
    }

    listen(s, BACKLOG);

    for (;;) {
        b = sizeof sa;

        if ((c = accept(s, (struct sockaddr *)&amp;sa, &amp;b)) &lt; 0) {
            perror("daytimed accept");
            return 4;
        }

        if ((client = fdopen(c, "w")) == NULL) {
            perror("daytimed fdopen");
            return 5;
        }

        if ((t = time(NULL)) &lt; 0) {
            perror("daytimed time");

            return 6;
        }

        tm = gmtime(&amp;t);
        fprintf(client, "%.4i-%.2i-%.2iT%.2i:%.2i:%.2iZ\n",
            tm-&gt;tm_year + 1900,
            tm-&gt;tm_mon + 1,
            tm-&gt;tm_mday,
            tm-&gt;tm_hour,
            tm-&gt;tm_min,
            tm-&gt;tm_sec);

        fclose(client);
    }
}
</programlisting>

          <para>We start by creating a socket. Then we fill out the
	    <varname>sockaddr_in</varname> structure in
	    <varname>sa</varname>.  Note the conditional use of
	    <symbol>INADDR_ANY</symbol>:</para>

<programlisting>
    if (INADDR_ANY)
        sa.sin_addr.s_addr = htonl(INADDR_ANY);
</programlisting>

          <para>Its value is <constant>0</constant>. Since we have
	    just used <function>bzero</function> on the entire
	    structure, it would be redundant to set it to
	    <constant>0</constant> again. But if we port our code to
	    some other system where <symbol>INADDR_ANY</symbol> is
	    perhaps not a zero, we need to assign it to
	    <varname>sa.sin_addr.s_addr</varname>. Most modern C
	    compilers are clever enough to notice that
	    <symbol>INADDR_ANY</symbol> is a constant. As long as it
	    is a zero, they will optimize the entire conditional
	    statement out of the code.</para>

	  <para>After we have called <function>bind</function>
	    successfully, we are ready to become a
	    <emphasis>daemon</emphasis>: We use
	    <function>fork</function> to create a child process. In
	    both, the parent and the child, the <varname>s</varname>
	    variable is our socket. The parent process will not need
	    it, so it calls <function>close</function>, then it
	    returns <constant>0</constant> to inform its own parent it
	    had terminated successfully.</para>

	  <para>Meanwhile, the child process continues working in the
	    background.  It calls <function>listen</function> and sets
	    its backlog to <constant>4</constant>. It does not need a
	    large value here because <emphasis>daytime</emphasis> is
	    not a protocol many clients request all the time, and
	    because it can process each request instantly anyway.</para>

	  <para>Finally, the daemon starts an endless loop, which
	    performs the following steps:</para>

	  <procedure>
	    <step><para> Call <function>accept</function>. It waits
	      here until a client contacts it. At that point, it
	      receives a new socket, <varname>c</varname>, which it
	      can use to communicate with this particular client.
	      </para></step>

            <step><para>It uses the C function
	      <function>fdopen</function> to turn the socket from a
	      low-level <emphasis>file descriptor</emphasis> to a
	      C-style <varname>FILE</varname> pointer. This will allow
	      the use of <function>fprintf</function> later on.
	      </para></step>

	    <step><para>It checks the time, and prints it in the
	      <emphasis><acronym>ISO</acronym> 8601</emphasis> format
	      to the <varname>client</varname> <quote>file</quote>. It
	      then uses <function>fclose</function> to close the
	      file. That will automatically close the socket as well.
	      </para></step>

	  </procedure>

	  <para>We can <emphasis>generalize</emphasis> this, and use
	    it as a model for many other servers:</para>

          <mediaobject>
            <imageobject>
              <imagedata fileref="sockets/serv"/>
            </imageobject>

            <textobject>
              <literallayout class="monospaced">+-----------------+
|  Create Socket  |
+-----------------+
          |
+-----------------+
|    Bind Port    |       Daemon Process
+-----------------+
          |                 +--------+
          +-------------+--&gt;|  Init  |
          |             |   +--------+
+-----------------+     |         |
|        Exit     |     |   +--------+
+-----------------+     |   | Listen |
                        |   +--------+
                        |         |
                        |   +--------+
                        |   | Accept |
                        |   +--------+
                        |         |
                        |   +--------+
                        |   | Serve  |
                        |   +--------+
                        |         |
                        |   +--------+
                        |   | Close  |
                        |&lt;--------+</literallayout>
            </textobject>

            <textobject>
              <phrase>Sequential Server</phrase>
            </textobject>
          </mediaobject>

          <para>This flowchart is good for <emphasis>sequential
	    servers</emphasis>, i.e., servers that can serve one
	    client at a time, just as we were able to with our
	    <emphasis>daytime</emphasis> server. This is only possible
	    whenever there is no real <quote>conversation</quote>
	    going on between the client and the server: As soon as the
	    server detects a connection to the client, it sends out
	    some data and closes the connection. The entire operation
	    may take nanoseconds, and it is finished.</para>

	  <para>The advantage of this flowchart is that, except for
	    the brief moment after the parent
	    <function>fork</function>s and before it exits, there is
	    always only one <emphasis>process</emphasis> active: Our
	    server does not take up much memory and other system
	    resources.</para>

	  <para>Note that we have added <emphasis>initialize
	    daemon</emphasis> in our flowchart. We did not need to
	    initialize our own daemon, but this is a good place in the
	    flow of the program to set up any
	    <function>signal</function> handlers, open any files we
	    may need, etc.</para>

	  <para>Just about everything in the flow chart can be used
	    literally on many different servers. The
	    <emphasis>serve</emphasis> entry is the exception.  We
	    think of it as a <emphasis><quote>black
	    box</quote></emphasis>, i.e., something you design
	    specifically for your own server, and just <quote>plug it
	    into the rest.</quote></para>

	  <para>Not all protocols are that simple. Many receive a
	    request from the client, reply to it, then receive another
	    request from the same client. Because of that, they do not
	    know in advance how long they will be serving the
	    client. Such servers usually start a new process for each
	    client. While the new process is serving its client, the
	    daemon can continue listening for more connections.</para>

	  <para>Now, go ahead, save the above source code as
	    <filename>daytimed.c</filename> (it is customary to end
	    the names of daemons with the letter
	    <constant>d</constant>). After you have compiled it, try
	    running it:</para>

<screen>&prompt.user; <userinput>./daytimed</userinput>
bind: Permission denied
&prompt.user;</screen>

          <para>What happened here? As you will recall, the
	    <emphasis>daytime</emphasis> protocol uses port 13. But
	    all ports below 1024 are reserved to the superuser
	    (otherwise, anyone could start a daemon pretending to
	    serve a commonly used port, while causing a security
	    breach).</para>

	  <para>Try again, this time as the superuser:</para>

<screen>&prompt.root; <userinput>./daytimed</userinput>
&prompt.root;</screen>

          <para>What... Nothing? Let us try again:</para>

<screen>&prompt.root; <userinput>./daytimed</userinput>

bind: Address already in use
&prompt.root;</screen>

          <para>Every port can only be bound by one program at a
	    time. Our first attempt was indeed successful: It started
	    the child daemon and returned quietly. It is still running
	    and will continue to run until you either kill it, or any
	    of its system calls fail, or you reboot the system.</para>

	  <para>Fine, we know it is running in the background. But is
	    it working?  How do we know it is a proper
	    <emphasis>daytime</emphasis> server?  Simple:</para>

<screen>&prompt.user; <userinput>telnet localhost 13</userinput>

Trying ::1...
telnet: connect to address ::1: Connection refused
Trying 127.0.0.1...
Connected to localhost.
Escape character is '^]'.
2001-06-19T21:04:42Z
Connection closed by foreign host.
&prompt.user;</screen>

          <para><application>telnet</application> tried the new
	    <acronym>IP</acronym>v6, and failed. It retried with
	    <acronym>IP</acronym>v4 and succeeded.  The daemon
	    works.</para>

          <para>If you have access to another &unix; system via
	    <application>telnet</application>, you can use it to test
	    accessing the server remotely. My computer does not have a
	    static <acronym>IP</acronym> address, so this is what I
	    did:</para>

<screen>&prompt.user; <userinput>who</userinput>

whizkid          ttyp0   Jun 19 16:59   (216.127.220.143)
xxx              ttyp1   Jun 19 16:06   (xx.xx.xx.xx)
&prompt.user; <userinput>telnet 216.127.220.143 13</userinput>

Trying 216.127.220.143...
Connected to r47.bfm.org.
Escape character is '^]'.
2001-06-19T21:31:11Z
Connection closed by foreign host.
&prompt.user;</screen>

          <para>Again, it worked. Will it work using the domain name?
	    </para>

<screen>&prompt.user; <userinput>telnet r47.bfm.org 13</userinput>

Trying 216.127.220.143...
Connected to r47.bfm.org.
Escape character is '^]'.
2001-06-19T21:31:40Z
Connection closed by foreign host.
&prompt.user;</screen>

          <para>By the way, <application>telnet</application> prints
	    the <emphasis>Connection closed by foreign host</emphasis>
	    message after our daemon has closed the socket. This shows
	    us that, indeed, using
	    <function>fclose(client);</function> in our code works as
	    advertised.</para>

        </sect4>

      </sect3>

    </sect2>

  </sect1>

  <sect1 xml:id="sockets-helper-functions">
    <title>Helper Functions</title>

    <para>FreeBSD C library contains many helper functions for sockets
      programming.  For example, in our sample client we hard coded
      the <systemitem class="fqdomainname">time.nist.gov</systemitem>
      <acronym>IP</acronym> address. But we do not always know the
      <acronym>IP</acronym> address.  Even if we do, our software is
      more flexible if it allows the user to enter the
      <acronym>IP</acronym> address, or even the domain name.
      </para>

    <sect2 xml:id="sockets-gethostbyname">
      <title><function>gethostbyname</function></title>

      <para>While there is no way to pass the domain name directly to
        any of the sockets functions, the FreeBSD C library comes with
        the &man.gethostbyname.3; and &man.gethostbyname2.3; functions,
        declared in <filename>netdb.h</filename>.</para>

<programlisting>
struct hostent * gethostbyname(const char *name);
struct hostent * gethostbyname2(const char *name, int af);
</programlisting>

      <para>Both return a pointer to the <varname>hostent</varname>
        structure, with much information about the domain. For our
        purposes, the <varname>h_addr_list[0]</varname> field of the
        structure points at <varname>h_length</varname> bytes of the
        correct address, already stored in the <emphasis>network byte
        order</emphasis>.</para>

      <para>This allows us to create a much more flexible&mdash;and
        much more useful&mdash;version of our
        <application>daytime</application> program:</para>

<programlisting>
/*
 * daytime.c
 *
 * Programmed by G. Adam Stanislav
 * 19 June 2001
 */
#include &lt;stdio.h&gt;
#include &lt;string.h&gt;
#include &lt;sys/types.h&gt;
#include &lt;sys/socket.h&gt;
#include &lt;netinet/in.h&gt;
#include &lt;netdb.h&gt;

int main(int argc, char *argv[]) {
  register int s;
  register int bytes;
  struct sockaddr_in sa;
  struct hostent *he;
  char buf[BUFSIZ+1];
  char *host;

  if ((s = socket(PF_INET, SOCK_STREAM, 0)) &lt; 0) {
    perror("socket");
    return 1;
  }

  bzero(&amp;sa, sizeof sa);

  sa.sin_family = AF_INET;
  sa.sin_port = htons(13);

  host = (argc &gt; 1) ? (char *)argv[1] : "time.nist.gov";

  if ((he = gethostbyname(host)) == NULL) {
    perror(host);
    return 2;
  }

  bcopy(he-&gt;h_addr_list[0],&amp;sa.sin_addr, he-&gt;h_length);

  if (connect(s, (struct sockaddr *)&amp;sa, sizeof sa) &lt; 0) {
    perror("connect");
    return 3;
  }

  while ((bytes = read(s, buf, BUFSIZ)) &gt; 0)
    write(1, buf, bytes);

  close(s);
  return 0;
}
</programlisting>

      <para>We now can type a domain name (or an <acronym>IP</acronym>
        address, it works both ways) on the command line, and the
        program will try to connect to its
        <emphasis>daytime</emphasis> server. Otherwise, it will still
        default to <systemitem class="fqdomainname">time.nist.gov</systemitem>. However, even in
        this case we will use <function>gethostbyname</function>
        rather than hard coding <systemitem class="ipaddress">192.43.244.18</systemitem>. That way, even if its
        <acronym>IP</acronym> address changes in the future, we will
        still find it.</para>

      <para>Since it takes virtually no time to get the time from your
        local server, you could run <application>daytime</application>
        twice in a row: First to get the time from <systemitem class="fqdomainname">time.nist.gov</systemitem>, the second time from
        your own system. You can then compare the results and see how
        exact your system clock is:</para>

<screen>&prompt.user; <userinput>daytime ; daytime localhost</userinput>


52080 01-06-20 04:02:33 50 0 0 390.2 UTC(NIST) *
2001-06-20T04:02:35Z
&prompt.user;</screen>

      <para>As you can see, my system was two seconds ahead of the
        <acronym>NIST</acronym> time.</para>

    </sect2>

    <sect2 xml:id="sockets-getservbyname">
      <title><function>getservbyname</function></title>

      <para>Sometimes you may not be sure what port a certain service
        uses.  The &man.getservbyname.3; function, also declared in
        <filename>netdb.h</filename> comes in very handy in those
        cases:</para>

<programlisting>
struct servent * getservbyname(const char *name, const char *proto);
</programlisting>

      <para>The <varname>servent</varname> structure contains the
        <varname>s_port</varname>, which contains the proper port,
        already in <emphasis>network byte order</emphasis>.</para>

      <para>Had we not known the correct port for the
        <emphasis>daytime</emphasis> service, we could have found it
        this way:</para>

<programlisting>
  struct servent *se;
  ...
  if ((se = getservbyname("daytime", "tcp")) == NULL {
    fprintf(stderr, "Cannot determine which port to use.\n");
    return 7;
  }
  sa.sin_port = se-&gt;s_port;
</programlisting>

      <para>You usually do know the port. But if you are developing a
        new protocol, you may be testing it on an unofficial
        port. Some day, you will register the protocol and its port
        (if nowhere else, at least in your
        <filename>/etc/services</filename>, which is where
        <function>getservbyname</function> looks). Instead of
        returning an error in the above code, you just use the
        temporary port number.  Once you have listed the protocol in
        <filename>/etc/services</filename>, your software will find
        its port without you having to rewrite the code.</para>

    </sect2>

  </sect1>

  <sect1 xml:id="sockets-concurrent-servers">
    <title>Concurrent Servers</title>

    <para>Unlike a sequential server, a <emphasis>concurrent
      server</emphasis> has to be able to serve more than one client
      at a time. For example, a <emphasis>chat server</emphasis> may
      be serving a specific client for hours&mdash;it cannot wait till
      it stops serving a client before it serves the next one.</para>

    <para>This requires a significant change in our flowchart:</para>

    <mediaobject>
      <imageobject>
        <imagedata fileref="sockets/serv2"/>
      </imageobject>

      <textobject>
        <literallayout class="monospaced">+-----------------+
|  Create Socket  |
+-----------------+
          |
+-----------------+
|    Bind Port    |       Daemon Process
+-----------------+
          |                 +--------+
          +-------------+--&gt;|  Init  |
          |             |   +--------+
+-----------------+     |         |
|        Exit     |     |   +--------+
+-----------------+     |   | Listen |
                        |   +--------+
                        |         |
                        |   +--------+
                        |   | Accept |
                        |   +--------+
                        |         |       +------------------+
                        |         +------&gt;| Close Top Socket |
                        |         |       +------------------+
                        |   +--------+             |
                        |   | Close  |    +------------------+
                        |   +--------+    |     Serve        |
                        |         |       +------------------+
                        |&lt;--------+                |
                                          +------------------+
                                          | Close Acc Socket |
                            +--------+    +------------------+
                            | Signal |             |
                            +--------+    +------------------+
                                          |      Exit        |
                                          +------------------+</literallayout>
      </textobject>

      <textobject>
        <phrase>Concurrent Server</phrase>
      </textobject>
    </mediaobject>

    <para>We moved the <emphasis>serve</emphasis> from the
      <emphasis>daemon process</emphasis> to its own <emphasis>server
      process</emphasis>. However, because each child process inherits
      all open files (and a socket is treated just like a file), the
      new process inherits not only the <emphasis><quote>accepted
      handle,</quote></emphasis> i.e., the socket returned by the
      <function>accept</function> call, but also the <emphasis>top
      socket</emphasis>, i.e., the one opened by the top process right
      at the beginning.</para>

    <para>However, the <emphasis>server process</emphasis> does not
      need this socket and should <function>close</function> it
      immediately. Similarly, the <emphasis>daemon process</emphasis>
      no longer needs the <emphasis>accepted socket</emphasis>, and
      not only should, but <emphasis>must</emphasis>
      <function>close</function> it&mdash;otherwise, it will run out
      of available <emphasis>file descriptors</emphasis> sooner or
      later.</para>

    <para>After the <emphasis>server process</emphasis> is done
      serving, it should close the <emphasis>accepted
      socket</emphasis>. Instead of returning to
      <function>accept</function>, it now exits.
      </para>

    <para>Under &unix;, a process does not really
      <emphasis>exit</emphasis>. Instead, it
      <emphasis>returns</emphasis> to its parent. Typically, a parent
      process <function>wait</function>s for its child process, and
      obtains a return value. However, our <emphasis>daemon
      process</emphasis> cannot simply stop and wait. That would
      defeat the whole purpose of creating additional processes. But
      if it never does <function>wait</function>, its children will
      become <emphasis>zombies</emphasis>&mdash;no longer functional
      but still roaming around.</para>

    <para>For that reason, the <emphasis>daemon process</emphasis>
      needs to set <emphasis>signal handlers</emphasis> in its
      <emphasis>initialize daemon</emphasis> phase. At least a
      <symbol>SIGCHLD</symbol> signal has to be processed, so the
      daemon can remove the zombie return values from the system and
      release the system resources they are taking up.</para>

    <para>That is why our flowchart now contains a <emphasis>process
      signals</emphasis> box, which is not connected to any other box.
      By the way, many servers also process <symbol>SIGHUP</symbol>,
      and typically interpret as the signal from the superuser that
      they should reread their configuration files. This allows us to
      change settings without having to kill and restart these
      servers.</para>

  </sect1>

</chapter>