path: root/lib/snmp/doc/src/snmp_def_instr_functions.xml

<?xml version="1.0" encoding="latin1" ?>
<!DOCTYPE chapter SYSTEM "chapter.dtd">

<chapter>
  <header>
    <copyright>
      <year>1997</year><year>2009</year>
      <holder>Ericsson AB. All Rights Reserved.</holder>
    </copyright>
    <legalnotice>
      The contents of this file are subject to the Erlang Public License,
      Version 1.1, (the "License"); you may not use this file except in
      compliance with the License. You should have received a copy of the
      Erlang Public License along with this software. If not, it can be
      retrieved online at http://www.erlang.org/.
    
      Software distributed under the License is distributed on an "AS IS"
      basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
      the License for the specific language governing rights and limitations
      under the License.
    
    </legalnotice>

    <title>Definition of Instrumentation Functions</title>
    <prepared></prepared>
    <responsible></responsible>
    <docno></docno>
    <approved></approved>
    <checked></checked>
    <date></date>
    <rev></rev>
    <file>snmp_def_instr_functions.xml</file>
  </header>
  <p>The section <em>Definition of Instrumentation Functions</em> describes
    the user defined functions, which the agent calls at different times.
    </p>

  <section>
    <title>Variable Instrumentation</title>
    <p>For scalar variables, a function <c>f(Operation, ...)</c> must
      be defined.
      </p>
    <p>The <c>Operation</c> can be <c>new</c>, <c>delete</c>,
      <c>get</c>, <c>is_set_ok</c>, <c>set</c>, or <c>undo</c>.
      </p>
    <p>In case of an error, all instrumentation functions may return
      either an SNMPv1 or an SNMPv2 error code. If it returns an SNMPv2
      code, it is converted into an SNMPv1 code before it is sent to a
      SNMPv1 manager.  It is  recommended to use the SNMPv2 error codes
      for all instrumentation functions, as these provide more
      details. See <seealso marker="snmp_app_a">Appendix A</seealso> for a
      description of error code conversions.
      </p>

    <section>
      <title>f(new [, ExtraArgs])</title>
      <p>The function <c>f(new [, ExtraArgs])</c> is called for each
        variable in the MIB when the
        MIB is loaded into the agent. This makes it possible to perform
        necessary initialization.
        </p>
      <p>This function is optional. The return value is discarded.</p>
    </section>

    <section>
      <title>f(delete [, ExtraArgs])</title>
      <p>The function <c>f(delete [, ExtraArgs])</c> is called
        for each object in an MIB when the
        MIB is unloaded from the agent. This makes it possible to
        perform necessary clean-up.
        </p>
      <p>This function is optional. The return value is discarded.</p>
    </section>

    <section>
      <title>f(get [, ExtraArgs])</title>
      <p>The function <c>f(get [, ExtraArgs])</c> is called when a
        get-request or a get-next
        request refers to the variable.
        </p>
      <p>This function is mandatory.
        </p>

      <section>
        <title>Valid Return Values</title>
        <list type="bulleted">
          <item><c>{value, Value}</c>. The <c>Value</c> must be of
           correct type, length and within ranges, otherwise
          <c>genErr</c> is returned in the response PDU.  If the
           object is an enumerated integer, the symbolic enum value
           may be used as an atom. If the object is of type BITS, the
           return value shall be an integer or a list of bits that
           are set.
          </item>
          <item><c>{noValue, noSuchName}</c>(SNMPv1)
          </item>
          <item><c>{noValue, noSuchObject | noSuchInstance} </c>(SNMPv2)
          </item>
          <item><c>genErr</c>.  Used if an error occurred.  Note,
           this should be an internal processing error, e.g. a caused
           by a programing fault somewhere.  If the variable does not
           exist, use <c>{noValue, noSuchName}</c> or
          <c>{noValue, noSuchInstance}</c>.
          </item>
        </list>
      </section>
    </section>

    <section>
      <title>f(is_set_ok, NewValue [, ExtraArgs])</title>
      <p>The function <c>f(is_set_ok, NewValue [, ExtraArgs])</c> is
        called in phase one of the set-request
        processing so that the new value can be checked for
        inconsistencies.
        </p>
      <p><c>NewValue</c> is guaranteed to be of the correct type,
        length and within ranges, as specified in the MIB.  If the
        object is an enumerated integer or of type BITS, the integer
        value is used.
        </p>
      <p>This function is optional.
        </p>
      <p>If this function is called, it will be called again, either
        with <c>undo</c> or with <c>set</c> as first argument.
        </p>

      <section>
        <title>Valid return values</title>
        <list type="bulleted">
          <item><c>noError</c></item>
          <item><c>badValue | noSuchName | genErr</c>(SNMPv1)
          </item>
          <item><c>noAccess | noCreation | inconsistentValue | resourceUnavailable | inconsistentName | genErr</c>(SNMPv2)</item>
        </list>
      </section>
    </section>

    <section>
      <title>f(undo, NewValue [, ExtraArgs])</title>
      <p>If an error occurred, this function is called after the
        <c>is_set_ok</c> function is called. If <c>set</c> is called for
        this object, <c>undo</c> is not called.
        </p>
      <p><c>NewValue</c> is guaranteed to be of the correct type,
        length and within ranges, as specified in the MIB.  If the
        object is an enumerated integer or of type BITS, the integer
        value is used.
        </p>
      <p>This function is optional.
        </p>

      <section>
        <title>Valid return values</title>
        <list type="bulleted">
          <item><c>noError</c></item>
          <item><c>genErr</c>(SNMPv1)
          </item>
          <item><c>undoFailed | genErr</c>(SNMPv2)</item>
        </list>
      </section>
    </section>

    <section>
      <title>f(set, NewValue [, ExtraArgs])</title>
      <p>This function is called to perform the set in phase two of
        the set-request processing. It is only called if the
        corresponding <c>is_set_ok</c> function is present and returns
        <c>noError</c>.
        </p>
      <p><c>NewValue</c> is guaranteed to be of the correct type,
        length and within ranges, as specified in the MIB.  If the
        object is an enumerated integer or of type BITS, the integer
        value is used.
        </p>
      <p>This function is mandatory.
        </p>

      <section>
        <title>Valid return values</title>
        <list type="bulleted">
          <item><c>noError</c></item>
          <item><c>genErr</c>(SNMPv1)
          </item>
          <item><c>commitFailed | undoFailed | genErr</c>(SNMPv2)</item>
        </list>
      </section>
    </section>
  </section>

  <section>
    <title>Table Instrumentation</title>
    <p>For tables, a <c>f(Operation, ...)</c> function should be
      defined (the function shown is exemplified with <c>f</c>).
      </p>
    <p>The <c>Operation</c> can be <c>new</c>, <c>delete</c>,
      <c>get</c>, <c>next</c>, <c>is_set_ok</c>, <c>undo</c> or
      <c>set</c>.
      </p>
    <p>In case of an error, all instrumentation functions may return
      either an SNMPv1 or an SNMPv2 error code. If it returns an SNMPv2
      code, it is converted into an SNMPv1 code before it is sent to a
      SNMPv1 manager. It is  recommended to use the SNMPv2 error codes
      for all instrumentation functions, as these provide more
      details. See <seealso marker="snmp_app_a">Appendix A</seealso> for a
      description of error code conversions.</p>

    <section>
      <title>f(new [, ExtraArgs])</title>
      <p>The function <c>f(new [, ExtraArgs])</c> is called for each object in an MIB when the
        MIB is loaded into the agent. This makes it possible to perform
        the necessary initialization.
        </p>
      <p>This function is optional. The return value is discarded.</p>
    </section>

    <section>
      <title>f(delete [, ExtraArgs])</title>
      <p>The function <c>f(delete [, ExtraArgs])</c> is called for each object in an MIB when the
        MIB is unloaded from the agent. This makes it possible to
        perform any necessary clean-up.
        </p>
      <p>This function is optional. The return value is discarded.</p>
    </section>

    <section>
      <title>f(get, RowIndex, Cols [, ExtraArgs])</title>
      <p>The function <c>f(get, RowIndex, Cols [, ExtraArgs])</c> is
        called when a get-request refers to a table.
        </p>
      <p>This function is mandatory.
        </p>

      <section>
        <title>Arguments</title>
        <list type="bulleted">
          <item><c>RowIndex</c> is a list of integers which define the
           key values for the row. The <c>RowIndex</c> is the list
           representation (list of integers) which follow the
          <c>Cols</c> integer in the OBJECT IDENTIFIER.
          </item>
          <item><c>Cols</c> is a list of integers which represent the
           column numbers. The <c>Cols</c> are sorted by increasing
           value and are guaranteed to be valid column numbers.</item>
        </list>
      </section>

      <section>
        <title>Valid Return Values</title>
        <list type="bulleted">
          <item>
            <p>A list with as many elements as the <c>Cols</c> list,
              where each element is the value of the corresponding
              column.  Each element can be:</p>
            <list type="bulleted">
              <item><c>{value, Value}</c>. The <c>Value</c> must be of
               correct type, length and within ranges, otherwise
              <c>genErr</c> is returned in the response PDU.  If the
               object is an enumerated integer, the symbolic enum
               value may be used (as an atom). If the object is of
               type BITS, the return value shall be an integer or a
               list of bits that are set.
              </item>
              <item><c>{noValue, noSuchName}</c>(SNMPv1)
              </item>
              <item><c>{noValue, noSuchObject | noSuchInstance}</c>(SNMPv2)
              </item>
            </list>
          </item>
          <item><c>{noValue, Error}</c>. If the row does not exist,
           because all columns have <c>{noValue, Error}</c>), the
           single tuple <c>{noValue, Error}</c> can be returned.
           This is a shorthand for a list with all elements
          <c>{noValue, Error}</c>.
          </item>
          <item><c>genErr</c>.  Used if an error occurred.  Note that
           this should be an internal processing error, e.g. a caused
           by a programing fault somewhere.  If some column does not
           exist, use <c>{noValue, noSuchName}</c> or
          <c>{noValue, noSuchInstance}</c>.
          </item>
        </list>
      </section>
    </section>

    <section>
      <title>f(get_next, RowIndex, Cols [, ExtraArgs])</title>
      <p>The function <c>f(get_next, RowIndex, Cols [, ExtraArgs])</c> is 
        called when a get-next- or a
        get-bulk-request refers to the table.
        </p>
      <p>The <c>RowIndex</c> argument may refer to an existing row or a
        non-existing row, or it may be unspecified. The <c>Cols</c> list
        may refer to inaccessible columns or non-existing columns. For
        each column in the <c>Cols</c> list, the corresponding next
        instance is determined, and the last part of its OBJECT
        IDENTIFIER and its value is returned.
        </p>
      <p>This function is mandatory.
        </p>

      <section>
        <title>Arguments</title>
        <list type="bulleted">
          <item><c>RowIndex</c> is a list of integers (possibly empty)
           that defines the key values for a row. The <c>RowIndex</c>
           is the list representation (list of integers), which follow
           the <c>Cols</c> integer in the OBJECT IDENTIFIER.
          </item>
          <item><c>Cols</c> is a list of integers, greater than or
           equal to zero, which represents the column numbers.</item>
        </list>
      </section>

      <section>
        <title>Valid Return Values</title>
        <list type="bulleted">
          <item>
            <p>A list with as many elements as the <c>Cols</c> list
              Each element can be:</p>
            <list type="bulleted">
              <item><c>{NextOid, NextValue}</c>, where <c>NextOid</c>
               is the lexicographic next OBJECT IDENTIFIER for the
               corresponding column. This should be specified as the
               OBJECT IDENTIFER part following the table entry. This
               means that the first integer is the column number and
               the rest is a specification of the keys.
              <c>NextValue</c> is the value of this element.
              </item>
              <item><c>endOfTable</c> if there are no accessible
               elements after this one.
              </item>
            </list>
          </item>
          <item><c>{genErr, Column}</c> where <c>Column</c> denotes
           the column that caused the error.  <c>Column</c> must be
           one of the columns in the <c>Cols</c> list.  Note that
           this should be an internal processing error, e.g. a caused
           by a programing fault somewhere.  If some column does not
           exist, you must return the next accessible element (or
          <c>endOfTable</c>). 
          </item>
        </list>
      </section>
    </section>

    <section>
      <title>f(is_set_ok, RowIndex, Cols [, ExtraArgs])</title>
      <p>The function <c>f(is_set_ok, RowIndex, Cols [, ExtraArgs])</c>
        is called in phase one of the set-request
        processing so that new values can be checked for
        inconsistencies.
        </p>
      <p>If the function is called, it will be called again with
        <c>undo</c>, or with <c>set</c> as first argument.
        </p>
      <p>This function is optional.
        </p>

      <section>
        <title>Arguments</title>
        <list type="bulleted">
          <item><c>RowIndex</c> is a list of integers which define the
           key values for the row. The <c>RowIndex</c> is the list
           representation (list of integers) which follow the
          <c>Cols</c> integer in the OBJECT IDENTIFIER.
          </item>
          <item><c>Cols</c> is a list of <c>{Column, NewValue}</c>,
           where <c>Column</c> is an integer, and <c>NewValue</c> is
           guaranteed to be of the correct type, length and within
           ranges, as specified in the MIB.  If the object is an
           enumerated integer or of type BITS, the integer value is
           used.  The list is sorted by <c>Column</c> (increasing) and
           each <c>Column</c> is guaranteed to be a valid column
           number.</item>
        </list>
      </section>

      <section>
        <title>Valid Return Values</title>
        <list type="bulleted">
          <item><c>{noError, 0}</c></item>
          <item><c>{Error, Column}</c>, where <c>Error</c> is the same
           as for <c>is_set_ok</c> for variables, and <c>Column</c>
           denotes the faulty column.  <c>Column</c> must be one of the
           columns in the <c>Cols</c> list.</item>
        </list>
      </section>
    </section>

    <section>
      <title>f(undo, RowIndex, Cols [, ExtraArgs])</title>
      <p>If an error occurs, The function
        <c>f(undo, RowIndex, Cols [, ExtraArgs])</c> is called after the
        <c>is_set_ok</c> function. If <c>set</c> is called for this
        object, <c>undo</c> is not called.
        </p>
      <p>This function is optional.
        </p>

      <section>
        <title>Arguments</title>
        <list type="bulleted">
          <item><c>RowIndex</c> is a list of integers which define the
           key values for the row. The <c>RowIndex</c> is the list
           representation (list of integers) which follow the
          <c>Cols</c> integer in the OBJECT IDENTIFIER.
          </item>
          <item><c>Cols</c> is a list of <c>{Column, NewValue}</c>,
           where <c>Column</c> is an integer, and <c>NewValue</c> is
           guaranteed to be of the correct type, length and within
           ranges, as specified in the MIB.  If the object is an
           enumerated integer or of type BITS, the integer value is
           used.  The list is sorted by <c>Column</c> (increasing) and
           each <c>Column</c> is guaranteed to be a valid column
           number.
          </item>
        </list>
      </section>

      <section>
        <title>Valid Return Values</title>
        <list type="bulleted">
          <item><c>{noError, 0}</c></item>
          <item><c>{Error, Column}</c> where <c>Error</c> is the same
           as for <c>undo</c> for variables, and <c>Column</c> denotes
           the faulty column.  <c>Column</c> must be one of the columns
           in the <c>Cols</c> list.
          </item>
        </list>
      </section>
    </section>

    <section>
      <title>f(set, RowIndex, Cols [, ExtraArgs])</title>
      <p>The function <c>f(set, RowIndex, Cols [, ExtraArgs])</c> is
        called to perform the set in phase two of
        the set-request processing. It is only called if the
        corresponding <c>is_set_ok</c> function did not exist, or
        returned <c>{noError, 0}</c>.
        </p>
      <p>This function is mandatory.
        </p>

      <section>
        <title>Arguments</title>
        <list type="bulleted">
          <item><c>RowIndex</c> is a list of integers that define the
           key values for the row. The <c>RowIndex</c> is the list
           representation (list of integers) which follow the
          <c>Cols</c> integer in the OBJECT IDENTIFIER.
          </item>
          <item><c>Cols</c> is a list of <c>{Column, NewValue}</c>,
           where <c>Column</c> is an integer, and <c>NewValue</c> is
           guaranteed to be of the correct type, length and within
           ranges, as specified in the MIB.  If the object is an
           enumerated integer or of type BITS, the integer value is
           used.  The list is sorted by <c>Column</c> (increasing) and
           each <c>Column</c> is guaranteed to be a valid column
           number.
          </item>
        </list>
      </section>

      <section>
        <title>Valid Return Values</title>
        <list type="bulleted">
          <item><c>{noError, 0}</c></item>
          <item><c>{Error, Column}</c> where <c>Error</c> is the same
           as <c>set</c> for variables, and <c>Column</c> denotes
           the faulty column.  <c>Column</c> must be one of the columns
           in the <c>Cols</c> list.</item>
        </list>
      </section>
    </section>
  </section>
</chapter>