Hello. Some new information: 1) Bug appears rarely, if installed IWD instead of wpasupplicant. Also, now bug appears rarely after reinstalling wpasupplicant.
2) I use gdb to find where Connman fallen into endless loop. In normal work, execution don't enter in function stats_file_history_update(). When bug occurs, execution entered in stats_file_history_update()->summarize()->process_file(), and in process_file() fallen into endless loop ( stats.c:476 ), because function get_next_record(iter) in stats.c:527 always return the same pointer. If I make a force return from process_file(), then Connman will continue normal execution. 3) Another backtrace (in attachment).
connman backtrace 2
Description: Binary data
