It was found that the host private key material could possibly leak to the privilege-separated child processes via re-allocated memory. An attacker able to compromise the privilege-separated process could therefore obtain the leaked key information.