Cool.
What version are you using? Can you do a test with the git tree (
https://www.gitorious.org/sofia-sip/sofia-sip), please?
I am commiting my patches and making a merge request in Gitorious, but the
project is half orphan.

Can I see your patch? Post the diff here, please.


2013/6/18 Anees K A <anees...@mistralsolutions.com>

> Hi,
>
> I have fixed the leak in the stack. How do I submit a patch?
>
> Regards
>
> anees k A
>
>
> On Tue, Jun 18, 2013 at 11:08 AM, Anees K A <anees...@mistralsolutions.com
> > wrote:
>
>> Alexsander,
>>
>> Please find the valgrind log for the following sample application.
>>
>> Operation performed was 1. register a URI 2. Re-register the same URI
>>
>> Regards
>>
>> anees k A
>>
>>
>> On Tue, Jun 18, 2013 at 10:47 AM, Anees K A <
>> anees...@mistralsolutions.com> wrote:
>>
>>> Dear Alexsander,
>>>
>>> Yes, I am re-using the same nua_handle for repeated calls.
>>>
>>> Mean time this is when the problem comes.
>>>
>>>    1. When REGISTER goes out *Contact: <sip:192.168.100.38>;q=0.1*
>>>    2. The 200OK for that will contain *Contact:
>>>    <sip:192.168.100.38>;q=0.3;expires=3600*
>>>    3. When this response comes, *msg_avlist_d()* gets called to parse
>>>    the parameters.
>>>    4. Inside that function, the parameters "*;q=0.3;expires=3600"* are
>>>    parsed and saved
>>>    5. This memory (32bytes) is never getting freed
>>>    6. If next register goes with *q=0.3* , I can see that another
>>>    32byte is getting allocated.
>>>
>>> I will post valgrind results soon.
>>>
>>>
>>> Regards
>>>
>>> anees k A
>>>
>>>
>>> On Mon, Jun 17, 2013 at 8:27 PM, Alexsander Petry <
>>> alexsanderpe...@gmail.com> wrote:
>>>
>>>> Are you using the same nua_handle (op->op_handle) for each REGISTER?
>>>> Try to run with valgrind and post the results.
>>>>
>>>>
>>>>  2013/6/17 Anees K A <anees...@mistralsolutions.com>
>>>>
>>>>>  Hi,
>>>>>
>>>>> I am using SIP REGISTER method to create a conference factory. For
>>>>> that I am using 'q' factor parameter in REGISTER to let the server know of
>>>>> the current load.
>>>>>
>>>>> When q value changes, I am calling re-registration with new q value. I
>>>>> am seeing that this second call is causing the application to allocate 
>>>>> more
>>>>> memory rather than using what is already allocated. If this operation is
>>>>> let to run for 1 day, the memory usage can be seen as increasing.
>>>>>
>>>>> The register call I use is as follows
>>>>>
>>>>> nua_register(op->op_handle,
>>>>>                     NUTAG_M_FEATURES('q=1.0'),
>>>>>                     TAG_NULL());
>>>>>
>>>>> Has anyone point me what I am doing wrong ?
>>>>>
>>>>> Regards
>>>>>
>>>>> anees k A
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> This SF.net email is sponsored by Windows:
>>>>>
>>>>> Build for Windows Store.
>>>>>
>>>>> http://p.sf.net/sfu/windows-dev2dev
>>>>> _______________________________________________
>>>>> Sofia-sip-devel mailing list
>>>>> Sofia-sip-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/sofia-sip-devel
>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> Alexsander Petry
>>>>
>>>
>>>
>>
>


-- 
Alexsander Petry
------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Sofia-sip-devel mailing list
Sofia-sip-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sofia-sip-devel

Reply via email to