[asterisk-bugs] [DAHDI-linux 0013866]: TDM400P FXO when configured for alaw fails to dialout

[Asterisk Bug Tracker]

A NOTE has been added to this issue. 
====================================================================== 
http://bugs.digium.com/view.php?id=13866 
====================================================================== 
Reported By:                alecdavis
Assigned To:                sruffell
====================================================================== 
Project:                    DAHDI-linux
Issue ID:                   13866
Category:                   Core/Channels
Reproducibility:            always
Severity:                   block
Priority:                   normal
Status:                     acknowledged
====================================================================== 
Date Submitted:             2008-11-09 19:44 CST
Last Modified:              2009-01-15 11:54 CST
====================================================================== 
Summary:                    TDM400P FXO when configured for alaw fails to
dialout
Description: 
In /etc/dahdi/system.conf selecting alaw=35 causes the FXO to fail
dialout.

All that is heard is constant ~1000Hz square wave, this estimated.

Using mulaw=35 the line works fine.

This is not new, also happens on our production box which is Asterisk
1.6.0-rc6 with DAHDI-2.0.0-rc2
====================================================================== 

---------------------------------------------------------------------- 
 (0097934) sruffell (administrator) - 2009-01-15 11:54
 http://bugs.digium.com/view.php?id=13866#c97934 
---------------------------------------------------------------------- 
As written, there is not currently a way at runtime for the wctdm24xxp
driver to be notified when the DAHDI_SETLAW ioctl is called on the channel.
 That ioctl is used to tell the dahdi how to control conversion to and from
signed linear.  The board driver and the framer doesn't actually perform
any companding and therefore doesn't care about that information.

In the case of the analog cards, they do actually need that information as
part of digital<->analog conversion.  So the board drivers would need some
hook into the ioctl in order to have the option of changing the companding.
 

But I agree that it would be helpful that even if the drivers are not
changed to make the switch from ulaw to alaw in response to the ioctl, then
maybe there is a way to report that not only do they have a default
companding method (dahdi_span.deflaw) but that it is locked and can not be
changed so that you do not get into a situation where your converting
samples in ulaw to signed linear as if they were alaw. 

Issue History 
Date Modified    Username       Field                    Change               
====================================================================== 
2009-01-15 11:54 sruffell       Note Added: 0097934                          
======================================================================