Discussion:
[Sofia-sip-devel] call-ID troubles
Francesco Lamonica
2016-05-14 16:44:49 UTC
Permalink
Hi all, i know this might be a shot in the dark however...
i am experiencing a strange issue, randomically on a UAC that uses
sofia-sip (latest stable from repository) when putting the remote on-hold.
the scenario is the following

sofia-uac -> invites
100
180
200
sofia-uac -> on hold (changing the sdp manually and resending the invite)
sofia-uac -> not on hold

this scenario 1 out of 20/30 calls does a very strange thing
the SDP that is sent out has the wrong callid so the remote party responds
correctly with a series of 481

the callid is mangled this way:
the correct one is like ***@172.16.17.15
while the one sent is ***@17217.117.15
as if the second octet is overwritten by following parts of the address

has anyone experienced something like that? or can give me some hints?
thanks a lot
Michael Jerris
2016-05-14 22:02:09 UTC
Permalink
Callid in sdp?
Post by Francesco Lamonica
Hi all, i know this might be a shot in the dark however...
i am experiencing a strange issue, randomically on a UAC that uses
sofia-sip (latest stable from repository) when putting the remote on-hold.
the scenario is the following
sofia-uac -> invites
100
180
200
sofia-uac -> on hold (changing the sdp manually and resending the invite)
sofia-uac -> not on hold
this scenario 1 out of 20/30 calls does a very strange thing
the SDP that is sent out has the wrong callid so the remote party responds
correctly with a series of 481
as if the second octet is overwritten by following parts of the address
has anyone experienced something like that? or can give me some hints?
thanks a lot
Francesco Lamonica
2016-05-25 15:16:07 UTC
Permalink
Hello Michael,
sorry, obviously i meant SIP :)
Post by Michael Jerris
Callid in sdp?
Post by Francesco Lamonica
Hi all, i know this might be a shot in the dark however...
i am experiencing a strange issue, randomically on a UAC that uses
sofia-sip (latest stable from repository) when putting the remote on-hold.
the scenario is the following
sofia-uac -> invites
100
180
200
sofia-uac -> on hold (changing the sdp manually and resending the invite)
sofia-uac -> not on hold
this scenario 1 out of 20/30 calls does a very strange thing
the SDP that is sent out has the wrong callid so the remote party
responds correctly with a series of 481
as if the second octet is overwritten by following parts of the address
has anyone experienced something like that? or can give me some hints?
thanks a lot
Michael Jerris
2016-05-25 15:26:25 UTC
Permalink
I would triple check the strings you are putting into the tags for this. I have never seen something like this unless it was actually me somehow buffer overflowing into it.
Post by Francesco Lamonica
Hello Michael,
sorry, obviously i meant SIP :)
Callid in sdp?
Hi all, i know this might be a shot in the dark however...
i am experiencing a strange issue, randomically on a UAC that uses sofia-sip (latest stable from repository) when putting the remote on-hold.
the scenario is the following
sofia-uac -> invites
100
180
200
sofia-uac -> on hold (changing the sdp manually and resending the invite)
sofia-uac -> not on hold
this scenario 1 out of 20/30 calls does a very strange thing
the SDP that is sent out has the wrong callid so the remote party responds correctly with a series of 481
as if the second octet is overwritten by following parts of the address
has anyone experienced something like that? or can give me some hints?
thanks a lot
Francesco Lamonica
2016-05-25 17:31:34 UTC
Permalink
Hello Michael,
you mean the "tag" in the from header?
or the sofia tags?

here is the code used to send the re-invite

nua_invite(nh,

TAG_IF(!proxyAddress.isEmpty(),
NUTAG_PROXY(proxyAddress.toLatin1().constData())),

NUTAG_AUTOANSWER(0),

TAG_IF(!options.isDataCall(),
SOATAG_USER_SDP_STR(sdp.toLatin1().constData())),

SIPTAG_PRIORITY_STR(m_mapPriorityStrings[prio].toLatin1().constData()),

TAG_END());


i am not manipulating the callid at all
Post by Michael Jerris
I would triple check the strings you are putting into the tags for this.
I have never seen something like this unless it was actually me somehow
buffer overflowing into it.
Hello Michael,
sorry, obviously i meant SIP :)
Post by Michael Jerris
Callid in sdp?
Post by Francesco Lamonica
Hi all, i know this might be a shot in the dark however...
i am experiencing a strange issue, randomically on a UAC that uses
sofia-sip (latest stable from repository) when putting the remote on-hold.
the scenario is the following
sofia-uac -> invites
100
180
200
sofia-uac -> on hold (changing the sdp manually and resending the invite)
sofia-uac -> not on hold
this scenario 1 out of 20/30 calls does a very strange thing
the SDP that is sent out has the wrong callid so the remote party
responds correctly with a series of 481
as if the second octet is overwritten by following parts of the address
has anyone experienced something like that? or can give me some hints?
thanks a lot
Continue reading on narkive:
Loading...