In the TEE.fail research paper, the authors demonstrate a method by which an Intel TDX platform’s PCK could potentially be extracted.
I have the following questions:
1. In the event of a successful PCK extraction, what recovery or re-provisioning mechanisms are officially supported by Intel?
2. For each supported mechanism, what are the exact operational steps, and does any of them require a full TCB Recovery cycle?
3. How long does the re-provisioning or TCB Recovery process typically take, and when will attestation collateral signed with the compromised PCK (or derived keys) be rejected?
4. Is it possible to use these procedures combined with a physical audit and physical sealing (by lock and key) of the platform to ensure the CPU’s PCK is not compromised prior to the sealing?
For context: I am evaluating these topics internally because, given the implications of the TEE.Fail attack, we cannot assume long-term trust in CPUs prior to deployment. The attack can be carried out while a system is offline, or potentially even months or years before installation (e.g., secondhand CPUs), with stolen PCK material being used later. This makes clarity on recovery and re-provisioning procedures essential.
