RE: [tkined] SNMPv3 support

From: Dunford, Martin (
Date: Thu Dec 07 2000 - 20:38:05 MET

Hi Juergen,

  Thanks once again for prompt reply.

  I am happy to supply any patches. However I
extensively reformatted tnmMibParser.c and
tnmMibFrozen.c (to aid readability, insert
my own comments, code to unravel internal
Tnm operation etc.) prior to any debugging.
I did this because I anticipated work might
need to be done on what is beta code.
It is all entirely superfluous to
the mainstream development effort. But I
will work a little and remove this so that
I can send a meaningful patch instead of
something enormous and mostly useless to
you all. Please bear with me a few days
on this one...

  I had one other question re. SNMPv3. Does it do
auto discovery of engineID, engineBoots, engineTime,
contextID etc.? The tnmnSnmp structure holds all that
info and I did notice the 'session' recording it in
tnmSnmpRecv.c. But using a sniffer (the GNU 'Ethereal'
sniffer decodes SNMPv3 pretty well) subsequent packets
don't seem to contain the correct info. I realize all
these are configurable from the command line but sometimes
it's a big advanatge to be able to auto discove them from
the agent.

 I think engineID is best a hex number. I tried using
0x80000... style input. Code does do a

 engineID = Tcl_GetStringFromObj(session->engineID, &engineIDLength);

and then proceeds to insert this as ASCII string into the packet. I
am a relative Tcl API novice so maybe there is easy solution to this
I missed. But thanks for all your help.

   Martin Dunford

-----Original Message-----
From: Juergen Schoenwaelder []
Sent: Thursday, December 07, 2000 12:03 PM
To: Dunford, Martin
Subject: Re: [tkined] SNMPv3 support

>>>>> Dunford, Martin writes:

Martin> - Range restriction Node information was not being recorded
Martin> in the binary 'frozen file' format that is used to record MIB
Martin> information once the MIB file has been parsed. (I fixed this)

Are you willing to share the patch?

Martin> - Size restrictions don't seem to be supported. e.g
Martin> DisplayString(SIZE(0..127)) etc.

Yes. We will fix this by moving to libsmi at some point in time.

Martin> - When initiating an SNMPv3 generator command as follows

Martin> Tnm::snmp -version SNMPv3 -engineID 80000E....

Martin> the engine ID is read in as an ASCII string instead of raw
Martin> hex data (I fixed this)

Well, you gave ASCII as input. I actually decided that it would be
appropriate to accept binary data as argument to engineID. So this is
primarily not a bug. Perhaps my design decision is questionable? What
do folks thing?


Juergen Schoenwaelder      Technical University Braunschweig
<>  Dept. Operating Systems & Computer Networks
Phone: +49 531 391 3289    Bueltenweg 74/75, 38106 Braunschweig, Germany
Fax:   +49 531 391 5936    <URL:>

-- !! This message is brought to you via the `tkined & scotty' mailing list. !! Please do not reply to this message to unsubscribe. To subscribe or !! unsubscribe, send a mail message to <>. !! See for more information.

This archive was generated by hypermail 2b29 : Mon Jan 08 2001 - 15:27:58 MET