Re: Extending spin lock concept to a table column

Thank you very much for your answer.

>> More seriously, if data is NULL, then you can't access data->data!

Of course, stupid mistake!

Last question: when adding a new row , the spinlock column will be null! How
can I initialize the spinlock column when a row is created ?

Something like:

if(data != NULL)
        data->data.voidp ++;
else
        // How can I set the column value to 1 ??????????????????????????

Robalo!




On 10/30/07, Dave Shield <[email protected]> wrote:
>
> On 30/10/2007, andre robalo <[email protected]> wrote:
> > In all examples i saw, the use of a spin lock object is only by scalar
> > objects.
> > I have a situation where I need to use this concept, but applied to a
> table.
>
> A spin lock is typically *implemented* as a scalar object,
> but it's perfectly common to use that (scalar) spin lock to protect a
> table.
>
> Just retrieve the scalar spinlock value at the same time as the table
> contents that you are interested in - and then include the (scalar)
> spinlock assignment as part of the update request:
>
>     $   snmpget localhost   snmpSetSerialNo.0   ifAdminStatus.2
>     snmpSetSerialNo.0 = 12345
>     ifAdminStatus.2= up(1)
>
>     $  snmpset localhost   snmpSetSerialNo.0 i 12345 ifAdminStatus.2 i 2
>     snmpSetSerialNo.0 = 12345
>     ifAdminStatus.2= down(2)
>
>
> You could implement the spinlock as a column object, but there's
> typically no real need.
>
>
> > In this situation, i would have a table column that is a spinlock
> > (testAndIncr object).
> >
> > My first question is : Is this possible ?
>
> Yes - it is possible.
> You probably can't use the spinlock helper, which is designed
> to implement a scalar object.   But there's nothing special about
> the code.  You could perfectly easily implement this as an
> integer column object.
>    The only special thing about a TestAndIncr object, is the check
> that the assigned value matches the current one, and incrementing
> this value at the end of a successful SET request.
>
>
> > Am I doing anything wrong in my code?
>
> In general, this looks mostly OK.
> Two things spring out though:
>
>
>
> >         case MODE_SET_ACTION:
> >                                         data =
> > netsnmp_table_data_set_find_column(rowData,COLUMN_SPINLOCK);
> >                                         if(data != NULL)
> >                                                 data->data.voidp ++;
>
> Shouldn't you be incrementing the value that voidp points to,
> rather than the pointer itself?
>
> >                                         else
> >                                                 data->data.voidp = 0;
>
> More seriously, if data is NULL, then you can't access data->data!
>
> Dave
>

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

_______________________________________________
Net-snmp-users mailing list
[email protected]
Please see the following page to unsubscribe or change other options:
https://lists.sourceforge.net/lists/listinfo/net-snmp-users