memory leak in prv_getParameters() reported by valgrind

[Kevin0626]

Hi, The Valgrind reported one definitely loss in the prv_getParameters() for version information allocation.

This function has big burden to free the memory for each exit port. We would suggest to use a label in the function and all exit must go to the label. That will easy manage the memory free situations.

==22838== 8 bytes in 2 blocks are definitely lost in loss record 12 of 246
==22838== at 0x4C2DB8F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==22838== by 0xE3D5EB5: lwm2m_malloc (platform.c:28)
==22838== by 0xE3CAB8E: prv_getParameters (registration.c:670)
==22838== by 0xE3CB990: registration_handleRequest (registration.c:1061)
==22838== by 0xE3C935F: handle_request (packet.c:172)
==22838== by 0xE3C9634: lwm2m_handle_packet (packet.c:299)
==22838== by 0xE3C0765: thread_lwm2m (lwm2mserver.c:237)
==22838== by 0x506A9FD: ThreadWrapper (in /media/beihai/WangXin-Fun/ssg-sto-crtl-iot-idrm/plugin-sdk/azure_base/host/lib/azure_iot_gateway_sdk-1.0.6/libgateway.so)
==22838== by 0x584D6B9: start_thread (pthread_create.c:333)

    else if (lwm2m_strncmp((char *)query->data, QUERY_VERSION, QUERY_VERSION_LEN) == 0)
    {
        if (*versionP != NULL) goto error;
        if (query->len == QUERY_VERSION_LEN) goto error;

line 670 --->        *versionP = (char *)lwm2m_malloc(query->len - QUERY_VERSION_LEN + 1);
        if (*versionP != NULL)
        {
            memcpy(*versionP, query->data + QUERY_VERSION_LEN, query->len - QUERY_VERSION_LEN);
            (*versionP)[query->len - QUERY_VERSION_LEN] = 0;
        }
    }

[davidgraeff]

I have reported this as well in #368.
And my solution would actually be to move the codebase to c++ (#369). It is just too hard to get object lifetime right in C and also maintain a correct code, while accepting contributions from external people with different backgrounds and knowledge of the code.

[sbertin-telular]

I believe this has been fixed by pull request #418

[sbernard31]

It should be fixed. Do not hesitate to reopen if this is not.
Maybe should be backport on 1.x branch (#487)