March 17, 2015 By Jaikumar Vijayan 2 min read

Security researchers at Google’s Project Zero vulnerability research group have found a way to reliably exploit a potentially serious hardware bug that exists in many modern dynamic random-access memory (DRAM) devices.

Tests conducted on DRAM devices in 29×86 laptop computers from multiple vendors suggest the problem is widespread and that vendors have known about it for a while, the researchers noted in a blog post this week. However, it may be that the vendors have so far treated the hardware bug as just a reliability problem and not as a security issue that can be exploited to corrupt a system.

The problem, dubbed “rowhammer” by the Google researchers, stems from the ever-shrinking nature of DRAM technologies. As DRAM cells have gotten smaller and closer together, it has become harder for manufacturers to prevent cells from interacting with each other electrically and from disturbing neighboring cells, the researchers said.

Experimental Study

In an experimental study on DRAM disturbance errors last year, researchers at Carnegie Mellon University and Intel showed how it is possible to corrupt data in DRAM chips by repeatedly activating and deactivating data in a single row. By simply reading from the same memory address on a DRAM chip repeatedly, it is possible to corrupt data in neighboring rows. It is this process that Google’s researchers describe as “row hammering” in their report.

The researchers investigated a total of 129 DRAM modules from different vendors and discovered disturbance errors in 110 of them. Interestingly, all DRAM chips manufactured between 2012 and 2013 were susceptible to the issue, suggesting the problem is especially prevalent in modern chipsets.

Modern manufacturing processes have allowed chipmakers to reduce memory costs by cramming more DRAM cells into the same amount of real estate. However, in the process, they have also introduced memory reliability issues.

Working Exploits

What Google’s Project Zero researchers have done is show how to exploit the hardware bug. They built two working exploits that let attackers corrupt a DRAM device and escalate their privileges on a compromised system. One of the exploits runs as a Native Client (NaCl) program that attackers can use to escalate privileges and escape the NaCl sandboxing system, the Google researchers said.

The second is a kernel privilege escalation exploit that basically uses row hammering to induce DRAM errors that would let an attacker gain read-write access to all physical memory. According to the researchers, this exploit is harder to mitigate than the one that involves the NaCl sandbox escape.

In order to increase the speed and their chances of corrupting memory cells in DRAM, the researchers showed how an approach they described as “double-sided hammering” can be used to cause memory bits to flip more quickly.

Addressing the Hardware Bug Issue

Chipmakers that are aware of the issue have tried to address it by improving the isolation between cells and by looking for disturbance errors during the post-production testing phase. There are indications that some DRAM manufacturers have begun to implement BIOS updates and other mitigations for the issue, the Google researchers said. At least one DRAM vendor has indicated publicly that it implements rowhammer mitigations within its devices, and some newer laptop models appear to be immune to the problem.

However, without more information from vendors, it is hard to say for sure how many laptop computers are currently vulnerable to the hardware bug. In treating the rowhammer problem as a reliability issue, vendors appear to have overlooked the security implications. It is time for hardware vendors to analyze the issue on their device and provide explanations of impact and mitigation strategies.

More from

A spotlight on Akira ransomware from X-Force Incident Response and Threat Intelligence

7 min read - This article was made possible thanks to contributions from Aaron Gdanski.IBM X-Force Incident Response and Threat Intelligence teams have investigated several Akira ransomware attacks since this threat actor group emerged in March 2023. This blog will share X-Force’s unique perspective on Akira gained while observing the threat actors behind this ransomware, including commands used to deploy the ransomware, active exploitation of CVE-2023-20269 and analysis of the ransomware binary.The Akira ransomware group has gained notoriety in the current cybersecurity landscape, underscored…

New proposed federal data privacy law suggests big changes

3 min read - After years of work and unsuccessful attempts at legislation, a draft of a federal data privacy law was recently released. The United States House Committee on Energy and Commerce released the American Privacy Rights Act on April 7, 2024. Several issues stood in the way of passing legislation in the past, such as whether states could issue tougher rules and if individuals could sue companies for privacy violations. With the American Privacy Rights Act of 2024, the U.S. government established…

AI cybersecurity solutions detect ransomware in under 60 seconds

2 min read - Worried about ransomware? If so, it’s not surprising. According to the World Economic Forum, for large cyber losses (€1 million+), the number of cases in which data is exfiltrated is increasing, doubling from 40% in 2019 to almost 80% in 2022. And more recent activity is tracking even higher.Meanwhile, other dangers are appearing on the horizon. For example, the 2024 IBM X-Force Threat Intelligence Index states that threat group investment is increasingly focused on generative AI attack tools.Criminals have been…

Topic updates

Get email updates and stay ahead of the latest threats to the security landscape, thought leadership and research.
Subscribe today