Re: Re: Termination of early Dialog

Actually I'm talking about the From tag being different between request and
response, which is prohibited.  Using SIPP I simulate the following
behavior:

INVITE - From tag: 11c1f1d0, To tag: <empty>
180 - From tag:  1000, To tag: 1
200 - From tag:  2000, To tag: 1

The 180 and 200 should have been rejected by the stack.  Instead they are
delivered to the SipListener, where I printed the following debug
information:

27059 [Thread-15] DEBUG - Received INVITE response: 180
27059 [Thread-15] DEBUG - Early dialog ID:
[email protected]:1000:1
27075 [Thread-15] DEBUG - Received INVITE response: 200
27075 [Thread-15] DEBUG - 200 OK dialog ID:
[email protected]:2000:1
27435 [EventScannerThread] DEBUG - Client transaction terminated:
z9hG4bK357671573a025ef25f0d7907e2a077a7

The consequence of this erroneous scenario is that one of the two dialogs
remains active even after the session has terminated, and memory
accumulates.

-Hai



On 3/15/07, M. (Neo-Ranga) Ranganathan <[email protected]> wrote:
>
> The stack is assuming that the response with different To tags is as a
> result of upstream forking. When such a situation occurs, the
> application is supposed to delete dialogs that it does not want to
> use.
>
> On 3/15/07, Hai Phan <[email protected]> wrote:
> > Oops, never mind.  I found that my UAS script is forcing a different
> From
> > tag in the 180.  That's why the dialog IDs are different.  The stack
> should
> > have rejected this 180 response then I suppose.  Could be a defect?
> >
> > -Hai
> >
> >
> >
> >
> > On 3/15/07, Hai Phan <[email protected]> wrote:
> > > Hello, I have a specific problem.
> > >
> > > When my UAC receives the 180 response with a To tag, it creates an
> early
> > dialog.  This dialog never terminates and hence my program has memory
> leak.
> > I'm not using reliable provisional responses, and I'm not sure what i
> have
> > to do to signal the lower layer to terminate the dialog.
> > >
> > > This early dialog has a different dialog ID than the dialog
> established by
> > the 200 OK, which gets terminated properly upon BYE.
> > >
> > > Any suggestion is appreciated.
> > >
> > > Thanks,
> > >
> > > -Hai
> > >
> >
> >
> >
> > --
> > Hai H. Phan
> > Veraz Networks
> > Work: (408) 750-9498
> > Cell: (408) 806-6472
> > _______________________________________________
> > nist-sip mailing list
> > [email protected]
> > http://www-x.antd.nist.gov/mailman/listinfo/nist-sip
> >
> >
>
>
> --
> M. Ranganathan
> "There are two ways to write error-free programs; only the third one
> works." --  Alan Perlis
>



-- 
Hai H. Phan
Veraz Networks
Work: (408) 750-9498
Cell: (408) 806-6472

_______________________________________________
nist-sip mailing list
[email protected]
http://www-x.antd.nist.gov/mailman/listinfo/nist-sip