Inter thread communication
ESP-AT middleware is only available with operating system.
For successful resources management, it uses
2 threads within library and allows multiple application threads to post new command to be processed.
Producing and Processing threads are part of library, its implementation is in
Processing thread is in charge of processing each and every received character from ESP device. It can process URC messages which are received from ESP device without any command request. Some of them are:
+IPD indicating new data packet received from remote side on active connection
WIFI CONNECTED indicating ESP has been just connected to access point
and more others
Received messages without any command (URC messages) are sent to application layer using events, where they can be processed and used in further steps
This thread also checks and processes specific received messages based on active command.
As an example, when application tries to make a new connection to remote server, it starts command with
Thread understands that active command is to connect to remote side and will wait for potential
indicating connection status. it will also wait for
indicating command finished status before it unlocks sync_sem to unblock producing thread.
When thread tries to unlock sync_sem, it first checks if it has been locked by producing thread.
Producing thread waits for command messages posted from application thread. When new message has been received, it sends initial AT message over AT port.
It checks if command is valid and if it has corresponding initial AT sequence, such as
It locks sync_sem semaphore and waits for processing thread to unlock it
Processing thread is in charge to read respone from ESP and react accordingly. See previous section for details.
If application uses blocking mode, it unlocks command sem semaphore and returns response
If application uses non-blocking mode, it frees memory for message and sends event with response message
Application thread is considered any thread which calls API functions and therefore writes new messages to producing message queue, later processed by producing thread.
A new message memory is allocated in this thread and type of command is assigned to it, together with required input data for command. It also sets blocking or non-blocking mode, how command shall be executed.
When application tries to execute command in blocking mode, it creates new sync semaphore sem, locks it, writes message to producing queue and waits for sem to get unlocked. This effectively puts thread to blocked state by operating system and removes it from scheduler until semaphore is unlocked again. Semaphore sem gets unlocked in producing thread when response has been received for specific command.
sem semaphore is unlocked in producing thread after sync_sem is unlocked in processing thread
Every command message uses its own sem semaphore to sync multiple application threads at the same time.
If message is to be executed in non-blocking mode, sem is not created as there is no need to block application thread. When this is the case, application thread will only write message command to producing queue and return status of writing to application.