Core Mqtt Agent API callbacks

[Litan (dBailos)]

Hi,

I’m working in an IoT project where I’m trying to use the MQTT Agent library for the first time. I’ve got some user app tasks running in a main thread and the MqttAgent running in another thread.

Reading the library functions I can see that the user App can use the thread safe API functions to publish, subscribe,… passing a callback function in the command info arguments.

So if I got it right, these callback functions will be called from the MQTT_Agent thread when the publish/subscribe command is completed, which means that are being executed from the Agent’s thread and will have the Agent’s task priority?

If I want to pass the data from the responses to the main user app (which is in different thread) should I use a queue from the callback to the user app main tasks or is there any better/smarter way do handle this?

My concern is that if I want for instance to execute a blocking piece of code in one of the callbacks that would prevent the MQTT Agent from handle its other commands. So should I treat the callbacks like a sort of ISR running only non blocking code?

Thanks,
Diego B

[aggarg (Gaurav Aggarwal)]

@Litan

which means that are being executed from the Agent’s thread and will have the Agent’s task priority?

Yes, that is correct.

@Litan

If I want to pass the data from the responses to the main user app (which is in different thread) should I use a queue from the callback to the user app main tasks or is there any better/smarter way do handle this?

That is one way. You can also copy to a buffer (which can be dynamically allocated or you can maintain a static pool of buffer) and pass the pointer to that buffer using queue.

@Litan

So should I treat the callbacks like a sort of ISR running only non blocking code?

Correct.

Thanks.

[Litan (dBailos)]

Hi Aggarg,

Thanks very much for clarifying that, it makes a lot of sense now.

Regards,
Diego