Apologies for the slow update, I was waiting for the recording to become available and was OOO Th/F last week. Key Takeaways: - Primary use case is to secure workloads that process/handle sensitive biometric data (e.g. fingerprints, face authentication). - SEAM is a poor fit as it doesn't provide mechanisms to restrict access to non-DRAM "memory", e.g. access to the hardware devices that provide biometric data. And there's no line of sight to an AMD equivalent. - pKVM support requires few changes outside of KVM (though the changes to KVM are extensive). Next Steps: - Re-assess in 3-4 weeks after people have had a chance to read through and review the RFC patches. Recording: https://drive.google.com/file/d/1JZ6e8ZgR2gUfB4uBYxsJUxp1KVL5YEA_/view?usp=drive_link&resourcekey=0-MGjMLec-8JEIFC3-vmZeLg
