Had another look: it posted the same trace again two hours after reboot without crashing, so it might be related, but it is not the direct cause.