Security researchers exploit Intel hyperthreading flaw to break encryption

News by Tom Reeve

Security researchers were able to steal an elliptic curve private key from an Intel processor by exploiting a contention flaw in the chip giant's hyperthreading technology.

Security researchers have discovered a side-channel vulnerability in Intel chips involving hyperthreading technology that would enable an attacker to break encryption.

According to a report on Ars Technica, the researchers used an exploit which they dubbed PortSmash to recover an elliptic curve private key from a TLS server running OpenSSL. They demonstrated it using servers running Intel Skylake and Kaby Lake chips and Ubuntu. The vulnerability works on OpenSSL 1.1.0h or less.

The exploit relies on port contention as two logical processors in a hyperthreading system contend with each other for resources on a shared physical processor.

The researchers are Alejandro Cabrera Aldaya of the Universidad Tecnológica de la Habana in Habana, Cuba and Billy Bob Brumley, Sohaib ul Hassan, Cesar Pereida García and Nicola Tuveri from the Tampere University of Technology, Tampere, Finland.

To exploit the vulnerability, the researchers sent instructions to one logical core while measuring the time it took it to process the instructions. From this they were able to deduce the key being processed in the other logical processor.

Writing on, Brumley said the attacker must be able to run a malicious process on the same core of the processor as the victim process.

The researchers wrote: "Our technique can choose among several configurations to target different configurations to target different ports in order to adapt to different scenarios, thus offering a very fine spatial granularity. Additionally, PortSmash is highly portable and its prerequisites for execution are minimal, i.e., does not require knowledge of memory cache-lines, eviction sets, machine learning techniques, nor reverse engineering techniques."

The researchers notified Intel on 1 October and have said that it appears the vulnerability has been fixed.

The vulnerability has been assigned to CVE-2018-5407.

The researchers are due to publish a paper on the vulnerability in the coming months but meanwhile they have published proof of concept code on Github.

Kevin Bocek, chief cyber security officer at Venafi, said: "PortSmash, and all the other processor vulnerabilities like Meltdown and Spectre, is a reminder that we have to rotate the keys and certificates that serve as machine identities, much more frequently than we do.

"Our machine identities are kept around for years, and it’s crazy to think that they won’t be attacked. This is especially true of cloud and microservices environments, where these kinds of vulnerabilities are most dangerous.

"Security and IT teams know we have to change passwords regularly and why. But we haven’t applied the same logic to machine identities, even though they provide even higher levels of access than most passwords.

"The reality is that most keys and certificates aren’t changed often, and a surprising number are never changed. These are the machine identities that are most at risk from PortSmash."

Find this article useful?

Get more great articles like this in your inbox every lunchtime

Video and interviews