∇in 6/6/06, Tiago Loureiro <[EMAIL PROTECTED]> wrote: > Hello again, > > I have sent my code in attachment, would be thankful if someone could have > a look at it...
> Yes, I have used valgrind but no memory leaks are detected. This happens > because the application is simply allocating memory but not freeing it when > call is terminated somehow keeping pointers to those piece of memories... > thus memory consumption goes to incredibily high values and never goes down, > even if I terminate all the calls. However, if I do destroy the stack, it is > able to free up all the memory that was being used meaning that there is in > fact no memory leak in the common sense but more of a strange memory > management... why is memory not free'd when call is terminated? I see that > handles are destroyed but still this does not cause memory to be free'd! I had a quick peek on the code and I'm not sure what is the reason of the memory consumption. Please note that the SIP stack keeps a copy of each BYE request and response for 30 seconds, just in case there is retransmission. The copy of request could be freed, but in expense of copying more data when transaction object is created. I would guess there might be some fragmentation issues. Using my test setup the memory consumption seems to stabilize at 13 Meg and then starts slowly decresing. The call rate is roughly 30 calls/second, there was 850 transactions in memory. 15 kB per call is a bit on the high side. -- Pekka.Pessi mail at nokia.com _______________________________________________ Sofia-sip-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/sofia-sip-devel
