[Asterisk-Users] Line death not recognized on TDM400P?

[matt.riddell at sineapps.com]

On 2 Sep 2004 at 15:34, Rich Adamson wrote:

> > > > > > A customer of mine has 3 TDM400P cards in a box running
> > > > > > asterisk. On each card he has four FXO modules.  
> > > > > > 
> > > > > > I have set up the dialplan to dial via group 1 for an
> > > > > > outgoing call.
> > > > > > 
> > > > > > Channels 1-12 are in group 1.
> > > > > > 
> > > > > > If he plugs a telephone cable into socket 2 or 3 etc, but
> > > > > > not 1, when he dials out, it still tries to make the call
> > > > > > via socket 1.
> > > > > > 
> > > > > > Straight away the console says that it has dialed the number
> > > > > > via g1 and that it is connecting sip/bla with zap/1-1 (or
> > > > > > some such)...
> > > > > > 
> > > > > > On my X100P I get a red alarm if the phone cable is not
> > > > > > plugged in. Is there any way to do this with the TDM400P?
> > > > > > 
> > > > > > They would like to be able to unplug lines and use them for
> > > > > > other purposes at times.
> > > > > > 
> > > > > > Make sense?
> > > > > > 
> > > > > > I kinda thought that asterisk would realise that nothing was
> > > > > > connected to the TDM card and try the second socket, the
> > > > > > third etc...
> > > > > 
> > > > > It makes sense, but the code to detect unused rj11's is not in
> > > > > * now.
> > > > > 
> > > > > In fact, you'll find that unplugging and replugging the rj11's
> > > > > will cause * to fail after a while. (At least that was the
> > > > > case about a month ago and there really haven't been any
> > > > > changes to the fxo software for some time.)
> > > > > 
> > > > > There are no alarms or other indicators available that would
> > > > > suggest a port has failed or is unavailable.
> > > > > 
> > > > So basically if there is a problem with line 1 out of 12, no
> > > > calls are going to get through...surely this isn't expected
> > > > behaviour.
> > > > 
> > > > Is there any way to fix this (I'm the New Zealand distributor of
> > > > Digium products and if this is the case, they will be returning
> > > > to their propreity pabx)?
> > > > 
> > > > Should I open a bugnote?
> > > > 
> > > > I spoke to a few people on IRC last night who confirmed that
> > > > this is the case.  It seems crazy.  If I send a fax using a fax
> > > > machine it works, if I dial a number with a modem it works (even
> > > > a $5 modem) so why can't it work in asterisk?  It used to work
> > > > on the X100P's...
> > > > 
> > > > I just want the card not to dial if there is no dialtone...
> > > 
> > > The only response that seems to have any real logic in it is
> > > this...
> > > 
> > > - The TDM card seems to have the hardware facilities (eg,
> > > chipsets) on
> > >   it to do substantially more then what it is today
> > > - There are very few * developers that truly understand the low
> > > level
> > >   pstn interface requirements necessary to code additional
> > >   functionality into the TDM drivers
> > > - Mark is one of the few that can translate pstn interface
> > > requirements
> > >   into real code, and his time seems to be distributed across a
> > >   lot of different areas (not sure when he gets to sleep)
> > > - the overall design of the TDM card relies on the host system
> > > processor 
> > >   to interpret dtmf tones (from digital pcm data flows), etc.
> > >   Therefore adding functionality to the drivers is a delicate
> > >   balance between the function and burning host cycles that might
> > >   be disruptive to other * components
> > > - detecting dial tone before dialing would seem to be trevial (I
> > > can
> > >   easily say that as a non programmer), but someone needs to add
> > >   the code to do it reliably
> > > - the chip set in use already has provisions to detect CO battery,
> > > line
> > >   off hook (presumably from a bridged analog phone), battery
> > >   reversal (supervision), etc, but someone needs to add the code
> > >   to use it.
> > > 
> > > Best guess is time constraints are having "the" biggest impact.
> > > The unwritten default condition for * development is that if a
> > > feature or bug fix is expected/required, there is a high
> > > probability it won't be addressed unless a bug report is
> > > generated, or, one of the developers has the same problem; sort of
> > > the squeaky wheel gets the grease.
> > > 
> > > It's my belief (right, wrong or indifferent) that digium is
> > > missing out on a major revenue stream due to a lack of
> > > understanding "marketing" and the prioritization of certain
> > > development functions (like the above) required to sustain/grow
> > > sales. If digium is going to rely on the sales of these cards to
> > > grow/support *, then something is going to have to change since
> > > user experience/opinion for those products is heading in a
> > > downward direction (not upward).
> > > 
> > > I'm certainly not ragging on anyone, just stating an opinion FWIW.
> > > 
> > > Rich
> > 
> > How's this for service!!!!
> > 
> > 
> > ===============================================
> > 09-02-04 14:00  ZX81  New Bug   
> > 09-02-04 15:04  markster  Bugnote Added: 0013839   
> > 09-02-04 15:04  markster  Assigned To   => markster  
> > 09-02-04 15:04  markster  Resolution  open => fixed  
> > 09-02-04 15:04  markster  Status  new => resolved  
> > ===============================================
> > 
> > And:
> > 
> > ===============================================
> > Update of /usr/cvsroot/asterisk/channels
> > In directory mongoose.digium.com:/tmp/cvs-serv14053/channels
> > 
> > Modified Files:
> >  chan_zap.c 
> > Log Message:
> > Don't use FXO's with no battery (bug #2359)
> > ===============================================
> > 
> > Amazing!  That's great!  My customer will be really happy!
> > 
> > Mark thanks for this.
> 
> That sounds excellent! But, the comments within the code imply the
> change is for T1 interfaces only (maybe not the TDM card?). 
> 
> Just sent a note off to Mark asking for clearification as to which
> interfaces it actually applies to.
> 
> Rich

Just finished chatting with Mark.  The change was at first for 
everything, but it broke someone's Channel Bank so now it is only for 
the TDM400P.

So for me the issue is resolved.  Maybe if other people were having 
problems previously with a channel bank's fxo cards not recognising 
the lack of battery on line they should post about it.

But for the problem I descibed above, asterisk has been patched.

Cheers,

Matt Riddell