Presumably execution is a feature that all modern CPUs have to do to improve the overall performance of the system. But the Spectre and Meltdown vulnerabilities exposed earlier in the years have been a wake-up call for the chip design industry. Today, a team of academics from universities around the world, along with Bitdefender’s vulnerability researchers, revealed a new security vulnerability in Intel processors, “Load Value Injection” (LVI).
Despite being considered a theoretical threat, Intel has released a firmware patch to mitigate the impact of an attack on the current CPU and will deploy fixes in next-generation hardware designs.
To understand what LVI companies are, first know the Spectre and Meltdown vulnerabilities — especially the latter!
Meltdown vulnerability released in January 2018, running an attacker to read data from the cache while running ‘presumptive execution’ code operations on the CPU.
A unique feature of modern CPU design that allows the CPU to pre-prepare the data to help increase processor speed and performance.
If you guess that the execution missed, it is discarded. The strangething about Meltdown and Spectre vulnerabilities, however, is the ability to target data between ‘transients’.
Meltdown and Spectre attacks, a bombshell in 2018, show significant flaws in modern CPU design.
Based on the original attack scheme, scholars from around the world have expanded their research to a similar set of “transient attacks”.
So over the past two years, we’ve seen a series of CPU-based exploits, including That’s The Time, Zombieload, RIDL, Fallout, and LazyFP.
In theory, a newly exposed LVI attack is the Meltdown vulnerability.
Although the original Meltdown vulnerability allows an attacker to read application data from the CPU cache during a transient transition, LVI allows an attacker to inject code into the CPU and execute it as a ‘temporary’ operation, giving the attacker more control.
Tests conducted by the two research groups showed that LVI attacks were independent of each other, successfully demonstrating how widespread the attack was.
For example, the academic research team focused on leaking data (encryption keys) from the Intel SGX security zone, while Bitdefender focused on demonstrating the impact of the attack on the cloud environment.
It indicates that such attacks are particularly disruptive in multi-tenant environments, such as enterprise workstations or servers in a data center.
In this environment, a less privileged tenant will be able to leak sensitive information from more privileged users, or other hypervisors.
When Meltdown was first exposed in January 2018, Intel initially said it only needed a firmware patch, while the Spectre attack required modifications to the CPU’s silicon design.
But now, researchers say it’s not that simple. Both the academic team and Bitvand ei, both the Meltdown and LVI attacks, will now need to implement hardware-level fixes, as LVI can even bypass Meltdown’s early fixes.