I'm looking for answer on such question.
Imagine that CPU bugged on the factory simply like Blue Pill, but for the network's interconnections.
How exactly firewall of OS could protect station against Hardware Trojan Horse intercommunication?
E.g. iptables blocks any new incoming packet.
iptables -A INPUT -p tcp -m state --state NEW -s 0.0.0.0/0 -j DROP
Ethernet packet with
message A switch CPU into the remote controllable mode. Such way that any next
message X with secret prefix would be accepted like a control command on the low level language(Assembler).
Are low level incoming packet goes through the whole of circuit to the CPU for the decision DROP/ACCEPT?
Internet -> Ethernet(PCI) -> North Bridge(Chip) -> CPU
Will CPU read the whole packet, before mark it as
message A would fill the registers of the CPU and switch it into the malicious mode.
HTH in my scenario injected into the CPU like this one.
Edit. I'm trying to clarify. Is there anything that OS can make with malicious control packet(Ethernet) on Bugged CPU?