Subversion Repositories HelenOS-doc

Rev

Rev 129 | Rev 160 | Go to most recent revision | Blame | Compare with Previous | Last modification | View Log | Download | RSS feed

  1. <?xml version="1.0" encoding="UTF-8"?>
  2. <chapter id="ipc">
  3.   <?dbhtml filename="ipc.html"?>
  4.  
  5.   <title>IPC</title>
  6.  
  7.   <para>Due to the high intertask communication traffic, IPC becomes critical
  8.   subsystem for microkernels, putting high demands on the speed, latency and
  9.   reliability of IPC model and implementation. Although theoretically the use
  10.   of asynchronous messaging system looks promising, it is not often
  11.   implemented because of a problematic implementation of end user
  12.   applications. HelenOS implements fully asynchronous messaging system with a
  13.   special layer providing a user application developer a reasonably
  14.   synchronous multithreaded environment sufficient to develop complex
  15.   protocols.</para>
  16.  
  17.   <section>
  18.     <title>Kernel Services</title>
  19.  
  20.     <para>Every message consists of 4 numeric arguments (32-bit and 64-bit on
  21.     the corresponding platforms), from which the first one is considered a
  22.     method number on message receipt and a return value on answer receipt. The
  23.     received message contains identification of the incoming connection, so
  24.     that the receiving application can distinguish the messages between
  25.     different senders. Internally the message contains pointer to the
  26.     originating task and to the source of the communication channel. If the
  27.     message is forwarded, the originating task identifies the recipient of the
  28.     answer, the source channel identifies the connection in case of a hangup
  29.     response.</para>
  30.  
  31.     <para>Every message must be eventually answered. The system keeps track of
  32.     all messages, so that it can answer them with appropriate error code
  33.     should one of the connection parties fail unexpectedly. To limit buffering
  34.     of the messages in the kernel, every process is has a limited account of
  35.     asynchronous messages it can send simultanously. If the limit is reached,
  36.     the kernel refuses to send any other message, until some active message is
  37.     answered.</para>
  38.  
  39.     <para>To facilitate kernel-to-user communication, the IPC subsystem
  40.     provides notification messages. The applications can subscribe to a
  41.     notification channel and receive messages directed to this channel. Such
  42.     messages can be freely sent even from interrupt context as they are
  43.     primarily destined to deliver IRQ events to userspace device drivers.
  44.     These messages need not be answered, there is no party that could receive
  45.     such response.</para>
  46.  
  47.     <section>
  48.       <title>Low Level IPC</title>
  49.  
  50.       <para>The whole IPC subsystem consists of one-way communication
  51.       channels. Each task has one associated message queue (answerbox). The
  52.       task can call other tasks and connect it's phones to their answerboxes.,
  53.      send and forward messages through these connections and answer received
  54.      messages. Every sent message is identified by a unique number, so that
  55.      the response can be later matched against it. The message is sent over
  56.      the phone to the target answerbox. Server application periodically
  57.      checks the answerbox and pulls messages from several queues associated
  58.      with it. After completing the requested action, server sends a reply
  59.      back to the answerbox of the originating task. If a need arises, it is
  60.      possible to <emphasis>forward</emphasis> a recevied message throught any
  61.      of the open phones to another task. This mechanism is used e.g. for
  62.      opening new connections.</para>
  63.  
  64.      <para>The answerbox contains four different message queues:</para>
  65.  
  66.      <itemizedlist>
  67.        <listitem>
  68.          <para>Incoming call queue</para>
  69.        </listitem>
  70.  
  71.        <listitem>
  72.          <para>Dispatched call queue</para>
  73.        </listitem>
  74.  
  75.        <listitem>
  76.          <para>Answer queue</para>
  77.        </listitem>
  78.  
  79.        <listitem>
  80.          <para>Notification queue</para>
  81.        </listitem>
  82.      </itemizedlist>
  83.  
  84.      <figure float="1">
  85.        <title>Low level IPC</title>
  86.  
  87.        <mediaobject id="ipc1">
  88.          <imageobject role="pdf">
  89.            <imagedata fileref="images/ipc1.pdf" format="PDF" />
  90.          </imageobject>
  91.  
  92.          <imageobject role="html">
  93.            <imagedata fileref="images/ipc1.png" format="PNG" />
  94.          </imageobject>
  95.  
  96.          <imageobject role="fop">
  97.            <imagedata fileref="images/ipc1.svg" format="SVG" />
  98.          </imageobject>
  99.        </mediaobject>
  100.      </figure>
  101.  
  102.      <para>The communication between task A, that is connected to task B
  103.      looks as follows: Task A sends a message over it's phone to the target
  104.       asnwerbox. The message is saved in task B incoming call queue. When task
  105.       B fetches the message for processing, it is automatically moved into the
  106.       dispatched call queue. After the server decides to answer the message,
  107.       it is removed from dispatched queue and the result is moved into the
  108.       answer queue of task A.</para>
  109.  
  110.       <para>The arguments contained in the message are completely arbitrary
  111.       and decided by the user. The low level part of kernel IPC fills in
  112.       appropriate error codes if there is an error during communication. It is
  113.       assured that the applications are correctly notified about communication
  114.       state. If a program closes the outgoing connection, the target answerbox
  115.       receives a hangup message. The connection identification is not reused,
  116.       until the hangup message is acknowledged and all other pending messages
  117.       are answered.</para>
  118.  
  119.       <para>Closing an incoming connection is done by responding to any
  120.       incoming message with an EHANGUP error code. The connection is then
  121.       immediately closed. The client connection identification (phone id) is
  122.       not reused, until the client issues closes it's own side of the
  123.      connection ("hangs his phone up").</para>
  124.  
  125.      <para>When a task dies (whether voluntarily or by being killed), cleanup
  126.      process is started.</para>
  127.  
  128.      <orderedlist>
  129.        <listitem>
  130.          <para>Hangs up all outgoing connections and sends hangup messages to
  131.          all target answerboxes.</para>
  132.        </listitem>
  133.  
  134.        <listitem>
  135.          <para>Disconnects all incoming connections.</para>
  136.        </listitem>
  137.  
  138.        <listitem>
  139.          <para>Disconnects from all notification channels.</para>
  140.        </listitem>
  141.  
  142.        <listitem>
  143.          <para>Answers all unanswered messages from answerbox queues with
  144.          appropriate error code.</para>
  145.        </listitem>
  146.  
  147.        <listitem>
  148.          <para>Waits until all outgoing messages are answered and all
  149.          remaining answerbox queues are empty.</para>
  150.        </listitem>
  151.      </orderedlist>
  152.    </section>
  153.  
  154.    <section>
  155.      <title>System Call IPC Layer</title>
  156.  
  157.      <para>On top of this simple protocol the kernel provides special
  158.      services closely related to the inter-process communication. A range of
  159.      method numbers is allocated and protocol is defined for these functions.
  160.      The messages are interpreted by the kernel layer and appropriate actions
  161.      are taken depending on the parameters of message and answer.</para>
  162.  
  163.      <para>The kernel provides the following services:</para>
  164.  
  165.      <itemizedlist>
  166.        <listitem>
  167.          <para>Creating new outgoing connection</para>
  168.        </listitem>
  169.  
  170.        <listitem>
  171.          <para>Creating a callback connection</para>
  172.        </listitem>
  173.  
  174.        <listitem>
  175.          <para>Sending an address space area</para>
  176.        </listitem>
  177.  
  178.        <listitem>
  179.          <para>Asking for an address space area</para>
  180.        </listitem>
  181.      </itemizedlist>
  182.  
  183.      <para>On startup every task is automatically connected to a
  184.      <emphasis>name service task</emphasis>, which provides a switchboard
  185.      functionality. To open a new outgoing connection, the client sends a
  186.      <constant>CONNECT_ME_TO</constant> message using any of his phones. If
  187.      the recepient of this message answers with an accepting answer, a new
  188.      connection is created. In itself, this mechanism would allow only
  189.      duplicating existing connection. However, if the message is forwarded,
  190.      the new connection is made to the final recipient.</para>
  191.  
  192.      <para>On startup every task is automatically connect to the name service
  193.      task, which acts as a switchboard and forwards requests for connection
  194.      to specific services. To be able to forward a message it must have a
  195.      phone connected to the service tasks. The task creates this connection
  196.      using a <constant>CONNECT_TO_ME</constant> message which creates a
  197.      callback connection. Every service that wants to receive connections
  198.      asks name service task to create a callback connection.</para>
  199.  
  200.      <para>Tasks can share their address space areas using IPC messages. The
  201.      2 message types - <constant>AS_AREA_SEND</constant> and <constant>AS_AREA_RECV</constant> are used for sending and
  202.      receiving an address area respectively. The shared area can be accessed
  203.      as soon as the message is acknowledged.</para>
  204.    </section>
  205.  </section>
  206.  
  207.  <section>
  208.    <title>Userspace View</title>
  209.  
  210.    <para>The conventional design of the asynchronous api seems to produce
  211.    applications with one event loop and several big switch statements.
  212.    However, by intensive utilization of user-space threads, it was possible
  213.    to create an environment that is not necesarilly restricted to this type
  214.    of event-driven programming and allows for more fluent expression of
  215.    application programs.</para>
  216.  
  217.    <section>
  218.      <title>Single Point of Entry</title>
  219.  
  220.      <para>Each tasks is associated with only one answerbox. If a
  221.      multi-threaded application needs to communicate, it must be not only
  222.      able to send a message, but it should be able to retrieve the answer as
  223.      well. If several threads pull messages from task answerbox, it is a
  224.      matter of fortune, which thread receives which message. If a particular
  225.      thread needs to wait for a message answer, an idle
  226.      <emphasis>manager</emphasis> task is found or a new one is created and
  227.      control is transfered to this manager task. The manager tasks pops
  228.      messages from the answerbox and puts them into appropriate queues of
  229.      running tasks. If a task waiting for a message is not running, the
  230.      control is transferred to it.</para>
  231.  
  232.      <figure float="1">
  233.        <title>Single point of entry</title>
  234.        <mediaobject id="ipc2">
  235.          <imageobject role="pdf">
  236.            <imagedata fileref="images/ipc2.pdf" format="PDF" />
  237.          </imageobject>
  238.  
  239.          <imageobject role="html">
  240.            <imagedata fileref="images/ipc2.png" format="PNG" />
  241.          </imageobject>
  242.  
  243.          <imageobject role="fop">
  244.            <imagedata fileref="images/ipc2.svg" format="SVG" />
  245.          </imageobject>
  246.        </mediaobject>
  247.  
  248.      </figure>
  249.  
  250.      <para>Very similar situation arises when a task decides to send a lot of
  251.      messages and reaches kernel limit of asynchronous messages. In such
  252.      situation 2 remedies are available - the userspace liberary can either
  253.      cache the message locally and resend the message when some answers
  254.      arrive, or it can block the thread and let it go on only after the
  255.      message is finally sent to the kernel layer. With one exception HelenOS
  256.      uses the second approach - when the kernel responds that maximum limit
  257.      of asynchronous messages was reached, control is transferred to manager
  258.      thread. The manager thread then handles incoming replies and when space
  259.      is available, sends the message to kernel and resumes application thread
  260.      execution.</para>
  261.  
  262.      <para>If a kernel notification is received, the servicing procedure is
  263.      run in the context of the manager thread. Although it wouldn't be
  264.       impossible to allow recursive calling, it could potentially lead to an
  265.       explosion of manager threads. Thus, the kernel notification procedures
  266.       are not allowed to wait for a message result, they can only answer
  267.       messages and send new ones without waiting for their results. If the
  268.       kernel limit for outgoing messages is reached, the data is automatically
  269.       cached within the application. This behaviour is enforced automatically
  270.       and the decision making is hidden from developers view.</para>
  271.  
  272.       <figure float="1">
  273.         <title>Single point of entry solution</title>
  274.         <mediaobject id="ipc3">
  275.           <imageobject role="pdf">
  276.             <imagedata fileref="images/ipc3.pdf" format="PDF" />
  277.           </imageobject>
  278.  
  279.           <imageobject role="html">
  280.             <imagedata fileref="images/ipc3.png" format="PNG" />
  281.           </imageobject>
  282.  
  283.           <imageobject role="fop">
  284.             <imagedata fileref="images/ipc3.svg" format="SVG" />
  285.           </imageobject>
  286.         </mediaobject>
  287.  
  288.       </figure>
  289.     </section>
  290.  
  291.     <section>
  292.       <title>Ordering Problem</title>
  293.  
  294.       <para>Unfortunately, the real world is is never so simple. E.g. if a
  295.       server handles incoming requests and as a part of its response sends
  296.       asynchronous messages, it can be easily prempted and other thread may
  297.       start intervening. This can happen even if the application utilizes only
  298.       1 kernel thread. Classical synchronization using semaphores is not
  299.       possible, as locking on them would block the thread completely so that
  300.       the answer couldn't be ever processed. The IPC framework allows a
  301.      developer to specify, that part of the code should not be preempted by
  302.      any other thread (except notification handlers) while still being able
  303.      to queue messages belonging to other threads and regain control when the
  304.      answer arrives.</para>
  305.  
  306.      <para>This mechanism works transparently in multithreaded environment,
  307.      where additional locking mechanism (futexes) should be used. The IPC
  308.      framework ensures that there will always be enough free kernel threads
  309.      to handle incoming answers and allow the application to run more
  310.      user-space threads inside the kernel threads without the danger of
  311.      locking all kernel threads in futexes.</para>
  312.    </section>
  313.  
  314.    <section>
  315.      <title>The Interface</title>
  316.  
  317.      <para>The interface was developed to be as simple to use as possible.
  318.      Classical applications simply send messages and occasionally wait for an
  319.      answer and check results. If the number of sent messages is higher than
  320.      kernel limit, the flow of application is stopped until some answers
  321.      arrive. On the other hand server applications are expected to work in a
  322.      multithreaded environment.</para>
  323.  
  324.      <para>The server interface requires developer to specify a
  325.      <function>connection_thread</function> function. When new connection is
  326.      detected, a new userspace thread is automatically created and control is
  327.      transferred to this function. The code then decides whether to accept
  328.      the connection and creates a normal event loop. The userspace IPC
  329.      library ensures correct switching between several userspace threads
  330.      within the kernel environment.</para>
  331.    </section>
  332.  </section>
  333. </chapter>