Issue when starting/stopping HTTP server

[ClMa_3563786]

Hello,

I'm using WICED SDK 6.1 and I have an issue when I start then stop then start again my HTTP server.

I'm using FreeRTOS and LwIP

I'm using HTTP_Server from WICED (librairies/daemons/HTTP_Server).

When my application goes into AP mode, I start my webserver using the function wiced_http_server_start.

Then I goes off the AP mode, I call the function wiced_http_server_stop.

A few minutes after, I goes back to AP mode, I call the wiced_http_server_start function again, but I have a result error ADDRESS_ALREADY_IN_USE.

I've check that I correctly call the function wiced_http_server_stop.

This return goes from the function err_t netconn_bind(struct netconn *conn, ip_addr_t *addr, u16_t port).

My issue is very similar to this one:

Re: Re: Problem stopping/starting webserver

The code I use:

void Network_start_AP(void)
{
    wiced_result_t    result_function = WICED_ERROR;
    // Power On the AP network
    if(WICED_TRUE != wiced_network_is_up(WICED_AP_INTERFACE)){
        result_function = wiced_network_up( WICED_AP_INTERFACE,
                                            WICED_USE_INTERNAL_DHCP_SERVER,
                                            &ap_ip_settings);
        // If success start the http server
        if(WICED_SUCCESS == result_function){
            result_function = wiced_http_server_start(
                &ap_http_server,
                PORT,
                NB_CLIENT,
                ap_web_pages,
                WICED_AP_INTERFACE,
                STACK_SIZE
                );
            // If start failed Then close http server properly
            if(WICED_SUCCESS != result_function){
                wiced_http_server_stop(&ap_http_server);
                wiced_network_down(WICED_AP_INTERFACE);
            }
        }
        // If Power On failed, then power off the AP network properly
        else{
            wiced_http_server_stop(&ap_http_server);
            wiced_network_down(WICED_AP_INTERFACE);
        }
    }
    else{
        // network already in AP mode
    }
}
void Network_close_AP(void){
    // Power Off the AP network if necessary
    if(WICED_FALSE != wiced_network_is_up(WICED_AP_INTERFACE)){
        WPRINT_APP_INFO( ( "Network_close_AP - stop http server\n" ) );
        wiced_http_server_stop(&ap_http_server);
        wiced_network_down(WICED_AP_INTERFACE);
        while(wiced_network_is_up(WICED_AP_INTERFACE) == WICED_TRUE ){}
    }
    else{
        // AP network already Power Off
    }
}

Unfortunately for me, the issue goes away by magic

If you have any suggestion, I would greatly appreciate

[PriyaM_16]

Kindly check the return value of wiced_http_server_stop() to make sure that the server has stopped before starting the server again.

[ClMa_3563786]

Thanks for this. I've done that first, but reading the code, the return can't be other than WICED_SUCCESS.

wiced_result_t wiced_http_server_stop( wiced_http_server_t* server )
{
    server_event_message_t current_event;
    wiced_assert( "bad arg", ( server != NULL ) );
    current_event.event_type = SERVER_STOP_EVENT;
    current_event.socket     = 0;
    wiced_rtos_push_to_queue( &server->event_queue, &current_event, WICED_NO_WAIT );
    if ( wiced_rtos_is_current_thread( &server->event_thread ) != WICED_SUCCESS )
    {
        /* Wakeup HTTP event thread */
        wiced_rtos_thread_force_awake( &server->event_thread );
    }
    /* Wait for the event to completely get processed for closing the server smoothly */
    wiced_rtos_thread_join( &server->event_thread );
    /* Delete the threads */
    wiced_rtos_delete_thread( &server->event_thread );
    wiced_rtos_delete_worker_thread( &server->connect_thread );
    linked_list_deinit( &server->inactive_stream_list );
    linked_list_deinit( &server->active_stream_list );
    free( server->streams );
    server->streams = NULL;
    return WICED_SUCCESS;
}

I've even checked in the http_server_event_thread_main function that managed the HTTP server.

We correctly receive the STOP message, and exit the thread after cleaning the socket using function:

wiced_result_t wiced_tcp_server_stop( wiced_tcp_server_t* tcp_server )
{
    ...
        netconn_delete( tcp_server->accept_socket[ i ].accept_handler );
    ...
    netconn_disconnect( tcp_server->listen_socket.conn_handler );
    netconn_delete( tcp_server->listen_socket.conn_handler );
}

[PriyaM_16 · Accepted solution]

Hello,

I have tried to replicate the mentioned problem and written a sample code for testing. Kindly check with the attached snip and revert back with the problems you are facing in starting the server again. I tested the snip and I am able to connect to the server after stop -> start cycle.

Console commands added for stopping and starting the server are similar to what you have added in response 1:

Network_close_AP and Network_start_AP

Also kindly attach the UART log for analysis.