[asterisk-dev] IAX Hardphone (again)

[Bill Shaw]

On 12/5/2011 10:10 AM, Kurt Lidl wrote:
> On 12/5/2011 10:07 AM, Kurt Lidl wrote:
>> On 12/4/2011 11:04 AM, Andreas Sikkema wrote:
>>> On 12/4/11 4:31 PM, Bill Shaw wrote:
>>>> Yes,  I agree that is what it is behaving like,  but there is no
>>>> firewall.  All the devices are on the same segment,  and call setup
>>>> works in the other direction.  I can setup the legs and transfer voice
>>>> packets bidirectionally just fine if the call originates from the 
>>>> other
>>>> end.  It looks to me like my 'new' is being dropped by Asterisk.  I'm
>>>> pretty sure I'm doing something wrong,  I just don't have the info I
>>>> need to figure out what I'm doing wrong.
>>> I'd build my own Asterisk and add debugging all over chan_iax.c, that's
>>> probably the fastest way to find out what's going on.
>>>
>> Are the protocol errors going up in the output of 'netstat -s'?  I've 
>> seen cases where
>> the generation of packets at a low level wasn't putting proper 
>> checksums on
>> the packets, so they all got received at the receiving end of the 
>> connection.
> Wow.  Not nearly enough coffee yet.
>
> Make that "so they all got DROPPED at the receiving end of the 
> connection".
>
>>
>> (IE -- the error is at the UDP level, long before Asterisk sees the 
>> packets)
>>
>> -Kurt
>>
>>

I've compiled in some ast_verb() statements.  The 'new' packet is not 
making it to socket_process(),  so it's not an IAX problem.  It's 
getting dropped somewhere before that.

Bill