Broadcom Chipset Bug in Android, iOS Smartphones Allows Remote Attack

BLACK HAT – Las Vegas - Android and iOS smartphones loaded with a Broadcom Wi-Fi chipset offer attackers a common means to launch a remote exploit that could affect millions of users, according to a presentation here today at Black Hat by security researcher Nitay Artenstein of Exodus Intelligence.

The discovery came about when Artenstein was looking for ways to launch a remote exploit from Android and iOS smartphones, but he knew it would be tough given the way the devices have been hardened with Address Space Layout Randomization (ASLR) and Data Execution Prevention (DEP).

"It's hard to get past ASLR and DEP, so I started looking around the neighborhood to see what would work," said Artenstein, who gave the Black Hat presentation Broadpwn: Remotely Compromising Android and IOS via a Bug in Broadcom's Wi-Fi Chipsets.

He looked at the application processor and toyed with the idea of looking for vulnerabilities in the baseband processor. But he noted that the iPhone, Samsung Galaxy and Note, Google Nexus, HTC, and other smartphones used different chipsets in the devices. However, when he turned his attention to exploring Wi-Fi chipsets, he found that Broadcom was used across the board.

"It's an interesting situation for attackers because they can write an exploit and repeat their work," Artenstein said.

He added that the Wi-Fi Broadcom chipsets have no ASLR or DEP to contend with.

A bug he found in the chipsets had the three necessary ingredients to launch a remote attack.

One is that the vulnerability did not require human interaction to launch an exploit. In this particular case, the smartphone would search for WiFi access points and when it found one, it would automatically connect, Artenstein explained.

The second characteristic is the bug did not require complex assumptions because a wrong assumption could reveal the exploit. "We wanted to find a bug that had static, consistent memory, if possible," Artenstein recalled.

And the third characteristic that's needed for a remote exploit is that its code could be cleaned up after the payload is installed to reduce the chance of it crashing or failing.

In this particular case, the security researcher searched for a location in the chipset where he could write large quantities of data for the payload, and he found that in the packet ring buffer.

With all the elements in place, Artenstein created an exploit that had the ability to be remotely launched without user interaction and could self-propagate, like a worm. Broadcom was informed of his discovery and patched the problem last month.

Related Content: