redundant check in mib.cpp

Faye Ly faye____pedestalnetworks.com
Fri Feb 28 02:39:40 CET 2003


Frank,

Can you please clarify the check in function:

Boolean MibTable::ready(Vbx* pvbs, int sz, MibTableRow* row)
{
...

         // check for changed value
         MibLeaf* l = generator.get_nth(i);
         if ((!l->valid()) &&
             (l->get_value() == pvbs[i])) {
            delete[] required;
            return FALSE;
         }
..
}

Why do we need to check the changed value here for the set?  It seems to
me that the value should be set per SNMP request even if the value is
the same.  This caused some problem with the implementation of the MIB
creation and setting on the following objects specified by RFC 2662:

        adslAtucConfMinSnrMgn  OBJECT-TYPE
             SYNTAX      INTEGER (0..310)
             UNITS       "tenth dB"
             MAX-ACCESS  read-create
             STATUS      current
             DESCRIPTION
                 "Configured Minimum acceptable Signal/Noise Margin.
                 If the noise margin falls below this level, the modem
                 should attempt to increase its power output.  If that
                 is not possible the modem will attempt to
                 re-initialize or shut down."
         ::= { adslLineConfProfileEntry 6 }

Setting (or as one of the vb in the creation request) this object to
zero will cause a "badValue" error.  Similar thing to all the
read-create objects defined in this table.

Same thing with the ATM MIB as defined by RFC 2515 and 2514 where the
transmit and receive traffic descriptor cannot be set to zero as
defined:

AtmTrafficDescrParamIndex ::= TEXTUAL-CONVENTION
        STATUS  current
        DESCRIPTION
            "The value of this object identifies a row in the
            atmTrafficDescrParamTable.  The value 0 signifies
            that no row has been identified."
        SYNTAX  INTEGER (0..2147483647)

Once I commented out the check, both scenarios worked.  Please advice
though as I don't fully know what is the intention of this check.
Thanks.

-faye




More information about the AGENTPP mailing list