Re: [Linphone-users] Bug in oRTP Version ortp-0.10.1

[Kannaiyan Natesan]

Hi Simon,

Thanks your kind reply.

Receive is in a seperate Thread.
Transmit is in a seperate Thread.

session_destroy is in a seperate thread.

WIth regardless of any condition I just called session_destroy, it must 
release the recv function right?

Regards,
Kannaiyan

----- Original Message ----- 
From: "Simon Morlat" <address@hidden>
To: "Kannaiyan Natesan" <address@hidden>
Cc: "linphone-users" <address@hidden>
Sent: Monday, August 21, 2006 8:06 PM
Subject: Re: [Linphone-users] Bug in oRTP Version ortp-0.10.1

You must wait the last call to rtp_session_*_with_ts to finish before 
calling
rtp_session_destroy().
I guess you are using this RtpSession object from two threads: if your
application does take care of serializing the calls to this RtpSession
object, you'll get problems.
If you don't do that the rtp_session_*_with_ts will lock since the mutex and
the condition will be destroyed by rtp_session_destroy().

Simon

Le lundi 31 juillet 2006 10:39, Kannaiyan Natesan a écrit :
> When restarting the session, it doesn't print anything. Since it is locked
> in the previous lock.
>
> pthread_cond_broadcast(&(&audiosession->recv_wp)->cond);
> rtp_session_destroy(audiosession);
>
> If I use the above pthread_cond_broadcast it works fine. Is that correct 
> to
> use or does it solves the lock?
>
> Kannaiyan
>
> ----- Original Message -----
> From: "Simon Morlat" <address@hidden>
> To: "Kannaiyan Natesan" <address@hidden>
> Cc: "linphone-users" <address@hidden>
> Sent: Monday, July 31, 2006 5:39 PM
> Subject: Re: [Linphone-users] Bug in oRTP Version ortp-0.10.1
>
>
> All those values seems correct to me.
> What are then the values when restarting the new session, the one that
> blocks ?
>
> Simon
>
> Le vendredi 28 juillet 2006 03:02, Kannaiyan Natesan a écrit :
> > ......
> > PT: 20010 ST: 20000
> > PT: 20020 ST: 20010
> > PT: 20030 ST: 20020
> > PT: 20040 ST: 20030
> > PT: 20050 ST: 20040
> > PT: 20060 ST: 20050
> > PT: 20070 ST: 20060
> > PT: 20080 ST: 20070
> > PT: 20090 ST: 20080
> > PT: 20100 ST: 20090
> > callback HookState: 0
> > INFO: Engine: HWHANGUPCALL
> > INFO: Sending SS CALLHANGUP
> > INFO: Sending RTP STOPAUDIORTP
> > INFO: Sending HW HWSTOPAUDIORTP
> > INFO: Sending RTP RTPAUDIOSESSIONEND
> > SIPStack: CALLHANGUP
> > Sending Bye
> > HW: HWSTOPAUDIORTP
> > RTP: STOPAUDIORTP
> > PT: 20110 ST: 20100                   <-- this is the last time which I
> > got
> > Audio RTP Stopped
> > RTP: RTPAUDIOSESSIONEND
> > INFO: RTP Session Ended    (rtp_session_destroy(session) ) <---- invoked
> > before this message.
> >
> >
> > Last time which I got is :  Packet time: 20110  Scheduled Time: 20100
> >
> > Let me know if you need more information.
> >
> > Regards,
> > Kannaiyan
> >
> >
> >
> > ----- Original Message -----
> > From: "Simon Morlat" <address@hidden>
> > To: <address@hidden>; "Kannaiyan Natesan"
> > <address@hidden>
> > Sent: Friday, July 28, 2006 4:24 AM
> > Subject: Re: [Linphone-users] Bug in oRTP Version ortp-0.10.1
> >
> >
> > Hello,
> >
> > I'm sure to understand what your problem is:
> > After destroying a session and creating a new one, this last one blocks
> > into the code you showed: is that correct ?
> >
> > If this is correct, then it would help debugging to know the values of
> > packet_time and sched->time_.
> >
> > Simon
> >
> > Le jeudi 27 juillet 2006 05:10, Kannaiyan Natesan a écrit :
> > > When I end a blocking call, and if I call
> > > rtp_session_destroy(session),
> > >
> > > In the next session, I'm not receiving the incoming packets, and it is
> > > becoz it is waiting in the following function code and my calling
> > > thread blocks,
> > >
> > >
> > >  if (TIME_IS_STRICTLY_NEWER_THAN (packet_time, sched->time_))
> > >                 {
> > >
> > > wait_point_wakeup_at(&session->recv_wp,packet_time,
> > > (session->flags & RTP_SESSION_BLOCKING_MODE)!=0);
> > >                         session_set_clr(&sched->r_sessions,session);
> > >                 }
> > >
> > >
> > > exact function is,
> > >
> > > void wait_point_wakeup_at(WaitPoint *wp, uint32_t t, bool_t dosleep){
> > >         wp->time=t;
> > >         wp->wakeup=TRUE;
> > >         if (dosleep) ortp_cond_wait(&wp->cond,&wp->lock);
> > > }
> > >
> > > I guess once the session is destroyed, the ortp_cond_wait must release
> > > condition.
> > >
> > > Please help.
> > >
> > > Regards,
> > > Kannaiyan
> > >
> > >
> > >
> > >
> > > _______________________________________________
> > > Linphone-users mailing list
> > > address@hidden
> > > http://lists.nongnu.org/mailman/listinfo/linphone-users