CWE-1278 is about information stored in hardware that an attacker can recover by imaging the integrated circuit, for example with scanning electron microscopy. It concerns chip designers whose products hold keys or proprietary designs.
How it happens
A device viewed at high enough magnification reveals what it stores. Typical steps are removing the chip packaging (decapsulation), then imaging with techniques ranging from high resolution x-ray microscopy to invasive methods that strip layers and image each one with a scanning electron microscope.
The goal is to recover secret keys, unique device identifiers, proprietary code and circuit designs that the attacker could not get any other way. Secrets sit in non-volatile memory or in the circuit netlist. MITRE notes that masked ROM makes extraction easier than one-time programmable (OTP) memory.
What an attacker gains
Confidentiality loss. MITRE notes a common goal is to produce and sell counterfeit versions of the chip.
How to find it
No detection methods are recorded. Assess which secrets live in masked ROM or the netlist and how much an attacker gains from each.
How to fix it
The record lists no mitigations. Following from the description, choose storage for secrets with imaging in mind, such as OTP rather than masked ROM, and avoid placing a secret that is shared across devices in the netlist.
Quick check
Which memory type is easier to extract from?
Masked ROM, according to MITRE, compared with one-time programmable memory.
