Memory Leak

[Stefan Newbold]

Has anyone else experienced a memory leak in the function:
rp_ack_fully_decode_service_request ?

It is in: h_rp_a.c

I've used valgrind to pinpoint it and its coming from the following calloc
call:

rp1_property = calloc(1, sizeof (BACNET_PROPERTY_REFERENCE));

I set a trap at the end of the function to call "free" if rp1_property is not
NULL.

Valgrind then complains about another leak at for the following line:

value = calloc(1, sizeof (BACNET_APPLICATION_DATA_VALUE));

Using the same trapping approach at the end of this function doesn't work for
that leak.

Any others experienced the same problem?

Thanks.

[vogel]

My situation is just the same as you! And now i am nearly sure that the function rp_ack_fully_decode_service_request is bad for the memory, because i wrote a small code-block to test, and the result is not so good!(after about half an hour my arm9 meet the OOM killer.)
Now i rewrite the function rp_ack_print_data() for my ReadProperty service, and everything goes well.

[Steve Karg]

I'd be happy to patch that handler code if you are interested in sharing your fixes.

[vogel]

Hello Steve,

I just simply rewrite the function rp_ack_print_data(BACNET_READ_PROPERTY_DATA * data) from h_rp_a.c, and replace the print_data (in bacrp) using the function rp_ack_kk_data(BACNET_READ_PROPERTY_DATA * data, BACNET_OBJECT_PROPERTY_VALUE * object_value).

thx for your job