We have waited for ~45 seconds to wait for response from server and client is restarted after this time period to avoid situation to stuck application for indefinite time.
In further analysis i can see that some bad logic is done for coordination of retry_count and no_timeout in code of Wait_for_ready_socket function that is actually causing indefinite waiting in poll for case of non-zero value in retry_count.
I think that logic should be corrected. Please let me know if there is any possibility to correct it.