Memory leak when frequent use of "connection->Connect()"

[EExuke]

• I've searched for previous similar issues and didn't find any solution

Known Issue

• I'm using ATS data type endpoint: the endpoint should look like <prefix>-ats.iot.<region>.amazonaws.com

Describe the bug
In order to keep the program automatic reconnection when dropped，we need to frequent use of "connection->Connect()" to reconnect. But we find that "connection->Connect()" will make memory leak when frequent called.

Platform/OS/Device
ARM/Linux/

SDK version number
v1.8.4

To Reproduce (observed behavior)
1.make sure network doesn't work.
2.put connect and disconnect in a loop which break until connected.
3.then running it can find memory leak.

[jmklix]

Can you provide a minimum code sample that reproduces this memory leak?

[EExuke]

Can you provide a minimum code sample that reproduces this memory leak?

For example, add "while" at "connection->Connect()" in .../sample/basic_pub_sub.cpp, and make sure network can not be connected.

......
/*
* Actually perform the connect dance.
* This will use default ping behavior of 1 hour and 3 second timeouts.
* If you want different behavior, those arguments go into slots 3 & 4.
*/
fprintf(stdout, "Connecting...\n");
int connect_flag = 0;
while(!connect_flag) {
if (!connection->Connect(clientId.c_str(), false, 1000))
{
fprintf(stderr, "MQTT Connection failed with error %s\n", ErrorDebugString(connection->LastError()));
//exit(-1);
sleep(5);
} else {
connect_flag = 1;
}
}
......

[jmklix]

Where in this function are you getting an error. I tried reproducing your error with my network disconnected, but I only ever called connect once because no error was generated in said function.

[EExuke]

Where in this function are you getting an error. I tried reproducing your error with my network disconnected, but I only ever called connect once because no error was generated in said function.

Maybe you shouled change the "exit(-1)" to "//exit(-1)"? To reproduce this need make sure repeated call "connect()" function.

[EExuke]

Now we find :
1.keep the program in a loop which can auto reconnect.
2.it will occupy more "2%" memery of device when every disconnect and auto connect (just control the network)
please help me, thanks.

[jmklix]

I can reproduce a memory leak on Ubuntu with the wifi off and using the following code snippet:

while (true)
    {
        fprintf(stderr, "Connection status: %s\n", connection->Connect(clientId.c_str(), false, 1000)? "true":"false");
        sleep(1);
    }

The following is the valgrind output that shows 71 bytes definitely lost and Aborted (core dumped):

==23500== Memcheck, a memory error detector
==23500== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==23500== Using Valgrind-3.13.0 and LibVEX; rerun with -h for copyright info
==23500== Command: ./basic-pub-sub <tags>
==23500== 
Connecting...
Connection status: true
Connection failed with error aws-c-io: AWS_IO_DNS_INVALID_NAME, Host name was invalid for dns resolution.
Connection status: true
Connection failed with error aws-c-io: AWS_IO_DNS_INVALID_NAME, Host name was invalid for dns resolution.
terminate called after throwing an instance of 'std::future_error'
  what():  std::future_error: Promise already satisfied
==23500== 
==23500== Process terminating with default action of signal 6 (SIGABRT)
==23500==    at 0x5A48FB7: raise (raise.c:51)
==23500==    by 0x5A4A920: abort (abort.c:79)
==23500==    by 0x52D6956: ??? (in /usr/lib/x86_64-linux-gnu/libstdc++.so.6.0.25)
==23500==    by 0x52DCAE5: ??? (in /usr/lib/x86_64-linux-gnu/libstdc++.so.6.0.25)
==23500==    by 0x52DCB20: std::terminate() (in /usr/lib/x86_64-linux-gnu/libstdc++.so.6.0.25)
==23500==    by 0x52DCD53: __cxa_throw (in /usr/lib/x86_64-linux-gnu/libstdc++.so.6.0.25)
==23500==    by 0x52D8930: ??? (in /usr/lib/x86_64-linux-gnu/libstdc++.so.6.0.25)
==23500==    by 0x1D205F: std::__future_base::_State_baseV2::_M_set_result(std::function<std::unique_ptr<std::__future_base::_Result_base, std::__future_base::_Result_base::_Deleter> ()>, bool) (in /home/joseph/Documents/sdk-cpp-workspace/aws-iot-device-sdk-cpp-v2/samples/mqtt/basic_pub_sub/build/basic-pub-sub)
==23500==    by 0x1D460B: std::promise<bool>::set_value(bool&&) (in /home/joseph/Documents/sdk-cpp-workspace/aws-iot-device-sdk-cpp-v2/samples/mqtt/basic_pub_sub/build/basic-pub-sub)
==23500==    by 0x1CC87F: main::{lambda(Aws::Crt::Mqtt::MqttConnection&, int, aws_mqtt_connect_return_code, bool)#1}::operator()(Aws::Crt::Mqtt::MqttConnection&, int, aws_mqtt_connect_return_code, bool) const (in /home/joseph/Documents/sdk-cpp-workspace/aws-iot-device-sdk-cpp-v2/samples/mqtt/basic_pub_sub/build/basic-pub-sub)
==23500==    by 0x1D0887: std::_Function_handler<void (Aws::Crt::Mqtt::MqttConnection&, int, aws_mqtt_connect_return_code, bool), main::{lambda(Aws::Crt::Mqtt::MqttConnection&, int, aws_mqtt_connect_return_code, bool)#1}>::_M_invoke(std::_Any_data const&, Aws::Crt::Mqtt::MqttConnection&, int&&, aws_mqtt_connect_return_code&&, bool&&) (in /home/joseph/Documents/sdk-cpp-workspace/aws-iot-device-sdk-cpp-v2/samples/mqtt/basic_pub_sub/build/basic-pub-sub)
==23500==    by 0x1E76F2: operator() (std_function.h:706)
==23500==    by 0x1E76F2: Aws::Crt::Mqtt::MqttConnection::s_onConnectionCompleted(aws_mqtt_client_connection*, int, aws_mqtt_connect_return_code, bool, void*) (MqttClient.cpp:51)
==23500== 
==23500== HEAP SUMMARY:
==23500==     in use at exit: 767,481 bytes in 13,687 blocks
==23500==   total heap usage: 30,479 allocs, 16,792 frees, 2,677,157 bytes allocated
==23500== 
==23500== LEAK SUMMARY:
==23500==    definitely lost: 71 bytes in 1 blocks
==23500==    indirectly lost: 0 bytes in 0 blocks
==23500==      possibly lost: 800 bytes in 3 blocks
==23500==    still reachable: 766,610 bytes in 13,683 blocks
==23500==                       of which reachable via heuristic:
==23500==                         stdstring          : 69 bytes in 1 blocks
==23500==         suppressed: 0 bytes in 0 blocks
==23500== Rerun with --leak-check=full to see details of leaked memory
==23500== 
==23500== For counts of detected and suppressed errors, rerun with: -v
==23500== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)
Aborted (core dumped)

[EExuke]

And if keep program running and break the Internet, why have memory leak when every aws disconnect ?