When Good Turns Evil: Using Intel SGX to Stealthily Steal Bitcoins
summary
Demonstrates that malware inside Intel SGX can stealthily exfiltrate keys by running an attacker enclave that builds an SGX-resident key extractor and uses Prime+Probe and Flush+Reload to observe secret-dependent memory accesses; reconstructs a 4096-bit RSA private key from a co-located enclave in practical runs and hides activity from OS-level detectors by running entirely in enclave context; also shows a cache-triggered race exploit that defeats DRM-like double-fetch checks and a TSX-based mitigation.
AI-written summary · about
chapters (11)
- 0:00Introduction to SGX attacks
- 3:33How Intel SGX works
- 6:55Side-channel attack concepts
- 11:45Attacking SGX enclaves
- 15:46Extracting private keys
- 20:59Countermeasures and defenses
- 25:02Exploiting DRM systems
- 30:28Detecting double fetches
- 34:05Exploiting double fetches
- 40:01Preventing double-fetch bugs
- 48:18Key takeaways and Q&A
related talks
25mForeshadow: Extracting the Keys to the Intel SGX Kingdom...
25mTelling Your Secrets without Page Faults...
33mHacking in Darkness: Return-oriented Programming against Secure Enclaves
23mThe Guard's Dilemma: Efficient Code-Reuse Attacks Against Intel SGX
48mSGX Secure Enclaves in Practice: Security and Crypto Review
49mBreaking Hardware-Enforced Security With Hypervisors
38mOne Ring -3 To Secure Them All: Computing with Hardware Enclaves
47mHow to Hack a Turned-Off Computer, or Running Unsigned Code in Intel Management Engine
