CWE-1248 is about security-sensitive hardware modules that contain semiconductor defects. It is a reminder that silicon fails, and that failure in the wrong block is a security problem and not just a reliability one.
How it happens
Semiconductor devices fail for many reasons. Some are manufacturing or packaging defects. Others come from long use or use under extreme conditions. MITRE lists mechanisms such as encapsulation failure, die-attach failure, wire-bond failure, bulk-silicon defects, oxide-layer faults, aluminum-metal faults including electromigration and corrosion, and thermal or electrical stress.
The visible effect is a fault on internal signals or registers. A signal is always 0 or always 1 and does not switch as expected.
MITRE's example is a network-on-chip that has a firewall controlling access to peripherals from every core that can start transactions. A manufacturing defect in that logic can set the filter output permanently to allow, so the firewall waves everything through. Nothing in the design is wrong. The silicon is.
What an attacker gains
The record lists availability through instability and access control. If a fault lands in a security-sensitive module, the security goals of that module can be compromised. The attacker does not need to create the defect. Defective parts can simply ship, and an attacker who learns about the weakness can exploit it.
How to find it
The record lists no detection methods, but the mitigations point to where defects get caught: testing.
- Use fault models such as stuck-at-0 and stuck-at-1 to build post-manufacturing test cases and reach good coverage.
- After packaging, run extensive post-silicon tests on the logic that implements security functions.
How to fix it
MITRE gives two mitigations.
- Testing. Manufacturers keep improving their processes, but some defects can only be found after manufacturing. Post-manufacturing testing of the silicon die is critical. Use stuck-at fault models and run thorough post-silicon testing to confirm the security logic is defect-free.
- Operation. Running hardware outside its specification, for example at extremely high temperature or voltage, speeds up degradation and creates defects. When those faults appear in security-critical modules, the security guarantees fail. Operating the device within specification matters.
Quick check
What is a stuck-at fault?
A defect where a signal is always 0 or always 1 and does not switch as expected. In a security filter, a stuck-at-allow fault would grant access to everything.
Why does operating conditions matter for security?
Extreme temperature or voltage accelerates semiconductor degradation, and the resulting defects can land in security-critical hardware.
