Hi Andreas,
I guess this is also related to the improper usage of uac_replace_from()
from branch route, right?
Regards,
Bogdan
Andreas Granig wrote:
Hi again,
This problem actually applies to the ACK for the 487 if the CANCEL as
well, and there it's even worse.
RFC3261 says in 17.1.1.3:
The ACK request constructed by the client transaction MUST contain
values for the Call-ID, From, and Request-URI that are equal to the
values of those header fields in the request passed to the transport
by the client transaction
On the other hand, in 17.2.3 it says regarding server transaction
matching:
The ACK request matches a transaction if the Request-
URI, From tag, Call-ID, CSeq number (not the method), and top Via
header field match those of the INVITE request which created the
transaction, and the To tag of the ACK matches the To tag of the
response sent by the server transaction.
Now if for some reason, the UAC in this case is OpenSER and the UAS is
a server which not only compares the From tag, but the overall
From-Header (like the Cisco PGW in version 9.6 seems to do), then both
of the parties don't behave correctly, and it's a little bit hard to
argue with vendors who has to move now.
The real problem I have now is that for the CANCEL I can at least
manually replace From, but for ACK, OpenSER refuses to do so. In
"auto" mode, uac_replace_from bails out with "ERROR:uac:replace_from:
decline FROM replacing in sequential request in auto mode (has TO
tag)". In "manual" mode, it doesn't complain, but the From still
isn't
replaced, and also an ugly hack like "remove_hf("From");
insert_hf("From: $var(my_from_hdr)\r\n", "To");" has no effect.
Any advice, since 1.3.1 doesn't seem to help also?
Andreas
Bai Shi wrote:
Hi Andreas,
I called without any branches.
Bogdan, I'd appreciate if you can spare sometime to look at this again.
Rgds,
Bai Shi
-----Original Message-----
From: Andreas Granig [mailto:agranig@sipwise.com] Sent: 2008年3月20日
1:34
To: Bogdan-Andrei Iancu
Cc: Bai Shi; Thomas Gelf; users(a)openser.org
Subject: Re: [OpenSER-Users] uac_replace_from and CANCEL
Or, to be more specific, I call it from a route which is called by a
branch_route. Not sure if that matters.
Andreas
Andreas Granig wrote:
> Bogdan,
>
> I just tested with 1.3.1, and there the From of the CANCEL is also
> NOT updated. A probably not so unimportant side-note is that I call
> uac_replace_from() from a branch_route.
>
> Andreas
>
> Bogdan-Andrei Iancu wrote:
>> Hi Bai Shi,
>>
>> I just tested (to be sure no bugs are present) and in 1.3, the
>> CANCELs are automatically updated (the FROM header), as expected.
>>
>> Regards,
>> Bogdan
>>
>> Bai Shi wrote:
>>> Hi, Thomas,
>>> This will do in openser 1.2, however, in openser 1.3 it won't do
>>> any help. The CANCEL will be sent out regardless what you have
>>> done to the request, but follow the information in the original
>>> INVITE. I tested it and suffered a lot ;(
>>> Rgds,
>>> BS
>>>
>>> -----Original Message-----
>>> From: users-bounces(a)lists.openser.org
>>> [mailto:users-bounces@lists.openser.org] On Behalf Of Andreas Granig
>>> Sent: 2008年3月13日 2:54
>>> To: Thomas Gelf
>>> Cc: users(a)openser.org
>>> Subject: Re: [OpenSER-Users] uac_replace_from and CANCEL
>>>
>>> Hi Thomas,
>>>
>>> Yes, I'm doing it that way now (without that "append_branch()"
>>> though), but if you have an uuid-based setup where you fetch the
>>> uuid from the subscriber table during authentication of the INVITE
>>> and then with that uuid fetch dynamic data from the
>>> usr_preferences table to be used for "uac_rewrite_host()", then
>>> you have to do some work-arounds and hacks to get that data for
>>> the CANCEL as well, which could be quite a pain.
>>>
>>> So yes, it would be really great if this could be done
>>> automatically *hint hint* :)
>>>
>>> Andreas
>>>
>>> Thomas Gelf wrote:
>>>
>>>> Applying uac_replace_from() to the CANCEL request should do the job,
>>>> it sadly doesn't happen automagically :'-(
>>>>
>>>> Try something like:
>>>>
>>>> if(is_method("CANCEL") && uri =~ "^sip:...")
>>>> {
>>>> if(t_check_trans())
>>>> {
>>>> rewritehostport("...");
>>>> uac_replace_from("...");
>>>> append_branch();
>>>> }
>>>> if(!t_relay())
>>>> {
>>>> sl_reply_error();
>>>> }
>>>> }
>>>>
>>>> Cheers,
>>>> Thomas Gelf
>>>>
>>>> Andreas Granig schrieb:
>>>>> Hi,
>>>>>
>>>>> I hit another problem with a Cisco PGW in combination with
>>>>> CANCEL, and I'm not sure which fault it is. If I do
>>>>> uac_replace_from in the INVITE, the From-Header is altered
>>>>> somehow and sent to the PGW. So if A calls B where B is a PGW,
>>>>> the From from A to OpenSER is for example sip:a@somedomain and
>>>>> from OpenSER to B it's sip:other-a@somedomain.
>>>>>
>>>>> If the call is cancelled, the From-header isn't altered, so the
>>>>> From from A to OpenSER is sip:a@somedomain and from OpenSER to B
>>>>> it's sip:a@somedomain as well.
>>>>>
>>>>> Now the PGW seems to ignore this CANCEL, and I guess it's
>>>>> because of the different From-usernames, since it works if I
>>>>> don't do any uac_replace_from.
>>>>>
>>>>> RFC3261 says in §9.1:
>>>>> The Request-URI, Call-ID, To, the numeric part of CSeq, and From
>>>>> header
>>>>> fields in the CANCEL request MUST be identical to those in the
>>>>> request being cancelled, including tags.
>>>>>
>>>>> So is it correct behavior of the PGW because the From header in
>>>>> the CANCEL is different from the From header in the INVITE?
>>>>> Should OpenSER alter the From in the CANCEL as well? Or should
>>>>> the PGW just check the From tags, but not the From URI?
>>>>>
>>>>> Andreas
>>>>>
>>>> _______________________________________________
>>>> Users mailing list
>>>> Users(a)lists.openser.org
>>>>
http://lists.openser.org/cgi-bin/mailman/listinfo/users
>>> _______________________________________________
>>> Users mailing list
>>> Users(a)lists.openser.org
>>>
http://lists.openser.org/cgi-bin/mailman/listinfo/users
>>>
>>> _______________________________________________
>>> Users mailing list
>>> Users(a)lists.openser.org
>>>
http://lists.openser.org/cgi-bin/mailman/listinfo/users