Translation Leak-aside Buffer: Defeating Cache Side-channel Protections
summary
Shows a practical TLB-based channel that bypasses cache defenses by reverse-engineering Intel TLB hashing, building a probe that times L1/L2 translation sets and classifies temporal patterns with an SVM. Prototype TLBleed recovers a 256-bit EdDSA key from one capture in ~17s with 98% success rate, and reconstructs ~92% of RSA bits from a FLUSH+RELOAD-hardened implementation; relies on hyperthread sharing and page-granularity TLB sets.
AI-written summary · about
chapters (15)
- 0:00Introduction and context
- 2:04Side-channel attack basics
- 4:24Analyzing cache traces
- 6:09Intel TSX cache defenses
- 7:33Understanding the TLB
- 9:00Reverse engineering the TLB
- 11:02Hash functions in the TLB
- 11:58Hyper-thread sharing
- 13:51Microarchitectural findings
- 14:36Latency-based monitoring
- 16:12TLB attack implementation
- 17:11SVM classifier application
- 19:12Experimental results
- 20:55Conclusion and reception
- 23:19Q&A session
related talks
15mI Know What You Said: Unveiling Hardware Cache Side-Channels in Local Large
36mStore-to-Leak Forwarding: There and Back Again
28mStrong and Efficient Cache Side-Channel Protection...
01h 01mTutorial: Side-Channel Attacks and Transient Execution Vulnerab...
43mUsing Undocumented CPU Behavior to See Into Kernel Mode and Break KASLR in the Process
28mPrime+Abort: A Timer-Free High-Precision L3 Cache Attack using Intel TSX
41mUsing Microarchitectural Design to Break KASLR and More
20mA Formal Model of Cache Speculation Side-Channels
