[asterisk-dev] Xorcom claims

[Tzafrir Cohen]

On Sat, Aug 04, 2007 at 10:20:45AM +0300, Oron Peled wrote:
> On Saturday, 4 בAugust 2007 09:08, bob_is_me at cox.net wrote:
> > Xorcom claims that:
> > from their press release at:
> >   http://www.xorcom.com/documentation/white_papers/fax_modem 

Yes, unfortunetly this is a bit low on technical details.

> > ...
> > Tzafrir has been pushing a patch that tries to help the USB / interrupt
> > latency issues of zaptel...
> 
> The patch has nothing to do with latency. It tries to enable *different*
> zaptel drivers to synchronize with each other.

The relevant patch is actually http://bugs.digium.com/9645

And it has been discussed before in this list:

http://lists.digium.com/pipermail/asterisk-dev/2007-May/thread.html#27437

However, the ability of Zaptel to send interrupts to a span is not good
enough. Another part of it is the fact that all the Astribank devices on
the system basically share the same clock: they are synchronized through
the xpp driver on the host. Generally one Astribank provides the clock,
and the xpp driver propagates this to the rest of the Astribanks on the
system.

A third part was the ability to synchronize this common clock with the
ticks we get from the Astribank sync master.

That last bit, synchronization to an external timing source, has 
basically already been implemented in the code included as of zaptel 
1.2.18 and 1.4.3 . It requires the patch from http://bugs.digium.com/9645
applied to Zaptel to be enabled (the tarballs from updates.xorcom.com,
as well as my patches branch in the svn, have this patch aplied).

> Have a nice summer everybody,

You too :-)

Viva Zapata! ;-)

-- 
               Tzafrir Cohen       
icq#16849755                    jabber:tzafrir at jabber.org
+972-50-7952406           mailto:tzafrir.cohen at xorcom.com       
http://www.xorcom.com  iax:guest at local.xorcom.com/tzafrir