Abstract

Covert channels based on the CPU front-end typically utilize time or power differences caused by contention. This requires that the sender and receiver are located on the same and quiet physical core. We propose one channel exploiting the macro-op fusion in the front-end, called MFCC, which has cross-core and robust properties. Macro-op fusion is one of the optimization strategies in x86 microarchitecture, which aims to fuse two macro-ops and decode them into a single micro-op. We reverse the constraints of macro-op fusion and find that decoders stop decoding after two macro-op fusions in one cycle. Thus time differences appear in two identical loops with different virtual addresses, represent-ing one and zero, respectively. We build two types of MFCC: the sender and receiver running in the same thread but operating at different privilege levels, and the two running in two different processes. The accuracy within a single thread is almost independent of the CPU load, while the accuracy of interprocess transmission maintains more than 0.8 even when the CPU load is 100%. Finally, we propose three possible protection strategies: eliminating the macro-op fusion mechanism, adding noise to the hardware counters, and making the events delivered at a defined point in time. This paper demonstrates that even minor optimizations in the front-end can lead to covert transmission.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.