Researchers exploited Intel CPU vulnerabilities to sniff their load ports. The CPU sends instructions through a series of ports, each serving a set of instruction types, such as arithmetic, memory loading/storage, vector instructions, and the sniffing of load ports means that all memory access to the system can be tracked.
The researchers exploited the Micro-Architecture Data Sampling (MDS) vulnerability, which was earlier exposed.
Because all loads have to pass through the load ports, the researchers say, what the load performs can theoretically be observed. This study will help to gain an in-depth understanding of how Intel CPU microarchitecture works.