Latest Rowhammer Feat Empowers Hackers to Avoid Modifications

The Rowhammer feat was first noticed in 2014 as well as was proved to let attackers take management of appliances by focusing on DRAM memory sections.

Rowhammer attacks take benefit of the nearby vicinity of memory sections, triggering them to pour out their charge as well as change the contents of nearby memory cells. The attack involves supplying continuous read-write operations utilizing cautiously shaped memory access shapes to continuously actuate the same memory lines, which can empower strong privilege escalation attacks.

Since the attack technique was revealed, security scientists have found the method has been used in several attacks. The attacks have even been carried out utilizing simple JavaScript, and have been proved to be effective on Linux-based virtual machines, Windows Machines, and Android appliances.

Substantial research into the Rowhammer exploit has empowered producers to apply several mitigations to avoid attacks; nevertheless, this week, latest research has been published proving that even if several modifications are positioned, the Rowhammer exploit can still be utilized and all existing modifications can be avoided.

Earlier attacks were carried out on several rows of memory cells, however the latest technique targets only one row – an attack technique called one-location hammering that keeps one DRAM row continuously open.

As per the researchers, “We substitute obvious and memory-exhausting grooming and spraying methods with a new dependable method known as memory waylaying. Memory waylaying uses system-level optimizations as well as a side channel to entice the operating system into positioning target pages at attacker selected physical sites.”

The team also fruitfully carried out an attack on Intel SGX and were capable to conceal the attack fully from the operating system. Although the attacks require more time to perform utilizing the new technique, they can still be efficient. The scientists say in their checks, an attack will require between 44.4 and 137.8 hours.

That would obviously be very long for attacks on most computer systems, however, the scientists say that there is a danger of attacks on online computer networks which are not turned off, or for attacks on cloud facilities which usually have over 99.9% uptime. The assailants might conduct refusal of facility attacks on cloud environs, but also privilege increase attacks on personal computer systems.